* [FFmpeg-devel] [PATCH v7 01/11] avcodec: add enum types for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 02/11] avcodec: add cbs " Thomas Siedel
` (10 subsequent siblings)
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
From: Nuo Mi <nuomi2021@gmail.com>
Add types as nal unit types, slice types and vvc typedefs needed
for parsing vvc high-level syntax.
---
libavcodec/h266.h | 142 ++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 142 insertions(+)
create mode 100644 libavcodec/h266.h
diff --git a/libavcodec/h266.h b/libavcodec/h266.h
new file mode 100644
index 0000000000..697592cf50
--- /dev/null
+++ b/libavcodec/h266.h
@@ -0,0 +1,142 @@
+/*
+ * H.266 / VVC shared code
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVCODEC_H266_H
+#define AVCODEC_H266_H
+
+/**
+ * Table 5 – NAL unit type codes and NAL unit type classes
+ * in T-REC-H.266-202008
+ */
+enum VVCNALUnitType {
+ VVC_TRAIL_NUT = 0,
+ VVC_STSA_NUT = 1,
+ VVC_RADL_NUT = 2,
+ VVC_RASL_NUT = 3,
+ VVC_RSV_VCL_4 = 4,
+ VVC_RSV_VCL_5 = 5,
+ VVC_RSV_VCL_6 = 6,
+ VVC_IDR_W_RADL = 7,
+ VVC_IDR_N_LP = 8,
+ VVC_CRA_NUT = 9,
+ VVC_GDR_NUT = 10,
+ VVC_RSV_IRAP_11 = 11,
+ VVC_OPI_NUT = 12,
+ VVC_DCI_NUT = 13,
+ VVC_VPS_NUT = 14,
+ VVC_SPS_NUT = 15,
+ VVC_PPS_NUT = 16,
+ VVC_PREFIX_APS_NUT = 17,
+ VVC_SUFFIX_APS_NUT = 18,
+ VVC_PH_NUT = 19,
+ VVC_AUD_NUT = 20,
+ VVC_EOS_NUT = 21,
+ VVC_EOB_NUT = 22,
+ VVC_PREFIX_SEI_NUT = 23,
+ VVC_SUFFIX_SEI_NUT = 24,
+ VVC_FD_NUT = 25,
+ VVC_RSV_NVCL_26 = 26,
+ VVC_RSV_NVCL_27 = 27,
+ VVC_UNSPEC_28 = 28,
+ VVC_UNSPEC_29 = 29,
+ VVC_UNSPEC_30 = 30,
+ VVC_UNSPEC_31 = 31,
+};
+
+enum VVCSliceType {
+ VVC_SLICE_TYPE_B = 0,
+ VVC_SLICE_TYPE_P = 1,
+ VVC_SLICE_TYPE_I = 2,
+};
+
+enum {
+ //6.2 we can have 3 sample arrays
+ VVC_MAX_SAMPLE_ARRAYS = 3,
+
+ //7.4.3.3 vps_max_layers_minus1 is u(6)
+ VVC_MAX_LAYERS = 64,
+
+ //7.4.3.3 The value of vps_max_sublayers_minus1 shall be in the range of 0 to 6, inclusive
+ VVC_MAX_SUBLAYERS = 7,
+
+ //7.4.3.3 vps_num_ptls_minus1 is u(8)
+ VVC_MAX_PTLS = 256,
+
+ //7.4.3.3 vps_num_output_layer_sets_minus2 is u(8)
+ VVC_MAX_TOTAL_NUM_OLSS = 257,
+
+ // 7.3.2.3: vps_video_parameter_set_id is u(4).
+ VVC_MAX_VPS_COUNT = 16,
+ // 7.3.2.4: sps_seq_parameter_set_id is u(4)
+ VVC_MAX_SPS_COUNT = 16,
+ // 7.3.2.5: pps_pic_parameter_set_id is u(6)
+ VVC_MAX_PPS_COUNT = 64,
+
+ // 7.4.4.1: ptl_num_sub_profiles is u(8)
+ VVC_MAX_SUB_PROFILES = 256,
+
+ // A.4.2: according to (1577), MaxDpbSize is bounded above by 2 * maxDpbPicBuf(8)
+ VVC_MAX_DPB_SIZE = 16,
+
+ //7.4.3.4 sps_num_ref_pic_lists in range [0, 64]
+ VVC_MAX_REF_PIC_LISTS = 64,
+
+ //7.4.11 num_ref_entries in range [0, MaxDpbSize + 13]
+ VVC_MAX_REF_ENTRIES = VVC_MAX_DPB_SIZE + 13,
+
+ //7.4.3.3 sps_num_points_in_qp_table_minus1[i] in range [0, 36 − sps_qp_table_start_minus26[i]],
+ //sps_qp_table_start_minus26[i] in range [sps_qp_table_start_minus26[i] −26 − QpBdOffset, 36]
+ //for 10 bitsQpBdOffset is 12, so sps_num_points_in_qp_table_minus1[i] in range [0, 74]
+ VVC_MAX_POINTS_IN_QP_TABLE = 75,
+
+ // 7.4.6.1: hrd_cpb_cnt_minus1 is in [0, 31].
+ VVC_MAX_CPB_CNT = 32,
+
+ // A.4.1: the highest level allows a MaxLumaPs of 35 651 584.
+ VVC_MAX_LUMA_PS = 35651584,
+
+ // A.4.1: pic_width_in_luma_samples and pic_height_in_luma_samples are
+ // constrained to be not greater than sqrt(MaxLumaPs * 8). Hence height/
+ // width are bounded above by sqrt(8 * 35651584) = 16888.2 samples.
+ VVC_MAX_WIDTH = 16888,
+ VVC_MAX_HEIGHT = 16888,
+
+ // A.4.1: table A.1 allows at most 440 tiles per au for any level.
+ VVC_MAX_TILES_PER_AU = 440,
+ // A.4.1: table A.1 did not define max tile rows.
+ // in worest a case, we can have 1x440 tiles picture.
+ VVC_MAX_TILE_ROWS = VVC_MAX_TILES_PER_AU,
+ // A.4.1: table A.1 allows at most 20 tile columns for any level.
+ VVC_MAX_TILE_COLUMNS = 20,
+
+ // A.4.1 table A.1 allows at most 600 slice for any level.
+ VVC_MAX_SLICES = 600,
+
+ // 7.4.8: in the worst case (!pps_no_pic_partition_flag and
+ // sps_entropy_coding_sync_enabled_flag are both true), entry points can be
+ // placed at the beginning of every Ctb row in every tile, giving an
+ // upper bound of (num_tile_columns_minus1 + 1) * PicHeightInCtbsY - 1.
+ // Only a stream with very high resolution and perverse parameters could
+ // get near that, though, so set a lower limit here with the maximum
+ // possible value for 8K video (at most 135 32x32 Ctb rows).
+ VVC_MAX_ENTRY_POINTS = VVC_MAX_TILE_COLUMNS * 135,
+};
+
+#endif /* AVCODEC_H266_H */
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 02/11] avcodec: add cbs for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 01/11] avcodec: add enum types " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 03/11] avcodec: add bitstream parser " Thomas Siedel
` (9 subsequent siblings)
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
From: Nuo Mi <nuomi2021@gmail.com>
Add CodedBitstreamContext to parse VPS,SPS,PPS in VVC nal units.
Implement parsing and writing of SPS,PPS,VPS,PH,AUD,SEI and slices.
Add ff_cbs_type_h266 to cbs types tables and AV_CODEC_ID_H266
to cbs codec ids.
Co-authored-by: Thomas Siedel <thomas.ff@spin-digital.com>
---
configure | 2 +
libavcodec/cbs.c | 6 +
libavcodec/cbs_h2645.c | 435 +++-
libavcodec/cbs_h266.h | 793 +++++++
libavcodec/cbs_h266_syntax_template.c | 3116 +++++++++++++++++++++++++
libavcodec/cbs_internal.h | 1 +
libavcodec/cbs_sei.c | 29 +
7 files changed, 4379 insertions(+), 3 deletions(-)
create mode 100644 libavcodec/cbs_h266.h
create mode 100644 libavcodec/cbs_h266_syntax_template.c
diff --git a/configure b/configure
index 8980cec7ee..91964edc8f 100755
--- a/configure
+++ b/configure
@@ -2427,6 +2427,7 @@ CONFIG_EXTRA="
cbs_av1
cbs_h264
cbs_h265
+ cbs_h266
cbs_jpeg
cbs_mpeg2
cbs_vp9
@@ -2709,6 +2710,7 @@ threads_if_any="$THREADS_LIST"
cbs_av1_select="cbs"
cbs_h264_select="cbs"
cbs_h265_select="cbs"
+cbs_h266_select="cbs"
cbs_jpeg_select="cbs"
cbs_mpeg2_select="cbs"
cbs_vp9_select="cbs"
diff --git a/libavcodec/cbs.c b/libavcodec/cbs.c
index 504197e06d..dc9cec443b 100644
--- a/libavcodec/cbs.c
+++ b/libavcodec/cbs.c
@@ -40,6 +40,9 @@ static const CodedBitstreamType *const cbs_type_table[] = {
#if CONFIG_CBS_H265
&ff_cbs_type_h265,
#endif
+#if CONFIG_CBS_H266
+ &ff_cbs_type_h266,
+#endif
#if CONFIG_CBS_JPEG
&ff_cbs_type_jpeg,
#endif
@@ -61,6 +64,9 @@ const enum AVCodecID ff_cbs_all_codec_ids[] = {
#if CONFIG_CBS_H265
AV_CODEC_ID_H265,
#endif
+#if CONFIG_CBS_H266
+ AV_CODEC_ID_H266,
+#endif
#if CONFIG_CBS_JPEG
AV_CODEC_ID_MJPEG,
#endif
diff --git a/libavcodec/cbs_h2645.c b/libavcodec/cbs_h2645.c
index 80e48829af..52639abe3d 100644
--- a/libavcodec/cbs_h2645.c
+++ b/libavcodec/cbs_h2645.c
@@ -24,10 +24,11 @@
#include "cbs_internal.h"
#include "cbs_h264.h"
#include "cbs_h265.h"
+#include "cbs_h266.h"
#include "h264.h"
#include "h2645_parse.h"
#include "hevc.h"
-
+#include "h266.h"
static int cbs_read_ue_golomb(CodedBitstreamContext *ctx, GetBitContext *gbc,
const char *name, const int *subscripts,
@@ -255,6 +256,7 @@ static int cbs_h265_payload_extension_present(GetBitContext *gbc, uint32_t paylo
#define FUNC_NAME1(rw, codec, name) FUNC_NAME2(rw, codec, name)
#define FUNC_H264(name) FUNC_NAME1(READWRITE, h264, name)
#define FUNC_H265(name) FUNC_NAME1(READWRITE, h265, name)
+#define FUNC_H266(name) FUNC_NAME1(READWRITE, h266, name)
#define FUNC_SEI(name) FUNC_NAME1(READWRITE, sei, name)
#define SUBSCRIPTS(subs, ...) (subs > 0 ? ((int[subs + 1]){ subs, __VA_ARGS__ }) : NULL)
@@ -369,6 +371,10 @@ static int cbs_h2645_read_more_rbsp_data(GetBitContext *gbc)
#include "cbs_h265_syntax_template.c"
#undef FUNC
+#define FUNC(name) FUNC_H266(name)
+#include "cbs_h266_syntax_template.c"
+#undef FUNC
+
#undef READ
#undef READWRITE
#undef RWContext
@@ -447,6 +453,10 @@ static int cbs_h2645_read_more_rbsp_data(GetBitContext *gbc)
#include "cbs_h265_syntax_template.c"
#undef FUNC
+#define FUNC(name) FUNC_H266(name)
+#include "cbs_h266_syntax_template.c"
+#undef FUNC
+
#undef WRITE
#undef READWRITE
#undef RWContext
@@ -476,8 +486,9 @@ static int cbs_h2645_fragment_add_nals(CodedBitstreamContext *ctx,
const H2645NAL *nal = &packet->nals[i];
AVBufferRef *ref;
size_t size = nal->size;
+ enum AVCodecID codec_id = ctx->codec->codec_id;
- if (nal->nuh_layer_id > 0)
+ if (codec_id != AV_CODEC_ID_VVC && nal->nuh_layer_id > 0)
continue;
// Remove trailing zeroes.
@@ -513,6 +524,12 @@ static int cbs_h2645_split_fragment(CodedBitstreamContext *ctx,
if (frag->data_size == 0)
return 0;
+ if (codec_id == AV_CODEC_ID_VVC) {
+ //we deactive picture header here to avoid reuse previous au's ph.
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+ h266->priv.ph = NULL;
+ }
+
if (header && frag->data[0] && codec_id == AV_CODEC_ID_H264) {
// AVCC header.
size_t size, start, end;
@@ -640,6 +657,71 @@ static int cbs_h2645_split_fragment(CodedBitstreamContext *ctx,
return err;
}
+ } else if(header && frag->data[0] && codec_id == AV_CODEC_ID_VVC) {
+ // VVCC header.
+ int ptl_present_flag, num_arrays;
+ int b, i, j;
+
+ priv->mp4 = 1;
+
+ bytestream2_init(&gbc, frag->data, frag->data_size);
+
+ b = bytestream2_get_byte(&gbc);
+ priv->nal_length_size = ((b >> 1) & 3) + 1;
+ ptl_present_flag = b & 1;
+
+ if(ptl_present_flag) {
+ int num_sublayers, num_bytes_constraint_info, num_sub_profiles;
+ num_sublayers = (bytestream2_get_be16u(&gbc) >> 4) & 7;
+ bytestream2_skip(&gbc, 1);
+
+ // begin VvcPTLRecord(num_sublayers);
+ num_bytes_constraint_info = bytestream2_get_byte(&gbc) & 0x3f;
+ bytestream2_skip(&gbc, 2 + num_bytes_constraint_info);
+ if(num_sublayers > 1) {
+ int count_present_flags = 0;
+ b = bytestream2_get_byte(&gbc);
+ for(i = num_sublayers - 2; i >= 0; i--) {
+ if((b >> (7 - (num_sublayers - 2 - i))) & 0x01)
+ count_present_flags++;
+ }
+ bytestream2_skip(&gbc, count_present_flags);
+ }
+ num_sub_profiles = bytestream2_get_byte(&gbc);
+ bytestream2_skip(&gbc, num_sub_profiles * 4);
+ // end VvcPTLRecord(num_sublayers);
+
+ bytestream2_skip(&gbc, 3 * 2);
+ }
+
+ num_arrays = bytestream2_get_byte(&gbc);
+ for(j = 0; j < num_arrays; j++) {
+ size_t start, end, size;
+ int nal_unit_type = bytestream2_get_byte(&gbc) & 0x1f;
+ unsigned int num_nalus = 1;
+ if(nal_unit_type != VVC_DCI_NUT && nal_unit_type != VVC_OPI_NUT)
+ num_nalus = bytestream2_get_be16u(&gbc);
+
+ start = bytestream2_tell(&gbc);
+ for(i = 0; i < num_nalus; i++) {
+ size = bytestream2_get_be16(&gbc);
+ bytestream2_skip(&gbc, size);
+ }
+ end = bytestream2_tell(&gbc);
+
+ err = ff_h2645_packet_split(&priv->read_packet,
+ frag->data + start, end - start,
+ ctx->log_ctx, 1, 2, AV_CODEC_ID_VVC, 1, 1);
+ if (err < 0) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "Failed to split "
+ "VVCC array %d (%d NAL units of type %d).\n",
+ i, num_nalus, nal_unit_type);
+ return err;
+ }
+ err = cbs_h2645_fragment_add_nals(ctx, frag, &priv->read_packet);
+ if (err < 0)
+ return err;
+ }
} else {
// Annex B, or later MP4 with already-known parameters.
@@ -685,6 +767,26 @@ cbs_h2645_replace_ps(4, PPS, pps, pic_parameter_set_id)
cbs_h2645_replace_ps(5, VPS, vps, vps_video_parameter_set_id)
cbs_h2645_replace_ps(5, SPS, sps, sps_seq_parameter_set_id)
cbs_h2645_replace_ps(5, PPS, pps, pps_pic_parameter_set_id)
+cbs_h2645_replace_ps(6, VPS, vps, vps_video_parameter_set_id)
+cbs_h2645_replace_ps(6, SPS, sps, sps_seq_parameter_set_id)
+cbs_h2645_replace_ps(6, PPS, pps, pps_pic_parameter_set_id)
+
+static int cbs_h266_replace_ph(CodedBitstreamContext *ctx,
+ CodedBitstreamUnit *unit)
+{
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+ int err;
+
+ h266->priv.ph = NULL;
+ err = ff_cbs_make_unit_refcounted(ctx, unit);
+ if (err < 0)
+ return err;
+ err = av_buffer_replace(&h266->priv.ph_ref, unit->content_ref);
+ if (err < 0)
+ return err;
+ h266->priv.ph = (H266RawPH*)h266->priv.ph_ref->data;
+ return 0;
+}
static int cbs_h264_read_nal_unit(CodedBitstreamContext *ctx,
CodedBitstreamUnit *unit)
@@ -926,6 +1028,130 @@ static int cbs_h265_read_nal_unit(CodedBitstreamContext *ctx,
return 0;
}
+static int cbs_h266_read_nal_unit(CodedBitstreamContext *ctx,
+ CodedBitstreamUnit *unit)
+{
+ GetBitContext gbc;
+ int err;
+
+ err = init_get_bits8(&gbc, unit->data, unit->data_size);
+ if (err < 0)
+ return err;
+
+ err = ff_cbs_alloc_unit_content(ctx, unit);
+ if (err < 0)
+ return err;
+
+ switch (unit->type) {
+ case VVC_VPS_NUT:
+ {
+ H266RawVPS *vps = unit->content;
+
+ err = cbs_h266_read_vps(ctx, &gbc, vps);
+ if (err < 0)
+ return err;
+
+ err = cbs_h266_replace_vps(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+ case VVC_SPS_NUT:
+ {
+ H266RawSPS *sps = unit->content;
+
+ err = cbs_h266_read_sps(ctx, &gbc, sps);
+ if (err < 0)
+ return err;
+
+ err = cbs_h266_replace_sps(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_PPS_NUT:
+ {
+ H266RawPPS *pps = unit->content;
+
+ err = cbs_h266_read_pps(ctx, &gbc, pps);
+ if (err < 0)
+ return err;
+
+ err = cbs_h266_replace_pps(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_PH_NUT:
+ {
+ H266RawPH *ph = unit->content;
+ err = cbs_h266_read_ph(ctx, &gbc, ph);
+ if (err < 0)
+ return err;
+ err = cbs_h266_replace_ph(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_TRAIL_NUT:
+ case VVC_STSA_NUT:
+ case VVC_RADL_NUT:
+ case VVC_RASL_NUT:
+ case VVC_IDR_W_RADL:
+ case VVC_IDR_N_LP:
+ case VVC_CRA_NUT:
+ case VVC_GDR_NUT:
+ {
+ H266RawSlice *slice = unit->content;
+ int pos, len;
+
+ err = cbs_h266_read_slice_header(ctx, &gbc, &slice->header);
+ if (err < 0)
+ return err;
+
+ if (!cbs_h2645_read_more_rbsp_data(&gbc))
+ return AVERROR_INVALIDDATA;
+
+ pos = get_bits_count(&gbc);
+ len = unit->data_size;
+
+ slice->data_size = len - pos / 8;
+ slice->data_ref = av_buffer_ref(unit->data_ref);
+ if (!slice->data_ref)
+ return AVERROR(ENOMEM);
+ slice->data = unit->data + pos / 8;
+ slice->data_bit_start = pos % 8;
+ }
+ break;
+
+ case VVC_AUD_NUT:
+ {
+ err = cbs_h266_read_aud(ctx, &gbc, unit->content);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_PREFIX_SEI_NUT:
+ case VVC_SUFFIX_SEI_NUT:
+ {
+ err = cbs_h266_read_sei(ctx, &gbc, unit->content,
+ unit->type == VVC_PREFIX_SEI_NUT);
+
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ default:
+ return AVERROR(ENOSYS);
+ }
+ return 0;
+}
+
static int cbs_h2645_write_slice_data(CodedBitstreamContext *ctx,
PutBitContext *pbc, const uint8_t *data,
size_t data_size, int data_bit_start)
@@ -1213,11 +1439,127 @@ static int cbs_h265_write_nal_unit(CodedBitstreamContext *ctx,
return 0;
}
+static int cbs_h266_write_nal_unit(CodedBitstreamContext *ctx,
+ CodedBitstreamUnit *unit,
+ PutBitContext *pbc)
+{
+ int err;
+
+ switch (unit->type) {
+ case VVC_VPS_NUT:
+ {
+ H266RawVPS *vps = unit->content;
+
+ err = cbs_h266_write_vps(ctx, pbc, vps);
+ if (err < 0)
+ return err;
+
+ err = cbs_h266_replace_vps(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+ case VVC_SPS_NUT:
+ {
+ H266RawSPS *sps = unit->content;
+
+ err = cbs_h266_write_sps(ctx, pbc, sps);
+ if (err < 0)
+ return err;
+
+ err = cbs_h266_replace_sps(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_PPS_NUT:
+ {
+ H266RawPPS *pps = unit->content;
+
+ err = cbs_h266_write_pps(ctx, pbc, pps);
+ if (err < 0)
+ return err;
+
+ err = cbs_h266_replace_pps(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_PH_NUT:
+ {
+ H266RawPH *ph = unit->content;
+ err = cbs_h266_write_ph(ctx, pbc, ph);
+ if (err < 0)
+ return err;
+
+ err = cbs_h266_replace_ph(ctx, unit);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_TRAIL_NUT:
+ case VVC_STSA_NUT:
+ case VVC_RADL_NUT:
+ case VVC_RASL_NUT:
+ case VVC_IDR_W_RADL:
+ case VVC_IDR_N_LP:
+ case VVC_CRA_NUT:
+ case VVC_GDR_NUT:
+ {
+ H266RawSlice *slice = unit->content;
+
+ err = cbs_h266_write_slice_header(ctx, pbc, &slice->header);
+ if (err < 0)
+ return err;
+
+ if (slice->data) {
+ err = cbs_h2645_write_slice_data(ctx, pbc, slice->data,
+ slice->data_size,
+ slice->data_bit_start);
+ if (err < 0)
+ return err;
+ } else {
+ // No slice data - that was just the header.
+ }
+ }
+ break;
+
+ case VVC_AUD_NUT:
+ {
+ err = cbs_h266_write_aud(ctx, pbc, unit->content);
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ case VVC_PREFIX_SEI_NUT:
+ case VVC_SUFFIX_SEI_NUT:
+ {
+ err = cbs_h266_write_sei(ctx, pbc, unit->content,
+ unit->type == VVC_PREFIX_SEI_NUT);
+
+ if (err < 0)
+ return err;
+ }
+ break;
+
+ default:
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "Write unimplemented for "
+ "NAL unit type %"PRIu32".\n", unit->type);
+ return AVERROR_PATCHWELCOME;
+ }
+
+ return 0;
+}
+
static int cbs_h2645_unit_requires_zero_byte(enum AVCodecID codec_id,
CodedBitstreamUnitType type,
int nal_unit_index)
{
- // Section B.1.2 in H.264, section B.2.2 in H.265.
+ // Section B.1.2 in H.264, section B.2.2 in H.265, H.266.
if (nal_unit_index == 0) {
// Assume that this is the first NAL unit in an access unit.
return 1;
@@ -1226,6 +1568,8 @@ static int cbs_h2645_unit_requires_zero_byte(enum AVCodecID codec_id,
return type == H264_NAL_SPS || type == H264_NAL_PPS;
if (codec_id == AV_CODEC_ID_HEVC)
return type == HEVC_NAL_VPS || type == HEVC_NAL_SPS || type == HEVC_NAL_PPS;
+ if (codec_id == AV_CODEC_ID_VVC)
+ return type >= VVC_OPI_NUT && type <= VVC_SUFFIX_APS_NUT;
return 0;
}
@@ -1377,6 +1721,35 @@ static void cbs_h265_close(CodedBitstreamContext *ctx)
av_buffer_unref(&h265->pps_ref[i]);
}
+static void cbs_h266_flush(CodedBitstreamContext *ctx)
+{
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+
+ for (int i = 0; i < FF_ARRAY_ELEMS(h266->vps); i++) {
+ av_buffer_unref(&h266->vps_ref[i]);
+ h266->vps[i] = NULL;
+ }
+
+ for (int i = 0; i < FF_ARRAY_ELEMS(h266->sps); i++) {
+ av_buffer_unref(&h266->sps_ref[i]);
+ h266->sps[i] = NULL;
+ }
+ for (int i = 0; i < FF_ARRAY_ELEMS(h266->pps); i++) {
+ av_buffer_unref(&h266->pps_ref[i]);
+ h266->pps[i] = NULL;
+ }
+ av_buffer_unref(&h266->priv.ph_ref);
+ h266->priv.ph = NULL;
+}
+
+static void cbs_h266_close(CodedBitstreamContext *ctx)
+{
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+
+ cbs_h266_flush(ctx);
+ ff_h2645_packet_uninit(&h266->common.read_packet);
+ }
+
static void cbs_h264_free_sei(void *opaque, uint8_t *content)
{
H264RawSEI *sei = (H264RawSEI*)content;
@@ -1431,6 +1804,33 @@ static const CodedBitstreamUnitTypeDescriptor cbs_h265_unit_types[] = {
CBS_UNIT_TYPE_END_OF_LIST
};
+static void cbs_h266_free_sei(void *opaque, uint8_t *content)
+{
+ H266RawSEI *sei = (H266RawSEI*)content;
+ ff_cbs_sei_free_message_list(&sei->message_list);
+ av_free(content);
+}
+
+static const CodedBitstreamUnitTypeDescriptor cbs_h266_unit_types[] = {
+ CBS_UNIT_TYPE_INTERNAL_REF(VVC_VPS_NUT, H266RawVPS, extension_data.data),
+ CBS_UNIT_TYPE_INTERNAL_REF(VVC_SPS_NUT, H266RawSPS, extension_data.data),
+ CBS_UNIT_TYPE_INTERNAL_REF(VVC_PPS_NUT, H266RawPPS, extension_data.data),
+
+ CBS_UNIT_TYPE_POD(VVC_PH_NUT , H266RawPH),
+ CBS_UNIT_TYPE_POD(VVC_AUD_NUT, H266RawAUD),
+
+ CBS_UNIT_RANGE_INTERNAL_REF(VVC_TRAIL_NUT, VVC_RASL_NUT,
+ H266RawSlice, data),
+
+ CBS_UNIT_RANGE_INTERNAL_REF(VVC_IDR_W_RADL, VVC_GDR_NUT,
+ H266RawSlice, data),
+
+ CBS_UNIT_TYPES_COMPLEX((VVC_PREFIX_SEI_NUT, VVC_SUFFIX_SEI_NUT),
+ H266RawSEI, cbs_h266_free_sei),
+
+ CBS_UNIT_TYPE_END_OF_LIST
+};
+
const CodedBitstreamType ff_cbs_type_h264 = {
.codec_id = AV_CODEC_ID_H264,
@@ -1463,6 +1863,22 @@ const CodedBitstreamType ff_cbs_type_h265 = {
.close = &cbs_h265_close,
};
+const CodedBitstreamType ff_cbs_type_h266 = {
+ .codec_id = AV_CODEC_ID_VVC,
+
+ .priv_data_size = sizeof(CodedBitstreamH266Context),
+
+ .unit_types = cbs_h266_unit_types,
+
+ .split_fragment = &cbs_h2645_split_fragment,
+ .read_unit = &cbs_h266_read_nal_unit,
+ .write_unit = &cbs_h266_write_nal_unit,
+ .assemble_fragment = &cbs_h2645_assemble_fragment,
+
+ .flush = &cbs_h266_flush,
+ .close = &cbs_h266_close,
+};
+
static const SEIMessageTypeDescriptor cbs_sei_common_types[] = {
{
SEI_TYPE_FILLER_PAYLOAD,
@@ -1613,6 +2029,16 @@ static const SEIMessageTypeDescriptor cbs_sei_h265_types[] = {
SEI_MESSAGE_TYPE_END
};
+static const SEIMessageTypeDescriptor cbs_sei_h266_types[] = {
+ {
+ SEI_TYPE_DECODED_PICTURE_HASH,
+ 0, 1,
+ sizeof(H266RawSEIDecodedPictureHash),
+ SEI_MESSAGE_RW(h266, sei_decoded_picture_hash),
+ },
+ SEI_MESSAGE_TYPE_END
+};
+
const SEIMessageTypeDescriptor *ff_cbs_sei_find_type(CodedBitstreamContext *ctx,
int payload_type)
{
@@ -1631,6 +2057,9 @@ const SEIMessageTypeDescriptor *ff_cbs_sei_find_type(CodedBitstreamContext *ctx,
case AV_CODEC_ID_H265:
codec_list = cbs_sei_h265_types;
break;
+ case AV_CODEC_ID_H266:
+ codec_list = cbs_sei_h266_types;
+ break;
default:
return NULL;
}
diff --git a/libavcodec/cbs_h266.h b/libavcodec/cbs_h266.h
new file mode 100644
index 0000000000..3016bbb89d
--- /dev/null
+++ b/libavcodec/cbs_h266.h
@@ -0,0 +1,793 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVCODEC_CBS_H266_H
+#define AVCODEC_CBS_H266_H
+
+#include <stddef.h>
+#include <stdint.h>
+
+#include "cbs_h2645.h"
+#include "cbs_sei.h"
+#include "h266.h"
+
+#define IS_H266_SLICE(nut) (nut <= VVC_RASL_NUT || (nut >= VVC_IDR_W_RADL && nut <= VVC_GDR_NUT))
+
+typedef struct H266RawNALUnitHeader {
+ uint8_t nuh_layer_id;
+ uint8_t nal_unit_type;
+ uint8_t nuh_temporal_id_plus1;
+ uint8_t nuh_reserved_zero_bit;
+} H266RawNALUnitHeader;
+
+typedef struct H266GeneralConstraintsInfo {
+ uint8_t gci_present_flag;
+ /* general */
+ uint8_t gci_intra_only_constraint_flag;
+ uint8_t gci_all_layers_independent_constraint_flag;
+ uint8_t gci_one_au_only_constraint_flag;
+
+ /* picture format */
+ uint8_t gci_sixteen_minus_max_bitdepth_constraint_idc;
+ uint8_t gci_three_minus_max_chroma_format_constraint_idc;
+
+ /* NAL unit type related */
+ uint8_t gci_no_mixed_nalu_types_in_pic_constraint_flag;
+ uint8_t gci_no_trail_constraint_flag;
+ uint8_t gci_no_stsa_constraint_flag;
+ uint8_t gci_no_rasl_constraint_flag;
+ uint8_t gci_no_radl_constraint_flag;
+ uint8_t gci_no_idr_constraint_flag;
+ uint8_t gci_no_cra_constraint_flag;
+ uint8_t gci_no_gdr_constraint_flag;
+ uint8_t gci_no_aps_constraint_flag;
+ uint8_t gci_no_idr_rpl_constraint_flag;
+
+ /* tile, slice, subpicture partitioning */
+ uint8_t gci_one_tile_per_pic_constraint_flag;
+ uint8_t gci_pic_header_in_slice_header_constraint_flag;
+ uint8_t gci_one_slice_per_pic_constraint_flag;
+ uint8_t gci_no_rectangular_slice_constraint_flag;
+ uint8_t gci_one_slice_per_subpic_constraint_flag;
+ uint8_t gci_no_subpic_info_constraint_flag;
+
+ /* CTU and block partitioning */
+ uint8_t gci_three_minus_max_log2_ctu_size_constraint_idc;
+ uint8_t gci_no_partition_constraints_override_constraint_flag;
+ uint8_t gci_no_mtt_constraint_flag;
+ uint8_t gci_no_qtbtt_dual_tree_intra_constraint_flag;
+
+ /* intra */
+ uint8_t gci_no_palette_constraint_flag;
+ uint8_t gci_no_ibc_constraint_flag;
+ uint8_t gci_no_isp_constraint_flag;
+ uint8_t gci_no_mrl_constraint_flag;
+ uint8_t gci_no_mip_constraint_flag;
+ uint8_t gci_no_cclm_constraint_flag;
+
+ /* inter */
+ uint8_t gci_no_ref_pic_resampling_constraint_flag;
+ uint8_t gci_no_res_change_in_clvs_constraint_flag;;
+ uint8_t gci_no_weighted_prediction_constraint_flag;
+ uint8_t gci_no_ref_wraparound_constraint_flag;
+ uint8_t gci_no_temporal_mvp_constraint_flag;
+ uint8_t gci_no_sbtmvp_constraint_flag;
+ uint8_t gci_no_amvr_constraint_flag;
+ uint8_t gci_no_bdof_constraint_flag;
+ uint8_t gci_no_smvd_constraint_flag;
+ uint8_t gci_no_dmvr_constraint_flag;
+ uint8_t gci_no_mmvd_constraint_flag;
+ uint8_t gci_no_affine_motion_constraint_flag;
+ uint8_t gci_no_prof_constraint_flag;
+ uint8_t gci_no_bcw_constraint_flag;
+ uint8_t gci_no_ciip_constraint_flag;
+ uint8_t gci_no_gpm_constraint_flag;
+
+ /* transform, quantization, residual */
+ uint8_t gci_no_luma_transform_size_64_constraint_flag;
+ uint8_t gci_no_transform_skip_constraint_flag;
+ uint8_t gci_no_bdpcm_constraint_flag;
+ uint8_t gci_no_mts_constraint_flag;
+ uint8_t gci_no_lfnst_constraint_flag;
+ uint8_t gci_no_joint_cbcr_constraint_flag;
+ uint8_t gci_no_sbt_constraint_flag;
+ uint8_t gci_no_act_constraint_flag;
+ uint8_t gci_no_explicit_scaling_list_constraint_flag;
+ uint8_t gci_no_dep_quant_constraint_flag;
+ uint8_t gci_no_sign_data_hiding_constraint_flag;
+ uint8_t gci_no_cu_qp_delta_constraint_flag;
+ uint8_t gci_no_chroma_qp_offset_constraint_flag;
+
+ /* loop filter */
+ uint8_t gci_no_sao_constraint_flag;
+ uint8_t gci_no_alf_constraint_flag;
+ uint8_t gci_no_ccalf_constraint_flag;
+ uint8_t gci_no_lmcs_constraint_flag;
+ uint8_t gci_no_ladf_constraint_flag;
+ uint8_t gci_no_virtual_boundaries_constraint_flag;
+ uint8_t gci_num_reserved_bits;
+ uint8_t gci_reserved_zero_bit[255];
+} H266GeneralConstraintsInfo;
+
+typedef struct H266RawProfileTierLevel {
+ uint8_t general_profile_idc;
+ uint8_t general_tier_flag;
+ uint8_t general_level_idc;
+ uint8_t ptl_frame_only_constraint_flag;
+ uint8_t ptl_multilayer_enabled_flag;
+ H266GeneralConstraintsInfo general_constraints_info;
+ uint8_t ptl_sublayer_level_present_flag[VVC_MAX_SUBLAYERS - 1];
+ uint8_t sublayer_level_idc[VVC_MAX_SUBLAYERS - 1];
+ uint8_t ptl_num_sub_profiles;
+ uint32_t general_sub_profile_idc[VVC_MAX_SUB_PROFILES];
+
+ uint8_t ptl_reserved_zero_bit;
+} H266RawProfileTierLevel;
+
+typedef struct H266RawExtensionData {
+ uint8_t *data;
+ AVBufferRef *data_ref;
+ size_t bit_length;
+} H266RawExtensionData;
+
+typedef struct H266DpbParameters {
+ uint8_t dpb_max_dec_pic_buffering_minus1[VVC_MAX_SUBLAYERS];
+ uint8_t dpb_max_num_reorder_pics[VVC_MAX_SUBLAYERS];
+ uint8_t dpb_max_latency_increase_plus1[VVC_MAX_SUBLAYERS];
+} H266DpbParameters;
+
+typedef struct H266RefPicListStruct {
+ uint8_t num_ref_entries;
+ uint8_t ltrp_in_header_flag;
+ uint8_t inter_layer_ref_pic_flag[VVC_MAX_REF_ENTRIES];
+ uint8_t st_ref_pic_flag[VVC_MAX_REF_ENTRIES];
+ uint8_t abs_delta_poc_st[VVC_MAX_REF_ENTRIES];
+ uint8_t strp_entry_sign_flag[VVC_MAX_REF_ENTRIES];
+ uint8_t rpls_poc_lsb_lt[VVC_MAX_REF_ENTRIES];
+ uint8_t ilrp_idx[VVC_MAX_REF_ENTRIES];
+} H266RefPicListStruct;
+
+typedef struct H266RefPicLists {
+ uint8_t rpl_sps_flag[2];
+ uint8_t rpl_idx[2];
+ H266RefPicListStruct rpl_ref_list[2];
+ uint16_t poc_lsb_lt[2][VVC_MAX_REF_ENTRIES];
+ uint8_t delta_poc_msb_cycle_present_flag[2][VVC_MAX_REF_ENTRIES];
+ uint16_t delta_poc_msb_cycle_lt[2][VVC_MAX_REF_ENTRIES];
+} H266RefPicLists;
+
+typedef struct H266RawGeneralTimingHrdParameters {
+ uint32_t num_units_in_tick;
+ uint32_t time_scale;
+ uint8_t general_nal_hrd_params_present_flag;
+ uint8_t general_vcl_hrd_params_present_flag;
+ uint8_t general_same_pic_timing_in_all_ols_flag;
+ uint8_t general_du_hrd_params_present_flag;
+ uint8_t tick_divisor_minus2;
+ uint8_t bit_rate_scale;
+ uint8_t cpb_size_scale;
+ uint8_t cpb_size_du_scale;
+ uint8_t hrd_cpb_cnt_minus1;
+} H266RawGeneralTimingHrdParameters;
+
+typedef struct H266RawSubLayerHRDParameters {
+ uint32_t bit_rate_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint32_t cpb_size_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint32_t cpb_size_du_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint32_t bit_rate_du_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint8_t cbr_flag[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+} H266RawSubLayerHRDParameters;
+
+typedef struct H266RawOlsTimingHrdParameters {
+ uint8_t fixed_pic_rate_general_flag[VVC_MAX_SUBLAYERS];
+ uint8_t fixed_pic_rate_within_cvs_flag[VVC_MAX_SUBLAYERS];
+ uint16_t elemental_duration_in_tc_minus1[VVC_MAX_SUBLAYERS];
+ uint8_t low_delay_hrd_flag[VVC_MAX_SUBLAYERS];
+ H266RawSubLayerHRDParameters nal_sub_layer_hrd_parameters;
+ H266RawSubLayerHRDParameters vcl_sub_layer_hrd_parameters;
+} H266RawOlsTimingHrdParameters;
+
+typedef struct H266RawVUI {
+ uint8_t vui_progressive_source_flag;
+ uint8_t vui_interlaced_source_flag;
+ uint8_t vui_non_packed_constraint_flag;
+ uint8_t vui_non_projected_constraint_flag;
+
+ uint8_t vui_aspect_ratio_info_present_flag;
+ uint8_t vui_aspect_ratio_constant_flag;
+ uint8_t vui_aspect_ratio_idc;
+
+ uint16_t vui_sar_width;
+ uint16_t vui_sar_height;;
+
+ uint8_t vui_overscan_info_present_flag;
+ uint8_t vui_overscan_appropriate_flag;
+
+ uint8_t vui_colour_description_present_flag;
+ uint8_t vui_colour_primaries;
+
+ uint8_t vui_transfer_characteristics;
+ uint8_t vui_matrix_coeffs;
+ uint8_t vui_full_range_flag;
+
+ uint8_t vui_chroma_loc_info_present_flag;
+ uint8_t vui_chroma_sample_loc_type_frame;
+ uint8_t vui_chroma_sample_loc_type_top_field;
+ uint8_t vui_chroma_sample_loc_type_bottom_field;
+ H266RawExtensionData extension_data;
+} H266RawVUI;
+
+typedef struct H266RawVPS {
+ H266RawNALUnitHeader nal_unit_header;
+
+ uint8_t vps_video_parameter_set_id;
+ uint8_t vps_max_layers_minus1;
+ uint8_t vps_max_sublayers_minus1;
+ uint8_t vps_default_ptl_dpb_hrd_max_tid_flag;
+ uint8_t vps_all_independent_layers_flag;
+ uint8_t vps_layer_id[VVC_MAX_LAYERS];
+ uint8_t vps_independent_layer_flag[VVC_MAX_LAYERS];
+ uint8_t vps_max_tid_ref_present_flag[VVC_MAX_LAYERS];
+ uint8_t vps_direct_ref_layer_flag[VVC_MAX_LAYERS][VVC_MAX_LAYERS - 1];
+ uint8_t vps_max_tid_il_ref_pics_plus1[VVC_MAX_LAYERS][VVC_MAX_LAYERS - 1];
+ uint8_t vps_each_layer_is_an_ols_flag;
+ uint8_t vps_ols_mode_idc;
+ uint8_t vps_num_output_layer_sets_minus2;
+ uint8_t vps_ols_output_layer_flag[VVC_MAX_TOTAL_NUM_OLSS][VVC_MAX_LAYERS];
+
+ uint8_t vps_num_ptls_minus1;
+ uint8_t vps_pt_present_flag[VVC_MAX_PTLS];
+ uint8_t vps_ptl_max_tid[VVC_MAX_PTLS];
+ H266RawProfileTierLevel vps_profile_tier_level[VVC_MAX_PTLS];
+ uint8_t vps_ols_ptl_idx[VVC_MAX_TOTAL_NUM_OLSS];
+
+ uint16_t vps_num_dpb_params_minus1;
+ uint8_t vps_sublayer_dpb_params_present_flag;
+ uint8_t vps_dpb_max_tid[VVC_MAX_TOTAL_NUM_OLSS];
+ H266DpbParameters vps_dpb_params[VVC_MAX_TOTAL_NUM_OLSS];
+ uint16_t vps_ols_dpb_pic_width[VVC_MAX_TOTAL_NUM_OLSS];
+ uint16_t vps_ols_dpb_pic_height[VVC_MAX_TOTAL_NUM_OLSS];
+ uint8_t vps_ols_dpb_chroma_format[VVC_MAX_TOTAL_NUM_OLSS];
+ uint8_t vps_ols_dpb_bitdepth_minus8[VVC_MAX_TOTAL_NUM_OLSS];
+ uint16_t vps_ols_dpb_params_idx[VVC_MAX_TOTAL_NUM_OLSS];
+
+ uint8_t vps_timing_hrd_params_present_flag;
+ H266RawGeneralTimingHrdParameters vps_general_timing_hrd_parameters;
+ uint8_t vps_sublayer_cpb_params_present_flag;
+ uint16_t vps_num_ols_timing_hrd_params_minus1;
+ uint8_t vps_hrd_max_tid[VVC_MAX_TOTAL_NUM_OLSS];
+ H266RawOlsTimingHrdParameters vps_ols_timing_hrd_parameters;
+ uint8_t vps_ols_timing_hrd_idx[VVC_MAX_TOTAL_NUM_OLSS];
+
+ uint8_t vps_extension_flag;
+ H266RawExtensionData extension_data;
+} H266RawVPS;
+
+typedef struct H266RawSPS {
+ H266RawNALUnitHeader nal_unit_header;
+
+ uint8_t sps_seq_parameter_set_id;
+ uint8_t sps_video_parameter_set_id;
+ uint8_t sps_max_sublayers_minus1;
+ uint8_t sps_chroma_format_idc;
+ uint8_t sps_log2_ctu_size_minus5;
+ uint8_t sps_ptl_dpb_hrd_params_present_flag;
+ H266RawProfileTierLevel profile_tier_level;
+ uint8_t sps_gdr_enabled_flag;
+ uint8_t sps_ref_pic_resampling_enabled_flag;
+ uint8_t sps_res_change_in_clvs_allowed_flag;
+
+ uint16_t sps_pic_width_max_in_luma_samples;
+ uint16_t sps_pic_height_max_in_luma_samples;
+
+ uint8_t sps_conformance_window_flag;
+ uint16_t sps_conf_win_left_offset;
+ uint16_t sps_conf_win_right_offset;
+ uint16_t sps_conf_win_top_offset;
+ uint16_t sps_conf_win_bottom_offset;
+
+ uint8_t sps_subpic_info_present_flag;
+ uint16_t sps_num_subpics_minus1;
+ uint8_t sps_independent_subpics_flag;
+ uint8_t sps_subpic_same_size_flag;
+ uint16_t sps_subpic_ctu_top_left_x[VVC_MAX_SLICES];
+ uint16_t sps_subpic_ctu_top_left_y[VVC_MAX_SLICES];
+ uint16_t sps_subpic_width_minus1[VVC_MAX_SLICES];
+ uint16_t sps_subpic_height_minus1[VVC_MAX_SLICES];
+ uint8_t sps_subpic_treated_as_pic_flag[VVC_MAX_SLICES];
+ uint8_t sps_loop_filter_across_subpic_enabled_flag[VVC_MAX_SLICES];
+ uint8_t sps_subpic_id_len_minus1;
+ uint8_t sps_subpic_id_mapping_explicitly_signalled_flag;
+ uint8_t sps_subpic_id_mapping_present_flag;
+ uint32_t sps_subpic_id[VVC_MAX_SLICES];
+
+
+ uint8_t sps_bitdepth_minus8;
+ uint8_t sps_entropy_coding_sync_enabled_flag;
+ uint8_t sps_entry_point_offsets_present_flag;
+
+ uint8_t sps_log2_max_pic_order_cnt_lsb_minus4;
+ uint8_t sps_poc_msb_cycle_flag;
+ uint8_t sps_poc_msb_cycle_len_minus1;
+
+ uint8_t sps_num_extra_ph_bytes;
+ uint8_t sps_extra_ph_bit_present_flag[16];
+
+ uint8_t sps_num_extra_sh_bytes;
+ uint8_t sps_extra_sh_bit_present_flag[16];
+
+ uint8_t sps_sublayer_dpb_params_flag;
+ H266DpbParameters sps_dpb_params;
+
+ uint8_t sps_log2_min_luma_coding_block_size_minus2;
+ uint8_t sps_partition_constraints_override_enabled_flag;
+ uint8_t sps_log2_diff_min_qt_min_cb_intra_slice_luma;
+ uint8_t sps_max_mtt_hierarchy_depth_intra_slice_luma;
+ uint8_t sps_log2_diff_max_bt_min_qt_intra_slice_luma;
+ uint8_t sps_log2_diff_max_tt_min_qt_intra_slice_luma;
+
+ uint8_t sps_qtbtt_dual_tree_intra_flag;
+ uint8_t sps_log2_diff_min_qt_min_cb_intra_slice_chroma;
+ uint8_t sps_max_mtt_hierarchy_depth_intra_slice_chroma;
+ uint8_t sps_log2_diff_max_bt_min_qt_intra_slice_chroma;
+ uint8_t sps_log2_diff_max_tt_min_qt_intra_slice_chroma;
+
+ uint8_t sps_log2_diff_min_qt_min_cb_inter_slice;
+ uint8_t sps_max_mtt_hierarchy_depth_inter_slice;
+ uint8_t sps_log2_diff_max_bt_min_qt_inter_slice;
+ uint8_t sps_log2_diff_max_tt_min_qt_inter_slice;
+
+ uint8_t sps_max_luma_transform_size_64_flag;
+
+ uint8_t sps_transform_skip_enabled_flag;
+ uint8_t sps_log2_transform_skip_max_size_minus2;
+ uint8_t sps_bdpcm_enabled_flag;
+
+ uint8_t sps_mts_enabled_flag;
+ uint8_t sps_explicit_mts_intra_enabled_flag;
+ uint8_t sps_explicit_mts_inter_enabled_flag;
+
+ uint8_t sps_lfnst_enabled_flag;
+
+ uint8_t sps_joint_cbcr_enabled_flag;
+ uint8_t sps_same_qp_table_for_chroma_flag;
+
+ int8_t sps_qp_table_start_minus26[VVC_MAX_SAMPLE_ARRAYS];
+ uint8_t sps_num_points_in_qp_table_minus1[VVC_MAX_SAMPLE_ARRAYS];
+ uint8_t sps_delta_qp_in_val_minus1[VVC_MAX_SAMPLE_ARRAYS][VVC_MAX_POINTS_IN_QP_TABLE];
+ uint8_t sps_delta_qp_diff_val[VVC_MAX_SAMPLE_ARRAYS][VVC_MAX_POINTS_IN_QP_TABLE];
+
+ uint8_t sps_sao_enabled_flag;
+ uint8_t sps_alf_enabled_flag;
+ uint8_t sps_ccalf_enabled_flag;
+ uint8_t sps_lmcs_enabled_flag;
+ uint8_t sps_weighted_pred_flag;
+ uint8_t sps_weighted_bipred_flag;
+ uint8_t sps_long_term_ref_pics_flag;
+ uint8_t sps_inter_layer_prediction_enabled_flag;
+ uint8_t sps_idr_rpl_present_flag;
+ uint8_t sps_rpl1_same_as_rpl0_flag;
+
+ uint8_t sps_num_ref_pic_lists[2];
+ H266RefPicListStruct sps_ref_pic_list_struct[2][VVC_MAX_REF_PIC_LISTS];
+
+ uint8_t sps_ref_wraparound_enabled_flag;
+ uint8_t sps_temporal_mvp_enabled_flag;
+ uint8_t sps_sbtmvp_enabled_flag;
+ uint8_t sps_amvr_enabled_flag;
+ uint8_t sps_bdof_enabled_flag;
+ uint8_t sps_bdof_control_present_in_ph_flag;
+ uint8_t sps_smvd_enabled_flag;
+ uint8_t sps_dmvr_enabled_flag;
+ uint8_t sps_dmvr_control_present_in_ph_flag;
+ uint8_t sps_mmvd_enabled_flag;
+ uint8_t sps_mmvd_fullpel_only_enabled_flag;
+ uint8_t sps_six_minus_max_num_merge_cand;
+ uint8_t sps_sbt_enabled_flag;
+ uint8_t sps_affine_enabled_flag;
+ uint8_t sps_five_minus_max_num_subblock_merge_cand;
+ uint8_t sps_6param_affine_enabled_flag;
+ uint8_t sps_affine_amvr_enabled_flag;
+ uint8_t sps_affine_prof_enabled_flag;
+ uint8_t sps_prof_control_present_in_ph_flag;
+ uint8_t sps_bcw_enabled_flag;
+ uint8_t sps_ciip_enabled_flag;
+ uint8_t sps_gpm_enabled_flag;
+ uint8_t sps_max_num_merge_cand_minus_max_num_gpm_cand;
+ uint8_t sps_log2_parallel_merge_level_minus2;
+ uint8_t sps_isp_enabled_flag;
+ uint8_t sps_mrl_enabled_flag;
+ uint8_t sps_mip_enabled_flag;
+ uint8_t sps_cclm_enabled_flag;
+ uint8_t sps_chroma_horizontal_collocated_flag;
+ uint8_t sps_chroma_vertical_collocated_flag;
+ uint8_t sps_palette_enabled_flag;
+ uint8_t sps_act_enabled_flag;
+ uint8_t sps_min_qp_prime_ts;
+ uint8_t sps_ibc_enabled_flag;
+ uint8_t sps_six_minus_max_num_ibc_merge_cand;
+ uint8_t sps_ladf_enabled_flag;
+ uint8_t sps_num_ladf_intervals_minus2;
+ int8_t sps_ladf_lowest_interval_qp_offset;
+ int8_t sps_ladf_qp_offset[4];
+ uint16_t sps_ladf_delta_threshold_minus1[4];
+
+ uint8_t sps_explicit_scaling_list_enabled_flag;
+ uint8_t sps_scaling_matrix_for_lfnst_disabled_flag;
+ uint8_t sps_scaling_matrix_for_alternative_colour_space_disabled_flag;
+ uint8_t sps_scaling_matrix_designated_colour_space_flag;
+ uint8_t sps_dep_quant_enabled_flag;
+ uint8_t sps_sign_data_hiding_enabled_flag;
+
+ uint8_t sps_virtual_boundaries_enabled_flag;
+ uint8_t sps_virtual_boundaries_present_flag;
+ uint8_t sps_num_ver_virtual_boundaries;
+ uint16_t sps_virtual_boundary_pos_x_minus1[3];
+ uint8_t sps_num_hor_virtual_boundaries;
+ uint16_t sps_virtual_boundary_pos_y_minus1[3];
+
+ uint8_t sps_timing_hrd_params_present_flag;
+ uint8_t sps_sublayer_cpb_params_present_flag;
+ H266RawGeneralTimingHrdParameters sps_general_timing_hrd_parameters;
+ H266RawOlsTimingHrdParameters sps_ols_timing_hrd_parameters;
+
+ uint8_t sps_field_seq_flag;
+ uint8_t sps_vui_parameters_present_flag;
+ uint16_t sps_vui_payload_size_minus1;
+ H266RawVUI vui;
+
+ uint8_t sps_extension_flag;
+
+ H266RawExtensionData extension_data;
+
+} H266RawSPS;
+
+typedef struct H266RawPPS {
+ H266RawNALUnitHeader nal_unit_header;
+
+ uint8_t pps_pic_parameter_set_id;
+ uint8_t pps_seq_parameter_set_id;
+ uint8_t pps_mixed_nalu_types_in_pic_flag;
+ uint16_t pps_pic_width_in_luma_samples;
+ uint16_t pps_pic_height_in_luma_samples;
+
+ uint8_t pps_conformance_window_flag;
+ uint16_t pps_conf_win_left_offset;
+ uint16_t pps_conf_win_right_offset;
+ uint16_t pps_conf_win_top_offset;
+ uint16_t pps_conf_win_bottom_offset;
+
+ uint8_t pps_scaling_window_explicit_signalling_flag;
+ int16_t pps_scaling_win_left_offset;
+ int16_t pps_scaling_win_right_offset;
+ int16_t pps_scaling_win_top_offset;
+ int16_t pps_scaling_win_bottom_offset;
+
+ uint8_t pps_output_flag_present_flag;
+ uint8_t pps_no_pic_partition_flag;
+
+ uint8_t pps_subpic_id_mapping_present_flag;
+ uint16_t pps_num_subpics_minus1;
+ uint8_t pps_subpic_id_len_minus1;
+ uint16_t pps_subpic_id[VVC_MAX_SLICES];
+
+ uint8_t pps_log2_ctu_size_minus5;
+ uint8_t pps_num_exp_tile_columns_minus1;
+ uint8_t pps_num_exp_tile_rows_minus1;
+ uint16_t pps_tile_column_width_minus1[VVC_MAX_TILE_COLUMNS];
+ uint16_t pps_tile_row_height_minus1[VVC_MAX_TILE_ROWS];
+
+ uint8_t pps_loop_filter_across_tiles_enabled_flag;
+ uint8_t pps_rect_slice_flag;
+ uint8_t pps_single_slice_per_subpic_flag;
+
+ uint16_t pps_num_slices_in_pic_minus1;
+ uint8_t pps_tile_idx_delta_present_flag;
+ uint16_t pps_slice_width_in_tiles_minus1[VVC_MAX_SLICES];
+ uint16_t pps_slice_height_in_tiles_minus1[VVC_MAX_SLICES];
+ uint16_t pps_num_exp_slices_in_tile[VVC_MAX_SLICES];
+ uint16_t pps_exp_slice_height_in_ctus_minus1[VVC_MAX_SLICES][VVC_MAX_TILE_ROWS];
+ int16_t pps_tile_idx_delta_val[VVC_MAX_SLICES];
+
+ uint8_t pps_loop_filter_across_slices_enabled_flag;
+ uint8_t pps_cabac_init_present_flag;
+ uint8_t pps_num_ref_idx_default_active_minus1[2];
+ uint8_t pps_rpl1_idx_present_flag;
+ uint8_t pps_weighted_pred_flag;
+ uint8_t pps_weighted_bipred_flag;
+ uint8_t pps_ref_wraparound_enabled_flag;
+ uint16_t pps_pic_width_minus_wraparound_offset;
+ int8_t pps_init_qp_minus26;
+ uint8_t pps_cu_qp_delta_enabled_flag;
+ uint8_t pps_chroma_tool_offsets_present_flag;
+ int8_t pps_cb_qp_offset;
+ int8_t pps_cr_qp_offset;
+ uint8_t pps_joint_cbcr_qp_offset_present_flag;
+ int8_t pps_joint_cbcr_qp_offset_value;
+ uint8_t pps_slice_chroma_qp_offsets_present_flag;
+ uint8_t pps_cu_chroma_qp_offset_list_enabled_flag;
+ uint8_t pps_chroma_qp_offset_list_len_minus1;
+ uint8_t pps_cb_qp_offset_list[6];
+ uint8_t pps_cr_qp_offset_list[6];
+ uint8_t pps_joint_cbcr_qp_offset_list[6];
+ uint8_t pps_deblocking_filter_control_present_flag;
+ uint8_t pps_deblocking_filter_override_enabled_flag;
+ uint8_t pps_deblocking_filter_disabled_flag;
+ uint8_t pps_dbf_info_in_ph_flag;
+
+ int8_t pps_luma_beta_offset_div2;
+ int8_t pps_luma_tc_offset_div2;
+ int8_t pps_cb_beta_offset_div2;
+ int8_t pps_cb_tc_offset_div2;
+ int8_t pps_cr_beta_offset_div2;
+ int8_t pps_cr_tc_offset_div2;
+
+ uint8_t pps_rpl_info_in_ph_flag;
+ uint8_t pps_sao_info_in_ph_flag;
+ uint8_t pps_alf_info_in_ph_flag;
+ uint8_t pps_wp_info_in_ph_flag;
+ uint8_t pps_qp_delta_info_in_ph_flag;
+
+ uint8_t pps_picture_header_extension_present_flag;
+ uint8_t pps_slice_header_extension_present_flag;
+ uint8_t pps_extension_flag;
+ H266RawExtensionData extension_data;
+
+ //calculated value;
+ uint16_t num_tile_columns;
+ uint16_t num_tile_rows;
+ uint16_t num_tiles_in_pic;
+ uint16_t slice_height_in_ctus[VVC_MAX_SLICES];
+ uint16_t num_slices_in_subpic[VVC_MAX_SLICES];
+} H266RawPPS;
+
+typedef struct H266RawAUD {
+ H266RawNALUnitHeader nal_unit_header;
+ uint8_t aud_irap_or_gdr_flag;
+ uint8_t aud_pic_type;
+} H266RawAUD;
+
+typedef struct H266RawPredWeightTable {
+ uint8_t luma_log2_weight_denom;
+ int8_t delta_chroma_log2_weight_denom;
+
+ uint8_t num_l0_weights;
+ uint8_t luma_weight_l0_flag[15];
+ uint8_t chroma_weight_l0_flag[15];
+ int8_t delta_luma_weight_l0[15];
+ int8_t luma_offset_l0[15];
+ int8_t delta_chroma_weight_l0[15][2];
+ int16_t delta_chroma_offset_l0[15][2];
+
+ uint8_t num_l1_weights;
+ uint8_t luma_weight_l1_flag[15];
+ uint8_t chroma_weight_l1_flag[15];
+ int8_t delta_luma_weight_l1[15];
+ int8_t luma_offset_l1[15];
+ int8_t delta_chroma_weight_l1[15][2];
+ int16_t delta_chroma_offset_l1[15][2];
+} H266RawPredWeightTable;
+
+typedef struct H266RawPH {
+ H266RawNALUnitHeader nal_unit_header;
+ uint8_t ph_gdr_or_irap_pic_flag;
+ uint8_t ph_non_ref_pic_flag;
+ uint8_t ph_gdr_pic_flag;
+ uint8_t ph_inter_slice_allowed_flag;
+ uint8_t ph_intra_slice_allowed_flag;
+ uint8_t ph_pic_parameter_set_id;
+ uint16_t ph_pic_order_cnt_lsb;
+ uint8_t ph_recovery_poc_cnt;
+ uint8_t ph_extra_bit[16];
+ uint8_t ph_poc_msb_cycle_present_flag;
+ uint8_t ph_poc_msb_cycle_val;
+
+ uint8_t ph_alf_enabled_flag;
+ uint8_t ph_num_alf_aps_ids_luma;
+ uint8_t ph_alf_aps_id_luma[8];
+ uint8_t ph_alf_cb_enabled_flag;
+ uint8_t ph_alf_cr_enabled_flag;
+ uint8_t ph_alf_aps_id_chroma;
+ uint8_t ph_alf_cc_cb_enabled_flag;
+ uint8_t ph_alf_cc_cb_aps_id;
+ uint8_t ph_alf_cc_cr_enabled_flag;
+ uint8_t ph_alf_cc_cr_aps_id;
+
+ uint8_t ph_lmcs_enabled_flag;
+ uint8_t ph_lmcs_aps_id;
+ uint8_t ph_chroma_residual_scale_flag;
+ uint8_t ph_explicit_scaling_list_enabled_flag;
+ uint8_t ph_scaling_list_aps_id;
+
+ uint8_t ph_virtual_boundaries_present_flag;
+ uint8_t ph_num_ver_virtual_boundaries;
+ uint16_t ph_virtual_boundary_pos_x_minus1[3];
+ uint8_t ph_num_hor_virtual_boundaries;
+ uint16_t ph_virtual_boundary_pos_y_minus1[3];
+
+ uint8_t ph_pic_output_flag;
+ H266RefPicLists ph_ref_pic_lists;
+
+ uint8_t ph_partition_constraints_override_flag;
+
+ uint8_t ph_log2_diff_min_qt_min_cb_intra_slice_luma;
+ uint8_t ph_max_mtt_hierarchy_depth_intra_slice_luma;
+ uint8_t ph_log2_diff_max_bt_min_qt_intra_slice_luma;
+ uint8_t ph_log2_diff_max_tt_min_qt_intra_slice_luma;
+ uint8_t ph_log2_diff_min_qt_min_cb_intra_slice_chroma;
+
+ uint8_t ph_max_mtt_hierarchy_depth_intra_slice_chroma;
+ uint8_t ph_log2_diff_max_bt_min_qt_intra_slice_chroma;
+ uint8_t ph_log2_diff_max_tt_min_qt_intra_slice_chroma;
+
+ uint8_t ph_cu_qp_delta_subdiv_intra_slice;
+ uint8_t ph_cu_chroma_qp_offset_subdiv_intra_slice;
+
+ uint8_t ph_log2_diff_min_qt_min_cb_inter_slice;
+ uint8_t ph_max_mtt_hierarchy_depth_inter_slice;
+ uint8_t ph_log2_diff_max_bt_min_qt_inter_slice;
+ uint8_t ph_log2_diff_max_tt_min_qt_inter_slice;
+ uint8_t ph_cu_qp_delta_subdiv_inter_slice;
+ uint8_t ph_cu_chroma_qp_offset_subdiv_inter_slice;
+
+ uint8_t ph_temporal_mvp_enabled_flag;
+ uint8_t ph_collocated_from_l0_flag;
+ uint8_t ph_collocated_ref_idx;
+ uint8_t ph_mmvd_fullpel_only_flag;
+ uint8_t ph_mvd_l1_zero_flag;
+ uint8_t ph_bdof_disabled_flag;
+ uint8_t ph_dmvr_disabled_flag;
+ uint8_t ph_prof_disabled_flag;
+
+ H266RawPredWeightTable ph_pred_weight_table;
+
+ int8_t ph_qp_delta;
+ uint8_t ph_joint_cbcr_sign_flag;
+ uint8_t ph_sao_luma_enabled_flag;
+ uint8_t ph_sao_chroma_enabled_flag;
+
+ uint8_t ph_deblocking_params_present_flag;
+ uint8_t ph_deblocking_filter_disabled_flag;
+ int8_t ph_luma_beta_offset_div2;
+ int8_t ph_luma_tc_offset_div2;
+ int8_t ph_cb_beta_offset_div2;
+ int8_t ph_cb_tc_offset_div2;
+ int8_t ph_cr_beta_offset_div2;
+ int8_t ph_cr_tc_offset_div2;
+
+ uint8_t ph_extension_length;
+ uint8_t ph_extension_data_byte[256];
+} H266RawPH;
+
+typedef struct H266RawSliceHeader {
+ H266RawNALUnitHeader nal_unit_header;
+ uint8_t sh_picture_header_in_slice_header_flag;
+ H266RawPH sh_picture_header;
+
+ uint16_t sh_subpic_id;
+ uint16_t sh_slice_address;
+ uint8_t sh_extra_bit[16];
+ uint8_t sh_num_tiles_in_slice_minus1;
+ uint8_t sh_slice_type;
+ uint8_t sh_no_output_of_prior_pics_flag;
+
+ uint8_t sh_alf_enabled_flag;
+ uint8_t sh_num_alf_aps_ids_luma;
+ uint8_t sh_alf_aps_id_luma[8];
+ uint8_t sh_alf_cb_enabled_flag;
+ uint8_t sh_alf_cr_enabled_flag;
+ uint8_t sh_alf_aps_id_chroma;
+ uint8_t sh_alf_cc_cb_enabled_flag;
+ uint8_t sh_alf_cc_cb_aps_id;
+ uint8_t sh_alf_cc_cr_enabled_flag;
+ uint8_t sh_alf_cc_cr_aps_id;
+
+ uint8_t sh_lmcs_used_flag;
+ uint8_t sh_explicit_scaling_list_used_flag;
+
+ H266RefPicLists sh_ref_pic_lists;
+
+ uint8_t sh_num_ref_idx_active_override_flag;
+ uint8_t sh_num_ref_idx_active_minus1[2];
+ uint8_t sh_cabac_init_flag;
+ uint8_t sh_collocated_from_l0_flag;
+ uint8_t sh_collocated_ref_idx;
+
+ H266RawPredWeightTable sh_pred_weight_table;
+
+ int8_t sh_qp_delta;
+ int8_t sh_cb_qp_offset;
+ int8_t sh_cr_qp_offset;
+ int8_t sh_joint_cbcr_qp_offset;
+ uint8_t sh_cu_chroma_qp_offset_enabled_flag;
+
+ uint8_t sh_sao_luma_used_flag;
+ uint8_t sh_sao_chroma_used_flag;
+
+ uint8_t sh_deblocking_params_present_flag;
+ uint8_t sh_deblocking_filter_disabled_flag;
+ int8_t sh_luma_beta_offset_div2;
+ int8_t sh_luma_tc_offset_div2;
+ int8_t sh_cb_beta_offset_div2;
+ int8_t sh_cb_tc_offset_div2;
+ int8_t sh_cr_beta_offset_div2;
+ int8_t sh_cr_tc_offset_div2;
+ uint8_t sh_dep_quant_used_flag;
+
+ uint8_t sh_sign_data_hiding_used_flag;
+ uint8_t sh_ts_residual_coding_disabled_flag;
+ uint16_t sh_slice_header_extension_length;
+ uint8_t sh_slice_header_extension_data_byte[256];
+
+ uint8_t sh_entry_offset_len_minus1;
+ uint32_t sh_entry_point_offset_minus1[VVC_MAX_ENTRY_POINTS];
+
+} H266RawSliceHeader;
+
+typedef struct H266RawSlice {
+ H266RawSliceHeader header;
+
+ uint8_t *data;
+ AVBufferRef *data_ref;
+ size_t data_size;
+ int data_bit_start;
+} H266RawSlice;
+
+typedef struct H266RawSEIDecodedPictureHash {
+ uint8_t dph_sei_hash_type;
+ uint8_t dph_sei_single_component_flag;
+ uint8_t dph_sei_picture_md5[3][16];
+ uint16_t dph_sei_picture_crc[3];
+ uint32_t dph_sei_picture_checksum[3];
+
+ uint8_t dph_sei_reserved_zero_7bits;
+} H266RawSEIDecodedPictureHash;
+
+typedef struct H266RawSEI {
+ H266RawNALUnitHeader nal_unit_header;
+ SEIRawMessageList message_list;
+} H266RawSEI;
+
+typedef struct CodedBitstreamH266Context {
+ // Reader/writer context in common with the H.264 implementation.
+ CodedBitstreamH2645Context common;
+
+ // All currently available parameter sets. These are updated when
+ // any parameter set NAL unit is read/written with this context.
+ AVBufferRef *vps_ref[VVC_MAX_VPS_COUNT];
+ AVBufferRef *sps_ref[VVC_MAX_SPS_COUNT];
+ AVBufferRef *pps_ref[VVC_MAX_PPS_COUNT];
+ H266RawVPS *vps[VVC_MAX_SPS_COUNT];
+ H266RawSPS *sps[VVC_MAX_SPS_COUNT];
+ H266RawPPS *pps[VVC_MAX_PPS_COUNT];
+
+ struct {
+ AVBufferRef *ph_ref;
+ H266RawPH *ph;
+ } priv;
+
+ // The currently active parameter sets. These are updated when any
+ // NAL unit refers to the relevant parameter set. These pointers
+ // must also be present in the arrays above.
+ const H266RawVPS *active_vps;
+ const H266RawSPS *active_sps;
+ const H266RawPPS *active_pps;
+
+} CodedBitstreamH266Context;
+
+#endif /* AVCODEC_CBS_H266_H */
diff --git a/libavcodec/cbs_h266_syntax_template.c b/libavcodec/cbs_h266_syntax_template.c
new file mode 100644
index 0000000000..06f9f29e08
--- /dev/null
+++ b/libavcodec/cbs_h266_syntax_template.c
@@ -0,0 +1,3116 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+static int FUNC(rbsp_trailing_bits) (CodedBitstreamContext *ctx,
+ RWContext *rw)
+{
+ int err;
+
+ fixed(1, rbsp_stop_one_bit, 1);
+ while (byte_alignment(rw) != 0)
+ fixed(1, rbsp_alignment_zero_bit, 0);
+ return 0;
+}
+
+static int FUNC(nal_unit_header) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawNALUnitHeader *current,
+ int expected_nal_unit_type)
+{
+ int err;
+
+ fixed(1, forbidden_zero_bit, 0);
+ flag(nuh_reserved_zero_bit);
+
+ u(6, nuh_layer_id, 0, 55);
+
+ if (expected_nal_unit_type >= 0)
+ u(5, nal_unit_type, expected_nal_unit_type, expected_nal_unit_type);
+ else
+ ub(5, nal_unit_type);
+
+ u(3, nuh_temporal_id_plus1, 1, 7);
+ return 0;
+}
+
+static int FUNC(byte_alignment) (CodedBitstreamContext *ctx, RWContext *rw)
+{
+ int err;
+
+ fixed(1, byte_alignment_bit_equal_to_one, 1);
+ while (byte_alignment(rw) != 0)
+ fixed(1, byte_alignment_bit_equal_to_zero, 0);
+ return 0;
+}
+
+static int FUNC(general_constraints_info) (CodedBitstreamContext *ctx,
+ RWContext *rw,
+ H266GeneralConstraintsInfo *current)
+{
+ int err, i;
+
+ flag(gci_present_flag);
+ if (current->gci_present_flag) {
+ /* general */
+ flag(gci_intra_only_constraint_flag);
+ flag(gci_all_layers_independent_constraint_flag);
+ flag(gci_one_au_only_constraint_flag);
+
+ /* picture format */
+ u(4, gci_sixteen_minus_max_bitdepth_constraint_idc, 0, 8);
+ ub(2, gci_three_minus_max_chroma_format_constraint_idc);
+
+ /* NAL unit type related */
+ flag(gci_no_mixed_nalu_types_in_pic_constraint_flag);
+ flag(gci_no_trail_constraint_flag);
+ flag(gci_no_stsa_constraint_flag);
+ flag(gci_no_rasl_constraint_flag);
+ flag(gci_no_radl_constraint_flag);
+ flag(gci_no_idr_constraint_flag);
+ flag(gci_no_cra_constraint_flag);
+ flag(gci_no_gdr_constraint_flag);
+ flag(gci_no_aps_constraint_flag);
+ flag(gci_no_idr_rpl_constraint_flag);
+
+ /* tile, slice, subpicture partitioning */
+ flag(gci_one_tile_per_pic_constraint_flag);
+ flag(gci_pic_header_in_slice_header_constraint_flag);
+ flag(gci_one_slice_per_pic_constraint_flag);
+ flag(gci_no_rectangular_slice_constraint_flag);
+ flag(gci_one_slice_per_subpic_constraint_flag);
+ flag(gci_no_subpic_info_constraint_flag);
+
+ /* CTU and block partitioning */
+ ub(2, gci_three_minus_max_log2_ctu_size_constraint_idc);
+ flag(gci_no_partition_constraints_override_constraint_flag);
+ flag(gci_no_mtt_constraint_flag);
+ flag(gci_no_qtbtt_dual_tree_intra_constraint_flag);
+
+ /* intra */
+ flag(gci_no_palette_constraint_flag);
+ flag(gci_no_ibc_constraint_flag);
+ flag(gci_no_isp_constraint_flag);
+ flag(gci_no_mrl_constraint_flag);
+ flag(gci_no_mip_constraint_flag);
+ flag(gci_no_cclm_constraint_flag);
+
+ /* inter */
+ flag(gci_no_ref_pic_resampling_constraint_flag);
+ flag(gci_no_res_change_in_clvs_constraint_flag);
+ flag(gci_no_weighted_prediction_constraint_flag);
+ flag(gci_no_ref_wraparound_constraint_flag);
+ flag(gci_no_temporal_mvp_constraint_flag);
+ flag(gci_no_sbtmvp_constraint_flag);
+ flag(gci_no_amvr_constraint_flag);
+ flag(gci_no_bdof_constraint_flag);
+ flag(gci_no_smvd_constraint_flag);
+ flag(gci_no_dmvr_constraint_flag);
+ flag(gci_no_mmvd_constraint_flag);
+ flag(gci_no_affine_motion_constraint_flag);
+ flag(gci_no_prof_constraint_flag);
+ flag(gci_no_bcw_constraint_flag);
+ flag(gci_no_ciip_constraint_flag);
+ flag(gci_no_gpm_constraint_flag);
+
+ /* transform, quantization, residual */
+ flag(gci_no_luma_transform_size_64_constraint_flag);
+ flag(gci_no_transform_skip_constraint_flag);
+ flag(gci_no_bdpcm_constraint_flag);
+ flag(gci_no_mts_constraint_flag);
+ flag(gci_no_lfnst_constraint_flag);
+ flag(gci_no_joint_cbcr_constraint_flag);
+ flag(gci_no_sbt_constraint_flag);
+ flag(gci_no_act_constraint_flag);
+ flag(gci_no_explicit_scaling_list_constraint_flag);
+ flag(gci_no_dep_quant_constraint_flag);
+ flag(gci_no_sign_data_hiding_constraint_flag);
+ flag(gci_no_cu_qp_delta_constraint_flag);
+ flag(gci_no_chroma_qp_offset_constraint_flag);
+
+ /* loop filter */
+ flag(gci_no_sao_constraint_flag);
+ flag(gci_no_alf_constraint_flag);
+ flag(gci_no_ccalf_constraint_flag);
+ flag(gci_no_lmcs_constraint_flag);
+ flag(gci_no_ladf_constraint_flag);
+ flag(gci_no_virtual_boundaries_constraint_flag);
+ ub(8, gci_num_reserved_bits);
+ for (i = 0; i < current->gci_num_reserved_bits; i++) {
+ flags(gci_reserved_zero_bit[i], 1, i);
+ }
+ }
+ while (byte_alignment(rw) != 0)
+ fixed(1, gci_alignment_zero_bit, 0);
+ return 0;
+}
+
+static int FUNC(profile_tier_level) (CodedBitstreamContext *ctx,
+ RWContext *rw,
+ H266RawProfileTierLevel *current,
+ int profile_tier_present_flag,
+ int max_num_sub_layers_minus1)
+{
+ int err, i;
+
+ if (profile_tier_present_flag) {
+ ub(7, general_profile_idc);
+ flag(general_tier_flag);
+ }
+ ub(8, general_level_idc);
+ flag(ptl_frame_only_constraint_flag);
+ flag(ptl_multilayer_enabled_flag);
+ if (profile_tier_present_flag) {
+ CHECK(FUNC(general_constraints_info) (ctx, rw,
+ ¤t->
+ general_constraints_info));
+ }
+ for (i = max_num_sub_layers_minus1 - 1; i >= 0; i--)
+ flags(ptl_sublayer_level_present_flag[i], 1, i);
+ while (byte_alignment(rw) != 0)
+ flag(ptl_reserved_zero_bit);
+ for (i = max_num_sub_layers_minus1 - 1; i >= 0; i--)
+ if (current->ptl_sublayer_level_present_flag[i])
+ ubs(8, sublayer_level_idc[i], 1, i);
+ if (profile_tier_present_flag) {
+ ub(8, ptl_num_sub_profiles);
+ for (i = 0; i < current->ptl_num_sub_profiles; i++)
+ ubs(32, general_sub_profile_idc[i], 1, i);
+ }
+ return 0;
+}
+
+static int FUNC(vui_parameters_default) (CodedBitstreamContext *ctx,
+ RWContext *rw, H266RawVUI *current)
+{
+ //defined in D.8
+ infer(vui_progressive_source_flag, 0);
+ infer(vui_interlaced_source_flag, 0);
+
+ infer(vui_non_packed_constraint_flag, 0);
+ infer(vui_non_projected_constraint_flag, 0);
+
+ infer(vui_aspect_ratio_constant_flag, 0);
+ infer(vui_aspect_ratio_idc, 0);
+
+ infer(vui_overscan_info_present_flag, 0);
+
+ infer(vui_colour_primaries, 2);
+ infer(vui_transfer_characteristics, 2);
+ infer(vui_matrix_coeffs, 2);
+ infer(vui_full_range_flag, 0);
+
+ infer(vui_chroma_sample_loc_type_frame, 6);
+ infer(vui_chroma_sample_loc_type_top_field, 6);
+ infer(vui_chroma_sample_loc_type_bottom_field, 6);
+ return 0;
+}
+
+static int FUNC(vui_parameters) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawVUI *current,
+ uint8_t chroma_format_idc)
+{
+ int err;
+
+ flag(vui_progressive_source_flag);
+ flag(vui_interlaced_source_flag);
+ flag(vui_non_packed_constraint_flag);
+ flag(vui_non_projected_constraint_flag);
+ flag(vui_aspect_ratio_info_present_flag);
+ if (current->vui_aspect_ratio_info_present_flag) {
+ flag(vui_aspect_ratio_constant_flag);
+ ub(8, vui_aspect_ratio_idc);
+ if (current->vui_aspect_ratio_idc == 255) {
+ ub(16, vui_sar_width);
+ ub(16, vui_sar_height);
+ }
+ } else {
+ infer(vui_aspect_ratio_constant_flag, 0);
+ infer(vui_aspect_ratio_idc, 0);
+ }
+ flag(vui_overscan_info_present_flag);
+ if (current->vui_overscan_info_present_flag)
+ flag(vui_overscan_appropriate_flag);
+ flag(vui_colour_description_present_flag);
+ if (current->vui_colour_description_present_flag) {
+ ub(8, vui_colour_primaries);
+ av_log(ctx->log_ctx, AV_LOG_DEBUG, "vui_colour_primaries == %d \n",
+ current->vui_colour_primaries);
+ ub(8, vui_transfer_characteristics);
+ av_log(ctx->log_ctx, AV_LOG_DEBUG,
+ "vui_transfer_characteristics == %d \n",
+ current->vui_transfer_characteristics);
+ ub(8, vui_matrix_coeffs);
+ av_log(ctx->log_ctx, AV_LOG_DEBUG, "vui_matrix_coeffs == %d \n",
+ current->vui_matrix_coeffs);
+ flag(vui_full_range_flag);
+ } else {
+ infer(vui_colour_primaries, 2);
+ infer(vui_transfer_characteristics, 2);
+ infer(vui_matrix_coeffs, 2);
+ infer(vui_full_range_flag, 0);
+ }
+ flag(vui_chroma_loc_info_present_flag);
+ if (chroma_format_idc != 1 && current->vui_chroma_loc_info_present_flag) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "chroma_format_idc == %d,"
+ "vui_chroma_loc_info_present_flag can't not be true",
+ chroma_format_idc);
+ return AVERROR_INVALIDDATA;
+ }
+ if (current->vui_chroma_loc_info_present_flag) {
+ if (current->vui_progressive_source_flag &&
+ !current->vui_interlaced_source_flag) {
+ ue(vui_chroma_sample_loc_type_frame, 0, 6);
+ } else {
+ ue(vui_chroma_sample_loc_type_top_field, 0, 6);
+ ue(vui_chroma_sample_loc_type_bottom_field, 0, 6);
+ }
+ } else {
+ if (chroma_format_idc == 1) {
+ infer(vui_chroma_sample_loc_type_frame, 6);
+ infer(vui_chroma_sample_loc_type_top_field,
+ current->vui_chroma_sample_loc_type_frame);
+ infer(vui_chroma_sample_loc_type_bottom_field,
+ current->vui_chroma_sample_loc_type_frame);
+ }
+ }
+ return 0;
+}
+
+static int FUNC(payload_extension) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawExtensionData *current,
+ uint32_t payload_size, int cur_pos)
+{
+ int err;
+ size_t byte_length, k;
+
+#ifdef READ
+ GetBitContext tmp;
+ int bits_left, payload_zero_bits;
+
+ if (!cbs_h265_payload_extension_present(rw, payload_size, cur_pos))
+ return 0;
+
+ bits_left = 8 * payload_size - cur_pos;
+ tmp = *rw;
+ if (bits_left > 8)
+ skip_bits_long(&tmp, bits_left - 8);
+ payload_zero_bits = get_bits(&tmp, FFMIN(bits_left, 8));
+ if (!payload_zero_bits)
+ return AVERROR_INVALIDDATA;
+ payload_zero_bits = ff_ctz(payload_zero_bits);
+ current->bit_length = bits_left - payload_zero_bits - 1;
+ allocate(current->data, (current->bit_length + 7) / 8);
+#endif
+
+ byte_length = (current->bit_length + 7) / 8;
+ for (k = 0; k < byte_length; k++) {
+ int length = FFMIN(current->bit_length - k * 8, 8);
+ xu(length, reserved_payload_extension_data, current->data[k],
+ 0, MAX_UINT_BITS(length), 0);
+ }
+
+ return 0;
+}
+
+static int FUNC(vui_payload) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawVUI *current, uint16_t vui_payload_size,
+ uint8_t chroma_format_idc)
+{
+ int err;
+ int start_position, current_position;
+
+ start_position = bit_position(rw);
+ CHECK(FUNC(vui_parameters) (ctx, rw, current, chroma_format_idc));
+ current_position = bit_position(rw) - start_position;
+
+ if (current_position < 8 * vui_payload_size) {
+ CHECK(FUNC(payload_extension) (ctx, rw, ¤t->extension_data,
+ vui_payload_size, current_position));
+ fixed(1, vui_payload_bit_equal_to_one, 1);
+ while (byte_alignment(rw) != 0)
+ fixed(1, vui_payload_bit_equal_to_zero, 0);
+ }
+ return 0;
+}
+
+static int FUNC(extension_data) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawExtensionData *current)
+{
+ int err;
+ size_t k;
+#ifdef READ
+ GetBitContext start;
+ uint8_t bit;
+ start = *rw;
+ for (k = 0; cbs_h2645_read_more_rbsp_data(rw); k++)
+ skip_bits(rw, 1);
+ current->bit_length = k;
+ if (k > 0) {
+ *rw = start;
+ allocate(current->data, (current->bit_length + 7) / 8);
+ for (k = 0; k < current->bit_length; k++) {
+ xu(1, extension_data, bit, 0, 1, 0);
+ current->data[k / 8] |= bit << (7 - k % 8);
+ }
+ }
+#else
+ for (k = 0; k < current->bit_length; k++)
+ xu(1, extension_data, current->data[k / 8] >> (7 - k % 8) & 1, 0, 1, 0);
+#endif
+ return 0;
+}
+
+static int FUNC(dpb_parameters) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266DpbParameters *current,
+ uint8_t max_sublayers_minus1,
+ uint8_t sublayer_info_flag)
+{
+ int err, i;
+ for (i = (sublayer_info_flag ? 0 : max_sublayers_minus1);
+ i <= max_sublayers_minus1; i++) {
+ ues(dpb_max_dec_pic_buffering_minus1[i], 0, VVC_MAX_DPB_SIZE - 1, 1, i);
+ ues(dpb_max_num_reorder_pics[i],
+ 0, current->dpb_max_dec_pic_buffering_minus1[i], 1, i);
+ ues(dpb_max_latency_increase_plus1[i], 0, UINT32_MAX - 1, 1, i);
+ }
+ return 0;
+}
+
+static int FUNC(ref_pic_list_struct) (CodedBitstreamContext *ctx,
+ RWContext *rw,
+ H266RefPicListStruct *current,
+ uint8_t list_idx, uint8_t rpls_idx,
+ const H266RawSPS *sps)
+{
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+ int err, i, j, general_layer_idx = -1, num_direct_ref_layers = 0;
+ const H266RawVPS *vps = h266->vps[sps->sps_video_parameter_set_id];
+
+ if (!vps) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "VPS id %d not available.\n", sps->sps_video_parameter_set_id);
+ return AVERROR_INVALIDDATA;
+ }
+ //7.4.3.3 (29)
+ for (i = 0; i <= vps->vps_max_layers_minus1; i++) {
+ if (sps->nal_unit_header.nuh_layer_id == vps->vps_layer_id[i]) {
+ general_layer_idx = i;
+ break;
+ }
+ }
+ if (general_layer_idx < 0) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "vps_layer_id %d not available.\n",
+ sps->nal_unit_header.nuh_layer_id);
+ return AVERROR_INVALIDDATA;
+ }
+ //7.4.3.3 (28)
+ for (j = 0; j <= vps->vps_max_layers_minus1; j++) {
+ if (vps->vps_direct_ref_layer_flag[general_layer_idx][j])
+ num_direct_ref_layers++;
+ }
+
+ ue(num_ref_entries, 0, VVC_MAX_REF_ENTRIES);
+ if (sps->sps_long_term_ref_pics_flag &&
+ rpls_idx < sps->sps_num_ref_pic_lists[list_idx] &&
+ current->num_ref_entries > 0)
+ flag(ltrp_in_header_flag);
+ if (sps->sps_long_term_ref_pics_flag &&
+ rpls_idx == sps->sps_num_ref_pic_lists[list_idx])
+ infer(ltrp_in_header_flag, 1);
+ for (i = 0, j = 0; i < current->num_ref_entries; i++) {
+ if (sps->sps_inter_layer_prediction_enabled_flag)
+ flags(inter_layer_ref_pic_flag[i], 1, i);
+ else
+ infer(inter_layer_ref_pic_flag[i], 0);
+
+ if (!current->inter_layer_ref_pic_flag[i]) {
+ if (sps->sps_long_term_ref_pics_flag)
+ flags(st_ref_pic_flag[i], 1, i);
+ else
+ infer(st_ref_pic_flag[i], 1);
+ if (current->st_ref_pic_flag[i]) {
+ int abs_delta_poc_st;
+ ues(abs_delta_poc_st[i], 0, MAX_UINT_BITS(15), 1, i);
+ if ((sps->sps_weighted_pred_flag ||
+ sps->sps_weighted_bipred_flag) && i != 0)
+ abs_delta_poc_st = current->abs_delta_poc_st[i];
+ else
+ abs_delta_poc_st = current->abs_delta_poc_st[i] + 1;
+ if (abs_delta_poc_st > 0)
+ flags(strp_entry_sign_flag[i], 1, i);
+ } else {
+ if (!current->ltrp_in_header_flag) {
+ uint8_t bits = sps->sps_log2_max_pic_order_cnt_lsb_minus4 + 4;
+ ubs(bits, rpls_poc_lsb_lt[j], 1, j);
+ j++;
+ }
+ }
+ } else {
+ if (num_direct_ref_layers == 0) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "num_direct_ref_layers needs > 0.\n");
+ return AVERROR_INVALIDDATA;
+ }
+ ues(ilrp_idx[i], 0, num_direct_ref_layers - 1, 1, i);
+ }
+ }
+ return 0;
+}
+
+static int FUNC(ref_pic_lists) (CodedBitstreamContext *ctx, RWContext *rw,
+ const H266RawSPS *sps, const H266RawPPS *pps,
+ H266RefPicLists *current) {
+ const H266RefPicListStruct * ref_list;
+ int err, i, j, num_ltrp_entries;
+ for (i = 0; i < 2; i++) {
+ if (sps->sps_num_ref_pic_lists[i] > 0 &&
+ (i == 0 || (i == 1 && pps->pps_rpl1_idx_present_flag))) {
+ flags(rpl_sps_flag[i], 1, i);
+ } else {
+ if (sps->sps_num_ref_pic_lists[i] == 0) {
+ infer(rpl_sps_flag[i], 0);
+ } else {
+ if (!pps->pps_rpl1_idx_present_flag && i == 1)
+ infer(rpl_sps_flag[1], current->rpl_sps_flag[0]);
+ }
+ }
+ if (current->rpl_sps_flag[i]) {
+ if (sps->sps_num_ref_pic_lists[i] > 1 &&
+ (i == 0 || (i == 1 && pps->pps_rpl1_idx_present_flag))) {
+ uint8_t bits = av_ceil_log2(sps->sps_num_ref_pic_lists[i]);
+ us(bits, rpl_idx[i], 0, sps->sps_num_ref_pic_lists[i] - 1, 1, i);
+ } else if (sps->sps_num_ref_pic_lists[i] == 1) {
+ infer(rpl_idx[i], 0);
+ } else if (i == 1 && !pps->pps_rpl1_idx_present_flag) {
+ infer(rpl_idx[1], current->rpl_idx[0]);
+ } else {
+ //how to handle this? or never happpend?
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "can't infer the rpl_idx[i]\n");
+ return AVERROR_PATCHWELCOME;
+ }
+ memcpy(¤t->rpl_ref_list[i],
+ &sps->sps_ref_pic_list_struct[i][current->rpl_idx[i]],
+ sizeof(current->rpl_ref_list[i]));
+ } else {
+ CHECK(FUNC(ref_pic_list_struct) (ctx, rw, ¤t->rpl_ref_list[i],
+ i, sps->sps_num_ref_pic_lists[i],
+ sps));
+ }
+ ref_list = ¤t->rpl_ref_list[i];
+
+ num_ltrp_entries = 0;
+ for (int i = 0; i < ref_list->num_ref_entries; i++) {
+ if (!ref_list->inter_layer_ref_pic_flag[i]) {
+ if (!ref_list->st_ref_pic_flag[i]) {
+ num_ltrp_entries++;
+ }
+ }
+ }
+
+ for (j = 0; j < num_ltrp_entries; j++) {
+ if (ref_list->ltrp_in_header_flag) {
+ ubs(sps->sps_log2_max_pic_order_cnt_lsb_minus4 + 4,
+ poc_lsb_lt[i][j], 2, i, j);
+ }
+ flags(delta_poc_msb_cycle_present_flag[i][j], 2, i, j);
+ if (current->delta_poc_msb_cycle_present_flag[i][j]) {
+ uint32_t max =
+ 1 << (32 - sps->sps_log2_max_pic_order_cnt_lsb_minus4 - 4);
+ ues(delta_poc_msb_cycle_lt[i][j], 0, max, 2, i, j);
+ }
+ }
+ }
+ return 0;
+}
+
+static int FUNC(general_timing_hrd_parameters)(CodedBitstreamContext *ctx,
+ RWContext *rw,
+ H266RawGeneralTimingHrdParameters *current)
+{
+ int err;
+ ub(32, num_units_in_tick);
+ u(32, time_scale, 1, MAX_UINT_BITS(32));
+ flag(general_nal_hrd_params_present_flag);
+ flag(general_vcl_hrd_params_present_flag);
+
+ if (current->general_nal_hrd_params_present_flag ||
+ current->general_vcl_hrd_params_present_flag) {
+ flag(general_same_pic_timing_in_all_ols_flag);
+ flag(general_du_hrd_params_present_flag);
+ if (current->general_du_hrd_params_present_flag)
+ ub(8, tick_divisor_minus2);
+ ub(4, bit_rate_scale);
+ ub(4, cpb_size_scale);
+ if (current->general_du_hrd_params_present_flag)
+ ub(4, cpb_size_du_scale);
+ ue(hrd_cpb_cnt_minus1, 0, 31);
+ } else {
+ //infer general_same_pic_timing_in_all_ols_flag?
+ infer(general_du_hrd_params_present_flag, 0);
+ }
+ return 0;
+}
+
+static int FUNC(sublayer_hrd_parameters) (CodedBitstreamContext *ctx,
+ RWContext *rw,
+ H266RawSubLayerHRDParameters *current,
+ int sublayer_id,
+ const H266RawGeneralTimingHrdParameters *general)
+{
+ int err, i;
+ for (i = 0; i <= general->hrd_cpb_cnt_minus1; i++) {
+ ues(bit_rate_value_minus1[sublayer_id][i], 0, UINT32_MAX - 1, 2,
+ sublayer_id, i);
+ ues(cpb_size_value_minus1[sublayer_id][i], 0, UINT32_MAX - 1, 2,
+ sublayer_id, i);
+ if (general->general_du_hrd_params_present_flag) {
+ ues(cpb_size_du_value_minus1[sublayer_id][i],
+ 0, UINT32_MAX - 1, 2, sublayer_id, i);
+ ues(bit_rate_du_value_minus1[sublayer_id][i],
+ 0, UINT32_MAX - 1, 2, sublayer_id, i);
+ }
+ flags(cbr_flag[sublayer_id][i], 2, sublayer_id, i);
+ }
+ return 0;
+}
+
+static int FUNC(ols_timing_hrd_parameters) (CodedBitstreamContext *ctx,
+ RWContext *rw, H266RawOlsTimingHrdParameters *current,
+ uint8_t first_sublayer, uint8_t max_sublayers_minus1,
+ const H266RawGeneralTimingHrdParameters *general)
+{
+ int err, i;
+ for (i = first_sublayer; i <= max_sublayers_minus1; i++) {
+ flags(fixed_pic_rate_general_flag[i], 1, i);
+ if (!current->fixed_pic_rate_general_flag[i])
+ flags(fixed_pic_rate_within_cvs_flag[i], 1, i);
+ else
+ infer(fixed_pic_rate_within_cvs_flag[i], 1);
+ if (current->fixed_pic_rate_within_cvs_flag[i]) {
+ ues(elemental_duration_in_tc_minus1[i], 0, 2047, 1, i);
+ infer(low_delay_hrd_flag[i], 0);
+ } else if ((general->general_nal_hrd_params_present_flag ||
+ general->general_vcl_hrd_params_present_flag) &&
+ general->hrd_cpb_cnt_minus1 == 0) {
+ flags(low_delay_hrd_flag[i], 1, i);
+ } else {
+ infer(low_delay_hrd_flag[i], 0);
+ }
+ if (general->general_nal_hrd_params_present_flag)
+ CHECK(FUNC(sublayer_hrd_parameters) (ctx, rw,
+ ¤t->nal_sub_layer_hrd_parameters,
+ i, general));
+ if (general->general_vcl_hrd_params_present_flag)
+ CHECK(FUNC(sublayer_hrd_parameters) (ctx, rw,
+ ¤t->nal_sub_layer_hrd_parameters,
+ i, general));
+ }
+ return 0;
+}
+
+static int FUNC(vps) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawVPS *current)
+{
+ int err, i, j, k;
+ uint16_t total_num_olss = 0;
+ uint8_t ols_mode_idc = 0;
+ uint16_t num_multi_layer_olss = 0;
+ uint8_t layer_included_in_ols_flag[VVC_MAX_TOTAL_NUM_OLSS][VVC_MAX_LAYERS];
+ uint8_t num_ref_layers[VVC_MAX_LAYERS];
+ uint8_t reference_layer_idx[VVC_MAX_LAYERS][VVC_MAX_LAYERS];
+
+ HEADER("Video Parameter Set");
+
+ CHECK(FUNC(nal_unit_header) (ctx, rw,
+ ¤t->nal_unit_header, VVC_VPS_NUT));
+
+ ub(4, vps_video_parameter_set_id);
+ if (current->vps_video_parameter_set_id == 0) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "vps_video_parameter_set_id should > 0.\n");
+ return AVERROR_INVALIDDATA;
+ }
+ ub(6, vps_max_layers_minus1);
+ u(3, vps_max_sublayers_minus1, 0, 6);
+ if (current->vps_max_layers_minus1 > 0
+ && current->vps_max_sublayers_minus1 > 0)
+ flag(vps_default_ptl_dpb_hrd_max_tid_flag);
+ else
+ infer(vps_default_ptl_dpb_hrd_max_tid_flag, 1);
+
+ if (current->vps_max_layers_minus1 > 0)
+ flag(vps_all_independent_layers_flag);
+ else
+ infer(vps_all_independent_layers_flag, 1);
+
+ for (i = 0; i <= current->vps_max_layers_minus1; i++) {
+ ubs(6, vps_layer_id[i], 1, i);
+ if (i > 0 && current->vps_layer_id[i] <= current->vps_layer_id[i - 1]) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "vps_layer_id[%d](%d) should > vps_layer_id[%d](%d).\n",
+ i, current->vps_layer_id[i], i - 1,
+ current->vps_layer_id[i - 1]);
+ return AVERROR_INVALIDDATA;
+ }
+ if (i > 0 && !current->vps_all_independent_layers_flag) {
+ flags(vps_independent_layer_flag[i], 1, i);
+ if (!current->vps_independent_layer_flag[i]) {
+ flags(vps_max_tid_ref_present_flag[i], 1, i);
+ for (j = 0; j < i; j++) {
+ flags(vps_direct_ref_layer_flag[i][j], 2, i, j);
+ if (current->vps_max_tid_ref_present_flag[i] &&
+ current->vps_direct_ref_layer_flag[i][j]) {
+ ubs(3, vps_max_tid_il_ref_pics_plus1[i][j], 2, i, j);
+ } else {
+ infer(vps_max_tid_il_ref_pics_plus1[i][j],
+ current->vps_max_sublayers_minus1 + 1);
+ }
+ }
+ } else {
+ for (j = 0; j < i; j++) {
+ infer(vps_direct_ref_layer_flag[i][j], 0);
+ }
+ }
+ } else {
+ infer(vps_independent_layer_flag[i], 1);
+ for (j = 0; j < i; j++) {
+ infer(vps_direct_ref_layer_flag[i][j], 0);
+ }
+ }
+ }
+
+ if (current->vps_max_layers_minus1 > 0) {
+ if (current->vps_all_independent_layers_flag)
+ flag(vps_each_layer_is_an_ols_flag);
+ else
+ infer(vps_each_layer_is_an_ols_flag, 0);
+ if (!current->vps_each_layer_is_an_ols_flag) {
+ if (!current->vps_all_independent_layers_flag)
+ ub(2, vps_ols_mode_idc);
+ else
+ infer(vps_ols_mode_idc, 2);
+ if (current->vps_ols_mode_idc == 2) {
+ ub(8, vps_num_output_layer_sets_minus2);
+ for (i = 1; i <= current->vps_num_output_layer_sets_minus2 + 1;
+ i++)
+ for (j = 0; j <= current->vps_max_layers_minus1; j++)
+ flags(vps_ols_output_layer_flag[i][j], 2, i, j);
+ }
+ ols_mode_idc = current->vps_ols_mode_idc;
+ } else {
+ ols_mode_idc = 4;
+ }
+ if (ols_mode_idc == 4 || ols_mode_idc == 0 || ols_mode_idc == 1)
+ total_num_olss = current->vps_max_layers_minus1 + 1;
+ else if (ols_mode_idc = 2)
+ total_num_olss = current->vps_num_output_layer_sets_minus2 + 2;
+ else
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "ols_mode_idc == 3, patch welcome");
+ u(8, vps_num_ptls_minus1, 0, total_num_olss - 1);
+ } else {
+ infer(vps_each_layer_is_an_ols_flag, 1);
+ infer(vps_num_ptls_minus1, 0);
+ }
+ {
+ //calc NumMultiLayerOlss
+ int m;
+ uint8_t dependency_flag[VVC_MAX_LAYERS][VVC_MAX_LAYERS];
+ uint16_t num_output_layers_in_ols[VVC_MAX_TOTAL_NUM_OLSS];
+ uint8_t num_sub_layers_in_layer_in_ols[VVC_MAX_TOTAL_NUM_OLSS][VVC_MAX_TOTAL_NUM_OLSS];
+ uint8_t output_layer_idx[VVC_MAX_TOTAL_NUM_OLSS][VVC_MAX_LAYERS];
+
+ //7.4.3.3 vps_direct_ref_layer_flag section
+ for (i = 0; i <= current->vps_max_layers_minus1; i++) {
+ for (j = 0; j <= current->vps_max_layers_minus1; j++) {
+ dependency_flag[i][j] = current->vps_direct_ref_layer_flag[i][j];
+ for (k = 0; k < i; k++) {
+ if (current->vps_direct_ref_layer_flag[i][k] &&
+ dependency_flag[k][j])
+ dependency_flag[i][j] = 1;
+ }
+ }
+ }
+ for (i = 0; i <= current->vps_max_layers_minus1; i++) {
+ int r;
+ for (j = 0, r = 0; j <= current->vps_max_layers_minus1; j++) {
+ if (dependency_flag[i][j])
+ reference_layer_idx[i][r++] = j;
+ }
+ num_ref_layers[i] = r;
+ }
+
+ //7.4.3.3 vps_ols_output_layer_flag section
+ num_output_layers_in_ols[0] = 1;
+ num_sub_layers_in_layer_in_ols[0][0] =
+ current->vps_ptl_max_tid[current->vps_ols_ptl_idx[0]] + 1;
+ for (i = 1; i < total_num_olss; i++) {
+ if (ols_mode_idc == 4 || ols_mode_idc == 0) {
+ num_output_layers_in_ols[i] = 1;
+ if (current->vps_each_layer_is_an_ols_flag) {
+ num_sub_layers_in_layer_in_ols[i][0] =
+ current->vps_ptl_max_tid[current->vps_ols_ptl_idx[i]] + 1;
+ } else {
+ num_sub_layers_in_layer_in_ols[i][i] =
+ current->vps_ptl_max_tid[current->vps_ols_ptl_idx[i]] + 1;
+ for (k = i - 1; k >= 0; k--) {
+ num_sub_layers_in_layer_in_ols[i][k] = 0;
+ for (m = k + 1; m <= i; m++) {
+ uint8_t max_sublayer_needed =
+ FFMIN(num_sub_layers_in_layer_in_ols[i][m],
+ current->vps_max_tid_il_ref_pics_plus1[m][k]);
+ if (current->vps_direct_ref_layer_flag[m][k] &&
+ num_sub_layers_in_layer_in_ols[i][k] < max_sublayer_needed)
+ num_sub_layers_in_layer_in_ols[i][k] = max_sublayer_needed;
+ }
+ }
+ }
+ } else if (current->vps_ols_mode_idc == 1) {
+ num_output_layers_in_ols[i] = i + 1;
+ for (j = 0; j < num_output_layers_in_ols[i]; j++) {
+ num_sub_layers_in_layer_in_ols[i][j] =
+ current->vps_ptl_max_tid[current->vps_ols_ptl_idx[i]] + 1;
+ }
+ } else if (current->vps_ols_mode_idc == 2) {
+ uint8_t highest_included_layer = 0;
+ for (j = 0; j <= current->vps_max_layers_minus1; j++) {
+ layer_included_in_ols_flag[i][j] = 0;
+ num_sub_layers_in_layer_in_ols[i][j] = 0;
+ }
+ for (k = 0, j = 0; k <= current->vps_max_layers_minus1; k++) {
+ if (current->vps_ols_output_layer_flag[i][k]) {
+ layer_included_in_ols_flag[i][k] = 1;
+ highest_included_layer = k;
+ output_layer_idx[i][j] = k;
+ num_sub_layers_in_layer_in_ols[i][k] =
+ current->vps_ptl_max_tid[current->
+ vps_ols_ptl_idx[i]] + 1;
+ j++;
+ }
+ }
+ num_output_layers_in_ols[i] = j;
+ for (j = 0; j < num_output_layers_in_ols[i]; j++) {
+ int idx = output_layer_idx[i][j];
+ for (k = 0; k < num_ref_layers[idx]; k++) {
+ if (!layer_included_in_ols_flag[i][reference_layer_idx[idx][k]])
+ layer_included_in_ols_flag[i][reference_layer_idx[idx][k]] = 1;
+ }
+ }
+ for (k = highest_included_layer - 1; k >= 0; k--) {
+ if (layer_included_in_ols_flag[i][k] &&
+ !current->vps_ols_output_layer_flag[i][k]) {
+ for (m = k + 1; m <= highest_included_layer; m++) {
+ uint8_t max_sublayer_needed =
+ FFMIN(num_sub_layers_in_layer_in_ols[i][m],
+ current->vps_max_tid_il_ref_pics_plus1[m][k]);
+ if (current->vps_direct_ref_layer_flag[m][k] &&
+ layer_included_in_ols_flag[i][m] &&
+ num_sub_layers_in_layer_in_ols[i][k] <
+ max_sublayer_needed)
+ num_sub_layers_in_layer_in_ols[i][k] =
+ max_sublayer_needed;
+ }
+ }
+ }
+ }
+ }
+ for (i = 1; i < total_num_olss; i++) {
+ int num_layers_in_ols = 0;
+ if (current->vps_each_layer_is_an_ols_flag) {
+ num_layers_in_ols = 1;
+ } else if (current->vps_ols_mode_idc == 0 ||
+ current->vps_ols_mode_idc == 1) {
+ num_layers_in_ols = i + 1;
+ } else if (current->vps_ols_mode_idc == 2) {
+ for (k = 0, j = 0; k <= current->vps_max_layers_minus1; k++) {
+ if (layer_included_in_ols_flag[i][k])
+ j++;
+ num_layers_in_ols = j;
+ }
+ }
+ if (num_layers_in_ols > 1) {
+ num_multi_layer_olss++;
+ }
+ }
+ }
+
+ for (i = 0; i <= current->vps_num_ptls_minus1; i++) {
+ if (i > 0)
+ flags(vps_pt_present_flag[i], 1, i);
+ else
+ infer(vps_pt_present_flag[i], 1);
+
+ if (!current->vps_default_ptl_dpb_hrd_max_tid_flag)
+ us(3, vps_ptl_max_tid[i], 0, current->vps_max_sublayers_minus1, 1, i);
+ else
+ infer(vps_ptl_max_tid[i], current->vps_max_sublayers_minus1);
+ }
+ while (byte_alignment(rw) != 0)
+ fixed(1, vps_ptl_alignment_zero_bit, 0);
+ for (i = 0; i <= current->vps_num_ptls_minus1; i++) {
+ CHECK(FUNC(profile_tier_level) (ctx, rw,
+ current->vps_profile_tier_level + i,
+ current->vps_pt_present_flag[i],
+ current->vps_ptl_max_tid[i]));
+ }
+ for (i = 0; i < total_num_olss; i++) {
+ if (current->vps_num_ptls_minus1 > 0 &&
+ current->vps_num_ptls_minus1 + 1 != total_num_olss) {
+ us(8, vps_ols_ptl_idx[i], 0, current->vps_num_ptls_minus1, 1, i);
+ } else if (current->vps_num_ptls_minus1 == 0) {
+ infer(vps_ols_ptl_idx[i], 0);
+ } else {
+ infer(vps_ols_ptl_idx[i], i);
+ }
+ }
+
+ if (!current->vps_each_layer_is_an_ols_flag) {
+ uint16_t vps_num_dpb_params;
+ ue(vps_num_dpb_params_minus1, 0, num_multi_layer_olss - 1);
+ if (current->vps_each_layer_is_an_ols_flag)
+ vps_num_dpb_params = 0;
+ else
+ vps_num_dpb_params = current->vps_num_dpb_params_minus1 + 1;
+
+ if (current->vps_max_sublayers_minus1 > 0)
+ flag(vps_sublayer_dpb_params_present_flag);
+ else
+ infer(vps_sublayer_dpb_params_present_flag, 0);
+
+ for (i = 0; i < vps_num_dpb_params; i++) {
+ if (!current->vps_default_ptl_dpb_hrd_max_tid_flag)
+ us(3, vps_dpb_max_tid[i], 0, current->vps_max_sublayers_minus1,
+ 1, i);
+ else
+ infer(vps_dpb_max_tid[i], current->vps_max_sublayers_minus1);
+ CHECK(FUNC(dpb_parameters) (ctx, rw, current->vps_dpb_params + i,
+ current->vps_dpb_max_tid[i],
+ current->
+ vps_sublayer_dpb_params_present_flag));
+ }
+ for (i = 0; i < num_multi_layer_olss; i++) {
+ ues(vps_ols_dpb_pic_width[i], 0, UINT16_MAX, 1, i);
+ ues(vps_ols_dpb_pic_height[i], 0, UINT16_MAX, 1, i);
+ ubs(2, vps_ols_dpb_chroma_format[i], 1, i);
+ ues(vps_ols_dpb_bitdepth_minus8[i], 0, 2, 1, i);
+ if (vps_num_dpb_params > 1
+ && vps_num_dpb_params != num_multi_layer_olss)
+ ues(vps_ols_dpb_params_idx[i], 0, vps_num_dpb_params - 1, 1, i);
+ else if (vps_num_dpb_params == 1)
+ infer(vps_ols_dpb_params_idx[i], 0);
+ else
+ infer(vps_ols_dpb_params_idx[i], i);
+ }
+ flag(vps_timing_hrd_params_present_flag);
+ if (current->vps_timing_hrd_params_present_flag) {
+ CHECK(FUNC(general_timing_hrd_parameters) (ctx, rw,
+ ¤t->
+ vps_general_timing_hrd_parameters));
+ if (current->vps_max_sublayers_minus1 > 0)
+ flag(vps_sublayer_cpb_params_present_flag);
+ else
+ infer(vps_sublayer_cpb_params_present_flag, 0);
+ ue(vps_num_ols_timing_hrd_params_minus1, 0,
+ num_multi_layer_olss - 1);
+ for (i = 0; i <= current->vps_num_ols_timing_hrd_params_minus1; i++) {
+ uint8_t first_sublayer;
+ if (!current->vps_default_ptl_dpb_hrd_max_tid_flag)
+ us(3, vps_hrd_max_tid[i], 0,
+ current->vps_max_sublayers_minus1, 1, i);
+ else
+ infer(vps_hrd_max_tid[i],
+ current->vps_max_sublayers_minus1);
+ first_sublayer = current->vps_sublayer_cpb_params_present_flag ?
+ 0 : current->vps_hrd_max_tid[i];
+ CHECK(FUNC(ols_timing_hrd_parameters)
+ (ctx, rw, ¤t->vps_ols_timing_hrd_parameters,
+ first_sublayer, current->vps_max_sublayers_minus1,
+ ¤t->vps_general_timing_hrd_parameters));
+
+ }
+ if (current->vps_num_ols_timing_hrd_params_minus1 > 0 &&
+ current->vps_num_ols_timing_hrd_params_minus1 + 1 !=
+ num_multi_layer_olss) {
+ for (i = 0; i < num_multi_layer_olss; i++) {
+ ues(vps_ols_timing_hrd_idx[i], 0,
+ current->vps_num_ols_timing_hrd_params_minus1, 1, i);
+ }
+ } else if (current->vps_num_ols_timing_hrd_params_minus1 == 0) {
+ for (i = 0; i < num_multi_layer_olss; i++)
+ infer(vps_ols_timing_hrd_idx[i], 0);
+ } else {
+ for (i = 0; i < num_multi_layer_olss; i++)
+ infer(vps_ols_timing_hrd_idx[i], i);
+ }
+ }
+ }
+
+ flag(vps_extension_flag);
+ if (current->vps_extension_flag)
+ CHECK(FUNC(extension_data) (ctx, rw, ¤t->extension_data));
+ CHECK(FUNC(rbsp_trailing_bits) (ctx, rw));
+
+ return 0;
+}
+
+
+static int FUNC(sps)(CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawSPS *current)
+{
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+ int err, i, j;
+ unsigned int ctb_log2_size_y, min_cb_log2_size_y,
+ min_qt_log2_size_intra_y, min_qt_log2_size_inter_y,
+ ctb_size_y, max_num_merge_cand, tmp_width_val, tmp_height_val;
+ uint8_t qp_bd_offset;
+
+ static const uint8_t h266_sub_width_c[] = {
+ 1, 2, 2, 1
+ };
+ static const uint8_t h266_sub_height_c[] = {
+ 1, 2, 1, 1
+ };
+
+ HEADER("Sequence Parameter Set");
+
+ CHECK(FUNC(nal_unit_header) (ctx, rw,
+ ¤t->nal_unit_header, VVC_SPS_NUT));
+
+ ub(4, sps_seq_parameter_set_id);
+ ub(4, sps_video_parameter_set_id);
+ if (current->sps_video_parameter_set_id == 0 && !h266->vps_ref[0]) {
+ H266RawVPS *vps;
+ AVBufferRef *ref = av_buffer_allocz(sizeof(H266RawVPS));
+ if (!ref) {
+ return AVERROR(ENOMEM);
+ }
+ vps = (H266RawVPS *) ref->data;
+ vps->vps_max_layers_minus1 = 0;
+ vps->vps_independent_layer_flag[0] = 1;
+ vps->vps_layer_id[0] = current->nal_unit_header.nuh_layer_id;
+ h266->vps_ref[0] = ref;
+ h266->vps[0] = vps;
+ }
+
+ u(3, sps_max_sublayers_minus1, 0, VVC_MAX_SUBLAYERS - 1);
+ u(2, sps_chroma_format_idc, 0, 3);
+ u(2, sps_log2_ctu_size_minus5, 0, 3);
+ ctb_log2_size_y = current->sps_log2_ctu_size_minus5 + 5;
+ ctb_size_y = 1 << ctb_log2_size_y;
+
+ flag(sps_ptl_dpb_hrd_params_present_flag);
+ if (current->sps_ptl_dpb_hrd_params_present_flag) {
+ CHECK(FUNC(profile_tier_level) (ctx, rw, ¤t->profile_tier_level,
+ 1, current->sps_max_sublayers_minus1));
+ }
+ flag(sps_gdr_enabled_flag);
+ flag(sps_ref_pic_resampling_enabled_flag);
+ if (current->sps_ref_pic_resampling_enabled_flag)
+ flag(sps_res_change_in_clvs_allowed_flag);
+ else
+ infer(sps_res_change_in_clvs_allowed_flag, 0);
+
+ ue(sps_pic_width_max_in_luma_samples, 1, VVC_MAX_WIDTH);
+ ue(sps_pic_height_max_in_luma_samples, 1, VVC_MAX_HEIGHT);
+
+ flag(sps_conformance_window_flag);
+ if (current->sps_conformance_window_flag) {
+ uint8_t sub_width_c = h266_sub_width_c[current->sps_chroma_format_idc];
+ uint8_t sub_height_c = h266_sub_height_c[current->sps_chroma_format_idc];
+ uint16_t width = current->sps_pic_width_max_in_luma_samples / sub_width_c;
+ uint16_t height = current->sps_pic_height_max_in_luma_samples / sub_height_c;
+ ue(sps_conf_win_left_offset, 0, width);
+ ue(sps_conf_win_right_offset, 0, width - current->sps_conf_win_left_offset);
+ ue(sps_conf_win_top_offset, 0, height);
+ ue(sps_conf_win_bottom_offset, 0, height - current->sps_conf_win_top_offset);
+ } else {
+ infer(sps_conf_win_left_offset, 0);
+ infer(sps_conf_win_right_offset, 0);
+ infer(sps_conf_win_top_offset, 0);
+ infer(sps_conf_win_bottom_offset, 0);
+ }
+
+ tmp_width_val = AV_CEIL_RSHIFT(current->sps_pic_width_max_in_luma_samples,
+ ctb_log2_size_y);
+ tmp_height_val = AV_CEIL_RSHIFT(current->sps_pic_height_max_in_luma_samples,
+ ctb_log2_size_y);
+
+ flag(sps_subpic_info_present_flag);
+ if (current->sps_subpic_info_present_flag) {
+ ue(sps_num_subpics_minus1, 1, VVC_MAX_SLICES - 1);
+ if (current->sps_num_subpics_minus1 > 0) {
+ flag(sps_independent_subpics_flag);
+ flag(sps_subpic_same_size_flag);
+ }
+
+ if (current->sps_num_subpics_minus1 > 0) {
+ int wlen = av_ceil_log2(tmp_width_val);
+ int hlen = av_ceil_log2(tmp_height_val);
+ if (current->sps_pic_width_max_in_luma_samples > ctb_size_y)
+ ubs(wlen, sps_subpic_width_minus1[0], 1, 0);
+ else
+ infer(sps_subpic_width_minus1[i], tmp_width_val - 1);
+ if (current->sps_pic_height_max_in_luma_samples > ctb_size_y)
+ ubs(hlen, sps_subpic_height_minus1[0], 1, 0);
+ else
+ infer(sps_subpic_height_minus1[0], tmp_height_val);
+ if (!current->sps_independent_subpics_flag) {
+ flags(sps_subpic_treated_as_pic_flag[0], 1, 0);
+ flags(sps_loop_filter_across_subpic_enabled_flag[0], 1, 0);
+ } else {
+ infer(sps_subpic_treated_as_pic_flag[0], 1);
+ infer(sps_loop_filter_across_subpic_enabled_flag[0], 1);
+ }
+ for (i = 1; i <= current->sps_num_subpics_minus1; i++) {
+ if (!current->sps_subpic_same_size_flag) {
+ if (current->sps_pic_width_max_in_luma_samples > ctb_size_y)
+ ubs(wlen, sps_subpic_ctu_top_left_x[i], 1, i);
+ else
+ infer(sps_subpic_ctu_top_left_x[i], 0);
+ if (current->sps_pic_height_max_in_luma_samples >
+ ctb_size_y)
+ ubs(hlen, sps_subpic_ctu_top_left_y[i], 1, i);
+ else
+ infer(sps_subpic_ctu_top_left_y[i], 0);
+ if (i < current->sps_num_subpics_minus1 &&
+ current->sps_pic_width_max_in_luma_samples >
+ ctb_size_y) {
+ ubs(wlen, sps_subpic_width_minus1[i], 1, i);
+ } else {
+ infer(sps_subpic_width_minus1[i],
+ tmp_width_val -
+ current->sps_subpic_ctu_top_left_x[i] - 1);
+ }
+ if (i < current->sps_num_subpics_minus1 &&
+ current->sps_pic_height_max_in_luma_samples >
+ ctb_size_y) {
+ ubs(hlen, sps_subpic_height_minus1[i], 1, i);
+ } else {
+ infer(sps_subpic_height_minus1[i],
+ tmp_height_val -
+ current->sps_subpic_ctu_top_left_y[i] - 1);
+ }
+ } else {
+ int num_subpic_cols = tmp_width_val /
+ (current->sps_subpic_width_minus1[0] + 1);
+ infer(sps_subpic_ctu_top_left_x[i],
+ (i % num_subpic_cols) *
+ (current->sps_subpic_width_minus1[0] + 1));
+ infer(sps_subpic_ctu_top_left_y[i],
+ (i / num_subpic_cols) *
+ (current->sps_subpic_height_minus1[0] + 1));
+ infer(sps_subpic_width_minus1[i],
+ current->sps_subpic_width_minus1[0]);
+ infer(sps_subpic_height_minus1[i],
+ current->sps_subpic_height_minus1[0]);
+ }
+ if (!current->sps_independent_subpics_flag) {
+ flags(sps_subpic_treated_as_pic_flag[i], 1, i);
+ flags(sps_loop_filter_across_subpic_enabled_flag[i], 1, i);
+ } else {
+ infer(sps_subpic_treated_as_pic_flag[i], 1);
+ infer(sps_loop_filter_across_subpic_enabled_flag[i], 0);
+ }
+ }
+ ue(sps_subpic_id_len_minus1, 0, 15);
+ if ((1 << (current->sps_subpic_id_len_minus1 + 1)) <
+ current->sps_num_subpics_minus1 + 1) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "sps_subpic_id_len_minus1(%d) is too small\n",
+ current->sps_subpic_id_len_minus1);
+ return AVERROR_INVALIDDATA;
+ }
+ flag(sps_subpic_id_mapping_explicitly_signalled_flag);
+ if (current->sps_subpic_id_mapping_explicitly_signalled_flag) {
+ flag(sps_subpic_id_mapping_present_flag);
+ if (current->sps_subpic_id_mapping_present_flag) {
+ for (i = 0; i <= current->sps_num_subpics_minus1; i++) {
+ ubs(current->sps_subpic_id_len_minus1 + 1,
+ sps_subpic_id[i], 1, i);
+ }
+ }
+ }
+ } else {
+ infer(sps_subpic_ctu_top_left_x[0], 0);
+ infer(sps_subpic_ctu_top_left_y[0], 0);
+ infer(sps_subpic_width_minus1[0], tmp_width_val - 1);
+ infer(sps_subpic_height_minus1[0], tmp_height_val - 1);
+ }
+ } else {
+ infer(sps_num_subpics_minus1, 0);
+ infer(sps_independent_subpics_flag, 1);
+ infer(sps_subpic_same_size_flag, 0);
+ infer(sps_subpic_id_mapping_explicitly_signalled_flag, 0);
+ infer(sps_subpic_ctu_top_left_x[0], 0);
+ infer(sps_subpic_ctu_top_left_y[0], 0);
+ infer(sps_subpic_width_minus1[0], tmp_width_val - 1);
+ infer(sps_subpic_height_minus1[0], tmp_height_val - 1);
+ }
+
+
+ ue(sps_bitdepth_minus8, 0, 2);
+ qp_bd_offset = 6 * current->sps_bitdepth_minus8;
+
+ flag(sps_entropy_coding_sync_enabled_flag);
+ flag(sps_entry_point_offsets_present_flag);
+
+ u(4, sps_log2_max_pic_order_cnt_lsb_minus4, 0, 12);
+ flag(sps_poc_msb_cycle_flag);
+ if (current->sps_poc_msb_cycle_flag)
+ ue(sps_poc_msb_cycle_len_minus1,
+ 0, 32 - current->sps_log2_max_pic_order_cnt_lsb_minus4 - 5);
+
+ u(2, sps_num_extra_ph_bytes, 0, 2);
+ for (i = 0; i < (current->sps_num_extra_ph_bytes * 8); i++) {
+ flags(sps_extra_ph_bit_present_flag[i], 1, i);
+ }
+
+ u(2, sps_num_extra_sh_bytes, 0, 2);
+ for (i = 0; i < (current->sps_num_extra_sh_bytes * 8); i++) {
+ flags(sps_extra_sh_bit_present_flag[i], 1, i);
+ }
+
+ if (current->sps_ptl_dpb_hrd_params_present_flag) {
+ if (current->sps_max_sublayers_minus1 > 0)
+ flag(sps_sublayer_dpb_params_flag);
+ else
+ infer(sps_sublayer_dpb_params_flag, 0);
+ CHECK(FUNC(dpb_parameters) (ctx, rw, ¤t->sps_dpb_params,
+ current->sps_max_sublayers_minus1,
+ current->sps_sublayer_dpb_params_flag));
+ }
+
+ ue(sps_log2_min_luma_coding_block_size_minus2,
+ 0, FFMIN(4, current->sps_log2_ctu_size_minus5 + 3));
+ min_cb_log2_size_y =
+ current->sps_log2_min_luma_coding_block_size_minus2 + 2;
+
+ flag(sps_partition_constraints_override_enabled_flag);
+
+ ue(sps_log2_diff_min_qt_min_cb_intra_slice_luma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_cb_log2_size_y);
+ min_qt_log2_size_intra_y =
+ current->sps_log2_diff_min_qt_min_cb_intra_slice_luma +
+ min_cb_log2_size_y;
+
+ ue(sps_max_mtt_hierarchy_depth_intra_slice_luma,
+ 0, 2 * (ctb_log2_size_y - min_cb_log2_size_y));
+
+ if (current->sps_max_mtt_hierarchy_depth_intra_slice_luma != 0) {
+ ue(sps_log2_diff_max_bt_min_qt_intra_slice_luma,
+ 0, ctb_log2_size_y - min_qt_log2_size_intra_y);
+ ue(sps_log2_diff_max_tt_min_qt_intra_slice_luma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_intra_y);
+ } else {
+ infer(sps_log2_diff_max_bt_min_qt_intra_slice_luma, 0);
+ infer(sps_log2_diff_max_tt_min_qt_intra_slice_luma, 0);
+ }
+
+ if (current->sps_chroma_format_idc != 0) {
+ flag(sps_qtbtt_dual_tree_intra_flag);
+ } else {
+ infer(sps_qtbtt_dual_tree_intra_flag, 0);
+ }
+
+ if (current->sps_qtbtt_dual_tree_intra_flag) {
+ ue(sps_log2_diff_min_qt_min_cb_intra_slice_chroma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_cb_log2_size_y);
+ ue(sps_max_mtt_hierarchy_depth_intra_slice_chroma,
+ 0, 2 * (ctb_log2_size_y - min_cb_log2_size_y));
+ if (current->sps_max_mtt_hierarchy_depth_intra_slice_chroma != 0) {
+ unsigned int min_qt_log2_size_intra_c =
+ current->sps_log2_diff_min_qt_min_cb_intra_slice_chroma +
+ min_cb_log2_size_y;
+ ue(sps_log2_diff_max_bt_min_qt_intra_slice_chroma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_intra_c);
+ ue(sps_log2_diff_max_tt_min_qt_intra_slice_chroma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_intra_c);
+ }
+ } else {
+ infer(sps_log2_diff_min_qt_min_cb_intra_slice_chroma, 0);
+ infer(sps_max_mtt_hierarchy_depth_intra_slice_chroma, 0);
+ }
+ if (current->sps_max_mtt_hierarchy_depth_intra_slice_chroma == 0) {
+ infer(sps_log2_diff_max_bt_min_qt_intra_slice_chroma, 0);
+ infer(sps_log2_diff_max_tt_min_qt_intra_slice_chroma, 0);
+ }
+
+ ue(sps_log2_diff_min_qt_min_cb_inter_slice,
+ 0, FFMIN(6, ctb_log2_size_y) - min_cb_log2_size_y);
+ min_qt_log2_size_inter_y =
+ current->sps_log2_diff_min_qt_min_cb_inter_slice + min_cb_log2_size_y;
+
+ ue(sps_max_mtt_hierarchy_depth_inter_slice,
+ 0, 2 * (ctb_log2_size_y - min_cb_log2_size_y));
+ if (current->sps_max_mtt_hierarchy_depth_inter_slice != 0) {
+ ue(sps_log2_diff_max_bt_min_qt_inter_slice,
+ 0, ctb_log2_size_y - min_qt_log2_size_inter_y);
+ ue(sps_log2_diff_max_tt_min_qt_inter_slice,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_inter_y);
+ } else {
+ infer(sps_log2_diff_max_bt_min_qt_inter_slice, 0);
+ infer(sps_log2_diff_max_tt_min_qt_inter_slice, 0);
+ }
+
+ if (ctb_size_y > 32)
+ flag(sps_max_luma_transform_size_64_flag);
+ else
+ infer(sps_max_luma_transform_size_64_flag, 0);
+
+ flag(sps_transform_skip_enabled_flag);
+ if (current->sps_transform_skip_enabled_flag) {
+ ue(sps_log2_transform_skip_max_size_minus2, 0, 3);
+ flag(sps_bdpcm_enabled_flag);
+ }
+
+ flag(sps_mts_enabled_flag);
+ if (current->sps_mts_enabled_flag) {
+ flag(sps_explicit_mts_intra_enabled_flag);
+ flag(sps_explicit_mts_inter_enabled_flag);
+ } else {
+ infer(sps_explicit_mts_intra_enabled_flag, 0);
+ infer(sps_explicit_mts_inter_enabled_flag, 0);
+ }
+
+ flag(sps_lfnst_enabled_flag);
+
+ if (current->sps_chroma_format_idc != 0) {
+ uint8_t num_qp_tables;
+ flag(sps_joint_cbcr_enabled_flag);
+ flag(sps_same_qp_table_for_chroma_flag);
+ num_qp_tables = current->sps_same_qp_table_for_chroma_flag ?
+ 1 : (current->sps_joint_cbcr_enabled_flag ? 3 : 2);
+ for (i = 0; i < num_qp_tables; i++) {
+ ses(sps_qp_table_start_minus26[i], -26 - qp_bd_offset, 36, 1, i);
+ ues(sps_num_points_in_qp_table_minus1[i],
+ 0, 36 - current->sps_qp_table_start_minus26[i], 1, i);
+ for (j = 0; j <= current->sps_num_points_in_qp_table_minus1[i]; j++) {
+ uint8_t max = MAX_UINT_BITS(8);
+ ues(sps_delta_qp_in_val_minus1[i][j], 0, max, 2, i, j);
+ ues(sps_delta_qp_diff_val[i][j], 0, max, 2, i, j);
+ }
+ }
+ } else {
+ infer(sps_joint_cbcr_enabled_flag, 0);
+ infer(sps_same_qp_table_for_chroma_flag, 0);
+ }
+
+ flag(sps_sao_enabled_flag);
+ flag(sps_alf_enabled_flag);
+ if (current->sps_alf_enabled_flag && current->sps_chroma_format_idc)
+ flag(sps_ccalf_enabled_flag);
+ else
+ infer(sps_ccalf_enabled_flag, 0);
+ flag(sps_lmcs_enabled_flag);
+ flag(sps_weighted_pred_flag);
+ flag(sps_weighted_bipred_flag);
+ flag(sps_long_term_ref_pics_flag);
+ if (current->sps_video_parameter_set_id > 0)
+ flag(sps_inter_layer_prediction_enabled_flag);
+ else
+ infer(sps_inter_layer_prediction_enabled_flag, 0);
+ flag(sps_idr_rpl_present_flag);
+ flag(sps_rpl1_same_as_rpl0_flag);
+
+ for (i = 0; i < (current->sps_rpl1_same_as_rpl0_flag ? 1 : 2); i++) {
+ ues(sps_num_ref_pic_lists[i], 0, VVC_MAX_REF_PIC_LISTS, 1, i);
+ for (j = 0; j < current->sps_num_ref_pic_lists[i]; j++)
+ CHECK(FUNC(ref_pic_list_struct) (ctx, rw,
+ ¤t->
+ sps_ref_pic_list_struct[i][j], i,
+ j, current));
+ }
+
+ if (current->sps_rpl1_same_as_rpl0_flag) {
+ current->sps_num_ref_pic_lists[1] = current->sps_num_ref_pic_lists[0];
+ for (j = 0; j < current->sps_num_ref_pic_lists[0]; j++)
+ memcpy(¤t->sps_ref_pic_list_struct[1][j],
+ ¤t->sps_ref_pic_list_struct[0][j],
+ sizeof(current->sps_ref_pic_list_struct[0][j]));
+ }
+
+ flag(sps_ref_wraparound_enabled_flag);
+
+ flag(sps_temporal_mvp_enabled_flag);
+ if (current->sps_temporal_mvp_enabled_flag)
+ flag(sps_sbtmvp_enabled_flag);
+ else
+ infer(sps_sbtmvp_enabled_flag, 0);
+
+ flag(sps_amvr_enabled_flag);
+ flag(sps_bdof_enabled_flag);
+ if (current->sps_bdof_enabled_flag)
+ flag(sps_bdof_control_present_in_ph_flag);
+ else
+ infer(sps_bdof_control_present_in_ph_flag, 0);
+
+ flag(sps_smvd_enabled_flag);
+ flag(sps_dmvr_enabled_flag);
+ if (current->sps_dmvr_enabled_flag)
+ flag(sps_dmvr_control_present_in_ph_flag);
+ else
+ infer(sps_dmvr_control_present_in_ph_flag, 0);
+
+ flag(sps_mmvd_enabled_flag);
+ if (current->sps_mmvd_enabled_flag)
+ flag(sps_mmvd_fullpel_only_enabled_flag);
+ else
+ infer(sps_mmvd_fullpel_only_enabled_flag, 0);
+
+ ue(sps_six_minus_max_num_merge_cand, 0, 5);
+ max_num_merge_cand = 6 - current->sps_six_minus_max_num_merge_cand;
+
+ flag(sps_sbt_enabled_flag);
+
+ flag(sps_affine_enabled_flag);
+ if (current->sps_affine_enabled_flag) {
+ ue(sps_five_minus_max_num_subblock_merge_cand,
+ 0, 5 - current->sps_sbtmvp_enabled_flag);
+ flag(sps_6param_affine_enabled_flag);
+ if (current->sps_amvr_enabled_flag)
+ flag(sps_affine_amvr_enabled_flag);
+ else
+ infer(sps_affine_amvr_enabled_flag, 0);
+ flag(sps_affine_prof_enabled_flag);
+ if (current->sps_affine_prof_enabled_flag)
+ flag(sps_prof_control_present_in_ph_flag);
+ else
+ infer(sps_prof_control_present_in_ph_flag, 0);
+ } else {
+ infer(sps_6param_affine_enabled_flag, 0);
+ infer(sps_affine_amvr_enabled_flag, 0);
+ infer(sps_affine_prof_enabled_flag, 0);
+ infer(sps_prof_control_present_in_ph_flag, 0);
+ }
+
+ flag(sps_bcw_enabled_flag);
+ flag(sps_ciip_enabled_flag);
+
+ if (max_num_merge_cand >= 2) {
+ flag(sps_gpm_enabled_flag);
+ if (current->sps_gpm_enabled_flag && max_num_merge_cand >= 3)
+ ue(sps_max_num_merge_cand_minus_max_num_gpm_cand,
+ 0, max_num_merge_cand - 2);
+ } else {
+ infer(sps_gpm_enabled_flag, 0);
+ }
+
+ ue(sps_log2_parallel_merge_level_minus2, 0, ctb_log2_size_y - 2);
+
+ flag(sps_isp_enabled_flag);
+ flag(sps_mrl_enabled_flag);
+ flag(sps_mip_enabled_flag);
+
+ if (current->sps_chroma_format_idc != 0)
+ flag(sps_cclm_enabled_flag);
+ else
+ infer(sps_cclm_enabled_flag, 0);
+ if (current->sps_chroma_format_idc == 1) {
+ flag(sps_chroma_horizontal_collocated_flag);
+ flag(sps_chroma_vertical_collocated_flag);
+ } else {
+ infer(sps_chroma_horizontal_collocated_flag, 1);
+ infer(sps_chroma_vertical_collocated_flag, 1);
+ }
+
+ flag(sps_palette_enabled_flag);
+ if (current->sps_chroma_format_idc == 3 &&
+ !current->sps_max_luma_transform_size_64_flag)
+ flag(sps_act_enabled_flag);
+ else
+ infer(sps_act_enabled_flag, 0);
+ if (current->sps_transform_skip_enabled_flag ||
+ current->sps_palette_enabled_flag)
+ ue(sps_min_qp_prime_ts, 0, 8);
+
+ flag(sps_ibc_enabled_flag);
+ if (current->sps_ibc_enabled_flag)
+ ue(sps_six_minus_max_num_ibc_merge_cand, 0, 5);
+
+ flag(sps_ladf_enabled_flag);
+ if (current->sps_ladf_enabled_flag) {
+ ub(2, sps_num_ladf_intervals_minus2);
+ se(sps_ladf_lowest_interval_qp_offset, -63, 63);
+ for (i = 0; i < current->sps_num_ladf_intervals_minus2 + 1; i++) {
+ ses(sps_ladf_qp_offset[i], -63, 63, 1, i);
+ ues(sps_ladf_delta_threshold_minus1[i],
+ 0, (2 << (8 + current->sps_bitdepth_minus8)) - 3, 1, i);
+ }
+ }
+
+ flag(sps_explicit_scaling_list_enabled_flag);
+ if (current->sps_lfnst_enabled_flag &&
+ current->sps_explicit_scaling_list_enabled_flag)
+ flag(sps_scaling_matrix_for_lfnst_disabled_flag);
+
+ if (current->sps_act_enabled_flag &&
+ current->sps_explicit_scaling_list_enabled_flag)
+ flag(sps_scaling_matrix_for_alternative_colour_space_disabled_flag);
+ else
+ infer(sps_scaling_matrix_for_alternative_colour_space_disabled_flag, 0);
+ if (current->sps_scaling_matrix_for_alternative_colour_space_disabled_flag)
+ flag(sps_scaling_matrix_designated_colour_space_flag);
+
+ flag(sps_dep_quant_enabled_flag);
+ flag(sps_sign_data_hiding_enabled_flag);
+
+ flag(sps_virtual_boundaries_enabled_flag);
+ if (current->sps_virtual_boundaries_enabled_flag) {
+ flag(sps_virtual_boundaries_present_flag);
+ if (current->sps_virtual_boundaries_present_flag) {
+ ue(sps_num_ver_virtual_boundaries,
+ 0, current->sps_pic_width_max_in_luma_samples <= 8 ? 0 : 3);
+ for (i = 0; i < current->sps_num_ver_virtual_boundaries; i++)
+ ues(sps_virtual_boundary_pos_x_minus1[i],
+ 0, (current->sps_pic_width_max_in_luma_samples + 7) / 8 - 2,
+ 1, i);
+ ue(sps_num_hor_virtual_boundaries,
+ 0, current->sps_pic_height_max_in_luma_samples <= 8 ? 0 : 3);
+ for (i = 0; i < current->sps_num_hor_virtual_boundaries; i++)
+ ues(sps_virtual_boundary_pos_y_minus1[i],
+ 0, (current->sps_pic_height_max_in_luma_samples + 7) /
+ 8 - 2, 1, i);
+ }
+ } else {
+ infer(sps_virtual_boundaries_present_flag, 0);
+ infer(sps_num_ver_virtual_boundaries, 0);
+ infer(sps_num_hor_virtual_boundaries, 0);
+ }
+
+ if (current->sps_ptl_dpb_hrd_params_present_flag) {
+ flag(sps_timing_hrd_params_present_flag);
+ if (current->sps_timing_hrd_params_present_flag) {
+ uint8_t first_sublayer;
+ CHECK(FUNC(general_timing_hrd_parameters) (ctx, rw,
+ ¤t->sps_general_timing_hrd_parameters));
+ if (current->sps_max_sublayers_minus1 > 0)
+ flag(sps_sublayer_cpb_params_present_flag);
+ else
+ infer(sps_sublayer_cpb_params_present_flag, 0);
+ first_sublayer = current->sps_sublayer_cpb_params_present_flag ?
+ 0 : current->sps_max_sublayers_minus1;
+ CHECK(FUNC(ols_timing_hrd_parameters) (ctx, rw,
+ ¤t->sps_ols_timing_hrd_parameters, first_sublayer,
+ current->sps_max_sublayers_minus1,
+ ¤t->sps_general_timing_hrd_parameters));
+ }
+ }
+
+ flag(sps_field_seq_flag);
+ flag(sps_vui_parameters_present_flag);
+ if (current->sps_vui_parameters_present_flag) {
+ ue(sps_vui_payload_size_minus1, 0, 1023);
+ while (byte_alignment(rw) != 0)
+ fixed(1, sps_vui_alignment_zero_bit, 0);
+ CHECK(FUNC(vui_payload) (ctx, rw, ¤t->vui,
+ current->sps_vui_payload_size_minus1 + 1,
+ current->sps_chroma_format_idc));
+ } else {
+ CHECK(FUNC(vui_parameters_default) (ctx, rw, ¤t->vui));
+ }
+ flag(sps_extension_flag);
+ if (current->sps_extension_flag)
+ CHECK(FUNC(extension_data) (ctx, rw, ¤t->extension_data));
+
+ CHECK(FUNC(rbsp_trailing_bits) (ctx, rw));
+
+ return 0;
+}
+
+static int FUNC(pps) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawPPS *current)
+{
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+ const H266RawSPS *sps;
+ int err, i;
+ unsigned int min_cb_size_y, divisor, ctb_size_y,
+ pic_width_in_ctbs_y, pic_height_in_ctbs_y;
+ uint8_t sub_width_c, sub_height_c, qp_bd_offset;
+
+ static const uint8_t h266_sub_width_c[] = {
+ 1, 2, 2, 1
+ };
+ static const uint8_t h266_sub_height_c[] = {
+ 1, 2, 1, 1
+ };
+
+ HEADER("Picture Parameter Set");
+
+ CHECK(FUNC(nal_unit_header) (ctx, rw,
+ ¤t->nal_unit_header, VVC_PPS_NUT));
+
+ ub(6, pps_pic_parameter_set_id);
+ ub(4, pps_seq_parameter_set_id);
+ sps = h266->sps[current->pps_seq_parameter_set_id];
+ if (!sps) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "SPS id %d not available.\n",
+ current->pps_seq_parameter_set_id);
+ return AVERROR_INVALIDDATA;
+ }
+
+ flag(pps_mixed_nalu_types_in_pic_flag);
+ ue(pps_pic_width_in_luma_samples,
+ 1, sps->sps_pic_width_max_in_luma_samples);
+ ue(pps_pic_height_in_luma_samples,
+ 1, sps->sps_pic_height_max_in_luma_samples);
+
+ min_cb_size_y = 1 << (sps->sps_log2_min_luma_coding_block_size_minus2 + 2);
+ divisor = FFMAX(min_cb_size_y, 8);
+ if (current->pps_pic_width_in_luma_samples % divisor ||
+ current->pps_pic_height_in_luma_samples % divisor) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Invalid dimensions: %ux%u not divisible "
+ "by %u, MinCbSizeY = %u.\n",
+ current->pps_pic_width_in_luma_samples,
+ current->pps_pic_height_in_luma_samples, divisor, min_cb_size_y);
+ return AVERROR_INVALIDDATA;
+ }
+ if (!sps->sps_res_change_in_clvs_allowed_flag &&
+ (current->pps_pic_width_in_luma_samples !=
+ sps->sps_pic_width_max_in_luma_samples ||
+ current->pps_pic_height_in_luma_samples !=
+ sps->sps_pic_height_max_in_luma_samples)) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Resoltuion change is not allowed, "
+ "in max resolution (%ux%u) mismatched with pps(%ux%u).\n",
+ sps->sps_pic_width_max_in_luma_samples,
+ sps->sps_pic_height_max_in_luma_samples,
+ current->pps_pic_width_in_luma_samples,
+ current->pps_pic_height_in_luma_samples);
+ return AVERROR_INVALIDDATA;
+ }
+
+ ctb_size_y = 1 << (sps->sps_log2_ctu_size_minus5 + 5);
+ if (sps->sps_ref_wraparound_enabled_flag) {
+ if ((ctb_size_y / min_cb_size_y + 1) >
+ (current->pps_pic_width_in_luma_samples / min_cb_size_y - 1)) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Invalid width(%u), ctb_size_y = %u, min_cb_size_y = %u.\n",
+ current->pps_pic_width_in_luma_samples,
+ ctb_size_y, min_cb_size_y);
+ return AVERROR_INVALIDDATA;
+ }
+ }
+
+ flag(pps_conformance_window_flag);
+ if (current->pps_pic_width_in_luma_samples ==
+ sps->sps_pic_width_max_in_luma_samples &&
+ current->pps_pic_height_in_luma_samples ==
+ sps->sps_pic_height_max_in_luma_samples &&
+ current->pps_conformance_window_flag) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Conformance window flag should not true.\n");
+ return AVERROR_INVALIDDATA;
+ }
+
+ sub_width_c = h266_sub_width_c[sps->sps_chroma_format_idc];
+ sub_height_c = h266_sub_height_c[sps->sps_chroma_format_idc];
+ if (current->pps_conformance_window_flag) {
+ ue(pps_conf_win_left_offset, 0, current->pps_pic_width_in_luma_samples);
+ ue(pps_conf_win_right_offset,
+ 0, current->pps_pic_width_in_luma_samples);
+ ue(pps_conf_win_top_offset, 0, current->pps_pic_height_in_luma_samples);
+ ue(pps_conf_win_bottom_offset,
+ 0, current->pps_pic_height_in_luma_samples);
+ if (sub_width_c *
+ (current->pps_conf_win_left_offset +
+ current->pps_conf_win_right_offset) >=
+ current->pps_pic_width_in_luma_samples ||
+ sub_height_c *
+ (current->pps_conf_win_top_offset +
+ current->pps_conf_win_bottom_offset) >=
+ current->pps_pic_height_in_luma_samples) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Invalid pps conformance window: (%u, %u, %u, %u), "
+ "resolution is %ux%u, sub wxh is %ux%u.\n",
+ current->pps_conf_win_left_offset,
+ current->pps_conf_win_right_offset,
+ current->pps_conf_win_top_offset,
+ current->pps_conf_win_bottom_offset,
+ current->pps_pic_width_in_luma_samples,
+ current->pps_pic_height_in_luma_samples,
+ sub_width_c, sub_height_c);
+ return AVERROR_INVALIDDATA;
+ }
+ } else {
+ if (current->pps_pic_width_in_luma_samples ==
+ sps->sps_pic_width_max_in_luma_samples &&
+ current->pps_pic_height_in_luma_samples ==
+ sps->sps_pic_height_max_in_luma_samples) {
+ infer(pps_conf_win_left_offset, sps->sps_conf_win_left_offset);
+ infer(pps_conf_win_right_offset, sps->sps_conf_win_right_offset);
+ infer(pps_conf_win_top_offset, sps->sps_conf_win_top_offset);
+ infer(pps_conf_win_bottom_offset, sps->sps_conf_win_bottom_offset);
+ } else {
+ infer(pps_conf_win_left_offset, 0);
+ infer(pps_conf_win_right_offset, 0);
+ infer(pps_conf_win_top_offset, 0);
+ infer(pps_conf_win_bottom_offset, 0);
+ }
+
+ }
+
+ flag(pps_scaling_window_explicit_signalling_flag);
+ if (!sps->sps_ref_pic_resampling_enabled_flag &&
+ current->pps_scaling_window_explicit_signalling_flag) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Invalid data: sps_ref_pic_resampling_enabled_flag is false, "
+ "but pps_scaling_window_explicit_signalling_flag is true.\n");
+ return AVERROR_INVALIDDATA;
+ }
+ if (current->pps_scaling_window_explicit_signalling_flag) {
+ se(pps_scaling_win_left_offset,
+ -current->pps_pic_width_in_luma_samples * 15 / sub_width_c,
+ current->pps_pic_width_in_luma_samples / sub_width_c);
+ se(pps_scaling_win_right_offset,
+ -current->pps_pic_width_in_luma_samples * 15 / sub_width_c,
+ current->pps_pic_width_in_luma_samples / sub_width_c);
+ se(pps_scaling_win_top_offset,
+ -current->pps_pic_height_in_luma_samples * 15 / sub_height_c,
+ current->pps_pic_height_in_luma_samples / sub_height_c);
+ se(pps_scaling_win_bottom_offset,
+ -current->pps_pic_height_in_luma_samples * 15 / sub_height_c,
+ current->pps_pic_height_in_luma_samples / sub_height_c);
+ } else {
+ infer(pps_scaling_win_left_offset, current->pps_conf_win_left_offset);
+ infer(pps_scaling_win_right_offset, current->pps_conf_win_right_offset);
+ infer(pps_scaling_win_top_offset, current->pps_conf_win_top_offset);
+ infer(pps_scaling_win_bottom_offset, current->pps_conf_win_bottom_offset);
+ }
+
+ flag(pps_output_flag_present_flag);
+ flag(pps_no_pic_partition_flag);
+ flag(pps_subpic_id_mapping_present_flag);
+
+ if (current->pps_subpic_id_mapping_present_flag) {
+ if (!current->pps_no_pic_partition_flag) {
+ ue(pps_num_subpics_minus1,
+ sps->sps_num_subpics_minus1, sps->sps_num_subpics_minus1);
+ } else {
+ infer(pps_num_subpics_minus1, 0);
+ }
+ ue(pps_subpic_id_len_minus1, sps->sps_subpic_id_len_minus1,
+ sps->sps_subpic_id_len_minus1);
+ for (i = 0; i <= current->pps_num_subpics_minus1; i++) {
+ ubs(sps->sps_subpic_id_len_minus1 + 1, pps_subpic_id[i], 1, i);
+ }
+ }
+
+ pic_width_in_ctbs_y = AV_CEIL_RSHIFT
+ (current->pps_pic_width_in_luma_samples, (sps->sps_log2_ctu_size_minus5 + 5));
+ pic_height_in_ctbs_y = AV_CEIL_RSHIFT(
+ current->pps_pic_height_in_luma_samples,(sps->sps_log2_ctu_size_minus5 + 5));
+ if (!current->pps_no_pic_partition_flag) {
+ unsigned int exp_tile_width = 0, exp_tile_height = 0;
+ unsigned int unified_size, remaining_size;
+
+ u(2, pps_log2_ctu_size_minus5,
+ sps->sps_log2_ctu_size_minus5, sps->sps_log2_ctu_size_minus5);
+ ue(pps_num_exp_tile_columns_minus1,
+ 0, FFMIN(pic_width_in_ctbs_y - 1, VVC_MAX_TILE_COLUMNS - 1));
+ ue(pps_num_exp_tile_rows_minus1,
+ 0, FFMIN(pic_height_in_ctbs_y - 1, VVC_MAX_TILE_ROWS - 1));
+
+ for (i = 0; i <= current->pps_num_exp_tile_columns_minus1; i++) {
+ ues(pps_tile_column_width_minus1[i],
+ 0, pic_width_in_ctbs_y - exp_tile_width - 1, 1, i);
+ exp_tile_width += current->pps_tile_column_width_minus1[i] + 1;
+ }
+ for (i = 0; i <= current->pps_num_exp_tile_rows_minus1; i++) {
+ ues(pps_tile_row_height_minus1[i],
+ 0, pic_height_in_ctbs_y - 1, 1, i);
+ exp_tile_height += current->pps_tile_row_height_minus1[i] + 1;
+ }
+
+ remaining_size = pic_width_in_ctbs_y;
+ for (i = 0; i <= current->pps_num_exp_tile_columns_minus1; i++) {
+ if (current->pps_tile_column_width_minus1[i] >= remaining_size) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Tile column width(%d) exceeds picture width\n",i);
+ return AVERROR_INVALIDDATA;
+ }
+ remaining_size -= (current->pps_tile_column_width_minus1[i] + 1);
+ }
+ unified_size = (i == 0 ? pic_width_in_ctbs_y :
+ (current->pps_tile_column_width_minus1[i - 1] + 1));
+ while (remaining_size > 0) {
+ if (current->num_tile_columns > VVC_MAX_TILE_COLUMNS) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "NumTileColumns(%d) > than VVC_MAX_TILE_COLUMNS(%d)\n",
+ current->num_tile_columns, VVC_MAX_TILE_COLUMNS);
+ return AVERROR_INVALIDDATA;
+ }
+ unified_size = FFMIN(remaining_size, unified_size);
+ current->pps_tile_column_width_minus1[i] = unified_size - 1;
+ remaining_size -= unified_size;
+ i++;
+ }
+ current->num_tile_columns = i;
+ if (current->num_tile_columns > VVC_MAX_TILE_COLUMNS) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "NumTileColumns(%d) > than VVC_MAX_TILE_COLUMNS(%d)\n",
+ current->num_tile_columns, VVC_MAX_TILE_COLUMNS);
+ return AVERROR_INVALIDDATA;
+ }
+
+ remaining_size = pic_height_in_ctbs_y;
+ for (i = 0; i <= current->pps_num_exp_tile_rows_minus1; i++) {
+ if (current->pps_tile_row_height_minus1[i] >= remaining_size) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Tile row height(%d) exceeds picture height\n",i);
+ return AVERROR_INVALIDDATA;
+ }
+ remaining_size -= (current->pps_tile_row_height_minus1[i] + 1);
+ }
+ unified_size = (i == 0 ? pic_height_in_ctbs_y :
+ (current->pps_tile_row_height_minus1[i - 1] + 1));
+
+ while (remaining_size > 0) {
+ unified_size = FFMIN(remaining_size, unified_size);
+ current->pps_tile_row_height_minus1[i] = unified_size - 1;
+ remaining_size -= unified_size;
+ i++;
+ }
+ current->num_tile_rows=i;
+ if (current->num_tile_rows > VVC_MAX_TILE_ROWS) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "NumTileRows(%d) > than VVC_MAX_TILE_ROWS(%d)\n",
+ current->num_tile_rows, VVC_MAX_TILE_ROWS);
+ return AVERROR_INVALIDDATA;
+ }
+
+ current->num_tiles_in_pic = current->num_tile_columns *
+ current->num_tile_rows;
+ if (current->num_tiles_in_pic > VVC_MAX_TILES_PER_AU) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "NumTilesInPic(%d) > than VVC_MAX_TILES_PER_AU(%d)\n",
+ current->num_tiles_in_pic, VVC_MAX_TILES_PER_AU);
+ return AVERROR_INVALIDDATA;
+ }
+
+ if (current->num_tiles_in_pic > 1) {
+ flag(pps_loop_filter_across_tiles_enabled_flag);
+ flag(pps_rect_slice_flag);
+ } else {
+ infer(pps_loop_filter_across_tiles_enabled_flag, 0);
+ infer(pps_rect_slice_flag, 1);
+ }
+ if (current->pps_rect_slice_flag)
+ flag(pps_single_slice_per_subpic_flag);
+ else
+ infer(pps_single_slice_per_subpic_flag, 1);
+ if (current->pps_rect_slice_flag &&
+ !current->pps_single_slice_per_subpic_flag) {
+ int j;
+ uint16_t tile_idx = 0, tile_x, tile_y, ctu_x, ctu_y;
+ uint16_t slice_top_left_ctu_x[VVC_MAX_SLICES];
+ uint16_t slice_top_left_ctu_y[VVC_MAX_SLICES];
+ ue(pps_num_slices_in_pic_minus1, 0, VVC_MAX_SLICES - 1);
+ if (current->pps_num_slices_in_pic_minus1 > 1)
+ flag(pps_tile_idx_delta_present_flag);
+ else
+ infer(pps_tile_idx_delta_present_flag, 0);
+ for (i = 0; i < current->pps_num_slices_in_pic_minus1; i++) {
+ tile_x = tile_idx % current->num_tile_columns;
+ tile_y = tile_idx / current->num_tile_columns;
+ if (tile_x != current->num_tile_columns - 1) {
+ ues(pps_slice_width_in_tiles_minus1[i],
+ 0, current->num_tile_columns - 1, 1, i);
+ } else {
+ infer(pps_slice_width_in_tiles_minus1[i], 0);
+ }
+ if (tile_y != current->num_tile_rows - 1 &&
+ (current->pps_tile_idx_delta_present_flag || tile_x == 0)) {
+ ues(pps_slice_height_in_tiles_minus1[i],
+ 0, current->num_tile_rows - 1, 1, i);
+ } else {
+ if (tile_y == current->num_tile_rows - 1)
+ infer(pps_slice_height_in_tiles_minus1[i], 0);
+ else
+ infer(pps_slice_height_in_tiles_minus1[i],
+ current->pps_slice_height_in_tiles_minus1[i - 1]);
+ }
+
+ ctu_x = ctu_y = 0;
+ for (j = 0; j < tile_x; j++) {
+ ctu_x += current->pps_tile_column_width_minus1[j] + 1;
+ }
+ for (j = 0; j < tile_y; j++) {
+ ctu_y += current->pps_tile_row_height_minus1[j] + 1;
+ }
+ if (current->pps_slice_width_in_tiles_minus1[i] == 0 &&
+ current->pps_slice_height_in_tiles_minus1[i] == 0 &&
+ current->pps_tile_row_height_minus1[tile_y] > 0) {
+ int num_slices_in_tile,
+ uniform_slice_height, remaining_height_in_ctbs_y;
+ remaining_height_in_ctbs_y =
+ current->pps_tile_row_height_minus1[tile_y] + 1;
+ ues(pps_num_exp_slices_in_tile[i],
+ 0, current->pps_tile_row_height_minus1[tile_y], 1, i);
+ if (current->pps_num_exp_slices_in_tile[i] == 0) {
+ num_slices_in_tile = 1;
+ slice_top_left_ctu_x[i] = ctu_x;
+ slice_top_left_ctu_y[i] = ctu_y;
+ } else {
+ uint16_t slice_height_in_ctus;
+ for (j = 0; j < current->pps_num_exp_slices_in_tile[i];
+ j++) {
+ ues(pps_exp_slice_height_in_ctus_minus1[i][j], 0,
+ current->pps_tile_row_height_minus1[tile_y], 2,
+ i, j);
+ slice_height_in_ctus =
+ current->
+ pps_exp_slice_height_in_ctus_minus1[i][j] + 1;
+
+ current->slice_height_in_ctus[i + j] =
+ slice_height_in_ctus;
+ slice_top_left_ctu_x[i + j] = ctu_x;
+ slice_top_left_ctu_y[i + j] = ctu_y;
+ ctu_y += slice_height_in_ctus;
+
+ remaining_height_in_ctbs_y -= slice_height_in_ctus;
+ }
+ uniform_slice_height = 1 +
+ (j == 0 ? current->pps_tile_row_height_minus1[tile_y] :
+ current->pps_exp_slice_height_in_ctus_minus1[i][j-1]);
+ while (remaining_height_in_ctbs_y > uniform_slice_height) {
+ current->slice_height_in_ctus[i + j] =
+ uniform_slice_height;
+ slice_top_left_ctu_x[i + j] = ctu_x;
+ slice_top_left_ctu_y[i + j] = ctu_y;
+ ctu_y += uniform_slice_height;
+
+ remaining_height_in_ctbs_y -= uniform_slice_height;
+ j++;
+ }
+ if (remaining_height_in_ctbs_y > 0) {
+ current->slice_height_in_ctus[i + j] =
+ remaining_height_in_ctbs_y;
+ slice_top_left_ctu_x[i + j] = ctu_x;
+ slice_top_left_ctu_y[i + j] = ctu_y;
+ j++;
+ }
+ num_slices_in_tile = j;
+ }
+ i += num_slices_in_tile - 1;
+ } else {
+ uint16_t height = 0;
+ infer(pps_num_exp_slices_in_tile[i], 0);
+ for (j = 0;
+ j <= current->pps_slice_height_in_tiles_minus1[i];
+ j++) {
+ height +=
+ current->pps_tile_row_height_minus1[tile_y + j] + 1;
+ }
+ current->slice_height_in_ctus[i] = height;
+
+ slice_top_left_ctu_x[i] = ctu_x;
+ slice_top_left_ctu_y[i] = ctu_y;
+ }
+ if (i < current->pps_num_slices_in_pic_minus1) {
+ if (current->pps_tile_idx_delta_present_flag) {
+ ses(pps_tile_idx_delta_val[i],
+ -current->num_tiles_in_pic + 1,
+ current->num_tiles_in_pic - 1, 1, i);
+ if (current->pps_tile_idx_delta_val[i] == 0) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "pps_tile_idx_delta_val[i] shall not be equal to 0.\n");
+ }
+ tile_idx += current->pps_tile_idx_delta_val[i];
+ } else {
+ infer(pps_tile_idx_delta_val[i], 0);
+ tile_idx +=
+ current->pps_slice_width_in_tiles_minus1[i] + 1;
+ if (tile_idx % current->num_tile_columns == 0) {
+ tile_idx +=
+ current->pps_slice_height_in_tiles_minus1[i] *
+ current->num_tile_columns;
+ }
+ }
+ }
+ }
+ if (i == current->pps_num_slices_in_pic_minus1) {
+ uint16_t height = 0;
+
+ tile_x = tile_idx % current->num_tile_columns;
+ tile_y = tile_idx / current->num_tile_columns;
+
+ ctu_x = 0, ctu_y = 0;
+ for (j = 0; j < tile_x; j++) {
+ ctu_x += current->pps_tile_column_width_minus1[j] + 1;
+ }
+ for (j = 0; j < tile_y; j++) {
+ ctu_y += current->pps_tile_row_height_minus1[j] + 1;
+ }
+ slice_top_left_ctu_x[i] = ctu_x;
+ slice_top_left_ctu_y[i] = ctu_y;
+
+ current->pps_slice_width_in_tiles_minus1[i] =
+ current->num_tile_columns - tile_x - 1;
+ current->pps_slice_height_in_tiles_minus1[i] =
+ current->num_tile_rows - tile_y - 1;
+
+ for (j = 0; j <= current->pps_slice_height_in_tiles_minus1[i];
+ j++) {
+ height +=
+ current->pps_tile_row_height_minus1[tile_y + j] + 1;
+ }
+ current->slice_height_in_ctus[i] = height;
+
+ infer(pps_num_exp_slices_in_tile[i], 0);
+ }
+ //now, we got all slice information, let's resolve NumSlicesInSubpic
+ for (i = 0; i <= sps->sps_num_subpics_minus1; i++) {
+ current->num_slices_in_subpic[i] = 0;
+ for (j = 0; j <= current->pps_num_slices_in_pic_minus1; j++) {
+ uint16_t pos_x = 0, pos_y = 0;
+ pos_x = slice_top_left_ctu_x[j];
+ pos_y = slice_top_left_ctu_y[j];
+ if ((pos_x >= sps->sps_subpic_ctu_top_left_x[i]) &&
+ (pos_x <
+ sps->sps_subpic_ctu_top_left_x[i] +
+ sps->sps_subpic_width_minus1[i] + 1) &&
+ (pos_y >= sps->sps_subpic_ctu_top_left_y[i]) &&
+ (pos_y < sps->sps_subpic_ctu_top_left_y[i] +
+ sps->sps_subpic_height_minus1[i] + 1)) {
+ current->num_slices_in_subpic[i]++;
+ }
+ }
+ }
+ } else {
+ if (current->pps_no_pic_partition_flag)
+ infer(pps_num_slices_in_pic_minus1, 0);
+ else if (current->pps_single_slice_per_subpic_flag)
+ infer(pps_num_slices_in_pic_minus1,
+ sps->sps_num_subpics_minus1);
+ // else?
+ }
+ if (!current->pps_rect_slice_flag ||
+ current->pps_single_slice_per_subpic_flag ||
+ current->pps_num_slices_in_pic_minus1 > 0)
+ flag(pps_loop_filter_across_slices_enabled_flag);
+ else
+ infer(pps_loop_filter_across_slices_enabled_flag, 0);
+ } else {
+ infer(pps_num_exp_tile_columns_minus1, 0);
+ infer(pps_tile_column_width_minus1[0], pic_width_in_ctbs_y - 1);
+ infer(pps_num_exp_tile_rows_minus1, 0);
+ infer(pps_tile_row_height_minus1[0], pic_height_in_ctbs_y - 1);
+ infer(num_tile_columns, 1);
+ infer(num_tile_rows, 1);
+ infer(num_tiles_in_pic, 1);
+ }
+
+ flag(pps_cabac_init_present_flag);
+ for (i = 0; i < 2; i++)
+ ues(pps_num_ref_idx_default_active_minus1[i], 0, 14, 1, i);
+ flag(pps_rpl1_idx_present_flag);
+ flag(pps_weighted_pred_flag);
+ flag(pps_weighted_bipred_flag);
+ flag(pps_ref_wraparound_enabled_flag);
+ if (current->pps_ref_wraparound_enabled_flag) {
+ ue(pps_pic_width_minus_wraparound_offset,
+ 0, (current->pps_pic_width_in_luma_samples / min_cb_size_y)
+ - (ctb_size_y / min_cb_size_y) - 2);
+ }
+
+ qp_bd_offset = 6 * sps->sps_bitdepth_minus8;
+ se(pps_init_qp_minus26, -(26 + qp_bd_offset), 37);
+ flag(pps_cu_qp_delta_enabled_flag);
+ flag(pps_chroma_tool_offsets_present_flag);
+ if (current->pps_chroma_tool_offsets_present_flag) {
+ se(pps_cb_qp_offset, -12, 12);
+ se(pps_cr_qp_offset, -12, 12);
+ flag(pps_joint_cbcr_qp_offset_present_flag);
+ if (current->pps_joint_cbcr_qp_offset_present_flag)
+ se(pps_joint_cbcr_qp_offset_value, -12, 12);
+ else
+ infer(pps_joint_cbcr_qp_offset_value, 0);
+ flag(pps_slice_chroma_qp_offsets_present_flag);
+ flag(pps_cu_chroma_qp_offset_list_enabled_flag);
+ if (current->pps_cu_chroma_qp_offset_list_enabled_flag) {
+ ue(pps_chroma_qp_offset_list_len_minus1, 0, 5);
+ for (i = 0; i <= current->pps_chroma_qp_offset_list_len_minus1; i++) {
+ ses(pps_cb_qp_offset_list[i], -12, 12, 1, i);
+ ses(pps_cr_qp_offset_list[i], -12, 12, 1, i);
+ if (current->pps_joint_cbcr_qp_offset_present_flag)
+ ses(pps_joint_cbcr_qp_offset_list[i], -12, 12, 1, i);
+ else
+ infer(pps_joint_cbcr_qp_offset_list[i], 0);
+ }
+ }
+ } else {
+ infer(pps_cb_qp_offset, 0);
+ infer(pps_cr_qp_offset, 0);
+ infer(pps_joint_cbcr_qp_offset_present_flag, 0);
+ infer(pps_joint_cbcr_qp_offset_value, 0);
+ infer(pps_slice_chroma_qp_offsets_present_flag, 0);
+ infer(pps_cu_chroma_qp_offset_list_enabled_flag, 0);
+ }
+ flag(pps_deblocking_filter_control_present_flag);
+ if (current->pps_deblocking_filter_control_present_flag) {
+ flag(pps_deblocking_filter_override_enabled_flag);
+ flag(pps_deblocking_filter_disabled_flag);
+ if (!current->pps_no_pic_partition_flag &&
+ current->pps_deblocking_filter_override_enabled_flag)
+ flag(pps_dbf_info_in_ph_flag);
+ else
+ infer(pps_dbf_info_in_ph_flag, 0);
+ if (!current->pps_deblocking_filter_disabled_flag) {
+ se(pps_luma_beta_offset_div2, -12, 12);
+ se(pps_luma_tc_offset_div2, -12, 12);
+ if (current->pps_chroma_tool_offsets_present_flag) {
+ se(pps_cb_beta_offset_div2, -12, 12);
+ se(pps_cb_tc_offset_div2, -12, 12);
+ se(pps_cr_beta_offset_div2, -12, 12);
+ se(pps_cr_tc_offset_div2, -12, 12);
+ } else {
+ infer(pps_cb_beta_offset_div2,
+ current->pps_luma_beta_offset_div2);
+ infer(pps_cb_tc_offset_div2, current->pps_luma_tc_offset_div2);
+ infer(pps_cr_beta_offset_div2,
+ current->pps_luma_beta_offset_div2);
+ infer(pps_cr_tc_offset_div2, current->pps_luma_tc_offset_div2);
+ }
+ } else {
+ infer(pps_luma_beta_offset_div2, 0);
+ infer(pps_luma_tc_offset_div2, 0);
+ infer(pps_cb_beta_offset_div2, 0);
+ infer(pps_cb_tc_offset_div2, 0);
+ infer(pps_cr_beta_offset_div2, 0);
+ infer(pps_cr_tc_offset_div2, 0);
+ }
+ } else {
+ infer(pps_deblocking_filter_override_enabled_flag, 0);
+ infer(pps_deblocking_filter_disabled_flag, 0);
+ infer(pps_dbf_info_in_ph_flag, 0);
+ infer(pps_luma_beta_offset_div2, 0);
+ infer(pps_luma_tc_offset_div2, 0);
+ infer(pps_cb_beta_offset_div2, 0);
+ infer(pps_cb_tc_offset_div2, 0);
+ infer(pps_cr_beta_offset_div2, 0);
+ infer(pps_cr_tc_offset_div2, 0);
+ }
+
+ if (!current->pps_no_pic_partition_flag) {
+ flag(pps_rpl_info_in_ph_flag);
+ flag(pps_sao_info_in_ph_flag);
+ flag(pps_alf_info_in_ph_flag);
+ if ((current->pps_weighted_pred_flag ||
+ current->pps_weighted_bipred_flag) &&
+ current->pps_rpl_info_in_ph_flag)
+ flag(pps_wp_info_in_ph_flag);
+ flag(pps_qp_delta_info_in_ph_flag);
+ }
+ flag(pps_picture_header_extension_present_flag);
+ flag(pps_slice_header_extension_present_flag);
+
+ flag(pps_extension_flag);
+ if (current->pps_extension_flag)
+ CHECK(FUNC(extension_data) (ctx, rw, ¤t->extension_data));
+
+ CHECK(FUNC(rbsp_trailing_bits) (ctx, rw));
+ return 0;
+}
+
+static int FUNC(aud) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawAUD *current)
+{
+ int err;
+
+ HEADER("Access Unit Delimiter");
+
+ CHECK(FUNC(nal_unit_header) (ctx, rw,
+ ¤t->nal_unit_header, VVC_AUD_NUT));
+
+ flag(aud_irap_or_gdr_flag);
+ u(3, aud_pic_type, 0, 2);
+
+ CHECK(FUNC(rbsp_trailing_bits) (ctx, rw));
+ return 0;
+}
+
+static int FUNC(pred_weight_table) (CodedBitstreamContext *ctx, RWContext *rw,
+ const H266RawSPS *sps,
+ const H266RawPPS *pps,
+ const H266RefPicLists *ref_lists,
+ uint8_t num_ref_idx_active[2],
+ H266RawPredWeightTable *current)
+{
+ int err, i, j;
+ uint8_t num_weights_l0, num_weights_l1;
+ ue(luma_log2_weight_denom, 0, 7);
+ if (sps->sps_chroma_format_idc != 0) {
+ se(delta_chroma_log2_weight_denom,
+ -current->luma_log2_weight_denom,
+ 7 - current->luma_log2_weight_denom);
+ } else {
+ infer(delta_chroma_log2_weight_denom, 0);
+ }
+ if (pps->pps_wp_info_in_ph_flag)
+ ue(num_l0_weights, 0,
+ FFMIN(15, ref_lists->rpl_ref_list[0].num_ref_entries));
+ else
+ infer(num_l0_weights, 0);
+ num_weights_l0 = pps->pps_wp_info_in_ph_flag ?
+ current->num_l0_weights : num_ref_idx_active[0];
+ for (i = 0; i < num_weights_l0; i++) {
+ flags(luma_weight_l0_flag[i], 1, i);
+ }
+ if (sps->sps_chroma_format_idc != 0) {
+ for (i = 0; i < num_weights_l0; i++)
+ flags(chroma_weight_l0_flag[i], 1, i);
+ }
+ for (i = 0; i < num_weights_l0; i++) {
+ if (current->luma_weight_l0_flag[i]) {
+ ses(delta_luma_weight_l0[i], -128, 127, 1, i);
+ ses(luma_offset_l0[i], -128, 127, 1, i);
+ } else {
+ infer(delta_luma_weight_l0[i], 0);
+ infer(luma_offset_l0[i], 0);
+ }
+ if (current->chroma_weight_l0_flag[i]) {
+ for (j = 0; j < 2; j++) {
+ ses(delta_chroma_weight_l0[i][j], -128, 127, 2, i, j);
+ ses(delta_chroma_offset_l0[i][j], -4 * 128, 4 * 127, 2, i, j);
+ }
+ }
+ }
+
+ if (pps->pps_weighted_bipred_flag && pps->pps_wp_info_in_ph_flag &&
+ ref_lists->rpl_ref_list[1].num_ref_entries > 0) {
+ ue(num_l1_weights, 0,
+ FFMIN(15, ref_lists->rpl_ref_list[1].num_ref_entries));
+ }
+ if (!pps->pps_weighted_bipred_flag ||
+ (pps->pps_wp_info_in_ph_flag &&
+ ref_lists->rpl_ref_list[1].num_ref_entries == 0)) {
+ num_weights_l1 = 0;
+ } else if (pps->pps_wp_info_in_ph_flag) {
+ num_weights_l1 = current->num_l1_weights;
+ } else {
+ num_weights_l1 = num_ref_idx_active[1];
+ }
+
+ for (i = 0; i < num_weights_l1; i++)
+ flags(luma_weight_l1_flag[i], 1, i);
+ if (sps->sps_chroma_format_idc != 0) {
+ for (i = 0; i < num_weights_l1; i++)
+ flags(chroma_weight_l1_flag[i], 1, i);
+ }
+ for (i = 0; i < num_weights_l1; i++) {
+ if (current->luma_weight_l1_flag[i]) {
+ ses(delta_luma_weight_l1[i], -128, 127, 1, i);
+ ses(luma_offset_l1[i], -128, 127, 1, i);
+ } else {
+ infer(delta_luma_weight_l1[i], 0);
+ infer(luma_offset_l1[i], 0);
+ }
+ if (current->chroma_weight_l1_flag[i]) {
+ for (j = 0; j < 2; j++) {
+ ses(delta_chroma_weight_l1[i][j], -128, 127, 2, i, j);
+ ses(delta_chroma_offset_l1[i][j], -4 * 128, 4 * 127, 2, i, j);
+ }
+ }
+ }
+ return 0;
+}
+
+static int FUNC(picture_header) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawPH *current){
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+ const H266RawVPS *vps;
+ const H266RawSPS *sps;
+ const H266RawPPS *pps;
+ int err, i;
+ unsigned int ctb_log2_size_y, min_cb_log2_size_y,
+ min_qt_log2_size_intra_y, min_qt_log2_size_inter_y;
+ uint8_t qp_bd_offset;
+
+ flag(ph_gdr_or_irap_pic_flag);
+ flag(ph_non_ref_pic_flag);
+ if (current->ph_gdr_or_irap_pic_flag)
+ flag(ph_gdr_pic_flag);
+ else
+ infer(ph_gdr_pic_flag, 0);
+ flag(ph_inter_slice_allowed_flag);
+ if (current->ph_inter_slice_allowed_flag)
+ flag(ph_intra_slice_allowed_flag);
+ else
+ infer(ph_intra_slice_allowed_flag, 1);
+ ue(ph_pic_parameter_set_id, 0, VVC_MAX_PPS_COUNT - 1);
+ pps = h266->pps[current->ph_pic_parameter_set_id];
+ if (!pps) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "PPS id %d not available.\n",
+ current->ph_pic_parameter_set_id);
+ return AVERROR_INVALIDDATA;
+ }
+ sps = h266->sps[pps->pps_seq_parameter_set_id];
+ if (!sps) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "SPS id %d not available.\n",
+ pps->pps_seq_parameter_set_id);
+ return AVERROR_INVALIDDATA;
+ }
+ vps = h266->vps[sps->sps_video_parameter_set_id];
+ if (!vps) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "VPS id %d not available.\n",
+ sps->sps_video_parameter_set_id);
+ return AVERROR_INVALIDDATA;
+ }
+
+ ub(sps->sps_log2_max_pic_order_cnt_lsb_minus4 + 4, ph_pic_order_cnt_lsb);
+ if (current->ph_gdr_pic_flag)
+ ue(ph_recovery_poc_cnt, 0,
+ 1 << (sps->sps_log2_max_pic_order_cnt_lsb_minus4 + 4));
+
+ for (i = 0; i < sps->sps_num_extra_ph_bytes * 8; i++) {
+ if (sps->sps_extra_ph_bit_present_flag[i])
+ flags(ph_extra_bit[i], 1, i);
+ }
+ if (sps->sps_poc_msb_cycle_flag) {
+ flag(ph_poc_msb_cycle_present_flag);
+ if (current->ph_poc_msb_cycle_present_flag)
+ ub(sps->sps_poc_msb_cycle_len_minus1 + 1, ph_poc_msb_cycle_val);
+ }
+ if (sps->sps_alf_enabled_flag && pps->pps_alf_info_in_ph_flag) {
+ flag(ph_alf_enabled_flag);
+ if (current->ph_alf_enabled_flag) {
+
+ ub(3, ph_num_alf_aps_ids_luma);
+ for (i = 0; i < current->ph_num_alf_aps_ids_luma; i++)
+ ubs(3, ph_alf_aps_id_luma[i], 1, i);
+
+ if (sps->sps_chroma_format_idc != 0) {
+ flag(ph_alf_cb_enabled_flag);
+ flag(ph_alf_cr_enabled_flag);
+ } else {
+ infer(ph_alf_cb_enabled_flag, 0);
+ infer(ph_alf_cr_enabled_flag, 0);
+ }
+
+ if (current->ph_alf_cb_enabled_flag
+ || current->ph_alf_cr_enabled_flag) {
+ ub(3, ph_alf_aps_id_chroma);
+ }
+
+ if (sps->sps_ccalf_enabled_flag) {
+ flag(ph_alf_cc_cb_enabled_flag);
+ if (current->ph_alf_cc_cb_enabled_flag)
+ ub(3, ph_alf_cc_cb_aps_id);
+ flag(ph_alf_cc_cr_enabled_flag);
+ if (current->ph_alf_cc_cr_enabled_flag)
+ ub(3, ph_alf_cc_cr_aps_id);
+ }
+ }
+ } else {
+ infer(ph_alf_enabled_flag, 0);
+ }
+ if (sps->sps_lmcs_enabled_flag) {
+ flag(ph_lmcs_enabled_flag);
+ if (current->ph_lmcs_enabled_flag) {
+ ub(2, ph_lmcs_aps_id);
+ if (sps->sps_chroma_format_idc != 0)
+ flag(ph_chroma_residual_scale_flag);
+ else
+ infer(ph_chroma_residual_scale_flag, 0);
+ }
+ } else {
+ infer(ph_lmcs_enabled_flag, 0);
+ infer(ph_chroma_residual_scale_flag, 0);
+ }
+
+ if (sps->sps_explicit_scaling_list_enabled_flag) {
+ flag(ph_explicit_scaling_list_enabled_flag);
+ if (current->ph_explicit_scaling_list_enabled_flag) {
+ //todo: check the ph_scaling_list_aps_id range, when aps ready
+ ub(3, ph_scaling_list_aps_id);
+ }
+ } else {
+ infer(ph_explicit_scaling_list_enabled_flag, 0);
+ }
+ if (sps->sps_virtual_boundaries_enabled_flag &&
+ !sps->sps_virtual_boundaries_present_flag) {
+ flag(ph_virtual_boundaries_present_flag);
+ if (current->ph_virtual_boundaries_present_flag) {
+ ue(ph_num_ver_virtual_boundaries,
+ 0, pps->pps_pic_width_in_luma_samples <= 8 ? 0 : 3);
+ for (i = 0; i < current->ph_num_ver_virtual_boundaries; i++) {
+ ues(ph_virtual_boundary_pos_x_minus1[i],
+ 0, (pps->pps_pic_width_in_luma_samples + 7) / 8 - 2, 1, i);
+ }
+ ue(ph_num_hor_virtual_boundaries,
+ 0, pps->pps_pic_height_in_luma_samples <= 8 ? 0 : 3);
+ for (i = 0; i < current->ph_num_hor_virtual_boundaries; i++) {
+ ues(ph_virtual_boundary_pos_y_minus1[i],
+ 0, (pps->pps_pic_height_in_luma_samples + 7) / 8 - 2, 1, i);
+ }
+ } else {
+ infer(ph_num_ver_virtual_boundaries, 0);
+ infer(ph_num_hor_virtual_boundaries, 0);
+ }
+ }
+ if (pps->pps_output_flag_present_flag && !current->ph_non_ref_pic_flag)
+ flag(ph_pic_output_flag);
+ else
+ infer(ph_pic_output_flag, 1);
+ if (pps->pps_rpl_info_in_ph_flag) {
+ CHECK(FUNC(ref_pic_lists)
+ (ctx, rw, sps, pps, ¤t->ph_ref_pic_lists));
+ }
+ if (sps->sps_partition_constraints_override_enabled_flag)
+ flag(ph_partition_constraints_override_flag);
+ else
+ infer(ph_partition_constraints_override_flag, 0);
+
+ ctb_log2_size_y = sps->sps_log2_ctu_size_minus5 + 5;
+ min_cb_log2_size_y = sps->sps_log2_min_luma_coding_block_size_minus2 + 2;
+ if (current->ph_intra_slice_allowed_flag) {
+ if (current->ph_partition_constraints_override_flag) {
+ ue(ph_log2_diff_min_qt_min_cb_intra_slice_luma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_cb_log2_size_y);
+ ue(ph_max_mtt_hierarchy_depth_intra_slice_luma,
+ 0, 2 * (ctb_log2_size_y - min_cb_log2_size_y));
+ if (current->ph_max_mtt_hierarchy_depth_intra_slice_luma != 0) {
+ ue(ph_log2_diff_max_bt_min_qt_intra_slice_luma,
+ 0, ctb_log2_size_y - min_qt_log2_size_intra_y);
+ ue(ph_log2_diff_max_tt_min_qt_intra_slice_luma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_intra_y);
+ } else {
+ infer(ph_log2_diff_max_bt_min_qt_intra_slice_luma,
+ sps->sps_log2_diff_max_bt_min_qt_intra_slice_luma);
+ infer(ph_log2_diff_max_tt_min_qt_intra_slice_luma,
+ sps->sps_log2_diff_max_tt_min_qt_intra_slice_luma);
+ }
+ if (sps->sps_qtbtt_dual_tree_intra_flag) {
+ ue(ph_log2_diff_min_qt_min_cb_intra_slice_chroma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_cb_log2_size_y);
+ ue(ph_max_mtt_hierarchy_depth_intra_slice_chroma,
+ 0, 2 * (ctb_log2_size_y - min_cb_log2_size_y));
+ if (sps->sps_max_mtt_hierarchy_depth_intra_slice_chroma != 0) {
+ unsigned int min_qt_log2_size_intra_c =
+ sps->sps_log2_diff_min_qt_min_cb_intra_slice_chroma +
+ min_cb_log2_size_y;
+ ue(ph_log2_diff_max_bt_min_qt_intra_slice_chroma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_intra_c);
+ ue(ph_log2_diff_max_tt_min_qt_intra_slice_chroma,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_intra_c);
+ } else {
+ infer(ph_log2_diff_max_bt_min_qt_intra_slice_chroma,
+ sps->sps_log2_diff_max_bt_min_qt_intra_slice_chroma);
+ infer(ph_log2_diff_max_tt_min_qt_intra_slice_chroma,
+ sps->sps_log2_diff_max_tt_min_qt_intra_slice_chroma);
+ }
+ }
+ } else {
+ infer(ph_log2_diff_min_qt_min_cb_intra_slice_luma,
+ sps->sps_log2_diff_min_qt_min_cb_intra_slice_luma);
+ infer(ph_max_mtt_hierarchy_depth_intra_slice_luma,
+ sps->sps_max_mtt_hierarchy_depth_intra_slice_luma);
+ infer(ph_log2_diff_max_bt_min_qt_intra_slice_luma,
+ sps->sps_log2_diff_max_bt_min_qt_intra_slice_luma);
+ infer(ph_log2_diff_max_tt_min_qt_intra_slice_luma,
+ sps->sps_log2_diff_max_tt_min_qt_intra_slice_luma);
+ infer(ph_log2_diff_min_qt_min_cb_intra_slice_chroma,
+ sps->sps_log2_diff_min_qt_min_cb_intra_slice_chroma);
+ infer(ph_max_mtt_hierarchy_depth_intra_slice_chroma,
+ sps->sps_max_mtt_hierarchy_depth_intra_slice_chroma);
+ infer(ph_log2_diff_max_bt_min_qt_intra_slice_chroma,
+ sps->sps_log2_diff_max_bt_min_qt_intra_slice_chroma);
+ infer(ph_log2_diff_max_tt_min_qt_intra_slice_chroma,
+ sps->sps_log2_diff_max_tt_min_qt_intra_slice_chroma);
+ }
+
+ min_qt_log2_size_intra_y =
+ current->ph_log2_diff_min_qt_min_cb_intra_slice_luma +
+ min_cb_log2_size_y;
+ if (pps->pps_cu_qp_delta_enabled_flag)
+ ue(ph_cu_qp_delta_subdiv_intra_slice, 0,
+ 2 * (ctb_log2_size_y - min_qt_log2_size_intra_y +
+ current->ph_max_mtt_hierarchy_depth_intra_slice_luma));
+ else
+ infer(ph_cu_qp_delta_subdiv_intra_slice, 0);
+
+ if (pps->pps_cu_chroma_qp_offset_list_enabled_flag)
+ ue(ph_cu_chroma_qp_offset_subdiv_intra_slice, 0,
+ 2 * (ctb_log2_size_y - min_qt_log2_size_intra_y +
+ current->ph_max_mtt_hierarchy_depth_intra_slice_luma));
+ else
+ infer(ph_cu_chroma_qp_offset_subdiv_intra_slice, 0);
+ }
+ if (current->ph_inter_slice_allowed_flag) {
+ if (current->ph_partition_constraints_override_flag) {
+ ue(ph_log2_diff_min_qt_min_cb_inter_slice,
+ 0, FFMIN(6, ctb_log2_size_y) - min_cb_log2_size_y);
+ min_qt_log2_size_inter_y =
+ current->ph_log2_diff_min_qt_min_cb_inter_slice +
+ min_cb_log2_size_y;
+ ue(ph_max_mtt_hierarchy_depth_inter_slice, 0,
+ 2 * (ctb_log2_size_y - min_cb_log2_size_y));
+ if (current->ph_max_mtt_hierarchy_depth_inter_slice != 0) {
+ ue(ph_log2_diff_max_bt_min_qt_inter_slice,
+ 0, ctb_log2_size_y - min_qt_log2_size_inter_y);
+ ue(ph_log2_diff_max_tt_min_qt_inter_slice,
+ 0, FFMIN(6, ctb_log2_size_y) - min_qt_log2_size_inter_y);
+ }
+ } else {
+ infer(ph_log2_diff_min_qt_min_cb_inter_slice,
+ sps->sps_log2_diff_min_qt_min_cb_inter_slice);
+ min_qt_log2_size_inter_y =
+ current->ph_log2_diff_min_qt_min_cb_inter_slice +
+ min_cb_log2_size_y;
+ infer(ph_max_mtt_hierarchy_depth_inter_slice,
+ sps->sps_max_mtt_hierarchy_depth_inter_slice);
+ infer(ph_log2_diff_max_bt_min_qt_inter_slice,
+ sps->sps_log2_diff_max_bt_min_qt_inter_slice);
+ infer(ph_log2_diff_max_tt_min_qt_inter_slice,
+ sps->sps_log2_diff_max_tt_min_qt_inter_slice);
+ }
+
+ if (pps->pps_cu_qp_delta_enabled_flag)
+ ue(ph_cu_qp_delta_subdiv_inter_slice, 0,
+ 2 * (ctb_log2_size_y - min_qt_log2_size_inter_y +
+ current->ph_max_mtt_hierarchy_depth_inter_slice));
+ else
+ infer(ph_cu_qp_delta_subdiv_inter_slice, 0);
+
+ if (pps->pps_cu_chroma_qp_offset_list_enabled_flag)
+ ue(ph_cu_chroma_qp_offset_subdiv_inter_slice, 0,
+ 2 * (ctb_log2_size_y - min_qt_log2_size_inter_y +
+ current->ph_max_mtt_hierarchy_depth_inter_slice));
+ else
+ infer(ph_cu_chroma_qp_offset_subdiv_inter_slice, 0);
+ if (sps->sps_temporal_mvp_enabled_flag) {
+ flag(ph_temporal_mvp_enabled_flag);
+ if (current->ph_temporal_mvp_enabled_flag &&
+ pps->pps_rpl_info_in_ph_flag) {
+ if (current->ph_ref_pic_lists.rpl_ref_list[1].num_ref_entries > 0)
+ flag(ph_collocated_from_l0_flag);
+ else
+ infer(ph_collocated_from_l0_flag, 1);
+ if ((current->ph_collocated_from_l0_flag &&
+ current->ph_ref_pic_lists.rpl_ref_list[0].num_ref_entries > 1)
+ || (!current->ph_collocated_from_l0_flag &&
+ current->ph_ref_pic_lists.rpl_ref_list[1].num_ref_entries > 1)) {
+ unsigned int idx =
+ current->ph_collocated_from_l0_flag ? 0 : 1;
+ ue(ph_collocated_ref_idx, 0,
+ current->ph_ref_pic_lists.rpl_ref_list[idx].
+ num_ref_entries - 1);
+ } else {
+ infer(ph_collocated_ref_idx, 0);
+ }
+ }
+ }
+ if (sps->sps_mmvd_fullpel_only_enabled_flag)
+ flag(ph_mmvd_fullpel_only_flag);
+ else
+ infer(ph_mmvd_fullpel_only_flag, 0);
+ if (!pps->pps_rpl_info_in_ph_flag ||
+ current->ph_ref_pic_lists.rpl_ref_list[1].num_ref_entries > 0) {
+ flag(ph_mvd_l1_zero_flag);
+ if (sps->sps_bdof_control_present_in_ph_flag) {
+ flag(ph_bdof_disabled_flag);
+ } else {
+ if (!sps->sps_bdof_control_present_in_ph_flag)
+ infer(ph_bdof_disabled_flag,
+ 1 - sps->sps_bdof_enabled_flag);
+ else
+ infer(ph_bdof_disabled_flag, 1);
+ }
+ if (sps->sps_dmvr_control_present_in_ph_flag) {
+ flag(ph_dmvr_disabled_flag);
+ } else {
+ if (!sps->sps_dmvr_control_present_in_ph_flag)
+ infer(ph_dmvr_disabled_flag,
+ 1 - sps->sps_dmvr_enabled_flag);
+ else
+ infer(ph_dmvr_disabled_flag, 1);
+ }
+ } else {
+ infer(ph_mvd_l1_zero_flag, 1);
+ }
+ if (sps->sps_prof_control_present_in_ph_flag)
+ flag(ph_prof_disabled_flag);
+ else
+ infer(ph_prof_disabled_flag, !sps->sps_affine_prof_enabled_flag);
+ if ((pps->pps_weighted_pred_flag ||
+ pps->pps_weighted_bipred_flag) && pps->pps_wp_info_in_ph_flag) {
+
+ // if pps->pps_wp_info_in_ph_fla == 1
+ // pred_weight_table will not use num_ref_idx_active
+ uint8_t num_ref_idx_active[2] = { 0, 0 };
+ CHECK(FUNC(pred_weight_table)
+ (ctx, rw, sps, pps, ¤t->ph_ref_pic_lists,
+ num_ref_idx_active, ¤t->ph_pred_weight_table));
+ }
+ }
+
+ qp_bd_offset = 6 * sps->sps_bitdepth_minus8;
+ if (pps->pps_qp_delta_info_in_ph_flag)
+ se(ph_qp_delta, -qp_bd_offset - (26 + pps->pps_init_qp_minus26),
+ 63 - (26 + pps->pps_init_qp_minus26));
+
+ if (sps->sps_joint_cbcr_enabled_flag)
+ flag(ph_joint_cbcr_sign_flag);
+ else
+ infer(ph_joint_cbcr_sign_flag, 0);
+ if (sps->sps_sao_enabled_flag && pps->pps_sao_info_in_ph_flag) {
+ flag(ph_sao_luma_enabled_flag);
+ if (sps->sps_chroma_format_idc != 0)
+ flag(ph_sao_chroma_enabled_flag);
+ else
+ infer(ph_sao_chroma_enabled_flag, 0);
+ } else {
+ infer(ph_sao_luma_enabled_flag, 0);
+ infer(ph_sao_chroma_enabled_flag, 0);
+ }
+
+ if (pps->pps_dbf_info_in_ph_flag) {
+ flag(ph_deblocking_params_present_flag);
+ if (current->ph_deblocking_params_present_flag) {
+ if (!pps->pps_deblocking_filter_disabled_flag) {
+ flag(ph_deblocking_filter_disabled_flag);
+ } else {
+ if (pps->pps_deblocking_filter_disabled_flag &&
+ current->ph_deblocking_params_present_flag) {
+ infer(ph_deblocking_filter_disabled_flag, 0);
+ } else {
+ infer(ph_deblocking_filter_disabled_flag,
+ pps->pps_deblocking_filter_disabled_flag);
+ }
+ }
+ if (!current->ph_deblocking_filter_disabled_flag) {
+ se(ph_luma_beta_offset_div2, -12, 12);
+ se(ph_luma_tc_offset_div2, -12, 12);
+ if (pps->pps_chroma_tool_offsets_present_flag) {
+ se(ph_cb_beta_offset_div2, -12, 12);
+ se(ph_cb_tc_offset_div2, -12, 12);
+ se(ph_cr_beta_offset_div2, -12, 12);
+ se(ph_cr_tc_offset_div2, -12, 12);
+ } else {
+ infer(ph_cb_beta_offset_div2,
+ current->ph_luma_beta_offset_div2);
+ infer(ph_cb_tc_offset_div2,
+ current->ph_luma_tc_offset_div2);
+ infer(ph_cr_beta_offset_div2,
+ current->ph_luma_beta_offset_div2);
+ infer(ph_cr_tc_offset_div2,
+ current->ph_luma_tc_offset_div2);
+ }
+ } else {
+ infer(ph_luma_beta_offset_div2, pps->pps_luma_beta_offset_div2);
+ infer(ph_luma_tc_offset_div2, pps->pps_luma_tc_offset_div2);
+ if (pps->pps_chroma_tool_offsets_present_flag) {
+ infer(ph_cb_beta_offset_div2, pps->pps_cb_beta_offset_div2);
+ infer(ph_cb_tc_offset_div2, pps->pps_cb_tc_offset_div2);
+ infer(ph_cr_beta_offset_div2, pps->pps_cr_beta_offset_div2);
+ infer(ph_cr_tc_offset_div2, pps->pps_cr_tc_offset_div2);
+ } else {
+ infer(ph_cb_beta_offset_div2,
+ current->ph_luma_beta_offset_div2);
+ infer(ph_cb_tc_offset_div2,
+ current->ph_luma_tc_offset_div2);
+ infer(ph_cr_beta_offset_div2,
+ current->ph_luma_beta_offset_div2);
+ infer(ph_cr_tc_offset_div2,
+ current->ph_luma_tc_offset_div2);
+ }
+ }
+ }
+ }
+
+ if (pps->pps_picture_header_extension_present_flag) {
+ ue(ph_extension_length, 0, 256);
+ for (i = 0; i < current->ph_extension_length; i++)
+ us(8, ph_extension_data_byte[i], 0x00, 0xff, 1, i);
+ }
+
+ return 0;
+}
+
+static int FUNC(ph) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawPH *current)
+{
+ int err;
+
+ HEADER("Picture Header");
+
+ CHECK(FUNC(nal_unit_header) (ctx, rw, ¤t->nal_unit_header, VVC_PH_NUT));
+ CHECK(FUNC(picture_header) (ctx, rw, current));
+ CHECK(FUNC(rbsp_trailing_bits) (ctx, rw));
+ return 0;
+}
+
+static int FUNC(slice_header) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawSliceHeader *current)
+{
+ CodedBitstreamH266Context *h266 = ctx->priv_data;
+ const H266RawSPS *sps;
+ const H266RawPPS *pps;
+ const H266RawPH *ph;
+ const H266RefPicLists *ref_pic_lists;
+ int err, i;
+ uint8_t nal_unit_type, qp_bd_offset;
+ uint16_t curr_subpic_idx;
+ uint16_t num_slices_in_subpic;
+
+ HEADER("Slice Header");
+
+ CHECK(FUNC(nal_unit_header) (ctx, rw, ¤t->nal_unit_header, -1));
+
+ flag(sh_picture_header_in_slice_header_flag);
+ if (current->sh_picture_header_in_slice_header_flag) {
+ CHECK(FUNC(picture_header) (ctx, rw, ¤t->sh_picture_header));
+ ph = ¤t->sh_picture_header;
+ //7.4.8 if sh_picture_header_in_slice_header_flag is true, we do not have PH NAL unit
+ h266->priv.ph = NULL;
+ } else {
+ ph = h266->priv.ph;
+ if (!ph) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "Picture header not available.\n");
+ return AVERROR_INVALIDDATA;
+ }
+ }
+
+ pps = h266->pps[ph->ph_pic_parameter_set_id];
+ if (!pps) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "PPS id %d not available.\n",
+ ph->ph_pic_parameter_set_id);
+ return AVERROR_INVALIDDATA;
+ }
+ sps = h266->sps[pps->pps_seq_parameter_set_id];
+ if (!sps) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "SPS id %d not available.\n",
+ pps->pps_seq_parameter_set_id);
+ return AVERROR_INVALIDDATA;
+ }
+
+ if (sps->sps_subpic_info_present_flag) {
+ ub(sps->sps_subpic_id_len_minus1 + 1, sh_subpic_id);
+ if (sps->sps_subpic_id_mapping_explicitly_signalled_flag) {
+ for (i = 0; i <= sps->sps_num_subpics_minus1; i++) {
+ uint16_t subpic_id_val =
+ pps->pps_subpic_id_mapping_present_flag ?
+ pps->pps_subpic_id[i] : sps->sps_subpic_id[i];
+ if (subpic_id_val == current->sh_subpic_id) {
+ curr_subpic_idx = i;
+ break;
+ }
+ }
+ } else {
+ curr_subpic_idx = current->sh_subpic_id;
+ if (curr_subpic_idx > sps->sps_num_subpics_minus1) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "sh_subpic_id(%d) should in range [0, %d]\n",
+ curr_subpic_idx, sps->sps_num_subpics_minus1);
+ return AVERROR_INVALIDDATA;
+ }
+ }
+ } else {
+ curr_subpic_idx = 0;
+ }
+
+ num_slices_in_subpic = pps->num_slices_in_subpic[curr_subpic_idx];
+
+ if ((pps->pps_rect_slice_flag && num_slices_in_subpic > 1) ||
+ (!pps->pps_rect_slice_flag && pps->num_tiles_in_pic > 1)) {
+ unsigned int bits, max;
+ if (!pps->pps_rect_slice_flag) {
+ bits = av_ceil_log2(pps->num_tiles_in_pic);
+ max = pps->num_tiles_in_pic - 1;
+ } else {
+ bits = av_ceil_log2(num_slices_in_subpic);
+ max = num_slices_in_subpic - 1;
+ }
+ u(bits, sh_slice_address, 0, max);
+ } else {
+ infer(sh_slice_address, 0);
+ }
+
+ for (i = 0; i < sps->sps_num_extra_sh_bytes * 8; i++) {
+ if (sps->sps_extra_sh_bit_present_flag[i])
+ flags(sh_extra_bit[i], 1, i);
+ }
+
+ if (!pps->pps_rect_slice_flag &&
+ pps->num_tiles_in_pic - current->sh_slice_address > 1)
+ ue(sh_num_tiles_in_slice_minus1, 0, pps->num_tiles_in_pic - 1);
+ else
+ infer(sh_num_tiles_in_slice_minus1, 0);
+
+ if (ph->ph_inter_slice_allowed_flag)
+ ue(sh_slice_type, 0, 2);
+ else
+ infer(sh_slice_type, 2);
+
+ nal_unit_type = current->nal_unit_header.nal_unit_type;
+ if (nal_unit_type == VVC_IDR_W_RADL || nal_unit_type == VVC_IDR_N_LP ||
+ nal_unit_type == VVC_CRA_NUT || nal_unit_type == VVC_GDR_NUT)
+ flag(sh_no_output_of_prior_pics_flag);
+ if (sps->sps_alf_enabled_flag && !pps->pps_alf_info_in_ph_flag) {
+ flag(sh_alf_enabled_flag);
+ if (current->sh_alf_enabled_flag) {
+ ub(3, sh_num_alf_aps_ids_luma);
+ for (i = 0; i < current->sh_num_alf_aps_ids_luma; i++)
+ ubs(3, sh_alf_aps_id_luma[i], 1, i);
+ if (sps->sps_chroma_format_idc != 0) {
+ flag(sh_alf_cb_enabled_flag);
+ flag(sh_alf_cr_enabled_flag);
+ } else {
+ infer(sh_alf_cb_enabled_flag, ph->ph_alf_cb_enabled_flag);
+ infer(sh_alf_cr_enabled_flag, ph->ph_alf_cr_enabled_flag);
+ }
+ if (current->sh_alf_cb_enabled_flag ||
+ current->sh_alf_cr_enabled_flag)
+ ub(3, sh_alf_aps_id_chroma);
+ else
+ infer(sh_alf_aps_id_chroma, ph->ph_alf_aps_id_chroma);
+ if (sps->sps_ccalf_enabled_flag) {
+ flag(sh_alf_cc_cb_enabled_flag);
+ if (current->sh_alf_cc_cb_enabled_flag)
+ ub(3, sh_alf_cc_cb_aps_id);
+ else
+ infer(sh_alf_cc_cb_aps_id, ph->ph_alf_cc_cb_aps_id);
+ flag(sh_alf_cc_cr_enabled_flag);
+ if (current->sh_alf_cc_cr_enabled_flag)
+ ub(3, sh_alf_cc_cr_aps_id);
+ else
+ infer(sh_alf_cc_cr_aps_id, ph->ph_alf_cc_cr_aps_id);
+ } else {
+ infer(sh_alf_cc_cb_enabled_flag, ph->ph_alf_cc_cb_enabled_flag);
+ infer(sh_alf_cc_cr_enabled_flag, ph->ph_alf_cc_cr_enabled_flag);
+ infer(sh_alf_cc_cb_aps_id, ph->ph_alf_cc_cb_aps_id);
+ infer(sh_alf_cc_cr_aps_id, ph->ph_alf_cc_cr_aps_id);
+ }
+ }
+ } else {
+ infer(sh_alf_enabled_flag, 0);
+ }
+
+ if (ph->ph_lmcs_enabled_flag &&
+ !current->sh_picture_header_in_slice_header_flag)
+ flag(sh_lmcs_used_flag);
+ else
+ infer(sh_lmcs_used_flag, 0);
+
+ if (ph->ph_explicit_scaling_list_enabled_flag &&
+ !current->sh_picture_header_in_slice_header_flag)
+ flag(sh_explicit_scaling_list_used_flag);
+ else
+ infer(sh_explicit_scaling_list_used_flag, 0);
+
+ if (!pps->pps_rpl_info_in_ph_flag &&
+ ((nal_unit_type != VVC_IDR_W_RADL &&
+ nal_unit_type != VVC_IDR_N_LP) || sps->sps_idr_rpl_present_flag)) {
+ CHECK(FUNC(ref_pic_lists)
+ (ctx, rw, sps, pps, ¤t->sh_ref_pic_lists));
+ ref_pic_lists = ¤t->sh_ref_pic_lists;
+ } else {
+ ref_pic_lists = &h266->priv.ph->ph_ref_pic_lists;
+ }
+ if ((current->sh_slice_type != VVC_SLICE_TYPE_I &&
+ ref_pic_lists->rpl_ref_list[0].num_ref_entries > 1) ||
+ (current->sh_slice_type == VVC_SLICE_TYPE_B &&
+ ref_pic_lists->rpl_ref_list[1].num_ref_entries > 1)) {
+ flag(sh_num_ref_idx_active_override_flag);
+ if (current->sh_num_ref_idx_active_override_flag) {
+ for (i = 0;
+ i < (current->sh_slice_type == VVC_SLICE_TYPE_B ? 2 : 1); i++)
+ if (ref_pic_lists->rpl_ref_list[i].num_ref_entries > 1)
+ ues(sh_num_ref_idx_active_minus1[i], 0, 14, 1, i);
+ else
+ infer(sh_num_ref_idx_active_minus1[i], 0);
+ }
+ } else {
+ infer(sh_num_ref_idx_active_override_flag, 1);
+ }
+ if (current->sh_slice_type != VVC_SLICE_TYPE_I) {
+ if (pps->pps_cabac_init_present_flag)
+ flag(sh_cabac_init_flag);
+ else
+ infer(sh_cabac_init_flag, 0);
+ if (ph->ph_temporal_mvp_enabled_flag && !pps->pps_rpl_info_in_ph_flag) {
+ uint8_t num_ref_idx_active[2];
+ for (i = 0; i < 2; i++) {
+ if (current->sh_slice_type == VVC_SLICE_TYPE_B ||
+ (current->sh_slice_type == VVC_SLICE_TYPE_P && i == 0)) {
+ if (current->sh_num_ref_idx_active_override_flag) {
+ num_ref_idx_active[i] =
+ current->sh_num_ref_idx_active_minus1[i] + 1;
+ } else {
+ num_ref_idx_active[i] =
+ FFMIN(ref_pic_lists->rpl_ref_list[i].num_ref_entries,
+ pps->pps_num_ref_idx_default_active_minus1[i] + 1);
+ }
+ } else {
+ num_ref_idx_active[i] = 0;
+ }
+ }
+
+ if (current->sh_slice_type == VVC_SLICE_TYPE_B)
+ flag(sh_collocated_from_l0_flag);
+ else
+ infer(sh_collocated_from_l0_flag, 1);
+ if ((current->sh_collocated_from_l0_flag &&
+ num_ref_idx_active[0] > 1) ||
+ (!current->sh_collocated_from_l0_flag &&
+ num_ref_idx_active[1] > 1)) {
+ unsigned int idx = current->sh_collocated_from_l0_flag ? 0 : 1;
+ ue(sh_collocated_ref_idx, 0, num_ref_idx_active[idx] - 1);
+ } else {
+ infer(sh_collocated_ref_idx, 0);
+ }
+ if (!pps->pps_wp_info_in_ph_flag &&
+ ((pps->pps_weighted_pred_flag &&
+ current->sh_slice_type == VVC_SLICE_TYPE_P) ||
+ (pps->pps_weighted_bipred_flag &&
+ current->sh_slice_type == VVC_SLICE_TYPE_B))) {
+ CHECK(FUNC(pred_weight_table) (ctx, rw, sps, pps, ref_pic_lists,
+ num_ref_idx_active,
+ ¤t->sh_pred_weight_table));
+ }
+ }
+
+ }
+ qp_bd_offset = 6 * sps->sps_bitdepth_minus8;
+ if (!pps->pps_qp_delta_info_in_ph_flag)
+ se(sh_qp_delta, -qp_bd_offset - (26 + pps->pps_init_qp_minus26),
+ 63 - (26 + pps->pps_init_qp_minus26));
+ if (pps->pps_slice_chroma_qp_offsets_present_flag) {
+ int8_t off;
+
+ se(sh_cb_qp_offset, -12, 12);
+ off = pps->pps_cb_qp_offset + current->sh_cb_qp_offset;
+ if (off < -12 || off > 12) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "pps_cb_qp_offset + sh_cb_qp_offset (%d) not in range [-12, 12].\n",
+ off);
+ return AVERROR_INVALIDDATA;
+ }
+
+ se(sh_cr_qp_offset, -12, 12);
+ off = pps->pps_cr_qp_offset + current->sh_cr_qp_offset;
+ if (off < -12 || off > 12) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "pps_cr_qp_offset + sh_cr_qp_offset (%d) not in range [-12, 12].\n",
+ off);
+ return AVERROR_INVALIDDATA;
+ }
+
+ if (sps->sps_joint_cbcr_enabled_flag) {
+ se(sh_joint_cbcr_qp_offset, -12, 12);
+ off =
+ pps->pps_joint_cbcr_qp_offset_value +
+ current->sh_joint_cbcr_qp_offset;
+ if (off < -12 || off > 12) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR,
+ "pps_joint_cbcr_qp_offset_value + sh_joint_cbcr_qp_offset (%d)"
+ "not in range [-12, 12]. \n", off);
+ return AVERROR_INVALIDDATA;
+ }
+ } else {
+ infer(sh_joint_cbcr_qp_offset, 0);
+ }
+ } else {
+ infer(sh_cb_qp_offset, 0);
+ infer(sh_cr_qp_offset, 0);
+ infer(sh_joint_cbcr_qp_offset, 0);
+ }
+ if (pps->pps_cu_chroma_qp_offset_list_enabled_flag)
+ flag(sh_cu_chroma_qp_offset_enabled_flag);
+ else
+ infer(sh_cu_chroma_qp_offset_enabled_flag, 0);
+ if (sps->sps_sao_enabled_flag && !pps->pps_sao_info_in_ph_flag) {
+ flag(sh_sao_luma_used_flag);
+ if (sps->sps_chroma_format_idc != 0)
+ flag(sh_sao_chroma_used_flag);
+ else
+ infer(sh_sao_chroma_used_flag, ph->ph_sao_chroma_enabled_flag);
+ } else {
+ infer(sh_sao_luma_used_flag, ph->ph_sao_luma_enabled_flag);
+ infer(sh_sao_chroma_used_flag, ph->ph_sao_chroma_enabled_flag);
+ }
+
+ if (pps->pps_deblocking_filter_override_enabled_flag &&
+ !pps->pps_dbf_info_in_ph_flag)
+ flag(sh_deblocking_params_present_flag);
+ else
+ infer(sh_deblocking_params_present_flag, 0);
+ if (current->sh_deblocking_params_present_flag) {
+ if (!pps->pps_deblocking_filter_disabled_flag)
+ flag(sh_deblocking_filter_disabled_flag);
+ else
+ infer(sh_deblocking_filter_disabled_flag,
+ !(pps->pps_deblocking_filter_disabled_flag &&
+ current->sh_deblocking_params_present_flag));
+ if (!current->sh_deblocking_filter_disabled_flag) {
+ se(sh_luma_beta_offset_div2, -12, 12);
+ se(sh_luma_tc_offset_div2, -12, 12);
+ if (pps->pps_chroma_tool_offsets_present_flag) {
+ se(sh_cb_beta_offset_div2, -12, 12);
+ se(sh_cb_tc_offset_div2, -12, 12);
+ se(sh_cr_beta_offset_div2, -12, 12);
+ se(sh_cr_tc_offset_div2, -12, 12);
+ } else {
+ infer(sh_cb_beta_offset_div2,
+ current->sh_luma_beta_offset_div2);
+ infer(sh_cb_tc_offset_div2, current->sh_luma_tc_offset_div2);
+ infer(sh_cr_beta_offset_div2,
+ current->sh_luma_beta_offset_div2);
+ infer(sh_cr_tc_offset_div2, current->sh_luma_tc_offset_div2);
+ }
+ } else {
+ infer(sh_luma_beta_offset_div2, ph->ph_luma_beta_offset_div2);
+ infer(sh_luma_tc_offset_div2, ph->ph_luma_tc_offset_div2);
+ if (pps->pps_chroma_tool_offsets_present_flag) {
+ infer(sh_cb_beta_offset_div2, ph->ph_cb_beta_offset_div2);
+ infer(sh_cb_tc_offset_div2, ph->ph_cb_tc_offset_div2);
+ infer(sh_cr_beta_offset_div2, ph->ph_cr_beta_offset_div2);
+ infer(sh_cr_tc_offset_div2, ph->ph_cr_beta_offset_div2);
+ } else {
+ infer(sh_cb_beta_offset_div2,
+ current->sh_luma_beta_offset_div2);
+ infer(sh_cb_tc_offset_div2, current->sh_luma_tc_offset_div2);
+ infer(sh_cr_beta_offset_div2,
+ current->sh_luma_beta_offset_div2);
+ infer(sh_cr_tc_offset_div2, current->sh_luma_tc_offset_div2);
+ }
+ }
+ }
+
+ if (sps->sps_dep_quant_enabled_flag)
+ flag(sh_dep_quant_used_flag);
+ else
+ infer(sh_dep_quant_used_flag, 0);
+
+ if (sps->sps_sign_data_hiding_enabled_flag &&
+ !current->sh_dep_quant_used_flag)
+ flag(sh_sign_data_hiding_used_flag);
+ else
+ infer(sh_sign_data_hiding_used_flag, 0);
+
+ if (sps->sps_transform_skip_enabled_flag &&
+ !current->sh_dep_quant_used_flag &&
+ !current->sh_sign_data_hiding_used_flag)
+ flag(sh_ts_residual_coding_disabled_flag);
+ else
+ infer(sh_ts_residual_coding_disabled_flag, 0);
+ if (pps->pps_slice_header_extension_present_flag) {
+ ue(sh_slice_header_extension_length, 0, 256);
+ for (i = 0; i < current->sh_slice_header_extension_length; i++)
+ us(8, sh_slice_header_extension_data_byte[i], 0x00, 0xff, 1, i);
+ }
+ if (sps->sps_entry_point_offsets_present_flag) {
+ int num_entry_points = 0;
+ uint8_t entropy_sync = sps->sps_entropy_coding_sync_enabled_flag;
+ int height;
+ if (pps->pps_rect_slice_flag) {
+ int width_in_tiles;
+ int slice_idx = current->sh_slice_address;
+ for (i = 0; i < curr_subpic_idx; i++) {
+ slice_idx += pps->num_slices_in_subpic[i];
+ }
+ width_in_tiles =
+ pps->pps_slice_width_in_tiles_minus1[slice_idx] + 1;
+
+ if (entropy_sync)
+ height = pps->slice_height_in_ctus[slice_idx];
+ else
+ height = pps->pps_slice_height_in_tiles_minus1[slice_idx] + 1;
+
+ num_entry_points = width_in_tiles * height;
+ } else {
+ int tile_idx;
+ int tile_y;
+ for (tile_idx = current->sh_slice_address;
+ tile_idx <=
+ current->sh_slice_address +
+ current->sh_num_tiles_in_slice_minus1; tile_idx++) {
+ tile_y = tile_idx / pps->num_tile_rows;
+ height = pps->pps_tile_row_height_minus1[tile_y] + 1;
+ num_entry_points += (entropy_sync ? height : 1);
+ }
+ }
+ num_entry_points--;
+ if (num_entry_points > VVC_MAX_ENTRY_POINTS) {
+ av_log(ctx->log_ctx, AV_LOG_ERROR, "Too many entry points: "
+ "%" PRIu16 ".\n", num_entry_points);
+ return AVERROR_PATCHWELCOME;
+ }
+ if (num_entry_points > 0) {
+ ue(sh_entry_offset_len_minus1, 0, 31);
+ for (i = 0; i < num_entry_points; i++) {
+ ubs(current->sh_entry_offset_len_minus1 + 1,
+ sh_entry_point_offset_minus1[i], 1, i);
+ }
+ }
+ }
+ CHECK(FUNC(byte_alignment) (ctx, rw));
+
+ return 0;
+}
+
+static int FUNC(sei_decoded_picture_hash) (CodedBitstreamContext *ctx,
+ RWContext *rw,
+ H266RawSEIDecodedPictureHash *
+ current)
+{
+ int err, c_idx, i;
+
+ HEADER("Decoded Picture Hash");
+
+ u(8, dph_sei_hash_type, 0, 2);
+ flag(dph_sei_single_component_flag);
+ ub(7, dph_sei_reserved_zero_7bits);
+
+ for (c_idx = 0; c_idx < (current->dph_sei_single_component_flag ? 1 : 3);
+ c_idx++) {
+ if (current->dph_sei_hash_type == 0) {
+ for (i = 0; i < 16; i++)
+ us(8, dph_sei_picture_md5[c_idx][i], 0x00, 0xff, 2, c_idx, i);
+ } else if (current->dph_sei_hash_type == 1) {
+ us(16, dph_sei_picture_crc[c_idx], 0x0000, 0xffff, 1, c_idx);
+ } else if (current->dph_sei_hash_type == 2) {
+ us(32, dph_sei_picture_checksum[c_idx], 0x00000000, 0xffffffff, 1,
+ c_idx);
+ }
+ }
+ return 0;
+}
+
+static int FUNC(sei) (CodedBitstreamContext *ctx, RWContext *rw,
+ H266RawSEI *current, int prefix)
+{
+ int err;
+
+ if (prefix)
+ HEADER("Prefix Supplemental Enhancement Information");
+ else
+ HEADER("Suffix Supplemental Enhancement Information");
+
+ CHECK(FUNC(nal_unit_header) (ctx, rw, ¤t->nal_unit_header,
+ prefix ? VVC_PREFIX_SEI_NUT
+ : VVC_SUFFIX_SEI_NUT));
+
+ CHECK(FUNC_SEI(message_list) (ctx, rw, ¤t->message_list, prefix));
+
+ CHECK(FUNC(rbsp_trailing_bits) (ctx, rw));
+
+ return 0;
+}
diff --git a/libavcodec/cbs_internal.h b/libavcodec/cbs_internal.h
index e585c77934..2de5c243b6 100644
--- a/libavcodec/cbs_internal.h
+++ b/libavcodec/cbs_internal.h
@@ -245,6 +245,7 @@ int ff_cbs_write_signed(CodedBitstreamContext *ctx, PutBitContext *pbc,
extern const CodedBitstreamType ff_cbs_type_av1;
extern const CodedBitstreamType ff_cbs_type_h264;
extern const CodedBitstreamType ff_cbs_type_h265;
+extern const CodedBitstreamType ff_cbs_type_h266;
extern const CodedBitstreamType ff_cbs_type_jpeg;
extern const CodedBitstreamType ff_cbs_type_mpeg2;
extern const CodedBitstreamType ff_cbs_type_vp9;
diff --git a/libavcodec/cbs_sei.c b/libavcodec/cbs_sei.c
index 50a513f592..bd7f6f4938 100644
--- a/libavcodec/cbs_sei.c
+++ b/libavcodec/cbs_sei.c
@@ -20,6 +20,7 @@
#include "cbs_internal.h"
#include "cbs_h264.h"
#include "cbs_h265.h"
+#include "cbs_h266.h"
#include "cbs_sei.h"
static void cbs_free_user_data_registered(void *opaque, uint8_t *data)
@@ -132,6 +133,13 @@ static int cbs_sei_get_unit(CodedBitstreamContext *ctx,
else
sei_type = HEVC_NAL_SEI_SUFFIX;
break;
+ case AV_CODEC_ID_H266:
+ highest_vcl_type = VVC_RSV_IRAP_11;
+ if (prefix)
+ sei_type = VVC_PREFIX_SEI_NUT;
+ else
+ sei_type = VVC_SUFFIX_SEI_NUT;
+ break;
default:
return AVERROR(EINVAL);
}
@@ -207,6 +215,18 @@ static int cbs_sei_get_unit(CodedBitstreamContext *ctx,
memcpy(unit->content, &sei, sizeof(sei));
}
break;
+ case AV_CODEC_ID_H266:
+ {
+ H266RawSEI sei = {
+ .nal_unit_header = {
+ .nal_unit_type = sei_type,
+ .nuh_layer_id = 0,
+ .nuh_temporal_id_plus1 = 1,
+ },
+ };
+ memcpy(unit->content, &sei, sizeof(sei));
+ }
+ break;
default:
av_assert0(0);
}
@@ -237,6 +257,15 @@ static int cbs_sei_get_message_list(CodedBitstreamContext *ctx,
*list = &sei->message_list;
}
break;
+ case AV_CODEC_ID_H266:
+ {
+ H266RawSEI *sei = unit->content;
+ if (unit->type != VVC_PREFIX_SEI_NUT &&
+ unit->type != VVC_SUFFIX_SEI_NUT)
+ return AVERROR(EINVAL);
+ *list = &sei->message_list;
+ }
+ break;
default:
return AVERROR(EINVAL);
}
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 03/11] avcodec: add bitstream parser for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 01/11] avcodec: add enum types " Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 02/11] avcodec: add cbs " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-06-29 20:35 ` Andreas Rheinhardt
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 04/11] avcodec: add h266_metadata_bsf support " Thomas Siedel
` (8 subsequent siblings)
11 siblings, 1 reply; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
From: Nuo Mi <nuomi2021@gmail.com>
Add nal parser ff_vvc_parser to parse vvc elementary bitstreams.
Co-authored-by: Thomas Siedel <thomas.ff@spin-digital.com>
---
configure | 1 +
libavcodec/Makefile | 2 +
libavcodec/h2645_parse.c | 71 ++++-
libavcodec/h266_parser.c | 601 +++++++++++++++++++++++++++++++++++++++
libavcodec/parsers.c | 1 +
5 files changed, 675 insertions(+), 1 deletion(-)
create mode 100644 libavcodec/h266_parser.c
diff --git a/configure b/configure
index 91964edc8f..2fb167c7c3 100755
--- a/configure
+++ b/configure
@@ -3266,6 +3266,7 @@ av1_amf_encoder_deps="amf"
aac_parser_select="adts_header mpeg4audio"
av1_parser_select="cbs_av1"
h264_parser_select="golomb h264dsp h264parse h264_sei"
+h266_parser_select="cbs_h266"
hevc_parser_select="hevcparse hevc_sei"
mpegaudio_parser_select="mpegaudioheader"
mpeg4video_parser_select="h263dsp mpegvideodec qpeldsp"
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index abae4909d2..79c4363f3d 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -76,6 +76,7 @@ OBJS-$(CONFIG_CBS) += cbs.o cbs_bsf.o
OBJS-$(CONFIG_CBS_AV1) += cbs_av1.o
OBJS-$(CONFIG_CBS_H264) += cbs_h2645.o cbs_sei.o h2645_parse.o
OBJS-$(CONFIG_CBS_H265) += cbs_h2645.o cbs_sei.o h2645_parse.o
+OBJS-$(CONFIG_CBS_H266) += cbs_h2645.o cbs_sei.o h2645_parse.o
OBJS-$(CONFIG_CBS_JPEG) += cbs_jpeg.o
OBJS-$(CONFIG_CBS_MPEG2) += cbs_mpeg2.o
OBJS-$(CONFIG_CBS_VP9) += cbs_vp9.o
@@ -1166,6 +1167,7 @@ OBJS-$(CONFIG_GSM_PARSER) += gsm_parser.o
OBJS-$(CONFIG_H261_PARSER) += h261_parser.o
OBJS-$(CONFIG_H263_PARSER) += h263_parser.o
OBJS-$(CONFIG_H264_PARSER) += h264_parser.o h264data.o
+OBJS-$(CONFIG_H266_PARSER) += h266_parser.o
OBJS-$(CONFIG_HEVC_PARSER) += hevc_parser.o hevc_data.o
OBJS-$(CONFIG_HDR_PARSER) += hdr_parser.o
OBJS-$(CONFIG_IPU_PARSER) += ipu_parser.o
diff --git a/libavcodec/h2645_parse.c b/libavcodec/h2645_parse.c
index 90944177c7..5261c3e568 100644
--- a/libavcodec/h2645_parse.c
+++ b/libavcodec/h2645_parse.c
@@ -30,6 +30,7 @@
#include "hevc.h"
#include "h264.h"
#include "h2645_parse.h"
+#include "h266.h"
int ff_h2645_extract_rbsp(const uint8_t *src, int length,
H2645RBSP *rbsp, H2645NAL *nal, int small_padding)
@@ -145,6 +146,47 @@ nsc:
return si;
}
+static const char *const h266_nal_type_name[32] = {
+ "TRAIL_NUT", // VVC_TRAIL_NUT
+ "STSA_NUT", // VVC_STSA_NUT
+ "RADL_NUT", // VVC_RADL_NUT
+ "RASL_NUT", // VVC_RASL_NUT
+ "RSV_VCL_4", // VVC_RSV_VCL_4
+ "RSV_VCL_5", // VVC_RSV_VCL_5
+ "RSV_VCL_6", // VVC_RSV_VCL_6
+ "IDR_W_RADL", // VVC_IDR_W_RADL
+ "IDR_N_LP", // VVC_IDR_N_LP
+ "CRA_NUT", // VVC_CRA_NUT
+ "GDR_NUT", // VVC_GDR_NUT
+ "RSV_IRAP_11", // VVC_RSV_IRAP_11
+ "OPI_NUT", // VVC_OPI_NUT
+ "DCI_NUT", // VVC_DCI_NUT
+ "VPS_NUT", // VVC_VPS_NUT
+ "SPS_NUT", // VVC_SPS_NUT
+ "PPS_NUT", // VVC_PPS_NUT
+ "PREFIX_APS_NUT",// VVC_PREFIX_APS_NUT
+ "SUFFIX_APS_NUT",// VVC_SUFFIX_APS_NUT
+ "PH_NUT", // VVC_PH_NUT
+ "AUD_NUT", // VVC_AUD_NUT
+ "EOS_NUT", // VVC_EOS_NUT
+ "EOB_NUT", // VVC_EOB_NUT
+ "PREFIX_SEI_NUT",// VVC_PREFIX_SEI_NUT
+ "SUFFIX_SEI_NUT",// VVC_SUFFIX_SEI_NUT
+ "FD_NUT", // VVC_FD_NUT
+ "RSV_NVCL_26", // VVC_RSV_NVCL_26
+ "RSV_NVCL_27", // VVC_RSV_NVCL_27
+ "UNSPEC_28", // VVC_UNSPEC_28
+ "UNSPEC_29", // VVC_UNSPEC_29
+ "UNSPEC_30", // VVC_UNSPEC_30
+ "UNSPEC_31", // VVC_UNSPEC_31
+};
+
+static const char *h266_nal_unit_name(int nal_type)
+{
+ av_assert0(nal_type >= 0 && nal_type < 32);
+ return h266_nal_type_name[nal_type];
+}
+
static const char *const hevc_nal_type_name[64] = {
"TRAIL_N", // HEVC_NAL_TRAIL_N
"TRAIL_R", // HEVC_NAL_TRAIL_R
@@ -293,6 +335,31 @@ static int get_bit_length(H2645NAL *nal, int min_size, int skip_trailing_zeros)
* @return AVERROR_INVALIDDATA if the packet is not a valid NAL unit,
* 0 otherwise
*/
+static int h266_parse_nal_header(H2645NAL *nal, void *logctx)
+{
+ GetBitContext *gb = &nal->gb;
+
+ if (get_bits1(gb) != 0) //forbidden_zero_bit
+ return AVERROR_INVALIDDATA;
+
+ skip_bits1(gb); //nuh_reserved_zero_bit
+
+ nal->nuh_layer_id = get_bits(gb, 6);
+ nal->type = get_bits(gb, 5);
+ nal->temporal_id = get_bits(gb, 3) - 1;
+ if (nal->temporal_id < 0)
+ return AVERROR_INVALIDDATA;
+
+ if ((nal->type >= VVC_IDR_W_RADL && nal->type <= VVC_RSV_IRAP_11) && nal->temporal_id)
+ return AVERROR_INVALIDDATA;
+
+ av_log(logctx, AV_LOG_DEBUG,
+ "nal_unit_type: %d(%s), nuh_layer_id: %d, temporal_id: %d\n",
+ nal->type, h266_nal_unit_name(nal->type), nal->nuh_layer_id, nal->temporal_id);
+
+ return 0;
+}
+
static int hevc_parse_nal_header(H2645NAL *nal, void *logctx)
{
GetBitContext *gb = &nal->gb;
@@ -509,7 +576,9 @@ int ff_h2645_packet_split(H2645Packet *pkt, const uint8_t *buf, int length,
/* Reset type in case it contains a stale value from a previously parsed NAL */
nal->type = 0;
- if (codec_id == AV_CODEC_ID_HEVC)
+ if (codec_id == AV_CODEC_ID_VVC)
+ ret = h266_parse_nal_header(nal, logctx);
+ else if (codec_id == AV_CODEC_ID_HEVC)
ret = hevc_parse_nal_header(nal, logctx);
else
ret = h264_parse_nal_header(nal, logctx);
diff --git a/libavcodec/h266_parser.c b/libavcodec/h266_parser.c
new file mode 100644
index 0000000000..fbf8177261
--- /dev/null
+++ b/libavcodec/h266_parser.c
@@ -0,0 +1,601 @@
+/*
+ * H.266 / VVC parser
+ *
+ * Copyright (C) 2021 Nuo Mi <nuomi2021@gmail.com>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "cbs.h"
+#include "cbs_h266.h"
+#include "internal.h"
+#include "parser.h"
+#include "decode.h"
+
+#define START_CODE 0x000001 ///< start_code_prefix_one_3bytes
+#define IS_IDR(nut) (nut == VVC_IDR_W_RADL || nut == VVC_IDR_N_LP)
+
+typedef struct PuInfo {
+ AVBufferRef *sps_ref;
+ AVBufferRef *pps_ref;
+ AVBufferRef *slice_ref;
+ AVBufferRef *ph_ref;
+
+ const H266RawPPS *pps;
+ const H266RawSPS *sps;
+ const H266RawPH *ph;
+ const H266RawSlice *slice;
+ int pic_type;
+} PuInfo;
+
+typedef struct AuDetector {
+ uint8_t prev_layer_id;
+ int prev_tid0_poc;
+ int prev_poc;
+} AuDetector;
+
+typedef struct H266ParserContext {
+ ParseContext pc;
+ CodedBitstreamContext *cbc;
+
+ CodedBitstreamFragment picture_unit;
+
+ PuInfo au_info;
+ AVPacket au;
+ AVPacket last_au;
+
+ AuDetector au_detector;
+
+ int parsed_extradata;
+} H266ParserContext;
+
+static const enum AVPixelFormat pix_fmts_8bit[] = {
+ AV_PIX_FMT_GRAY8, AV_PIX_FMT_YUV420P,
+ AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV444P
+};
+
+static const enum AVPixelFormat pix_fmts_10bit[] = {
+ AV_PIX_FMT_GRAY10, AV_PIX_FMT_YUV420P10,
+ AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10
+};
+
+static int get_format(const H266RawSPS *sps)
+{
+ switch (sps->sps_bitdepth_minus8) {
+ case 0:
+ return pix_fmts_8bit[sps->sps_chroma_format_idc];
+ case 2:
+ return pix_fmts_10bit[sps->sps_chroma_format_idc];
+ }
+ return AV_PIX_FMT_NONE;
+}
+
+/**
+ * Find the end of the current frame in the bitstream.
+ * @return the position of the first byte of the next frame, or END_NOT_FOUND
+ */
+static int find_frame_end(AVCodecParserContext *s, const uint8_t *buf,
+ int buf_size)
+{
+ H266ParserContext *ctx = s->priv_data;
+ ParseContext *pc = &ctx->pc;
+ int i;
+
+ for (i = 0; i < buf_size; i++) {
+ int nut, code_len;
+
+ pc->state64 = (pc->state64 << 8) | buf[i];
+
+ if (((pc->state64 >> 3 * 8) & 0xFFFFFF) != START_CODE)
+ continue;
+
+ code_len = ((pc->state64 >> 3 * 8) & 0xFFFFFFFF) == 0x01 ? 4 : 3;
+
+ nut = (pc->state64 >> (8 + 3)) & 0x1F;
+ // 7.4.2.4.3 and 7.4.2.4.4
+ if ((nut >= VVC_OPI_NUT && nut <= VVC_PREFIX_APS_NUT &&
+ nut != VVC_PH_NUT) || nut == VVC_AUD_NUT
+ || (nut == VVC_PREFIX_SEI_NUT && !pc->frame_start_found)
+ || nut == VVC_RSV_NVCL_26 || nut == VVC_UNSPEC_28
+ || nut == VVC_UNSPEC_29) {
+ if (pc->frame_start_found) {
+ pc->frame_start_found = 0;
+ return i - (code_len + 2);
+ }
+ } else if (nut == VVC_PH_NUT || IS_H266_SLICE(nut)) {
+ int sh_picture_header_in_slice_header_flag = buf[i] >> 7;
+
+ if (nut == VVC_PH_NUT || sh_picture_header_in_slice_header_flag) {
+ if (!pc->frame_start_found) {
+ pc->frame_start_found = 1;
+ } else { // First slice of next frame found
+ pc->frame_start_found = 0;
+ return i - (code_len + 2);
+ }
+ }
+ }
+ }
+ return END_NOT_FOUND;
+}
+
+static int get_pict_type(const CodedBitstreamFragment *pu)
+{
+ int has_p = 0;
+ for (int i = 0; i < pu->nb_units; i++) {
+ CodedBitstreamUnit *unit = &pu->units[i];
+ if (IS_H266_SLICE(unit->type)) {
+ const H266RawSlice *slice = unit->content;
+ uint8_t type = slice->header.sh_slice_type;
+ if (type == VVC_SLICE_TYPE_B) {
+ return AV_PICTURE_TYPE_B;
+ }
+ if (type == VVC_SLICE_TYPE_P) {
+ has_p = 1;
+ }
+ }
+ }
+ return has_p ? AV_PICTURE_TYPE_P : AV_PICTURE_TYPE_I;
+}
+
+static void pu_info_unref(PuInfo *info)
+{
+ av_buffer_unref(&info->slice_ref);
+ av_buffer_unref(&info->ph_ref);
+ av_buffer_unref(&info->pps_ref);
+ av_buffer_unref(&info->sps_ref);
+ info->slice = NULL;
+ info->ph = NULL;
+ info->pps = NULL;
+ info->sps = NULL;
+ info->pic_type = AV_PICTURE_TYPE_NONE;
+}
+
+static int pu_info_ref(PuInfo *dest, const PuInfo *src)
+{
+ pu_info_unref(dest);
+ dest->sps_ref = av_buffer_ref(src->sps_ref);
+ dest->pps_ref = av_buffer_ref(src->pps_ref);
+ if (src->ph_ref)
+ dest->ph_ref = av_buffer_ref(src->ph_ref);
+ dest->slice_ref = av_buffer_ref(src->slice_ref);
+ if (!dest->sps_ref || !dest->pps_ref || (src->ph_ref && !dest->ph_ref)
+ || !dest->slice_ref) {
+ pu_info_unref(dest);
+ return AVERROR(ENOMEM);
+ }
+
+ dest->sps = src->sps;
+ dest->pps = src->pps;
+ dest->ph = src->ph;
+ dest->slice = src->slice;
+ dest->pic_type = src->pic_type;
+ return 0;
+}
+
+static int set_parser_ctx(AVCodecParserContext *s, AVCodecContext *avctx,
+ const PuInfo *pu)
+{
+ int ret, num = 0, den = 0;
+ static const uint8_t h266_sub_width_c[] = {
+ 1, 2, 2, 1
+ };
+ static const uint8_t h266_sub_height_c[] = {
+ 1, 2, 1, 1
+ };
+ const H266RawSPS *sps = pu->sps;
+ const H266RawPPS *pps = pu->pps;
+ //const H266RawPH *ph = pu->ph;
+ const H266RawNALUnitHeader *nal = &pu->slice->header.nal_unit_header;
+
+ /* set some sane default values */
+ s->pict_type = AV_PICTURE_TYPE_I;
+ s->key_frame = 0;
+ s->picture_structure = AV_PICTURE_STRUCTURE_FRAME;
+
+ s->key_frame = nal->nal_unit_type == VVC_IDR_W_RADL ||
+ nal->nal_unit_type == VVC_IDR_N_LP ||
+ nal->nal_unit_type == VVC_CRA_NUT ||
+ nal->nal_unit_type == VVC_GDR_NUT;
+
+ s->coded_width = pps->pps_pic_width_in_luma_samples;
+ s->coded_height = pps->pps_pic_height_in_luma_samples;
+ s->width = pps->pps_pic_width_in_luma_samples -
+ (pps->pps_conf_win_left_offset + pps->pps_conf_win_right_offset) *
+ h266_sub_width_c[sps->sps_chroma_format_idc];
+ s->height = pps->pps_pic_height_in_luma_samples -
+ (pps->pps_conf_win_top_offset + pps->pps_conf_win_bottom_offset) *
+ h266_sub_height_c[sps->sps_chroma_format_idc];;
+ s->pict_type = pu->pic_type;
+ s->format = get_format(sps);
+
+ avctx->profile = sps->profile_tier_level.general_profile_idc;
+ avctx->level = sps->profile_tier_level.general_level_idc;
+
+ avctx->colorspace = (enum AVColorSpace) sps->vui.vui_matrix_coeffs;
+ avctx->color_primaries = (enum AVColorPrimaries) sps->vui.vui_colour_primaries;
+ avctx->color_trc = (enum AVColorTransferCharacteristic) sps->vui.vui_transfer_characteristics;
+ avctx->color_range =
+ sps->vui.vui_full_range_flag ? AVCOL_RANGE_JPEG : AVCOL_RANGE_MPEG;
+
+ if (s->width != avctx->width || s->height != avctx->height) {
+ ret = ff_set_dimensions(avctx, s->width, s->height);
+ if (ret < 0)
+ return ret;
+ }
+ avctx->pix_fmt = s->format;
+ avctx->has_b_frames = (sps->sps_max_sublayers_minus1 + 1) > 2 ? 2 :
+ sps->sps_max_sublayers_minus1;
+ avctx->max_b_frames = sps->sps_max_sublayers_minus1;
+
+ if (sps->sps_ptl_dpb_hrd_params_present_flag &&
+ sps->sps_timing_hrd_params_present_flag) {
+ num = sps->sps_general_timing_hrd_parameters.num_units_in_tick;
+ den = sps->sps_general_timing_hrd_parameters.time_scale;
+ } else {
+ return 1;
+ }
+ if (num != 0 && den != 0)
+ av_reduce(&avctx->framerate.den, &avctx->framerate.num,
+ num, den, 1 << 30);
+
+ if (avctx->framerate.num)
+ avctx->time_base = av_inv_q(av_mul_q(avctx->framerate, (AVRational) {
+ avctx->ticks_per_frame, 1} ));
+
+ return 1;
+}
+
+static int set_ctx(AVCodecParserContext *s, AVCodecContext *avctx,
+ const PuInfo *next_pu)
+{
+ H266ParserContext *ctx = s->priv_data;
+ int ret;
+ if (ctx->au_info.slice) {
+ if ((ret = set_parser_ctx(s, avctx, &ctx->au_info)) < 0)
+ return ret;
+ }
+ ret = pu_info_ref(&ctx->au_info, next_pu);
+ return ret;
+}
+
+//8.3.1 Decoding process for picture order count.
+//VTM did not follow the spec, and it's much simpler than spec.
+//We follow the VTM.
+static void get_slice_poc(H266ParserContext *s, int *poc,
+ const H266RawSPS *sps,
+ const H266RawPH *ph,
+ const H266RawSliceHeader *slice, void *log_ctx)
+{
+ int poc_msb, max_poc_lsb, poc_lsb;
+ AuDetector *d = &s->au_detector;
+ max_poc_lsb = 1 << (sps->sps_log2_max_pic_order_cnt_lsb_minus4 + 4);
+ poc_lsb = ph->ph_pic_order_cnt_lsb;
+ if (IS_IDR(slice->nal_unit_header.nal_unit_type)) {
+ if (ph->ph_poc_msb_cycle_present_flag)
+ poc_msb = ph->ph_poc_msb_cycle_val * max_poc_lsb;
+ else
+ poc_msb = 0;
+ } else {
+ int prev_poc = d->prev_tid0_poc;
+ int prev_poc_lsb = prev_poc & (max_poc_lsb - 1);
+ int prev_poc_msb = prev_poc - prev_poc_lsb;
+ if (ph->ph_poc_msb_cycle_present_flag) {
+ poc_msb = ph->ph_poc_msb_cycle_val * max_poc_lsb;
+ } else {
+ if ((poc_lsb < prev_poc_lsb) && ((prev_poc_lsb - poc_lsb) >=
+ (max_poc_lsb / 2)))
+ poc_msb = prev_poc_msb + max_poc_lsb;
+ else if ((poc_lsb > prev_poc_lsb) && ((poc_lsb - prev_poc_lsb) >
+ (max_poc_lsb / 2)))
+ poc_msb = prev_poc_msb - max_poc_lsb;
+ else
+ poc_msb = prev_poc_msb;
+ }
+ }
+
+ *poc = poc_msb + poc_lsb;
+}
+
+static void au_detector_init(AuDetector *d)
+{
+ d->prev_layer_id = UINT8_MAX;
+ d->prev_poc = INT_MAX;
+ d->prev_tid0_poc = INT_MAX;
+}
+
+static int is_au_start(H266ParserContext *s, const PuInfo *pu, void *log_ctx)
+{
+ //7.4.2.4.3
+ AuDetector *d = &s->au_detector;
+ const H266RawSPS *sps = pu->sps;
+ const H266RawNALUnitHeader *nal = &pu->slice->header.nal_unit_header;
+ const H266RawPH *ph = pu->ph;
+ const H266RawSlice *slice = pu->slice;
+ int ret, poc, nut;
+
+ get_slice_poc(s, &poc, sps, ph, &slice->header, log_ctx);
+
+ ret = (nal->nuh_layer_id <= d->prev_layer_id) || (poc != d->prev_poc);
+
+ nut = nal->nal_unit_type;
+ d->prev_layer_id = nal->nuh_layer_id;
+ d->prev_poc = poc;
+ if (nal->nuh_temporal_id_plus1 == 1 &&
+ !ph->ph_non_ref_pic_flag && nut != VVC_RADL_NUT
+ && nut != VVC_RASL_NUT) {
+ d->prev_tid0_poc = poc;
+ }
+ return ret;
+}
+
+static int get_pu_info(PuInfo *info, const CodedBitstreamH266Context *h266,
+ const CodedBitstreamFragment *pu, void *logctx)
+{
+ const H266RawNALUnitHeader *nal;
+ int ret;
+
+ memset(info, 0, sizeof(*info));
+ for (int i = 0; i < pu->nb_units; i++) {
+ nal = pu->units[i].content;
+ if (!nal)
+ continue;
+ if ( nal->nal_unit_type == VVC_PH_NUT ) {
+ info->ph = pu->units[i].content;
+ info->ph_ref = pu->units[i].content_ref;
+ } else if (IS_H266_SLICE(nal->nal_unit_type)) {
+ info->slice = pu->units[i].content;
+ info->slice_ref = pu->units[i].content_ref;
+ if (info->slice->header.sh_picture_header_in_slice_header_flag)
+ info->ph = &info->slice->header.sh_picture_header;
+ if (!info->ph) {
+ av_log(logctx, AV_LOG_ERROR,
+ "can't find picture header in picture unit.\n");
+ ret = AVERROR_INVALIDDATA;
+ goto error;
+ }
+ break;
+ }
+ }
+ if (!info->slice) {
+ av_log(logctx, AV_LOG_ERROR, "can't find slice in picture unit.\n");
+ ret = AVERROR_INVALIDDATA;
+ goto error;
+ }
+ info->pps = h266->pps[info->ph->ph_pic_parameter_set_id];
+ if (!info->pps) {
+ av_log(logctx, AV_LOG_ERROR, "PPS id %d is not avaliable.\n",
+ info->ph->ph_pic_parameter_set_id);
+ ret = AVERROR_INVALIDDATA;
+ goto error;
+ }
+ info->pps_ref = h266->pps_ref[info->ph->ph_pic_parameter_set_id];
+ info->sps = h266->sps[info->pps->pps_seq_parameter_set_id];
+ if (!info->sps) {
+ av_log(logctx, AV_LOG_ERROR, "SPS id %d is not avaliable.\n",
+ info->pps->pps_seq_parameter_set_id);
+ ret = AVERROR_INVALIDDATA;
+ goto error;
+ }
+ info->sps_ref = h266->sps_ref[info->pps->pps_seq_parameter_set_id];
+ info->pic_type = get_pict_type(pu);
+ return 0;
+ error:
+ memset(info, 0, sizeof(*info));
+ return ret;
+}
+
+static int append_au(AVPacket *pkt, const uint8_t *buf, int buf_size)
+{
+ int offset = pkt->size;
+ int ret;
+ if ((ret = av_grow_packet(pkt, buf_size)) < 0)
+ goto end;
+ memcpy(pkt->data + offset, buf, buf_size);
+ end:
+ return ret;
+}
+
+/**
+ * Parse NAL units of found picture and decode some basic information.
+ *
+ * @param s parser context.
+ * @param avctx codec context.
+ * @param buf buffer with field/frame data.
+ * @param buf_size size of the buffer.
+ * @return < 0 for error, == 0 for a complete au, > 0 is not a completed au.
+ */
+static int parse_nal_units(AVCodecParserContext *s, const uint8_t *buf,
+ int buf_size, AVCodecContext *avctx)
+{
+ H266ParserContext *ctx = s->priv_data;
+ const CodedBitstreamH266Context *h266 = ctx->cbc->priv_data;
+
+ CodedBitstreamFragment *pu = &ctx->picture_unit;
+ int ret;
+ PuInfo info;
+
+ if (!buf_size) {
+ if (ctx->au.size) {
+ if ((ret = av_packet_ref(&ctx->last_au, &ctx->au)) < 0)
+ goto end;
+ av_packet_unref(&ctx->au);
+ return 0;
+ }
+ return 1;
+ }
+
+ if ((ret = ff_cbs_read(ctx->cbc, pu, buf, buf_size)) < 0) {
+ av_log(avctx, AV_LOG_ERROR, "Failed to parse picture unit.\n");
+ goto end;
+ }
+ if ((ret = get_pu_info(&info, h266, pu, avctx)) < 0)
+ goto end;
+ if (append_au(&ctx->au, buf, buf_size) < 0){
+ ret = AVERROR(ENOMEM);
+ goto end;
+ }
+ if (is_au_start(ctx, &info, avctx)) {
+ if ((ret = set_ctx(s, avctx, &info)) < 0)
+ goto end;
+ if ((ret = av_packet_ref(&ctx->last_au, &ctx->au)) < 0)
+ goto end;
+ av_packet_unref(&ctx->au);
+ } else {
+ ret = 1; //not a completed au
+ }
+ end:
+ ff_cbs_fragment_reset(pu);
+ return ret;
+}
+
+/**
+ * Combine PU to AU
+ *
+ * @param s parser context.
+ * @param avctx codec context.
+ * @param buf buffer to a PU.
+ * @param buf_size size of the buffer.
+ * @return < 0 for error, == 0 a complete au, > 0 not a completed au.
+ */
+static int combine_au(AVCodecParserContext *s, AVCodecContext *avctx,
+ const uint8_t **buf, int *buf_size)
+{
+ H266ParserContext *ctx = s->priv_data;
+ int ret;
+
+ ctx->cbc->log_ctx = avctx;
+
+ av_packet_unref(&ctx->last_au);
+ ret = parse_nal_units(s, *buf, *buf_size, avctx);
+ if (ret == 0) {
+ if (ctx->last_au.size) {
+ *buf = ctx->last_au.data;
+ *buf_size = ctx->last_au.size;
+ } else {
+ ret = 1; //no output
+ }
+ }
+ ctx->cbc->log_ctx = NULL;
+ return ret;
+}
+
+static int h266_parser_parse(AVCodecParserContext *s, AVCodecContext *avctx,
+ const uint8_t **poutbuf, int *poutbuf_size,
+ const uint8_t *buf, int buf_size)
+{
+ int next, ret;
+ H266ParserContext *ctx = s->priv_data;
+ ParseContext *pc = &ctx->pc;
+ CodedBitstreamFragment *pu = &ctx->picture_unit;
+
+ int is_dummy_buf = !buf_size;
+ int flush = !buf_size;
+ const uint8_t *dummy_buf = buf;
+
+ if (avctx->extradata_size && !ctx->parsed_extradata) {
+ ctx->parsed_extradata = 1;
+
+ ret = ff_cbs_read_extradata_from_codec(ctx->cbc, pu, avctx);
+ if (ret < 0)
+ av_log(avctx, AV_LOG_WARNING, "Failed to parse extradata.\n");
+
+ ff_cbs_fragment_reset(pu);
+ }
+
+ if (s->flags & PARSER_FLAG_COMPLETE_FRAMES) {
+ next = buf_size;
+ } else {
+ next = find_frame_end(s, buf, buf_size);
+ if (ff_combine_frame(pc, next, &buf, &buf_size) < 0)
+ goto no_out;
+ }
+
+ is_dummy_buf &= (dummy_buf == buf);
+
+ if (!is_dummy_buf) {
+ ret = combine_au(s, avctx, &buf, &buf_size);
+ if (ret > 0 && flush) {
+ buf_size = 0;
+ ret = combine_au(s, avctx, &buf, &buf_size);
+ }
+ if (ret != 0) {
+ buf_size = next;
+ goto no_out;
+ }
+ }
+
+ *poutbuf = buf;
+ *poutbuf_size = buf_size;
+ return next;
+ no_out:
+ *poutbuf = NULL;
+ *poutbuf_size = 0;
+ return buf_size;
+}
+
+static const CodedBitstreamUnitType decompose_unit_types[] = {
+ VVC_TRAIL_NUT,
+ VVC_STSA_NUT,
+ VVC_RADL_NUT,
+ VVC_RASL_NUT,
+ VVC_IDR_W_RADL,
+ VVC_IDR_N_LP,
+ VVC_CRA_NUT,
+ VVC_GDR_NUT,
+ VVC_VPS_NUT,
+ VVC_SPS_NUT,
+ VVC_PPS_NUT,
+ VVC_PH_NUT,
+ VVC_AUD_NUT,
+};
+
+static av_cold int h266_parser_init(AVCodecParserContext *s)
+{
+ H266ParserContext *ctx = s->priv_data;
+ int ret;
+
+ ret = ff_cbs_init(&ctx->cbc, AV_CODEC_ID_VVC, NULL);
+ if (ret < 0)
+ return ret;
+ au_detector_init(&ctx->au_detector);
+
+ ctx->cbc->decompose_unit_types = decompose_unit_types;
+ ctx->cbc->nb_decompose_unit_types = FF_ARRAY_ELEMS(decompose_unit_types);
+
+ return ret;
+}
+
+static void h266_parser_close(AVCodecParserContext *s)
+{
+ H266ParserContext *ctx = s->priv_data;
+
+ pu_info_unref(&ctx->au_info);
+ av_packet_unref(&ctx->au);
+ av_packet_unref(&ctx->last_au);
+ ff_cbs_fragment_free(&ctx->picture_unit);
+
+ ff_cbs_close(&ctx->cbc);
+ av_freep(&ctx->pc.buffer);
+}
+
+AVCodecParser ff_h266_parser = {
+ .codec_ids = { AV_CODEC_ID_VVC },
+ .priv_data_size = sizeof(H266ParserContext),
+ .parser_init = h266_parser_init,
+ .parser_close = h266_parser_close,
+ .parser_parse = h266_parser_parse,
+};
diff --git a/libavcodec/parsers.c b/libavcodec/parsers.c
index d355808018..aa2043ee1b 100644
--- a/libavcodec/parsers.c
+++ b/libavcodec/parsers.c
@@ -50,6 +50,7 @@ extern const AVCodecParser ff_gsm_parser;
extern const AVCodecParser ff_h261_parser;
extern const AVCodecParser ff_h263_parser;
extern const AVCodecParser ff_h264_parser;
+extern const AVCodecParser ff_h266_parser;
extern const AVCodecParser ff_hevc_parser;
extern const AVCodecParser ff_hdr_parser;
extern const AVCodecParser ff_ipu_parser;
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [FFmpeg-devel] [PATCH v7 03/11] avcodec: add bitstream parser for H266/VVC
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 03/11] avcodec: add bitstream parser " Thomas Siedel
@ 2023-06-29 20:35 ` Andreas Rheinhardt
0 siblings, 0 replies; 19+ messages in thread
From: Andreas Rheinhardt @ 2023-06-29 20:35 UTC (permalink / raw)
To: ffmpeg-devel
Thomas Siedel:
> From: Nuo Mi <nuomi2021@gmail.com>
>
> Add nal parser ff_vvc_parser to parse vvc elementary bitstreams.
>
> Co-authored-by: Thomas Siedel <thomas.ff@spin-digital.com>
> ---
> configure | 1 +
> libavcodec/Makefile | 2 +
> libavcodec/h2645_parse.c | 71 ++++-
> libavcodec/h266_parser.c | 601 +++++++++++++++++++++++++++++++++++++++
> libavcodec/parsers.c | 1 +
> 5 files changed, 675 insertions(+), 1 deletion(-)
> create mode 100644 libavcodec/h266_parser.c
>
> diff --git a/configure b/configure
> index 91964edc8f..2fb167c7c3 100755
> --- a/configure
> +++ b/configure
> @@ -3266,6 +3266,7 @@ av1_amf_encoder_deps="amf"
> aac_parser_select="adts_header mpeg4audio"
> av1_parser_select="cbs_av1"
> h264_parser_select="golomb h264dsp h264parse h264_sei"
> +h266_parser_select="cbs_h266"
> hevc_parser_select="hevcparse hevc_sei"
> mpegaudio_parser_select="mpegaudioheader"
> mpeg4video_parser_select="h263dsp mpegvideodec qpeldsp"
> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
> index abae4909d2..79c4363f3d 100644
> --- a/libavcodec/Makefile
> +++ b/libavcodec/Makefile
> @@ -76,6 +76,7 @@ OBJS-$(CONFIG_CBS) += cbs.o cbs_bsf.o
> OBJS-$(CONFIG_CBS_AV1) += cbs_av1.o
> OBJS-$(CONFIG_CBS_H264) += cbs_h2645.o cbs_sei.o h2645_parse.o
> OBJS-$(CONFIG_CBS_H265) += cbs_h2645.o cbs_sei.o h2645_parse.o
> +OBJS-$(CONFIG_CBS_H266) += cbs_h2645.o cbs_sei.o h2645_parse.o
> OBJS-$(CONFIG_CBS_JPEG) += cbs_jpeg.o
> OBJS-$(CONFIG_CBS_MPEG2) += cbs_mpeg2.o
> OBJS-$(CONFIG_CBS_VP9) += cbs_vp9.o
> @@ -1166,6 +1167,7 @@ OBJS-$(CONFIG_GSM_PARSER) += gsm_parser.o
> OBJS-$(CONFIG_H261_PARSER) += h261_parser.o
> OBJS-$(CONFIG_H263_PARSER) += h263_parser.o
> OBJS-$(CONFIG_H264_PARSER) += h264_parser.o h264data.o
> +OBJS-$(CONFIG_H266_PARSER) += h266_parser.o
> OBJS-$(CONFIG_HEVC_PARSER) += hevc_parser.o hevc_data.o
> OBJS-$(CONFIG_HDR_PARSER) += hdr_parser.o
> OBJS-$(CONFIG_IPU_PARSER) += ipu_parser.o
> diff --git a/libavcodec/h2645_parse.c b/libavcodec/h2645_parse.c
> index 90944177c7..5261c3e568 100644
> --- a/libavcodec/h2645_parse.c
> +++ b/libavcodec/h2645_parse.c
> @@ -30,6 +30,7 @@
> #include "hevc.h"
> #include "h264.h"
> #include "h2645_parse.h"
> +#include "h266.h"
>
> int ff_h2645_extract_rbsp(const uint8_t *src, int length,
> H2645RBSP *rbsp, H2645NAL *nal, int small_padding)
> @@ -145,6 +146,47 @@ nsc:
> return si;
> }
>
> +static const char *const h266_nal_type_name[32] = {
> + "TRAIL_NUT", // VVC_TRAIL_NUT
> + "STSA_NUT", // VVC_STSA_NUT
> + "RADL_NUT", // VVC_RADL_NUT
> + "RASL_NUT", // VVC_RASL_NUT
> + "RSV_VCL_4", // VVC_RSV_VCL_4
> + "RSV_VCL_5", // VVC_RSV_VCL_5
> + "RSV_VCL_6", // VVC_RSV_VCL_6
> + "IDR_W_RADL", // VVC_IDR_W_RADL
> + "IDR_N_LP", // VVC_IDR_N_LP
> + "CRA_NUT", // VVC_CRA_NUT
> + "GDR_NUT", // VVC_GDR_NUT
> + "RSV_IRAP_11", // VVC_RSV_IRAP_11
> + "OPI_NUT", // VVC_OPI_NUT
> + "DCI_NUT", // VVC_DCI_NUT
> + "VPS_NUT", // VVC_VPS_NUT
> + "SPS_NUT", // VVC_SPS_NUT
> + "PPS_NUT", // VVC_PPS_NUT
> + "PREFIX_APS_NUT",// VVC_PREFIX_APS_NUT
> + "SUFFIX_APS_NUT",// VVC_SUFFIX_APS_NUT
> + "PH_NUT", // VVC_PH_NUT
> + "AUD_NUT", // VVC_AUD_NUT
> + "EOS_NUT", // VVC_EOS_NUT
> + "EOB_NUT", // VVC_EOB_NUT
> + "PREFIX_SEI_NUT",// VVC_PREFIX_SEI_NUT
> + "SUFFIX_SEI_NUT",// VVC_SUFFIX_SEI_NUT
> + "FD_NUT", // VVC_FD_NUT
> + "RSV_NVCL_26", // VVC_RSV_NVCL_26
> + "RSV_NVCL_27", // VVC_RSV_NVCL_27
> + "UNSPEC_28", // VVC_UNSPEC_28
> + "UNSPEC_29", // VVC_UNSPEC_29
> + "UNSPEC_30", // VVC_UNSPEC_30
> + "UNSPEC_31", // VVC_UNSPEC_31
> +};
> +
> +static const char *h266_nal_unit_name(int nal_type)
> +{
> + av_assert0(nal_type >= 0 && nal_type < 32);
> + return h266_nal_type_name[nal_type];
> +}
> +
> static const char *const hevc_nal_type_name[64] = {
> "TRAIL_N", // HEVC_NAL_TRAIL_N
> "TRAIL_R", // HEVC_NAL_TRAIL_R
> @@ -293,6 +335,31 @@ static int get_bit_length(H2645NAL *nal, int min_size, int skip_trailing_zeros)
> * @return AVERROR_INVALIDDATA if the packet is not a valid NAL unit,
> * 0 otherwise
> */
> +static int h266_parse_nal_header(H2645NAL *nal, void *logctx)
> +{
> + GetBitContext *gb = &nal->gb;
> +
> + if (get_bits1(gb) != 0) //forbidden_zero_bit
> + return AVERROR_INVALIDDATA;
> +
> + skip_bits1(gb); //nuh_reserved_zero_bit
> +
> + nal->nuh_layer_id = get_bits(gb, 6);
> + nal->type = get_bits(gb, 5);
> + nal->temporal_id = get_bits(gb, 3) - 1;
> + if (nal->temporal_id < 0)
> + return AVERROR_INVALIDDATA;
> +
> + if ((nal->type >= VVC_IDR_W_RADL && nal->type <= VVC_RSV_IRAP_11) && nal->temporal_id)
> + return AVERROR_INVALIDDATA;
> +
> + av_log(logctx, AV_LOG_DEBUG,
> + "nal_unit_type: %d(%s), nuh_layer_id: %d, temporal_id: %d\n",
> + nal->type, h266_nal_unit_name(nal->type), nal->nuh_layer_id, nal->temporal_id);
> +
> + return 0;
> +}
> +
> static int hevc_parse_nal_header(H2645NAL *nal, void *logctx)
> {
> GetBitContext *gb = &nal->gb;
> @@ -509,7 +576,9 @@ int ff_h2645_packet_split(H2645Packet *pkt, const uint8_t *buf, int length,
> /* Reset type in case it contains a stale value from a previously parsed NAL */
> nal->type = 0;
>
> - if (codec_id == AV_CODEC_ID_HEVC)
> + if (codec_id == AV_CODEC_ID_VVC)
> + ret = h266_parse_nal_header(nal, logctx);
> + else if (codec_id == AV_CODEC_ID_HEVC)
> ret = hevc_parse_nal_header(nal, logctx);
> else
> ret = h264_parse_nal_header(nal, logctx);
> diff --git a/libavcodec/h266_parser.c b/libavcodec/h266_parser.c
> new file mode 100644
> index 0000000000..fbf8177261
> --- /dev/null
> +++ b/libavcodec/h266_parser.c
> @@ -0,0 +1,601 @@
> +/*
> + * H.266 / VVC parser
> + *
> + * Copyright (C) 2021 Nuo Mi <nuomi2021@gmail.com>
> + *
> + * This file is part of FFmpeg.
> + *
> + * FFmpeg is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU Lesser General Public
> + * License as published by the Free Software Foundation; either
> + * version 2.1 of the License, or (at your option) any later version.
> + *
> + * FFmpeg is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + * Lesser General Public License for more details.
> + *
> + * You should have received a copy of the GNU Lesser General Public
> + * License along with FFmpeg; if not, write to the Free Software
> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
> + */
> +
> +#include "cbs.h"
> +#include "cbs_h266.h"
> +#include "internal.h"
> +#include "parser.h"
> +#include "decode.h"
> +
> +#define START_CODE 0x000001 ///< start_code_prefix_one_3bytes
> +#define IS_IDR(nut) (nut == VVC_IDR_W_RADL || nut == VVC_IDR_N_LP)
> +
> +typedef struct PuInfo {
> + AVBufferRef *sps_ref;
> + AVBufferRef *pps_ref;
> + AVBufferRef *slice_ref;
> + AVBufferRef *ph_ref;
> +
> + const H266RawPPS *pps;
> + const H266RawSPS *sps;
> + const H266RawPH *ph;
> + const H266RawSlice *slice;
> + int pic_type;
> +} PuInfo;
> +
> +typedef struct AuDetector {
> + uint8_t prev_layer_id;
> + int prev_tid0_poc;
> + int prev_poc;
> +} AuDetector;
> +
> +typedef struct H266ParserContext {
> + ParseContext pc;
> + CodedBitstreamContext *cbc;
> +
> + CodedBitstreamFragment picture_unit;
> +
> + PuInfo au_info;
> + AVPacket au;
> + AVPacket last_au;
> +
> + AuDetector au_detector;
> +
> + int parsed_extradata;
> +} H266ParserContext;
> +
> +static const enum AVPixelFormat pix_fmts_8bit[] = {
> + AV_PIX_FMT_GRAY8, AV_PIX_FMT_YUV420P,
> + AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV444P
> +};
> +
> +static const enum AVPixelFormat pix_fmts_10bit[] = {
> + AV_PIX_FMT_GRAY10, AV_PIX_FMT_YUV420P10,
> + AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10
> +};
> +
> +static int get_format(const H266RawSPS *sps)
> +{
> + switch (sps->sps_bitdepth_minus8) {
> + case 0:
> + return pix_fmts_8bit[sps->sps_chroma_format_idc];
> + case 2:
> + return pix_fmts_10bit[sps->sps_chroma_format_idc];
> + }
> + return AV_PIX_FMT_NONE;
> +}
> +
> +/**
> + * Find the end of the current frame in the bitstream.
> + * @return the position of the first byte of the next frame, or END_NOT_FOUND
> + */
> +static int find_frame_end(AVCodecParserContext *s, const uint8_t *buf,
> + int buf_size)
> +{
> + H266ParserContext *ctx = s->priv_data;
> + ParseContext *pc = &ctx->pc;
> + int i;
> +
> + for (i = 0; i < buf_size; i++) {
> + int nut, code_len;
> +
> + pc->state64 = (pc->state64 << 8) | buf[i];
> +
> + if (((pc->state64 >> 3 * 8) & 0xFFFFFF) != START_CODE)
> + continue;
Reading one byte at a time is the slowest way to search for a startcode.
We have special functions for that.
> +
> + code_len = ((pc->state64 >> 3 * 8) & 0xFFFFFFFF) == 0x01 ? 4 : 3;
> +
> + nut = (pc->state64 >> (8 + 3)) & 0x1F;
> + // 7.4.2.4.3 and 7.4.2.4.4
> + if ((nut >= VVC_OPI_NUT && nut <= VVC_PREFIX_APS_NUT &&
> + nut != VVC_PH_NUT) || nut == VVC_AUD_NUT
> + || (nut == VVC_PREFIX_SEI_NUT && !pc->frame_start_found)
> + || nut == VVC_RSV_NVCL_26 || nut == VVC_UNSPEC_28
> + || nut == VVC_UNSPEC_29) {
> + if (pc->frame_start_found) {
> + pc->frame_start_found = 0;
> + return i - (code_len + 2);
> + }
> + } else if (nut == VVC_PH_NUT || IS_H266_SLICE(nut)) {
> + int sh_picture_header_in_slice_header_flag = buf[i] >> 7;
> +
> + if (nut == VVC_PH_NUT || sh_picture_header_in_slice_header_flag) {
> + if (!pc->frame_start_found) {
> + pc->frame_start_found = 1;
> + } else { // First slice of next frame found
> + pc->frame_start_found = 0;
> + return i - (code_len + 2);
> + }
> + }
> + }
> + }
> + return END_NOT_FOUND;
> +}
> +
> +static int get_pict_type(const CodedBitstreamFragment *pu)
> +{
> + int has_p = 0;
> + for (int i = 0; i < pu->nb_units; i++) {
> + CodedBitstreamUnit *unit = &pu->units[i];
> + if (IS_H266_SLICE(unit->type)) {
> + const H266RawSlice *slice = unit->content;
> + uint8_t type = slice->header.sh_slice_type;
> + if (type == VVC_SLICE_TYPE_B) {
> + return AV_PICTURE_TYPE_B;
> + }
> + if (type == VVC_SLICE_TYPE_P) {
> + has_p = 1;
> + }
> + }
> + }
> + return has_p ? AV_PICTURE_TYPE_P : AV_PICTURE_TYPE_I;
> +}
> +
> +static void pu_info_unref(PuInfo *info)
> +{
> + av_buffer_unref(&info->slice_ref);
> + av_buffer_unref(&info->ph_ref);
> + av_buffer_unref(&info->pps_ref);
> + av_buffer_unref(&info->sps_ref);
> + info->slice = NULL;
> + info->ph = NULL;
> + info->pps = NULL;
> + info->sps = NULL;
> + info->pic_type = AV_PICTURE_TYPE_NONE;
> +}
> +
> +static int pu_info_ref(PuInfo *dest, const PuInfo *src)
> +{
> + pu_info_unref(dest);
> + dest->sps_ref = av_buffer_ref(src->sps_ref);
> + dest->pps_ref = av_buffer_ref(src->pps_ref);
> + if (src->ph_ref)
> + dest->ph_ref = av_buffer_ref(src->ph_ref);
> + dest->slice_ref = av_buffer_ref(src->slice_ref);
> + if (!dest->sps_ref || !dest->pps_ref || (src->ph_ref && !dest->ph_ref)
> + || !dest->slice_ref) {
> + pu_info_unref(dest);
> + return AVERROR(ENOMEM);
> + }
> +
> + dest->sps = src->sps;
> + dest->pps = src->pps;
> + dest->ph = src->ph;
> + dest->slice = src->slice;
> + dest->pic_type = src->pic_type;
> + return 0;
> +}
> +
> +static int set_parser_ctx(AVCodecParserContext *s, AVCodecContext *avctx,
> + const PuInfo *pu)
> +{
> + int ret, num = 0, den = 0;
> + static const uint8_t h266_sub_width_c[] = {
> + 1, 2, 2, 1
> + };
> + static const uint8_t h266_sub_height_c[] = {
> + 1, 2, 1, 1
> + };
> + const H266RawSPS *sps = pu->sps;
> + const H266RawPPS *pps = pu->pps;
> + //const H266RawPH *ph = pu->ph;
> + const H266RawNALUnitHeader *nal = &pu->slice->header.nal_unit_header;
> +
> + /* set some sane default values */
> + s->pict_type = AV_PICTURE_TYPE_I;
> + s->key_frame = 0;
> + s->picture_structure = AV_PICTURE_STRUCTURE_FRAME;
> +
> + s->key_frame = nal->nal_unit_type == VVC_IDR_W_RADL ||
> + nal->nal_unit_type == VVC_IDR_N_LP ||
> + nal->nal_unit_type == VVC_CRA_NUT ||
> + nal->nal_unit_type == VVC_GDR_NUT;
> +
> + s->coded_width = pps->pps_pic_width_in_luma_samples;
> + s->coded_height = pps->pps_pic_height_in_luma_samples;
> + s->width = pps->pps_pic_width_in_luma_samples -
> + (pps->pps_conf_win_left_offset + pps->pps_conf_win_right_offset) *
> + h266_sub_width_c[sps->sps_chroma_format_idc];
> + s->height = pps->pps_pic_height_in_luma_samples -
> + (pps->pps_conf_win_top_offset + pps->pps_conf_win_bottom_offset) *
> + h266_sub_height_c[sps->sps_chroma_format_idc];;
> + s->pict_type = pu->pic_type;
> + s->format = get_format(sps);
> +
> + avctx->profile = sps->profile_tier_level.general_profile_idc;
> + avctx->level = sps->profile_tier_level.general_level_idc;
> +
> + avctx->colorspace = (enum AVColorSpace) sps->vui.vui_matrix_coeffs;
> + avctx->color_primaries = (enum AVColorPrimaries) sps->vui.vui_colour_primaries;
> + avctx->color_trc = (enum AVColorTransferCharacteristic) sps->vui.vui_transfer_characteristics;
> + avctx->color_range =
> + sps->vui.vui_full_range_flag ? AVCOL_RANGE_JPEG : AVCOL_RANGE_MPEG;
> +
> + if (s->width != avctx->width || s->height != avctx->height) {
> + ret = ff_set_dimensions(avctx, s->width, s->height);
> + if (ret < 0)
> + return ret;
> + }
> + avctx->pix_fmt = s->format;
> + avctx->has_b_frames = (sps->sps_max_sublayers_minus1 + 1) > 2 ? 2 :
> + sps->sps_max_sublayers_minus1;
> + avctx->max_b_frames = sps->sps_max_sublayers_minus1;
> +
> + if (sps->sps_ptl_dpb_hrd_params_present_flag &&
> + sps->sps_timing_hrd_params_present_flag) {
> + num = sps->sps_general_timing_hrd_parameters.num_units_in_tick;
> + den = sps->sps_general_timing_hrd_parameters.time_scale;
> + } else {
> + return 1;
> + }
> + if (num != 0 && den != 0)
> + av_reduce(&avctx->framerate.den, &avctx->framerate.num,
> + num, den, 1 << 30);
> +
> + if (avctx->framerate.num)
> + avctx->time_base = av_inv_q(av_mul_q(avctx->framerate, (AVRational) {
> + avctx->ticks_per_frame, 1} ));
> +
> + return 1;
> +}
> +
> +static int set_ctx(AVCodecParserContext *s, AVCodecContext *avctx,
> + const PuInfo *next_pu)
> +{
> + H266ParserContext *ctx = s->priv_data;
> + int ret;
> + if (ctx->au_info.slice) {
> + if ((ret = set_parser_ctx(s, avctx, &ctx->au_info)) < 0)
> + return ret;
> + }
> + ret = pu_info_ref(&ctx->au_info, next_pu);
> + return ret;
> +}
> +
> +//8.3.1 Decoding process for picture order count.
> +//VTM did not follow the spec, and it's much simpler than spec.
> +//We follow the VTM.
> +static void get_slice_poc(H266ParserContext *s, int *poc,
> + const H266RawSPS *sps,
> + const H266RawPH *ph,
> + const H266RawSliceHeader *slice, void *log_ctx)
> +{
> + int poc_msb, max_poc_lsb, poc_lsb;
> + AuDetector *d = &s->au_detector;
> + max_poc_lsb = 1 << (sps->sps_log2_max_pic_order_cnt_lsb_minus4 + 4);
> + poc_lsb = ph->ph_pic_order_cnt_lsb;
> + if (IS_IDR(slice->nal_unit_header.nal_unit_type)) {
> + if (ph->ph_poc_msb_cycle_present_flag)
> + poc_msb = ph->ph_poc_msb_cycle_val * max_poc_lsb;
> + else
> + poc_msb = 0;
> + } else {
> + int prev_poc = d->prev_tid0_poc;
> + int prev_poc_lsb = prev_poc & (max_poc_lsb - 1);
> + int prev_poc_msb = prev_poc - prev_poc_lsb;
> + if (ph->ph_poc_msb_cycle_present_flag) {
> + poc_msb = ph->ph_poc_msb_cycle_val * max_poc_lsb;
> + } else {
> + if ((poc_lsb < prev_poc_lsb) && ((prev_poc_lsb - poc_lsb) >=
> + (max_poc_lsb / 2)))
> + poc_msb = prev_poc_msb + max_poc_lsb;
> + else if ((poc_lsb > prev_poc_lsb) && ((poc_lsb - prev_poc_lsb) >
> + (max_poc_lsb / 2)))
> + poc_msb = prev_poc_msb - max_poc_lsb;
> + else
> + poc_msb = prev_poc_msb;
> + }
> + }
> +
> + *poc = poc_msb + poc_lsb;
> +}
> +
> +static void au_detector_init(AuDetector *d)
> +{
> + d->prev_layer_id = UINT8_MAX;
> + d->prev_poc = INT_MAX;
> + d->prev_tid0_poc = INT_MAX;
> +}
> +
> +static int is_au_start(H266ParserContext *s, const PuInfo *pu, void *log_ctx)
> +{
> + //7.4.2.4.3
> + AuDetector *d = &s->au_detector;
> + const H266RawSPS *sps = pu->sps;
> + const H266RawNALUnitHeader *nal = &pu->slice->header.nal_unit_header;
> + const H266RawPH *ph = pu->ph;
> + const H266RawSlice *slice = pu->slice;
> + int ret, poc, nut;
> +
> + get_slice_poc(s, &poc, sps, ph, &slice->header, log_ctx);
> +
> + ret = (nal->nuh_layer_id <= d->prev_layer_id) || (poc != d->prev_poc);
> +
> + nut = nal->nal_unit_type;
> + d->prev_layer_id = nal->nuh_layer_id;
> + d->prev_poc = poc;
> + if (nal->nuh_temporal_id_plus1 == 1 &&
> + !ph->ph_non_ref_pic_flag && nut != VVC_RADL_NUT
> + && nut != VVC_RASL_NUT) {
> + d->prev_tid0_poc = poc;
> + }
> + return ret;
> +}
> +
> +static int get_pu_info(PuInfo *info, const CodedBitstreamH266Context *h266,
> + const CodedBitstreamFragment *pu, void *logctx)
> +{
> + const H266RawNALUnitHeader *nal;
> + int ret;
> +
> + memset(info, 0, sizeof(*info));
> + for (int i = 0; i < pu->nb_units; i++) {
> + nal = pu->units[i].content;
> + if (!nal)
> + continue;
> + if ( nal->nal_unit_type == VVC_PH_NUT ) {
> + info->ph = pu->units[i].content;
> + info->ph_ref = pu->units[i].content_ref;
> + } else if (IS_H266_SLICE(nal->nal_unit_type)) {
> + info->slice = pu->units[i].content;
> + info->slice_ref = pu->units[i].content_ref;
> + if (info->slice->header.sh_picture_header_in_slice_header_flag)
> + info->ph = &info->slice->header.sh_picture_header;
> + if (!info->ph) {
> + av_log(logctx, AV_LOG_ERROR,
> + "can't find picture header in picture unit.\n");
> + ret = AVERROR_INVALIDDATA;
> + goto error;
> + }
> + break;
> + }
> + }
> + if (!info->slice) {
> + av_log(logctx, AV_LOG_ERROR, "can't find slice in picture unit.\n");
> + ret = AVERROR_INVALIDDATA;
> + goto error;
> + }
> + info->pps = h266->pps[info->ph->ph_pic_parameter_set_id];
> + if (!info->pps) {
> + av_log(logctx, AV_LOG_ERROR, "PPS id %d is not avaliable.\n",
> + info->ph->ph_pic_parameter_set_id);
> + ret = AVERROR_INVALIDDATA;
> + goto error;
> + }
> + info->pps_ref = h266->pps_ref[info->ph->ph_pic_parameter_set_id];
Copying AVBufferRef* is not nice, although I see that it is correct here.
> + info->sps = h266->sps[info->pps->pps_seq_parameter_set_id];
> + if (!info->sps) {
> + av_log(logctx, AV_LOG_ERROR, "SPS id %d is not avaliable.\n",
> + info->pps->pps_seq_parameter_set_id);
> + ret = AVERROR_INVALIDDATA;
> + goto error;
> + }
> + info->sps_ref = h266->sps_ref[info->pps->pps_seq_parameter_set_id];
> + info->pic_type = get_pict_type(pu);
> + return 0;
> + error:
> + memset(info, 0, sizeof(*info));
> + return ret;
> +}
> +
> +static int append_au(AVPacket *pkt, const uint8_t *buf, int buf_size)
> +{
> + int offset = pkt->size;
> + int ret;
> + if ((ret = av_grow_packet(pkt, buf_size)) < 0)
> + goto end;
> + memcpy(pkt->data + offset, buf, buf_size);
> + end:
> + return ret;
> +}
> +
> +/**
> + * Parse NAL units of found picture and decode some basic information.
> + *
> + * @param s parser context.
> + * @param avctx codec context.
> + * @param buf buffer with field/frame data.
> + * @param buf_size size of the buffer.
> + * @return < 0 for error, == 0 for a complete au, > 0 is not a completed au.
> + */
> +static int parse_nal_units(AVCodecParserContext *s, const uint8_t *buf,
> + int buf_size, AVCodecContext *avctx)
> +{
> + H266ParserContext *ctx = s->priv_data;
> + const CodedBitstreamH266Context *h266 = ctx->cbc->priv_data;
> +
> + CodedBitstreamFragment *pu = &ctx->picture_unit;
> + int ret;
> + PuInfo info;
> +
> + if (!buf_size) {
> + if (ctx->au.size) {
> + if ((ret = av_packet_ref(&ctx->last_au, &ctx->au)) < 0)
> + goto end;
> + av_packet_unref(&ctx->au);
> + return 0;
> + }
> + return 1;
> + }
> +
> + if ((ret = ff_cbs_read(ctx->cbc, pu, buf, buf_size)) < 0) {
> + av_log(avctx, AV_LOG_ERROR, "Failed to parse picture unit.\n");
> + goto end;
> + }
> + if ((ret = get_pu_info(&info, h266, pu, avctx)) < 0)
> + goto end;
> + if (append_au(&ctx->au, buf, buf_size) < 0){
Why is this done unconditionally, even with PARSER_FLAG_COMPLETE_FRAMES?
> + ret = AVERROR(ENOMEM);
> + goto end;
> + }
> + if (is_au_start(ctx, &info, avctx)) {
Why are you checking again? Does this mean that find_frame_end() is not
correct?
> + if ((ret = set_ctx(s, avctx, &info)) < 0)
> + goto end;
> + if ((ret = av_packet_ref(&ctx->last_au, &ctx->au)) < 0)
> + goto end;
> + av_packet_unref(&ctx->au);
> + } else {
> + ret = 1; //not a completed au
> + }
> + end:
> + ff_cbs_fragment_reset(pu);
> + return ret;
> +}
> +
> +/**
> + * Combine PU to AU
> + *
> + * @param s parser context.
> + * @param avctx codec context.
> + * @param buf buffer to a PU.
> + * @param buf_size size of the buffer.
> + * @return < 0 for error, == 0 a complete au, > 0 not a completed au.
> + */
> +static int combine_au(AVCodecParserContext *s, AVCodecContext *avctx,
> + const uint8_t **buf, int *buf_size)
> +{
> + H266ParserContext *ctx = s->priv_data;
> + int ret;
> +
> + ctx->cbc->log_ctx = avctx;
> +
> + av_packet_unref(&ctx->last_au);
> + ret = parse_nal_units(s, *buf, *buf_size, avctx);
> + if (ret == 0) {
> + if (ctx->last_au.size) {
> + *buf = ctx->last_au.data;
> + *buf_size = ctx->last_au.size;
> + } else {
> + ret = 1; //no output
> + }
> + }
> + ctx->cbc->log_ctx = NULL;
> + return ret;
> +}
> +
> +static int h266_parser_parse(AVCodecParserContext *s, AVCodecContext *avctx,
> + const uint8_t **poutbuf, int *poutbuf_size,
> + const uint8_t *buf, int buf_size)
> +{
> + int next, ret;
> + H266ParserContext *ctx = s->priv_data;
> + ParseContext *pc = &ctx->pc;
> + CodedBitstreamFragment *pu = &ctx->picture_unit;
> +
> + int is_dummy_buf = !buf_size;
> + int flush = !buf_size;
> + const uint8_t *dummy_buf = buf;
Your dummy_buf might not be a dummy_buf at all; better call it
"orig_buf" or so.
> +
> + if (avctx->extradata_size && !ctx->parsed_extradata) {
> + ctx->parsed_extradata = 1;
> +
> + ret = ff_cbs_read_extradata_from_codec(ctx->cbc, pu, avctx);
> + if (ret < 0)
> + av_log(avctx, AV_LOG_WARNING, "Failed to parse extradata.\n");
> +
> + ff_cbs_fragment_reset(pu);
> + }
> +
> + if (s->flags & PARSER_FLAG_COMPLETE_FRAMES) {
> + next = buf_size;
> + } else {
> + next = find_frame_end(s, buf, buf_size);
> + if (ff_combine_frame(pc, next, &buf, &buf_size) < 0)
> + goto no_out;
> + }
> +
> + is_dummy_buf &= (dummy_buf == buf);
> +
> + if (!is_dummy_buf) {
I don't really understand this. The above check is only relevant if we
are flushing; but if we are flushing, then we should always analyze
everything we have. It seems to me that the above check is actually
meant to be equivalent to "if (buf_size)". (If
PARSER_FLAG_COMPLETE_FRAMES is set, then we have no leftover stuff from
last time and is_dummy_buf is 1 for flushing; if
PARSER_FLAG_COMPLETE_FRAMES is not true, then find_frame_end() won't
find an end in our 0-sized buffer, but ff_combine_frame() will (as a
special case when parsing) return a value >= 0 and the line above will
make is_dummy_buf 0, so we analyze the leftover.)
> + ret = combine_au(s, avctx, &buf, &buf_size);
> + if (ret > 0 && flush) {
> + buf_size = 0;
> + ret = combine_au(s, avctx, &buf, &buf_size);
> + }
> + if (ret != 0) {
> + buf_size = next;
> + goto no_out;
> + }
> + }
> +
> + *poutbuf = buf;
> + *poutbuf_size = buf_size;
> + return next;
> + no_out:
> + *poutbuf = NULL;
> + *poutbuf_size = 0;
> + return buf_size;
> +}
> +
> +static const CodedBitstreamUnitType decompose_unit_types[] = {
> + VVC_TRAIL_NUT,
> + VVC_STSA_NUT,
> + VVC_RADL_NUT,
> + VVC_RASL_NUT,
> + VVC_IDR_W_RADL,
> + VVC_IDR_N_LP,
> + VVC_CRA_NUT,
> + VVC_GDR_NUT,
> + VVC_VPS_NUT,
> + VVC_SPS_NUT,
> + VVC_PPS_NUT,
> + VVC_PH_NUT,
> + VVC_AUD_NUT,
> +};
> +
> +static av_cold int h266_parser_init(AVCodecParserContext *s)
> +{
> + H266ParserContext *ctx = s->priv_data;
> + int ret;
> +
> + ret = ff_cbs_init(&ctx->cbc, AV_CODEC_ID_VVC, NULL);
> + if (ret < 0)
> + return ret;
> + au_detector_init(&ctx->au_detector);
> +
> + ctx->cbc->decompose_unit_types = decompose_unit_types;
> + ctx->cbc->nb_decompose_unit_types = FF_ARRAY_ELEMS(decompose_unit_types);
> +
> + return ret;
> +}
> +
> +static void h266_parser_close(AVCodecParserContext *s)
> +{
> + H266ParserContext *ctx = s->priv_data;
> +
> + pu_info_unref(&ctx->au_info);
> + av_packet_unref(&ctx->au);
> + av_packet_unref(&ctx->last_au);
> + ff_cbs_fragment_free(&ctx->picture_unit);
> +
> + ff_cbs_close(&ctx->cbc);
> + av_freep(&ctx->pc.buffer);
> +}
> +
> +AVCodecParser ff_h266_parser = {
> + .codec_ids = { AV_CODEC_ID_VVC },
> + .priv_data_size = sizeof(H266ParserContext),
> + .parser_init = h266_parser_init,
> + .parser_close = h266_parser_close,
> + .parser_parse = h266_parser_parse,
> +};
> diff --git a/libavcodec/parsers.c b/libavcodec/parsers.c
> index d355808018..aa2043ee1b 100644
> --- a/libavcodec/parsers.c
> +++ b/libavcodec/parsers.c
> @@ -50,6 +50,7 @@ extern const AVCodecParser ff_gsm_parser;
> extern const AVCodecParser ff_h261_parser;
> extern const AVCodecParser ff_h263_parser;
> extern const AVCodecParser ff_h264_parser;
> +extern const AVCodecParser ff_h266_parser;
> extern const AVCodecParser ff_hevc_parser;
> extern const AVCodecParser ff_hdr_parser;
> extern const AVCodecParser ff_ipu_parser;
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread
* [FFmpeg-devel] [PATCH v7 04/11] avcodec: add h266_metadata_bsf support for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (2 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 03/11] avcodec: add bitstream parser " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 05/11] avcodec: add MP4 to annexb " Thomas Siedel
` (7 subsequent siblings)
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
From: Nuo Mi <nuomi2021@gmail.com>
Add H.266/VVC metadata bsf.
---
configure | 1 +
libavcodec/Makefile | 1 +
libavcodec/bitstream_filters.c | 1 +
libavcodec/h266_metadata_bsf.c | 147 +++++++++++++++++++++++++++++++++
4 files changed, 150 insertions(+)
create mode 100644 libavcodec/h266_metadata_bsf.c
diff --git a/configure b/configure
index 2fb167c7c3..297ce6e27f 100755
--- a/configure
+++ b/configure
@@ -3283,6 +3283,7 @@ filter_units_bsf_select="cbs"
h264_metadata_bsf_deps="const_nan"
h264_metadata_bsf_select="cbs_h264"
h264_redundant_pps_bsf_select="cbs_h264"
+h266_metadata_bsf_select="cbs_h266"
hevc_metadata_bsf_select="cbs_h265"
mjpeg2jpeg_bsf_select="jpegtables"
mpeg2_metadata_bsf_select="cbs_mpeg2"
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 79c4363f3d..c3989df358 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -1219,6 +1219,7 @@ OBJS-$(CONFIG_H264_METADATA_BSF) += h264_metadata_bsf.o h264_levels.o \
h2645data.o
OBJS-$(CONFIG_H264_MP4TOANNEXB_BSF) += h264_mp4toannexb_bsf.o
OBJS-$(CONFIG_H264_REDUNDANT_PPS_BSF) += h264_redundant_pps_bsf.o
+OBJS-$(CONFIG_H266_METADATA_BSF) += h266_metadata_bsf.o
OBJS-$(CONFIG_HAPQA_EXTRACT_BSF) += hapqa_extract_bsf.o hap.o
OBJS-$(CONFIG_HEVC_METADATA_BSF) += h265_metadata_bsf.o h265_profile_level.o \
h2645data.o
diff --git a/libavcodec/bitstream_filters.c b/libavcodec/bitstream_filters.c
index e8216819ca..848f430014 100644
--- a/libavcodec/bitstream_filters.c
+++ b/libavcodec/bitstream_filters.c
@@ -39,6 +39,7 @@ extern const FFBitStreamFilter ff_filter_units_bsf;
extern const FFBitStreamFilter ff_h264_metadata_bsf;
extern const FFBitStreamFilter ff_h264_mp4toannexb_bsf;
extern const FFBitStreamFilter ff_h264_redundant_pps_bsf;
+extern const FFBitStreamFilter ff_h266_metadata_bsf;
extern const FFBitStreamFilter ff_hapqa_extract_bsf;
extern const FFBitStreamFilter ff_hevc_metadata_bsf;
extern const FFBitStreamFilter ff_hevc_mp4toannexb_bsf;
diff --git a/libavcodec/h266_metadata_bsf.c b/libavcodec/h266_metadata_bsf.c
new file mode 100644
index 0000000000..05e476071f
--- /dev/null
+++ b/libavcodec/h266_metadata_bsf.c
@@ -0,0 +1,147 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/common.h"
+#include "libavutil/opt.h"
+
+#include "bsf.h"
+#include "bsf_internal.h"
+#include "cbs.h"
+#include "cbs_bsf.h"
+#include "cbs_h266.h"
+#include "h266.h"
+
+typedef struct H266MetadataContext {
+ CBSBSFContext common;
+
+ H266RawAUD aud_nal;
+
+ int aud;
+} H266MetadataContext;
+
+static int h266_metadata_update_fragment(AVBSFContext *bsf, AVPacket *pkt,
+ CodedBitstreamFragment *pu)
+{
+ H266MetadataContext *ctx = bsf->priv_data;
+ int err, i;
+
+ // If an AUD is present, it must be the first NAL unit.
+ if (pu->units[0].type == VVC_AUD_NUT) {
+ if (ctx->aud == BSF_ELEMENT_REMOVE)
+ ff_cbs_delete_unit(pu, 0);
+ } else if ( pkt && ctx->aud == BSF_ELEMENT_INSERT) {
+ const H266RawSlice *first_slice = NULL;
+ const H266RawPH *ph = NULL;
+ H266RawAUD *aud = &ctx->aud_nal;
+ int pic_type = 0, temporal_id = 8, layer_id = 0;
+ for (i = 0; i < pu->nb_units; i++) {
+ const H266RawNALUnitHeader *nal = pu->units[i].content;
+ if (!nal)
+ continue;
+ if (nal->nuh_temporal_id_plus1 < temporal_id + 1)
+ temporal_id = nal->nuh_temporal_id_plus1 - 1;
+ if ( nal->nal_unit_type == VVC_PH_NUT ) {
+ ph = pu->units[i].content;
+ } else if (IS_H266_SLICE(nal->nal_unit_type)) {
+ const H266RawSlice *slice = pu->units[i].content;
+ layer_id = nal->nuh_layer_id;
+ if (slice->header.sh_slice_type == VVC_SLICE_TYPE_B &&
+ pic_type < 2)
+ pic_type = 2;
+ if (slice->header.sh_slice_type == VVC_SLICE_TYPE_P &&
+ pic_type < 1)
+ pic_type = 1;
+ if (!first_slice) {
+ first_slice = slice;
+ if (first_slice->header.
+ sh_picture_header_in_slice_header_flag)
+ ph = &first_slice->header.sh_picture_header;
+ else if (!ph)
+ break;
+ }
+ }
+ }
+ if (!ph) {
+ av_log(bsf, AV_LOG_ERROR, "no avaliable picture header");
+ return AVERROR_INVALIDDATA;
+ }
+
+ aud->nal_unit_header = (H266RawNALUnitHeader) {
+ .nal_unit_type = VVC_AUD_NUT,
+ .nuh_layer_id = layer_id,
+ .nuh_temporal_id_plus1 = temporal_id + 1,
+ };
+ aud->aud_pic_type = pic_type;
+ aud->aud_irap_or_gdr_flag = ph->ph_gdr_or_irap_pic_flag;
+
+ err = ff_cbs_insert_unit_content(pu, 0, VVC_AUD_NUT, aud, NULL);
+ if (err < 0) {
+ av_log(bsf, AV_LOG_ERROR, "Failed to insert AUD.\n");
+ return err;
+ }
+ }
+
+ /* TODO: implement more metadata parsing, like VUI, Levels etc. */
+ //for (i = 0; i < pu->nb_units; i++) {
+ // if (pu->units[i].type == VVC_SPS_NUT) {
+ // }
+ //}
+ return 0;
+}
+
+static const CBSBSFType h266_metadata_type = {
+ .codec_id = AV_CODEC_ID_VVC,
+ .fragment_name = "access unit",
+ .unit_name = "NAL unit",
+ .update_fragment = &h266_metadata_update_fragment,
+};
+
+static int h266_metadata_init(AVBSFContext *bsf)
+{
+ return ff_cbs_bsf_generic_init(bsf, &h266_metadata_type);
+}
+
+#define OFFSET(x) offsetof(H266MetadataContext, x)
+#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
+static const AVOption h266_metadata_options[] = {
+ BSF_ELEMENT_OPTIONS_PIR("aud", "Access Unit Delimiter NAL units",
+ aud, FLAGS),
+
+ { NULL }
+};
+
+static const AVClass h266_metadata_class = {
+ .class_name = "h266_metadata_bsf",
+ .item_name = av_default_item_name,
+ .option = h266_metadata_options,
+ .version = LIBAVUTIL_VERSION_INT,
+};
+
+static const enum AVCodecID h266_metadata_codec_ids[] = {
+ AV_CODEC_ID_VVC, AV_CODEC_ID_NONE,
+};
+
+const FFBitStreamFilter ff_h266_metadata_bsf = {
+ .p.name = "h266_metadata",
+ .p.codec_ids = h266_metadata_codec_ids,
+ .p.priv_class = &h266_metadata_class,
+ .priv_data_size = sizeof(H266MetadataContext),
+ .init = &h266_metadata_init,
+ .close = &ff_cbs_bsf_generic_close,
+ .filter = &ff_cbs_bsf_generic_filter,
+};
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 05/11] avcodec: add MP4 to annexb support for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (3 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 04/11] avcodec: add h266_metadata_bsf support " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 06/11] avformat: add demuxer and probe " Thomas Siedel
` (6 subsequent siblings)
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
Add parser for H.266/VVC MP4 to Annex B byte stream format.
---
libavcodec/Makefile | 1 +
libavcodec/bitstream_filters.c | 1 +
libavcodec/h266_mp4toannexb_bsf.c | 329 ++++++++++++++++++++++++++++++
3 files changed, 331 insertions(+)
create mode 100644 libavcodec/h266_mp4toannexb_bsf.c
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index c3989df358..81fda09597 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -1220,6 +1220,7 @@ OBJS-$(CONFIG_H264_METADATA_BSF) += h264_metadata_bsf.o h264_levels.o \
OBJS-$(CONFIG_H264_MP4TOANNEXB_BSF) += h264_mp4toannexb_bsf.o
OBJS-$(CONFIG_H264_REDUNDANT_PPS_BSF) += h264_redundant_pps_bsf.o
OBJS-$(CONFIG_H266_METADATA_BSF) += h266_metadata_bsf.o
+OBJS-$(CONFIG_H266_MP4TOANNEXB_BSF) += h266_mp4toannexb_bsf.o
OBJS-$(CONFIG_HAPQA_EXTRACT_BSF) += hapqa_extract_bsf.o hap.o
OBJS-$(CONFIG_HEVC_METADATA_BSF) += h265_metadata_bsf.o h265_profile_level.o \
h2645data.o
diff --git a/libavcodec/bitstream_filters.c b/libavcodec/bitstream_filters.c
index 848f430014..e0e3116eb4 100644
--- a/libavcodec/bitstream_filters.c
+++ b/libavcodec/bitstream_filters.c
@@ -40,6 +40,7 @@ extern const FFBitStreamFilter ff_h264_metadata_bsf;
extern const FFBitStreamFilter ff_h264_mp4toannexb_bsf;
extern const FFBitStreamFilter ff_h264_redundant_pps_bsf;
extern const FFBitStreamFilter ff_h266_metadata_bsf;
+extern const FFBitStreamFilter ff_h266_mp4toannexb_bsf;
extern const FFBitStreamFilter ff_hapqa_extract_bsf;
extern const FFBitStreamFilter ff_hevc_metadata_bsf;
extern const FFBitStreamFilter ff_hevc_mp4toannexb_bsf;
diff --git a/libavcodec/h266_mp4toannexb_bsf.c b/libavcodec/h266_mp4toannexb_bsf.c
new file mode 100644
index 0000000000..1f7bf0618d
--- /dev/null
+++ b/libavcodec/h266_mp4toannexb_bsf.c
@@ -0,0 +1,329 @@
+/*
+ * H.266/VVC MP4 to Annex B byte stream format filter
+ * Copyright (c) 2022, Thomas Siedel
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <string.h>
+
+#include "libavutil/intreadwrite.h"
+#include "libavutil/mem.h"
+
+#include "avcodec.h"
+#include "bsf.h"
+#include "bsf_internal.h"
+#include "bytestream.h"
+#include "h266.h"
+
+#include "libavcodec/get_bits.h"
+
+#define MIN_VVCC_LENGTH 23
+
+typedef struct H266BSFContext {
+ uint8_t length_size;
+ int extradata_parsed;
+} H266BSFContext;
+
+static int h266_extradata_to_annexb(AVBSFContext *ctx)
+{
+ GetByteContext gb;
+ int length_size, num_arrays, i, j;
+ int ret = 0;
+ int temp = 0;
+ int ptl_present;
+
+ uint8_t *new_extradata = NULL;
+ size_t new_extradata_size = 0;
+
+ int max_picture_width = 0;
+ int max_picture_height = 0;
+ int avg_frame_rate = 0;
+
+ bytestream2_init(&gb, ctx->par_in->extradata, ctx->par_in->extradata_size);
+ temp = bytestream2_get_byte(&gb);
+ length_size = ((temp & 6) >> 1) + 1;
+ ptl_present = temp & 1;
+ if (ptl_present) {
+ int num_bytes_constraint_info;
+ int general_profile_idc;
+ int general_tier_flag;
+ int general_level_idc;
+ int ptl_frame_only_constraint_flag;
+ int ptl_multi_layer_enabled_flag;
+ int ptl_num_sub_profiles;
+ int temp3, temp4, temp5;
+ int temp2 = bytestream2_get_be16(&gb);
+ int ols_idx = (temp2 >> 7) & 0x1ff;
+ int num_sublayers = (temp2 >> 4) & 0x7;
+ int constant_frame_rate = (temp2 >> 2) & 0x3;
+ int chroma_format_idc = temp2 & 0x3;
+ int bit_depth_minus8 = (bytestream2_get_byte(&gb) >> 5) & 0x7;
+ av_log(ctx, AV_LOG_DEBUG,
+ "bit_depth_minus8 %d chroma_format_idc %d\n", bit_depth_minus8,
+ chroma_format_idc);
+ av_log(ctx, AV_LOG_DEBUG, "constant_frame_rate %d, ols_idx %d\n",
+ constant_frame_rate, ols_idx);
+ // VvcPTLRecord(num_sublayers) native_ptl
+ temp3 = bytestream2_get_byte(&gb);
+ num_bytes_constraint_info = (temp3) & 0x3f;
+ temp4 = bytestream2_get_byte(&gb);
+ general_profile_idc = (temp4 >> 1) & 0x7f;
+ general_tier_flag = (temp4) & 1;
+ general_level_idc = bytestream2_get_byte(&gb);
+ av_log(ctx, AV_LOG_DEBUG,
+ "general_profile_idc %d, general_tier_flag %d, general_level_idc %d, num_sublayers %d num_bytes_constraint_info %d\n",
+ general_profile_idc, general_tier_flag, general_level_idc,
+ num_sublayers, num_bytes_constraint_info);
+
+ temp5 = bytestream2_get_byte(&gb);
+ ptl_frame_only_constraint_flag = (temp5 >> 7) & 0x1;
+ ptl_multi_layer_enabled_flag = (temp5 >> 6) & 0x1;
+ for (i = 0; i < num_bytes_constraint_info - 1; i++) {
+ // unsigned int(8*num_bytes_constraint_info - 2) general_constraint_info;
+ bytestream2_get_byte(&gb);
+ }
+
+ av_log(ctx, AV_LOG_DEBUG,
+ "ptl_multi_layer_enabled_flag %d, ptl_frame_only_constraint_flag %d\n",
+ ptl_multi_layer_enabled_flag, ptl_frame_only_constraint_flag);
+
+ if (num_sublayers > 1) {
+ int temp6 = bytestream2_get_byte(&gb);
+ uint8_t ptl_sublayer_level_present_flag[8] = { 0 };
+ //uint8_t sublayer_level_idc[8] = {0};
+ for (i = num_sublayers - 2; i >= 0; i--) {
+ ptl_sublayer_level_present_flag[i] =
+ (temp6 >> (7 - (num_sublayers - 2 - i))) & 0x01;
+ }
+ // for (j=num_sublayers; j<=8 && num_sublayers > 1; j++)
+ // bit(1) ptl_reserved_zero_bit = 0;
+ for (i = num_sublayers - 2; i >= 0; i--) {
+ if (ptl_sublayer_level_present_flag[i]) {
+ //sublayer_level_idc[i] = bytestream2_get_byte(&gb);
+ }
+ }
+ }
+
+ ptl_num_sub_profiles = bytestream2_get_byte(&gb);
+ for (j = 0; j < ptl_num_sub_profiles; j++) {
+ // unsigned int(32) general_sub_profile_idc[j];
+ bytestream2_get_be16(&gb);
+ bytestream2_get_be16(&gb);
+ }
+
+ max_picture_width = bytestream2_get_be16(&gb); // unsigned_int(16) max_picture_width;
+ max_picture_height = bytestream2_get_be16(&gb); // unsigned_int(16) max_picture_height;
+ avg_frame_rate = bytestream2_get_be16(&gb); // unsigned int(16) avg_frame_rate; }
+ av_log(ctx, AV_LOG_DEBUG,
+ "max_picture_width %d, max_picture_height %d, avg_frame_rate %d\n",
+ max_picture_width, max_picture_height, avg_frame_rate);
+ }
+
+ num_arrays = bytestream2_get_byte(&gb);
+
+ for (i = 0; i < num_arrays; i++) {
+ int cnt;
+ int type = bytestream2_get_byte(&gb) & 0x1f;
+
+ if (type == VVC_OPI_NUT || type == VVC_DCI_NUT)
+ cnt = 1;
+ else
+ cnt = bytestream2_get_be16(&gb);
+
+ av_log(ctx, AV_LOG_DEBUG, "nalu_type %d cnt %d\n", type, cnt);
+
+ if (!(type == VVC_OPI_NUT || type == VVC_DCI_NUT ||
+ type == VVC_VPS_NUT || type == VVC_SPS_NUT || type == VVC_PPS_NUT
+ || type == VVC_PREFIX_SEI_NUT || type == VVC_SUFFIX_SEI_NUT)) {
+ av_log(ctx, AV_LOG_ERROR,
+ "Invalid NAL unit type in extradata: %d\n", type);
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ for (j = 0; j < cnt; j++) {
+ int nalu_len = bytestream2_get_be16(&gb);
+
+ if (4 + AV_INPUT_BUFFER_PADDING_SIZE + nalu_len >
+ SIZE_MAX - new_extradata_size) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+ ret = av_reallocp(&new_extradata, new_extradata_size + nalu_len + 4
+ + AV_INPUT_BUFFER_PADDING_SIZE);
+ if (ret < 0)
+ goto fail;
+
+ AV_WB32(new_extradata + new_extradata_size, 1); // add the startcode
+ bytestream2_get_buffer(&gb, new_extradata + new_extradata_size + 4,
+ nalu_len);
+ new_extradata_size += 4 + nalu_len;
+ memset(new_extradata + new_extradata_size, 0,
+ AV_INPUT_BUFFER_PADDING_SIZE);
+ }
+ }
+
+ av_freep(&ctx->par_out->extradata);
+ ctx->par_out->extradata = new_extradata;
+ ctx->par_out->extradata_size = new_extradata_size;
+
+ if (!new_extradata_size)
+ av_log(ctx, AV_LOG_WARNING, "No parameter sets in the extradata\n");
+
+ return length_size;
+ fail:
+ av_freep(&new_extradata);
+ return ret;
+}
+
+static int h266_mp4toannexb_init(AVBSFContext *ctx)
+{
+ H266BSFContext *s = ctx->priv_data;
+ int ret;
+
+ if (ctx->par_in->extradata_size < MIN_VVCC_LENGTH ||
+ AV_RB24(ctx->par_in->extradata) == 1 ||
+ AV_RB32(ctx->par_in->extradata) == 1) {
+ av_log(ctx, AV_LOG_VERBOSE,
+ "The input looks like it is Annex B already\n");
+ } else {
+ ret = h266_extradata_to_annexb(ctx);
+ if (ret < 0)
+ return ret;
+ s->length_size = ret;
+ s->extradata_parsed = 1;
+ }
+
+ return 0;
+}
+
+static int h266_mp4toannexb_filter(AVBSFContext *ctx, AVPacket *out)
+{
+ H266BSFContext *s = ctx->priv_data;
+ AVPacket *in;
+ GetByteContext gb;
+
+ int is_irap = 0;
+ int added_extra = 0;
+ int i, ret = 0;
+
+ ret = ff_bsf_get_packet(ctx, &in);
+ if (ret < 0)
+ return ret;
+
+ if (!s->extradata_parsed) {
+ av_packet_move_ref(out, in);
+ av_packet_free(&in);
+ return 0;
+ }
+
+ bytestream2_init(&gb, in->data, in->size);
+
+ /* check if this packet contains an IRAP. The extradata will need to be added before any potential PH_NUT */
+ while (bytestream2_get_bytes_left(&gb)) {
+ uint32_t nalu_size = 0;
+ int nalu_type;
+
+ if (bytestream2_get_bytes_left(&gb) < s->length_size) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ for (i = 0; i < s->length_size; i++)
+ nalu_size = (nalu_size << 8) | bytestream2_get_byte(&gb);
+
+ if (nalu_size < 2 || nalu_size > bytestream2_get_bytes_left(&gb)) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ nalu_type = (bytestream2_peek_be16(&gb) >> 3) & 0x1f;
+ is_irap = nalu_type >= VVC_IDR_W_RADL && nalu_type <= VVC_RSV_IRAP_11;
+ if (is_irap) {
+ break;
+ }
+ bytestream2_seek(&gb, nalu_size, SEEK_CUR);
+ }
+
+ bytestream2_seek(&gb, 0, SEEK_SET);
+ while (bytestream2_get_bytes_left(&gb)) {
+ uint32_t nalu_size = 0;
+ int nalu_type;
+ int add_extradata, extra_size, prev_size;
+
+ if (bytestream2_get_bytes_left(&gb) < s->length_size) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ for (i = 0; i < s->length_size; i++)
+ nalu_size = (nalu_size << 8) | bytestream2_get_byte(&gb);
+
+ if (nalu_size < 2 || nalu_size > bytestream2_get_bytes_left(&gb)) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ nalu_type = (bytestream2_peek_be16(&gb) >> 3) & 0x1f;
+
+ /* prepend extradata to IRAP frames */
+ add_extradata = is_irap && nalu_type != VVC_AUD_NUT && !added_extra;
+ extra_size = add_extradata * ctx->par_out->extradata_size;
+ added_extra |= add_extradata;
+
+ if (FFMIN(INT_MAX, SIZE_MAX) < 4ULL + nalu_size + extra_size) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ prev_size = out->size;
+
+ ret = av_grow_packet(out, 4 + nalu_size + extra_size);
+ if (ret < 0)
+ goto fail;
+
+ if (extra_size)
+ memcpy(out->data + prev_size, ctx->par_out->extradata, extra_size);
+ AV_WB32(out->data + prev_size + extra_size, 1);
+ bytestream2_get_buffer(&gb, out->data + prev_size + 4 + extra_size,
+ nalu_size);
+ }
+
+ ret = av_packet_copy_props(out, in);
+ if (ret < 0)
+ goto fail;
+
+ fail:
+ if (ret < 0)
+ av_packet_unref(out);
+ av_packet_free(&in);
+
+ return ret;
+}
+
+static const enum AVCodecID codec_ids[] = {
+ AV_CODEC_ID_VVC, AV_CODEC_ID_NONE,
+};
+
+const FFBitStreamFilter ff_h266_mp4toannexb_bsf = {
+ .p.name = "h266_mp4toannexb",
+ .p.codec_ids = codec_ids,
+ .priv_data_size = sizeof(H266BSFContext),
+ .init = h266_mp4toannexb_init,
+ .filter = h266_mp4toannexb_filter,
+};
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 06/11] avformat: add demuxer and probe support for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (4 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 05/11] avcodec: add MP4 to annexb " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-12-18 11:43 ` Zhao Zhili
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 07/11] avformat: add muxer " Thomas Siedel
` (5 subsequent siblings)
11 siblings, 1 reply; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
Add demuxer to probe raw vvc and parse vvcc byte stream format.
Co-authored-by: Nuo Mi <nuomi2021@gmail.com>
---
libavformat/Makefile | 1 +
libavformat/allformats.c | 1 +
libavformat/demux.c | 7 +-
libavformat/h266.c | 998 +++++++++++++++++++++++++++++++++++++++
libavformat/h266.h | 99 ++++
libavformat/h266dec.c | 61 +++
6 files changed, 1165 insertions(+), 2 deletions(-)
create mode 100644 libavformat/h266.c
create mode 100644 libavformat/h266.h
create mode 100644 libavformat/h266dec.c
diff --git a/libavformat/Makefile b/libavformat/Makefile
index 47bbbbfb2a..9d8efe1936 100644
--- a/libavformat/Makefile
+++ b/libavformat/Makefile
@@ -245,6 +245,7 @@ OBJS-$(CONFIG_H263_DEMUXER) += h263dec.o rawdec.o
OBJS-$(CONFIG_H263_MUXER) += rawenc.o
OBJS-$(CONFIG_H264_DEMUXER) += h264dec.o rawdec.o
OBJS-$(CONFIG_H264_MUXER) += rawenc.o
+OBJS-$(CONFIG_H266_DEMUXER) += h266dec.o rawdec.o
OBJS-$(CONFIG_HASH_MUXER) += hashenc.o
OBJS-$(CONFIG_HCA_DEMUXER) += hca.o
OBJS-$(CONFIG_HCOM_DEMUXER) += hcom.o pcm.o
diff --git a/libavformat/allformats.c b/libavformat/allformats.c
index cb5b69e9cd..855849449d 100644
--- a/libavformat/allformats.c
+++ b/libavformat/allformats.c
@@ -199,6 +199,7 @@ extern const AVInputFormat ff_h263_demuxer;
extern const FFOutputFormat ff_h263_muxer;
extern const AVInputFormat ff_h264_demuxer;
extern const FFOutputFormat ff_h264_muxer;
+extern const AVInputFormat ff_h266_demuxer;
extern const FFOutputFormat ff_hash_muxer;
extern const AVInputFormat ff_hca_demuxer;
extern const AVInputFormat ff_hcom_demuxer;
diff --git a/libavformat/demux.c b/libavformat/demux.c
index b19ab86d08..f0c6d7895f 100644
--- a/libavformat/demux.c
+++ b/libavformat/demux.c
@@ -120,6 +120,7 @@ static int set_codec_from_probe_data(AVFormatContext *s, AVStream *st,
{ "mp3", AV_CODEC_ID_MP3, AVMEDIA_TYPE_AUDIO },
{ "mpegvideo", AV_CODEC_ID_MPEG2VIDEO, AVMEDIA_TYPE_VIDEO },
{ "truehd", AV_CODEC_ID_TRUEHD, AVMEDIA_TYPE_AUDIO },
+ { "vvc", AV_CODEC_ID_VVC, AVMEDIA_TYPE_VIDEO },
{ 0 }
};
int score;
@@ -743,7 +744,8 @@ static int64_t select_from_pts_buffer(AVStream *st, int64_t *pts_buffer, int64_t
{
FFStream *const sti = ffstream(st);
int onein_oneout = st->codecpar->codec_id != AV_CODEC_ID_H264 &&
- st->codecpar->codec_id != AV_CODEC_ID_HEVC;
+ st->codecpar->codec_id != AV_CODEC_ID_HEVC &&
+ st->codecpar->codec_id != AV_CODEC_ID_VVC;
if (!onein_oneout) {
int delay = sti->avctx->has_b_frames;
@@ -933,7 +935,8 @@ static void compute_pkt_fields(AVFormatContext *s, AVStream *st,
int64_t offset;
AVRational duration;
int onein_oneout = st->codecpar->codec_id != AV_CODEC_ID_H264 &&
- st->codecpar->codec_id != AV_CODEC_ID_HEVC;
+ st->codecpar->codec_id != AV_CODEC_ID_HEVC &&
+ st->codecpar->codec_id != AV_CODEC_ID_VVC;
if (s->flags & AVFMT_FLAG_NOFILLIN)
return;
diff --git a/libavformat/h266.c b/libavformat/h266.c
new file mode 100644
index 0000000000..fe2c2532a9
--- /dev/null
+++ b/libavformat/h266.c
@@ -0,0 +1,998 @@
+/*
+ * H.266/VVC helper functions for muxers
+ *
+ * Copyright (C) 2022, Thomas Siedel
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavcodec/get_bits.h"
+#include "libavcodec/golomb.h"
+#include "libavcodec/h266.h"
+#include "libavutil/intreadwrite.h"
+#include "avc.h"
+#include "avio.h"
+#include "avio_internal.h"
+#include "h266.h"
+
+typedef struct VVCCNALUnitArray {
+ uint8_t array_completeness;
+ uint8_t NAL_unit_type;
+ uint16_t num_nalus;
+ uint16_t *nal_unit_length;
+ uint8_t **nal_unit;
+} VVCCNALUnitArray;
+
+typedef struct VVCPTLRecord {
+ uint8_t num_bytes_constraint_info;
+ uint8_t general_profile_idc;
+ uint8_t general_tier_flag;
+ uint8_t general_level_idc;
+ uint8_t ptl_frame_only_constraint_flag;
+ uint8_t ptl_multilayer_enabled_flag;
+ uint8_t general_constraint_info[9];
+ uint8_t *ptl_sublayer_level_present_flag;
+ uint8_t *sublayer_level_idc;
+ uint8_t ptl_num_sub_profiles;
+ uint32_t *general_sub_profile_idc;
+} VVCPTLRecord;
+
+typedef struct VVCDecoderConfigurationRecord {
+ uint8_t lengthSizeMinusOne;
+ uint8_t ptl_present_flag;
+ uint16_t ols_idx;
+ uint8_t num_sublayers;
+ uint8_t constant_frame_rate;
+ uint8_t chroma_format_idc;
+ uint8_t bit_depth_minus8;
+ VVCPTLRecord ptl;
+ uint16_t max_picture_width;
+ uint16_t max_picture_height;
+ uint16_t avg_frame_rate;
+ uint8_t num_of_arrays;
+ VVCCNALUnitArray *array;
+} VVCDecoderConfigurationRecord;
+
+typedef struct VVCCProfileTierLevel {
+ uint8_t profile_idc;
+ uint8_t tier_flag;
+ uint8_t general_level_idc;
+ uint8_t ptl_frame_only_constraint_flag;
+ uint8_t ptl_multilayer_enabled_flag;
+// general_constraint_info
+ uint8_t gci_present_flag;
+ uint8_t gci_general_constraints[9];
+ uint8_t gci_num_reserved_bits;
+// end general_constraint_info
+ uint8_t *ptl_sublayer_level_present_flag;
+ uint8_t *sublayer_level_idc;
+ uint8_t ptl_num_sub_profiles;
+ uint32_t *general_sub_profile_idc;
+} VVCCProfileTierLevel;
+
+static void vvcc_update_ptl(VVCDecoderConfigurationRecord *vvcc,
+ VVCCProfileTierLevel *ptl)
+{
+ /*
+ * The level indication general_level_idc must indicate a level of
+ * capability equal to or greater than the highest level indicated for the
+ * highest tier in all the parameter sets.
+ */
+ if (vvcc->ptl.general_tier_flag < ptl->tier_flag)
+ vvcc->ptl.general_level_idc = ptl->general_level_idc;
+ else
+ vvcc->ptl.general_level_idc =
+ FFMAX(vvcc->ptl.general_level_idc, ptl->general_level_idc);
+
+ /*
+ * The tier indication general_tier_flag must indicate a tier equal to or
+ * greater than the highest tier indicated in all the parameter sets.
+ */
+ vvcc->ptl.general_tier_flag =
+ FFMAX(vvcc->ptl.general_tier_flag, ptl->tier_flag);
+
+ /*
+ * The profile indication general_profile_idc must indicate a profile to
+ * which the stream associated with this configuration record conforms.
+ *
+ * If the sequence parameter sets are marked with different profiles, then
+ * the stream may need examination to determine which profile, if any, the
+ * entire stream conforms to. If the entire stream is not examined, or the
+ * examination reveals that there is no profile to which the entire stream
+ * conforms, then the entire stream must be split into two or more
+ * sub-streams with separate configuration records in which these rules can
+ * be met.
+ *
+ * Note: set the profile to the highest value for the sake of simplicity.
+ */
+ vvcc->ptl.general_profile_idc =
+ FFMAX(vvcc->ptl.general_profile_idc, ptl->profile_idc);
+
+ /*
+ * Each bit in flags may only be set if all
+ * the parameter sets set that bit.
+ */
+ vvcc->ptl.ptl_frame_only_constraint_flag &=
+ ptl->ptl_frame_only_constraint_flag;
+ vvcc->ptl.ptl_multilayer_enabled_flag &= ptl->ptl_multilayer_enabled_flag;
+
+ /*
+ * Constraints Info
+ */
+ if (ptl->gci_present_flag) {
+ vvcc->ptl.num_bytes_constraint_info = 9;
+ memcpy(&vvcc->ptl.general_constraint_info[0],
+ &ptl->gci_general_constraints[0], sizeof(uint8_t) * 9);
+
+ } else {
+ vvcc->ptl.num_bytes_constraint_info = 1;
+ memset(&vvcc->ptl.general_constraint_info[0], 0, sizeof(uint8_t) * 9);
+ }
+
+ /*
+ * Each bit in flags may only be set if one of
+ * the parameter sets set that bit.
+ */
+ vvcc->ptl.ptl_sublayer_level_present_flag =
+ (uint8_t *) malloc(sizeof(uint8_t) * vvcc->num_sublayers - 1);
+ vvcc->ptl.sublayer_level_idc =
+ (uint8_t *) malloc(sizeof(uint8_t) * vvcc->num_sublayers - 1);
+
+ memset(vvcc->ptl.ptl_sublayer_level_present_flag, 0,
+ sizeof(uint8_t) * vvcc->num_sublayers - 1);
+ memset(vvcc->ptl.sublayer_level_idc, 0,
+ sizeof(uint8_t) * vvcc->num_sublayers - 1);
+
+ for (int i = vvcc->num_sublayers - 2; i >= 0; i--) {
+ vvcc->ptl.ptl_sublayer_level_present_flag[i] |=
+ ptl->ptl_sublayer_level_present_flag[i];
+ if (vvcc->ptl.ptl_sublayer_level_present_flag[i]) {
+ vvcc->ptl.sublayer_level_idc[i] =
+ FFMAX(vvcc->ptl.sublayer_level_idc[i],
+ ptl->sublayer_level_idc[i]);
+ } else {
+ if (i == vvcc->num_sublayers - 1) {
+ vvcc->ptl.sublayer_level_idc[i] = vvcc->ptl.general_level_idc;
+ } else {
+ vvcc->ptl.sublayer_level_idc[i] =
+ vvcc->ptl.sublayer_level_idc[i + 1];
+ }
+ }
+ }
+
+ vvcc->ptl.ptl_num_sub_profiles =
+ FFMAX(vvcc->ptl.ptl_num_sub_profiles, ptl->ptl_num_sub_profiles);
+ if (vvcc->ptl.ptl_num_sub_profiles) {
+ vvcc->ptl.general_sub_profile_idc =
+ (uint32_t *) malloc(sizeof(uint32_t) *
+ vvcc->ptl.ptl_num_sub_profiles);
+ for (int i = 0; i < vvcc->ptl.ptl_num_sub_profiles; i++) {
+ vvcc->ptl.general_sub_profile_idc[i] =
+ ptl->general_sub_profile_idc[i];
+ }
+ } else {
+ vvcc->ptl.general_sub_profile_idc =
+ (uint32_t *) malloc(sizeof(uint32_t));
+ }
+}
+
+static void vvcc_parse_ptl(GetBitContext *gb,
+ VVCDecoderConfigurationRecord *vvcc,
+ unsigned int profileTierPresentFlag,
+ unsigned int max_sub_layers_minus1)
+{
+ VVCCProfileTierLevel general_ptl;
+ int j;
+
+ if (profileTierPresentFlag) {
+ general_ptl.profile_idc = get_bits(gb, 7);
+ general_ptl.tier_flag = get_bits1(gb);
+ }
+ general_ptl.general_level_idc = get_bits(gb, 8);
+
+ general_ptl.ptl_frame_only_constraint_flag = get_bits1(gb);
+ general_ptl.ptl_multilayer_enabled_flag = get_bits1(gb);
+ if (profileTierPresentFlag) { // parse constraint info
+ general_ptl.gci_present_flag = get_bits1(gb);
+ if (general_ptl.gci_present_flag) {
+ for (j = 0; j < 8; j++)
+ general_ptl.gci_general_constraints[j] = get_bits(gb, 8);
+ general_ptl.gci_general_constraints[8] = 0;
+ general_ptl.gci_general_constraints[8] = get_bits(gb, 7);
+
+ general_ptl.gci_num_reserved_bits = get_bits(gb, 8);
+ skip_bits(gb, general_ptl.gci_num_reserved_bits);
+ }
+ while (gb->index % 8 != 0)
+ skip_bits1(gb);
+ }
+
+ general_ptl.ptl_sublayer_level_present_flag =
+ (uint8_t *) malloc(sizeof(uint8_t) * max_sub_layers_minus1);
+ for (int i = max_sub_layers_minus1 - 1; i >= 0; i--) {
+ general_ptl.ptl_sublayer_level_present_flag[i] = get_bits1(gb);
+ }
+ while (gb->index % 8 != 0)
+ skip_bits1(gb);
+
+ general_ptl.sublayer_level_idc =
+ (uint8_t *) malloc(sizeof(uint8_t) * max_sub_layers_minus1);
+ for (int i = max_sub_layers_minus1 - 1; i >= 0; i--) {
+ if (general_ptl.ptl_sublayer_level_present_flag[i])
+ general_ptl.sublayer_level_idc[i] = get_bits(gb, 8);
+ }
+
+ if (profileTierPresentFlag) {
+ general_ptl.ptl_num_sub_profiles = get_bits(gb, 8);
+ if (general_ptl.ptl_num_sub_profiles) {
+ general_ptl.general_sub_profile_idc =
+ (uint32_t *) malloc(sizeof(uint32_t) *
+ general_ptl.ptl_num_sub_profiles);
+ for (int i = 0; i < general_ptl.ptl_num_sub_profiles; i++) {
+ general_ptl.general_sub_profile_idc[i] = get_bits_long(gb, 32);
+ }
+ } else {
+ general_ptl.general_sub_profile_idc =
+ (uint32_t *) malloc(sizeof(uint32_t));
+ }
+ }
+
+ vvcc_update_ptl(vvcc, &general_ptl);
+
+ free(general_ptl.ptl_sublayer_level_present_flag);
+ free(general_ptl.sublayer_level_idc);
+ free(general_ptl.general_sub_profile_idc);
+}
+
+static int vvcc_parse_vps(GetBitContext *gb,
+ VVCDecoderConfigurationRecord *vvcc)
+{
+ unsigned int vps_max_layers_minus1;
+ unsigned int vps_max_sub_layers_minus1;
+ unsigned int vps_default_ptl_dpb_hrd_max_tid_flag;
+ unsigned int vps_all_independant_layer_flag;
+ unsigned int vps_each_layer_is_an_ols_flag;
+ unsigned int vps_ols_mode_idc;
+
+ unsigned int *vps_pt_present_flag;
+ unsigned int *vps_ptl_max_tid;
+ unsigned int vps_num_ptls_minus1 = 0;
+
+ /*
+ * vps_video_parameter_set_id u(4)
+ */
+ skip_bits(gb, 4);
+
+ vps_max_layers_minus1 = get_bits(gb, 6);
+ vps_max_sub_layers_minus1 = get_bits(gb, 3);
+
+ /*
+ * numTemporalLayers greater than 1 indicates that the stream to which this
+ * configuration record applies is temporally scalable and the contained
+ * number of temporal layers (also referred to as temporal sub-layer or
+ * sub-layer in ISO/IEC 23008-2) is equal to numTemporalLayers. Value 1
+ * indicates that the stream is not temporally scalable. Value 0 indicates
+ * that it is unknown whether the stream is temporally scalable.
+ */
+ vvcc->num_sublayers = FFMAX(vvcc->num_sublayers,
+ vps_max_sub_layers_minus1 + 1);
+
+ if (vps_max_layers_minus1 > 0 && vps_max_sub_layers_minus1 > 0)
+ vps_default_ptl_dpb_hrd_max_tid_flag = get_bits1(gb);
+ if (vps_max_layers_minus1 > 0)
+ vps_all_independant_layer_flag = get_bits1(gb);
+
+ for (int i = 0; i <= vps_max_layers_minus1; i++) {
+ skip_bits(gb, 6); //vps_default_ptl_dpb_hrd_max_tid_flag[i]
+ if (i > 0 && !vps_all_independant_layer_flag) {
+ if (get_bits1(gb)) { // vps_independant_layer_flag
+ unsigned int vps_max_tid_ref_present_flag = get_bits1(gb);
+ for (int j = 0; j < i; j++) {
+ if (vps_max_tid_ref_present_flag && get_bits1(gb)) // vps_direct_ref_layer_flag[i][j]
+ skip_bits(gb, 3); // vps_max_tid_il_ref_pics_plus1
+ }
+ }
+ }
+ }
+
+ if (vps_max_layers_minus1 > 0) {
+ if (vps_all_independant_layer_flag)
+ vps_each_layer_is_an_ols_flag = get_bits1(gb);
+ if (vps_each_layer_is_an_ols_flag) {
+ if (!vps_all_independant_layer_flag)
+ vps_ols_mode_idc = get_bits(gb, 2);
+ if (vps_ols_mode_idc == 2) {
+ unsigned int vps_num_output_layer_sets_minus2 = get_bits(gb, 8);
+ for (int i = 1; i <= vps_num_output_layer_sets_minus2 + 1; i++) {
+ for (int j = 0; j <= vps_max_layers_minus1; j++) {
+ skip_bits1(gb);
+ }
+ }
+ }
+ }
+ vps_num_ptls_minus1 = get_bits(gb, 8);
+ }
+
+ vps_pt_present_flag =
+ (unsigned int *) malloc(sizeof(unsigned int) *
+ (vps_num_ptls_minus1 + 1));
+ vps_ptl_max_tid =
+ (unsigned int *) malloc(sizeof(unsigned int) *
+ (vps_num_ptls_minus1 + 1));
+ for (int i = 0; i <= vps_num_ptls_minus1; i++) {
+ if (i > 0)
+ vps_pt_present_flag[i] = get_bits1(gb);
+ if (!vps_default_ptl_dpb_hrd_max_tid_flag)
+ vps_ptl_max_tid[i] = get_bits(gb, 3);
+ }
+
+ while (gb->index % 8 != 0)
+ skip_bits1(gb);
+
+ for (int i = 0; i <= vps_num_ptls_minus1; i++) {
+ vvcc_parse_ptl(gb, vvcc, vps_pt_present_flag[i], vps_ptl_max_tid[i]);
+ }
+
+ free(vps_pt_present_flag);
+ free(vps_ptl_max_tid);
+
+ /* nothing useful for vvcc past this point */
+ return 0;
+}
+
+static int vvcc_parse_sps(GetBitContext *gb,
+ VVCDecoderConfigurationRecord *vvcc)
+{
+ unsigned int sps_max_sub_layers_minus1, log2_ctu_size_minus5;
+ //unsigned int num_short_term_ref_pic_sets, num_delta_pocs[VVC_MAX_REF_PIC_LISTS];
+ //unsigned int sps_chroma_format_idc;
+ unsigned int sps_subpic_same_size_flag, sps_pic_height_max_in_luma_sample,
+ sps_pic_width_max_in_luma_sample;
+ unsigned int sps_independant_subpics_flag;
+ unsigned int flag;
+
+ skip_bits(gb, 8); // sps_seq_parameter_set_id && sps_video_parameter_set_id
+ sps_max_sub_layers_minus1 = get_bits(gb, 3);
+
+ /*
+ * numTemporalLayers greater than 1 indicates that the stream to which this
+ * configuration record applies is temporally scalable and the contained
+ * number of temporal layers (also referred to as temporal sub-layer or
+ * sub-layer in ISO/IEC 23008-2) is equal to numTemporalLayers. Value 1
+ * indicates that the stream is not temporally scalable. Value 0 indicates
+ * that it is unknown whether the stream is temporally scalable.
+ */
+ vvcc->num_sublayers = FFMAX(vvcc->num_sublayers,
+ sps_max_sub_layers_minus1 + 1);
+
+ vvcc->chroma_format_idc = get_bits(gb, 2);
+ log2_ctu_size_minus5 = get_bits(gb, 2);
+
+ if (get_bits1(gb)) //sps_ptl_dpb_hrd_params_present_flag
+ vvcc_parse_ptl(gb, vvcc, 1, sps_max_sub_layers_minus1);
+
+ flag = get_bits(gb, 1); //skip_bits1(gb); //sps_gdr_enabled_flag
+ flag = get_bits(gb, 1); //sps_ref_pic_resampling_enabled_flag
+ if (flag) { //sps_ref_pic_resampling_enabled_flag
+ flag = get_bits(gb, 1); //skip_bits1(gb); //sps_res_change_in_clvs_allowed_flag
+ }
+
+ sps_pic_width_max_in_luma_sample = get_ue_golomb_long(gb);
+ vvcc->max_picture_width =
+ FFMAX(vvcc->max_picture_width, sps_pic_width_max_in_luma_sample);
+ sps_pic_height_max_in_luma_sample = get_ue_golomb_long(gb);
+ vvcc->max_picture_height =
+ FFMAX(vvcc->max_picture_height, sps_pic_height_max_in_luma_sample);
+
+ if (get_bits1(gb)) {
+ get_ue_golomb_long(gb); // sps_conf_win_left_offset
+ get_ue_golomb_long(gb); // sps_conf_win_right_offset
+ get_ue_golomb_long(gb); // sps_conf_win_top_offset
+ get_ue_golomb_long(gb); // sps_conf_win_bottom_offset
+ }
+
+ if (get_bits1(gb)) { // sps_subpic_info_present_flag
+ unsigned int sps_num_subpics_minus1 = get_ue_golomb_long(gb);
+ if (sps_num_subpics_minus1 > 0) { // sps_num_subpics_minus1
+ sps_independant_subpics_flag = get_bits1(gb);
+ sps_subpic_same_size_flag = get_bits1(gb);
+ }
+ for (int i = 0;
+ sps_num_subpics_minus1 > 0 && i <= sps_num_subpics_minus1; i++) {
+ if (!sps_subpic_same_size_flag || i == 0) {
+ int len = FFMIN(log2_ctu_size_minus5 + 5, 16);
+ if (i > 0 && sps_pic_width_max_in_luma_sample > 128)
+ skip_bits(gb, len);
+ if (i > 0 && sps_pic_height_max_in_luma_sample > 128)
+ skip_bits(gb, len);
+ if (i < sps_num_subpics_minus1
+ && sps_pic_width_max_in_luma_sample > 128)
+ skip_bits(gb, len);
+ if (i < sps_num_subpics_minus1
+ && sps_pic_height_max_in_luma_sample > 128)
+ skip_bits(gb, len);
+ }
+ if (!sps_independant_subpics_flag) {
+ skip_bits(gb, 2); // sps_subpic_treated_as_pic_flag && sps_loop_filter_across_subpic_enabled_flag
+ }
+ }
+ get_ue_golomb_long(gb); // sps_subpic_id_len_minus1
+ if (get_bits1(gb)) { // sps_subpic_id_mapping_explicitly_signalled_flag
+ if (get_bits1(gb)) // sps_subpic_id_mapping_present_flag
+ for (int i = 0; i <= sps_num_subpics_minus1; i++) {
+ skip_bits1(gb); // sps_subpic_id[i]
+ }
+ }
+ }
+ vvcc->bit_depth_minus8 = get_ue_golomb_long(gb);
+
+ /* nothing useful for vvcc past this point */
+ return 0;
+}
+
+static int vvcc_parse_pps(GetBitContext *gb,
+ VVCDecoderConfigurationRecord *vvcc)
+{
+
+ // Nothing of importance to parse in PPS
+ /* nothing useful for vvcc past this point */
+ return 0;
+}
+
+static void nal_unit_parse_header(GetBitContext *gb, uint8_t *nal_type)
+{
+ /*
+ * forbidden_zero_bit u(1)
+ * nuh_reserved_zero_bit u(1)
+ * nuh_layer_id u(6)
+ */
+ skip_bits(gb, 8);
+ *nal_type = get_bits(gb, 5);
+
+ /*
+ * nuh_temporal_id_plus1 u(3)
+ */
+ skip_bits(gb, 3);
+}
+
+static int vvcc_array_add_nal_unit(uint8_t *nal_buf, uint32_t nal_size,
+ uint8_t nal_type, int ps_array_completeness,
+ VVCDecoderConfigurationRecord *vvcc)
+{
+ int ret;
+ uint8_t index;
+ uint16_t num_nalus;
+ VVCCNALUnitArray *array;
+
+ for (index = 0; index < vvcc->num_of_arrays; index++)
+ if (vvcc->array[index].NAL_unit_type == nal_type)
+ break;
+
+ if (index >= vvcc->num_of_arrays) {
+ uint8_t i;
+
+ ret =
+ av_reallocp_array(&vvcc->array, index + 1,
+ sizeof(VVCCNALUnitArray));
+ if (ret < 0)
+ return ret;
+
+ for (i = vvcc->num_of_arrays; i <= index; i++)
+ memset(&vvcc->array[i], 0, sizeof(VVCCNALUnitArray));
+ vvcc->num_of_arrays = index + 1;
+ }
+
+ array = &vvcc->array[index];
+ num_nalus = array->num_nalus;
+
+ ret = av_reallocp_array(&array->nal_unit, num_nalus + 1, sizeof(uint8_t *));
+ if (ret < 0)
+ return ret;
+
+ ret =
+ av_reallocp_array(&array->nal_unit_length, num_nalus + 1,
+ sizeof(uint16_t));
+ if (ret < 0)
+ return ret;
+
+ array->nal_unit[num_nalus] = nal_buf;
+ array->nal_unit_length[num_nalus] = nal_size;
+ array->NAL_unit_type = nal_type;
+ array->num_nalus++;
+
+ /*
+ * When the sample entry name is 'vvc1', the following applies:
+ * • The value of array_completeness shall be equal to 1 for arrays of SPS,
+ * and PPS NAL units.
+ * • If a VVC bitstream includes DCI NAL unit(s), the value of
+ * array_completeness shall be equal to 1 for the array of DCI units.
+ * Otherwise, NAL_unit_type shall not indicate DCI NAL units.
+ * • If a VVC bitstream includes VPS NAL unit(s), the value of
+ * array_completeness shall be equal to 1 for the array of VPS NAL units.
+ * Otherwise, NAL_unit_type shall not indicate VPS NAL units.
+ * When the value of array_completeness is equal to 1 for an array of a
+ * particular NAL_unit_type value, NAL units of that NAL_unit_type value
+ * cannot be updated without causing a different sample entry to be used.
+ * When the sample entry name is 'vvi1', the value of array_completeness
+ * of at least one of the following arrays shall be equal to 0:
+ • The array of DCI NAL units, if present.
+ • The array of VPS NAL units, if present.
+ • The array of SPS NAL units
+ • The array of PPS NAL units.
+ */
+ if (nal_type == VVC_VPS_NUT || nal_type == VVC_SPS_NUT ||
+ nal_type == VVC_PPS_NUT || nal_type == VVC_DCI_NUT )
+ array->array_completeness = ps_array_completeness;
+
+ return 0;
+}
+
+static int vvcc_add_nal_unit(uint8_t *nal_buf, uint32_t nal_size,
+ int ps_array_completeness,
+ VVCDecoderConfigurationRecord *vvcc)
+{
+ int ret = 0;
+ GetBitContext gbc;
+ uint8_t nal_type;
+ uint8_t *rbsp_buf;
+ uint32_t rbsp_size;
+
+ rbsp_buf = ff_nal_unit_extract_rbsp(nal_buf, nal_size, &rbsp_size, 2);
+ if (!rbsp_buf) {
+ ret = AVERROR(ENOMEM);
+ goto end;
+ }
+
+ ret = init_get_bits8(&gbc, rbsp_buf, rbsp_size);
+ if (ret < 0)
+ goto end;
+
+ nal_unit_parse_header(&gbc, &nal_type);
+
+ /*
+ * Note: only 'declarative' SEI messages are allowed in
+ * vvcc. Perhaps the SEI playload type should be checked
+ * and non-declarative SEI messages discarded?
+ */
+ switch (nal_type) {
+ case VVC_OPI_NUT:
+ case VVC_VPS_NUT:
+ case VVC_SPS_NUT:
+ case VVC_PPS_NUT:
+ case VVC_PREFIX_SEI_NUT:
+ case VVC_SUFFIX_SEI_NUT:
+ ret = vvcc_array_add_nal_unit(nal_buf, nal_size, nal_type,
+ ps_array_completeness, vvcc);
+ if (ret < 0)
+ goto end;
+ else if (nal_type == VVC_VPS_NUT)
+ ret = vvcc_parse_vps(&gbc, vvcc);
+ else if (nal_type == VVC_SPS_NUT)
+ ret = vvcc_parse_sps(&gbc, vvcc);
+ else if (nal_type == VVC_PPS_NUT)
+ ret = vvcc_parse_pps(&gbc, vvcc);
+ else if (nal_type == VVC_OPI_NUT) {
+ // not yet supported
+ }
+ if (ret < 0)
+ goto end;
+ break;
+ default:
+ ret = AVERROR_INVALIDDATA;
+ goto end;
+ }
+
+ end:
+ av_free(rbsp_buf);
+ return ret;
+}
+
+static void vvcc_init(VVCDecoderConfigurationRecord *vvcc)
+{
+ memset(vvcc, 0, sizeof(VVCDecoderConfigurationRecord));
+ vvcc->lengthSizeMinusOne = 3; // 4 bytes
+
+ vvcc->ptl.num_bytes_constraint_info = 1;
+
+ vvcc->ptl_present_flag = 1;
+}
+
+static void vvcc_close(VVCDecoderConfigurationRecord *vvcc)
+{
+ uint8_t i;
+
+ for (i = 0; i < vvcc->num_of_arrays; i++) {
+ vvcc->array[i].num_nalus = 0;
+ av_freep(&vvcc->array[i].nal_unit);
+ av_freep(&vvcc->array[i].nal_unit_length);
+ }
+
+ free(vvcc->ptl.ptl_sublayer_level_present_flag);
+ free(vvcc->ptl.sublayer_level_idc);
+ free(vvcc->ptl.general_sub_profile_idc);
+
+ vvcc->num_of_arrays = 0;
+ av_freep(&vvcc->array);
+}
+
+static int vvcc_write(AVIOContext *pb, VVCDecoderConfigurationRecord *vvcc)
+{
+ uint8_t i;
+ uint16_t j, vps_count = 0, sps_count = 0, pps_count = 0;
+ unsigned char *buf = NULL;
+ /*
+ * It's unclear how to properly compute these fields, so
+ * let's always set them to values meaning 'unspecified'.
+ */
+ vvcc->avg_frame_rate = 0;
+ vvcc->constant_frame_rate = 1;
+
+ av_log(NULL, AV_LOG_TRACE,
+ "lengthSizeMinusOne: %" PRIu8 "\n",
+ vvcc->lengthSizeMinusOne);
+ av_log(NULL, AV_LOG_TRACE,
+ "ptl_present_flag: %" PRIu8 "\n",
+ vvcc->ptl_present_flag);
+ av_log(NULL, AV_LOG_TRACE,
+ "ols_idx: %" PRIu16 "\n", vvcc->ols_idx);
+ av_log(NULL, AV_LOG_TRACE,
+ "num_sublayers: %" PRIu8 "\n",
+ vvcc->num_sublayers);
+ av_log(NULL, AV_LOG_TRACE,
+ "constant_frame_rate: %" PRIu8 "\n",
+ vvcc->constant_frame_rate);
+ av_log(NULL, AV_LOG_TRACE,
+ "chroma_format_idc: %" PRIu8 "\n",
+ vvcc->chroma_format_idc);
+
+ av_log(NULL, AV_LOG_TRACE,
+ "bit_depth_minus8: %" PRIu8 "\n",
+ vvcc->bit_depth_minus8);
+ av_log(NULL, AV_LOG_TRACE,
+ "num_bytes_constraint_info: %" PRIu8 "\n",
+ vvcc->ptl.num_bytes_constraint_info);
+ av_log(NULL, AV_LOG_TRACE,
+ "general_profile_idc: %" PRIu8 "\n",
+ vvcc->ptl.general_profile_idc);
+ av_log(NULL, AV_LOG_TRACE,
+ "general_tier_flag: %" PRIu8 "\n",
+ vvcc->ptl.general_tier_flag);
+ av_log(NULL, AV_LOG_TRACE,
+ "general_level_idc: %" PRIu8 "\n",
+ vvcc->ptl.general_level_idc);
+ av_log(NULL, AV_LOG_TRACE,
+ "ptl_frame_only_constraint_flag: %" PRIu8 "\n",
+ vvcc->ptl.ptl_frame_only_constraint_flag);
+ av_log(NULL, AV_LOG_TRACE,
+ "ptl_multilayer_enabled_flag: %" PRIu8 "\n",
+ vvcc->ptl.ptl_multilayer_enabled_flag);
+ for (i = 0; i < vvcc->ptl.num_bytes_constraint_info; i++) {
+ av_log(NULL, AV_LOG_TRACE,
+ "general_constraint_info[%d]: %" PRIu8 "\n", i,
+ vvcc->ptl.general_constraint_info[i]);
+ }
+
+ for (i = 0; i < vvcc->num_sublayers - 1; i++) {
+ av_log(NULL, AV_LOG_TRACE,
+ "ptl_sublayer_level_present_flag[%" PRIu8 "]: %" PRIu8 "\n", i,
+ vvcc->ptl.ptl_sublayer_level_present_flag[i]);
+ av_log(NULL, AV_LOG_TRACE,
+ "sublayer_level_idc[%" PRIu8 "]: %" PRIu8 "\n", i,
+ vvcc->ptl.sublayer_level_idc[i]);
+ }
+
+ av_log(NULL, AV_LOG_TRACE,
+ "num_sub_profiles: %" PRIu8 "\n",
+ vvcc->ptl.ptl_num_sub_profiles);
+
+ for (i = 0; i < vvcc->ptl.ptl_num_sub_profiles; i++) {
+ av_log(NULL, AV_LOG_TRACE,
+ "general_sub_profile_idc[%" PRIu8 "]: %" PRIx32 "\n", i,
+ vvcc->ptl.general_sub_profile_idc[i]);
+ }
+
+ av_log(NULL, AV_LOG_TRACE,
+ "max_picture_width: %" PRIu16 "\n",
+ vvcc->max_picture_width);
+ av_log(NULL, AV_LOG_TRACE,
+ "max_picture_height: %" PRIu16 "\n",
+ vvcc->max_picture_height);
+ av_log(NULL, AV_LOG_TRACE,
+ "avg_frame_rate: %" PRIu16 "\n",
+ vvcc->avg_frame_rate);
+
+ av_log(NULL, AV_LOG_TRACE,
+ "num_of_arrays: %" PRIu8 "\n",
+ vvcc->num_of_arrays);
+ for (i = 0; i < vvcc->num_of_arrays; i++) {
+ av_log(NULL, AV_LOG_TRACE,
+ "array_completeness[%" PRIu8 "]: %" PRIu8 "\n", i,
+ vvcc->array[i].array_completeness);
+ av_log(NULL, AV_LOG_TRACE,
+ "NAL_unit_type[%" PRIu8 "]: %" PRIu8 "\n", i,
+ vvcc->array[i].NAL_unit_type);
+ av_log(NULL, AV_LOG_TRACE,
+ "num_nalus[%" PRIu8 "]: %" PRIu16 "\n", i,
+ vvcc->array[i].num_nalus);
+ for (j = 0; j < vvcc->array[i].num_nalus; j++)
+ av_log(NULL, AV_LOG_TRACE,
+ "nal_unit_length[%" PRIu8 "][%" PRIu16 "]: %"
+ PRIu16 "\n", i, j, vvcc->array[i].nal_unit_length[j]);
+ }
+
+ /*
+ * We need at least one of each: VPS and SPS.
+ */
+ for (i = 0; i < vvcc->num_of_arrays; i++)
+ switch (vvcc->array[i].NAL_unit_type) {
+ case VVC_VPS_NUT:
+ vps_count += vvcc->array[i].num_nalus;
+ break;
+ case VVC_SPS_NUT:
+ sps_count += vvcc->array[i].num_nalus;
+ break;
+ case VVC_PPS_NUT:
+ pps_count += vvcc->array[i].num_nalus;
+ break;
+ default:
+ break;
+ }
+
+ if (!sps_count || sps_count > VVC_MAX_SPS_COUNT)
+ return AVERROR_INVALIDDATA;
+
+ /* bit(5) reserved = ‘11111’b;
+ unsigned int (2) LengthSizeMinusOne
+ unsigned int (1) ptl_present_flag */
+ avio_w8(pb, vvcc->lengthSizeMinusOne << 1 | vvcc->ptl_present_flag | 0xf8);
+
+ if (vvcc->ptl_present_flag) {
+ /*
+ * unsigned int(9) ols_idx;
+ * unsigned int(3) num_sublayers;
+ * unsigned int(2) constant_frame_rate;
+ * unsigned int(2) chroma_format_idc; */
+ avio_wb16(pb,
+ vvcc->ols_idx << 7 | vvcc->num_sublayers << 4 | vvcc->
+ constant_frame_rate << 2 | vvcc->chroma_format_idc);
+
+ /* unsigned int(3) bit_depth_minus8;
+ bit(5) reserved = ‘11111’b; */
+ avio_w8(pb, vvcc->bit_depth_minus8 << 5 | 0x1f);
+
+ //VVCPTLRecord
+
+ /* bit(2) reserved = ‘00’b;
+ unsigned int (6) num_bytes_constraint_info */
+ avio_w8(pb, vvcc->ptl.num_bytes_constraint_info & 0x3f);
+
+ /* unsigned int (7) general_profile_idc
+ unsigned int (1) general_tier_flag */
+ avio_w8(pb,
+ vvcc->ptl.general_profile_idc << 1 | vvcc->ptl.general_tier_flag);
+
+ /* unsigned int (8) general_level_idc */
+ avio_w8(pb, vvcc->ptl.general_level_idc);
+
+ /*
+ * unsigned int (1) ptl_frame_only_constraint_flag
+ * unsigned int (1) ptl_multilayer_enabled_flag
+ * unsigned int (8*num_bytes_constraint_info -2) general_constraint_info */
+ buf =
+ (unsigned char *) malloc(sizeof(unsigned char) *
+ vvcc->ptl.num_bytes_constraint_info);
+ *buf = vvcc->ptl.ptl_frame_only_constraint_flag << vvcc->ptl.
+ num_bytes_constraint_info * 8 - 1 | vvcc->ptl.
+ ptl_multilayer_enabled_flag << vvcc->ptl.num_bytes_constraint_info *
+ 8 - 2 | *vvcc->ptl.general_constraint_info >> 2;
+ avio_write(pb, buf, vvcc->ptl.num_bytes_constraint_info);
+ free(buf);
+
+ if (vvcc->num_sublayers > 1) {
+ uint8_t ptl_sublayer_level_present_flags = 0;
+ for (int i = vvcc->num_sublayers - 2; i >= 0; i--) {
+ ptl_sublayer_level_present_flags =
+ (ptl_sublayer_level_present_flags << 1 | vvcc->ptl.
+ ptl_sublayer_level_present_flag[i]);
+ }
+ avio_w8(pb, ptl_sublayer_level_present_flags);
+ }
+
+ for (int i = vvcc->num_sublayers - 2; i >= 0; i--) {
+ if (vvcc->ptl.ptl_sublayer_level_present_flag[i])
+ avio_w8(pb, vvcc->ptl.sublayer_level_idc[i]);
+ }
+
+ /* unsigned int(8) num_sub_profiles; */
+ avio_w8(pb, vvcc->ptl.ptl_num_sub_profiles);
+
+ for (int j = 0; j < vvcc->ptl.ptl_num_sub_profiles; j++) {
+ /* unsigned int(32) general_sub_profile_idc[j]; */
+ avio_wb32(pb, vvcc->ptl.general_sub_profile_idc[j]);
+ }
+
+ //End of VvcPTLRecord
+
+ /*
+ * unsigned int(16) max_picture_width;*/
+ avio_wb16(pb, vvcc->max_picture_width);
+
+ /*
+ * unsigned int(16) max_picture_height;*/
+ avio_wb16(pb, vvcc->max_picture_height);
+
+ /*
+ * unsigned int(16) avg_frame_rate; */
+ avio_wb16(pb, vvcc->avg_frame_rate);
+ }
+
+ /* unsigned int(8) num_of_arrays; */
+ avio_w8(pb, vvcc->num_of_arrays);
+
+ for (i = 0; i < vvcc->num_of_arrays; i++) {
+ /*
+ * bit(1) array_completeness;
+ * unsigned int(2) reserved = 0;
+ * unsigned int(5) NAL_unit_type;
+ */
+ avio_w8(pb, vvcc->array[i].array_completeness << 7 |
+ vvcc->array[i].NAL_unit_type & 0x1f);
+ /* unsigned int(16) num_nalus; */
+ if (vvcc->array[i].NAL_unit_type != VVC_DCI_NUT &&
+ vvcc->array[i].NAL_unit_type != VVC_OPI_NUT)
+ avio_wb16(pb, vvcc->array[i].num_nalus);
+ for (j = 0; j < vvcc->array[i].num_nalus; j++) {
+ /* unsigned int(16) nal_unit_length; */
+ avio_wb16(pb, vvcc->array[i].nal_unit_length[j]);
+
+ /* bit(8*nal_unit_length) nal_unit; */
+ avio_write(pb, vvcc->array[i].nal_unit[j],
+ vvcc->array[i].nal_unit_length[j]);
+ }
+ }
+
+ return 0;
+}
+
+int ff_h266_annexb2mp4(AVIOContext *pb, const uint8_t *buf_in,
+ int size, int filter_ps, int *ps_count)
+{
+ int num_ps = 0, ret = 0;
+ uint8_t *buf, *end, *start = NULL;
+
+ if (!filter_ps) {
+ ret = ff_avc_parse_nal_units(pb, buf_in, size);
+ goto end;
+ }
+
+ ret = ff_avc_parse_nal_units_buf(buf_in, &start, &size);
+ if (ret < 0)
+ goto end;
+
+ ret = 0;
+ buf = start;
+ end = start + size;
+
+ while (end - buf > 4) {
+ uint32_t len = FFMIN(AV_RB32(buf), end - buf - 4);
+ uint8_t type = (buf[5] >> 3);
+
+ buf += 4;
+
+ switch (type) {
+ case VVC_VPS_NUT:
+ case VVC_SPS_NUT:
+ case VVC_PPS_NUT:
+ num_ps++;
+ break;
+ default:
+ ret += 4 + len;
+ avio_wb32(pb, len);
+ avio_write(pb, buf, len);
+ break;
+ }
+
+ buf += len;
+ }
+
+ end:
+ av_free(start);
+ if (ps_count)
+ *ps_count = num_ps;
+ return ret;
+}
+
+int ff_h266_annexb2mp4_buf(const uint8_t *buf_in, uint8_t **buf_out,
+ int *size, int filter_ps, int *ps_count)
+{
+ AVIOContext *pb;
+ int ret;
+
+ ret = avio_open_dyn_buf(&pb);
+ if (ret < 0)
+ return ret;
+
+ ret = ff_h266_annexb2mp4(pb, buf_in, *size, filter_ps, ps_count);
+ if (ret < 0) {
+ ffio_free_dyn_buf(&pb);
+ return ret;
+ }
+
+ *size = avio_close_dyn_buf(pb, buf_out);
+
+ return 0;
+}
+
+int ff_isom_write_vvcc(AVIOContext *pb, const uint8_t *data,
+ int size, int ps_array_completeness)
+{
+ VVCDecoderConfigurationRecord vvcc;
+ uint8_t *buf, *end, *start;
+ int ret;
+
+ if (size < 6) {
+ /* We can't write a valid vvcc from the provided data */
+ return AVERROR_INVALIDDATA;
+ } else if (*data == 1) {
+ /* Data is already vvcc-formatted */
+ avio_write(pb, data, size);
+ return 0;
+ } else if (!(AV_RB24(data) == 1 || AV_RB32(data) == 1)) {
+ /* Not a valid Annex B start code prefix */
+ return AVERROR_INVALIDDATA;
+ }
+
+ ret = ff_avc_parse_nal_units_buf(data, &start, &size);
+ if (ret < 0)
+ return ret;
+
+ vvcc_init(&vvcc);
+
+ buf = start;
+ end = start + size;
+
+ while (end - buf > 4) {
+ uint32_t len = FFMIN(AV_RB32(buf), end - buf - 4);
+ uint8_t type = (buf[5] >> 3);
+
+ buf += 4;
+
+ switch (type) {
+ case VVC_OPI_NUT:
+ case VVC_VPS_NUT:
+ case VVC_SPS_NUT:
+ case VVC_PPS_NUT:
+ case VVC_PREFIX_SEI_NUT:
+ case VVC_SUFFIX_SEI_NUT:
+ ret = vvcc_add_nal_unit(buf, len, ps_array_completeness, &vvcc);
+ if (ret < 0)
+ goto end;
+ break;
+ default:
+ break;
+ }
+
+ buf += len;
+ }
+
+ ret = vvcc_write(pb, &vvcc);
+
+ end:
+ vvcc_close(&vvcc);
+ av_free(start);
+ return ret;
+}
diff --git a/libavformat/h266.h b/libavformat/h266.h
new file mode 100644
index 0000000000..f58145e4ae
--- /dev/null
+++ b/libavformat/h266.h
@@ -0,0 +1,99 @@
+/*
+ * H.266 / VVC helper functions for muxers
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * internal header for H.266/VVC (de)muxer utilities
+ */
+
+#ifndef AVFORMAT_H266_H
+#define AVFORMAT_H266_H
+
+#include <stdint.h>
+#include "avio.h"
+
+/**
+ * Writes Annex B formatted H.266/VVC NAL units to the provided AVIOContext.
+ *
+ * The NAL units are converted to an MP4-compatible format (start code prefixes
+ * are replaced by 4-byte size fields, as per ISO/IEC 14496-15).
+ *
+ * If filter_ps is non-zero, any VVC parameter sets found in the input will be
+ * discarded, and *ps_count will be set to the number of discarded PS NAL units.
+ *
+ * @param pb address of the AVIOContext where the data shall be written
+ * @param buf_in address of the buffer holding the input data
+ * @param size size (in bytes) of the input buffer
+ * @param filter_ps whether to write parameter set NAL units to the output (0)
+ * or to discard them (non-zero)
+ * @param ps_count address of the variable where the number of discarded
+ * parameter set NAL units shall be written, may be NULL
+ * @return the amount (in bytes) of data written in case of success, a negative
+ * value corresponding to an AVERROR code in case of failure
+ */
+int ff_h266_annexb2mp4(AVIOContext *pb, const uint8_t *buf_in,
+ int size, int filter_ps, int *ps_count);
+
+/**
+ * Writes Annex B formatted H.266/VVC NAL units to a data buffer.
+ *
+ * The NAL units are converted to an MP4-compatible format (start code prefixes
+ * are replaced by 4-byte size fields, as per ISO/IEC 14496-15).
+ *
+ * If filter_ps is non-zero, any VVC parameter sets found in the input will be
+ * discarded, and *ps_count will be set to the number of discarded PS NAL units.
+ *
+ * On success, *size holds the size (in bytes) of the output data buffer.
+ *
+ * @param buf_in address of the buffer holding the input data
+ * @param size address of the variable holding the size (in bytes) of the input
+ * buffer (on input) and of the output buffer (on success)
+ * @param buf_out on success, address of the variable holding the address of
+ * the output buffer
+ * @param filter_ps whether to write parameter set NAL units to the output (0)
+ * or to discard them (non-zero)
+ * @param ps_count address of the variable where the number of discarded
+ * parameter set NAL units shall be written, may be NULL
+ * @return 0 in case of success, a negative value corresponding to an AVERROR
+ * code in case of failure
+ * @note *buf_out will be treated as uninitialized on input and won't be freed.
+ */
+int ff_h266_annexb2mp4_buf(const uint8_t *buf_in, uint8_t **buf_out,
+ int *size, int filter_ps, int *ps_count);
+
+/**
+ * Writes H.266/VVC extradata (parameter sets, declarative SEI NAL units) to
+ * the provided AVIOContext.
+ *
+ * If the extradata is Annex B format, it gets converted to vvcC format before
+ * writing.
+ *
+ * @param pb address of the AVIOContext where the vvcC shall be written
+ * @param data address of the buffer holding the data needed to write the vvcC
+ * @param size size (in bytes) of the data buffer
+ * @param ps_array_completeness whether all parameter sets are in the vvcC (1)
+ * or there may be additional parameter sets in the bitstream (0)
+ * @return >=0 in case of success, a negative value corresponding to an AVERROR
+ * code in case of failure
+ */
+int ff_isom_write_vvcc(AVIOContext *pb, const uint8_t *data,
+ int size, int ps_array_completeness);
+
+#endif /* AVFORMAT_H266_H */
diff --git a/libavformat/h266dec.c b/libavformat/h266dec.c
new file mode 100644
index 0000000000..dbda4b868a
--- /dev/null
+++ b/libavformat/h266dec.c
@@ -0,0 +1,61 @@
+/*
+ * RAW H.266 / VVC video demuxer
+ * Copyright (c) 2020 Nuo Mi <nuomi2021@gmail.com>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavcodec/h266.h"
+
+#include "avformat.h"
+#include "rawdec.h"
+
+static int h266_probe(const AVProbeData *p)
+{
+ uint32_t code = -1;
+ int sps = 0, pps = 0, irap = 0;
+ int i;
+
+ for (i = 0; i < p->buf_size - 1; i++) {
+ code = (code << 8) + p->buf[i];
+ if ((code & 0xffffff00) == 0x100) {
+ uint8_t nal2 = p->buf[i + 1];
+ int type = (nal2 & 0xF8) >> 3;
+
+ if (code & 0x80) // forbidden_zero_bit
+ return 0;
+
+ if ((nal2 & 0x7) == 0) // nuh_temporal_id_plus1
+ return 0;
+
+ switch (type) {
+ case VVC_SPS_NUT: sps++; break;
+ case VVC_PPS_NUT: pps++; break;
+ case VVC_IDR_N_LP:
+ case VVC_IDR_W_RADL:
+ case VVC_CRA_NUT:
+ case VVC_GDR_NUT: irap++; break;
+ }
+ }
+ }
+
+ if (sps && pps && irap)
+ return AVPROBE_SCORE_EXTENSION + 1; // 1 more than .mpg
+ return 0;
+}
+
+FF_DEF_RAWVIDEO_DEMUXER(h266, "raw H.266/VVC video", h266_probe, "h266,266,vvc", AV_CODEC_ID_VVC)
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [FFmpeg-devel] [PATCH v7 06/11] avformat: add demuxer and probe support for H266/VVC
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 06/11] avformat: add demuxer and probe " Thomas Siedel
@ 2023-12-18 11:43 ` Zhao Zhili
2023-12-18 12:25 ` Nuo Mi
0 siblings, 1 reply; 19+ messages in thread
From: Zhao Zhili @ 2023-12-18 11:43 UTC (permalink / raw)
To: FFmpeg development discussions and patches
> On Mar 21, 2023, at 23:01, Thomas Siedel <thomas.ff@spin-digital.com> wrote:
>
> Add demuxer to probe raw vvc and parse vvcc byte stream format.
>
> Co-authored-by: Nuo Mi <nuomi2021@gmail.com>
> ---
>
> +
> +static int h266_probe(const AVProbeData *p)
> +{
> + uint32_t code = -1;
> + int sps = 0, pps = 0, irap = 0;
> + int i;
> +
> + for (i = 0; i < p->buf_size - 1; i++) {
> + code = (code << 8) + p->buf[i];
> + if ((code & 0xffffff00) == 0x100) {
> + uint8_t nal2 = p->buf[i + 1];
> + int type = (nal2 & 0xF8) >> 3;
> +
> + if (code & 0x80) // forbidden_zero_bit
> + return 0;
> +
> + if ((nal2 & 0x7) == 0) // nuh_temporal_id_plus1
> + return 0;
> +
> + switch (type) {
> + case VVC_SPS_NUT: sps++; break;
> + case VVC_PPS_NUT: pps++; break;
> + case VVC_IDR_N_LP:
> + case VVC_IDR_W_RADL:
> + case VVC_CRA_NUT:
> + case VVC_GDR_NUT: irap++; break;
> + }
> + }
> + }
> +
> + if (sps && pps && irap)
> + return AVPROBE_SCORE_EXTENSION + 1; // 1 more than .mpg
The probe isn’t reliable enough. It mistreated an TS as vvc,
https://trac.ffmpeg.org/ticket/10731#comment:1
Any idea to reduce false positive?
> + return 0;
> +}
> +
> +FF_DEF_RAWVIDEO_DEMUXER(h266, "raw H.266/VVC video", h266_probe, "h266,266,vvc", AV_CODEC_ID_VVC)
> --
> 2.25.1
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [FFmpeg-devel] [PATCH v7 06/11] avformat: add demuxer and probe support for H266/VVC
2023-12-18 11:43 ` Zhao Zhili
@ 2023-12-18 12:25 ` Nuo Mi
2023-12-19 4:01 ` Zhao Zhili
0 siblings, 1 reply; 19+ messages in thread
From: Nuo Mi @ 2023-12-18 12:25 UTC (permalink / raw)
To: FFmpeg development discussions and patches
On Mon, Dec 18, 2023 at 7:43 PM Zhao Zhili <quinkblack@foxmail.com> wrote:
>
>
> > On Mar 21, 2023, at 23:01, Thomas Siedel <thomas.ff@spin-digital.com>
> wrote:
> >
> > Add demuxer to probe raw vvc and parse vvcc byte stream format.
> >
> > Co-authored-by: Nuo Mi <nuomi2021@gmail.com>
> > ---
> >
>
>
> > +
> > +static int h266_probe(const AVProbeData *p)
> > +{
> > + uint32_t code = -1;
> > + int sps = 0, pps = 0, irap = 0;
> > + int i;
> > +
> > + for (i = 0; i < p->buf_size - 1; i++) {
> > + code = (code << 8) + p->buf[i];
> > + if ((code & 0xffffff00) == 0x100) {
> > + uint8_t nal2 = p->buf[i + 1];
> > + int type = (nal2 & 0xF8) >> 3;
> > +
> > + if (code & 0x80) // forbidden_zero_bit
> > + return 0;
> > +
> > + if ((nal2 & 0x7) == 0) // nuh_temporal_id_plus1
> > + return 0;
> > +
> > + switch (type) {
> > + case VVC_SPS_NUT: sps++; break;
> > + case VVC_PPS_NUT: pps++; break;
> > + case VVC_IDR_N_LP:
> > + case VVC_IDR_W_RADL:
> > + case VVC_CRA_NUT:
> > + case VVC_GDR_NUT: irap++; break;
> > + }
> > + }
> > + }
> > +
> > + if (sps && pps && irap)
> > + return AVPROBE_SCORE_EXTENSION + 1; // 1 more than .mpg
>
> The probe isn’t reliable enough. It mistreated an TS as vvc,
>
> https://trac.ffmpeg.org/ticket/10731#comment:1
>
> Any idea to reduce false positive?
>
> > + return 0;
> > +}
> > +
> > +FF_DEF_RAWVIDEO_DEMUXER(h266, "raw H.266/VVC video", h266_probe,
> "h266,266,vvc", AV_CODEC_ID_VVC)
> > --
> > 2.25.1
It will fix by
https://patchwork.ffmpeg.org/project/ffmpeg/patch/20231205094038.12293-1-post@frankplowman.com/
Could you help merge and backport to 6.1
https://trac.ffmpeg.org/ticket/10703
thank you.
> >
> > _______________________________________________
> > ffmpeg-devel mailing list
> > ffmpeg-devel@ffmpeg.org
> > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> >
> > To unsubscribe, visit link above, or email
> > ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [FFmpeg-devel] [PATCH v7 06/11] avformat: add demuxer and probe support for H266/VVC
2023-12-18 12:25 ` Nuo Mi
@ 2023-12-19 4:01 ` Zhao Zhili
0 siblings, 0 replies; 19+ messages in thread
From: Zhao Zhili @ 2023-12-19 4:01 UTC (permalink / raw)
To: FFmpeg development discussions and patches
> On Dec 18, 2023, at 20:25, Nuo Mi <nuomi2021@gmail.com> wrote:
>
> On Mon, Dec 18, 2023 at 7:43 PM Zhao Zhili <quinkblack@foxmail.com> wrote:
>
>>
>>
>>> On Mar 21, 2023, at 23:01, Thomas Siedel <thomas.ff@spin-digital.com>
>> wrote:
>>>
>>> Add demuxer to probe raw vvc and parse vvcc byte stream format.
>>>
>>> Co-authored-by: Nuo Mi <nuomi2021@gmail.com>
>>> ---
>>>
>>
>
> It will fix by
> https://patchwork.ffmpeg.org/project/ffmpeg/patch/20231205094038.12293-1-post@frankplowman.com/
> Could you help merge and backport to 6.1
> https://trac.ffmpeg.org/ticket/10703
Applied on master branch and cherry-picked to 6.1.
>
> thank you.
>
>
>>>
>>> _______________________________________________
>>> ffmpeg-devel mailing list
>>> ffmpeg-devel@ffmpeg.org
>>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>>
>>> To unsubscribe, visit link above, or email
>>> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>>
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel@ffmpeg.org
>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>
>> To unsubscribe, visit link above, or email
>> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread
* [FFmpeg-devel] [PATCH v7 07/11] avformat: add muxer support for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (5 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 06/11] avformat: add demuxer and probe " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 08/11] avcodec: add external decoder libvvdec " Thomas Siedel
` (4 subsequent siblings)
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
Add muxer for vvcc byte stream format.
Add AV_CODEC_ID_VVC to ff_mp4_obj_type.
Add AV_CODEC_ID_VVC to ISO Media codec (VvcConfigurationBox vvi1,
vvc1 defined in ISO/IEC 14496-15:2021).
Add VvcConfigurationBox vvcC which extends FullBox type in
ISO/IEC 14496-15:2021.
Add ff_vvc_muxer to RAW muxers.
---
libavformat/Makefile | 7 ++++---
libavformat/allformats.c | 1 +
libavformat/isom.c | 1 +
libavformat/isom_tags.c | 3 +++
libavformat/mov.c | 6 ++++++
libavformat/movenc.c | 41 +++++++++++++++++++++++++++++++++++++++-
libavformat/rawenc.c | 23 ++++++++++++++++++++++
7 files changed, 78 insertions(+), 4 deletions(-)
diff --git a/libavformat/Makefile b/libavformat/Makefile
index 9d8efe1936..15a013f3fc 100644
--- a/libavformat/Makefile
+++ b/libavformat/Makefile
@@ -246,6 +246,7 @@ OBJS-$(CONFIG_H263_MUXER) += rawenc.o
OBJS-$(CONFIG_H264_DEMUXER) += h264dec.o rawdec.o
OBJS-$(CONFIG_H264_MUXER) += rawenc.o
OBJS-$(CONFIG_H266_DEMUXER) += h266dec.o rawdec.o
+OBJS-$(CONFIG_H266_MUXER) += rawenc.o
OBJS-$(CONFIG_HASH_MUXER) += hashenc.o
OBJS-$(CONFIG_HCA_DEMUXER) += hca.o
OBJS-$(CONFIG_HCOM_DEMUXER) += hcom.o pcm.o
@@ -337,7 +338,7 @@ OBJS-$(CONFIG_MATROSKA_DEMUXER) += matroskadec.o matroska.o \
oggparsevorbis.o vorbiscomment.o \
qtpalette.o replaygain.o dovi_isom.o
OBJS-$(CONFIG_MATROSKA_MUXER) += matroskaenc.o matroska.o \
- av1.o avc.o hevc.o \
+ av1.o avc.o hevc.o h266.o\
flacenc_header.o avlanguage.o \
vorbiscomment.o wv.o dovi_isom.o
OBJS-$(CONFIG_MCA_DEMUXER) += mca.o
@@ -359,7 +360,7 @@ OBJS-$(CONFIG_MODS_DEMUXER) += mods.o
OBJS-$(CONFIG_MOFLEX_DEMUXER) += moflex.o
OBJS-$(CONFIG_MOV_DEMUXER) += mov.o mov_chan.o mov_esds.o \
qtpalette.o replaygain.o dovi_isom.o
-OBJS-$(CONFIG_MOV_MUXER) += movenc.o av1.o avc.o hevc.o vpcc.o \
+OBJS-$(CONFIG_MOV_MUXER) += movenc.o av1.o avc.o hevc.o h266.o vpcc.o \
movenchint.o mov_chan.o rtp.o \
movenccenc.o movenc_ttml.o rawutils.o \
dovi_isom.o
@@ -510,7 +511,7 @@ OBJS-$(CONFIG_RTP_MUXER) += rtp.o \
rtpenc_vp8.o \
rtpenc_vp9.o \
rtpenc_xiph.o \
- avc.o hevc.o
+ avc.o hevc.o h266.o
OBJS-$(CONFIG_RTSP_DEMUXER) += rtsp.o rtspdec.o httpauth.o \
urldecode.o
OBJS-$(CONFIG_RTSP_MUXER) += rtsp.o rtspenc.o httpauth.o \
diff --git a/libavformat/allformats.c b/libavformat/allformats.c
index 855849449d..2ab0061139 100644
--- a/libavformat/allformats.c
+++ b/libavformat/allformats.c
@@ -200,6 +200,7 @@ extern const FFOutputFormat ff_h263_muxer;
extern const AVInputFormat ff_h264_demuxer;
extern const FFOutputFormat ff_h264_muxer;
extern const AVInputFormat ff_h266_demuxer;
+extern const FFOutputFormat ff_h266_muxer;
extern const FFOutputFormat ff_hash_muxer;
extern const AVInputFormat ff_hca_demuxer;
extern const AVInputFormat ff_hcom_demuxer;
diff --git a/libavformat/isom.c b/libavformat/isom.c
index 6d019881e5..9fbccd4437 100644
--- a/libavformat/isom.c
+++ b/libavformat/isom.c
@@ -36,6 +36,7 @@ const AVCodecTag ff_mp4_obj_type[] = {
{ AV_CODEC_ID_MPEG4 , 0x20 },
{ AV_CODEC_ID_H264 , 0x21 },
{ AV_CODEC_ID_HEVC , 0x23 },
+ { AV_CODEC_ID_VVC , 0x33 },
{ AV_CODEC_ID_AAC , 0x40 },
{ AV_CODEC_ID_MP4ALS , 0x40 }, /* 14496-3 ALS */
{ AV_CODEC_ID_MPEG2VIDEO , 0x61 }, /* MPEG-2 Main */
diff --git a/libavformat/isom_tags.c b/libavformat/isom_tags.c
index 86c7272525..5da9d03b90 100644
--- a/libavformat/isom_tags.c
+++ b/libavformat/isom_tags.c
@@ -123,6 +123,9 @@ const AVCodecTag ff_codec_movvideo_tags[] = {
{ AV_CODEC_ID_HEVC, MKTAG('d', 'v', 'h', 'e') }, /* HEVC-based Dolby Vision derived from hev1 */
/* dvh1 is handled within mov.c */
+ { AV_CODEC_ID_VVC, MKTAG('v', 'v', 'i', '1') }, /* VVC/H.266 which indicates parameter sets may be in ES */
+ { AV_CODEC_ID_VVC, MKTAG('v', 'v', 'c', '1') }, /* VVC/H.266 which indicates parameter shall not be in ES */
+
{ AV_CODEC_ID_H264, MKTAG('a', 'v', 'c', '1') }, /* AVC-1/H.264 */
{ AV_CODEC_ID_H264, MKTAG('a', 'v', 'c', '2') },
{ AV_CODEC_ID_H264, MKTAG('a', 'v', 'c', '3') },
diff --git a/libavformat/mov.c b/libavformat/mov.c
index 057fd872b1..5567593771 100644
--- a/libavformat/mov.c
+++ b/libavformat/mov.c
@@ -2107,6 +2107,11 @@ static int mov_read_glbl(MOVContext *c, AVIOContext *pb, MOVAtom atom)
if ((uint64_t)atom.size > (1<<30))
return AVERROR_INVALIDDATA;
+ if (atom.type == MKTAG('v','v','c','C')) {
+ avio_rb32(pb);
+ atom.size -= 4;
+ }
+
if (atom.size >= 10) {
// Broken files created by legacy versions of libavformat will
// wrap a whole fiel atom inside of a glbl atom.
@@ -7905,6 +7910,7 @@ static const MOVParseTableEntry mov_default_parse_table[] = {
{ MKTAG('s','g','p','d'), mov_read_sgpd },
{ MKTAG('s','b','g','p'), mov_read_sbgp },
{ MKTAG('h','v','c','C'), mov_read_glbl },
+{ MKTAG('v','v','c','C'), mov_read_glbl },
{ MKTAG('u','u','i','d'), mov_read_uuid },
{ MKTAG('C','i','n', 0x8e), mov_read_targa_y216 },
{ MKTAG('f','r','e','e'), mov_read_free },
diff --git a/libavformat/movenc.c b/libavformat/movenc.c
index 68e7f8222b..4844d54990 100644
--- a/libavformat/movenc.c
+++ b/libavformat/movenc.c
@@ -58,6 +58,7 @@
#include "libavutil/timecode.h"
#include "libavutil/dovi_meta.h"
#include "libavutil/uuid.h"
+#include "h266.h"
#include "hevc.h"
#include "rtpenc.h"
#include "mov_chan.h"
@@ -1455,6 +1456,23 @@ static int mov_write_hvcc_tag(AVIOContext *pb, MOVTrack *track)
return update_size(pb, pos);
}
+static int mov_write_vvcc_tag(AVIOContext *pb, MOVTrack *track)
+{
+ int64_t pos = avio_tell(pb);
+
+ avio_wb32(pb, 0);
+ ffio_wfourcc(pb, "vvcC");
+
+ avio_w8 (pb, 0); /* version */
+ avio_wb24(pb, 0); /* flags */
+
+ if (track->tag == MKTAG('v','v','c','1'))
+ ff_isom_write_vvcc(pb, track->vos_data, track->vos_len, 1);
+ else
+ ff_isom_write_vvcc(pb, track->vos_data, track->vos_len, 0);
+ return update_size(pb, pos);
+}
+
/* also used by all avid codecs (dv, imx, meridien) and their variants */
static int mov_write_avid_tag(AVIOContext *pb, MOVTrack *track)
{
@@ -2348,6 +2366,8 @@ static int mov_write_video_tag(AVFormatContext *s, AVIOContext *pb, MOVMuxContex
avid = 1;
} else if (track->par->codec_id == AV_CODEC_ID_HEVC)
mov_write_hvcc_tag(pb, track);
+ else if (track->par->codec_id == AV_CODEC_ID_VVC)
+ mov_write_vvcc_tag(pb, track);
else if (track->par->codec_id == AV_CODEC_ID_H264 && !TAG_IS_AVCI(track->tag)) {
mov_write_avcc_tag(pb, track);
if (track->mode == MODE_IPOD)
@@ -6122,6 +6142,7 @@ int ff_mov_write_packet(AVFormatContext *s, AVPacket *pkt)
if ((par->codec_id == AV_CODEC_ID_DNXHD ||
par->codec_id == AV_CODEC_ID_H264 ||
par->codec_id == AV_CODEC_ID_HEVC ||
+ par->codec_id == AV_CODEC_ID_VVC ||
par->codec_id == AV_CODEC_ID_VP9 ||
par->codec_id == AV_CODEC_ID_TRUEHD) && !trk->vos_len &&
!TAG_IS_AVCI(trk->tag)) {
@@ -6186,6 +6207,18 @@ int ff_mov_write_packet(AVFormatContext *s, AVPacket *pkt)
size = ff_hevc_annexb2mp4(pb, pkt->data, pkt->size, 0, NULL);
}
}
+ } else if (par->codec_id == AV_CODEC_ID_VVC && trk->vos_len > 6 &&
+ (AV_RB24(trk->vos_data) == 1 || AV_RB32(trk->vos_data) == 1)) {
+ /* extradata is Annex B, assume the bitstream is too and convert it */
+ if (trk->hint_track >= 0 && trk->hint_track < mov->nb_streams) {
+ ret = ff_h266_annexb2mp4_buf(pkt->data, &reformatted_data,
+ &size, 0, NULL);
+ if (ret < 0)
+ return ret;
+ avio_write(pb, reformatted_data, size);
+ } else {
+ size = ff_h266_annexb2mp4(pb, pkt->data, pkt->size, 0, NULL);
+ }
} else if (par->codec_id == AV_CODEC_ID_AV1) {
if (trk->hint_track >= 0 && trk->hint_track < mov->nb_streams) {
ret = ff_av1_filter_obus_buf(pkt->data, &reformatted_data,
@@ -6232,6 +6265,9 @@ int ff_mov_write_packet(AVFormatContext *s, AVPacket *pkt)
} else if(par->codec_id == AV_CODEC_ID_HEVC && par->extradata_size > 21) {
int nal_size_length = (par->extradata[21] & 0x3) + 1;
ret = ff_mov_cenc_avc_write_nal_units(s, &trk->cenc, nal_size_length, pb, pkt->data, size);
+ } else if(par->codec_id == AV_CODEC_ID_VVC && par->extradata_size > 21) {
+ int nal_size_length = (par->extradata[21] & 0x3) + 1;
+ ret = ff_mov_cenc_avc_write_nal_units(s, &trk->cenc, nal_size_length, pb, pkt->data, size);
} else {
ret = ff_mov_cenc_write_packet(&trk->cenc, pb, pkt->data, size);
}
@@ -7314,7 +7350,8 @@ static int mov_init(AVFormatContext *s)
if (mov->encryption_scheme == MOV_ENC_CENC_AES_CTR) {
ret = ff_mov_cenc_init(&track->cenc, mov->encryption_key,
- (track->par->codec_id == AV_CODEC_ID_H264 || track->par->codec_id == AV_CODEC_ID_HEVC),
+ (track->par->codec_id == AV_CODEC_ID_H264 || track->par->codec_id == AV_CODEC_ID_HEVC ||
+ track->par->codec_id == AV_CODEC_ID_VVC),
s->flags & AVFMT_FLAG_BITEXACT);
if (ret)
return ret;
@@ -7782,6 +7819,8 @@ static const AVCodecTag codec_mp4_tags[] = {
{ AV_CODEC_ID_H264, MKTAG('a', 'v', 'c', '3') },
{ AV_CODEC_ID_HEVC, MKTAG('h', 'e', 'v', '1') },
{ AV_CODEC_ID_HEVC, MKTAG('h', 'v', 'c', '1') },
+ { AV_CODEC_ID_VVC, MKTAG('v', 'v', 'c', '1') },
+ { AV_CODEC_ID_VVC, MKTAG('v', 'v', 'i', '1') },
{ AV_CODEC_ID_MPEG2VIDEO, MKTAG('m', 'p', '4', 'v') },
{ AV_CODEC_ID_MPEG1VIDEO, MKTAG('m', 'p', '4', 'v') },
{ AV_CODEC_ID_MJPEG, MKTAG('m', 'p', '4', 'v') },
diff --git a/libavformat/rawenc.c b/libavformat/rawenc.c
index 78fadda967..c5176d9a5e 100644
--- a/libavformat/rawenc.c
+++ b/libavformat/rawenc.c
@@ -378,6 +378,29 @@ const FFOutputFormat ff_h264_muxer = {
};
#endif
+#if CONFIG_H266_MUXER
+static int h266_check_bitstream(AVFormatContext *s, AVStream *st,
+ const AVPacket *pkt)
+{
+ if (pkt->size >= 5 && AV_RB32(pkt->data) != 0x0000001 &&
+ AV_RB24(pkt->data) != 0x000001)
+ return ff_stream_add_bitstream_filter(st, "h266_mp4toannexb", NULL);
+ return 1;
+}
+
+const FFOutputFormat ff_h266_muxer = {
+ .p.name = "vvc",
+ .p.long_name = NULL_IF_CONFIG_SMALL("raw H.266/VVC video"),
+ .p.extensions = "vvc,h266,266",
+ .p.audio_codec = AV_CODEC_ID_NONE,
+ .p.video_codec = AV_CODEC_ID_VVC,
+ .init = force_one_stream,
+ .write_packet = ff_raw_write_packet,
+ .check_bitstream = h266_check_bitstream,
+ .p.flags = AVFMT_NOTIMESTAMPS,
+};
+#endif
+
#if CONFIG_HEVC_MUXER
static int hevc_check_bitstream(AVFormatContext *s, AVStream *st,
const AVPacket *pkt)
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 08/11] avcodec: add external decoder libvvdec for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (6 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 07/11] avformat: add muxer " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 09/11] avcodec: add external encoder libvvenc " Thomas Siedel
` (3 subsequent siblings)
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
Add external decoder VVdeC for H266/VVC decoding.
Register new decoder libvvdec.
Add vvc_parse_extradata to support parse/probe of vvcC stream input.
Add vvc_paramset that implements the parser of vvcC configuration boxes.
Add libvvdec to wrap the vvdec interface.
Enable decoder by adding --enable-libvvdec in configure step.
---
configure | 5 +
libavcodec/Makefile | 1 +
libavcodec/allcodecs.c | 1 +
libavcodec/h266_paramset.c | 1005 +++++++++++++++++++++++++++++
libavcodec/h266_paramset.h | 307 +++++++++
| 246 +++++++
| 36 ++
libavcodec/libvvdec.c | 563 ++++++++++++++++
8 files changed, 2164 insertions(+)
create mode 100644 libavcodec/h266_paramset.c
create mode 100644 libavcodec/h266_paramset.h
create mode 100644 libavcodec/h266_parse_extradata.c
create mode 100644 libavcodec/h266_parse_extradata.h
create mode 100644 libavcodec/libvvdec.c
diff --git a/configure b/configure
index 297ce6e27f..6568493849 100755
--- a/configure
+++ b/configure
@@ -288,6 +288,7 @@ External library support:
--enable-libvorbis enable Vorbis en/decoding via libvorbis,
native implementation exists [no]
--enable-libvpx enable VP8 and VP9 de/encoding via libvpx [no]
+ --enable-libvvdec enable H.266/VVC decoding via vvdec [no]
--enable-libwebp enable WebP encoding via libwebp [no]
--enable-libx264 enable H.264 encoding via x264 [no]
--enable-libx265 enable HEVC encoding via x265 [no]
@@ -1861,6 +1862,7 @@ EXTERNAL_LIBRARY_LIST="
libvmaf
libvorbis
libvpx
+ libvvdec
libwebp
libxml2
libzimg
@@ -3391,6 +3393,8 @@ libvpx_vp8_decoder_deps="libvpx"
libvpx_vp8_encoder_deps="libvpx"
libvpx_vp9_decoder_deps="libvpx"
libvpx_vp9_encoder_deps="libvpx"
+libvvdec_decoder_deps="libvvdec"
+libvvdec_decoder_select="h266_mp4toannexb_bsf"
libwebp_encoder_deps="libwebp"
libwebp_anim_encoder_deps="libwebp"
libx262_encoder_deps="libx262"
@@ -6724,6 +6728,7 @@ enabled libvpx && {
die "libvpx enabled but no supported decoders found"
fi
}
+enabled libvvdec && require_pkg_config libvvdec "libvvdec >= 1.6.0" "vvdec/vvdec.h" vvdec_get_version
enabled libwebp && {
enabled libwebp_encoder && require_pkg_config libwebp "libwebp >= 0.2.0" webp/encode.h WebPGetEncoderVersion
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 81fda09597..2e7fb2ae12 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -1124,6 +1124,7 @@ OBJS-$(CONFIG_LIBVPX_VP8_DECODER) += libvpxdec.o
OBJS-$(CONFIG_LIBVPX_VP8_ENCODER) += libvpxenc.o
OBJS-$(CONFIG_LIBVPX_VP9_DECODER) += libvpxdec.o
OBJS-$(CONFIG_LIBVPX_VP9_ENCODER) += libvpxenc.o
+OBJS-$(CONFIG_LIBVVDEC_DECODER) += libvvdec.o h266_parse_extradata.o h266_paramset.o
OBJS-$(CONFIG_LIBWEBP_ENCODER) += libwebpenc_common.o libwebpenc.o
OBJS-$(CONFIG_LIBWEBP_ANIM_ENCODER) += libwebpenc_common.o libwebpenc_animencoder.o
OBJS-$(CONFIG_LIBX262_ENCODER) += libx264.o
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index 385ee34803..18bbe62808 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -799,6 +799,7 @@ extern const FFCodec ff_libvpx_vp8_encoder;
extern const FFCodec ff_libvpx_vp8_decoder;
extern FFCodec ff_libvpx_vp9_encoder;
extern const FFCodec ff_libvpx_vp9_decoder;
+extern const FFCodec ff_libvvdec_decoder;
/* preferred over libwebp */
extern const FFCodec ff_libwebp_anim_encoder;
extern const FFCodec ff_libwebp_encoder;
diff --git a/libavcodec/h266_paramset.c b/libavcodec/h266_paramset.c
new file mode 100644
index 0000000000..4e0e958750
--- /dev/null
+++ b/libavcodec/h266_paramset.c
@@ -0,0 +1,1005 @@
+/*
+ * H.266 / VVC Parameter Set decoding
+ *
+ * Copyright (c) 2022, Thomas Siedel
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/imgutils.h"
+#include "golomb.h"
+#include "h266_paramset.h"
+
+static void remove_sps(H266ParamSets *s, int id)
+{
+ if (s->sps_list[id]) {
+ if (s->sps == (const H266SPS *)s->sps_list[id]->data)
+ s->sps = NULL;
+
+ av_assert0(!(s->sps_list[id] &&
+ s->sps == (H266SPS *)s->sps_list[id]->data));
+ }
+ av_buffer_unref(&s->sps_list[id]);
+}
+
+static int decode_general_constraints_info(GetBitContext *gb,
+ AVCodecContext *avctx,
+ H266GeneralConstraintsInfo *gci)
+{
+ int i;
+ gci->gci_present_flag = get_bits1(gb);
+
+ if (gci->gci_present_flag) {
+ /* general */
+ gci->gci_intra_only_constraint_flag = get_bits1(gb);
+ gci->gci_all_layers_independent_constraint_flag = get_bits1(gb);
+ gci->gci_one_au_only_constraint_flag = get_bits1(gb);
+
+ /* picture format */
+ gci->gci_sixteen_minus_max_bitdepth_constraint_idc = get_bits(gb, 4);
+ gci->gci_three_minus_max_chroma_format_constraint_idc = get_bits(gb, 2);
+
+ /* NAL unit type related */
+ gci->gci_no_mixed_nalu_types_in_pic_constraint_flag = get_bits1(gb);
+ gci->gci_no_trail_constraint_flag = get_bits1(gb);
+ gci->gci_no_stsa_constraint_flag = get_bits1(gb);
+ gci->gci_no_rasl_constraint_flag = get_bits1(gb);
+ gci->gci_no_radl_constraint_flag = get_bits1(gb);
+ gci->gci_no_idr_constraint_flag = get_bits1(gb);
+ gci->gci_no_cra_constraint_flag = get_bits1(gb);
+ gci->gci_no_gdr_constraint_flag = get_bits1(gb);
+ gci->gci_no_aps_constraint_flag = get_bits1(gb);
+ gci->gci_no_idr_rpl_constraint_flag = get_bits1(gb);
+
+ /* tile, slice, subpicture partitioning */
+ gci->gci_one_tile_per_pic_constraint_flag = get_bits1(gb);
+ gci->gci_pic_header_in_slice_header_constraint_flag = get_bits1(gb);
+ gci->gci_one_slice_per_pic_constraint_flag = get_bits1(gb);
+ gci->gci_no_rectangular_slice_constraint_flag = get_bits1(gb);
+ gci->gci_one_slice_per_subpic_constraint_flag = get_bits1(gb);
+ gci->gci_no_subpic_info_constraint_flag = get_bits1(gb);
+
+ /* CTU and block partitioning */
+ gci->gci_three_minus_max_log2_ctu_size_constraint_idc = get_bits(gb, 2);
+ gci->gci_no_partition_constraints_override_constraint_flag =
+ get_bits1(gb);
+ gci->gci_no_mtt_constraint_flag = get_bits1(gb);
+ gci->gci_no_qtbtt_dual_tree_intra_constraint_flag = get_bits1(gb);
+
+ /* intra */
+ gci->gci_no_palette_constraint_flag = get_bits1(gb);
+ gci->gci_no_ibc_constraint_flag = get_bits1(gb);
+ gci->gci_no_isp_constraint_flag = get_bits1(gb);
+ gci->gci_no_mrl_constraint_flag = get_bits1(gb);
+ gci->gci_no_mip_constraint_flag = get_bits1(gb);
+ gci->gci_no_cclm_constraint_flag = get_bits1(gb);
+
+ /* inter */
+ gci->gci_no_ref_pic_resampling_constraint_flag = get_bits1(gb);
+ gci->gci_no_res_change_in_clvs_constraint_flag = get_bits1(gb);
+ gci->gci_no_weighted_prediction_constraint_flag = get_bits1(gb);
+ gci->gci_no_ref_wraparound_constraint_flag = get_bits1(gb);
+ gci->gci_no_temporal_mvp_constraint_flag = get_bits1(gb);
+ gci->gci_no_sbtmvp_constraint_flag = get_bits1(gb);
+ gci->gci_no_amvr_constraint_flag = get_bits1(gb);
+ gci->gci_no_bdof_constraint_flag = get_bits1(gb);
+ gci->gci_no_smvd_constraint_flag = get_bits1(gb);
+ gci->gci_no_dmvr_constraint_flag = get_bits1(gb);
+ gci->gci_no_mmvd_constraint_flag = get_bits1(gb);
+ gci->gci_no_affine_motion_constraint_flag = get_bits1(gb);
+ gci->gci_no_prof_constraint_flag = get_bits1(gb);
+ gci->gci_no_bcw_constraint_flag = get_bits1(gb);
+ gci->gci_no_ciip_constraint_flag = get_bits1(gb);
+ gci->gci_no_gpm_constraint_flag = get_bits1(gb);
+
+ /* transform, quantization, residual */
+ gci->gci_no_luma_transform_size_64_constraint_flag = get_bits1(gb);
+ gci->gci_no_transform_skip_constraint_flag = get_bits1(gb);
+ gci->gci_no_bdpcm_constraint_flag = get_bits1(gb);
+ gci->gci_no_mts_constraint_flag = get_bits1(gb);
+ gci->gci_no_lfnst_constraint_flag = get_bits1(gb);
+ gci->gci_no_joint_cbcr_constraint_flag = get_bits1(gb);
+ gci->gci_no_sbt_constraint_flag = get_bits1(gb);
+ gci->gci_no_act_constraint_flag = get_bits1(gb);
+ gci->gci_no_explicit_scaling_list_constraint_flag = get_bits1(gb);
+ gci->gci_no_dep_quant_constraint_flag = get_bits1(gb);
+ gci->gci_no_sign_data_hiding_constraint_flag = get_bits1(gb);
+ gci->gci_no_cu_qp_delta_constraint_flag = get_bits1(gb);
+ gci->gci_no_chroma_qp_offset_constraint_flag = get_bits1(gb);
+
+ /* loop filter */
+ gci->gci_no_sao_constraint_flag = get_bits1(gb);
+ gci->gci_no_alf_constraint_flag = get_bits1(gb);
+ gci->gci_no_ccalf_constraint_flag = get_bits1(gb);
+ gci->gci_no_lmcs_constraint_flag = get_bits1(gb);
+ gci->gci_no_ladf_constraint_flag = get_bits1(gb);
+ gci->gci_no_virtual_boundaries_constraint_flag = get_bits1(gb);
+ gci->gci_num_reserved_bits = get_bits(gb,8);
+ for (i = 0; i < gci->gci_num_reserved_bits; i++) {
+ gci->gci_reserved_zero_bit[i] = get_bits1(gb);
+ }
+ }
+
+ align_get_bits(gb);
+
+ return 0;
+}
+
+
+static int decode_profile_tier_level(GetBitContext *gb, AVCodecContext *avctx,
+ H266RawProfileTierLevel *ptl,
+ int profile_tier_present_flag,
+ int max_num_sub_layers_minus1)
+{
+ int i;
+
+ if (profile_tier_present_flag) {
+ ptl->general_profile_idc = get_bits(gb, 7);
+ ptl->general_tier_flag = get_bits1(gb);
+ }
+ ptl->general_level_idc = get_bits(gb, 8);
+ ptl->ptl_frame_only_constraint_flag = get_bits1(gb);
+ ptl->ptl_multilayer_enabled_flag = get_bits1(gb);
+
+ if (profile_tier_present_flag) {
+ decode_general_constraints_info(gb, avctx,
+ &ptl->general_constraints_info);
+ }
+
+ for (i = max_num_sub_layers_minus1 - 1; i >= 0; i--) {
+ ptl->ptl_sublayer_level_present_flag[i] = get_bits1(gb);
+ }
+
+ align_get_bits(gb);
+
+ for (i = max_num_sub_layers_minus1 - 1; i >= 0; i--) {
+ if (ptl->ptl_sublayer_level_present_flag[i])
+ ptl->sublayer_level_idc[i] = get_bits(gb, 8);
+ }
+
+ if (profile_tier_present_flag) {
+ ptl->ptl_num_sub_profiles = get_bits(gb, 8);
+ for (i = 0; i < ptl->ptl_num_sub_profiles; i++)
+ ptl->general_sub_profile_idc[i] = get_bits_long(gb, 32);
+ }
+
+ return 0;
+}
+
+static int decode_dpb_parameters(GetBitContext *gb, AVCodecContext *avctx,
+ H266DpbParameters *dpb,
+ uint8_t max_sublayers_minus1,
+ uint8_t sublayer_info_flag)
+{
+ int i;
+ for (i = (sublayer_info_flag ? 0 : max_sublayers_minus1);
+ i <= max_sublayers_minus1; i++) {
+ dpb->dpb_max_dec_pic_buffering_minus1[i] = get_ue_golomb(gb);
+ dpb->dpb_max_num_reorder_pics[i] = get_ue_golomb(gb);
+ dpb->dpb_max_latency_increase_plus1[i] = get_ue_golomb(gb);
+ }
+ return 0;
+}
+
+static int decode_ref_pic_list_struct(GetBitContext *gb, AVCodecContext *avctx,
+ H266RefPicListStruct *current,
+ uint8_t list_idx, uint8_t rpls_idx,
+ const H266SPS *sps)
+{
+ int i, j, num_direct_ref_layers = 0;
+
+ current->num_ref_entries = get_ue_golomb(gb);
+ if (sps->long_term_ref_pics_flag &&
+ rpls_idx < sps->num_ref_pic_lists[list_idx] &&
+ current->num_ref_entries > 0)
+ current->ltrp_in_header_flag = get_bits1(gb);
+ if (sps->long_term_ref_pics_flag &&
+ rpls_idx == sps->num_ref_pic_lists[list_idx])
+ current->ltrp_in_header_flag = 1;
+ for (i = 0, j = 0; i < current->num_ref_entries; i++) {
+ if (sps->inter_layer_prediction_enabled_flag)
+ current->inter_layer_ref_pic_flag[i] = get_bits1(gb);
+ else
+ current->inter_layer_ref_pic_flag[i] = 0;
+
+ if (!current->inter_layer_ref_pic_flag[i]) {
+ if (sps->long_term_ref_pics_flag)
+ current->st_ref_pic_flag[i] = get_bits1(gb);
+ else
+ current->st_ref_pic_flag[i] = 1;
+ if (current->st_ref_pic_flag[i]) {
+ int abs_delta_poc_st;
+ current->abs_delta_poc_st[i] = get_ue_golomb(gb);
+ if ((sps->weighted_pred_flag ||
+ sps->weighted_bipred_flag) && i != 0)
+ abs_delta_poc_st = current->abs_delta_poc_st[i];
+ else
+ abs_delta_poc_st = current->abs_delta_poc_st[i] + 1;
+ if (abs_delta_poc_st > 0)
+ current->strp_entry_sign_flag[i] = get_bits1(gb);
+ } else {
+ if (!current->ltrp_in_header_flag) {
+ uint8_t bits = sps->log2_max_pic_order_cnt_lsb_minus4 + 4;
+ current->rpls_poc_lsb_lt[j] = get_bits(gb, bits);
+ j++;
+ }
+ }
+ } else {
+ if (num_direct_ref_layers == 0) {
+ av_log(avctx, AV_LOG_ERROR,
+ "num_direct_ref_layers needs > 0.\n");
+ return AVERROR_INVALIDDATA;
+ }
+ current->ilrp_idx[i] = get_ue_golomb(gb);
+ }
+ }
+ return 0;
+}
+
+static int decode_general_timing_hrd_parameters(GetBitContext *gb,
+ H266GeneralTimingHrdParameters *current)
+{
+ current->num_units_in_tick = get_bits_long(gb, 32);
+ current->time_scale = get_bits_long(gb, 32);
+ current->general_nal_hrd_params_present_flag = get_bits1(gb);
+ current->general_vcl_hrd_params_present_flag = get_bits1(gb);
+
+ if (current->general_nal_hrd_params_present_flag ||
+ current->general_vcl_hrd_params_present_flag) {
+ current->general_same_pic_timing_in_all_ols_flag = get_bits1(gb);
+ current->general_du_hrd_params_present_flag = get_bits1(gb);
+ if (current->general_du_hrd_params_present_flag)
+ current->tick_divisor_minus2 = get_bits(gb, 8);
+ current->bit_rate_scale = get_bits(gb, 4);
+ current->cpb_size_scale = get_bits(gb, 4);
+ if (current->general_du_hrd_params_present_flag)
+ current->cpb_size_du_scale = get_bits(gb, 4);
+ current->hrd_cpb_cnt_minus1 = get_ue_golomb_long(gb);
+ } else {
+ current->general_du_hrd_params_present_flag = 0;
+ }
+ return 0;
+}
+
+static int decode_sublayer_hrd_parameters(GetBitContext *gb,
+ H266SubLayerHRDParameters *current,
+ int sublayer_id,
+ const H266GeneralTimingHrdParameters *general)
+{
+ int i;
+ for (i = 0; i <= general->hrd_cpb_cnt_minus1; i++) {
+ current->bit_rate_value_minus1[sublayer_id][i] = get_ue_golomb_long(gb);
+ current->cpb_size_value_minus1[sublayer_id][i] = get_ue_golomb_long(gb);
+ if (general->general_du_hrd_params_present_flag) {
+ current->cpb_size_du_value_minus1[sublayer_id][i] =
+ get_ue_golomb_long(gb);
+ current->bit_rate_du_value_minus1[sublayer_id][i] =
+ get_ue_golomb_long(gb);
+ }
+ current->cbr_flag[sublayer_id][i] = get_bits1(gb);
+ }
+ return 0;
+}
+
+static int decode_ols_timing_hrd_parameters(GetBitContext *gb,
+ H266OlsTimingHrdParameters *current,
+ uint8_t first_sublayer,
+ uint8_t max_sublayers_minus1,
+ const H266GeneralTimingHrdParameters * general)
+{
+ int i;
+ for (i = first_sublayer; i <= max_sublayers_minus1; i++) {
+ current->fixed_pic_rate_general_flag[i] = get_bits1(gb);
+ if (!current->fixed_pic_rate_general_flag[i])
+ current->fixed_pic_rate_within_cvs_flag[i] = get_bits1(gb);
+ else
+ current->fixed_pic_rate_within_cvs_flag[i] = 1;
+ if (current->fixed_pic_rate_within_cvs_flag[i]) {
+ current->elemental_duration_in_tc_minus1[i] = get_ue_golomb_long(gb);
+ current->low_delay_hrd_flag[i] = 0;
+ } else if ((general->general_nal_hrd_params_present_flag ||
+ general->general_vcl_hrd_params_present_flag) &&
+ general->hrd_cpb_cnt_minus1 == 0) {
+ current->low_delay_hrd_flag[i] = get_bits1(gb);
+ } else {
+ current->low_delay_hrd_flag[i] = 0;
+ }
+ if (general->general_nal_hrd_params_present_flag)
+ decode_sublayer_hrd_parameters(gb,
+ ¤t->
+ nal_sub_layer_hrd_parameters, i,
+ general);
+ if (general->general_vcl_hrd_params_present_flag)
+ decode_sublayer_hrd_parameters(gb,
+ ¤t->
+ nal_sub_layer_hrd_parameters, i,
+ general);
+ }
+ return 0;
+}
+
+static int decode_vui(GetBitContext *gb, AVCodecContext *avctx,
+ H266VUI* vui, uint8_t chroma_format_idc )
+{
+ vui->progressive_source_flag = get_bits1(gb);
+ vui->interlaced_source_flag = get_bits1(gb);
+ vui->non_packed_constraint_flag = get_bits1(gb);
+ vui->non_projected_constraint_flag = get_bits1(gb);
+ vui->aspect_ratio_info_present_flag = get_bits1(gb);
+ if (vui->aspect_ratio_info_present_flag) {
+ vui->aspect_ratio_constant_flag = get_bits1(gb);
+ vui->aspect_ratio_idc = get_bits(gb, 8);
+ if (vui->aspect_ratio_idc == 255) {
+ vui->sar_width = get_bits(gb, 16);
+ vui->sar_height = get_bits(gb, 16);
+ }
+ } else {
+ vui->aspect_ratio_constant_flag = 0;
+ vui->aspect_ratio_idc = 0;
+ }
+ vui->overscan_info_present_flag = get_bits1(gb);
+ if (vui->overscan_info_present_flag)
+ vui->overscan_appropriate_flag = get_bits1(gb);
+ vui->colour_description_present_flag = get_bits1(gb);
+ if (vui->colour_description_present_flag) {
+ vui->colour_primaries = get_bits(gb, 8);
+ vui->transfer_characteristics = get_bits(gb, 8);
+ vui->matrix_coeffs = get_bits(gb, 8);
+ av_log(avctx, AV_LOG_DEBUG,
+ "colour_primaries: %d transfer_characteristics: %d matrix_coeffs: %d \n",
+ vui->colour_primaries, vui->transfer_characteristics,
+ vui->matrix_coeffs);
+
+ vui->full_range_flag = get_bits1(gb);
+ } else {
+ vui->colour_primaries = 2;
+ vui->transfer_characteristics = 2;
+ vui->matrix_coeffs = 2;
+ vui->full_range_flag = 0;
+ }
+ vui->chroma_loc_info_present_flag = get_bits1(gb);
+ if (chroma_format_idc != 1 && vui->chroma_loc_info_present_flag) {
+ av_log(avctx, AV_LOG_ERROR, "chroma_format_idc == %d,"
+ "chroma_loc_info_present_flag can't not be true",
+ chroma_format_idc);
+ return AVERROR_INVALIDDATA;
+ }
+ if (vui->chroma_loc_info_present_flag) {
+ if (vui->progressive_source_flag && !vui->interlaced_source_flag) {
+ vui->chroma_sample_loc_type_frame = get_ue_golomb(gb);
+ } else {
+ vui->chroma_sample_loc_type_top_field = get_ue_golomb(gb);
+ vui->chroma_sample_loc_type_bottom_field = get_ue_golomb(gb);
+ }
+ } else {
+ if (chroma_format_idc == 1) {
+ vui->chroma_sample_loc_type_frame = get_ue_golomb(gb);
+ vui->chroma_sample_loc_type_top_field = get_ue_golomb(gb);
+ vui->chroma_sample_loc_type_bottom_field = get_ue_golomb(gb);
+ }
+ }
+ return 0;
+}
+
+static int map_pixel_format(AVCodecContext *avctx, H266SPS *sps)
+{
+ const AVPixFmtDescriptor *desc;
+ switch (sps->bit_depth) {
+ case 8:
+ if (sps->chroma_format_idc == 0)
+ sps->pix_fmt = AV_PIX_FMT_GRAY8;
+ if (sps->chroma_format_idc == 1)
+ sps->pix_fmt = AV_PIX_FMT_YUV420P;
+ if (sps->chroma_format_idc == 2)
+ sps->pix_fmt = AV_PIX_FMT_YUV422P;
+ if (sps->chroma_format_idc == 3)
+ sps->pix_fmt = AV_PIX_FMT_YUV444P;
+ break;
+ case 9:
+ if (sps->chroma_format_idc == 0)
+ sps->pix_fmt = AV_PIX_FMT_GRAY9;
+ if (sps->chroma_format_idc == 1)
+ sps->pix_fmt = AV_PIX_FMT_YUV420P9;
+ if (sps->chroma_format_idc == 2)
+ sps->pix_fmt = AV_PIX_FMT_YUV422P9;
+ if (sps->chroma_format_idc == 3)
+ sps->pix_fmt = AV_PIX_FMT_YUV444P9;
+ break;
+ case 10:
+ if (sps->chroma_format_idc == 0)
+ sps->pix_fmt = AV_PIX_FMT_GRAY10;
+ if (sps->chroma_format_idc == 1)
+ sps->pix_fmt = AV_PIX_FMT_YUV420P10;
+ if (sps->chroma_format_idc == 2)
+ sps->pix_fmt = AV_PIX_FMT_YUV422P10;
+ if (sps->chroma_format_idc == 3)
+ sps->pix_fmt = AV_PIX_FMT_YUV444P10;
+ break;
+ case 12:
+ if (sps->chroma_format_idc == 0)
+ sps->pix_fmt = AV_PIX_FMT_GRAY12;
+ if (sps->chroma_format_idc == 1)
+ sps->pix_fmt = AV_PIX_FMT_YUV420P12;
+ if (sps->chroma_format_idc == 2)
+ sps->pix_fmt = AV_PIX_FMT_YUV422P12;
+ if (sps->chroma_format_idc == 3)
+ sps->pix_fmt = AV_PIX_FMT_YUV444P12;
+ break;
+ default:
+ av_log(avctx, AV_LOG_ERROR,
+ "The following bit-depths are currently specified: 8, 9, 10 and 12 bits, "
+ "chroma_format_idc is %d, depth is %d\n",
+ sps->chroma_format_idc, sps->bit_depth);
+ return AVERROR_INVALIDDATA;
+ }
+
+ desc = av_pix_fmt_desc_get(sps->pix_fmt);
+ if (!desc)
+ return AVERROR(EINVAL);
+
+ return 0;
+}
+
+int ff_h266_parse_sps(H266SPS *sps, GetBitContext *gb, unsigned int *sps_id,
+ int apply_defdispwin, AVCodecContext *avctx)
+{
+ int i, j;
+ unsigned int ctb_log2_size_y, ctb_size_y, max_num_merge_cand,
+ tmp_width_val, tmp_height_val;
+
+ sps->sps_id = get_bits(gb, 4);
+ sps->vps_id = get_bits(gb, 4);
+
+ *sps_id = sps->sps_id;
+
+ sps->max_sublayers = get_bits(gb, 3) + 1;
+ sps->chroma_format_idc = get_bits(gb, 2);
+ sps->log2_ctu_size = get_bits(gb, 2) + 5;
+
+ ctb_log2_size_y = sps->log2_ctu_size;
+ ctb_size_y = 1 << ctb_log2_size_y;
+
+ sps->ptl_dpb_hrd_params_present_flag = get_bits1(gb);
+ if (sps->ptl_dpb_hrd_params_present_flag)
+ decode_profile_tier_level(gb, avctx, &sps->profile_tier_level, 1,
+ sps->max_sublayers - 1);
+
+ sps->gdr_enabled_flag = get_bits1(gb);
+ sps->ref_pic_resampling_enabled_flag = get_bits1(gb);
+
+ if (sps->ref_pic_resampling_enabled_flag)
+ sps->res_change_in_clvs_allowed_flag = get_bits1(gb);
+ else
+ sps->res_change_in_clvs_allowed_flag = 0;
+
+ sps->pic_width_max_in_luma_samples = get_ue_golomb(gb);
+ sps->pic_height_max_in_luma_samples = get_ue_golomb(gb);
+
+ sps->conformance_window_flag = get_bits1(gb);
+
+ if (sps->conformance_window_flag) {
+ sps->conf_win_left_offset = get_ue_golomb(gb);
+ sps->conf_win_right_offset = get_ue_golomb(gb);
+ sps->conf_win_top_offset = get_ue_golomb(gb);
+ sps->conf_win_bottom_offset = get_ue_golomb(gb);
+ } else {
+ sps->conf_win_left_offset = 0;
+ sps->conf_win_right_offset = 0;
+ sps->conf_win_top_offset = 0;
+ sps->conf_win_bottom_offset = 0;
+ }
+
+ tmp_width_val =
+ (sps->pic_width_max_in_luma_samples + ctb_size_y - 1) / ctb_size_y;
+ tmp_height_val =
+ (sps->pic_height_max_in_luma_samples + ctb_size_y - 1) / ctb_size_y;
+
+ sps->subpic_info_present_flag = get_bits1(gb);
+ if (sps->subpic_info_present_flag) {
+ sps->num_subpics_minus1 = get_ue_golomb(gb);
+ if (sps->num_subpics_minus1 > 0) {
+ sps->independent_subpics_flag = get_bits1(gb);
+ sps->subpic_same_size_flag = get_bits1(gb);
+ }
+ }
+
+ if (sps->num_subpics_minus1 > 0) {
+ int wlen = av_ceil_log2(tmp_width_val);
+ int hlen = av_ceil_log2(tmp_height_val);
+ if (sps->pic_width_max_in_luma_samples > ctb_size_y)
+ sps->subpic_width_minus1[0] = get_bits(gb, wlen);
+ else
+ sps->subpic_width_minus1[0] = tmp_width_val - 1;
+
+ if (sps->pic_height_max_in_luma_samples > ctb_size_y)
+ sps->subpic_height_minus1[0] = get_bits(gb, hlen);
+ else
+ sps->subpic_height_minus1[0] = tmp_height_val;
+
+ if (!sps->independent_subpics_flag) {
+ sps->subpic_treated_as_pic_flag[0] = get_bits1(gb);
+ sps->loop_filter_across_subpic_enabled_flag[0] = get_bits1(gb);
+ } else {
+ sps->subpic_treated_as_pic_flag[0] = 1;
+ sps->loop_filter_across_subpic_enabled_flag[0] = 1;
+ }
+
+ for (i = 1; i <= sps->num_subpics_minus1; i++) {
+ if (!sps->subpic_same_size_flag) {
+ if (sps->pic_width_max_in_luma_samples > ctb_size_y)
+ sps->subpic_ctu_top_left_x[i] = get_bits(gb, wlen);
+ else
+ sps->subpic_ctu_top_left_x[i] = 0;
+
+ if (sps->pic_height_max_in_luma_samples > ctb_size_y)
+ sps->subpic_ctu_top_left_y[i] = get_bits(gb, hlen);
+ else
+ sps->subpic_ctu_top_left_y[i] = 0;
+
+ if (i < sps->num_subpics_minus1 &&
+ sps->pic_width_max_in_luma_samples > ctb_size_y) {
+ sps->subpic_width_minus1[i] = get_bits(gb, wlen);
+ } else {
+ sps->subpic_width_minus1[i] =
+ tmp_width_val - sps->subpic_ctu_top_left_x[i] - 1;
+ }
+ if (i < sps->num_subpics_minus1 &&
+ sps->pic_height_max_in_luma_samples > ctb_size_y) {
+ sps->subpic_height_minus1[i] = get_bits(gb, hlen);
+ } else {
+ sps->subpic_height_minus1[i] =
+ tmp_height_val - sps->subpic_ctu_top_left_y[i] - 1;
+ }
+ } else {
+ int num_subpic_cols =
+ tmp_width_val / (sps->subpic_width_minus1[0] + 1);
+ sps->subpic_ctu_top_left_x[i] = (i % num_subpic_cols) *(sps->subpic_width_minus1[0] + 1);
+ sps->subpic_ctu_top_left_y[i] = (i / num_subpic_cols) *(sps->subpic_height_minus1[0] + 1);
+ sps->subpic_width_minus1[i] = sps->subpic_width_minus1[0];
+ sps->subpic_height_minus1[i] = sps->subpic_height_minus1[0];
+ }
+ if (!sps->independent_subpics_flag) {
+ sps->subpic_treated_as_pic_flag[i] = get_bits1(gb);
+ sps->loop_filter_across_subpic_enabled_flag[i] = get_bits1(gb);
+ } else {
+ sps->subpic_treated_as_pic_flag[i] = 1;
+ sps->loop_filter_across_subpic_enabled_flag[i] = 0;
+ }
+ }
+ sps->subpic_id_len_minus1 = get_ue_golomb(gb);
+
+ if ((1 << (sps->subpic_id_len_minus1 + 1)) <
+ sps->num_subpics_minus1 + 1) {
+ av_log(avctx, AV_LOG_ERROR,
+ "sps->subpic_id_len_minus1(%d) is too small\n",
+ sps->subpic_id_len_minus1);
+ return AVERROR_INVALIDDATA;
+ }
+ sps->subpic_id_mapping_explicitly_signalled_flag = get_bits1(gb);
+ if (sps->subpic_id_mapping_explicitly_signalled_flag) {
+ sps->subpic_id_mapping_present_flag = get_bits1(gb);
+ if (sps->subpic_id_mapping_present_flag) {
+ for (i = 0; i <= sps->num_subpics_minus1; i++)
+ sps->subpic_id[i] =
+ get_bits(gb, sps->subpic_id_len_minus1 + 1);
+ }
+ } else {
+ sps->subpic_ctu_top_left_x[0] = 0;
+ sps->subpic_ctu_top_left_y[0] = 0;
+ sps->subpic_width_minus1[0] = tmp_width_val - 1;
+ sps->subpic_height_minus1[0] = tmp_height_val - 1;
+ }
+ } else {
+ sps->num_subpics_minus1 = 0;
+ sps->independent_subpics_flag = 1;
+ sps->subpic_same_size_flag = 0;
+ sps->subpic_id_mapping_explicitly_signalled_flag = 0;
+ sps->subpic_ctu_top_left_x[0] = 0;
+ sps->subpic_ctu_top_left_y[0] = 0;
+ sps->subpic_width_minus1[0] = tmp_width_val - 1;
+ sps->subpic_height_minus1[0] = tmp_height_val - 1;
+ }
+
+ sps->bit_depth = get_ue_golomb(gb) + 8;
+
+ sps->entropy_coding_sync_enabled_flag = get_bits1(gb);
+ sps->entry_point_offsets_present_flag = get_bits1(gb);
+
+ sps->log2_max_pic_order_cnt_lsb_minus4 = get_bits(gb, 4);
+
+ sps->poc_msb_cycle_flag = get_bits1(gb);
+ if (sps->poc_msb_cycle_flag)
+ sps->poc_msb_cycle_len_minus1 = get_ue_golomb(gb);
+
+ sps->num_extra_ph_bytes = get_bits(gb, 2);
+
+ for (i = 0; i < FFMIN(16, (sps->num_extra_ph_bytes * 8)); i++) {
+ sps->extra_ph_bit_present_flag[i] = get_bits1(gb);
+ }
+
+ sps->num_extra_sh_bytes = get_bits(gb, 2);
+ for (i = 0; i < FFMIN(16, (sps->num_extra_sh_bytes * 8)); i++) {
+ sps->extra_sh_bit_present_flag[i] = get_bits1(gb);
+ }
+
+ if (sps->ptl_dpb_hrd_params_present_flag) {
+ if (sps->max_sublayers > 1)
+ sps->sublayer_dpb_params_flag = get_bits1(gb);
+ else
+ sps->sublayer_dpb_params_flag = 0;
+
+ decode_dpb_parameters(gb, avctx, &sps->dpb_params,
+ sps->max_sublayers - 1,
+ sps->sublayer_dpb_params_flag);
+ }
+
+ sps->log2_min_luma_coding_block_size_minus2 = get_ue_golomb(gb);
+ sps->partition_constraints_override_enabled_flag = get_bits1(gb);
+ sps->log2_diff_min_qt_min_cb_intra_slice_luma = get_ue_golomb(gb);
+ sps->max_mtt_hierarchy_depth_intra_slice_luma = get_ue_golomb(gb);
+
+ if (sps->max_mtt_hierarchy_depth_intra_slice_luma != 0) {
+ sps->log2_diff_max_bt_min_qt_intra_slice_luma = get_ue_golomb(gb);
+ sps->log2_diff_max_tt_min_qt_intra_slice_luma = get_ue_golomb(gb);
+ } else {
+ sps->log2_diff_max_bt_min_qt_intra_slice_luma = 0;
+ sps->log2_diff_max_tt_min_qt_intra_slice_luma = 0;
+ }
+
+ if (sps->chroma_format_idc != 0) {
+ sps->qtbtt_dual_tree_intra_flag = get_bits1(gb);
+ } else {
+ sps->qtbtt_dual_tree_intra_flag = 0;
+ }
+
+ if (sps->qtbtt_dual_tree_intra_flag) {
+ sps->log2_diff_min_qt_min_cb_intra_slice_chroma = get_ue_golomb(gb);
+ sps->max_mtt_hierarchy_depth_intra_slice_chroma = get_ue_golomb(gb);
+ if (sps->max_mtt_hierarchy_depth_intra_slice_chroma != 0) {
+ sps->log2_diff_max_bt_min_qt_intra_slice_chroma = get_ue_golomb(gb);
+ sps->log2_diff_max_tt_min_qt_intra_slice_chroma = get_ue_golomb(gb);
+ }
+ } else {
+ sps->log2_diff_min_qt_min_cb_intra_slice_chroma = 0;
+ sps->max_mtt_hierarchy_depth_intra_slice_chroma = 0;
+ }
+ if (sps->max_mtt_hierarchy_depth_intra_slice_chroma == 0) {
+ sps->log2_diff_max_bt_min_qt_intra_slice_chroma = 0;
+ sps->log2_diff_max_tt_min_qt_intra_slice_chroma = 0;
+ }
+
+ sps->log2_diff_min_qt_min_cb_inter_slice = get_ue_golomb(gb);
+
+ sps->max_mtt_hierarchy_depth_inter_slice = get_ue_golomb(gb);
+ if (sps->max_mtt_hierarchy_depth_inter_slice != 0) {
+ sps->log2_diff_max_bt_min_qt_inter_slice = get_ue_golomb(gb);
+ sps->log2_diff_max_tt_min_qt_inter_slice = get_ue_golomb(gb);
+ } else {
+ sps->log2_diff_max_bt_min_qt_inter_slice = 0;
+ sps->log2_diff_max_tt_min_qt_inter_slice = 0;
+ }
+
+ if (ctb_size_y > 32)
+ sps->max_luma_transform_size_64_flag = get_bits1(gb);
+ else
+ sps->max_luma_transform_size_64_flag = 0;
+
+ sps->transform_skip_enabled_flag = get_bits1(gb);
+ if (sps->transform_skip_enabled_flag) {
+ sps->log2_transform_skip_max_size_minus2 = get_ue_golomb(gb);
+ sps->bdpcm_enabled_flag = get_bits1(gb);
+ }
+
+ sps->mts_enabled_flag = get_bits1(gb);
+ if (sps->mts_enabled_flag) {
+ sps->explicit_mts_intra_enabled_flag = get_bits1(gb);
+ sps->explicit_mts_inter_enabled_flag = get_bits1(gb);
+ } else {
+ sps->explicit_mts_intra_enabled_flag = 0;
+ sps->explicit_mts_inter_enabled_flag = 0;
+ }
+
+ sps->lfnst_enabled_flag = get_bits1(gb);
+
+ if (sps->chroma_format_idc != 0) {
+ uint8_t num_qp_tables;
+ sps->joint_cbcr_enabled_flag = get_bits1(gb);
+ sps->same_qp_table_for_chroma_flag = get_bits1(gb);
+ num_qp_tables = sps->same_qp_table_for_chroma_flag ?
+ 1 : (sps->joint_cbcr_enabled_flag ? 3 : 2);
+ for (i = 0; i < num_qp_tables; i++) {
+ sps->qp_table_start_minus26[i] = get_se_golomb(gb);
+ sps->num_points_in_qp_table_minus1[i] = get_ue_golomb(gb);
+ for (j = 0; j <= sps->num_points_in_qp_table_minus1[i]; j++) {
+ sps->delta_qp_in_val_minus1[i][j] = get_ue_golomb(gb);
+ sps->delta_qp_diff_val[i][j] = get_ue_golomb(gb);
+ }
+ }
+ } else {
+ sps->joint_cbcr_enabled_flag = 0;
+ sps->same_qp_table_for_chroma_flag = 0;
+ }
+
+ sps->sao_enabled_flag = get_bits1(gb);
+ sps->alf_enabled_flag = get_bits1(gb);
+ if (sps->alf_enabled_flag && sps->chroma_format_idc)
+ sps->ccalf_enabled_flag = get_bits1(gb);
+ else
+ sps->ccalf_enabled_flag = 0;
+
+ sps->lmcs_enabled_flag = get_bits1(gb);
+ sps->weighted_pred_flag = get_bits1(gb);
+ sps->weighted_bipred_flag = get_bits1(gb);
+ sps->long_term_ref_pics_flag = get_bits1(gb);
+ if (sps->vps_id > 0)
+ sps->inter_layer_prediction_enabled_flag = get_bits1(gb);
+ else
+ sps->inter_layer_prediction_enabled_flag = 0;
+ sps->idr_rpl_present_flag = get_bits1(gb);
+ sps->rpl1_same_as_rpl0_flag = get_bits1(gb);
+
+ for (i = 0; i < (sps->rpl1_same_as_rpl0_flag ? 1 : 2); i++) {
+ sps->num_ref_pic_lists[i] = get_ue_golomb(gb);
+ for (j = 0; j < sps->num_ref_pic_lists[i]; j++)
+ decode_ref_pic_list_struct(gb, avctx,
+ &sps->ref_pic_list_struct[i][j],
+ i, j, sps);
+ }
+
+ if (sps->rpl1_same_as_rpl0_flag) {
+ sps->num_ref_pic_lists[1] = sps->num_ref_pic_lists[0];
+ for (j = 0; j < sps->num_ref_pic_lists[0]; j++)
+ memcpy(&sps->ref_pic_list_struct[1][j],
+ &sps->ref_pic_list_struct[0][j],
+ sizeof(sps->ref_pic_list_struct[0][j]));
+ }
+
+ sps->ref_wraparound_enabled_flag = get_bits1(gb);
+
+ sps->temporal_mvp_enabled_flag = get_bits1(gb);
+ if (sps->temporal_mvp_enabled_flag)
+ sps->sbtmvp_enabled_flag = get_bits1(gb);
+ else
+ sps->sbtmvp_enabled_flag = 0;
+
+ sps->amvr_enabled_flag = get_bits1(gb);
+ sps->bdof_enabled_flag = get_bits1(gb);
+ if (sps->bdof_enabled_flag)
+ sps->bdof_control_present_in_ph_flag = get_bits1(gb);
+ else
+ sps->bdof_control_present_in_ph_flag = 0;
+
+ sps->smvd_enabled_flag = get_bits1(gb);
+ sps->dmvr_enabled_flag = get_bits1(gb);
+ if (sps->dmvr_enabled_flag)
+ sps->dmvr_control_present_in_ph_flag = get_bits1(gb);
+ else
+ sps->dmvr_control_present_in_ph_flag = 0;
+
+ sps->mmvd_enabled_flag = get_bits1(gb);
+ if (sps->mmvd_enabled_flag)
+ sps->mmvd_fullpel_only_enabled_flag = get_bits1(gb);
+ else
+ sps->mmvd_fullpel_only_enabled_flag = 0;
+
+ sps->six_minus_max_num_merge_cand = get_ue_golomb(gb);
+ max_num_merge_cand = 6 - sps->six_minus_max_num_merge_cand;
+
+ sps->sbt_enabled_flag = get_bits1(gb);
+
+ sps->affine_enabled_flag = get_bits1(gb);
+ if (sps->affine_enabled_flag) {
+ sps->five_minus_max_num_subblock_merge_cand = get_ue_golomb(gb);
+ sps->param_affine_enabled_flag = get_bits1(gb);
+ if (sps->amvr_enabled_flag)
+ sps->affine_amvr_enabled_flag = get_bits1(gb);
+ else
+ sps->affine_amvr_enabled_flag = 0;
+ sps->affine_prof_enabled_flag = get_bits1(gb);
+ if (sps->affine_prof_enabled_flag)
+ sps->prof_control_present_in_ph_flag = get_bits1(gb);
+ else
+ sps->prof_control_present_in_ph_flag = 0;
+ } else {
+ sps->param_affine_enabled_flag = 0;
+ sps->affine_amvr_enabled_flag = 0;
+ sps->affine_prof_enabled_flag = 0;
+ sps->prof_control_present_in_ph_flag = 0;
+ }
+
+ sps->bcw_enabled_flag = get_bits1(gb);
+ sps->ciip_enabled_flag = get_bits1(gb);
+
+ if (max_num_merge_cand >= 2) {
+ sps->gpm_enabled_flag = get_bits1(gb);
+ if (sps->gpm_enabled_flag && max_num_merge_cand >= 3)
+ sps->max_num_merge_cand_minus_max_num_gpm_cand = get_ue_golomb(gb);
+ } else {
+ sps->gpm_enabled_flag = 0;
+ }
+
+ sps->log2_parallel_merge_level_minus2 = get_ue_golomb(gb);
+
+ sps->isp_enabled_flag = get_bits1(gb);
+ sps->mrl_enabled_flag = get_bits1(gb);
+ sps->mip_enabled_flag = get_bits1(gb);
+
+ if (sps->chroma_format_idc != 0)
+ sps->cclm_enabled_flag = get_bits1(gb);
+ else
+ sps->cclm_enabled_flag = 0;
+ if (sps->chroma_format_idc == 1) {
+ sps->chroma_horizontal_collocated_flag = get_bits1(gb);
+ sps->chroma_vertical_collocated_flag = get_bits1(gb);
+ } else {
+ sps->chroma_horizontal_collocated_flag = 0;
+ sps->chroma_vertical_collocated_flag = 0;
+ }
+
+ sps->palette_enabled_flag = get_bits1(gb);
+ if (sps->chroma_format_idc == 3 && !sps->max_luma_transform_size_64_flag)
+ sps->act_enabled_flag = get_bits1(gb);
+ else
+ sps->act_enabled_flag = 0;
+ if (sps->transform_skip_enabled_flag || sps->palette_enabled_flag)
+ sps->min_qp_prime_ts = get_ue_golomb(gb);
+
+ sps->ibc_enabled_flag = get_bits1(gb);
+ if (sps->ibc_enabled_flag)
+ sps->six_minus_max_num_ibc_merge_cand = get_ue_golomb(gb);
+
+ sps->ladf_enabled_flag = get_bits1(gb);
+ if (sps->ladf_enabled_flag) {
+ sps->num_ladf_intervals_minus2 = get_bits(gb, 2);
+ sps->ladf_lowest_interval_qp_offset = get_se_golomb(gb);
+ for (i = 0; i < sps->num_ladf_intervals_minus2 + 1; i++) {
+ sps->ladf_qp_offset[i] = get_se_golomb(gb);
+ sps->ladf_delta_threshold_minus1[i] = get_ue_golomb(gb);
+ }
+ }
+
+ sps->explicit_scaling_list_enabled_flag = get_bits1(gb);
+ if (sps->lfnst_enabled_flag && sps->explicit_scaling_list_enabled_flag)
+ sps->scaling_matrix_for_lfnst_disabled_flag = get_bits1(gb);
+
+ if (sps->act_enabled_flag && sps->explicit_scaling_list_enabled_flag)
+ sps->scaling_matrix_for_alternative_colour_space_disabled_flag =
+ get_bits1(gb);
+ else
+ sps->scaling_matrix_for_alternative_colour_space_disabled_flag = 0;
+ if (sps->scaling_matrix_for_alternative_colour_space_disabled_flag)
+ sps->scaling_matrix_designated_colour_space_flag = get_bits1(gb);
+
+ sps->dep_quant_enabled_flag = get_bits1(gb);
+ sps->sign_data_hiding_enabled_flag = get_bits1(gb);
+
+ sps->virtual_boundaries_enabled_flag = get_bits1(gb);
+ if (sps->virtual_boundaries_enabled_flag) {
+ sps->virtual_boundaries_present_flag = get_bits1(gb);
+ if (sps->virtual_boundaries_present_flag) {
+ sps->num_ver_virtual_boundaries = get_ue_golomb(gb);
+ for (i = 0; i < sps->num_ver_virtual_boundaries; i++)
+ sps->virtual_boundary_pos_x_minus1[i] = get_ue_golomb(gb);
+ for (i = 0; i < sps->num_hor_virtual_boundaries; i++)
+ sps->virtual_boundary_pos_y_minus1[i] = get_ue_golomb(gb);
+ }
+ } else {
+ sps->virtual_boundaries_present_flag = 0;
+ sps->num_ver_virtual_boundaries = 0;
+ sps->num_hor_virtual_boundaries = 0;
+ }
+
+ if (sps->ptl_dpb_hrd_params_present_flag) {
+ sps->timing_hrd_params_present_flag = get_bits1(gb);
+ if (sps->timing_hrd_params_present_flag) {
+ uint8_t first_sublayer;
+ decode_general_timing_hrd_parameters(gb,
+ &sps->general_timing_hrd_parameters);
+
+ if (sps->max_sublayers > 1)
+ sps->sublayer_cpb_params_present_flag = get_bits1(gb);
+ else
+ sps->sublayer_cpb_params_present_flag = 0;
+
+ first_sublayer = sps->sublayer_cpb_params_present_flag ?
+ 0 : sps->max_sublayers - 1;
+
+ decode_ols_timing_hrd_parameters(gb,
+ &sps->ols_timing_hrd_parameters,
+ first_sublayer,
+ sps->max_sublayers - 1,
+ &sps->
+ general_timing_hrd_parameters);
+ }
+ }
+
+ sps->field_seq_flag = get_bits1(gb);
+ sps->vui_parameters_present_flag = get_bits1(gb);
+ if (sps->vui_parameters_present_flag) {
+ sps->vui_payload_size_minus1 = get_ue_golomb(gb);
+ align_get_bits(gb);
+ decode_vui(gb, avctx, &sps->vui, sps->chroma_format_idc);
+ }
+
+ sps->extension_flag = get_bits1(gb);
+ // TODO: parse sps extension flag and read extension data
+
+ map_pixel_format(avctx, sps);
+
+ return 0;
+}
+
+int ff_h266_decode_nal_sps( GetBitContext *gb, AVCodecContext *avctx,
+ H266ParamSets *ps, int apply_defdispwin)
+{
+ unsigned int sps_id;
+ int ret;
+ H266SPS *sps;
+ AVBufferRef *sps_buf = av_buffer_allocz(sizeof(*sps));
+
+ if (!sps_buf)
+ return AVERROR(ENOMEM);
+ sps = (H266SPS *) sps_buf->data;
+
+ ret = ff_h266_parse_sps(sps, gb, &sps_id, apply_defdispwin, avctx);
+ if (ret < 0) {
+ av_buffer_unref(&sps_buf);
+ return ret;
+ }
+
+ if (avctx->debug & FF_DEBUG_BITSTREAM) {
+ av_log(avctx, AV_LOG_DEBUG,
+ "Parsed SPS: id %d; coded wxh: %dx%d; "
+ "pix_fmt: %s.\n",
+ sps->sps_id, sps->pic_width_max_in_luma_samples,
+ sps->pic_height_max_in_luma_samples,
+ av_get_pix_fmt_name(sps->pix_fmt));
+ }
+
+ /* check if this is a repeat of an already parsed SPS, then keep the
+ * original one otherwise drop all PPSes that depend on it (PPS,VPS not implemented yet) */
+ if (ps->sps_list[sps_id] &&
+ !memcmp(ps->sps_list[sps_id]->data, sps_buf->data, sps_buf->size)) {
+ av_buffer_unref(&sps_buf);
+ } else {
+ remove_sps(ps, sps_id);
+ ps->sps_list[sps_id] = sps_buf;
+ ps->sps = (H266SPS *) ps->sps_list[sps_id]->data;
+ }
+
+ // TODO: read PPS flag and data
+
+ return 0;
+}
+
+void ff_h266_ps_uninit(H266ParamSets *ps)
+{
+ int i;
+
+ for (i = 0; i < FF_ARRAY_ELEMS(ps->sps_list); i++)
+ av_buffer_unref(&ps->sps_list[i]);
+
+ // TODO: if PPS, VPS is implemended it must be unrefed as well
+ // for (i = 0; i < FF_ARRAY_ELEMS(ps->vps_list); i++)
+ // av_buffer_unref(&ps->vps_list[i]);
+ // for (i = 0; i < FF_ARRAY_ELEMS(ps->pps_list); i++)
+ // av_buffer_unref(&ps->pps_list[i]);
+
+ ps->sps = NULL;
+ //ps->pps = NULL;
+ //ps->vps = NULL;
+}
diff --git a/libavcodec/h266_paramset.h b/libavcodec/h266_paramset.h
new file mode 100644
index 0000000000..d70247550d
--- /dev/null
+++ b/libavcodec/h266_paramset.h
@@ -0,0 +1,307 @@
+/*
+ * H.266 / VVC parameter set parsing
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVCODEC_H266_PARAMSET_H
+#define AVCODEC_H266_PARAMSET_H
+
+#include <stdint.h>
+
+#include "libavutil/buffer.h"
+#include "libavutil/pixfmt.h"
+#include "libavutil/rational.h"
+
+#include "avcodec.h"
+#include "cbs_h266.h"
+#include "get_bits.h"
+#include "h266.h"
+
+typedef struct H266GeneralTimingHrdParameters {
+ uint32_t num_units_in_tick;
+ uint32_t time_scale;
+ uint8_t general_nal_hrd_params_present_flag;
+ uint8_t general_vcl_hrd_params_present_flag;
+ uint8_t general_same_pic_timing_in_all_ols_flag;
+ uint8_t general_du_hrd_params_present_flag;
+ uint8_t tick_divisor_minus2;
+ uint8_t bit_rate_scale;
+ uint8_t cpb_size_scale;
+ uint8_t cpb_size_du_scale;
+ uint8_t hrd_cpb_cnt_minus1;
+} H266GeneralTimingHrdParameters;
+
+typedef struct H266SubLayerHRDParameters {
+ uint32_t bit_rate_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint32_t cpb_size_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint32_t cpb_size_du_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint32_t bit_rate_du_value_minus1[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+ uint8_t cbr_flag[VVC_MAX_SUBLAYERS][VVC_MAX_CPB_CNT];
+} H266SubLayerHRDParameters;
+
+typedef struct H266OlsTimingHrdParameters {
+ uint8_t fixed_pic_rate_general_flag[VVC_MAX_SUBLAYERS];
+ uint8_t fixed_pic_rate_within_cvs_flag[VVC_MAX_SUBLAYERS];
+ uint16_t elemental_duration_in_tc_minus1[VVC_MAX_SUBLAYERS];
+ uint8_t low_delay_hrd_flag[VVC_MAX_SUBLAYERS];
+ H266SubLayerHRDParameters nal_sub_layer_hrd_parameters;
+ H266SubLayerHRDParameters vcl_sub_layer_hrd_parameters;
+} H266OlsTimingHrdParameters;
+
+
+typedef struct H266VUI {
+ uint8_t progressive_source_flag;
+ uint8_t interlaced_source_flag;
+ uint8_t non_packed_constraint_flag;
+ uint8_t non_projected_constraint_flag;
+
+ uint8_t aspect_ratio_info_present_flag;
+ uint8_t aspect_ratio_constant_flag;
+ uint8_t aspect_ratio_idc;
+
+ uint16_t sar_width;
+ uint16_t sar_height;
+
+ uint8_t overscan_info_present_flag;
+ uint8_t overscan_appropriate_flag;
+
+ uint8_t colour_description_present_flag;
+ uint8_t colour_primaries;
+
+ uint8_t transfer_characteristics;
+ uint8_t matrix_coeffs;
+ uint8_t full_range_flag;
+
+ uint8_t chroma_loc_info_present_flag;
+ uint8_t chroma_sample_loc_type_frame;
+ uint8_t chroma_sample_loc_type_top_field;
+ uint8_t chroma_sample_loc_type_bottom_field;
+ //H266ExtensionData extension_data;
+} H266VUI;
+
+typedef struct H266SPS {
+ uint8_t sps_id;
+ uint8_t vps_id;
+ uint8_t max_sublayers;
+ uint8_t chroma_format_idc;
+ uint8_t log2_ctu_size;
+ uint8_t ptl_dpb_hrd_params_present_flag;
+ H266RawProfileTierLevel profile_tier_level;
+ uint8_t gdr_enabled_flag;
+ uint8_t ref_pic_resampling_enabled_flag;
+ uint8_t res_change_in_clvs_allowed_flag;
+
+ uint16_t pic_width_max_in_luma_samples;
+ uint16_t pic_height_max_in_luma_samples;
+
+ uint8_t conformance_window_flag;
+ uint16_t conf_win_left_offset;
+ uint16_t conf_win_right_offset;
+ uint16_t conf_win_top_offset;
+ uint16_t conf_win_bottom_offset;
+
+ uint8_t subpic_info_present_flag;
+ uint16_t num_subpics_minus1;
+ uint8_t independent_subpics_flag;
+ uint8_t subpic_same_size_flag;
+ uint16_t subpic_ctu_top_left_x[VVC_MAX_SLICES];
+ uint16_t subpic_ctu_top_left_y[VVC_MAX_SLICES];
+ uint16_t subpic_width_minus1[VVC_MAX_SLICES];
+ uint16_t subpic_height_minus1[VVC_MAX_SLICES];
+ uint8_t subpic_treated_as_pic_flag[VVC_MAX_SLICES];
+ uint8_t loop_filter_across_subpic_enabled_flag[VVC_MAX_SLICES];
+ uint8_t subpic_id_len_minus1;
+ uint8_t subpic_id_mapping_explicitly_signalled_flag;
+ uint8_t subpic_id_mapping_present_flag;
+ uint32_t subpic_id[VVC_MAX_SLICES];
+
+
+ uint8_t bit_depth;
+ uint8_t entropy_coding_sync_enabled_flag;
+ uint8_t entry_point_offsets_present_flag;
+
+ uint8_t log2_max_pic_order_cnt_lsb_minus4;
+ uint8_t poc_msb_cycle_flag;
+ uint8_t poc_msb_cycle_len_minus1;
+
+ uint8_t num_extra_ph_bytes;
+ uint8_t extra_ph_bit_present_flag[16];
+
+ uint8_t num_extra_sh_bytes;
+ uint8_t extra_sh_bit_present_flag[16];
+
+ uint8_t sublayer_dpb_params_flag;
+ H266DpbParameters dpb_params;
+
+ uint8_t log2_min_luma_coding_block_size_minus2;
+ uint8_t partition_constraints_override_enabled_flag;
+ uint8_t log2_diff_min_qt_min_cb_intra_slice_luma;
+ uint8_t max_mtt_hierarchy_depth_intra_slice_luma;
+ uint8_t log2_diff_max_bt_min_qt_intra_slice_luma;
+ uint8_t log2_diff_max_tt_min_qt_intra_slice_luma;
+
+ uint8_t qtbtt_dual_tree_intra_flag;
+ uint8_t log2_diff_min_qt_min_cb_intra_slice_chroma;
+ uint8_t max_mtt_hierarchy_depth_intra_slice_chroma;
+ uint8_t log2_diff_max_bt_min_qt_intra_slice_chroma;
+ uint8_t log2_diff_max_tt_min_qt_intra_slice_chroma;
+
+ uint8_t log2_diff_min_qt_min_cb_inter_slice;
+ uint8_t max_mtt_hierarchy_depth_inter_slice;
+ uint8_t log2_diff_max_bt_min_qt_inter_slice;
+ uint8_t log2_diff_max_tt_min_qt_inter_slice;
+
+ uint8_t max_luma_transform_size_64_flag;
+
+ uint8_t transform_skip_enabled_flag;
+ uint8_t log2_transform_skip_max_size_minus2;
+ uint8_t bdpcm_enabled_flag;
+
+ uint8_t mts_enabled_flag;
+ uint8_t explicit_mts_intra_enabled_flag;
+ uint8_t explicit_mts_inter_enabled_flag;
+
+ uint8_t lfnst_enabled_flag;
+
+ uint8_t joint_cbcr_enabled_flag;
+ uint8_t same_qp_table_for_chroma_flag;
+
+ int8_t qp_table_start_minus26[VVC_MAX_SAMPLE_ARRAYS];
+ uint8_t num_points_in_qp_table_minus1[VVC_MAX_SAMPLE_ARRAYS];
+ uint8_t delta_qp_in_val_minus1[VVC_MAX_SAMPLE_ARRAYS][VVC_MAX_POINTS_IN_QP_TABLE];
+ uint8_t delta_qp_diff_val[VVC_MAX_SAMPLE_ARRAYS][VVC_MAX_POINTS_IN_QP_TABLE];
+
+ uint8_t sao_enabled_flag;
+ uint8_t alf_enabled_flag;
+ uint8_t ccalf_enabled_flag;
+ uint8_t lmcs_enabled_flag;
+ uint8_t weighted_pred_flag;
+ uint8_t weighted_bipred_flag;
+ uint8_t long_term_ref_pics_flag;
+ uint8_t inter_layer_prediction_enabled_flag;
+ uint8_t idr_rpl_present_flag;
+ uint8_t rpl1_same_as_rpl0_flag;
+
+ uint8_t num_ref_pic_lists[2];
+ H266RefPicListStruct ref_pic_list_struct[2][VVC_MAX_REF_PIC_LISTS];
+
+ uint8_t ref_wraparound_enabled_flag;
+ uint8_t temporal_mvp_enabled_flag;
+ uint8_t sbtmvp_enabled_flag;
+ uint8_t amvr_enabled_flag;
+ uint8_t bdof_enabled_flag;
+ uint8_t bdof_control_present_in_ph_flag;
+ uint8_t smvd_enabled_flag;
+ uint8_t dmvr_enabled_flag;
+ uint8_t dmvr_control_present_in_ph_flag;
+ uint8_t mmvd_enabled_flag;
+ uint8_t mmvd_fullpel_only_enabled_flag;
+ uint8_t six_minus_max_num_merge_cand;
+ uint8_t sbt_enabled_flag;
+ uint8_t affine_enabled_flag;
+ uint8_t five_minus_max_num_subblock_merge_cand;
+ uint8_t param_affine_enabled_flag;
+ uint8_t affine_amvr_enabled_flag;
+ uint8_t affine_prof_enabled_flag;
+ uint8_t prof_control_present_in_ph_flag;
+ uint8_t bcw_enabled_flag;
+ uint8_t ciip_enabled_flag;
+ uint8_t gpm_enabled_flag;
+ uint8_t max_num_merge_cand_minus_max_num_gpm_cand;
+ uint8_t log2_parallel_merge_level_minus2;
+ uint8_t isp_enabled_flag;
+ uint8_t mrl_enabled_flag;
+ uint8_t mip_enabled_flag;
+ uint8_t cclm_enabled_flag;
+ uint8_t chroma_horizontal_collocated_flag;
+ uint8_t chroma_vertical_collocated_flag;
+ uint8_t palette_enabled_flag;
+ uint8_t act_enabled_flag;
+ uint8_t min_qp_prime_ts;
+ uint8_t ibc_enabled_flag;
+ uint8_t six_minus_max_num_ibc_merge_cand;
+ uint8_t ladf_enabled_flag;
+ uint8_t num_ladf_intervals_minus2;
+ int8_t ladf_lowest_interval_qp_offset;
+ int8_t ladf_qp_offset[4];
+ uint16_t ladf_delta_threshold_minus1[4];
+
+ uint8_t explicit_scaling_list_enabled_flag;
+ uint8_t scaling_matrix_for_lfnst_disabled_flag;
+ uint8_t scaling_matrix_for_alternative_colour_space_disabled_flag;
+ uint8_t scaling_matrix_designated_colour_space_flag;
+ uint8_t dep_quant_enabled_flag;
+ uint8_t sign_data_hiding_enabled_flag;
+
+ uint8_t virtual_boundaries_enabled_flag;
+ uint8_t virtual_boundaries_present_flag;
+ uint8_t num_ver_virtual_boundaries;
+ uint16_t virtual_boundary_pos_x_minus1[3];
+ uint8_t num_hor_virtual_boundaries;
+ uint16_t virtual_boundary_pos_y_minus1[3];
+
+ uint8_t timing_hrd_params_present_flag;
+ uint8_t sublayer_cpb_params_present_flag;
+ H266GeneralTimingHrdParameters general_timing_hrd_parameters;
+ H266OlsTimingHrdParameters ols_timing_hrd_parameters;
+
+ uint8_t field_seq_flag;
+ uint8_t vui_parameters_present_flag;
+ uint16_t vui_payload_size_minus1;
+ H266VUI vui;
+
+ uint8_t extension_flag;
+ //H266RawExtensionData extension_data; /* TODO: read extension flag and data*/
+
+ enum AVPixelFormat pix_fmt;
+} H266SPS;
+
+
+typedef struct H266ParamSets {
+ AVBufferRef *sps_list[VVC_MAX_SPS_COUNT];
+ //AVBufferRef *vps_list[VVC_MAX_VPS_COUNT]; // TODO: since not needed, not implemented yet
+ //AVBufferRef *pps_list[VVC_MAX_PPS_COUNT]; // TODO: since not needed, not implemented yet
+
+ /* currently active parameter sets */
+ const H266SPS *sps;
+ //const H266VPS *vps; // TODO: since not needed, not implemented yet
+ //const H266PPS *pps; // TODO: since not needed, not implemented yet
+} H266ParamSets;
+
+/**
+ * Parse the SPS from the bitstream into the provided HEVCSPS struct.
+ *
+ * @param sps_id the SPS id will be written here
+ * @param apply_defdispwin if set 1, the default display window from the VUI
+ * will be applied to the video dimensions
+ */
+int ff_h266_parse_sps(H266SPS *sps, GetBitContext *gb, unsigned int *sps_id,
+ int apply_defdispwin, AVCodecContext *avctx);
+
+int ff_h266_decode_nal_sps(GetBitContext *gb, AVCodecContext *avctx,
+ H266ParamSets *ps, int apply_defdispwin);
+
+// TODO: since not needed, not implemented yet
+//int ff_h266_decode_nal_vps(GetBitContext *gb, AVCodecContext *avctx,
+// H266ParamSets *ps);
+//int ff_h266_decode_nal_pps(GetBitContext *gb, AVCodecContext *avctx,
+// H266ParamSets *ps);
+
+void ff_h266_ps_uninit(H266ParamSets *ps);
+
+#endif /* AVCODEC_H266_PARAMSET_H */
--git a/libavcodec/h266_parse_extradata.c b/libavcodec/h266_parse_extradata.c
new file mode 100644
index 0000000000..559ae0a4c5
--- /dev/null
+++ b/libavcodec/h266_parse_extradata.c
@@ -0,0 +1,246 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "bytestream.h"
+#include "h2645_parse.h"
+#include "h266.h"
+#include "h266_parse_extradata.h"
+
+static int h266_decode_nal_units(const uint8_t *buf, int buf_size,
+ H266ParamSets *ps, int is_nalff,
+ int nal_length_size, int err_recognition,
+ int apply_defdispwin, void *logctx)
+{
+ int i;
+ int ret = 0;
+ H2645Packet pkt = { 0 };
+
+ ret = ff_h2645_packet_split(&pkt, buf, buf_size, logctx, is_nalff,
+ nal_length_size, AV_CODEC_ID_VVC, 1, 0);
+ if (ret < 0) {
+ goto done;
+ }
+
+ for (i = 0; i < pkt.nb_nals; i++) {
+ H2645NAL *nal = &pkt.nals[i];
+ if (nal->nuh_layer_id > 0)
+ continue;
+
+ /* ignore everything except parameter sets and VCL NALUs */
+ switch (nal->type) {
+ case VVC_VPS_NUT:
+ // TODO: since not needed yet, not implemented
+ //ret = ff_h266_decode_nal_vps(&nal->gb, logctx, ps);
+ //if (ret < 0)
+ // goto done;
+ break;
+ case VVC_SPS_NUT:
+ ret = ff_h266_decode_nal_sps(&nal->gb, logctx, ps, apply_defdispwin);
+ if (ret < 0)
+ goto done;
+ break;
+ case VVC_PPS_NUT:
+ // TODO: since not needed yet, not implemented
+ //ret = ff_h266_decode_nal_pps(&nal->gb, logctx, ps);
+ //if (ret < 0)
+ // goto done;
+ break;
+ case VVC_PREFIX_SEI_NUT:
+ case VVC_SUFFIX_SEI_NUT:
+ // TODO: SEI decoding not implemented yet
+ break;
+ default:
+ av_log(logctx, AV_LOG_VERBOSE,
+ "Ignoring NAL type %d in extradata\n", nal->type);
+ break;
+ }
+ }
+
+ done:
+ ff_h2645_packet_uninit(&pkt);
+ if (err_recognition & AV_EF_EXPLODE)
+ return ret;
+
+ return 0;
+}
+
+int ff_h266_decode_extradata(const uint8_t *data, int size, H266ParamSets *ps,
+ int *is_nalff, int *nal_length_size,
+ int err_recognition, int apply_defdispwin,
+ void *logctx)
+{
+ int ret = 0;
+ GetByteContext gb;
+
+ bytestream2_init(&gb, data, size);
+
+ if (size > 3 && (data[0] || data[1] || data[2] > 1)) {
+ /* extradata is encoded as vvcC format. */
+ int i, j, b, num_arrays, nal_len_size, has_ptl;
+
+ b = bytestream2_get_byte(&gb);
+ nal_len_size = ((b >> 1) & 3) + 1;
+ has_ptl = b & 1;
+
+ if (has_ptl) {
+ int max_picture_width = 0;
+ int max_picture_height = 0;
+ int avg_frame_rate = 0;
+
+ int num_sublayers;
+ int num_bytes_constraint_info;
+ int general_profile_idc;
+ int general_tier_flag;
+ int general_level_idc;
+ int ptl_frame_only_constraint_flag;
+ int ptl_multi_layer_enabled_flag;
+ int ptl_num_sub_profiles;
+ int ols_idx;
+ int constant_frame_rate;
+ int chroma_format_idc;
+ int bit_depth_minus8;
+
+ b = bytestream2_get_be16(&gb);
+ ols_idx = (b >> 7) & 0x1ff;
+ num_sublayers = (b >> 4) & 7;
+ constant_frame_rate = (b >> 2) & 3;
+ chroma_format_idc = b & 0x3;
+ bit_depth_minus8 = (bytestream2_get_byte(&gb) >> 5) & 7;
+ av_log(logctx, AV_LOG_TRACE,
+ "bit_depth_minus8 %d chroma_format_idc %d\n",
+ bit_depth_minus8, chroma_format_idc);
+ av_log(logctx, AV_LOG_TRACE,
+ "constant_frame_rate %d, ols_idx %d\n",
+ constant_frame_rate, ols_idx);
+ // VvcPTLRecord(num_sublayers) native_ptl
+ b = bytestream2_get_byte(&gb);
+ num_bytes_constraint_info = (b) & 0x3f;
+ b = bytestream2_get_byte(&gb);
+ general_profile_idc = (b >> 1) & 0x7f;
+ general_tier_flag = b & 1;
+ general_level_idc = bytestream2_get_byte(&gb);
+ av_log(logctx, AV_LOG_TRACE,
+ "general_profile_idc %d, general_tier_flag %d, "
+ "general_level_idc %d, num_sublayers %d num_bytes_constraint_info %d\n",
+ general_profile_idc, general_tier_flag, general_level_idc,
+ num_sublayers, num_bytes_constraint_info);
+
+ b = bytestream2_get_byte(&gb);
+ ptl_frame_only_constraint_flag = (b >> 7) & 1;
+ ptl_multi_layer_enabled_flag = (b >> 6) & 1;
+ for (i = 0; i < num_bytes_constraint_info - 1; i++) {
+ // unsigned int(8*num_bytes_constraint_info - 2) general_constraint_info;
+ bytestream2_get_byte(&gb);
+ }
+
+ av_log(logctx, AV_LOG_TRACE,
+ "ptl_multi_layer_enabled_flag %d, ptl_frame_only_constraint_flag %d\n",
+ ptl_multi_layer_enabled_flag,
+ ptl_frame_only_constraint_flag);
+
+ if (num_sublayers > 1) {
+ int temp6 = bytestream2_get_byte(&gb);
+ uint8_t ptl_sublayer_level_present_flag[8] = { 0 };
+ //uint8_t sublayer_level_idc[8] = {0};
+ for (i = num_sublayers - 2; i >= 0; i--) {
+ ptl_sublayer_level_present_flag[i] =
+ (temp6 >> (7 - (num_sublayers - 2 - i))) & 0x01;
+ }
+ for (i = num_sublayers - 2; i >= 0; i--) {
+ if (ptl_sublayer_level_present_flag[i]) {
+ bytestream2_get_byte(&gb); // sublayer_level_idc[8]
+ }
+ }
+ }
+
+ ptl_num_sub_profiles = bytestream2_get_byte(&gb);
+ for (j = 0; j < ptl_num_sub_profiles; j++) {
+ // unsigned int(32) general_sub_profile_idc[j];
+ bytestream2_get_be16(&gb);
+ bytestream2_get_be16(&gb);
+ }
+
+ max_picture_width = bytestream2_get_be16(&gb);
+ max_picture_height = bytestream2_get_be16(&gb);
+ avg_frame_rate = bytestream2_get_be16(&gb);
+ av_log(logctx, AV_LOG_TRACE,
+ "max_picture_width %d, max_picture_height %d, avg_frame_rate %d\n",
+ max_picture_width, max_picture_height, avg_frame_rate);
+ }
+
+ *is_nalff = 1;
+
+ /* nal units in the vvcC always have length coded with 2 bytes,
+ * so set nal_length_size = 2 while parsing them */
+ *nal_length_size = 2;
+
+ num_arrays = bytestream2_get_byte(&gb);
+ for (i = 0; i < num_arrays; i++) {
+ int cnt;
+ int type = bytestream2_get_byte(&gb) & 0x1f;
+
+ if (type == VVC_OPI_NUT || type == VVC_DCI_NUT)
+ cnt = 1;
+ else
+ cnt = bytestream2_get_be16(&gb);
+
+ av_log(logctx, AV_LOG_DEBUG, "nalu_type %d cnt %d\n", type, cnt);
+
+ if (!(type == VVC_OPI_NUT || type == VVC_DCI_NUT ||
+ type == VVC_VPS_NUT || type == VVC_SPS_NUT
+ || type == VVC_PPS_NUT || type == VVC_PREFIX_SEI_NUT
+ || type == VVC_SUFFIX_SEI_NUT)) {
+ av_log(logctx, AV_LOG_ERROR,
+ "Invalid NAL unit type in extradata: %d\n", type);
+ ret = AVERROR_INVALIDDATA;
+ return ret;
+ }
+
+ for (j = 0; j < cnt; j++) {
+ // +2 for the nal size field
+ int nalsize = bytestream2_peek_be16(&gb) + 2;
+ if (bytestream2_get_bytes_left(&gb) < nalsize) {
+ av_log(logctx, AV_LOG_ERROR,
+ "Invalid NAL unit size in extradata.\n");
+ return AVERROR_INVALIDDATA;
+ }
+
+ ret = h266_decode_nal_units(gb.buffer, nalsize, ps, *is_nalff,
+ *nal_length_size, err_recognition,
+ apply_defdispwin, logctx);
+ if (ret < 0) {
+ av_log(logctx, AV_LOG_ERROR,
+ "Decoding nal unit %d %d from vvcC failed\n", type, i);
+ return ret;
+ }
+ bytestream2_skip(&gb, nalsize);
+ }
+ }
+
+ /* store nal length size, that will be used to parse all other nals */
+ *nal_length_size = nal_len_size;
+ } else {
+ *is_nalff = 0;
+ ret = h266_decode_nal_units(data, size, ps, *is_nalff, *nal_length_size,
+ err_recognition, apply_defdispwin, logctx);
+ if (ret < 0)
+ return ret;
+ }
+
+ return ret;
+}
--git a/libavcodec/h266_parse_extradata.h b/libavcodec/h266_parse_extradata.h
new file mode 100644
index 0000000000..735a950488
--- /dev/null
+++ b/libavcodec/h266_parse_extradata.h
@@ -0,0 +1,36 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * H.266 parser code
+ */
+
+#ifndef AVCODEC_H266_PARSE_EXTRADATA_H
+#define AVCODEC_H266_PARSE_EXTRADATA_H
+
+#include <stdint.h>
+
+#include "h266_paramset.h"
+
+
+int ff_h266_decode_extradata(const uint8_t *data, int size, H266ParamSets *ps,
+ int *is_nalff, int *nal_length_size,
+ int err_recognition, int apply_defdispwin, void *logctx);
+
+#endif /* AVCODEC_266_PARSE_EXTRADATA_H */
diff --git a/libavcodec/libvvdec.c b/libavcodec/libvvdec.c
new file mode 100644
index 0000000000..8f621bbf05
--- /dev/null
+++ b/libavcodec/libvvdec.c
@@ -0,0 +1,563 @@
+/*
+ * H.266 decoding using the VVdeC library
+ *
+ * Copyright (C) 2022, Thomas Siedel
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "config_components.h"
+
+#include <vvdec/vvdec.h>
+
+#include "libavutil/common.h"
+#include "libavutil/avutil.h"
+#include "libavutil/pixdesc.h"
+#include "libavutil/opt.h"
+#include "libavutil/imgutils.h"
+#include "libavutil/frame.h"
+#include "libavutil/mastering_display_metadata.h"
+#include "libavutil/log.h"
+
+#include "avcodec.h"
+#include "codec_internal.h"
+#include "decode.h"
+#include "internal.h"
+#include "profiles.h"
+
+#include "h266_paramset.h"
+#include "h266_parse_extradata.h"
+
+typedef struct VVdeCContext {
+ AVClass *av_class;
+ vvdecDecoder *vvdecDec;
+ vvdecParams vvdecParams;
+ H266ParamSets ps;
+ int is_nalff;
+ int nal_length_size;
+ bool bFlush;
+ AVBufferPool *pools[3]; /** Pools for each data plane. */
+ int pool_size[3];
+} VVdeCContext;
+
+
+static void ff_vvdec_log_callback(void *avctx, int level, const char *fmt,
+ va_list args)
+{
+ vfprintf(level == 1 ? stderr : stdout, fmt, args);
+}
+
+static void *ff_vvdec_buffer_allocator(void *ctx, vvdecComponentType comp,
+ uint32_t size, uint32_t alignment,
+ void **allocator)
+{
+ AVBufferRef *buf;
+ VVdeCContext *s;
+ int plane;
+
+ uint32_t alignedsize = FFALIGN(size, alignment);
+ s = (VVdeCContext *) ctx;
+ plane = (int) comp;
+
+ if (plane < 0 || plane > 3)
+ return NULL;
+
+ if (alignedsize != s->pool_size[plane]) {
+ av_buffer_pool_uninit(&s->pools[plane]);
+ s->pools[plane] = av_buffer_pool_init(alignedsize, NULL);
+ if (!s->pools[plane]) {
+ s->pool_size[plane] = 0;
+ return NULL;
+ }
+ s->pool_size[plane] = alignedsize;
+ }
+
+ buf = av_buffer_pool_get(s->pools[plane]);
+ if (!buf)
+ return NULL;
+
+ *allocator = (void *) buf;
+ return buf->data;
+}
+
+static void ff_vvdec_buffer_unref(void *ctx, void *allocator)
+{
+ AVBufferRef *buf = (AVBufferRef *) allocator;
+ av_buffer_unref(&buf);
+}
+
+static void ff_vvdec_printParameterInfo(AVCodecContext *avctx,
+ vvdecParams *params)
+{
+ av_log(avctx, AV_LOG_DEBUG, "Version info: vvdec %s ( threads %d)\n",
+ vvdec_get_version(), params->threads);
+}
+
+static int ff_vvdec_set_pix_fmt(AVCodecContext *avctx, vvdecFrame *frame)
+{
+ if (NULL != frame->picAttributes && NULL != frame->picAttributes->vui &&
+ frame->picAttributes->vui->colourDescriptionPresentFlag) {
+ avctx->color_trc = frame->picAttributes->vui->transferCharacteristics;
+ avctx->color_primaries = frame->picAttributes->vui->colourPrimaries;
+ avctx->colorspace = frame->picAttributes->vui->matrixCoefficients;
+ } else {
+ avctx->color_primaries = AVCOL_PRI_UNSPECIFIED;
+ avctx->color_trc = AVCOL_TRC_UNSPECIFIED;
+ avctx->colorspace = AVCOL_SPC_UNSPECIFIED;
+ }
+
+ if (NULL != frame->picAttributes && NULL != frame->picAttributes->vui &&
+ frame->picAttributes->vui->videoSignalTypePresentFlag) {
+ avctx->color_range = frame->picAttributes->vui->videoFullRangeFlag ?
+ AVCOL_RANGE_JPEG : AVCOL_RANGE_MPEG;
+ } else {
+ avctx->color_range = AVCOL_RANGE_MPEG;
+ }
+
+ switch (frame->colorFormat) {
+ case VVDEC_CF_YUV420_PLANAR:
+ case VVDEC_CF_YUV400_PLANAR:
+ if (frame->bitDepth == 8) {
+ avctx->pix_fmt = frame->numPlanes == 1 ?
+ AV_PIX_FMT_GRAY8 : AV_PIX_FMT_YUV420P;
+ avctx->profile = FF_PROFILE_VVC_MAIN_10;
+ return 0;
+ } else if (frame->bitDepth == 10) {
+ avctx->pix_fmt = frame->numPlanes == 1 ?
+ AV_PIX_FMT_GRAY10 : AV_PIX_FMT_YUV420P10;
+ avctx->profile = FF_PROFILE_VVC_MAIN_10;
+ return 0;
+ } else {
+ return AVERROR_INVALIDDATA;
+ }
+ case VVDEC_CF_YUV422_PLANAR:
+ case VVDEC_CF_YUV444_PLANAR:
+ if (frame->bitDepth == 8) {
+ avctx->pix_fmt = frame->colorFormat == VVDEC_CF_YUV444_PLANAR ?
+ AV_PIX_FMT_YUV444P : AV_PIX_FMT_YUV422P;
+ if (avctx->profile != FF_PROFILE_VVC_MAIN_10_444)
+ avctx->profile = FF_PROFILE_VVC_MAIN_10_444;
+ return 0;
+ } else if (frame->bitDepth == 10) {
+ avctx->pix_fmt = frame->colorFormat == VVDEC_CF_YUV444_PLANAR ?
+ AV_PIX_FMT_YUV444P10 : AV_PIX_FMT_YUV422P10;
+ if (avctx->profile != FF_PROFILE_VVC_MAIN_10_444)
+ avctx->profile = FF_PROFILE_VVC_MAIN_10_444;
+ return 0;
+ } else {
+ return AVERROR_INVALIDDATA;
+ }
+ default:
+ return AVERROR_INVALIDDATA;
+ }
+}
+
+static int set_side_data(AVCodecContext *avctx, AVFrame *avframe,
+ vvdecFrame *frame)
+{
+ vvdecSEI *sei;
+ VVdeCContext *s = (VVdeCContext *) avctx->priv_data;
+
+ sei = vvdec_find_frame_sei(s->vvdecDec,
+ VVDEC_MASTERING_DISPLAY_COLOUR_VOLUME, frame);
+ if (sei) {
+ // VVC uses a g,b,r ordering, which we convert to a more natural r,g,b
+ const int mapping[3] = { 2, 0, 1 };
+ const int chroma_den = 50000;
+ const int luma_den = 10000;
+ int i;
+ vvdecSEIMasteringDisplayColourVolume *p;
+ AVMasteringDisplayMetadata *metadata =
+ av_mastering_display_metadata_create_side_data(avframe);
+ p = (vvdecSEIMasteringDisplayColourVolume *) (sei->payload);
+ if (p && metadata) {
+ for (i = 0; i < 3; i++) {
+ const int j = mapping[i];
+ metadata->display_primaries[i][0].num = p->primaries[j][0];
+ metadata->display_primaries[i][0].den = chroma_den;
+ metadata->display_primaries[i][1].num = p->primaries[j][1];
+ metadata->display_primaries[i][1].den = chroma_den;
+ }
+ metadata->white_point[0].num = p->whitePoint[0];
+ metadata->white_point[0].den = chroma_den;
+ metadata->white_point[1].num = p->whitePoint[1];
+ metadata->white_point[1].den = chroma_den;
+
+ metadata->max_luminance.num = p->maxLuminance;
+ metadata->max_luminance.den = luma_den;
+ metadata->min_luminance.num = p->minLuminance;
+ metadata->min_luminance.den = luma_den;
+ metadata->has_luminance = 1;
+ metadata->has_primaries = 1;
+
+ av_log(avctx, AV_LOG_DEBUG, "Mastering Display Metadata:\n");
+ av_log(avctx, AV_LOG_DEBUG,
+ "r(%5.4f,%5.4f) g(%5.4f,%5.4f) b(%5.4f %5.4f) wp(%5.4f, %5.4f)\n",
+ av_q2d(metadata->display_primaries[0][0]),
+ av_q2d(metadata->display_primaries[0][1]),
+ av_q2d(metadata->display_primaries[1][0]),
+ av_q2d(metadata->display_primaries[1][1]),
+ av_q2d(metadata->display_primaries[2][0]),
+ av_q2d(metadata->display_primaries[2][1]),
+ av_q2d(metadata->white_point[0]),
+ av_q2d(metadata->white_point[1]));
+ av_log(avctx, AV_LOG_DEBUG, "min_luminance=%f, max_luminance=%f\n",
+ av_q2d(metadata->min_luminance),
+ av_q2d(metadata->max_luminance));
+ }
+ return 0;
+ }
+
+ sei = vvdec_find_frame_sei(s->vvdecDec, VVDEC_CONTENT_LIGHT_LEVEL_INFO,
+ frame);
+ if (sei) {
+ vvdecSEIContentLightLevelInfo *p = NULL;
+ AVContentLightMetadata *light =
+ av_content_light_metadata_create_side_data(avframe);
+ p = (vvdecSEIContentLightLevelInfo *) (sei->payload);
+ if (p && light) {
+ light->MaxCLL = p->maxContentLightLevel;
+ light->MaxFALL = p->maxPicAverageLightLevel;
+ }
+
+ av_log(avctx, AV_LOG_DEBUG, "Content Light Level Metadata:\n");
+ av_log(avctx, AV_LOG_DEBUG, "MaxCLL=%d, MaxFALL=%d\n",
+ light->MaxCLL, light->MaxFALL);
+ }
+
+ return 0;
+}
+
+static void export_stream_params(AVCodecContext *avctx, const H266SPS *sps)
+{
+ avctx->coded_width = sps->pic_width_max_in_luma_samples;
+ avctx->coded_height = sps->pic_height_max_in_luma_samples;
+ avctx->width = sps->pic_width_max_in_luma_samples -
+ sps->conf_win_left_offset -
+ sps->conf_win_right_offset;
+ avctx->height = sps->pic_height_max_in_luma_samples -
+ sps->conf_win_top_offset -
+ sps->conf_win_bottom_offset;
+ avctx->has_b_frames = sps->max_sublayers;
+ avctx->profile = sps->profile_tier_level.general_profile_idc;
+ avctx->level = sps->profile_tier_level.general_level_idc;
+ avctx->pix_fmt = sps->pix_fmt;
+
+ avctx->color_range = sps->vui.full_range_flag ? AVCOL_RANGE_JPEG :
+ AVCOL_RANGE_MPEG;
+
+ if (sps->vui.colour_description_present_flag) {
+ avctx->color_primaries = sps->vui.colour_primaries;
+ avctx->color_trc = sps->vui.transfer_characteristics;
+ avctx->colorspace = sps->vui.matrix_coeffs;
+ } else {
+ avctx->color_primaries = AVCOL_PRI_UNSPECIFIED;
+ avctx->color_trc = AVCOL_TRC_UNSPECIFIED;
+ avctx->colorspace = AVCOL_SPC_UNSPECIFIED;
+ }
+
+ avctx->chroma_sample_location = AVCHROMA_LOC_UNSPECIFIED;
+ if (sps->chroma_format_idc == 1) {
+ if (sps->vui.chroma_loc_info_present_flag) {
+ if (sps->vui.chroma_sample_loc_type_top_field <= 5)
+ avctx->chroma_sample_location =
+ sps->vui.chroma_sample_loc_type_top_field + 1;
+ } else
+ avctx->chroma_sample_location = AVCHROMA_LOC_LEFT;
+ }
+
+ if (sps->timing_hrd_params_present_flag &&
+ sps->general_timing_hrd_parameters.num_units_in_tick &&
+ sps->general_timing_hrd_parameters.time_scale) {
+ av_reduce(&avctx->framerate.den, &avctx->framerate.num,
+ sps->general_timing_hrd_parameters.num_units_in_tick,
+ sps->general_timing_hrd_parameters.time_scale, INT_MAX);
+ }
+}
+
+static int h266_decode_extradata(AVCodecContext *avctx, uint8_t *buf,
+ int length, int first)
+{
+ VVdeCContext *s = (VVdeCContext *) avctx->priv_data;
+ int ret;
+
+ ret = ff_h266_decode_extradata(buf, length, &s->ps, &s->is_nalff,
+ &s->nal_length_size, avctx->err_recognition,
+ false, avctx);
+ if (ret < 0)
+ return ret;
+
+ if (s->ps.sps != NULL)
+ export_stream_params(avctx, s->ps.sps);
+
+ return 0;
+}
+
+static av_cold int ff_vvdec_decode_init(AVCodecContext *avctx)
+{
+ int i;
+ VVdeCContext *s = (VVdeCContext *) avctx->priv_data;
+
+ vvdec_params_default(&s->vvdecParams);
+ s->vvdecParams.logLevel = VVDEC_DETAILS;
+
+ if (av_log_get_level() >= AV_LOG_DEBUG)
+ s->vvdecParams.logLevel = VVDEC_DETAILS;
+ else if (av_log_get_level() >= AV_LOG_VERBOSE)
+ s->vvdecParams.logLevel = VVDEC_INFO; // VVDEC_INFO will output per picture info
+ else if (av_log_get_level() >= AV_LOG_INFO)
+ s->vvdecParams.logLevel = VVDEC_WARNING; // AV_LOG_INFO is ffmpeg default
+ else
+ s->vvdecParams.logLevel = VVDEC_SILENT;
+
+ if (avctx->thread_count > 0)
+ s->vvdecParams.threads = avctx->thread_count; // number of worker threads (should not exceed the number of physical cpu's)
+ else
+ s->vvdecParams.threads = -1; // get max cpus
+
+ ff_vvdec_printParameterInfo(avctx, &s->vvdecParams);
+
+ // using buffer allocation by using AVBufferPool
+ s->vvdecParams.opaque = avctx->priv_data;
+ s->vvdecDec = vvdec_decoder_open_with_allocator(&s->vvdecParams,
+ ff_vvdec_buffer_allocator,
+ ff_vvdec_buffer_unref);
+
+
+ if (!s->vvdecDec) {
+ av_log(avctx, AV_LOG_ERROR, "cannot init vvc decoder\n");
+ return -1;
+ }
+
+ vvdec_set_logging_callback(s->vvdecDec, ff_vvdec_log_callback);
+
+ s->bFlush = false;
+ s->is_nalff = 0;
+ s->nal_length_size = 0;
+
+ for (i = 0; i < FF_ARRAY_ELEMS(s->pools); i++) {
+ s->pools[i] = NULL;
+ s->pool_size[i] = 0;
+ }
+
+ if (!avctx->internal->is_copy) {
+ if (avctx->extradata_size > 0 && avctx->extradata) {
+ int ret = h266_decode_extradata(avctx, avctx->extradata,
+ avctx->extradata_size, 1);
+ if (ret < 0) {
+ return ret;
+ }
+ }
+ }
+
+ return 0;
+}
+
+static av_cold int ff_vvdec_decode_close(AVCodecContext *avctx)
+{
+ VVdeCContext *s = (VVdeCContext *) avctx->priv_data;
+
+ for (int i = 0; i < FF_ARRAY_ELEMS(s->pools); i++) {
+ av_buffer_pool_uninit(&s->pools[i]);
+ s->pool_size[i] = 0;
+ }
+
+ if (0 != vvdec_decoder_close(s->vvdecDec)) {
+ av_log(avctx, AV_LOG_ERROR, "cannot close vvdec\n");
+ return -1;
+ }
+
+ ff_h266_ps_uninit(&s->ps);
+ s->bFlush = false;
+ return 0;
+}
+
+static av_cold int ff_vvdec_decode_frame(AVCodecContext *avctx, AVFrame *data,
+ int *got_frame, AVPacket *avpkt)
+{
+ VVdeCContext *s = avctx->priv_data;
+ AVFrame *avframe = data;
+
+ int ret = 0;
+ vvdecFrame *frame = NULL;
+
+ if (avframe) {
+ if (!avpkt->size && !s->bFlush)
+ s->bFlush = true;
+
+ if (s->bFlush)
+ ret = vvdec_flush(s->vvdecDec, &frame);
+ else {
+ vvdecAccessUnit accessUnit;
+ vvdec_accessUnit_default(&accessUnit);
+ accessUnit.payload = avpkt->data;
+ accessUnit.payloadSize = avpkt->size;
+ accessUnit.payloadUsedSize = avpkt->size;
+
+ accessUnit.cts = avpkt->pts;
+ accessUnit.ctsValid = true;
+ accessUnit.dts = avpkt->dts;
+ accessUnit.dtsValid = true;
+
+ ret = vvdec_decode(s->vvdecDec, &accessUnit, &frame);
+ }
+
+ if (ret < 0) {
+ if (ret == VVDEC_EOF)
+ s->bFlush = true;
+ else if (ret != VVDEC_TRY_AGAIN) {
+ av_log(avctx, AV_LOG_ERROR,
+ "error in vvdec::decode - ret:%d - %s %s\n", ret,
+ vvdec_get_last_error(s->vvdecDec), vvdec_get_last_additional_error( s->vvdecDec));
+ ret=AVERROR_EXTERNAL;
+ goto fail;
+ }
+ } else if (NULL != frame) {
+ const uint8_t *src_data[4] = { frame->planes[0].ptr,
+ frame->planes[1].ptr,
+ frame->planes[2].ptr, NULL };
+ const int src_linesizes[4] = { (int) frame->planes[0].stride,
+ (int) frame->planes[1].stride,
+ (int) frame->planes[2].stride, 0 };
+
+ if ((ret = ff_vvdec_set_pix_fmt(avctx, frame)) < 0) {
+ av_log(avctx, AV_LOG_ERROR,
+ "Unsupported output colorspace (%d) / bit_depth (%d)\n",
+ frame->colorFormat, frame->bitDepth);
+ goto fail;
+ }
+
+ if ((int) frame->width != avctx->width ||
+ (int) frame->height != avctx->height) {
+ av_log(avctx, AV_LOG_INFO, "dimension change! %dx%d -> %dx%d\n",
+ avctx->width, avctx->height, frame->width, frame->height);
+
+ ret = ff_set_dimensions(avctx, frame->width, frame->height);
+ if (ret < 0)
+ goto fail;
+ }
+
+ if (frame->planes[0].allocator)
+ avframe->buf[0] =
+ av_buffer_ref((AVBufferRef *) frame->planes[0].allocator);
+ if (frame->planes[1].allocator)
+ avframe->buf[1] =
+ av_buffer_ref((AVBufferRef *) frame->planes[1].allocator);
+ if (frame->planes[2].allocator)
+ avframe->buf[2] =
+ av_buffer_ref((AVBufferRef *) frame->planes[2].allocator);
+
+ for (int i = 0; i < 4; i++) {
+ avframe->data[i] = (uint8_t *) src_data[i];
+ avframe->linesize[i] = src_linesizes[i];
+ }
+
+ ret = ff_decode_frame_props(avctx, avframe);
+ if (ret < 0)
+ goto fail;
+
+ if (frame->picAttributes) {
+ avframe->key_frame = frame->picAttributes->isRefPic;
+ avframe->pict_type = (frame->picAttributes->sliceType !=
+ VVDEC_SLICETYPE_UNKNOWN) ?
+ frame->picAttributes->sliceType + 1 : AV_PICTURE_TYPE_NONE;
+ }
+
+ if (frame->ctsValid)
+ avframe->pts = frame->cts;
+
+ ret = set_side_data(avctx, avframe, frame);
+ if (ret < 0)
+ goto fail;
+
+ if (0 != vvdec_frame_unref(s->vvdecDec, frame))
+ av_log(avctx, AV_LOG_ERROR, "cannot free picture memory\n");
+
+ *got_frame = 1;
+ }
+ }
+
+ return avpkt->size;
+
+ fail:
+ if (frame) {
+ if (frame->planes[0].allocator)
+ av_buffer_unref((AVBufferRef **) &frame->planes[0].allocator);
+ if (frame->planes[1].allocator)
+ av_buffer_unref((AVBufferRef **) &frame->planes[1].allocator);
+ if (frame->planes[2].allocator)
+ av_buffer_unref((AVBufferRef **) &frame->planes[2].allocator);
+
+ vvdec_frame_unref(s->vvdecDec, frame);
+ }
+ return ret;
+}
+
+static av_cold void ff_vvdec_decode_flush(AVCodecContext *avctx)
+{
+ VVdeCContext *s = (VVdeCContext *) avctx->priv_data;
+
+ if (0 != vvdec_decoder_close(s->vvdecDec))
+ av_log(avctx, AV_LOG_ERROR, "cannot close vvdec during flush\n");
+
+ s->vvdecDec = vvdec_decoder_open_with_allocator(&s->vvdecParams,
+ ff_vvdec_buffer_allocator,
+ ff_vvdec_buffer_unref);
+ if (!s->vvdecDec)
+ av_log(avctx, AV_LOG_ERROR, "cannot reinit vvdec during flush\n");
+
+ vvdec_set_logging_callback(s->vvdecDec, ff_vvdec_log_callback);
+
+ s->bFlush = false;
+}
+
+static const enum AVPixelFormat pix_fmts_vvdec[] = {
+ AV_PIX_FMT_GRAY8,
+ AV_PIX_FMT_GRAY10,
+ AV_PIX_FMT_YUV420P,
+ AV_PIX_FMT_YUV422P,
+ AV_PIX_FMT_YUV444P,
+ AV_PIX_FMT_YUV420P10LE,
+ AV_PIX_FMT_YUV422P10LE,
+ AV_PIX_FMT_YUV444P10LE,
+ AV_PIX_FMT_NONE
+};
+
+static const AVClass class_libvvdec = {
+ .class_name = "libvvdec-vvc decoder",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+};
+
+FFCodec ff_libvvdec_decoder = {
+ .p.name = "libvvdec",
+ CODEC_LONG_NAME("H.266 / VVC Decoder VVdeC"),
+ .p.type = AVMEDIA_TYPE_VIDEO,
+ .p.id = AV_CODEC_ID_VVC,
+ .p.capabilities = AV_CODEC_CAP_DELAY | AV_CODEC_CAP_OTHER_THREADS,
+ .p.profiles = NULL_IF_CONFIG_SMALL(ff_vvc_profiles),
+ .p.priv_class = &class_libvvdec,
+ .p.wrapper_name = "libvvdec",
+ .priv_data_size = sizeof(VVdeCContext),
+ .p.pix_fmts = pix_fmts_vvdec,
+ .init = ff_vvdec_decode_init,
+ FF_CODEC_DECODE_CB(ff_vvdec_decode_frame),
+ .close = ff_vvdec_decode_close,
+ .flush = ff_vvdec_decode_flush,
+ .bsfs = "h266_mp4toannexb",
+ .caps_internal = FF_CODEC_CAP_AUTO_THREADS,
+};
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 09/11] avcodec: add external encoder libvvenc for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (7 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 08/11] avcodec: add external decoder libvvdec " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 10/11] avformat: add ts stream types " Thomas Siedel
` (2 subsequent siblings)
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
Add external encoder VVenC for H266/VVC encoding.
Register new encoder libvvenc.
Add libvvenc to wrap the vvenc interface.
libvvenc implements encoder option: preset,qp,period,subjopt,
vvenc-params,levelidc,tier.
Enable encoder by adding --enable-libvvenc in configure step.
---
configure | 5 +
libavcodec/Makefile | 1 +
libavcodec/allcodecs.c | 1 +
libavcodec/libvvenc.c | 469 +++++++++++++++++++++++++++++++++++++++++
4 files changed, 476 insertions(+)
create mode 100644 libavcodec/libvvenc.c
diff --git a/configure b/configure
index 6568493849..0444422b18 100755
--- a/configure
+++ b/configure
@@ -289,6 +289,7 @@ External library support:
native implementation exists [no]
--enable-libvpx enable VP8 and VP9 de/encoding via libvpx [no]
--enable-libvvdec enable H.266/VVC decoding via vvdec [no]
+ --enable-libvvenc enable H.266/VVC encoding via vvenc [no]
--enable-libwebp enable WebP encoding via libwebp [no]
--enable-libx264 enable H.264 encoding via x264 [no]
--enable-libx265 enable HEVC encoding via x265 [no]
@@ -1863,6 +1864,7 @@ EXTERNAL_LIBRARY_LIST="
libvorbis
libvpx
libvvdec
+ libvvenc
libwebp
libxml2
libzimg
@@ -3395,6 +3397,8 @@ libvpx_vp9_decoder_deps="libvpx"
libvpx_vp9_encoder_deps="libvpx"
libvvdec_decoder_deps="libvvdec"
libvvdec_decoder_select="h266_mp4toannexb_bsf"
+libvvenc_encoder_deps="libvvenc"
+libvvenc_encoder_select="atsc_a53"
libwebp_encoder_deps="libwebp"
libwebp_anim_encoder_deps="libwebp"
libx262_encoder_deps="libx262"
@@ -6729,6 +6733,7 @@ enabled libvpx && {
fi
}
enabled libvvdec && require_pkg_config libvvdec "libvvdec >= 1.6.0" "vvdec/vvdec.h" vvdec_get_version
+enabled libvvenc && require_pkg_config libvvenc "libvvenc >= 1.6.1" "vvenc/vvenc.h" vvenc_get_version
enabled libwebp && {
enabled libwebp_encoder && require_pkg_config libwebp "libwebp >= 0.2.0" webp/encode.h WebPGetEncoderVersion
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 2e7fb2ae12..0c21b5f015 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -1125,6 +1125,7 @@ OBJS-$(CONFIG_LIBVPX_VP8_ENCODER) += libvpxenc.o
OBJS-$(CONFIG_LIBVPX_VP9_DECODER) += libvpxdec.o
OBJS-$(CONFIG_LIBVPX_VP9_ENCODER) += libvpxenc.o
OBJS-$(CONFIG_LIBVVDEC_DECODER) += libvvdec.o h266_parse_extradata.o h266_paramset.o
+OBJS-$(CONFIG_LIBVVENC_ENCODER) += libvvenc.o
OBJS-$(CONFIG_LIBWEBP_ENCODER) += libwebpenc_common.o libwebpenc.o
OBJS-$(CONFIG_LIBWEBP_ANIM_ENCODER) += libwebpenc_common.o libwebpenc_animencoder.o
OBJS-$(CONFIG_LIBX262_ENCODER) += libx264.o
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index 18bbe62808..8d501e5f1c 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -800,6 +800,7 @@ extern const FFCodec ff_libvpx_vp8_decoder;
extern FFCodec ff_libvpx_vp9_encoder;
extern const FFCodec ff_libvpx_vp9_decoder;
extern const FFCodec ff_libvvdec_decoder;
+extern const FFCodec ff_libvvenc_encoder;
/* preferred over libwebp */
extern const FFCodec ff_libwebp_anim_encoder;
extern const FFCodec ff_libwebp_encoder;
diff --git a/libavcodec/libvvenc.c b/libavcodec/libvvenc.c
new file mode 100644
index 0000000000..52097aeaa8
--- /dev/null
+++ b/libavcodec/libvvenc.c
@@ -0,0 +1,469 @@
+/*
+ * H.266 encoding using the VVenC library
+ *
+ * Copyright (C) 2022, Thomas Siedel
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "config_components.h"
+
+#include <vvenc/vvenc.h>
+#include <vvenc/vvencCfg.h>
+
+#include "avcodec.h"
+#include "codec_internal.h"
+#include "encode.h"
+#include "internal.h"
+#include "packet_internal.h"
+#include "profiles.h"
+
+#include "libavutil/avutil.h"
+#include "libavutil/pixdesc.h"
+#include "libavutil/opt.h"
+#include "libavutil/common.h"
+#include "libavutil/imgutils.h"
+#include "libavutil/frame.h"
+#include "libavutil/log.h"
+
+typedef struct VVenCOptions {
+ int preset; // preset 0: faster 4: slower
+ int qp; // quantization parameter 0-63
+ int subjectiveOptimization; // perceptually motivated QP adaptation, XPSNR based
+ int intraRefreshSec; // intra period/refresh in seconds
+ int levelIdc; // vvc level_idc
+ int tier; // vvc tier
+ AVDictionary *vvenc_opts;
+} VVenCOptions;
+
+typedef struct VVenCContext {
+ AVClass *av_class;
+ VVenCOptions options; // encoder options
+ vvencEncoder *vvencEnc;
+ vvencAccessUnit *pAU;
+ bool encodeDone;
+} VVenCContext;
+
+
+static av_cold void ff_vvenc_log_callback(void *avctx, int level,
+ const char *fmt, va_list args)
+{
+ vfprintf(level == 1 ? stderr : stdout, fmt, args);
+}
+
+static void ff_vvenc_internalLog(void *ctx, int level, const char *fmt, ...)
+{
+ va_list args;
+ va_start(args, fmt);
+ ff_vvenc_log_callback(ctx, level, fmt, args);
+ va_end(args);
+}
+
+static av_cold int ff_vvenc_encode_init(AVCodecContext *avctx)
+{
+ int ret;
+ int framerate, qp, parse_ret;
+ VVenCContext *s;
+ vvenc_config params;
+ vvencPresetMode preset;
+ AVDictionaryEntry *en;
+ char statsfile[1024] = "vvenc-rcstats.json";
+
+ s = (VVenCContext *) avctx->priv_data;
+ qp = (vvencPresetMode) s->options.qp;
+ preset = (vvencPresetMode) s->options.preset;
+
+ if (avctx->flags & AV_CODEC_FLAG_INTERLACED_DCT) {
+ av_log(avctx, AV_LOG_ERROR,
+ "ff_vvenc_encode_init::init() interlaced encoding not supported yet\n");
+ return AVERROR_INVALIDDATA;
+ }
+
+ vvenc_config_default(¶ms);
+
+ // set desired encoding options
+ framerate = avctx->time_base.den / avctx->time_base.num;
+ vvenc_init_default(¶ms, avctx->width, avctx->height, framerate,
+ avctx->bit_rate, qp, preset);
+ params.m_FrameRate = avctx->time_base.den;
+ params.m_FrameScale = avctx->time_base.num;
+
+ params.m_verbosity = VVENC_VERBOSE;
+ if (av_log_get_level() >= AV_LOG_DEBUG)
+ params.m_verbosity = VVENC_DETAILS;
+ else if (av_log_get_level() >= AV_LOG_VERBOSE)
+ params.m_verbosity = VVENC_NOTICE; // output per picture info
+ else if (av_log_get_level() >= AV_LOG_INFO)
+ params.m_verbosity = VVENC_WARNING; // ffmpeg default ffmpeg loglevel
+ else
+ params.m_verbosity = VVENC_SILENT;
+
+ if (avctx->ticks_per_frame == 1) {
+ params.m_TicksPerSecond = -1; // auto mode for ticks per frame = 1
+ } else {
+ params.m_TicksPerSecond =
+ ceil((avctx->time_base.den / (double) avctx->time_base.num) *
+ (double) avctx->ticks_per_frame);
+ }
+
+ if (avctx->thread_count > 0)
+ params.m_numThreads = avctx->thread_count;
+
+ // GOP settings (IDR/CRA)
+ if (avctx->flags & AV_CODEC_FLAG_CLOSED_GOP)
+ params.m_DecodingRefreshType = VVENC_DRT_IDR;
+
+ if (avctx->gop_size == 1) {
+ params.m_GOPSize = 1;
+ params.m_IntraPeriod = 1;
+ } else {
+ params.m_IntraPeriodSec = s->options.intraRefreshSec;
+ }
+
+ params.m_usePerceptQPA = s->options.subjectiveOptimization;
+ params.m_level = (vvencLevel) s->options.levelIdc;
+ params.m_levelTier = (vvencTier) s->options.tier;
+
+ params.m_AccessUnitDelimiter = true;
+
+ params.m_internChromaFormat = VVENC_CHROMA_420;
+ switch (avctx->pix_fmt) {
+ case AV_PIX_FMT_YUV420P:
+ params.m_inputBitDepth[0] = 8;
+ break;
+ case AV_PIX_FMT_YUV420P10LE:
+ params.m_inputBitDepth[0] = 10;
+ break;
+ default:{
+ av_log(avctx, AV_LOG_ERROR,
+ "unsupported pixel format %s, choose yuv420p or yuv420p10le\n",
+ av_get_pix_fmt_name(avctx->pix_fmt));
+ return AVERROR(EINVAL);
+ break;
+ }
+ }
+
+ if (avctx->color_primaries != AVCOL_PRI_UNSPECIFIED)
+ params.m_colourPrimaries = (int) avctx->color_primaries;
+ if (avctx->colorspace != AVCOL_SPC_UNSPECIFIED)
+ params.m_matrixCoefficients = (int) avctx->colorspace;
+ if (avctx->color_trc != AVCOL_TRC_UNSPECIFIED) {
+ params.m_transferCharacteristics = (int) avctx->color_trc;
+
+ if (avctx->color_trc == AVCOL_TRC_SMPTE2084)
+ params.m_HdrMode = (avctx->color_primaries == AVCOL_PRI_BT2020) ?
+ VVENC_HDR_PQ_BT2020 : VVENC_HDR_PQ;
+ else if (avctx->color_trc == AVCOL_TRC_BT2020_10
+ || avctx->color_trc == AVCOL_TRC_ARIB_STD_B67)
+ params.m_HdrMode = (avctx->color_trc == AVCOL_TRC_BT2020_10 ||
+ avctx->color_primaries == AVCOL_PRI_BT2020 ||
+ avctx->colorspace == AVCOL_SPC_BT2020_NCL ||
+ avctx->colorspace == AVCOL_SPC_BT2020_CL) ?
+ VVENC_HDR_HLG_BT2020 : VVENC_HDR_HLG;
+ }
+
+ if (params.m_HdrMode == VVENC_HDR_OFF
+ && (avctx->color_primaries != AVCOL_PRI_UNSPECIFIED
+ || avctx->colorspace != AVCOL_SPC_UNSPECIFIED)) {
+ params.m_vuiParametersPresent = 1;
+ params.m_colourDescriptionPresent = true;
+ }
+
+ params.m_RCNumPasses = 1;
+ en = NULL;
+ while ((en = av_dict_get(s->options.vvenc_opts, "", en,
+ AV_DICT_IGNORE_SUFFIX))) {
+ av_log(avctx, AV_LOG_DEBUG, "vvenc_set_param: '%s:%s'\n", en->key,
+ en->value);
+ parse_ret = vvenc_set_param(¶ms, en->key, en->value);
+ switch (parse_ret) {
+ case VVENC_PARAM_BAD_NAME:
+ av_log(avctx, AV_LOG_WARNING, "Unknown vvenc option: %s.\n",
+ en->key);
+ break;
+ case VVENC_PARAM_BAD_VALUE:
+ av_log(avctx, AV_LOG_WARNING,
+ "Invalid vvenc value for %s: %s.\n", en->key, en->value);
+ break;
+ default:
+ break;
+ }
+
+ if (memcmp(en->key, "rcstatsfile", 11) == 0 ||
+ memcmp(en->key, "RCStatsFile", 11) == 0) {
+ strncpy(statsfile, en->value, sizeof(statsfile) - 1);
+ statsfile[sizeof(statsfile) - 1] = '\0';
+ }
+ }
+
+ if (params.m_RCPass != -1 && params.m_RCNumPasses == 1)
+ params.m_RCNumPasses = 2; // enable 2pass mode
+
+ s->vvencEnc = vvenc_encoder_create();
+ if (NULL == s->vvencEnc) {
+ av_log(avctx, AV_LOG_ERROR, "cannot create vvc encoder (vvenc)\n");
+ return AVERROR(ENOMEM);
+ }
+
+ vvenc_set_msg_callback(¶ms, s->vvencEnc, ff_vvenc_log_callback);
+ ret = vvenc_encoder_open(s->vvencEnc, ¶ms);
+ if (0 != ret) {
+ av_log(avctx, AV_LOG_ERROR, "cannot open vvc encoder (vvenc): %s\n",
+ vvenc_get_last_error(s->vvencEnc));
+ vvenc_encoder_close(s->vvencEnc);
+ return AVERROR(EINVAL);
+ }
+
+ vvenc_get_config(s->vvencEnc, ¶ms); // get the adapted config
+
+ if (params.m_verbosity >= VVENC_DETAILS
+ && av_log_get_level() < AV_LOG_DEBUG) {
+ ff_vvenc_internalLog(avctx, params.m_verbosity, "vvenc version: %s\n",
+ vvenc_get_version());
+ ff_vvenc_internalLog(avctx, params.m_verbosity, "vvenc info:\n%s\n",
+ vvenc_get_config_as_string(¶ms,
+ VVENC_DETAILS));
+ } else {
+ av_log(avctx, AV_LOG_DEBUG, "vvenc version: %s\n", vvenc_get_version());
+ av_log(avctx, AV_LOG_DEBUG, "vvenc info:\n%s\n",
+ vvenc_get_config_as_string(¶ms, VVENC_DETAILS));
+ }
+
+ if (params.m_RCNumPasses == 2) {
+ ret = vvenc_init_pass(s->vvencEnc, params.m_RCPass - 1, &statsfile[0]);
+ if (0 != ret) {
+ av_log(avctx, AV_LOG_ERROR,
+ "cannot init pass %d for vvc encoder (vvenc): %s\n",
+ params.m_RCPass, vvenc_get_last_error(s->vvencEnc));
+ vvenc_encoder_close(s->vvencEnc);
+ return AVERROR(EINVAL);
+ }
+ }
+
+ s->pAU = vvenc_accessUnit_alloc();
+ vvenc_accessUnit_alloc_payload(s->pAU, avctx->width * avctx->height);
+
+ if (avctx->flags & AV_CODEC_FLAG_GLOBAL_HEADER) {
+ ret = vvenc_get_headers(s->vvencEnc, s->pAU);
+ if (0 != ret) {
+ av_log(avctx, AV_LOG_ERROR,
+ "cannot get headers (SPS,PPS) from vvc encoder(vvenc): %s\n",
+ vvenc_get_last_error(s->vvencEnc));
+ vvenc_encoder_close(s->vvencEnc);
+ return AVERROR(EINVAL);
+ }
+
+ if (s->pAU->payloadUsedSize <= 0) {
+ vvenc_encoder_close(s->vvencEnc);
+ return AVERROR_INVALIDDATA;
+ }
+
+ avctx->extradata_size = s->pAU->payloadUsedSize;
+ avctx->extradata =
+ av_malloc(avctx->extradata_size + AV_INPUT_BUFFER_PADDING_SIZE);
+ if (!avctx->extradata) {
+ av_log(avctx, AV_LOG_ERROR,
+ "Cannot allocate VVC header of size %d.\n",
+ avctx->extradata_size);
+ vvenc_encoder_close(s->vvencEnc);
+ return AVERROR(ENOMEM);
+ }
+
+ memcpy(avctx->extradata, s->pAU->payload, avctx->extradata_size);
+ memset(avctx->extradata + avctx->extradata_size, 0,
+ AV_INPUT_BUFFER_PADDING_SIZE);
+ }
+ s->encodeDone = false;
+ return 0;
+}
+
+static av_cold int ff_vvenc_encode_close(AVCodecContext * avctx)
+{
+ VVenCContext *s = (VVenCContext *) avctx->priv_data;
+ if (s->vvencEnc) {
+ if (av_log_get_level() >= AV_LOG_VERBOSE)
+ vvenc_print_summary(s->vvencEnc);
+
+ if (0 != vvenc_encoder_close(s->vvencEnc)) {
+ av_log(avctx, AV_LOG_ERROR, "cannot close vvenc\n");
+ return -1;
+ }
+ }
+
+ vvenc_accessUnit_free(s->pAU, true);
+
+ return 0;
+}
+
+static av_cold int ff_vvenc_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
+ const AVFrame *frame, int *got_packet)
+{
+ VVenCContext *s = (VVenCContext *) avctx->priv_data;
+ vvencYUVBuffer *pyuvbuf;
+ vvencYUVBuffer yuvbuf;
+ int pict_type;
+ int ret;
+
+ pyuvbuf = NULL;
+ if (frame) {
+ if (avctx->pix_fmt == AV_PIX_FMT_YUV420P10LE) {
+ vvenc_YUVBuffer_default(&yuvbuf);
+ yuvbuf.planes[0].ptr = (int16_t *) frame->data[0];
+ yuvbuf.planes[1].ptr = (int16_t *) frame->data[1];
+ yuvbuf.planes[2].ptr = (int16_t *) frame->data[2];
+
+ yuvbuf.planes[0].width = frame->width;
+ yuvbuf.planes[0].height = frame->height;
+ yuvbuf.planes[0].stride = frame->linesize[0] >> 1; // stride is used in samples (16bit) in vvenc, ffmpeg uses stride in bytes
+
+ yuvbuf.planes[1].width = frame->width >> 1;
+ yuvbuf.planes[1].height = frame->height >> 1;
+ yuvbuf.planes[1].stride = frame->linesize[1] >> 1;
+
+ yuvbuf.planes[2].width = frame->width >> 1;
+ yuvbuf.planes[2].height = frame->height >> 1;
+ yuvbuf.planes[2].stride = frame->linesize[2] >> 1;
+
+ yuvbuf.cts = frame->pts;
+ yuvbuf.ctsValid = true;
+ pyuvbuf = &yuvbuf;
+ } else {
+ av_log(avctx, AV_LOG_ERROR,
+ "unsupported input colorspace! input must be yuv420p10le");
+ return AVERROR(EINVAL);
+ }
+ }
+
+ if (!s->encodeDone) {
+ ret = vvenc_encode(s->vvencEnc, pyuvbuf, s->pAU, &s->encodeDone);
+ if (ret != 0) {
+ av_log(avctx, AV_LOG_ERROR, "error in vvenc::encode - ret:%d\n",
+ ret);
+ return AVERROR(EINVAL);
+ }
+ } else {
+ *got_packet = 0;
+ return 0;
+ }
+
+ if (s->pAU->payloadUsedSize > 0) {
+ ret = ff_get_encode_buffer(avctx, pkt, s->pAU->payloadUsedSize, 0);
+ if (ret < 0) {
+ av_log(avctx, AV_LOG_ERROR, "Error getting output packet.\n");
+ return ret;
+ }
+
+ memcpy(pkt->data, s->pAU->payload, s->pAU->payloadUsedSize);
+
+ if (s->pAU->ctsValid)
+ pkt->pts = s->pAU->cts;
+ if (s->pAU->dtsValid)
+ pkt->dts = s->pAU->dts;
+ pkt->flags |= AV_PKT_FLAG_KEY * s->pAU->rap;
+
+ switch (s->pAU->sliceType) {
+ case VVENC_I_SLICE:
+ pict_type = AV_PICTURE_TYPE_I;
+ break;
+ case VVENC_P_SLICE:
+ pict_type = AV_PICTURE_TYPE_P;
+ break;
+ case VVENC_B_SLICE:
+ pict_type = AV_PICTURE_TYPE_B;
+ break;
+ default:
+ av_log(avctx, AV_LOG_ERROR, "Unknown picture type encountered.\n");
+ return AVERROR_EXTERNAL;
+ }
+
+ ff_side_data_set_encoder_stats(pkt, 0, NULL, 0, pict_type);
+
+ *got_packet = 1;
+
+ return 0;
+ } else {
+ *got_packet = 0;
+ return 0;
+ }
+
+ return 0;
+}
+
+static const enum AVPixelFormat pix_fmts_vvenc[] = {
+ //AV_PIX_FMT_YUV420P, // TODO
+ AV_PIX_FMT_YUV420P10LE,
+ AV_PIX_FMT_NONE
+};
+
+#define OFFSET(x) offsetof(VVenCContext, x)
+#define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
+static const AVOption libvvenc_options[] = {
+ {"preset", "set encoding preset(0: faster - 4: slower", OFFSET( options.preset), AV_OPT_TYPE_INT, {.i64 = 2} , 0 , 4 , VE, "preset"},
+ { "faster", "0", 0, AV_OPT_TYPE_CONST, {.i64 = VVENC_FASTER}, INT_MIN, INT_MAX, VE, "preset" },
+ { "fast", "1", 0, AV_OPT_TYPE_CONST, {.i64 = VVENC_FAST}, INT_MIN, INT_MAX, VE, "preset" },
+ { "medium", "2", 0, AV_OPT_TYPE_CONST, {.i64 = VVENC_MEDIUM}, INT_MIN, INT_MAX, VE, "preset" },
+ { "slow", "3", 0, AV_OPT_TYPE_CONST, {.i64 = VVENC_SLOW}, INT_MIN, INT_MAX, VE, "preset" },
+ { "slower", "4", 0, AV_OPT_TYPE_CONST, {.i64 = VVENC_SLOWER}, INT_MIN, INT_MAX, VE, "preset" },
+ { "qp" , "set quantization", OFFSET(options.qp), AV_OPT_TYPE_INT, {.i64 = 32}, 0 , 63 ,VE, "qp_mode" },
+ { "period" , "set (intra) refresh period in seconds", OFFSET(options.intraRefreshSec), AV_OPT_TYPE_INT, {.i64 = 1}, 1 , INT_MAX ,VE,"irefreshsec" },
+ { "subjopt", "set subjective (perceptually motivated) optimization", OFFSET(options.subjectiveOptimization), AV_OPT_TYPE_BOOL, {.i64 = 1}, 0 , 1, VE},
+ { "vvenc-params", "set the vvenc configuration using a :-separated list of key=value parameters", OFFSET(options.vvenc_opts), AV_OPT_TYPE_DICT, { 0 }, 0, 0, VE },
+ { "levelidc", "vvc level_idc", OFFSET( options.levelIdc), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 105, VE, "levelidc"},
+ { "0", "auto", 0, AV_OPT_TYPE_CONST, {.i64 = 0}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "1", "1" , 0, AV_OPT_TYPE_CONST, {.i64 = 16}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "2", "2" , 0, AV_OPT_TYPE_CONST, {.i64 = 32}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "2.1", "2.1" , 0, AV_OPT_TYPE_CONST, {.i64 = 35}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "3", "3" , 0, AV_OPT_TYPE_CONST, {.i64 = 48}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "3.1", "3.1" , 0, AV_OPT_TYPE_CONST, {.i64 = 51}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "4", "4" , 0, AV_OPT_TYPE_CONST, {.i64 = 64}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "4.1", "4.1" , 0, AV_OPT_TYPE_CONST, {.i64 = 67}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "5", "5" , 0, AV_OPT_TYPE_CONST, {.i64 = 80}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "5.1", "5.1" , 0, AV_OPT_TYPE_CONST, {.i64 = 83}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "5.2", "5.2" , 0, AV_OPT_TYPE_CONST, {.i64 = 86}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "6", "6" , 0, AV_OPT_TYPE_CONST, {.i64 = 96}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "6.1", "6.1" , 0, AV_OPT_TYPE_CONST, {.i64 = 99}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "6.2", "6.2" , 0, AV_OPT_TYPE_CONST, {.i64 = 102}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "6.3", "6.3" , 0, AV_OPT_TYPE_CONST, {.i64 = 105}, INT_MIN, INT_MAX, VE, "levelidc"},
+ { "tier", "set vvc tier", OFFSET( options.tier), AV_OPT_TYPE_INT, {.i64 = 0}, 0 , 1 , VE, "tier"},
+ { "main", "main", 0, AV_OPT_TYPE_CONST, {.i64 = 0}, INT_MIN, INT_MAX, VE, "tier"},
+ { "high", "high", 0, AV_OPT_TYPE_CONST, {.i64 = 1}, INT_MIN, INT_MAX, VE, "tier"},
+ {NULL}
+};
+
+static const AVClass class_libvvenc = {
+ .class_name = "libvvenc-vvc encoder",
+ .item_name = av_default_item_name,
+ .option = libvvenc_options,
+ .version = LIBAVUTIL_VERSION_INT,
+};
+
+FFCodec ff_libvvenc_encoder = {
+ .p.name = "libvvenc",
+ CODEC_LONG_NAME("H.266 / VVC Encoder VVenC"),
+ .p.type = AVMEDIA_TYPE_VIDEO,
+ .p.id = AV_CODEC_ID_VVC,
+ .p.capabilities = AV_CODEC_CAP_DELAY | AV_CODEC_CAP_OTHER_THREADS,
+ .p.profiles = NULL_IF_CONFIG_SMALL(ff_vvc_profiles),
+ .p.priv_class = &class_libvvenc,
+ .p.wrapper_name = "libvvenc",
+ .priv_data_size = sizeof(VVenCContext),
+ .p.pix_fmts = pix_fmts_vvenc,
+ .init = ff_vvenc_encode_init,
+ FF_CODEC_ENCODE_CB(ff_vvenc_encode_frame),
+ .close = ff_vvenc_encode_close,
+ .caps_internal = FF_CODEC_CAP_AUTO_THREADS,
+};
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 10/11] avformat: add ts stream types for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (8 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 09/11] avcodec: add external encoder libvvenc " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 11/11] avcodec: increase minor version " Thomas Siedel
2023-06-29 17:38 ` [FFmpeg-devel] [PATCH v7 00/11] Add support " James Almer
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
Add transport stream stream type 0x33 for vvc.
Add STREAM_TYPE_VIDEO_VVC to MPEG-1/2 and MPEG-2 transport stream.
Add basic transport stream support for TS mux/demux.
---
configure | 2 +-
libavformat/mpeg.c | 3 ++
libavformat/mpeg.h | 1 +
libavformat/mpegts.c | 2 ++
libavformat/mpegts.h | 1 +
libavformat/mpegtsenc.c | 65 +++++++++++++++++++++++++++++++++++++++++
6 files changed, 73 insertions(+), 1 deletion(-)
diff --git a/configure b/configure
index 0444422b18..b0161e7682 100755
--- a/configure
+++ b/configure
@@ -3472,7 +3472,7 @@ mp3_demuxer_select="mpegaudio_parser"
mp3_muxer_select="mpegaudioheader"
mp4_muxer_select="mov_muxer"
mpegts_demuxer_select="iso_media"
-mpegts_muxer_select="ac3_parser adts_muxer latm_muxer h264_mp4toannexb_bsf hevc_mp4toannexb_bsf"
+mpegts_muxer_select="ac3_parser adts_muxer latm_muxer h264_mp4toannexb_bsf hevc_mp4toannexb_bsf h266_mp4toannexb_bsf"
mpegtsraw_demuxer_select="mpegts_demuxer"
mxf_muxer_select="pcm_rechunk_bsf"
mxf_d10_muxer_select="mxf_muxer"
diff --git a/libavformat/mpeg.c b/libavformat/mpeg.c
index 864b08d8f8..970926df6b 100644
--- a/libavformat/mpeg.c
+++ b/libavformat/mpeg.c
@@ -546,6 +546,9 @@ redo:
} else if (es_type == STREAM_TYPE_VIDEO_HEVC) {
codec_id = AV_CODEC_ID_HEVC;
type = AVMEDIA_TYPE_VIDEO;
+ } else if (es_type == STREAM_TYPE_VIDEO_VVC) {
+ codec_id = AV_CODEC_ID_VVC;
+ type = AVMEDIA_TYPE_VIDEO;
} else if (es_type == STREAM_TYPE_AUDIO_AC3) {
codec_id = AV_CODEC_ID_AC3;
type = AVMEDIA_TYPE_AUDIO;
diff --git a/libavformat/mpeg.h b/libavformat/mpeg.h
index b635295776..20592eb184 100644
--- a/libavformat/mpeg.h
+++ b/libavformat/mpeg.h
@@ -56,6 +56,7 @@
#define STREAM_TYPE_VIDEO_MPEG4 0x10
#define STREAM_TYPE_VIDEO_H264 0x1b
#define STREAM_TYPE_VIDEO_HEVC 0x24
+#define STREAM_TYPE_VIDEO_VVC 0x33
#define STREAM_TYPE_VIDEO_CAVS 0x42
#define STREAM_TYPE_AUDIO_AC3 0x81
diff --git a/libavformat/mpegts.c b/libavformat/mpegts.c
index d97702fcd7..61a800c85f 100644
--- a/libavformat/mpegts.c
+++ b/libavformat/mpegts.c
@@ -811,6 +811,7 @@ static const StreamType ISO_types[] = {
{ 0x20, AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_H264 },
{ 0x21, AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_JPEG2000 },
{ 0x24, AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_HEVC },
+ { 0x33, AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_VVC },
{ 0x42, AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_CAVS },
{ 0xd1, AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_DIRAC },
{ 0xd2, AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_AVS2 },
@@ -865,6 +866,7 @@ static const StreamType REGD_types[] = {
{ MKTAG('D', 'T', 'S', '3'), AVMEDIA_TYPE_AUDIO, AV_CODEC_ID_DTS },
{ MKTAG('E', 'A', 'C', '3'), AVMEDIA_TYPE_AUDIO, AV_CODEC_ID_EAC3 },
{ MKTAG('H', 'E', 'V', 'C'), AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_HEVC },
+ { MKTAG('V', 'V', 'C', ' '), AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_VVC },
{ MKTAG('K', 'L', 'V', 'A'), AVMEDIA_TYPE_DATA, AV_CODEC_ID_SMPTE_KLV },
{ MKTAG('I', 'D', '3', ' '), AVMEDIA_TYPE_DATA, AV_CODEC_ID_TIMED_ID3 },
{ MKTAG('V', 'C', '-', '1'), AVMEDIA_TYPE_VIDEO, AV_CODEC_ID_VC1 },
diff --git a/libavformat/mpegts.h b/libavformat/mpegts.h
index a48f14e768..14ae312c50 100644
--- a/libavformat/mpegts.h
+++ b/libavformat/mpegts.h
@@ -128,6 +128,7 @@
#define STREAM_TYPE_METADATA 0x15
#define STREAM_TYPE_VIDEO_H264 0x1b
#define STREAM_TYPE_VIDEO_HEVC 0x24
+#define STREAM_TYPE_VIDEO_VVC 0x33
#define STREAM_TYPE_VIDEO_CAVS 0x42
#define STREAM_TYPE_VIDEO_AVS2 0xd2
#define STREAM_TYPE_VIDEO_AVS3 0xd4
diff --git a/libavformat/mpegtsenc.c b/libavformat/mpegtsenc.c
index 3f201cce1d..8a3995ca0c 100644
--- a/libavformat/mpegtsenc.c
+++ b/libavformat/mpegtsenc.c
@@ -368,6 +368,9 @@ static int get_dvb_stream_type(AVFormatContext *s, AVStream *st)
case AV_CODEC_ID_HEVC:
stream_type = STREAM_TYPE_VIDEO_HEVC;
break;
+ case AV_CODEC_ID_VVC:
+ stream_type = STREAM_TYPE_VIDEO_VVC;
+ break;
case AV_CODEC_ID_CAVS:
stream_type = STREAM_TYPE_VIDEO_CAVS;
break;
@@ -464,6 +467,11 @@ static int get_m2ts_stream_type(AVFormatContext *s, AVStream *st)
case AV_CODEC_ID_HEVC:
stream_type = STREAM_TYPE_VIDEO_HEVC;
break;
+ case AV_CODEC_ID_VVC:
+ av_log(s, AV_LOG_ERROR,
+ "MPEGTS VVC %s.\n", avcodec_get_name(st->codecpar->codec_id));
+ stream_type = STREAM_TYPE_VIDEO_VVC;
+ break;
case AV_CODEC_ID_PCM_BLURAY:
stream_type = 0x80;
break;
@@ -1786,6 +1794,21 @@ static int check_hevc_startcode(AVFormatContext *s, const AVStream *st, const AV
return 0;
}
+static int check_vvc_startcode(AVFormatContext *s, const AVStream *st, const AVPacket *pkt)
+{
+ if (pkt->size < 5 || AV_RB32(pkt->data) != 0x0000001 && AV_RB24(pkt->data) != 0x000001) {
+ if (!st->nb_frames) {
+ av_log(s, AV_LOG_ERROR, "VVC bitstream malformed, no startcode found\n");
+ return AVERROR_PATCHWELCOME;
+ }
+ av_log(s, AV_LOG_WARNING, "VVC bitstream error, startcode missing, size %d", pkt->size);
+ if (pkt->size)
+ av_log(s, AV_LOG_WARNING, " data %08"PRIX32, AV_RB32(pkt->data));
+ av_log(s, AV_LOG_WARNING, "\n");
+ }
+ return 0;
+}
+
/* Based on GStreamer's gst-plugins-base/ext/ogg/gstoggstream.c
* Released under the LGPL v2.1+, written by
* Vincent Penquerc'h <vincent.penquerch@collabora.co.uk>
@@ -2010,6 +2033,42 @@ static int mpegts_write_packet_internal(AVFormatContext *s, AVPacket *pkt)
buf = data;
size = pkt->size + 7 + extradd;
}
+ } else if (st->codecpar->codec_id == AV_CODEC_ID_VVC) {
+ const uint8_t *p = buf, *buf_end = p + size;
+ uint32_t state = -1;
+ uint32_t naltype = -1;
+ int extradd = (pkt->flags & AV_PKT_FLAG_KEY) ? st->codecpar->extradata_size : 0;
+ int ret = check_vvc_startcode(s, st, pkt);
+ if (ret < 0)
+ return ret;
+
+ if (extradd && AV_RB24(st->codecpar->extradata) > 1)
+ extradd = 0;
+
+ do {
+ p = avpriv_find_start_code(p, buf_end, &state);
+ // state contains byte behind start code, p points 2 bytes behind start code
+ naltype = ((AV_RB8(p)>>3) & 0x1F);
+ av_log(s, AV_LOG_TRACE, "nal %"PRId32"\n", naltype );
+ if (naltype == 14 ) // VPS
+ extradd = 0;
+ } while (p < buf_end && naltype != 20 && naltype >= 12);
+
+ if (naltype >= 12)
+ extradd = 0;
+ if (naltype != 20) { // AUD NAL
+ data = av_malloc(pkt->size + 7 + extradd);
+ if (!data)
+ return AVERROR(ENOMEM);
+ memcpy(data + 7, st->codecpar->extradata, extradd);
+ memcpy(data + 7 + extradd, pkt->data, pkt->size);
+ AV_WB32(data, 0x00000001);
+ data[4] = 20;
+ data[5] = 1;
+ data[6] = 0x50; // any slice type (0x4) + rbsp stop one bit
+ buf = data;
+ size = pkt->size + 7 + extradd;
+ }
} else if (st->codecpar->codec_id == AV_CODEC_ID_OPUS) {
if (pkt->size < 2) {
av_log(s, AV_LOG_ERROR, "Opus packet too short\n");
@@ -2266,6 +2325,12 @@ static int mpegts_check_bitstream(AVFormatContext *s, AVStream *st,
(st->codecpar->extradata_size > 0 &&
st->codecpar->extradata[0] == 1)))
ret = ff_stream_add_bitstream_filter(st, "hevc_mp4toannexb", NULL);
+ } else if (st->codecpar->codec_id == AV_CODEC_ID_VVC) {
+ if (pkt->size >= 5 && AV_RB32(pkt->data) != 0x0000001 &&
+ (AV_RB24(pkt->data) != 0x000001 ||
+ (st->codecpar->extradata_size > 0 &&
+ st->codecpar->extradata[0] == 1)))
+ ret = ff_stream_add_bitstream_filter(st, "h266_mp4toannexb", NULL);
}
return ret;
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* [FFmpeg-devel] [PATCH v7 11/11] avcodec: increase minor version for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (9 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 10/11] avformat: add ts stream types " Thomas Siedel
@ 2023-03-21 15:01 ` Thomas Siedel
2023-06-29 17:38 ` [FFmpeg-devel] [PATCH v7 00/11] Add support " James Almer
11 siblings, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-03-21 15:01 UTC (permalink / raw)
To: ffmpeg-devel
Increase avcodec minor version for vvc support.
---
libavcodec/version.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/version.h b/libavcodec/version.h
index 39dbec0208..ecb096f38b 100644
--- a/libavcodec/version.h
+++ b/libavcodec/version.h
@@ -29,7 +29,7 @@
#include "version_major.h"
-#define LIBAVCODEC_VERSION_MINOR 6
+#define LIBAVCODEC_VERSION_MINOR 7
#define LIBAVCODEC_VERSION_MICRO 101
#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC
2023-03-21 15:01 [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC Thomas Siedel
` (10 preceding siblings ...)
2023-03-21 15:01 ` [FFmpeg-devel] [PATCH v7 11/11] avcodec: increase minor version " Thomas Siedel
@ 2023-06-29 17:38 ` James Almer
2023-07-01 11:24 ` Nuo Mi
2023-07-03 15:22 ` Thomas Siedel
11 siblings, 2 replies; 19+ messages in thread
From: James Almer @ 2023-06-29 17:38 UTC (permalink / raw)
To: ffmpeg-devel
On 3/21/2023 12:01 PM, Thomas Siedel wrote:
> This patch set adds H266/VVC support.
> This includes parsing, muxing, demuxing, decoding and encoding.
> Decoding is done using the external library VVdeC
> (https://github.com/fraunhoferhhi/vvdec.git) and can be enabled with
> --enable-libvvdec.
> Encoding is done using the external library VVenC
> (https://github.com/fraunhoferhhi/vvenc.git) and can be enabled with
> --enable-libvvenc.
>
> For conformance testing, the following JVET bitstreams can be used:
> https://www.itu.int/wftp3/av-arch/jvet-site/bitstream_exchange/VVC/draft_conformance/draft6/
> DVB test streams can be found here:
> https://dvb.org/specifications/verification-validation/vvc-test-content/
> All JVET conformance bitstreams that are supported by VVdeC can be decoded.
> Pallete mode and multi-layer streams are unsupported.
> The DVB MP4 and TS reference files can be decoded as well.
I have pushed patches 1 to 6, and the raw muxer part of 7. Will look at
the ISOBMFF and ts stuff better later, as well as the encoder wrapper.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC
2023-06-29 17:38 ` [FFmpeg-devel] [PATCH v7 00/11] Add support " James Almer
@ 2023-07-01 11:24 ` Nuo Mi
2023-07-03 15:22 ` Thomas Siedel
1 sibling, 0 replies; 19+ messages in thread
From: Nuo Mi @ 2023-07-01 11:24 UTC (permalink / raw)
To: FFmpeg development discussions and patches
On Fri, Jun 30, 2023 at 1:39 AM James Almer <jamrial@gmail.com> wrote:
> On 3/21/2023 12:01 PM, Thomas Siedel wrote:
> > This patch set adds H266/VVC support.
> > This includes parsing, muxing, demuxing, decoding and encoding.
> > Decoding is done using the external library VVdeC
> > (https://github.com/fraunhoferhhi/vvdec.git) and can be enabled with
> > --enable-libvvdec.
> > Encoding is done using the external library VVenC
> > (https://github.com/fraunhoferhhi/vvenc.git) and can be enabled with
> > --enable-libvvenc.
> >
> > For conformance testing, the following JVET bitstreams can be used:
> >
> https://www.itu.int/wftp3/av-arch/jvet-site/bitstream_exchange/VVC/draft_conformance/draft6/
> > DVB test streams can be found here:
> > https://dvb.org/specifications/verification-validation/vvc-test-content/
> > All JVET conformance bitstreams that are supported by VVdeC can be
> decoded.
> > Pallete mode and multi-layer streams are unsupported.
> > The DVB MP4 and TS reference files can be decoded as well.
>
> I have pushed patches 1 to 6, and the raw muxer part of 7. Will look at
> the ISOBMFF and ts stuff better later, as well as the encoder wrapper.
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>
Great! Really appreciate it.
I'll send the native decoder v2 very soon.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [FFmpeg-devel] [PATCH v7 00/11] Add support for H266/VVC
2023-06-29 17:38 ` [FFmpeg-devel] [PATCH v7 00/11] Add support " James Almer
2023-07-01 11:24 ` Nuo Mi
@ 2023-07-03 15:22 ` Thomas Siedel
1 sibling, 0 replies; 19+ messages in thread
From: Thomas Siedel @ 2023-07-03 15:22 UTC (permalink / raw)
To: FFmpeg development discussions and patches
On Thu, 29 Jun 2023 at 19:39, James Almer <jamrial@gmail.com> wrote:
> On 3/21/2023 12:01 PM, Thomas Siedel wrote:
> > This patch set adds H266/VVC support.
> > This includes parsing, muxing, demuxing, decoding and encoding.
> > Decoding is done using the external library VVdeC
> > (https://github.com/fraunhoferhhi/vvdec.git) and can be enabled with
> > --enable-libvvdec.
> > Encoding is done using the external library VVenC
> > (https://github.com/fraunhoferhhi/vvenc.git) and can be enabled with
> > --enable-libvvenc.
> >
> > For conformance testing, the following JVET bitstreams can be used:
> >
> https://www.itu.int/wftp3/av-arch/jvet-site/bitstream_exchange/VVC/draft_conformance/draft6/
> > DVB test streams can be found here:
> > https://dvb.org/specifications/verification-validation/vvc-test-content/
> > All JVET conformance bitstreams that are supported by VVdeC can be
> decoded.
> > Pallete mode and multi-layer streams are unsupported.
> > The DVB MP4 and TS reference files can be decoded as well.
>
> I have pushed patches 1 to 6, and the raw muxer part of 7. Will look at
> the ISOBMFF and ts stuff better later, as well as the encoder wrapper.
>
That is great, thank you for pushing the patches.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 19+ messages in thread