* Re: [FFmpeg-devel] [FFmpeg-devel v10 7/9] Extended MOV muxer to handle EVC video content
2022-08-25 13:35 ` [FFmpeg-devel] [FFmpeg-devel v10 7/9] Extended MOV muxer to handle EVC video content Dawid Kozinski
@ 2022-08-25 19:57 ` Andreas Rheinhardt
0 siblings, 0 replies; 2+ messages in thread
From: Andreas Rheinhardt @ 2022-08-25 19:57 UTC (permalink / raw)
To: ffmpeg-devel
Dawid Kozinski:
> - Changes in mov_write_video_tag function to handle EVC elementary stream
> - Provided structure EVCDecoderConfigurationRecord that specifies the decoder configuration information for ISO/IEC 23094-1 video content
>
> Signed-off-by: Dawid Kozinski <d.kozinski@samsung.com>
> ---
> libavformat/Makefile | 2 +-
> libavformat/evc.c | 460 ++++++++++++++++++++++++++++++++++++++++
> libavformat/evc.h | 147 +++++++++++++
> libavformat/isom_tags.c | 2 +
> libavformat/movenc.c | 35 ++-
> 5 files changed, 644 insertions(+), 2 deletions(-)
> create mode 100644 libavformat/evc.c
> create mode 100644 libavformat/evc.h
>
> diff --git a/libavformat/Makefile b/libavformat/Makefile
> index a320cac22b..bb4e96e460 100644
> --- a/libavformat/Makefile
> +++ b/libavformat/Makefile
> @@ -360,7 +360,7 @@ OBJS-$(CONFIG_MOV_DEMUXER) += mov.o mov_chan.o mov_esds.o \
> OBJS-$(CONFIG_MOV_MUXER) += movenc.o av1.o avc.o hevc.o vpcc.o \
> movenchint.o mov_chan.o rtp.o \
> movenccenc.o movenc_ttml.o rawutils.o \
> - dovi_isom.o
> + dovi_isom.o evc.o
> OBJS-$(CONFIG_MP2_MUXER) += rawenc.o
> OBJS-$(CONFIG_MP3_DEMUXER) += mp3dec.o replaygain.o
> OBJS-$(CONFIG_MP3_MUXER) += mp3enc.o rawenc.o id3v2enc.o
> diff --git a/libavformat/evc.c b/libavformat/evc.c
> new file mode 100644
> index 0000000000..153fe6daeb
> --- /dev/null
> +++ b/libavformat/evc.c
> @@ -0,0 +1,460 @@
> +/*
> + * EVC helper functions for muxers
> + * Copyright (c) 2022 Dawid Kozinski <d.kozinski@samsung.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 "libavutil/intreadwrite.h"
> +#include "libavcodec/get_bits.h"
> +#include "libavcodec/golomb.h"
> +#include "avformat.h"
> +#include "avio.h"
> +#include "evc.h"
> +#include "avio_internal.h"
> +
> +// The length field that indicates the length in bytes of the following NAL unit is configured to be of 4 bytes
> +#define EVC_NAL_UNIT_LENGTH_BYTE (4) /* byte */
> +#define EVC_NAL_HEADER_SIZE (2) /* byte */
> +
> +// rpl structure
> +typedef struct RefPicListStruct {
> + int poc;
> + int tid;
> + int ref_pic_num;
> + int ref_pic_active_num;
> + int ref_pics[EVC_MAX_NUM_REF_PICS];
> + char pic_type;
> +
> +} RefPicListStruct;
> +
> +// The sturcture reflects SPS RBSP(raw byte sequence payload) layout
> +// @see ISO_IEC_23094-1 section 7.3.2.1
> +//
> +// The following descriptors specify the parsing process of each element
> +// u(n) - unsigned integer using n bits
> +// ue(v) - unsigned integer 0-th order Exp_Golomb-coded syntax element with the left bit first
> +typedef struct EVCSPS {
> + int sps_seq_parameter_set_id; // ue(v)
> + int profile_idc; // u(8)
> + int level_idc; // u(8)
> + int toolset_idc_h; // u(32)
> + int toolset_idc_l; // u(32)
> + int chroma_format_idc; // ue(v)
> + int pic_width_in_luma_samples; // ue(v)
> + int pic_height_in_luma_samples; // ue(v)
> + int bit_depth_luma_minus8; // ue(v)
> + int bit_depth_chroma_minus8; // ue(v)
> +
> + // @note
> + // Currently the structure does not reflect the entire SPS RBSP layout.
> + // It contains only the fields that are necessary to read from the NAL unit all the values
> + // necessary for the correct initialization of EVCDecoderConfigurationRecord
> +
> + // @note
> + // If necessary, add the missing fields to the structure to reflect
> + // the contents of the entire NAL unit of the SPS type
> +
> +} EVCSPS;
> +
> +typedef struct EVCNALUnitArray {
> + uint8_t array_completeness;
> + uint8_t NAL_unit_type;
> + uint16_t numNalus;
> + uint16_t *nalUnitLength;
> + uint8_t **nalUnit;
> +} EVCNALUnitArray;
> +
> +/**
> + * @brief Specifies the decoder configuration information for ISO/IEC 23094-1 video content.
> + * @see ISO/IEC 14496-15:2021 Coding of audio-visual objects - Part 15:
> + * Carriage of network abstraction layer (NAL) unit structured video in the ISO base media file format
> + */
> +typedef struct EVCDecoderConfigurationRecord {
> + uint8_t configurationVersion; // 8 bits
> + uint8_t profile_idc; // 8 bits
> + uint8_t level_idc; // 8 bits
> + uint32_t toolset_idc_h; // 32 bits
> + uint32_t toolset_idc_l; // 32 bits
> + uint8_t chroma_format_idc; // 2 bits
> + uint8_t bit_depth_luma_minus8; // 3 bits
> + uint8_t bit_depth_chroma_minus8; // 3 bits
> + uint16_t pic_width_in_luma_samples; // 16 bits
> + uint16_t pic_height_in_luma_samples; // 16 bits
> + uint8_t reserved; // 6 bits '000000'b
> + uint8_t lengthSizeMinusOne; // 2 bits
> + uint8_t num_of_arrays; // 8 bits
> + EVCNALUnitArray *array;
> +} EVCDecoderConfigurationRecord;
> +
> +typedef struct NALU {
> + int offset;
> + uint32_t size;
> +} NALU;
> +
> +typedef struct NALUList {
> + NALU *nalus;
> + unsigned nalus_array_size;
> + unsigned nb_nalus; ///< valid entries in nalus
> +} NALUList;
> +
> +static int get_nalu_type(const uint8_t *bits, int bits_size)
> +{
> + int unit_type_plus1 = 0;
> +
> + if (bits_size >= EVC_NAL_HEADER_SIZE) {
> + unsigned char *p = (unsigned char *)bits;
> + // forbidden_zero_bit
> + if ((p[0] & 0x80) != 0)
> + return -1;
> +
> + // nal_unit_type
> + unit_type_plus1 = (p[0] >> 1) & 0x3F;
> + }
> +
> + return unit_type_plus1 - 1;
> +}
> +
> +static uint32_t read_nal_unit_length(const uint8_t *bits, int bits_size)
> +{
> + uint32_t nalu_len = 0;
> +
> + if (bits_size >= EVC_NAL_UNIT_LENGTH_BYTE) {
> +
> + int t = 0;
> + unsigned char *p = (unsigned char *)bits;
Don't cast const away! And don't add superfluous variables.
> +
> + for (int i = 0; i < EVC_NAL_UNIT_LENGTH_BYTE; i++)
> + t = (t << 8) | p[i];
> +
> + nalu_len = t;
> + if (nalu_len == 0)
> + return 0;
> + }
> +
> + return nalu_len;
> +}
> +
> +// @see ISO_IEC_23094-1 (7.3.2.1 SPS RBSP syntax)
> +static int evcc_parse_sps(const uint8_t *bs, int bs_size, EVCDecoderConfigurationRecord *evcc)
> +{
> + GetBitContext gb;
> + int sps_seq_parameter_set_id;
> + EVCSPS sps;
> +
> + if (init_get_bits8(&gb, bs, bs_size) < 0)
> + return 0;
> +
> + sps.sps_seq_parameter_set_id = get_ue_golomb(&gb);
> +
> + if (sps_seq_parameter_set_id >= EVC_MAX_SPS_COUNT)
> + return 0;
> +
> + // the Baseline profile is indicated by profile_idc eqal to 0
> + // the Main profile is indicated by profile_idc eqal to 1
> + sps.profile_idc = get_bits(&gb, 8);
> +
> + sps.level_idc = get_bits(&gb, 8);
> +
> + sps.toolset_idc_h = get_bits(&gb, 32);
> + sps.toolset_idc_l = get_bits(&gb, 32);
> +
> + // 0 - monochrome
> + // 1 - 4:2:0
> + // 2 - 4:2:2
> + // 3 - 4:4:4
> + sps.chroma_format_idc = get_ue_golomb(&gb);
> +
> + sps.pic_width_in_luma_samples = get_ue_golomb(&gb);
> + sps.pic_height_in_luma_samples = get_ue_golomb(&gb);
> +
> + sps.bit_depth_luma_minus8 = get_ue_golomb(&gb);
> + sps.bit_depth_chroma_minus8 = get_ue_golomb(&gb);
> +
> + evcc->profile_idc = sps.profile_idc;
> + evcc->level_idc = sps.level_idc;
> + evcc->toolset_idc_h = sps.toolset_idc_h;
> + evcc->toolset_idc_l = sps.toolset_idc_l;
> + evcc->chroma_format_idc = sps.chroma_format_idc;
> + evcc->bit_depth_luma_minus8 = sps.bit_depth_luma_minus8;
> + evcc->bit_depth_chroma_minus8 = sps.bit_depth_chroma_minus8;
> + evcc->pic_width_in_luma_samples = sps.pic_width_in_luma_samples;
> + evcc->pic_height_in_luma_samples = sps.pic_height_in_luma_samples;
> +
> + return 0;
> +}
> +
> +static int evcc_array_add_nal_unit(uint8_t *nal_buf, uint32_t nal_size,
> + uint8_t nal_type, int ps_array_completeness,
> + EVCDecoderConfigurationRecord *evcc)
> +{
This function here (and the whole approach) is basically a clone of the
(horrible) HEVC muxer. We don't like duplicate code.
> + int ret;
> + uint8_t index;
> + uint16_t numNalus;
> + EVCNALUnitArray *array;
> +
> + for (index = 0; index < evcc->num_of_arrays; index++)
> + if (evcc->array[index].NAL_unit_type == nal_type)
> + break;
> +
> + if (index >= evcc->num_of_arrays) {
> + uint8_t i;
> +
> + ret = av_reallocp_array(&evcc->array, index + 1, sizeof(EVCNALUnitArray));
> + if (ret < 0)
> + return ret;
> +
> + for (i = evcc->num_of_arrays; i <= index; i++)
> + memset(&evcc->array[i], 0, sizeof(EVCNALUnitArray));
> + evcc->num_of_arrays = index + 1;
> + }
> +
> + array = &evcc->array[index];
> + numNalus = array->numNalus;
> +
> + ret = av_reallocp_array(&array->nalUnit, numNalus + 1, sizeof(uint8_t *));
> + if (ret < 0)
> + return ret;
> +
> + ret = av_reallocp_array(&array->nalUnitLength, numNalus + 1, sizeof(uint16_t));
> + if (ret < 0)
> + return ret;
> +
> + array->nalUnit [numNalus] = nal_buf;
> + array->nalUnitLength[numNalus] = nal_size;
> + array->NAL_unit_type = nal_type;
> + array->numNalus++;
> +
> + /*
> + * When the sample entry name is 'evc1', the default and mandatory value of
> + * array_completeness is 1 for arrays of all types of parameter sets, and 0
> + * for all other arrays.
> + */
> + if (nal_type == EVC_SPS_NUT || nal_type == EVC_PPS_NUT)
> + array->array_completeness = ps_array_completeness;
> +
> + return 0;
> +}
> +
> +static void evcc_init(EVCDecoderConfigurationRecord *evcc)
> +{
> + memset(evcc, 0, sizeof(EVCDecoderConfigurationRecord));
> + evcc->configurationVersion = 1;
> + evcc->lengthSizeMinusOne = 3; // 4 bytes
> +}
> +
> +static void evcc_close(EVCDecoderConfigurationRecord *evcc)
> +{
> + uint8_t i;
> +
> + for (i = 0; i < evcc->num_of_arrays; i++) {
> + evcc->array[i].numNalus = 0;
> + av_freep(&evcc->array[i].nalUnit);
> + av_freep(&evcc->array[i].nalUnitLength);
> + }
> +
> + evcc->num_of_arrays = 0;
> + av_freep(&evcc->array);
> +}
> +
> +static int evcc_write(AVIOContext *pb, EVCDecoderConfigurationRecord *evcc)
> +{
> + uint8_t i;
> + uint16_t j, aps_count = 0, sps_count = 0, pps_count = 0;
> +
> + av_log(NULL, AV_LOG_TRACE, "configurationVersion: %"PRIu8"\n",
> + evcc->configurationVersion);
> + av_log(NULL, AV_LOG_TRACE, "profile_idc: %"PRIu8"\n",
> + evcc->profile_idc);
> + av_log(NULL, AV_LOG_TRACE, "level_idc: %"PRIu8"\n",
> + evcc->level_idc);
> + av_log(NULL, AV_LOG_TRACE, "toolset_idc_h: %"PRIu32"\n",
> + evcc->toolset_idc_h);
> + av_log(NULL, AV_LOG_TRACE, "toolset_idc_l: %"PRIu32"\n",
> + evcc->toolset_idc_l);
> + av_log(NULL, AV_LOG_TRACE, "chroma_format_idc: %"PRIu8"\n",
> + evcc->chroma_format_idc);
> + av_log(NULL, AV_LOG_TRACE, "bit_depth_luma_minus8: %"PRIu8"\n",
> + evcc->bit_depth_luma_minus8);
> + av_log(NULL, AV_LOG_TRACE, "bit_depth_chroma_minus8: %"PRIu8"\n",
> + evcc->bit_depth_chroma_minus8);
> + av_log(NULL, AV_LOG_TRACE, "pic_width_in_luma_samples: %"PRIu16"\n",
> + evcc->pic_width_in_luma_samples);
> + av_log(NULL, AV_LOG_TRACE, "pic_height_in_luma_samples: %"PRIu16"\n",
> + evcc->pic_height_in_luma_samples);
> + av_log(NULL, AV_LOG_TRACE, "lengthSizeMinusOne: %"PRIu8"\n",
> + evcc->lengthSizeMinusOne);
> + av_log(NULL, AV_LOG_TRACE, "num_of_arrays: %"PRIu8"\n",
> + evcc->num_of_arrays);
> + for (i = 0; i < evcc->num_of_arrays; i++) {
> + av_log(NULL, AV_LOG_TRACE, "array_completeness[%"PRIu8"]: %"PRIu8"\n",
> + i, evcc->array[i].array_completeness);
> + av_log(NULL, AV_LOG_TRACE, "NAL_unit_type[%"PRIu8"]: %"PRIu8"\n",
> + i, evcc->array[i].NAL_unit_type);
> + av_log(NULL, AV_LOG_TRACE, "numNalus[%"PRIu8"]: %"PRIu16"\n",
> + i, evcc->array[i].numNalus);
> + for (j = 0; j < evcc->array[i].numNalus; j++)
> + av_log(NULL, AV_LOG_TRACE,
> + "nalUnitLength[%"PRIu8"][%"PRIu16"]: %"PRIu16"\n",
> + i, j, evcc->array[i].nalUnitLength[j]);
> + }
> +
> + /*
> + * We need at least one SPS.
> + */
> + for (i = 0; i < evcc->num_of_arrays; i++)
> + switch (evcc->array[i].NAL_unit_type) {
> + case EVC_APS_NUT:
> + aps_count += evcc->array[i].numNalus;
> + break;
> + case EVC_SPS_NUT:
> + sps_count += evcc->array[i].numNalus;
> + break;
> + case EVC_PPS_NUT:
> + pps_count += evcc->array[i].numNalus;
> + break;
> + default:
> + break;
> + }
> + if (!sps_count || sps_count > EVC_MAX_SPS_COUNT)
> + return AVERROR_INVALIDDATA;
> +
> + /* unsigned int(8) configurationVersion = 1; */
> + avio_w8(pb, evcc->configurationVersion);
> +
> + /* unsigned int(8) profile_idc */
> + avio_w8(pb, evcc->profile_idc);
> +
> + /* unsigned int(8) profile_idc */
> + avio_w8(pb, evcc->level_idc);
> +
> + /* unsigned int(32) toolset_idc_h */
> + avio_wb32(pb, evcc->toolset_idc_h);
> +
> + /* unsigned int(32) toolset_idc_l */
> + avio_wb32(pb, evcc->toolset_idc_l);
> +
> + /*
> + * unsigned int(2) chroma_format_idc;
> + * unsigned int(3) bit_depth_luma_minus8;
> + * unsigned int(3) bit_depth_chroma_minus8;
> + */
> + avio_w8(pb, evcc->chroma_format_idc << 6 |
> + evcc->bit_depth_luma_minus8 << 3 |
> + evcc->bit_depth_chroma_minus8);
> +
> + /* unsigned int(16) pic_width_in_luma_samples; */
> + avio_wb16(pb, evcc->pic_width_in_luma_samples);
> +
> + /* unsigned int(16) pic_width_in_luma_samples; */
> + avio_wb16(pb, evcc->pic_height_in_luma_samples);
> +
> + /*
> + * bit(6) reserved = '111111'b;
> + * unsigned int(2) chromaFormat;
> + */
> + avio_w8(pb, evcc->lengthSizeMinusOne | 0xfc);
> +
> + /* unsigned int(8) numOfArrays; */
> + avio_w8(pb, evcc->num_of_arrays);
> +
> + for (i = 0; i < evcc->num_of_arrays; i++) {
> + /*
> + * bit(1) array_completeness;
> + * unsigned int(1) reserved = 0;
> + * unsigned int(6) NAL_unit_type;
> + */
> + avio_w8(pb, evcc->array[i].array_completeness << 7 |
> + evcc->array[i].NAL_unit_type & 0x3f);
> +
> + /* unsigned int(16) numNalus; */
> + avio_wb16(pb, evcc->array[i].numNalus);
> +
> + for (j = 0; j < evcc->array[i].numNalus; j++) {
> + /* unsigned int(16) nalUnitLength; */
> + avio_wb16(pb, evcc->array[i].nalUnitLength[j]);
> +
> + /* bit(8*nalUnitLength) nalUnit; */
> + avio_write(pb, evcc->array[i].nalUnit[j],
> + evcc->array[i].nalUnitLength[j]);
> + }
> + }
> +
> + return 0;
> +}
> +
> +int ff_isom_write_evcc(AVIOContext *pb, const uint8_t *data,
> + int size, int ps_array_completeness)
> +{
> + EVCDecoderConfigurationRecord evcc;
> + int nalu_type;
> + size_t nalu_size;
> + unsigned char *bits = (unsigned char *)data;
Don't cast const away. In fact, I don't see a reason for a separate
variable at all.
> + int bytes_to_read = size;
> +
> + int ret = 0;
> +
> + if (size < 8) {
> + /* We can't write a valid evcC from the provided data */
> + return AVERROR_INVALIDDATA;
> + } else if (*data == 1) {
> + /* Data is already evcC-formatted */
> + avio_write(pb, data, size);
> + return 0;
> + }
> +
> + evcc_init(&evcc);
> +
> + while (bytes_to_read > EVC_NAL_UNIT_LENGTH_BYTE) {
> + uint8_t *nalu_buf = NULL;
> + nalu_size = read_nal_unit_length(bits, EVC_NAL_UNIT_LENGTH_BYTE);
> + if (nalu_size == 0) break;
> +
> + bits += EVC_NAL_UNIT_LENGTH_BYTE;
> + bytes_to_read -= EVC_NAL_UNIT_LENGTH_BYTE;
> +
> + if (bytes_to_read < nalu_size) break;
> +
> + nalu_type = get_nalu_type(bits, bytes_to_read);
> + nalu_buf = bits;
> +
> + switch (nalu_type) {
> + case EVC_APS_NUT:
> + case EVC_SPS_NUT:
> + case EVC_PPS_NUT:
> + ret = evcc_array_add_nal_unit(nalu_buf, nalu_size, nalu_type, ps_array_completeness, &evcc);
> + if (ret < 0)
> + goto end;
> + else if (nalu_type == EVC_SPS_NUT)
> + ret = evcc_parse_sps(nalu_buf, nalu_size, &evcc);
> + if (ret < 0)
> + goto end;
> + break;
> + default:
> + break;
> + }
> +
> + bits += nalu_size;
> + bytes_to_read -= nalu_size;
> + }
> +
> + ret = evcc_write(pb, &evcc);
> +
> +end:
> + evcc_close(&evcc);
> + return ret;
> +}
> diff --git a/libavformat/evc.h b/libavformat/evc.h
> new file mode 100644
> index 0000000000..dfe038cc80
> --- /dev/null
> +++ b/libavformat/evc.h
> @@ -0,0 +1,147 @@
> +/*
> + * EVC helper functions for muxers
> + * Copyright (c) 2022 Dawid Kozinski <d.kozinski@samsung.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
> + */
> +
> +#ifndef AVFORMAT_EVC_H
> +#define AVFORMAT_EVC_H
> +
> +#include <stdint.h>
> +#include "libavutil/rational.h"
> +#include "avio.h"
> +
> +/**
> + * @see ISO_IEC_23094-1_2020, 7.4.2.2 NAL unit header semantic
> + * Table 4 - NAL unit type codes and NAL unit type classes
> + */
> +enum EVCNALUnitType {
> + EVC_NOIDR_NUT = 0,
> + EVC_IDR_NUT = 1,
> + EVC_RSV_VCL_NUT02 = 2,
> + EVC_RSV_VCL_NUT03 = 3,
> + EVC_RSV_VCL_NUT04 = 4,
> + EVC_RSV_VCL_NUT05 = 5,
> + EVC_RSV_VCL_NUT06 = 6,
> + EVC_RSV_VCL_NUT07 = 7,
> + EVC_RSV_VCL_NUT08 = 8,
> + EVC_RSV_VCL_NUT09 = 9,
> + EVC_RSV_VCL_NUT10 = 10,
> + EVC_RSV_VCL_NUT11 = 11,
> + EVC_RSV_VCL_NUT12 = 12,
> + EVC_RSV_VCL_NUT13 = 13,
> + EVC_RSV_VCL_NUT14 = 14,
> + EVC_RSV_VCL_NUT15 = 15,
> + EVC_RSV_VCL_NUT16 = 16,
> + EVC_RSV_VCL_NUT17 = 17,
> + EVC_RSV_VCL_NUT18 = 18,
> + EVC_RSV_VCL_NUT19 = 19,
> + EVC_RSV_VCL_NUT20 = 20,
> + EVC_RSV_VCL_NUT21 = 21,
> + EVC_RSV_VCL_NUT22 = 22,
> + EVC_RSV_VCL_NUT23 = 23,
> + EVC_SPS_NUT = 24,
> + EVC_PPS_NUT = 25,
> + EVC_APS_NUT = 26,
> + EVC_FD_NUT = 27,
> + EVC_SEI_NUT = 28,
> + EVC_RSV_NONVCL29 = 29,
> + EVC_RSV_NONVCL30 = 30,
> + EVC_RSV_NONVCL31 = 31,
> + EVC_RSV_NONVCL32 = 32,
> + EVC_RSV_NONVCL33 = 33,
> + EVC_RSV_NONVCL34 = 34,
> + EVC_RSV_NONVCL35 = 35,
> + EVC_RSV_NONVCL36 = 36,
> + EVC_RSV_NONVCL37 = 37,
> + EVC_RSV_NONVCL38 = 38,
> + EVC_RSV_NONVCL39 = 39,
> + EVC_RSV_NONVCL40 = 40,
> + EVC_RSV_NONVCL41 = 41,
> + EVC_RSV_NONVCL42 = 42,
> + EVC_RSV_NONVCL43 = 43,
> + EVC_RSV_NONVCL44 = 44,
> + EVC_RSV_NONVCL45 = 45,
> + EVC_RSV_NONVCL46 = 46,
> + EVC_RSV_NONVCL47 = 47,
> + EVC_RSV_NONVCL48 = 48,
> + EVC_RSV_NONVCL49 = 49,
> + EVC_RSV_NONVCL50 = 50,
> + EVC_RSV_NONVCL51 = 51,
> + EVC_RSV_NONVCL52 = 52,
> + EVC_RSV_NONVCL53 = 53,
> + EVC_RSV_NONVCL54 = 54,
> + EVC_RSV_NONVCL55 = 55,
> + EVC_UNSPEC_NUT56 = 56,
> + EVC_UNSPEC_NUT57 = 57,
> + EVC_UNSPEC_NUT58 = 58,
> + EVC_UNSPEC_NUT59 = 59,
> + EVC_UNSPEC_NUT60 = 60,
> + EVC_UNSPEC_NUT61 = 61,
> + EVC_UNSPEC_NUT62 = 62
These enums are not muxer-dependent; it should be in a separate header
in libavcodec. In fact, you should use this in the parser instead of
adding custom "#define EVC_NUT_SPS (24)" for it.
> +};
> +
> +enum {
> + // 7.4.3.2: aps_video_parameter_set_id is u(4).
> + EVC_MAX_APS_COUNT = 32,
> +
> + // 7.4.3.1: sps_seq_parameter_set_id is in [0, 15].
> + EVC_MAX_SPS_COUNT = 16,
> +
> + // 7.4.3.2: pps_pic_parameter_set_id is in [0, 63].
> + EVC_MAX_PPS_COUNT = 64,
> +
> + // 7.4.5: slice header slice_pic_parameter_set_id in [0, 63]
> + EVC_MAX_SH_COUNT = 64,
> +
> + // E.3.2: cpb_cnt_minus1[i] is in [0, 31].
> + EVC_MAX_CPB_CNT = 32,
> +
> + // A.4.1: in table A.1 the highest level allows a MaxLumaPs of 35 651 584.
> + EVC_MAX_LUMA_PS = 35651584,
> +
> + EVC_MAX_NUM_REF_PICS = 21,
> +
> + // 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.
> + EVC_MAX_WIDTH = 16888,
> + EVC_MAX_HEIGHT = 16888,
> +
> + // A.4.1: table A.1 allows at most 22 tile rows for any level.
> + EVC_MAX_TILE_ROWS = 22,
> + // A.4.1: table A.1 allows at most 20 tile columns for any level.
> + EVC_MAX_TILE_COLUMNS = 20,
> +
> + // A.4.1: table A.1 allows at most 600 slice segments for any level.
> + EVC_MAX_SLICE_SEGMENTS = 600,
> +
> + // 7.4.7.1: in the worst case (tiles_enabled_flag and
> + // entropy_coding_sync_enabled_flag are both set), 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 4K video (at most 135 16x16 Ctb rows).
> + HEVC_MAX_ENTRY_POINT_OFFSETS = EVC_MAX_TILE_COLUMNS * 135,
> +};
> +
> +int ff_isom_write_evcc(AVIOContext *pb, const uint8_t *data,
> + int size, int ps_array_completeness);
> +
> +#endif // AVFORMAT_EVC_H
> diff --git a/libavformat/isom_tags.c b/libavformat/isom_tags.c
> index 362cb77e8f..98f866094b 100644
> --- a/libavformat/isom_tags.c
> +++ b/libavformat/isom_tags.c
> @@ -147,6 +147,8 @@ const AVCodecTag ff_codec_movvideo_tags[] = {
> { AV_CODEC_ID_H264, MKTAG('d', 'v', 'a', '1') }, /* AVC-based Dolby Vision derived from avc1 */
> { AV_CODEC_ID_H264, MKTAG('d', 'v', 'a', 'v') }, /* AVC-based Dolby Vision derived from avc3 */
>
> + { AV_CODEC_ID_EVC, MKTAG('e', 'v', 'c', '1') }, /* EVC/MPEG-5 */
> +
> { AV_CODEC_ID_VP8, MKTAG('v', 'p', '0', '8') }, /* VP8 */
> { AV_CODEC_ID_VP9, MKTAG('v', 'p', '0', '9') }, /* VP9 */
> { AV_CODEC_ID_AV1, MKTAG('a', 'v', '0', '1') }, /* AV1 */
> diff --git a/libavformat/movenc.c b/libavformat/movenc.c
> index c8b2e141cb..93e4f64a63 100644
> --- a/libavformat/movenc.c
> +++ b/libavformat/movenc.c
> @@ -35,6 +35,7 @@
> #include "isom.h"
> #include "av1.h"
> #include "avc.h"
> +#include "evc.h"
> #include "libavcodec/ac3_parser_internal.h"
> #include "libavcodec/dnxhddata.h"
> #include "libavcodec/flac.h"
> @@ -1392,6 +1393,21 @@ static int mov_write_hvcc_tag(AVIOContext *pb, MOVTrack *track)
> return update_size(pb, pos);
> }
>
> +static int mov_write_evcc_tag(AVIOContext *pb, MOVTrack *track)
> +{
> + int64_t pos = avio_tell(pb);
> +
> + avio_wb32(pb, 0);
> + ffio_wfourcc(pb, "evcC");
> +
> + if (track->tag == MKTAG('e','v','c','1'))
> + ff_isom_write_evcc(pb, track->vos_data, track->vos_len, 1);
> + else
> + ff_isom_write_evcc(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)
> {
> @@ -1641,6 +1657,16 @@ static int mov_get_h264_codec_tag(AVFormatContext *s, MOVTrack *track)
> return tag;
> }
>
> +static int mov_get_evc_codec_tag(AVFormatContext *s, MOVTrack *track)
> +{
> + int tag = track->par->codec_tag;
> +
> + if (!tag)
> + tag = MKTAG('e', 'v', 'c', 'i');
> +
> + return tag;
> +}
> +
> static const struct {
> enum AVPixelFormat pix_fmt;
> uint32_t tag;
> @@ -1722,6 +1748,8 @@ static unsigned int mov_get_codec_tag(AVFormatContext *s, MOVTrack *track)
> tag = mov_get_mpeg2_xdcam_codec_tag(s, track);
> else if (track->par->codec_id == AV_CODEC_ID_H264)
> tag = mov_get_h264_codec_tag(s, track);
> + else if (track->par->codec_id == AV_CODEC_ID_EVC)
> + tag = mov_get_evc_codec_tag(s, track);
> else if (track->par->codec_id == AV_CODEC_ID_DNXHD)
> tag = mov_get_dnxhd_codec_tag(s, track);
> else if (track->par->codec_type == AVMEDIA_TYPE_VIDEO) {
> @@ -2280,6 +2308,9 @@ static int mov_write_video_tag(AVFormatContext *s, AVIOContext *pb, MOVMuxContex
> mov_write_avcc_tag(pb, track);
> if (track->mode == MODE_IPOD)
> mov_write_uuid_tag_ipod(pb);
> + }
> + else if (track->par->codec_id ==AV_CODEC_ID_EVC) {
> + mov_write_evcc_tag(pb, track);
> } else if (track->par->codec_id == AV_CODEC_ID_VP9) {
> mov_write_vpcc_tag(mov->fc, pb, track);
> } else if (track->par->codec_id == AV_CODEC_ID_AV1) {
> @@ -6030,7 +6061,8 @@ 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_TRUEHD) && !trk->vos_len &&
> + par->codec_id == AV_CODEC_ID_TRUEHD ||
> + par->codec_id == AV_CODEC_ID_EVC) && !trk->vos_len &&
> !TAG_IS_AVCI(trk->tag)) {
> /* copy frame to create needed atoms */
> trk->vos_len = size;
> @@ -7689,6 +7721,7 @@ 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_EVC, MKTAG('e', 'v', 'c', '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') },
_______________________________________________
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] 2+ messages in thread