* [FFmpeg-devel] [PATCH 1/8] avutil: introduce an Immersive Audio Model and Formats API
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
@ 2023-12-14 20:14 ` James Almer
2023-12-18 11:04 ` Anton Khirnov
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 2/8] avformat: introduce AVStreamGroup James Almer
` (6 subsequent siblings)
7 siblings, 1 reply; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Signed-off-by: James Almer <jamrial@gmail.com>
---
libavutil/Makefile | 2 +
libavutil/iamf.c | 563 ++++++++++++++++++++++++++++++++++++++++
libavutil/iamf.h | 620 +++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 1185 insertions(+)
create mode 100644 libavutil/iamf.c
create mode 100644 libavutil/iamf.h
diff --git a/libavutil/Makefile b/libavutil/Makefile
index 4711f8cde8..62cc1a1831 100644
--- a/libavutil/Makefile
+++ b/libavutil/Makefile
@@ -51,6 +51,7 @@ HEADERS = adler32.h \
hwcontext_videotoolbox.h \
hwcontext_vdpau.h \
hwcontext_vulkan.h \
+ iamf.h \
imgutils.h \
intfloat.h \
intreadwrite.h \
@@ -140,6 +141,7 @@ OBJS = adler32.o \
hdr_dynamic_vivid_metadata.o \
hmac.o \
hwcontext.o \
+ iamf.o \
imgutils.o \
integer.o \
intmath.o \
diff --git a/libavutil/iamf.c b/libavutil/iamf.c
new file mode 100644
index 0000000000..62b6051049
--- /dev/null
+++ b/libavutil/iamf.c
@@ -0,0 +1,563 @@
+/*
+ * Immersive Audio Model and Formats helper functions and defines
+ *
+ * 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 <limits.h>
+#include <stddef.h>
+#include <stdint.h>
+
+#include "avassert.h"
+#include "error.h"
+#include "iamf.h"
+#include "log.h"
+#include "mem.h"
+#include "opt.h"
+
+#define IAMF_ADD_FUNC_TEMPLATE(parent_type, parent_name, child_type, child_name, suffix) \
+child_type *av_iamf_ ## parent_name ## _add_ ## child_name(parent_type *parent_name) \
+{ \
+ child_type **child_name ## suffix, *child_name; \
+ \
+ if (parent_name->nb_## child_name ## suffix == UINT_MAX) \
+ return NULL; \
+ \
+ child_name ## suffix = av_realloc_array(parent_name->child_name ## suffix, \
+ parent_name->nb_## child_name ## suffix + 1, \
+ sizeof(*parent_name->child_name ## suffix)); \
+ if (!child_name ## suffix) \
+ return NULL; \
+ \
+ parent_name->child_name ## suffix = child_name ## suffix; \
+ \
+ child_name = parent_name->child_name ## suffix[parent_name->nb_## child_name ## suffix] \
+ = av_mallocz(sizeof(*child_name)); \
+ if (!child_name) \
+ return NULL; \
+ \
+ child_name->av_class = &child_name ## _class; \
+ av_opt_set_defaults(child_name); \
+ parent_name->nb_## child_name ## suffix++; \
+ \
+ return child_name; \
+}
+
+#define FLAGS AV_OPT_FLAG_ENCODING_PARAM
+
+//
+// Param Definition
+//
+#define OFFSET(x) offsetof(AVIAMFMixGain, x)
+static const AVOption mix_gain_options[] = {
+ { "subblock_duration", "set subblock_duration", OFFSET(subblock_duration), AV_OPT_TYPE_INT64, {.i64 = 1 }, 1, UINT_MAX, FLAGS },
+ { "animation_type", "set animation_type", OFFSET(animation_type), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 2, FLAGS },
+ { "start_point_value", "set start_point_value", OFFSET(animation_type), AV_OPT_TYPE_RATIONAL, {.dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "end_point_value", "set end_point_value", OFFSET(animation_type), AV_OPT_TYPE_RATIONAL, {.dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "control_point_value", "set control_point_value", OFFSET(animation_type), AV_OPT_TYPE_RATIONAL, {.dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "control_point_relative_time", "set control_point_relative_time", OFFSET(animation_type), AV_OPT_TYPE_RATIONAL, {.dbl = 0 }, 0.0, 1.0, FLAGS },
+ { NULL },
+};
+
+static const AVClass mix_gain_class = {
+ .class_name = "AVIAMFSubmixElement",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = mix_gain_options,
+};
+
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFDemixingInfo, x)
+static const AVOption demixing_info_options[] = {
+ { "subblock_duration", "set subblock_duration", OFFSET(subblock_duration), AV_OPT_TYPE_INT64, {.i64 = 1 }, 1, UINT_MAX, FLAGS },
+ { "dmixp_mode", "set dmixp_mode", OFFSET(dmixp_mode), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 6, FLAGS },
+ { NULL },
+};
+
+static const AVClass demixing_info_class = {
+ .class_name = "AVIAMFDemixingInfo",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = demixing_info_options,
+};
+
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFReconGain, x)
+static const AVOption recon_gain_options[] = {
+ { "subblock_duration", "set subblock_duration", OFFSET(subblock_duration), AV_OPT_TYPE_INT64, {.i64 = 1 }, 1, UINT_MAX, FLAGS },
+ { NULL },
+};
+
+static const AVClass recon_gain_class = {
+ .class_name = "AVIAMFReconGain",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = recon_gain_options,
+};
+
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFParamDefinition, x)
+static const AVOption param_definition_options[] = {
+ { "parameter_id", "set parameter_id", OFFSET(parameter_id), AV_OPT_TYPE_INT64, {.i64 = 0 }, 0, UINT_MAX, FLAGS },
+ { "parameter_rate", "set parameter_rate", OFFSET(parameter_rate), AV_OPT_TYPE_INT64, {.i64 = 0 }, 0, UINT_MAX, FLAGS },
+ { "duration", "set duration", OFFSET(duration), AV_OPT_TYPE_INT64, {.i64 = 0 }, 0, UINT_MAX, FLAGS },
+ { "constant_subblock_duration", "set constant_subblock_duration", OFFSET(constant_subblock_duration), AV_OPT_TYPE_INT64, {.i64 = 0 }, 0, UINT_MAX, FLAGS },
+ { NULL },
+};
+
+static const AVClass *param_definition_child_iterate(void **opaque)
+{
+ uintptr_t i = (uintptr_t)*opaque;
+ const AVClass *ret = NULL;
+
+ switch(i) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN:
+ ret = &mix_gain_class;
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING:
+ ret = &demixing_info_class;
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN:
+ ret = &recon_gain_class;
+ break;
+ default:
+ break;
+ }
+
+ if (ret)
+ *opaque = (void*)(i + 1);
+ return ret;
+}
+
+static const AVClass param_definition_class = {
+ .class_name = "AVIAMFParamDefinition",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = param_definition_options,
+ .child_class_iterate = param_definition_child_iterate,
+};
+
+const AVClass *av_iamf_param_definition_get_class(void)
+{
+ return ¶m_definition_class;
+}
+
+AVIAMFParamDefinition *av_iamf_param_definition_alloc(enum AVIAMFParamDefinitionType type,
+ unsigned int nb_subblocks, size_t *out_size)
+{
+
+ struct MixGainStruct {
+ AVIAMFParamDefinition p;
+ AVIAMFMixGain m;
+ };
+ struct DemixStruct {
+ AVIAMFParamDefinition p;
+ AVIAMFDemixingInfo d;
+ };
+ struct ReconGainStruct {
+ AVIAMFParamDefinition p;
+ AVIAMFReconGain r;
+ };
+ size_t subblocks_offset, subblock_size;
+ size_t size;
+ AVIAMFParamDefinition *par;
+
+ switch (type) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN:
+ subblocks_offset = offsetof(struct MixGainStruct, m);
+ subblock_size = sizeof(AVIAMFMixGain);
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING:
+ subblocks_offset = offsetof(struct DemixStruct, d);
+ subblock_size = sizeof(AVIAMFDemixingInfo);
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN:
+ subblocks_offset = offsetof(struct ReconGainStruct, r);
+ subblock_size = sizeof(AVIAMFReconGain);
+ break;
+ default:
+ return NULL;
+ }
+
+ size = subblocks_offset;
+ if (nb_subblocks > (SIZE_MAX - size) / subblock_size)
+ return NULL;
+ size += subblock_size * nb_subblocks;
+
+ par = av_mallocz(size);
+ if (!par)
+ return NULL;
+
+ par->av_class = ¶m_definition_class;
+ av_opt_set_defaults(par);
+
+ par->type = type;
+ par->nb_subblocks = nb_subblocks;
+ par->subblock_size = subblock_size;
+ par->subblocks_offset = subblocks_offset;
+
+ for (int i = 0; i < nb_subblocks; i++) {
+ void *subblock = av_iamf_param_definition_get_subblock(par, i);
+
+ switch (type) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN:
+ ((AVIAMFMixGain *)subblock)->av_class = &mix_gain_class;
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING:
+ ((AVIAMFDemixingInfo *)subblock)->av_class = &demixing_info_class;
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN:
+ ((AVIAMFReconGain *)subblock)->av_class = &recon_gain_class;
+ break;
+ default:
+ av_assert0(0);
+ }
+
+ av_opt_set_defaults(subblock);
+ }
+
+ if (out_size)
+ *out_size = size;
+
+ return par;
+}
+
+//
+// Audio Element
+//
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFLayer, x)
+static const AVOption layer_options[] = {
+ { "ch_layout", "set ch_layout", OFFSET(ch_layout), AV_OPT_TYPE_CHLAYOUT, {.str = NULL }, 0, 0, FLAGS },
+ { "flags", "set flags", OFFSET(flags), AV_OPT_TYPE_FLAGS,
+ {.i64 = 0 }, 0, AV_IAMF_LAYER_FLAG_RECON_GAIN, FLAGS, "flags" },
+ {"recon_gain", "Recon gain is present", 0, AV_OPT_TYPE_CONST,
+ {.i64 = AV_IAMF_LAYER_FLAG_RECON_GAIN }, INT_MIN, INT_MAX, FLAGS, "flags"},
+ { "output_gain_flags", "set output_gain_flags", OFFSET(output_gain_flags), AV_OPT_TYPE_FLAGS,
+ {.i64 = 0 }, 0, (1 << 6) - 1, FLAGS, "output_gain_flags" },
+ {"FL", "Left channel", 0, AV_OPT_TYPE_CONST,
+ {.i64 = 1 << 5 }, INT_MIN, INT_MAX, FLAGS, "output_gain_flags"},
+ {"FR", "Right channel", 0, AV_OPT_TYPE_CONST,
+ {.i64 = 1 << 4 }, INT_MIN, INT_MAX, FLAGS, "output_gain_flags"},
+ {"BL", "Left surround channel", 0, AV_OPT_TYPE_CONST,
+ {.i64 = 1 << 3 }, INT_MIN, INT_MAX, FLAGS, "output_gain_flags"},
+ {"BR", "Right surround channel", 0, AV_OPT_TYPE_CONST,
+ {.i64 = 1 << 2 }, INT_MIN, INT_MAX, FLAGS, "output_gain_flags"},
+ {"TFL", "Left top front channel", 0, AV_OPT_TYPE_CONST,
+ {.i64 = 1 << 1 }, INT_MIN, INT_MAX, FLAGS, "output_gain_flags"},
+ {"TFR", "Right top front channel", 0, AV_OPT_TYPE_CONST,
+ {.i64 = 1 << 0 }, INT_MIN, INT_MAX, FLAGS, "output_gain_flags"},
+ { "output_gain", "set output_gain", OFFSET(output_gain), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "ambisonics_mode", "set ambisonics_mode", OFFSET(ambisonics_mode), AV_OPT_TYPE_INT,
+ { .i64 = AV_IAMF_AMBISONICS_MODE_MONO },
+ AV_IAMF_AMBISONICS_MODE_MONO, AV_IAMF_AMBISONICS_MODE_PROJECTION, FLAGS, "ambisonics_mode" },
+ { "mono", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_AMBISONICS_MODE_MONO }, .unit = "ambisonics_mode" },
+ { "projection", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_AMBISONICS_MODE_PROJECTION }, .unit = "ambisonics_mode" },
+ { NULL },
+};
+
+static const AVClass layer_class = {
+ .class_name = "AVIAMFLayer",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = layer_options,
+};
+
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFAudioElement, x)
+static const AVOption audio_element_options[] = {
+ { "audio_element_type", "set audio_element_type", OFFSET(audio_element_type), AV_OPT_TYPE_INT,
+ {.i64 = AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL },
+ AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL, AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE, FLAGS, "audio_element_type" },
+ { "channel", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL }, .unit = "audio_element_type" },
+ { "scene", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE }, .unit = "audio_element_type" },
+ { "default_w", "set default_w", OFFSET(default_w), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 10, FLAGS },
+ { NULL },
+};
+
+static const AVClass *audio_element_child_iterate(void **opaque)
+{
+ uintptr_t i = (uintptr_t)*opaque;
+ const AVClass *ret = NULL;
+
+ if (i)
+ ret = &layer_class;
+
+ if (ret)
+ *opaque = (void*)(i + 1);
+ return ret;
+}
+
+static const AVClass audio_element_class = {
+ .class_name = "AVIAMFAudioElement",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = audio_element_options,
+ .child_class_iterate = audio_element_child_iterate,
+};
+
+const AVClass *av_iamf_audio_element_get_class(void)
+{
+ return &audio_element_class;
+}
+
+AVIAMFAudioElement *av_iamf_audio_element_alloc(void)
+{
+ AVIAMFAudioElement *audio_element = av_mallocz(sizeof(*audio_element));
+
+ if (audio_element) {
+ audio_element->av_class = &audio_element_class;
+ av_opt_set_defaults(audio_element);
+ }
+
+ return audio_element;
+}
+
+IAMF_ADD_FUNC_TEMPLATE(AVIAMFAudioElement, audio_element, AVIAMFLayer, layer, s)
+
+void av_iamf_audio_element_free(AVIAMFAudioElement **paudio_element)
+{
+ AVIAMFAudioElement *audio_element = *paudio_element;
+
+ if (!audio_element)
+ return;
+
+ for (int i = 0; i < audio_element->nb_layers; i++) {
+ AVIAMFLayer *layer = audio_element->layers[i];
+ av_opt_free(layer);
+ av_free(layer->demixing_matrix);
+ av_free(layer);
+ }
+ av_free(audio_element->layers);
+
+ av_free(audio_element->demixing_info);
+ av_free(audio_element->recon_gain_info);
+ av_freep(paudio_element);
+}
+
+//
+// Mix Presentation
+//
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFSubmixElement, x)
+static const AVOption submix_element_options[] = {
+ { "headphones_rendering_mode", "Headphones rendering mode", OFFSET(headphones_rendering_mode), AV_OPT_TYPE_INT,
+ { .i64 = AV_IAMF_HEADPHONES_MODE_STEREO },
+ AV_IAMF_HEADPHONES_MODE_STEREO, AV_IAMF_HEADPHONES_MODE_BINAURAL, FLAGS, "headphones_rendering_mode" },
+ { "stereo", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_HEADPHONES_MODE_STEREO }, .unit = "headphones_rendering_mode" },
+ { "binaural", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_HEADPHONES_MODE_BINAURAL }, .unit = "headphones_rendering_mode" },
+ { "default_mix_gain", "Default mix gain", OFFSET(default_mix_gain), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "annotations", "Annotations", OFFSET(annotations), AV_OPT_TYPE_DICT, { .str = NULL }, 0, 0, FLAGS },
+ { NULL },
+};
+
+static void *submix_element_child_next(void *obj, void *prev)
+{
+ AVIAMFSubmixElement *submix_element = obj;
+ if (!prev)
+ return submix_element->element_mix_config;
+
+ return NULL;
+}
+
+static const AVClass *submix_element_child_iterate(void **opaque)
+{
+ uintptr_t i = (uintptr_t)*opaque;
+ const AVClass *ret = NULL;
+
+ if (i)
+ ret = ¶m_definition_class;
+
+ if (ret)
+ *opaque = (void*)(i + 1);
+ return ret;
+}
+
+static const AVClass element_class = {
+ .class_name = "AVIAMFSubmixElement",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = submix_element_options,
+ .child_next = submix_element_child_next,
+ .child_class_iterate = submix_element_child_iterate,
+};
+
+IAMF_ADD_FUNC_TEMPLATE(AVIAMFSubmix, submix, AVIAMFSubmixElement, element, s)
+
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFSubmixLayout, x)
+static const AVOption submix_layout_options[] = {
+ { "layout_type", "Layout type", OFFSET(layout_type), AV_OPT_TYPE_INT,
+ { .i64 = AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS },
+ AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS, AV_IAMF_SUBMIX_LAYOUT_TYPE_BINAURAL, FLAGS, "layout_type" },
+ { "loudspeakers", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS }, .unit = "layout_type" },
+ { "binaural", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_IAMF_SUBMIX_LAYOUT_TYPE_BINAURAL }, .unit = "layout_type" },
+ { "sound_system", "Sound System", OFFSET(sound_system), AV_OPT_TYPE_CHLAYOUT, { .str = NULL }, 0, 0, FLAGS },
+ { "integrated_loudness", "Integrated loudness", OFFSET(integrated_loudness), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "digital_peak", "Digital peak", OFFSET(digital_peak), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "true_peak", "True peak", OFFSET(true_peak), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "dialog_anchored_loudness", "Anchored loudness (Dialog)", OFFSET(dialogue_anchored_loudness), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { "album_anchored_loudness", "Anchored loudness (Album)", OFFSET(album_anchored_loudness), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { NULL },
+};
+
+static const AVClass layout_class = {
+ .class_name = "AVIAMFSubmixLayout",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = submix_layout_options,
+};
+
+IAMF_ADD_FUNC_TEMPLATE(AVIAMFSubmix, submix, AVIAMFSubmixLayout, layout, s)
+
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFSubmix, x)
+static const AVOption submix_presentation_options[] = {
+ { "default_mix_gain", "Default mix gain", OFFSET(default_mix_gain), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, -128.0, 128.0, FLAGS },
+ { NULL },
+};
+
+static void *submix_presentation_child_next(void *obj, void *prev)
+{
+ AVIAMFSubmix *sub_mix = obj;
+ if (!prev)
+ return sub_mix->output_mix_config;
+
+ return NULL;
+}
+
+static const AVClass *submix_presentation_child_iterate(void **opaque)
+{
+ uintptr_t i = (uintptr_t)*opaque;
+ const AVClass *ret = NULL;
+
+ switch(i) {
+ case 0:
+ ret = &element_class;
+ break;
+ case 1:
+ ret = &layout_class;
+ break;
+ case 2:
+ ret = ¶m_definition_class;
+ break;
+ default:
+ break;
+ }
+
+ if (ret)
+ *opaque = (void*)(i + 1);
+ return ret;
+}
+
+static const AVClass submix_class = {
+ .class_name = "AVIAMFSubmix",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = submix_presentation_options,
+ .child_next = submix_presentation_child_next,
+ .child_class_iterate = submix_presentation_child_iterate,
+};
+
+#undef OFFSET
+#define OFFSET(x) offsetof(AVIAMFMixPresentation, x)
+static const AVOption mix_presentation_options[] = {
+ { "annotations", "set annotations", OFFSET(annotations), AV_OPT_TYPE_DICT, {.str = NULL }, 0, 0, FLAGS },
+ { NULL },
+};
+
+#undef OFFSET
+#undef FLAGS
+
+static const AVClass *mix_presentation_child_iterate(void **opaque)
+{
+ uintptr_t i = (uintptr_t)*opaque;
+ const AVClass *ret = NULL;
+
+ if (i)
+ ret = &submix_class;
+
+ if (ret)
+ *opaque = (void*)(i + 1);
+ return ret;
+}
+
+static const AVClass mix_presentation_class = {
+ .class_name = "AVIAMFMixPresentation",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = mix_presentation_options,
+ .child_class_iterate = mix_presentation_child_iterate,
+};
+
+const AVClass *av_iamf_mix_presentation_get_class(void)
+{
+ return &mix_presentation_class;
+}
+
+AVIAMFMixPresentation *av_iamf_mix_presentation_alloc(void)
+{
+ AVIAMFMixPresentation *mix_presentation = av_mallocz(sizeof(*mix_presentation));
+
+ if (mix_presentation) {
+ mix_presentation->av_class = &mix_presentation_class;
+ av_opt_set_defaults(mix_presentation);
+ }
+
+ return mix_presentation;
+}
+
+IAMF_ADD_FUNC_TEMPLATE(AVIAMFMixPresentation, mix_presentation, AVIAMFSubmix, submix, es)
+
+void av_iamf_mix_presentation_free(AVIAMFMixPresentation **pmix_presentation)
+{
+ AVIAMFMixPresentation *mix_presentation = *pmix_presentation;
+
+ if (!mix_presentation)
+ return;
+
+ for (int i = 0; i < mix_presentation->nb_submixes; i++) {
+ AVIAMFSubmix *sub_mix = mix_presentation->submixes[i];
+ for (int j = 0; j < sub_mix->nb_elements; j++) {
+ AVIAMFSubmixElement *submix_element = sub_mix->elements[j];
+ av_opt_free(submix_element);
+ av_free(submix_element->element_mix_config);
+ av_free(submix_element);
+ }
+ av_free(sub_mix->elements);
+ for (int j = 0; j < sub_mix->nb_layouts; j++) {
+ AVIAMFSubmixLayout *submix_layout = sub_mix->layouts[j];
+ av_opt_free(submix_layout);
+ av_free(submix_layout);
+ }
+ av_free(sub_mix->layouts);
+ av_free(sub_mix->output_mix_config);
+ av_free(sub_mix);
+ }
+ av_opt_free(mix_presentation);
+ av_free(mix_presentation->submixes);
+
+ av_freep(pmix_presentation);
+}
diff --git a/libavutil/iamf.h b/libavutil/iamf.h
new file mode 100644
index 0000000000..7038b71a27
--- /dev/null
+++ b/libavutil/iamf.h
@@ -0,0 +1,620 @@
+/*
+ * Immersive Audio Model and Formats helper functions and defines
+ *
+ * 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 AVUTIL_IAMF_H
+#define AVUTIL_IAMF_H
+
+/**
+ * @file
+ * Immersive Audio Model and Formats API header
+ * @see <a href="https://aomediacodec.github.io/iamf/">Immersive Audio Model and Formats</a>
+ */
+
+#include <stdint.h>
+#include <stddef.h>
+
+#include "attributes.h"
+#include "avassert.h"
+#include "channel_layout.h"
+#include "dict.h"
+#include "rational.h"
+
+/**
+ * @defgroup lavf_iamf_params Parameter Definition
+ * @{
+ * Parameters as defined in section 3.6.1 and 3.8 of IAMF.
+ * @}
+ * @defgroup lavf_iamf_audio Audio Element
+ * @{
+ * Audio Elements as defined in section 3.6 of IAMF.
+ * @}
+ * @defgroup lavf_iamf_mix Mix Presentation
+ * @{
+ * Mix Presentations as defined in section 3.7 of IAMF.
+ * @}
+ *
+ * @}
+ * @addtogroup lavf_iamf_params
+ * @{
+ */
+enum AVIAMFAnimationType {
+ AV_IAMF_ANIMATION_TYPE_STEP,
+ AV_IAMF_ANIMATION_TYPE_LINEAR,
+ AV_IAMF_ANIMATION_TYPE_BEZIER,
+};
+
+/**
+ * Mix Gain Parameter Data as defined in section 3.8.1 of IAMF.
+ */
+typedef struct AVIAMFMixGain {
+ const AVClass *av_class;
+
+ /**
+ * Duration for the given subblock. It must not be 0.
+ */
+ unsigned int subblock_duration;
+ /**
+ * The type of animation applied to the parameter values.
+ */
+ enum AVIAMFAnimationType animation_type;
+ /**
+ * Parameter value that is applied at the start of the subblock.
+ * Applies to all defined Animation Types.
+ *
+ * Valid range of values is -128.0 to 128.0
+ */
+ AVRational start_point_value;
+ /**
+ * Parameter value that is applied at the end of the subblock.
+ * Applies only to AV_IAMF_ANIMATION_TYPE_LINEAR and
+ * AV_IAMF_ANIMATION_TYPE_BEZIER Animation Types.
+ *
+ * Valid range of values is -128.0 to 128.0
+ */
+ AVRational end_point_value;
+ /**
+ * Parameter value of the middle control point of a quadratic Bezier
+ * curve, i.e., its y-axis value.
+ * Applies only to AV_IAMF_ANIMATION_TYPE_BEZIER Animation Type.
+ *
+ * Valid range of values is -128.0 to 128.0
+ */
+ AVRational control_point_value;
+ /**
+ * Parameter value of the time of the middle control point of a
+ * quadratic Bezier curve, i.e., its x-axis value.
+ * Applies only to AV_IAMF_ANIMATION_TYPE_BEZIER Animation Type.
+ *
+ * Valid range of values is 0.0 to 1.0
+ */
+ AVRational control_point_relative_time;
+} AVIAMFMixGain;
+
+/**
+ * Demixing Info Parameter Data as defined in section 3.8.2 of IAMF.
+ */
+typedef struct AVIAMFDemixingInfo {
+ const AVClass *av_class;
+
+ /**
+ * Duration for the given subblock. It must not be 0.
+ */
+ unsigned int subblock_duration;
+ /**
+ * Pre-defined combination of demixing parameters.
+ */
+ unsigned int dmixp_mode;
+} AVIAMFDemixingInfo;
+
+/**
+ * Recon Gain Info Parameter Data as defined in section 3.8.3 of IAMF.
+ */
+typedef struct AVIAMFReconGain {
+ const AVClass *av_class;
+
+ /**
+ * Duration for the given subblock. It must not be 0.
+ */
+ unsigned int subblock_duration;
+
+ /**
+ * Array of gain values to be applied to each channel for each layer
+ * defined in the Audio Element referencing the parent Parameter Definition.
+ * Values for layers where the AV_IAMF_LAYER_FLAG_RECON_GAIN flag is not set
+ * are undefined.
+ *
+ * Channel order is: FL, C, FR, SL, SR, TFL, TFR, BL, BR, TBL, TBR, LFE
+ */
+ uint8_t recon_gain[6][12];
+} AVIAMFReconGain;
+
+enum AVIAMFParamDefinitionType {
+ /**
+ * Subblocks are of struct type AVIAMFMixGain
+ */
+ AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN,
+ /**
+ * Subblocks are of struct type AVIAMFDemixingInfo
+ */
+ AV_IAMF_PARAMETER_DEFINITION_DEMIXING,
+ /**
+ * Subblocks are of struct type AVIAMFReconGain
+ */
+ AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN,
+};
+
+/**
+ * Parameters as defined in section 3.6.1 of IAMF.
+ *
+ * The struct is allocated by av_iamf_param_definition_alloc() along with an
+ * array of subblocks, its type depending on the value of type.
+ * This array is placed subblocks_offset bytes after the start of this struct.
+ */
+typedef struct AVIAMFParamDefinition {
+ const AVClass *av_class;
+
+ /**
+ * Offset in bytes from the start of this struct, at which the subblocks
+ * array is located.
+ */
+ size_t subblocks_offset;
+ /**
+ * Size in bytes of each element in the subblocks array.
+ */
+ size_t subblock_size;
+ /**
+ * Number of subblocks in the array.
+ *
+ * Must be 0 if @ref constant_subblock_duration is not 0.
+ */
+ unsigned int nb_subblocks;
+
+ /**
+ * Parameters type. Determines the type of the subblock elements.
+ */
+ enum AVIAMFParamDefinitionType type;
+
+ /**
+ * Identifier for the paremeter substream.
+ */
+ unsigned int parameter_id;
+ /**
+ * Sample rate for the paremeter substream. It must not be 0.
+ */
+ unsigned int parameter_rate;
+
+ /**
+ * The duration of the all subblocks in this parameter definition.
+ *
+ * May be 0, in which case all duration values should be specified in
+ * another parameter definition referencing the same parameter_id.
+ */
+ unsigned int duration;
+ /**
+ * The duration of every subblock in the case where all subblocks, with
+ * the optional exception of the last subblock, have equal durations.
+ *
+ * Must be 0 if subblocks have different durations.
+ */
+ unsigned int constant_subblock_duration;
+} AVIAMFParamDefinition;
+
+const AVClass *av_iamf_param_definition_get_class(void);
+
+/**
+ * Allocates memory for AVIAMFParamDefinition, plus an array of {@code nb_subblocks}
+ * amount of subblocks of the given type and initializes the variables. Can be
+ * freed with a normal av_free() call.
+ *
+ * @param size if non-NULL, the size in bytes of the resulting data array is written here.
+ */
+AVIAMFParamDefinition *av_iamf_param_definition_alloc(enum AVIAMFParamDefinitionType type,
+ unsigned int nb_subblocks, size_t *size);
+
+/**
+ * Get the subblock at the specified {@code idx}. Must be between 0 and nb_subblocks - 1.
+ *
+ * The @ref AVIAMFParamDefinition.type "param definition type" defines
+ * the struct type of the returned pointer.
+ */
+static av_always_inline void*
+av_iamf_param_definition_get_subblock(const AVIAMFParamDefinition *par, unsigned int idx)
+{
+ av_assert0(idx < par->nb_subblocks);
+ return (void *)((uint8_t *)par + par->subblocks_offset + idx * par->subblock_size);
+}
+
+/**
+ * @}
+ * @addtogroup lavf_iamf_audio
+ * @{
+ */
+
+enum AVIAMFAmbisonicsMode {
+ AV_IAMF_AMBISONICS_MODE_MONO,
+ AV_IAMF_AMBISONICS_MODE_PROJECTION,
+};
+
+/**
+ * Recon gain information for the layer is present in AVIAMFReconGain
+ */
+#define AV_IAMF_LAYER_FLAG_RECON_GAIN (1 << 0)
+
+/**
+ * A layer defining a Channel Layout in the Audio Element.
+ *
+ * When @ref AVIAMFAudioElement.audio_element_type "the parent's Audio Element type"
+ * is AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL, this corresponds to an Scalable Channel
+ * Layout layer as defined in section 3.6.2 of IAMF.
+ * For AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE, it is an Ambisonics channel
+ * layout as defined in section 3.6.3 of IAMF.
+ */
+typedef struct AVIAMFLayer {
+ const AVClass *av_class;
+
+ AVChannelLayout ch_layout;
+
+ /**
+ * A bitmask which may contain a combination of AV_IAMF_LAYER_FLAG_* flags.
+ */
+ unsigned int flags;
+ /**
+ * Output gain channel flags as defined in section 3.6.2 of IAMF.
+ *
+ * This field is defined only if @ref AVIAMFAudioElement.audio_element_type
+ * "the parent's Audio Element type" is AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL,
+ * must be 0 otherwise.
+ */
+ unsigned int output_gain_flags;
+ /**
+ * Output gain as defined in section 3.6.2 of IAMF.
+ *
+ * Must be 0 if @ref output_gain_flags is 0.
+ */
+ AVRational output_gain;
+ /**
+ * Ambisonics mode as defined in section 3.6.3 of IAMF.
+ *
+ * This field is defined only if @ref AVIAMFAudioElement.audio_element_type
+ * "the parent's Audio Element type" is AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE.
+ *
+ * If AV_IAMF_AMBISONICS_MODE_MONO, channel_mapping is defined implicitly
+ * (Ambisonic Order) or explicitly (Custom Order with ambi channels) in
+ * @ref ch_layout.
+ * If AV_IAMF_AMBISONICS_MODE_PROJECTION, @ref demixing_matrix must be set.
+ */
+ enum AVIAMFAmbisonicsMode ambisonics_mode;
+
+ /**
+ * Demixing matrix as defined in section 3.6.3 of IAMF.
+ *
+ * The length of the array is ch_layout.nb_channels multiplied by the sum of
+ * the amount of streams in the group plus the amount of streams in the group
+ * that are stereo.
+ *
+ * May be set only if @ref ambisonics_mode == AV_IAMF_AMBISONICS_MODE_PROJECTION,
+ * must be NULL otherwise.
+ */
+ AVRational *demixing_matrix;
+} AVIAMFLayer;
+
+
+enum AVIAMFAudioElementType {
+ AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL,
+ AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE,
+};
+
+typedef struct AVIAMFAudioElement {
+ const AVClass *av_class;
+
+ AVIAMFLayer **layers;
+ /**
+ * Number of layers, or channel groups, in the Audio Element.
+ * There may be 6 layers at most, and for @ref audio_element_type
+ * AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE, there may be exactly 1.
+ *
+ * Set by av_iamf_audio_element_add_layer(), must not be
+ * modified by any other code.
+ */
+ unsigned int nb_layers;
+
+ /**
+ * Demixing information used to reconstruct a scalable channel audio
+ * representation.
+ * The @ref AVIAMFParamDefinition.type "type" must be
+ * AV_IAMF_PARAMETER_DEFINITION_DEMIXING.
+ */
+ AVIAMFParamDefinition *demixing_info;
+ /**
+ * Recon gain information used to reconstruct a scalable channel audio
+ * representation.
+ * The @ref AVIAMFParamDefinition.type "type" must be
+ * AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN.
+ */
+ AVIAMFParamDefinition *recon_gain_info;
+
+ /**
+ * Audio element type as defined in section 3.6 of IAMF.
+ */
+ enum AVIAMFAudioElementType audio_element_type;
+
+ /**
+ * Default weight value as defined in section 3.6 of IAMF.
+ */
+ unsigned int default_w;
+} AVIAMFAudioElement;
+
+const AVClass *av_iamf_audio_element_get_class(void);
+
+/**
+ * Allocates a AVIAMFAudioElement, and initializes its fields with default values.
+ * No layers are allocated. Must be freed with av_iamf_audio_element_free().
+ *
+ * @see av_iamf_audio_element_add_layer()
+ */
+AVIAMFAudioElement *av_iamf_audio_element_alloc(void);
+
+/**
+ * Allocate a layer and add it to a given AVIAMFAudioElement.
+ * It is freed by av_iamf_audio_element_free() alongside the rest of the parent
+ * AVIAMFAudioElement.
+ *
+ * @return a pointer to the allocated layer.
+ */
+AVIAMFLayer *av_iamf_audio_element_add_layer(AVIAMFAudioElement *audio_element);
+
+void av_iamf_audio_element_free(AVIAMFAudioElement **audio_element);
+
+/**
+ * @}
+ * @addtogroup lavf_iamf_mix
+ * @{
+ */
+
+enum AVIAMFHeadphonesMode {
+ /**
+ * The referenced Audio Element shall be rendered to stereo loudspeakers.
+ */
+ AV_IAMF_HEADPHONES_MODE_STEREO,
+ /**
+ * The referenced Audio Element shall be rendered with a binaural renderer.
+ */
+ AV_IAMF_HEADPHONES_MODE_BINAURAL,
+};
+
+typedef struct AVIAMFSubmixElement {
+ const AVClass *av_class;
+
+ /**
+ * The id of the Audio Element this submix element references.
+ */
+ unsigned int audio_element_id;
+
+ /**
+ * Information required required for applying any processing to the
+ * referenced and rendered Audio Element before being summed with other
+ * processed Audio Elements.
+ * The @ref AVIAMFParamDefinition.type "type" must be
+ * AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN.
+ */
+ AVIAMFParamDefinition *element_mix_config;
+
+ /**
+ * Default mix gain value to apply when there are no AVIAMFParamDefinition
+ * with @ref element_mix_config "element_mix_config's"
+ * @ref AVIAMFParamDefinition.parameter_id "parameter_id" available for a
+ * given audio frame.
+ */
+ AVRational default_mix_gain;
+
+ /**
+ * A value that indicates whether the referenced channel-based Audio Element
+ * shall be rendered to stereo loudspeakers or spatialized with a binaural
+ * renderer when played back on headphones.
+ * If the Audio Element is not of @ref AVIAMFAudioElement.audio_element_type
+ * "type" AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL, then this field is undefined.
+ */
+ enum AVIAMFHeadphonesMode headphones_rendering_mode;
+
+ /**
+ * A dictionary of strings describing the submix in different languages.
+ * Must have the same amount of entries as
+ * @ref AVIAMFMixPresentation.annotations "the mix's annotations", stored
+ * in the same order, and with the same key strings.
+ *
+ * @ref AVDictionaryEntry.key "key" is a string conforming to BCP-47 that
+ * specifies the language for the string stored in
+ * @ref AVDictionaryEntry.value "value".
+ */
+ AVDictionary *annotations;
+} AVIAMFSubmixElement;
+
+enum AVIAMFSubmixLayoutType {
+ /**
+ * The layout follows the loudspeaker sound system convention of ITU-2051-3.
+ */
+ AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS = 2,
+ /**
+ * The layout is binaural.
+ */
+ AV_IAMF_SUBMIX_LAYOUT_TYPE_BINAURAL = 3,
+};
+
+typedef struct AVIAMFSubmixLayout {
+ const AVClass *av_class;
+
+ enum AVIAMFSubmixLayoutType layout_type;
+
+ /**
+ * Channel layout matching one of Sound Systems A to J of ITU-2051-3, plus
+ * 7.1.2ch and 3.1.2ch
+ * If layout_type is not AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS, this field
+ * is undefined.
+ */
+ AVChannelLayout sound_system;
+ /**
+ * The program integrated loudness information, as defined in
+ * ITU-1770-4.
+ */
+ AVRational integrated_loudness;
+ /**
+ * The digital (sampled) peak value of the audio signal, as defined
+ * in ITU-1770-4.
+ */
+ AVRational digital_peak;
+ /**
+ * The true peak of the audio signal, as defined in ITU-1770-4.
+ */
+ AVRational true_peak;
+ /**
+ * The Dialogue loudness information, as defined in ITU-1770-4.
+ */
+ AVRational dialogue_anchored_loudness;
+ /**
+ * The Album loudness information, as defined in ITU-1770-4.
+ */
+ AVRational album_anchored_loudness;
+} AVIAMFSubmixLayout;
+
+typedef struct AVIAMFSubmix {
+ const AVClass *av_class;
+
+ /**
+ * Array of submix elements.
+ *
+ * Set by av_iamf_submix_add_element(), must not be modified by any
+ * other code.
+ */
+ AVIAMFSubmixElement **elements;
+ /**
+ * Number of elements in the submix.
+ *
+ * Set by av_iamf_submix_add_element(), must not be modified by any
+ * other code.
+ */
+ unsigned int nb_elements;
+
+ /**
+ * Array of submix layouts.
+ *
+ * Set by av_iamf_submix_add_layout(), must not be modified by any
+ * other code.
+ */
+ AVIAMFSubmixLayout **layouts;
+ /**
+ * Number of layouts in the submix.
+ *
+ * Set by av_iamf_submix_add_layout(), must not be modified by any
+ * other code.
+ */
+ unsigned int nb_layouts;
+
+ /**
+ * Information required for post-processing the mixed audio signal to
+ * generate the audio signal for playback.
+ * The @ref AVIAMFParamDefinition.type "type" must be
+ * AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN.
+ */
+ AVIAMFParamDefinition *output_mix_config;
+
+ /**
+ * Default mix gain value to apply when there are no AVIAMFParamDefinition
+ * with @ref output_mix_config "output_mix_config's"
+ * @ref AVIAMFParamDefinition.parameter_id "parameter_id" available for a
+ * given audio frame.
+ */
+ AVRational default_mix_gain;
+} AVIAMFSubmix;
+
+typedef struct AVIAMFMixPresentation {
+ const AVClass *av_class;
+
+ /**
+ * Array of submixes.
+ *
+ * Set by av_iamf_mix_presentation_add_submix(), must not be modified
+ * by any other code.
+ */
+ AVIAMFSubmix **submixes;
+ /**
+ * Number of submixes in the presentation.
+ *
+ * Set by av_iamf_mix_presentation_add_submix(), must not be modified
+ * by any other code.
+ */
+ unsigned int nb_submixes;
+
+ /**
+ * A dictionary of strings describing the mix in different languages.
+ * Must have the same amount of entries as every
+ * @ref AVIAMFSubmixElement.annotations "Submix element annotations",
+ * stored in the same order, and with the same key strings.
+ *
+ * @ref AVDictionaryEntry.key "key" is a string conforming to BCP-47
+ * that specifies the language for the string stored in
+ * @ref AVDictionaryEntry.value "value".
+ */
+ AVDictionary *annotations;
+} AVIAMFMixPresentation;
+
+const AVClass *av_iamf_mix_presentation_get_class(void);
+
+/**
+ * Allocates a AVIAMFMixPresentation, and initializes its fields with default
+ * values. No submixes are allocated.
+ * Must be freed with av_iamf_mix_presentation_free().
+ *
+ * @see av_iamf_mix_presentation_add_submix()
+ */
+AVIAMFMixPresentation *av_iamf_mix_presentation_alloc(void);
+
+/**
+ * Allocate a submix and add it to a given AVIAMFMixPresentation.
+ * It is freed by av_iamf_mix_presentation_free() alongside the rest of the
+ * parent AVIAMFMixPresentation.
+ *
+ * @return a pointer to the allocated submix.
+ */
+AVIAMFSubmix *av_iamf_mix_presentation_add_submix(AVIAMFMixPresentation *mix_presentation);
+
+/**
+ * Allocate a submix element and add it to a given AVIAMFSubmix.
+ * It is freed by av_iamf_mix_presentation_free() alongside the rest of the
+ * parent AVIAMFSubmix.
+ *
+ * @return a pointer to the allocated submix.
+ */
+AVIAMFSubmixElement *av_iamf_submix_add_element(AVIAMFSubmix *submix);
+
+/**
+ * Allocate a submix layout and add it to a given AVIAMFSubmix.
+ * It is freed by av_iamf_mix_presentation_free() alongside the rest of the
+ * parent AVIAMFSubmix.
+ *
+ * @return a pointer to the allocated submix.
+ */
+AVIAMFSubmixLayout *av_iamf_submix_add_layout(AVIAMFSubmix *submix);
+
+void av_iamf_mix_presentation_free(AVIAMFMixPresentation **mix_presentation);
+/**
+ * @}
+ */
+
+#endif /* AVUTIL_IAMF_H */
--
2.43.0
_______________________________________________
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] 12+ messages in thread* Re: [FFmpeg-devel] [PATCH 1/8] avutil: introduce an Immersive Audio Model and Formats API
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 1/8] avutil: introduce an Immersive Audio Model and Formats API James Almer
@ 2023-12-18 11:04 ` Anton Khirnov
2023-12-18 18:10 ` James Almer
0 siblings, 1 reply; 12+ messages in thread
From: Anton Khirnov @ 2023-12-18 11:04 UTC (permalink / raw)
To: FFmpeg development discussions and patches
Quoting James Almer (2023-12-14 21:14:26)
> +/**
> + * Mix Gain Parameter Data as defined in section 3.8.1 of IAMF.
> + */
> +typedef struct AVIAMFMixGain {
> + const AVClass *av_class;
> +
> + /**
> + * Duration for the given subblock. It must not be 0.
In what units? Same for all durations in this patch.
> +typedef struct AVIAMFParamDefinition {
> + const AVClass *av_class;
> +
> + /**
> + * Offset in bytes from the start of this struct, at which the subblocks
> + * array is located.
> + */
> + size_t subblocks_offset;
> + /**
> + * Size in bytes of each element in the subblocks array.
> + */
> + size_t subblock_size;
> + /**
> + * Number of subblocks in the array.
> + *
> + * Must be 0 if @ref constant_subblock_duration is not 0.
> + */
> + unsigned int nb_subblocks;
> +
> + /**
> + * Parameters type. Determines the type of the subblock elements.
> + */
> + enum AVIAMFParamDefinitionType type;
> +
> + /**
> + * Identifier for the paremeter substream.
> + */
> + unsigned int parameter_id;
> + /**
> + * Sample rate for the paremeter substream. It must not be 0.
> + */
> + unsigned int parameter_rate;
> +
> + /**
> + * The duration of the all subblocks in this parameter definition.
> + *
> + * May be 0, in which case all duration values should be specified in
> + * another parameter definition referencing the same parameter_id.
> + */
> + unsigned int duration;
> + /**
> + * The duration of every subblock in the case where all subblocks, with
> + * the optional exception of the last subblock, have equal durations.
> + *
> + * Must be 0 if subblocks have different durations.
> + */
> + unsigned int constant_subblock_duration;
This also seems like should be a flags field.
Otherwise looks good.
--
Anton Khirnov
_______________________________________________
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] 12+ messages in thread* Re: [FFmpeg-devel] [PATCH 1/8] avutil: introduce an Immersive Audio Model and Formats API
2023-12-18 11:04 ` Anton Khirnov
@ 2023-12-18 18:10 ` James Almer
0 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-18 18:10 UTC (permalink / raw)
To: ffmpeg-devel
On 12/18/2023 8:04 AM, Anton Khirnov wrote:
> Quoting James Almer (2023-12-14 21:14:26)
>> +/**
>> + * Mix Gain Parameter Data as defined in section 3.8.1 of IAMF.
>> + */
>> +typedef struct AVIAMFMixGain {
>> + const AVClass *av_class;
>> +
>> + /**
>> + * Duration for the given subblock. It must not be 0.
>
> In what units? Same for all durations in this patch.
parameter_rate. Amended.
>
>> +typedef struct AVIAMFParamDefinition {
>> + const AVClass *av_class;
>> +
>> + /**
>> + * Offset in bytes from the start of this struct, at which the subblocks
>> + * array is located.
>> + */
>> + size_t subblocks_offset;
>> + /**
>> + * Size in bytes of each element in the subblocks array.
>> + */
>> + size_t subblock_size;
>> + /**
>> + * Number of subblocks in the array.
>> + *
>> + * Must be 0 if @ref constant_subblock_duration is not 0.
Removed this line as it's bogus.
>> + */
>> + unsigned int nb_subblocks;
>> +
>> + /**
>> + * Parameters type. Determines the type of the subblock elements.
>> + */
>> + enum AVIAMFParamDefinitionType type;
>> +
>> + /**
>> + * Identifier for the paremeter substream.
>> + */
>> + unsigned int parameter_id;
>> + /**
>> + * Sample rate for the paremeter substream. It must not be 0.
>> + */
>> + unsigned int parameter_rate;
>> +
>> + /**
>> + * The duration of the all subblocks in this parameter definition.
>> + *
>> + * May be 0, in which case all duration values should be specified in
>> + * another parameter definition referencing the same parameter_id.
>> + */
>> + unsigned int duration;
>> + /**
>> + * The duration of every subblock in the case where all subblocks, with
>> + * the optional exception of the last subblock, have equal durations.
>> + *
>> + * Must be 0 if subblocks have different durations.
>> + */
>> + unsigned int constant_subblock_duration;
>
> This also seems like should be a flags field.
No, duration and subblock duration are not the same thing. The former is
the accumulated duration of all subblocks in a given parameter
definition. subblock durations can be smaller, and only if they are
constant will constant_subblock_duration be set to a value other than 0.
>
> Otherwise looks good.
>
Pushed. Thanks a lot for looking at it.
_______________________________________________
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] 12+ messages in thread
* [FFmpeg-devel] [PATCH 2/8] avformat: introduce AVStreamGroup
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 1/8] avutil: introduce an Immersive Audio Model and Formats API James Almer
@ 2023-12-14 20:14 ` James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 3/8] ffmpeg: add support for muxing AVStreamGroups James Almer
` (5 subsequent siblings)
7 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Signed-off-by: James Almer <jamrial@gmail.com>
---
doc/fftools-common-opts.texi | 17 +++-
libavformat/avformat.c | 91 +++++++++++++++++++--
libavformat/avformat.h | 153 +++++++++++++++++++++++++++++++++++
libavformat/dump.c | 147 +++++++++++++++++++++++++++------
libavformat/internal.h | 33 ++++++++
libavformat/options.c | 139 +++++++++++++++++++++++++++++++
6 files changed, 546 insertions(+), 34 deletions(-)
diff --git a/doc/fftools-common-opts.texi b/doc/fftools-common-opts.texi
index d9145704d6..f459bfdc1d 100644
--- a/doc/fftools-common-opts.texi
+++ b/doc/fftools-common-opts.texi
@@ -37,9 +37,9 @@ Matches the stream with this index. E.g. @code{-threads:1 4} would set the
thread count for the second stream to 4. If @var{stream_index} is used as an
additional stream specifier (see below), then it selects stream number
@var{stream_index} from the matching streams. Stream numbering is based on the
-order of the streams as detected by libavformat except when a program ID is
-also specified. In this case it is based on the ordering of the streams in the
-program.
+order of the streams as detected by libavformat except when a stream group
+specifier or program ID is also specified. In this case it is based on the
+ordering of the streams in the group or program.
@item @var{stream_type}[:@var{additional_stream_specifier}]
@var{stream_type} is one of following: 'v' or 'V' for video, 'a' for audio, 's'
for subtitle, 'd' for data, and 't' for attachments. 'v' matches all video
@@ -48,6 +48,17 @@ thumbnails or cover arts. If @var{additional_stream_specifier} is used, then
it matches streams which both have this type and match the
@var{additional_stream_specifier}. Otherwise, it matches all streams of the
specified type.
+@item g:@var{group_specifier}[:@var{additional_stream_specifier}]
+Matches streams which are in the group with the specifier @var{group_specifier}.
+if @var{additional_stream_specifier} is used, then it matches streams which both
+are part of the group and match the @var{additional_stream_specifier}.
+@var{group_specifier} may be one of the following:
+@table @option
+@item @var{group_index}
+Match the stream with this group index.
+@item #@var{group_id} or i:@var{group_id}
+Match the stream with this group id.
+@end table
@item p:@var{program_id}[:@var{additional_stream_specifier}]
Matches streams which are in the program with the id @var{program_id}. If
@var{additional_stream_specifier} is used, then it matches streams which both
diff --git a/libavformat/avformat.c b/libavformat/avformat.c
index 5b8bb7879e..7e747c43d5 100644
--- a/libavformat/avformat.c
+++ b/libavformat/avformat.c
@@ -24,6 +24,7 @@
#include "libavutil/avstring.h"
#include "libavutil/channel_layout.h"
#include "libavutil/frame.h"
+#include "libavutil/iamf.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/mem.h"
#include "libavutil/opt.h"
@@ -80,6 +81,32 @@ FF_ENABLE_DEPRECATION_WARNINGS
av_freep(pst);
}
+void ff_free_stream_group(AVStreamGroup **pstg)
+{
+ AVStreamGroup *stg = *pstg;
+
+ if (!stg)
+ return;
+
+ av_freep(&stg->streams);
+ av_dict_free(&stg->metadata);
+ av_freep(&stg->priv_data);
+ switch (stg->type) {
+ case AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT: {
+ av_iamf_audio_element_free(&stg->params.iamf_audio_element);
+ break;
+ }
+ case AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION: {
+ av_iamf_mix_presentation_free(&stg->params.iamf_mix_presentation);
+ break;
+ }
+ default:
+ break;
+ }
+
+ av_freep(pstg);
+}
+
void ff_remove_stream(AVFormatContext *s, AVStream *st)
{
av_assert0(s->nb_streams>0);
@@ -88,6 +115,14 @@ void ff_remove_stream(AVFormatContext *s, AVStream *st)
ff_free_stream(&s->streams[ --s->nb_streams ]);
}
+void ff_remove_stream_group(AVFormatContext *s, AVStreamGroup *stg)
+{
+ av_assert0(s->nb_stream_groups > 0);
+ av_assert0(s->stream_groups[ s->nb_stream_groups - 1 ] == stg);
+
+ ff_free_stream_group(&s->stream_groups[ --s->nb_stream_groups ]);
+}
+
/* XXX: suppress the packet queue */
void ff_flush_packet_queue(AVFormatContext *s)
{
@@ -118,6 +153,9 @@ void avformat_free_context(AVFormatContext *s)
for (unsigned i = 0; i < s->nb_streams; i++)
ff_free_stream(&s->streams[i]);
+ for (unsigned i = 0; i < s->nb_stream_groups; i++)
+ ff_free_stream_group(&s->stream_groups[i]);
+ s->nb_stream_groups = 0;
s->nb_streams = 0;
for (unsigned i = 0; i < s->nb_programs; i++) {
@@ -139,6 +177,7 @@ void avformat_free_context(AVFormatContext *s)
av_packet_free(&si->pkt);
av_packet_free(&si->parse_pkt);
av_freep(&s->streams);
+ av_freep(&s->stream_groups);
ff_flush_packet_queue(s);
av_freep(&s->url);
av_free(s);
@@ -464,7 +503,7 @@ int av_find_best_stream(AVFormatContext *ic, enum AVMediaType type,
*/
static int match_stream_specifier(const AVFormatContext *s, const AVStream *st,
const char *spec, const char **indexptr,
- const AVProgram **p)
+ const AVStreamGroup **g, const AVProgram **p)
{
int match = 1; /* Stores if the specifier matches so far. */
while (*spec) {
@@ -493,6 +532,46 @@ static int match_stream_specifier(const AVFormatContext *s, const AVStream *st,
match = 0;
if (nopic && (st->disposition & AV_DISPOSITION_ATTACHED_PIC))
match = 0;
+ } else if (*spec == 'g' && *(spec + 1) == ':') {
+ int64_t group_idx = -1, group_id = -1;
+ int found = 0;
+ char *endptr;
+ spec += 2;
+ if (*spec == '#' || (*spec == 'i' && *(spec + 1) == ':')) {
+ spec += 1 + (*spec == 'i');
+ group_id = strtol(spec, &endptr, 0);
+ if (spec == endptr || (*endptr && *endptr++ != ':'))
+ return AVERROR(EINVAL);
+ spec = endptr;
+ } else {
+ group_idx = strtol(spec, &endptr, 0);
+ /* Disallow empty id and make sure that if we are not at the end, then another specifier must follow. */
+ if (spec == endptr || (*endptr && *endptr++ != ':'))
+ return AVERROR(EINVAL);
+ spec = endptr;
+ }
+ if (match) {
+ if (group_id > 0) {
+ for (unsigned i = 0; i < s->nb_stream_groups; i++) {
+ if (group_id == s->stream_groups[i]->id) {
+ group_idx = i;
+ break;
+ }
+ }
+ }
+ if (group_idx < 0 || group_idx > s->nb_stream_groups)
+ return AVERROR(EINVAL);
+ for (unsigned j = 0; j < s->stream_groups[group_idx]->nb_streams; j++) {
+ if (st->index == s->stream_groups[group_idx]->streams[j]->index) {
+ found = 1;
+ if (g)
+ *g = s->stream_groups[group_idx];
+ break;
+ }
+ }
+ }
+ if (!found)
+ match = 0;
} else if (*spec == 'p' && *(spec + 1) == ':') {
int prog_id;
int found = 0;
@@ -591,10 +670,11 @@ int avformat_match_stream_specifier(AVFormatContext *s, AVStream *st,
int ret, index;
char *endptr;
const char *indexptr = NULL;
+ const AVStreamGroup *g = NULL;
const AVProgram *p = NULL;
int nb_streams;
- ret = match_stream_specifier(s, st, spec, &indexptr, &p);
+ ret = match_stream_specifier(s, st, spec, &indexptr, &g, &p);
if (ret < 0)
goto error;
@@ -612,10 +692,11 @@ int avformat_match_stream_specifier(AVFormatContext *s, AVStream *st,
return (index == st->index);
/* If we requested a matching stream index, we have to ensure st is that. */
- nb_streams = p ? p->nb_stream_indexes : s->nb_streams;
+ nb_streams = g ? g->nb_streams : (p ? p->nb_stream_indexes : s->nb_streams);
for (int i = 0; i < nb_streams && index >= 0; i++) {
- const AVStream *candidate = s->streams[p ? p->stream_index[i] : i];
- ret = match_stream_specifier(s, candidate, spec, NULL, NULL);
+ unsigned idx = g ? g->streams[i]->index : (p ? p->stream_index[i] : i);
+ const AVStream *candidate = s->streams[idx];
+ ret = match_stream_specifier(s, candidate, spec, NULL, NULL, NULL);
if (ret < 0)
goto error;
if (ret > 0 && index-- == 0 && st == candidate)
diff --git a/libavformat/avformat.h b/libavformat/avformat.h
index 9e7eca007e..5d0fe82250 100644
--- a/libavformat/avformat.h
+++ b/libavformat/avformat.h
@@ -1018,6 +1018,83 @@ typedef struct AVStream {
int pts_wrap_bits;
} AVStream;
+enum AVStreamGroupParamsType {
+ AV_STREAM_GROUP_PARAMS_NONE,
+ AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT,
+ AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION,
+};
+
+struct AVIAMFAudioElement;
+struct AVIAMFMixPresentation;
+
+typedef struct AVStreamGroup {
+ /**
+ * A class for @ref avoptions. Set by avformat_stream_group_create().
+ */
+ const AVClass *av_class;
+
+ void *priv_data;
+
+ /**
+ * Group index in AVFormatContext.
+ */
+ unsigned int index;
+
+ /**
+ * Group type-specific group ID.
+ *
+ * decoding: set by libavformat
+ * encoding: may set by the user
+ */
+ int64_t id;
+
+ /**
+ * Group type
+ *
+ * decoding: set by libavformat on group creation
+ * encoding: set by avformat_stream_group_create()
+ */
+ enum AVStreamGroupParamsType type;
+
+ /**
+ * Group type-specific parameters
+ */
+ union {
+ struct AVIAMFAudioElement *iamf_audio_element;
+ struct AVIAMFMixPresentation *iamf_mix_presentation;
+ } params;
+
+ /**
+ * Metadata that applies to the whole group.
+ *
+ * - demuxing: set by libavformat on group creation
+ * - muxing: may be set by the caller before avformat_write_header()
+ *
+ * Freed by libavformat in avformat_free_context().
+ */
+ AVDictionary *metadata;
+
+ /**
+ * Number of elements in AVStreamGroup.streams.
+ *
+ * Set by avformat_stream_group_add_stream() must not be modified by any other code.
+ */
+ unsigned int nb_streams;
+
+ /**
+ * A list of streams in the group. New entries are created with
+ * avformat_stream_group_add_stream().
+ *
+ * - demuxing: entries are created by libavformat on group creation.
+ * If AVFMTCTX_NOHEADER is set in ctx_flags, then new entries may also
+ * appear in av_read_frame().
+ * - muxing: entries are created by the user before avformat_write_header().
+ *
+ * Freed by libavformat in avformat_free_context().
+ */
+ AVStream **streams;
+} AVStreamGroup;
+
struct AVCodecParserContext *av_stream_get_parser(const AVStream *s);
#if FF_API_GET_END_PTS
@@ -1726,6 +1803,26 @@ typedef struct AVFormatContext {
* @return 0 on success, a negative AVERROR code on failure
*/
int (*io_close2)(struct AVFormatContext *s, AVIOContext *pb);
+
+ /**
+ * Number of elements in AVFormatContext.stream_groups.
+ *
+ * Set by avformat_stream_group_create(), must not be modified by any other code.
+ */
+ unsigned int nb_stream_groups;
+
+ /**
+ * A list of all stream groups in the file. New groups are created with
+ * avformat_stream_group_create(), and filled with avformat_stream_group_add_stream().
+ *
+ * - demuxing: groups may be created by libavformat in avformat_open_input().
+ * If AVFMTCTX_NOHEADER is set in ctx_flags, then new groups may also
+ * appear in av_read_frame().
+ * - muxing: groups may be created by the user before avformat_write_header().
+ *
+ * Freed by libavformat in avformat_free_context().
+ */
+ AVStreamGroup **stream_groups;
} AVFormatContext;
/**
@@ -1844,6 +1941,37 @@ const AVClass *avformat_get_class(void);
*/
const AVClass *av_stream_get_class(void);
+/**
+ * Get the AVClass for AVStreamGroup. It can be used in combination with
+ * AV_OPT_SEARCH_FAKE_OBJ for examining options.
+ *
+ * @see av_opt_find().
+ */
+const AVClass *av_stream_group_get_class(void);
+
+/**
+ * Add a new empty stream group to a media file.
+ *
+ * When demuxing, it may be called by the demuxer in read_header(). If the
+ * flag AVFMTCTX_NOHEADER is set in s.ctx_flags, then it may also
+ * be called in read_packet().
+ *
+ * When muxing, may be called by the user before avformat_write_header().
+ *
+ * User is required to call avformat_free_context() to clean up the allocation
+ * by avformat_stream_group_create().
+ *
+ * New streams can be added to the group with avformat_stream_group_add_stream().
+ *
+ * @param s media file handle
+ *
+ * @return newly created group or NULL on error.
+ * @see avformat_new_stream, avformat_stream_group_add_stream.
+ */
+AVStreamGroup *avformat_stream_group_create(AVFormatContext *s,
+ enum AVStreamGroupParamsType type,
+ AVDictionary **options);
+
/**
* Add a new stream to a media file.
*
@@ -1863,6 +1991,31 @@ const AVClass *av_stream_get_class(void);
*/
AVStream *avformat_new_stream(AVFormatContext *s, const struct AVCodec *c);
+/**
+ * Add an already allocated stream to a stream group.
+ *
+ * When demuxing, it may be called by the demuxer in read_header(). If the
+ * flag AVFMTCTX_NOHEADER is set in s.ctx_flags, then it may also
+ * be called in read_packet().
+ *
+ * When muxing, may be called by the user before avformat_write_header() after
+ * having allocated a new group with avformat_stream_group_create() and stream with
+ * avformat_new_stream().
+ *
+ * User is required to call avformat_free_context() to clean up the allocation
+ * by avformat_stream_group_add_stream().
+ *
+ * @param stg stream group belonging to a media file.
+ * @param st stream in the media file to add to the group.
+ *
+ * @retval 0 success
+ * @retval AVERROR(EEXIST) the stream was already in the group
+ * @retval "another negative error code" legitimate errors
+ *
+ * @see avformat_new_stream, avformat_stream_group_create.
+ */
+int avformat_stream_group_add_stream(AVStreamGroup *stg, AVStream *st);
+
#if FF_API_AVSTREAM_SIDE_DATA
/**
* Wrap an existing array as stream side data.
diff --git a/libavformat/dump.c b/libavformat/dump.c
index c0868a1bb3..cc179f284f 100644
--- a/libavformat/dump.c
+++ b/libavformat/dump.c
@@ -24,6 +24,7 @@
#include "libavutil/channel_layout.h"
#include "libavutil/display.h"
+#include "libavutil/iamf.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/log.h"
#include "libavutil/mastering_display_metadata.h"
@@ -134,28 +135,36 @@ static void print_fps(double d, const char *postfix)
av_log(NULL, AV_LOG_INFO, "%1.0fk %s", d / 1000, postfix);
}
-static void dump_metadata(void *ctx, const AVDictionary *m, const char *indent)
+static void dump_dictionary(void *ctx, const AVDictionary *m,
+ const char *name, const char *indent)
{
- if (m && !(av_dict_count(m) == 1 && av_dict_get(m, "language", NULL, 0))) {
- const AVDictionaryEntry *tag = NULL;
-
- av_log(ctx, AV_LOG_INFO, "%sMetadata:\n", indent);
- while ((tag = av_dict_iterate(m, tag)))
- if (strcmp("language", tag->key)) {
- const char *p = tag->value;
- av_log(ctx, AV_LOG_INFO,
- "%s %-16s: ", indent, tag->key);
- while (*p) {
- size_t len = strcspn(p, "\x8\xa\xb\xc\xd");
- av_log(ctx, AV_LOG_INFO, "%.*s", (int)(FFMIN(255, len)), p);
- p += len;
- if (*p == 0xd) av_log(ctx, AV_LOG_INFO, " ");
- if (*p == 0xa) av_log(ctx, AV_LOG_INFO, "\n%s %-16s: ", indent, "");
- if (*p) p++;
- }
- av_log(ctx, AV_LOG_INFO, "\n");
+ const AVDictionaryEntry *tag = NULL;
+
+ if (!m)
+ return;
+
+ av_log(ctx, AV_LOG_INFO, "%s%s:\n", indent, name);
+ while ((tag = av_dict_iterate(m, tag)))
+ if (strcmp("language", tag->key)) {
+ const char *p = tag->value;
+ av_log(ctx, AV_LOG_INFO,
+ "%s %-16s: ", indent, tag->key);
+ while (*p) {
+ size_t len = strcspn(p, "\x8\xa\xb\xc\xd");
+ av_log(ctx, AV_LOG_INFO, "%.*s", (int)(FFMIN(255, len)), p);
+ p += len;
+ if (*p == 0xd) av_log(ctx, AV_LOG_INFO, " ");
+ if (*p == 0xa) av_log(ctx, AV_LOG_INFO, "\n%s %-16s: ", indent, "");
+ if (*p) p++;
}
- }
+ av_log(ctx, AV_LOG_INFO, "\n");
+ }
+}
+
+static void dump_metadata(void *ctx, const AVDictionary *m, const char *indent)
+{
+ if (m && !(av_dict_count(m) == 1 && av_dict_get(m, "language", NULL, 0)))
+ dump_dictionary(ctx, m, "Metadata", indent);
}
/* param change side data*/
@@ -509,7 +518,7 @@ static void dump_sidedata(void *ctx, const AVStream *st, const char *indent)
/* "user interface" functions */
static void dump_stream_format(const AVFormatContext *ic, int i,
- int index, int is_output)
+ int group_index, int index, int is_output)
{
char buf[256];
int flags = (is_output ? ic->oformat->flags : ic->iformat->flags);
@@ -517,6 +526,8 @@ static void dump_stream_format(const AVFormatContext *ic, int i,
const FFStream *const sti = cffstream(st);
const AVDictionaryEntry *lang = av_dict_get(st->metadata, "language", NULL, 0);
const char *separator = ic->dump_separator;
+ const char *group_indent = group_index >= 0 ? " " : "";
+ const char *extra_indent = group_index >= 0 ? " " : " ";
AVCodecContext *avctx;
int ret;
@@ -543,7 +554,8 @@ static void dump_stream_format(const AVFormatContext *ic, int i,
avcodec_string(buf, sizeof(buf), avctx, is_output);
avcodec_free_context(&avctx);
- av_log(NULL, AV_LOG_INFO, " Stream #%d:%d", index, i);
+ av_log(NULL, AV_LOG_INFO, "%s Stream #%d", group_indent, index);
+ av_log(NULL, AV_LOG_INFO, ":%d", i);
/* the pid is an important information, so we display it */
/* XXX: add a generic system */
@@ -621,9 +633,89 @@ static void dump_stream_format(const AVFormatContext *ic, int i,
av_log(NULL, AV_LOG_INFO, " (non-diegetic)");
av_log(NULL, AV_LOG_INFO, "\n");
- dump_metadata(NULL, st->metadata, " ");
+ dump_metadata(NULL, st->metadata, extra_indent);
+
+ dump_sidedata(NULL, st, extra_indent);
+}
+
+static void dump_stream_group(const AVFormatContext *ic, uint8_t *printed,
+ int i, int index, int is_output)
+{
+ const AVStreamGroup *stg = ic->stream_groups[i];
+ int flags = (is_output ? ic->oformat->flags : ic->iformat->flags);
+ char buf[512];
+ int ret;
- dump_sidedata(NULL, st, " ");
+ av_log(NULL, AV_LOG_INFO, " Stream group #%d:%d", index, i);
+ if (flags & AVFMT_SHOW_IDS)
+ av_log(NULL, AV_LOG_INFO, "[0x%"PRIx64"]", stg->id);
+ av_log(NULL, AV_LOG_INFO, ":");
+
+ switch (stg->type) {
+ case AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT: {
+ const AVIAMFAudioElement *audio_element = stg->params.iamf_audio_element;
+ av_log(NULL, AV_LOG_INFO, " IAMF Audio Element\n");
+ dump_metadata(NULL, stg->metadata, " ");
+ for (int j = 0; j < audio_element->nb_layers; j++) {
+ const AVIAMFLayer *layer = audio_element->layers[j];
+ int channel_count = layer->ch_layout.nb_channels;
+ av_log(NULL, AV_LOG_INFO, " Layer %d:", j);
+ ret = av_channel_layout_describe(&layer->ch_layout, buf, sizeof(buf));
+ if (ret >= 0)
+ av_log(NULL, AV_LOG_INFO, " %s", buf);
+ av_log(NULL, AV_LOG_INFO, "\n");
+ for (int k = 0; channel_count > 0 && k < stg->nb_streams; k++) {
+ AVStream *st = stg->streams[k];
+ dump_stream_format(ic, st->index, i, index, is_output);
+ printed[st->index] = 1;
+ channel_count -= st->codecpar->ch_layout.nb_channels;
+ }
+ }
+ break;
+ }
+ case AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION: {
+ const AVIAMFMixPresentation *mix_presentation = stg->params.iamf_mix_presentation;
+ av_log(NULL, AV_LOG_INFO, " IAMF Mix Presentation\n");
+ dump_metadata(NULL, stg->metadata, " ");
+ dump_dictionary(NULL, mix_presentation->annotations, "Annotations", " ");
+ for (int j = 0; j < mix_presentation->nb_submixes; j++) {
+ AVIAMFSubmix *sub_mix = mix_presentation->submixes[j];
+ av_log(NULL, AV_LOG_INFO, " Submix %d:\n", j);
+ for (int k = 0; k < sub_mix->nb_elements; k++) {
+ const AVIAMFSubmixElement *submix_element = sub_mix->elements[k];
+ const AVStreamGroup *audio_element = NULL;
+ for (int l = 0; l < ic->nb_stream_groups; l++)
+ if (ic->stream_groups[l]->type == AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT &&
+ ic->stream_groups[l]->id == submix_element->audio_element_id) {
+ audio_element = ic->stream_groups[l];
+ break;
+ }
+ if (audio_element) {
+ av_log(NULL, AV_LOG_INFO, " IAMF Audio Element #%d:%d",
+ index, audio_element->index);
+ if (flags & AVFMT_SHOW_IDS)
+ av_log(NULL, AV_LOG_INFO, "[0x%"PRIx64"]", audio_element->id);
+ av_log(NULL, AV_LOG_INFO, "\n");
+ dump_dictionary(NULL, submix_element->annotations, "Annotations", " ");
+ }
+ }
+ for (int k = 0; k < sub_mix->nb_layouts; k++) {
+ const AVIAMFSubmixLayout *submix_layout = sub_mix->layouts[k];
+ av_log(NULL, AV_LOG_INFO, " Layout #%d:", k);
+ if (submix_layout->layout_type == 2) {
+ ret = av_channel_layout_describe(&submix_layout->sound_system, buf, sizeof(buf));
+ if (ret >= 0)
+ av_log(NULL, AV_LOG_INFO, " %s", buf);
+ } else if (submix_layout->layout_type == 3)
+ av_log(NULL, AV_LOG_INFO, " Binaural");
+ av_log(NULL, AV_LOG_INFO, "\n");
+ }
+ }
+ break;
+ }
+ default:
+ break;
+ }
}
void av_dump_format(AVFormatContext *ic, int index,
@@ -699,7 +791,7 @@ void av_dump_format(AVFormatContext *ic, int index,
dump_metadata(NULL, program->metadata, " ");
for (k = 0; k < program->nb_stream_indexes; k++) {
dump_stream_format(ic, program->stream_index[k],
- index, is_output);
+ -1, index, is_output);
printed[program->stream_index[k]] = 1;
}
total += program->nb_stream_indexes;
@@ -708,9 +800,12 @@ void av_dump_format(AVFormatContext *ic, int index,
av_log(NULL, AV_LOG_INFO, " No Program\n");
}
+ for (i = 0; i < ic->nb_stream_groups; i++)
+ dump_stream_group(ic, printed, i, index, is_output);
+
for (i = 0; i < ic->nb_streams; i++)
if (!printed[i])
- dump_stream_format(ic, i, index, is_output);
+ dump_stream_format(ic, i, -1, index, is_output);
av_free(printed);
}
diff --git a/libavformat/internal.h b/libavformat/internal.h
index 7702986c9c..c6181683ef 100644
--- a/libavformat/internal.h
+++ b/libavformat/internal.h
@@ -202,6 +202,7 @@ typedef struct FFStream {
*/
AVStream pub;
+ AVFormatContext *fmtctx;
/**
* Set to 1 if the codec allows reordering, so pts can be different
* from dts.
@@ -427,6 +428,26 @@ static av_always_inline const FFStream *cffstream(const AVStream *st)
return (const FFStream*)st;
}
+typedef struct FFStreamGroup {
+ /**
+ * The public context.
+ */
+ AVStreamGroup pub;
+
+ AVFormatContext *fmtctx;
+} FFStreamGroup;
+
+
+static av_always_inline FFStreamGroup *ffstreamgroup(AVStreamGroup *stg)
+{
+ return (FFStreamGroup*)stg;
+}
+
+static av_always_inline const FFStreamGroup *cffstreamgroup(const AVStreamGroup *stg)
+{
+ return (const FFStreamGroup*)stg;
+}
+
#ifdef __GNUC__
#define dynarray_add(tab, nb_ptr, elem)\
do {\
@@ -608,6 +629,18 @@ void ff_free_stream(AVStream **st);
*/
void ff_remove_stream(AVFormatContext *s, AVStream *st);
+/**
+ * Frees a stream group without modifying the corresponding AVFormatContext.
+ * Must only be called if the latter doesn't matter or if the stream
+ * is not yet attached to an AVFormatContext.
+ */
+void ff_free_stream_group(AVStreamGroup **pstg);
+/**
+ * Remove a stream group from its AVFormatContext and free it.
+ * The group must be the last stream of the AVFormatContext.
+ */
+void ff_remove_stream_group(AVFormatContext *s, AVStreamGroup *stg);
+
unsigned int ff_codec_get_tag(const AVCodecTag *tags, enum AVCodecID id);
enum AVCodecID ff_codec_get_id(const AVCodecTag *tags, unsigned int tag);
diff --git a/libavformat/options.c b/libavformat/options.c
index 1d8c52246b..bf6113ca95 100644
--- a/libavformat/options.c
+++ b/libavformat/options.c
@@ -26,6 +26,7 @@
#include "libavcodec/codec_par.h"
#include "libavutil/avassert.h"
+#include "libavutil/iamf.h"
#include "libavutil/internal.h"
#include "libavutil/intmath.h"
#include "libavutil/opt.h"
@@ -271,6 +272,7 @@ AVStream *avformat_new_stream(AVFormatContext *s, const AVCodec *c)
if (!st->codecpar)
goto fail;
+ sti->fmtctx = s;
sti->avctx = avcodec_alloc_context3(NULL);
if (!sti->avctx)
goto fail;
@@ -325,6 +327,143 @@ fail:
return NULL;
}
+static void *stream_group_child_next(void *obj, void *prev)
+{
+ AVStreamGroup *stg = obj;
+ if (!prev) {
+ switch(stg->type) {
+ case AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT:
+ return stg->params.iamf_audio_element;
+ case AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION:
+ return stg->params.iamf_mix_presentation;
+ default:
+ break;
+ }
+ }
+ return NULL;
+}
+
+static const AVClass *stream_group_child_iterate(void **opaque)
+{
+ uintptr_t i = (uintptr_t)*opaque;
+ const AVClass *ret = NULL;
+
+ switch(i) {
+ case AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT:
+ ret = av_iamf_audio_element_get_class();
+ break;
+ case AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION:
+ ret = av_iamf_mix_presentation_get_class();
+ break;
+ default:
+ break;
+ }
+
+ if (ret)
+ *opaque = (void*)(i + 1);
+ return ret;
+}
+
+static const AVOption stream_group_options[] = {
+ {"id", "Set group id", offsetof(AVStreamGroup, id), AV_OPT_TYPE_INT64, {.i64 = 0}, 0, INT64_MAX, AV_OPT_FLAG_ENCODING_PARAM },
+ { NULL }
+};
+
+static const AVClass stream_group_class = {
+ .class_name = "AVStreamGroup",
+ .item_name = av_default_item_name,
+ .version = LIBAVUTIL_VERSION_INT,
+ .option = stream_group_options,
+ .child_next = stream_group_child_next,
+ .child_class_iterate = stream_group_child_iterate,
+};
+
+const AVClass *av_stream_group_get_class(void)
+{
+ return &stream_group_class;
+}
+
+AVStreamGroup *avformat_stream_group_create(AVFormatContext *s,
+ enum AVStreamGroupParamsType type,
+ AVDictionary **options)
+{
+ AVStreamGroup **stream_groups;
+ AVStreamGroup *stg;
+ FFStreamGroup *stgi;
+
+ stream_groups = av_realloc_array(s->stream_groups, s->nb_stream_groups + 1,
+ sizeof(*stream_groups));
+ if (!stream_groups)
+ return NULL;
+ s->stream_groups = stream_groups;
+
+ stgi = av_mallocz(sizeof(*stgi));
+ if (!stgi)
+ return NULL;
+ stg = &stgi->pub;
+
+ stg->av_class = &stream_group_class;
+ av_opt_set_defaults(stg);
+ stg->type = type;
+ switch (type) {
+ case AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT:
+ stg->params.iamf_audio_element = av_iamf_audio_element_alloc();
+ if (!stg->params.iamf_audio_element)
+ goto fail;
+ break;
+ case AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION:
+ stg->params.iamf_mix_presentation = av_iamf_mix_presentation_alloc();
+ if (!stg->params.iamf_mix_presentation)
+ goto fail;
+ break;
+ default:
+ goto fail;
+ }
+
+ if (options) {
+ if (av_opt_set_dict2(stg, options, AV_OPT_SEARCH_CHILDREN))
+ goto fail;
+ }
+
+ stgi->fmtctx = s;
+ stg->index = s->nb_stream_groups;
+
+ s->stream_groups[s->nb_stream_groups++] = stg;
+
+ return stg;
+fail:
+ ff_free_stream_group(&stg);
+ return NULL;
+}
+
+static int stream_group_add_stream(AVStreamGroup *stg, AVStream *st)
+{
+ AVStream **streams = av_realloc_array(stg->streams, stg->nb_streams + 1,
+ sizeof(*stg->streams));
+ if (!streams)
+ return AVERROR(ENOMEM);
+
+ stg->streams = streams;
+ stg->streams[stg->nb_streams++] = st;
+
+ return 0;
+}
+
+int avformat_stream_group_add_stream(AVStreamGroup *stg, AVStream *st)
+{
+ const FFStreamGroup *stgi = cffstreamgroup(stg);
+ const FFStream *sti = cffstream(st);
+
+ if (stgi->fmtctx != sti->fmtctx)
+ return AVERROR(EINVAL);
+
+ for (int i = 0; i < stg->nb_streams; i++)
+ if (stg->streams[i]->index == st->index)
+ return AVERROR(EEXIST);
+
+ return stream_group_add_stream(stg, st);
+}
+
static int option_is_disposition(const AVOption *opt)
{
return opt->type == AV_OPT_TYPE_CONST &&
--
2.43.0
_______________________________________________
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] 12+ messages in thread* [FFmpeg-devel] [PATCH 3/8] ffmpeg: add support for muxing AVStreamGroups
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 1/8] avutil: introduce an Immersive Audio Model and Formats API James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 2/8] avformat: introduce AVStreamGroup James Almer
@ 2023-12-14 20:14 ` James Almer
2023-12-15 21:28 ` James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 4/8] avcodec/packet: add IAMF Parameters side data types James Almer
` (4 subsequent siblings)
7 siblings, 1 reply; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Starting with IAMF support.
Signed-off-by: James Almer <jamrial@gmail.com>
---
doc/ffmpeg.texi | 200 ++++++++++++++++++++++
fftools/ffmpeg.h | 2 +
fftools/ffmpeg_mux_init.c | 341 ++++++++++++++++++++++++++++++++++++++
fftools/ffmpeg_opt.c | 2 +
4 files changed, 545 insertions(+)
diff --git a/doc/ffmpeg.texi b/doc/ffmpeg.texi
index c503963941..1fadb20686 100644
--- a/doc/ffmpeg.texi
+++ b/doc/ffmpeg.texi
@@ -623,6 +623,206 @@ Not all muxers support embedded thumbnails, and those who do, only support a few
Creates a program with the specified @var{title}, @var{program_num} and adds the specified
@var{stream}(s) to it.
+@item -stream_group type=@var{type}:st=@var{stream}[:st=@var{stream}][:stg=@var{stream_group}][:id=@var{stream_group_id}...] (@emph{output})
+
+Creates a stream group of the specified @var{type}, @var{stream_group_id} and adds the specified
+@var{stream}(s) and/or previously defined @var{stream_group}(s) to it.
+
+@var{type} can be one of the following:
+@table @option
+
+@item iamf_audio_element
+Groups @var{stream}s that belong to the same IAMF Audio Element
+
+For this group @var{type}, the following options are available
+@table @option
+@item audio_element_type
+The Audio Element type. The following values are supported:
+
+@table @option
+@item channel
+Scalable channel audio representation
+@item scene
+Ambisonics representation
+@end table
+
+@item demixing
+Demixing information used to reconstruct a scalable channel audio representation.
+This option must be separated from the rest with a ',', and takes the following
+key=value options
+
+@table @option
+@item parameter_id
+An identifier parameters blocks in frames may refer to
+@item dmixp_mode
+A pre-defined combination of demixing parameters
+@end table
+
+@item recon_gain
+Recon gain information used to reconstruct a scalable channel audio representation.
+This option must be separated from the rest with a ',', and takes the following
+key=value options
+
+@table @option
+@item parameter_id
+An identifier parameters blocks in frames may refer to
+@end table
+
+@item layer
+A layer defining a Channel Layout in the Audio Element.
+This option must be separated from the rest with a ','. Several ',' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options
+
+@table @option
+@item ch_layout
+The layer's channel layout
+@item flags
+The following flags are available:
+
+@table @option
+@item recon_gain
+Wether to signal if recon_gain is present as metadata in parameter blocks within frames
+@end table
+
+@item output_gain
+@item output_gain_flags
+Which channels output_gain applies to. The following flags are available:
+
+@table @option
+@item FL
+@item FR
+@item BL
+@item BR
+@item TFL
+@item TFR
+@end table
+
+@item ambisonics_mode
+The ambisonics mode. This has no effect if audio_element_type is set to channel.
+
+The following values are supported:
+
+@table @option
+@item mono
+Each ambisonics channel is coded as an individual mono stream in the group
+@end table
+
+@end table
+
+@item default_w
+Default weight value
+
+@end table
+
+@item iamf_mix_presentation
+Groups @var{stream}s that belong to all IAMF Audio Element the same
+IAMF Mix Presentation references
+
+For this group @var{type}, the following options are available
+
+@table @option
+@item submix
+A sub-mix within the Mix Presentation.
+This option must be separated from the rest with a ','. Several ',' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options
+
+@table @option
+@item parameter_id
+An identifier parameters blocks in frames may refer to, for post-processing the mixed
+audio signal to generate the audio signal for playback
+@item parameter_rate
+The sample rate duration fields in parameters blocks in frames that refer to this
+@var{parameter_id} are expressed as
+@item default_mix_gain
+Default mix gain value to apply when there are no parameter blocks sharing the same
+@var{parameter_id} for a given frame
+
+@item element
+References an Audio Element used in this Mix Presentation to generate the final output
+audio signal for playback.
+This option must be separated from the rest with a '|'. Several '|' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options:
+
+@table @option
+@item stg
+The @var{stream_group_id} for an Audio Element which this sub-mix refers to
+@item parameter_id
+An identifier parameters blocks in frames may refer to, for applying any processing to
+the referenced and rendered Audio Element before being summed with other processed Audio
+Elements
+@item parameter_rate
+The sample rate duration fields in parameters blocks in frames that refer to this
+@var{parameter_id} are expressed as
+@item default_mix_gain
+Default mix gain value to apply when there are no parameter blocks sharing the same
+@var{parameter_id} for a given frame
+@item annotations
+A key=value string describing the sub-mix element where "key" is a string conforming to
+BCP-47 that specifies the language for the "value" string. "key" must be the same as the
+one in the mix's @var{annotations}
+@item headphones_rendering_mode
+Indicates whether the input channel-based Audio Element is rendered to stereo loudspeakers
+or spatialized with a binaural renderer when played back on headphones.
+This has no effect if the referenced Audio Element's @var{audio_element_type} is set to
+channel.
+
+The following values are supported:
+
+@table @option
+@item stereo
+@item binaural
+@end table
+
+@end table
+
+@item layout
+Specifies the layouts for this sub-mix on which the loudness information was measured.
+This option must be separated from the rest with a '|'. Several '|' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options:
+
+@table @option
+@item layout_type
+
+@table @option
+@item loudspeakers
+The layout follows the loudspeaker sound system convention of ITU-2051-3.
+@item binaural
+The layout is binaural.
+@end table
+
+@item sound_system
+Channel layout matching one of Sound Systems A to J of ITU-2051-3, plus 7.1.2 and 3.1.2
+This has no effect if @var{layout_type} is set to binaural.
+@item integrated_loudness
+The program integrated loudness information, as defined in ITU-1770-4.
+@item digital_peak
+The digital (sampled) peak value of the audio signal, as defined in ITU-1770-4.
+@item true_peak
+The true peak of the audio signal, as defined in ITU-1770-4.
+@item dialog_anchored_loudness
+The Dialogue loudness information, as defined in ITU-1770-4.
+@item album_anchored_loudness
+The Album loudness information, as defined in ITU-1770-4.
+@end table
+
+@end table
+
+@item annotations
+A key=value string string describing the mix where "key" is a string conforming to BCP-47
+that specifies the language for the "value" string. "key" must be the same as the ones in
+all sub-mix element's @var{annotations}s
+@end table
+
+@end table
+
@item -target @var{type} (@emph{output})
Specify target file type (@code{vcd}, @code{svcd}, @code{dvd}, @code{dv},
@code{dv50}). @var{type} may be prefixed with @code{pal-}, @code{ntsc-} or
diff --git a/fftools/ffmpeg.h b/fftools/ffmpeg.h
index affa80856a..1169f723d1 100644
--- a/fftools/ffmpeg.h
+++ b/fftools/ffmpeg.h
@@ -281,6 +281,8 @@ typedef struct OptionsContext {
int nb_disposition;
SpecifierOpt *program;
int nb_program;
+ SpecifierOpt *stream_groups;
+ int nb_stream_groups;
SpecifierOpt *time_bases;
int nb_time_bases;
SpecifierOpt *enc_time_bases;
diff --git a/fftools/ffmpeg_mux_init.c b/fftools/ffmpeg_mux_init.c
index f527a083db..0f03ee092e 100644
--- a/fftools/ffmpeg_mux_init.c
+++ b/fftools/ffmpeg_mux_init.c
@@ -40,6 +40,7 @@
#include "libavutil/dict.h"
#include "libavutil/display.h"
#include "libavutil/getenv_utf8.h"
+#include "libavutil/iamf.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/log.h"
#include "libavutil/mem.h"
@@ -2008,6 +2009,342 @@ static int setup_sync_queues(Muxer *mux, AVFormatContext *oc, int64_t buf_size_u
return 0;
}
+static int of_parse_iamf_audio_element_layers(Muxer *mux, AVStreamGroup *stg, char *ptr)
+{
+ AVIAMFAudioElement *audio_element = stg->params.iamf_audio_element;
+ AVDictionary *dict = NULL;
+ const char *token;
+ int ret = 0;
+
+ audio_element->demixing_info =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_DEMIXING, 1, NULL);
+ audio_element->recon_gain_info =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN, 1, NULL);
+
+ if (!audio_element->demixing_info ||
+ !audio_element->recon_gain_info)
+ return AVERROR(ENOMEM);
+
+ /* process manually set layers and parameters */
+ token = av_strtok(NULL, ",", &ptr);
+ while (token) {
+ const AVDictionaryEntry *e;
+ int demixing = 0, recon_gain = 0;
+ int layer = 0;
+
+ if (av_strstart(token, "layer=", &token))
+ layer = 1;
+ else if (av_strstart(token, "demixing=", &token))
+ demixing = 1;
+ else if (av_strstart(token, "recon_gain=", &token))
+ recon_gain = 1;
+
+ av_dict_free(&dict);
+ ret = av_dict_parse_string(&dict, token, "=", ":", 0);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Error parsing audio element specification %s\n", token);
+ goto fail;
+ }
+
+ if (layer) {
+ AVIAMFLayer *audio_layer = av_iamf_audio_element_add_layer(audio_element);
+ if (!audio_layer) {
+ av_log(mux, AV_LOG_ERROR, "Error adding layer to stream group %d\n", stg->index);
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ av_opt_set_dict(audio_layer, &dict);
+ } else if (demixing || recon_gain) {
+ AVIAMFParamDefinition *param = demixing ? audio_element->demixing_info
+ : audio_element->recon_gain_info;
+ void *subblock = av_iamf_param_definition_get_subblock(param, 0);
+
+ av_opt_set_dict(param, &dict);
+ av_opt_set_dict(subblock, &dict);
+ }
+
+ // make sure that no entries are left in the dict
+ e = NULL;
+ if (e = av_dict_iterate(dict, e)) {
+ av_log(mux, AV_LOG_FATAL, "Unknown layer key %s.\n", e->key);
+ ret = AVERROR(EINVAL);
+ goto fail;
+ }
+ token = av_strtok(NULL, ",", &ptr);
+ }
+
+fail:
+ av_dict_free(&dict);
+ if (!ret && !audio_element->nb_layers) {
+ av_log(mux, AV_LOG_ERROR, "No layer in audio element specification\n");
+ ret = AVERROR(EINVAL);
+ }
+
+ return ret;
+}
+
+static int of_parse_iamf_submixes(Muxer *mux, AVStreamGroup *stg, char *ptr)
+{
+ AVFormatContext *oc = mux->fc;
+ AVIAMFMixPresentation *mix = stg->params.iamf_mix_presentation;
+ AVDictionary *dict = NULL;
+ const char *token;
+ char *submix_str = NULL;
+ int ret = 0;
+
+ /* process manually set submixes */
+ token = av_strtok(NULL, ",", &ptr);
+ while (token) {
+ AVIAMFSubmix *submix = NULL;
+ const char *subtoken;
+ char *subptr = NULL;
+
+ if (!av_strstart(token, "submix=", &token)) {
+ av_log(mux, AV_LOG_ERROR, "No submix in mix presentation specification \"%s\"\n", token);
+ goto fail;
+ }
+
+ submix_str = av_strdup(token);
+ if (!submix_str)
+ goto fail;
+
+ submix = av_iamf_mix_presentation_add_submix(mix);
+ if (!submix) {
+ av_log(mux, AV_LOG_ERROR, "Error adding submix to stream group %d\n", stg->index);
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ submix->output_mix_config =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN, 0, NULL);
+ if (!submix->output_mix_config) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ subptr = NULL;
+ subtoken = av_strtok(submix_str, "|", &subptr);
+ while (subtoken) {
+ const AVDictionaryEntry *e;
+ int element = 0, layout = 0;
+
+ if (av_strstart(subtoken, "element=", &subtoken))
+ element = 1;
+ else if (av_strstart(subtoken, "layout=", &subtoken))
+ layout = 1;
+
+ av_dict_free(&dict);
+ ret = av_dict_parse_string(&dict, subtoken, "=", ":", 0);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Error parsing submix specification \"%s\"\n", subtoken);
+ goto fail;
+ }
+
+ if (element) {
+ AVIAMFSubmixElement *submix_element;
+ int idx = -1;
+
+ if (e = av_dict_get(dict, "stg", NULL, 0))
+ idx = strtol(e->value, NULL, 0);
+ av_dict_set(&dict, "stg", NULL, 0);
+ if (idx < 0 || idx >= oc->nb_stream_groups) {
+ av_log(mux, AV_LOG_ERROR, "Invalid or missing stream group index in "
+ "submix element specification \"%s\"\n", subtoken);
+ ret = AVERROR(EINVAL);
+ goto fail;
+ }
+ submix_element = av_iamf_submix_add_element(submix);
+ if (!submix_element) {
+ av_log(mux, AV_LOG_ERROR, "Error adding element to submix\n");
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ submix_element->audio_element_id = oc->stream_groups[idx]->id;
+
+ submix_element->element_mix_config =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN, 0, NULL);
+ if (!submix_element->element_mix_config)
+ ret = AVERROR(ENOMEM);
+ av_opt_set_dict2(submix_element, &dict, AV_OPT_SEARCH_CHILDREN);
+ } else if (layout) {
+ AVIAMFSubmixLayout *submix_layout = av_iamf_submix_add_layout(submix);
+ if (!submix_layout) {
+ av_log(mux, AV_LOG_ERROR, "Error adding layout to submix\n");
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ av_opt_set_dict(submix_layout, &dict);
+ } else
+ av_opt_set_dict2(submix, &dict, AV_OPT_SEARCH_CHILDREN);
+
+ if (ret < 0) {
+ goto fail;
+ }
+
+ // make sure that no entries are left in the dict
+ e = NULL;
+ while (e = av_dict_iterate(dict, e)) {
+ av_log(mux, AV_LOG_FATAL, "Unknown submix key %s.\n", e->key);
+ ret = AVERROR(EINVAL);
+ goto fail;
+ }
+ subtoken = av_strtok(NULL, "|", &subptr);
+ }
+ av_freep(&submix_str);
+
+ if (!submix->nb_elements) {
+ av_log(mux, AV_LOG_ERROR, "No audio elements in submix specification \"%s\"\n", token);
+ ret = AVERROR(EINVAL);
+ }
+ token = av_strtok(NULL, ",", &ptr);
+ }
+
+fail:
+ av_dict_free(&dict);
+ av_free(submix_str);
+
+ return ret;
+}
+
+static int of_parse_group_token(Muxer *mux, const char *token, char *ptr)
+{
+ AVFormatContext *oc = mux->fc;
+ AVStreamGroup *stg;
+ AVDictionary *dict = NULL, *tmp = NULL;
+ const AVDictionaryEntry *e;
+ const AVOption opts[] = {
+ { "type", "Set group type", offsetof(AVStreamGroup, type), AV_OPT_TYPE_INT,
+ { .i64 = 0 }, 0, INT_MAX, AV_OPT_FLAG_ENCODING_PARAM, "type" },
+ { "iamf_audio_element", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT }, .unit = "type" },
+ { "iamf_mix_presentation", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION }, .unit = "type" },
+ { NULL },
+ };
+ const AVClass class = {
+ .class_name = "StreamGroupType",
+ .item_name = av_default_item_name,
+ .option = opts,
+ .version = LIBAVUTIL_VERSION_INT,
+ };
+ const AVClass *pclass = &class;
+ int type, ret;
+
+ ret = av_dict_parse_string(&dict, token, "=", ":", AV_DICT_MULTIKEY);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Error parsing group specification %s\n", token);
+ return ret;
+ }
+
+ // "type" is not a user settable AVOption in AVStreamGroup, so handle it here
+ e = av_dict_get(dict, "type", NULL, 0);
+ if (!e) {
+ av_log(mux, AV_LOG_ERROR, "No type specified for Stream Group in \"%s\"\n", token);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+
+ ret = av_opt_eval_int(&pclass, opts, e->value, &type);
+ if (!ret && type == AV_STREAM_GROUP_PARAMS_NONE)
+ ret = AVERROR(EINVAL);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Invalid group type \"%s\"\n", e->value);
+ goto end;
+ }
+
+ av_dict_copy(&tmp, dict, 0);
+ stg = avformat_stream_group_create(oc, type, &tmp);
+ if (!stg) {
+ ret = AVERROR(ENOMEM);
+ goto end;
+ }
+
+ e = NULL;
+ while (e = av_dict_get(dict, "st", e, 0)) {
+ unsigned int idx = strtol(e->value, NULL, 0);
+ if (idx >= oc->nb_streams) {
+ av_log(mux, AV_LOG_ERROR, "Invalid stream index %d\n", idx);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+ ret = avformat_stream_group_add_stream(stg, oc->streams[idx]);
+ if (ret < 0)
+ goto end;
+ }
+ while (e = av_dict_get(dict, "stg", e, 0)) {
+ unsigned int idx = strtol(e->value, NULL, 0);
+ if (idx >= oc->nb_stream_groups || idx == stg->index) {
+ av_log(mux, AV_LOG_ERROR, "Invalid stream group index %u\n", idx);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+ for (int i = 0; i < oc->stream_groups[idx]->nb_streams; i++) {
+ ret = avformat_stream_group_add_stream(stg, oc->stream_groups[idx]->streams[i]);
+ if (ret < 0)
+ goto end;
+ }
+ }
+
+ switch(type) {
+ case AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT:
+ ret = of_parse_iamf_audio_element_layers(mux, stg, ptr);
+ break;
+ case AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION:
+ ret = of_parse_iamf_submixes(mux, stg, ptr);
+ break;
+ default:
+ av_log(mux, AV_LOG_FATAL, "Unknown group type %d.\n", type);
+ ret = AVERROR(EINVAL);
+ break;
+ }
+
+ if (ret < 0)
+ goto end;
+
+ // make sure that nothing but "st" and "stg" entries are left in the dict
+ e = NULL;
+ av_dict_set(&tmp, "type", NULL, 0);
+ while (e = av_dict_iterate(tmp, e)) {
+ if (!strcmp(e->key, "st") || !strcmp(e->key, "stg"))
+ continue;
+
+ av_log(mux, AV_LOG_FATAL, "Unknown group key %s.\n", e->key);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+
+ ret = 0;
+end:
+ av_dict_free(&dict);
+ av_dict_free(&tmp);
+
+ return ret;
+}
+
+static int of_add_groups(Muxer *mux, const OptionsContext *o)
+{
+ /* process manually set groups */
+ for (int i = 0; i < o->nb_stream_groups; i++) {
+ const char *token;
+ char *str, *ptr = NULL;
+ int ret = 0;
+
+ str = av_strdup(o->stream_groups[i].u.str);
+ if (!str)
+ return ret;
+
+ token = av_strtok(str, ",", &ptr);
+ if (token)
+ ret = of_parse_group_token(mux, token, ptr);
+
+ av_free(str);
+ if (ret < 0)
+ return ret;
+ }
+
+ return 0;
+}
+
static int of_add_programs(Muxer *mux, const OptionsContext *o)
{
AVFormatContext *oc = mux->fc;
@@ -2793,6 +3130,10 @@ int of_open(const OptionsContext *o, const char *filename, Scheduler *sch)
if (err < 0)
return err;
+ err = of_add_groups(mux, o);
+ if (err < 0)
+ return err;
+
err = of_add_programs(mux, o);
if (err < 0)
return err;
diff --git a/fftools/ffmpeg_opt.c b/fftools/ffmpeg_opt.c
index 6177a96a4e..915f8e3ea0 100644
--- a/fftools/ffmpeg_opt.c
+++ b/fftools/ffmpeg_opt.c
@@ -1493,6 +1493,8 @@ const OptionDef options[] = {
"add metadata", "string=string" },
{ "program", HAS_ARG | OPT_STRING | OPT_SPEC | OPT_OUTPUT, { .off = OFFSET(program) },
"add program with specified streams", "title=string:st=number..." },
+ { "stream_group", HAS_ARG | OPT_STRING | OPT_SPEC | OPT_OUTPUT, { .off = OFFSET(stream_groups) },
+ "add stream group with specified streams and group type-specific arguments", "id=number:st=number..." },
{ "dframes", HAS_ARG | OPT_PERFILE | OPT_EXPERT |
OPT_OUTPUT, { .func_arg = opt_data_frames },
"set the number of data frames to output", "number" },
--
2.43.0
_______________________________________________
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] 12+ messages in thread* [FFmpeg-devel] [PATCH 3/8] ffmpeg: add support for muxing AVStreamGroups
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 3/8] ffmpeg: add support for muxing AVStreamGroups James Almer
@ 2023-12-15 21:28 ` James Almer
0 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-15 21:28 UTC (permalink / raw)
To: ffmpeg-devel
Starting with IAMF support.
Signed-off-by: James Almer <jamrial@gmail.com>
---
doc/ffmpeg.texi | 200 ++++++++++++++++++++++
fftools/ffmpeg.h | 2 +
fftools/ffmpeg_mux_init.c | 342 ++++++++++++++++++++++++++++++++++++++
fftools/ffmpeg_opt.c | 2 +
4 files changed, 546 insertions(+)
diff --git a/doc/ffmpeg.texi b/doc/ffmpeg.texi
index c503963941..1fadb20686 100644
--- a/doc/ffmpeg.texi
+++ b/doc/ffmpeg.texi
@@ -623,6 +623,206 @@ Not all muxers support embedded thumbnails, and those who do, only support a few
Creates a program with the specified @var{title}, @var{program_num} and adds the specified
@var{stream}(s) to it.
+@item -stream_group type=@var{type}:st=@var{stream}[:st=@var{stream}][:stg=@var{stream_group}][:id=@var{stream_group_id}...] (@emph{output})
+
+Creates a stream group of the specified @var{type}, @var{stream_group_id} and adds the specified
+@var{stream}(s) and/or previously defined @var{stream_group}(s) to it.
+
+@var{type} can be one of the following:
+@table @option
+
+@item iamf_audio_element
+Groups @var{stream}s that belong to the same IAMF Audio Element
+
+For this group @var{type}, the following options are available
+@table @option
+@item audio_element_type
+The Audio Element type. The following values are supported:
+
+@table @option
+@item channel
+Scalable channel audio representation
+@item scene
+Ambisonics representation
+@end table
+
+@item demixing
+Demixing information used to reconstruct a scalable channel audio representation.
+This option must be separated from the rest with a ',', and takes the following
+key=value options
+
+@table @option
+@item parameter_id
+An identifier parameters blocks in frames may refer to
+@item dmixp_mode
+A pre-defined combination of demixing parameters
+@end table
+
+@item recon_gain
+Recon gain information used to reconstruct a scalable channel audio representation.
+This option must be separated from the rest with a ',', and takes the following
+key=value options
+
+@table @option
+@item parameter_id
+An identifier parameters blocks in frames may refer to
+@end table
+
+@item layer
+A layer defining a Channel Layout in the Audio Element.
+This option must be separated from the rest with a ','. Several ',' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options
+
+@table @option
+@item ch_layout
+The layer's channel layout
+@item flags
+The following flags are available:
+
+@table @option
+@item recon_gain
+Wether to signal if recon_gain is present as metadata in parameter blocks within frames
+@end table
+
+@item output_gain
+@item output_gain_flags
+Which channels output_gain applies to. The following flags are available:
+
+@table @option
+@item FL
+@item FR
+@item BL
+@item BR
+@item TFL
+@item TFR
+@end table
+
+@item ambisonics_mode
+The ambisonics mode. This has no effect if audio_element_type is set to channel.
+
+The following values are supported:
+
+@table @option
+@item mono
+Each ambisonics channel is coded as an individual mono stream in the group
+@end table
+
+@end table
+
+@item default_w
+Default weight value
+
+@end table
+
+@item iamf_mix_presentation
+Groups @var{stream}s that belong to all IAMF Audio Element the same
+IAMF Mix Presentation references
+
+For this group @var{type}, the following options are available
+
+@table @option
+@item submix
+A sub-mix within the Mix Presentation.
+This option must be separated from the rest with a ','. Several ',' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options
+
+@table @option
+@item parameter_id
+An identifier parameters blocks in frames may refer to, for post-processing the mixed
+audio signal to generate the audio signal for playback
+@item parameter_rate
+The sample rate duration fields in parameters blocks in frames that refer to this
+@var{parameter_id} are expressed as
+@item default_mix_gain
+Default mix gain value to apply when there are no parameter blocks sharing the same
+@var{parameter_id} for a given frame
+
+@item element
+References an Audio Element used in this Mix Presentation to generate the final output
+audio signal for playback.
+This option must be separated from the rest with a '|'. Several '|' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options:
+
+@table @option
+@item stg
+The @var{stream_group_id} for an Audio Element which this sub-mix refers to
+@item parameter_id
+An identifier parameters blocks in frames may refer to, for applying any processing to
+the referenced and rendered Audio Element before being summed with other processed Audio
+Elements
+@item parameter_rate
+The sample rate duration fields in parameters blocks in frames that refer to this
+@var{parameter_id} are expressed as
+@item default_mix_gain
+Default mix gain value to apply when there are no parameter blocks sharing the same
+@var{parameter_id} for a given frame
+@item annotations
+A key=value string describing the sub-mix element where "key" is a string conforming to
+BCP-47 that specifies the language for the "value" string. "key" must be the same as the
+one in the mix's @var{annotations}
+@item headphones_rendering_mode
+Indicates whether the input channel-based Audio Element is rendered to stereo loudspeakers
+or spatialized with a binaural renderer when played back on headphones.
+This has no effect if the referenced Audio Element's @var{audio_element_type} is set to
+channel.
+
+The following values are supported:
+
+@table @option
+@item stereo
+@item binaural
+@end table
+
+@end table
+
+@item layout
+Specifies the layouts for this sub-mix on which the loudness information was measured.
+This option must be separated from the rest with a '|'. Several '|' separated entries
+can be defined, and at least one must be set.
+
+It takes the following ":"-separated key=value options:
+
+@table @option
+@item layout_type
+
+@table @option
+@item loudspeakers
+The layout follows the loudspeaker sound system convention of ITU-2051-3.
+@item binaural
+The layout is binaural.
+@end table
+
+@item sound_system
+Channel layout matching one of Sound Systems A to J of ITU-2051-3, plus 7.1.2 and 3.1.2
+This has no effect if @var{layout_type} is set to binaural.
+@item integrated_loudness
+The program integrated loudness information, as defined in ITU-1770-4.
+@item digital_peak
+The digital (sampled) peak value of the audio signal, as defined in ITU-1770-4.
+@item true_peak
+The true peak of the audio signal, as defined in ITU-1770-4.
+@item dialog_anchored_loudness
+The Dialogue loudness information, as defined in ITU-1770-4.
+@item album_anchored_loudness
+The Album loudness information, as defined in ITU-1770-4.
+@end table
+
+@end table
+
+@item annotations
+A key=value string string describing the mix where "key" is a string conforming to BCP-47
+that specifies the language for the "value" string. "key" must be the same as the ones in
+all sub-mix element's @var{annotations}s
+@end table
+
+@end table
+
@item -target @var{type} (@emph{output})
Specify target file type (@code{vcd}, @code{svcd}, @code{dvd}, @code{dv},
@code{dv50}). @var{type} may be prefixed with @code{pal-}, @code{ntsc-} or
diff --git a/fftools/ffmpeg.h b/fftools/ffmpeg.h
index affa80856a..1169f723d1 100644
--- a/fftools/ffmpeg.h
+++ b/fftools/ffmpeg.h
@@ -281,6 +281,8 @@ typedef struct OptionsContext {
int nb_disposition;
SpecifierOpt *program;
int nb_program;
+ SpecifierOpt *stream_groups;
+ int nb_stream_groups;
SpecifierOpt *time_bases;
int nb_time_bases;
SpecifierOpt *enc_time_bases;
diff --git a/fftools/ffmpeg_mux_init.c b/fftools/ffmpeg_mux_init.c
index f527a083db..2134b28512 100644
--- a/fftools/ffmpeg_mux_init.c
+++ b/fftools/ffmpeg_mux_init.c
@@ -40,6 +40,7 @@
#include "libavutil/dict.h"
#include "libavutil/display.h"
#include "libavutil/getenv_utf8.h"
+#include "libavutil/iamf.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/log.h"
#include "libavutil/mem.h"
@@ -2008,6 +2009,343 @@ static int setup_sync_queues(Muxer *mux, AVFormatContext *oc, int64_t buf_size_u
return 0;
}
+static int of_parse_iamf_audio_element_layers(Muxer *mux, AVStreamGroup *stg, char *ptr)
+{
+ AVIAMFAudioElement *audio_element = stg->params.iamf_audio_element;
+ AVDictionary *dict = NULL;
+ const char *token;
+ int ret = 0;
+
+ audio_element->demixing_info =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_DEMIXING, 1, NULL);
+ audio_element->recon_gain_info =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN, 1, NULL);
+
+ if (!audio_element->demixing_info ||
+ !audio_element->recon_gain_info)
+ return AVERROR(ENOMEM);
+
+ /* process manually set layers and parameters */
+ token = av_strtok(NULL, ",", &ptr);
+ while (token) {
+ const AVDictionaryEntry *e;
+ int demixing = 0, recon_gain = 0;
+ int layer = 0;
+
+ if (av_strstart(token, "layer=", &token))
+ layer = 1;
+ else if (av_strstart(token, "demixing=", &token))
+ demixing = 1;
+ else if (av_strstart(token, "recon_gain=", &token))
+ recon_gain = 1;
+
+ av_dict_free(&dict);
+ ret = av_dict_parse_string(&dict, token, "=", ":", 0);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Error parsing audio element specification %s\n", token);
+ goto fail;
+ }
+
+ if (layer) {
+ AVIAMFLayer *audio_layer = av_iamf_audio_element_add_layer(audio_element);
+ if (!audio_layer) {
+ av_log(mux, AV_LOG_ERROR, "Error adding layer to stream group %d\n", stg->index);
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ av_opt_set_dict(audio_layer, &dict);
+ } else if (demixing || recon_gain) {
+ AVIAMFParamDefinition *param = demixing ? audio_element->demixing_info
+ : audio_element->recon_gain_info;
+ void *subblock = av_iamf_param_definition_get_subblock(param, 0);
+
+ av_opt_set_dict(param, &dict);
+ av_opt_set_dict(subblock, &dict);
+ }
+
+ // make sure that no entries are left in the dict
+ e = NULL;
+ if (e = av_dict_iterate(dict, e)) {
+ av_log(mux, AV_LOG_FATAL, "Unknown layer key %s.\n", e->key);
+ ret = AVERROR(EINVAL);
+ goto fail;
+ }
+ token = av_strtok(NULL, ",", &ptr);
+ }
+
+fail:
+ av_dict_free(&dict);
+ if (!ret && !audio_element->nb_layers) {
+ av_log(mux, AV_LOG_ERROR, "No layer in audio element specification\n");
+ ret = AVERROR(EINVAL);
+ }
+
+ return ret;
+}
+
+static int of_parse_iamf_submixes(Muxer *mux, AVStreamGroup *stg, char *ptr)
+{
+ AVFormatContext *oc = mux->fc;
+ AVIAMFMixPresentation *mix = stg->params.iamf_mix_presentation;
+ AVDictionary *dict = NULL;
+ const char *token;
+ char *submix_str = NULL;
+ int ret = 0;
+
+ /* process manually set submixes */
+ token = av_strtok(NULL, ",", &ptr);
+ while (token) {
+ AVIAMFSubmix *submix = NULL;
+ const char *subtoken;
+ char *subptr = NULL;
+
+ if (!av_strstart(token, "submix=", &token)) {
+ av_log(mux, AV_LOG_ERROR, "No submix in mix presentation specification \"%s\"\n", token);
+ goto fail;
+ }
+
+ submix_str = av_strdup(token);
+ if (!submix_str)
+ goto fail;
+
+ submix = av_iamf_mix_presentation_add_submix(mix);
+ if (!submix) {
+ av_log(mux, AV_LOG_ERROR, "Error adding submix to stream group %d\n", stg->index);
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ submix->output_mix_config =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN, 0, NULL);
+ if (!submix->output_mix_config) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ subptr = NULL;
+ subtoken = av_strtok(submix_str, "|", &subptr);
+ while (subtoken) {
+ const AVDictionaryEntry *e;
+ int element = 0, layout = 0;
+
+ if (av_strstart(subtoken, "element=", &subtoken))
+ element = 1;
+ else if (av_strstart(subtoken, "layout=", &subtoken))
+ layout = 1;
+
+ av_dict_free(&dict);
+ ret = av_dict_parse_string(&dict, subtoken, "=", ":", 0);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Error parsing submix specification \"%s\"\n", subtoken);
+ goto fail;
+ }
+
+ if (element) {
+ AVIAMFSubmixElement *submix_element;
+ int64_t idx = -1;
+
+ if (e = av_dict_get(dict, "stg", NULL, 0))
+ idx = strtol(e->value, NULL, 0);
+ av_dict_set(&dict, "stg", NULL, 0);
+ if (idx < 0 || idx >= oc->nb_stream_groups - 1 ||
+ oc->stream_groups[idx]->type != AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT) {
+ av_log(mux, AV_LOG_ERROR, "Invalid or missing stream group index in "
+ "submix element specification \"%s\"\n", subtoken);
+ ret = AVERROR(EINVAL);
+ goto fail;
+ }
+ submix_element = av_iamf_submix_add_element(submix);
+ if (!submix_element) {
+ av_log(mux, AV_LOG_ERROR, "Error adding element to submix\n");
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ submix_element->audio_element_id = oc->stream_groups[idx]->id;
+
+ submix_element->element_mix_config =
+ av_iamf_param_definition_alloc(AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN, 0, NULL);
+ if (!submix_element->element_mix_config)
+ ret = AVERROR(ENOMEM);
+ av_opt_set_dict2(submix_element, &dict, AV_OPT_SEARCH_CHILDREN);
+ } else if (layout) {
+ AVIAMFSubmixLayout *submix_layout = av_iamf_submix_add_layout(submix);
+ if (!submix_layout) {
+ av_log(mux, AV_LOG_ERROR, "Error adding layout to submix\n");
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ av_opt_set_dict(submix_layout, &dict);
+ } else
+ av_opt_set_dict2(submix, &dict, AV_OPT_SEARCH_CHILDREN);
+
+ if (ret < 0) {
+ goto fail;
+ }
+
+ // make sure that no entries are left in the dict
+ e = NULL;
+ while (e = av_dict_iterate(dict, e)) {
+ av_log(mux, AV_LOG_FATAL, "Unknown submix key %s.\n", e->key);
+ ret = AVERROR(EINVAL);
+ goto fail;
+ }
+ subtoken = av_strtok(NULL, "|", &subptr);
+ }
+ av_freep(&submix_str);
+
+ if (!submix->nb_elements) {
+ av_log(mux, AV_LOG_ERROR, "No audio elements in submix specification \"%s\"\n", token);
+ ret = AVERROR(EINVAL);
+ }
+ token = av_strtok(NULL, ",", &ptr);
+ }
+
+fail:
+ av_dict_free(&dict);
+ av_free(submix_str);
+
+ return ret;
+}
+
+static int of_parse_group_token(Muxer *mux, const char *token, char *ptr)
+{
+ AVFormatContext *oc = mux->fc;
+ AVStreamGroup *stg;
+ AVDictionary *dict = NULL, *tmp = NULL;
+ const AVDictionaryEntry *e;
+ const AVOption opts[] = {
+ { "type", "Set group type", offsetof(AVStreamGroup, type), AV_OPT_TYPE_INT,
+ { .i64 = 0 }, 0, INT_MAX, AV_OPT_FLAG_ENCODING_PARAM, "type" },
+ { "iamf_audio_element", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT }, .unit = "type" },
+ { "iamf_mix_presentation", NULL, 0, AV_OPT_TYPE_CONST,
+ { .i64 = AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION }, .unit = "type" },
+ { NULL },
+ };
+ const AVClass class = {
+ .class_name = "StreamGroupType",
+ .item_name = av_default_item_name,
+ .option = opts,
+ .version = LIBAVUTIL_VERSION_INT,
+ };
+ const AVClass *pclass = &class;
+ int type, ret;
+
+ ret = av_dict_parse_string(&dict, token, "=", ":", AV_DICT_MULTIKEY);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Error parsing group specification %s\n", token);
+ return ret;
+ }
+
+ // "type" is not a user settable AVOption in AVStreamGroup, so handle it here
+ e = av_dict_get(dict, "type", NULL, 0);
+ if (!e) {
+ av_log(mux, AV_LOG_ERROR, "No type specified for Stream Group in \"%s\"\n", token);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+
+ ret = av_opt_eval_int(&pclass, opts, e->value, &type);
+ if (!ret && type == AV_STREAM_GROUP_PARAMS_NONE)
+ ret = AVERROR(EINVAL);
+ if (ret < 0) {
+ av_log(mux, AV_LOG_ERROR, "Invalid group type \"%s\"\n", e->value);
+ goto end;
+ }
+
+ av_dict_copy(&tmp, dict, 0);
+ stg = avformat_stream_group_create(oc, type, &tmp);
+ if (!stg) {
+ ret = AVERROR(ENOMEM);
+ goto end;
+ }
+
+ e = NULL;
+ while (e = av_dict_get(dict, "st", e, 0)) {
+ int64_t idx = strtol(e->value, NULL, 0);
+ if (idx < 0 || idx >= oc->nb_streams) {
+ av_log(mux, AV_LOG_ERROR, "Invalid stream index %"PRId64"\n", idx);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+ ret = avformat_stream_group_add_stream(stg, oc->streams[idx]);
+ if (ret < 0)
+ goto end;
+ }
+ while (e = av_dict_get(dict, "stg", e, 0)) {
+ int64_t idx = strtol(e->value, NULL, 0);
+ if (idx < 0 || idx >= oc->nb_stream_groups - 1) {
+ av_log(mux, AV_LOG_ERROR, "Invalid stream group index %"PRId64"\n", idx);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+ for (unsigned i = 0; i < oc->stream_groups[idx]->nb_streams; i++) {
+ ret = avformat_stream_group_add_stream(stg, oc->stream_groups[idx]->streams[i]);
+ if (ret < 0)
+ goto end;
+ }
+ }
+
+ switch(type) {
+ case AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT:
+ ret = of_parse_iamf_audio_element_layers(mux, stg, ptr);
+ break;
+ case AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION:
+ ret = of_parse_iamf_submixes(mux, stg, ptr);
+ break;
+ default:
+ av_log(mux, AV_LOG_FATAL, "Unknown group type %d.\n", type);
+ ret = AVERROR(EINVAL);
+ break;
+ }
+
+ if (ret < 0)
+ goto end;
+
+ // make sure that nothing but "st" and "stg" entries are left in the dict
+ e = NULL;
+ av_dict_set(&tmp, "type", NULL, 0);
+ while (e = av_dict_iterate(tmp, e)) {
+ if (!strcmp(e->key, "st") || !strcmp(e->key, "stg"))
+ continue;
+
+ av_log(mux, AV_LOG_FATAL, "Unknown group key %s.\n", e->key);
+ ret = AVERROR(EINVAL);
+ goto end;
+ }
+
+ ret = 0;
+end:
+ av_dict_free(&dict);
+ av_dict_free(&tmp);
+
+ return ret;
+}
+
+static int of_add_groups(Muxer *mux, const OptionsContext *o)
+{
+ /* process manually set groups */
+ for (int i = 0; i < o->nb_stream_groups; i++) {
+ const char *token;
+ char *str, *ptr = NULL;
+ int ret = 0;
+
+ str = av_strdup(o->stream_groups[i].u.str);
+ if (!str)
+ return ret;
+
+ token = av_strtok(str, ",", &ptr);
+ if (token)
+ ret = of_parse_group_token(mux, token, ptr);
+
+ av_free(str);
+ if (ret < 0)
+ return ret;
+ }
+
+ return 0;
+}
+
static int of_add_programs(Muxer *mux, const OptionsContext *o)
{
AVFormatContext *oc = mux->fc;
@@ -2793,6 +3131,10 @@ int of_open(const OptionsContext *o, const char *filename, Scheduler *sch)
if (err < 0)
return err;
+ err = of_add_groups(mux, o);
+ if (err < 0)
+ return err;
+
err = of_add_programs(mux, o);
if (err < 0)
return err;
diff --git a/fftools/ffmpeg_opt.c b/fftools/ffmpeg_opt.c
index 6177a96a4e..915f8e3ea0 100644
--- a/fftools/ffmpeg_opt.c
+++ b/fftools/ffmpeg_opt.c
@@ -1493,6 +1493,8 @@ const OptionDef options[] = {
"add metadata", "string=string" },
{ "program", HAS_ARG | OPT_STRING | OPT_SPEC | OPT_OUTPUT, { .off = OFFSET(program) },
"add program with specified streams", "title=string:st=number..." },
+ { "stream_group", HAS_ARG | OPT_STRING | OPT_SPEC | OPT_OUTPUT, { .off = OFFSET(stream_groups) },
+ "add stream group with specified streams and group type-specific arguments", "id=number:st=number..." },
{ "dframes", HAS_ARG | OPT_PERFILE | OPT_EXPERT |
OPT_OUTPUT, { .func_arg = opt_data_frames },
"set the number of data frames to output", "number" },
--
2.43.0
_______________________________________________
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] 12+ messages in thread
* [FFmpeg-devel] [PATCH 4/8] avcodec/packet: add IAMF Parameters side data types
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
` (2 preceding siblings ...)
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 3/8] ffmpeg: add support for muxing AVStreamGroups James Almer
@ 2023-12-14 20:14 ` James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 5/8] avcodec/get_bits: add get_leb() James Almer
` (3 subsequent siblings)
7 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Signed-off-by: James Almer <jamrial@gmail.com>
---
libavcodec/avpacket.c | 3 +++
libavcodec/packet.h | 24 ++++++++++++++++++++++++
2 files changed, 27 insertions(+)
diff --git a/libavcodec/avpacket.c b/libavcodec/avpacket.c
index e29725c2d2..0f8c9b77ae 100644
--- a/libavcodec/avpacket.c
+++ b/libavcodec/avpacket.c
@@ -301,6 +301,9 @@ const char *av_packet_side_data_name(enum AVPacketSideDataType type)
case AV_PKT_DATA_DOVI_CONF: return "DOVI configuration record";
case AV_PKT_DATA_S12M_TIMECODE: return "SMPTE ST 12-1:2014 timecode";
case AV_PKT_DATA_DYNAMIC_HDR10_PLUS: return "HDR10+ Dynamic Metadata (SMPTE 2094-40)";
+ case AV_PKT_DATA_IAMF_MIX_GAIN_PARAM: return "IAMF Mix Gain Parameter Data";
+ case AV_PKT_DATA_IAMF_DEMIXING_INFO_PARAM: return "IAMF Demixing Info Parameter Data";
+ case AV_PKT_DATA_IAMF_RECON_GAIN_INFO_PARAM: return "IAMF Recon Gain Info Parameter Data";
}
return NULL;
}
diff --git a/libavcodec/packet.h b/libavcodec/packet.h
index b19409b719..2c57d262c6 100644
--- a/libavcodec/packet.h
+++ b/libavcodec/packet.h
@@ -299,6 +299,30 @@ enum AVPacketSideDataType {
*/
AV_PKT_DATA_DYNAMIC_HDR10_PLUS,
+ /**
+ * IAMF Mix Gain Parameter Data associated with the audio frame. This metadata
+ * is in the form of the AVIAMFParamDefinition struct and contains information
+ * defined in sections 3.6.1 and 3.8.1 of the Immersive Audio Model and
+ * Formats standard.
+ */
+ AV_PKT_DATA_IAMF_MIX_GAIN_PARAM,
+
+ /**
+ * IAMF Demixing Info Parameter Data associated with the audio frame. This
+ * metadata is in the form of the AVIAMFParamDefinition struct and contains
+ * information defined in sections 3.6.1 and 3.8.2 of the Immersive Audio Model
+ * and Formats standard.
+ */
+ AV_PKT_DATA_IAMF_DEMIXING_INFO_PARAM,
+
+ /**
+ * IAMF Recon Gain Info Parameter Data associated with the audio frame. This
+ * metadata is in the form of the AVIAMFParamDefinition struct and contains
+ * information defined in sections 3.6.1 and 3.8.3 of the Immersive Audio Model
+ * and Formats standard.
+ */
+ AV_PKT_DATA_IAMF_RECON_GAIN_INFO_PARAM,
+
/**
* The number of side data types.
* This is not part of the public API/ABI in the sense that it may
--
2.43.0
_______________________________________________
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] 12+ messages in thread* [FFmpeg-devel] [PATCH 5/8] avcodec/get_bits: add get_leb()
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
` (3 preceding siblings ...)
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 4/8] avcodec/packet: add IAMF Parameters side data types James Almer
@ 2023-12-14 20:14 ` James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 6/8] avformat/aviobuf: add ffio_read_leb() and ffio_write_leb() James Almer
` (2 subsequent siblings)
7 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Signed-off-by: James Almer <jamrial@gmail.com>
---
libavcodec/bitstream.h | 2 ++
libavcodec/bitstream_template.h | 23 +++++++++++++++++++++++
libavcodec/get_bits.h | 24 ++++++++++++++++++++++++
3 files changed, 49 insertions(+)
diff --git a/libavcodec/bitstream.h b/libavcodec/bitstream.h
index 35b7873b9c..17f8a5da83 100644
--- a/libavcodec/bitstream.h
+++ b/libavcodec/bitstream.h
@@ -103,6 +103,7 @@
# define bits_apply_sign bits_apply_sign_le
# define bits_read_vlc bits_read_vlc_le
# define bits_read_vlc_multi bits_read_vlc_multi_le
+# define bits_read_leb bits_read_leb_le
#elif defined(BITS_DEFAULT_BE)
@@ -132,6 +133,7 @@
# define bits_apply_sign bits_apply_sign_be
# define bits_read_vlc bits_read_vlc_be
# define bits_read_vlc_multi bits_read_vlc_multi_be
+# define bits_read_leb bits_read_leb_be
#endif
diff --git a/libavcodec/bitstream_template.h b/libavcodec/bitstream_template.h
index 4f3d07275f..4c7101632f 100644
--- a/libavcodec/bitstream_template.h
+++ b/libavcodec/bitstream_template.h
@@ -562,6 +562,29 @@ static inline int BS_FUNC(read_vlc_multi)(BSCTX *bc, uint8_t dst[8],
return ret;
}
+/**
+ * Read a unsigned integer coded as a variable number of up to eight
+ * little-endian bytes, where the MSB in a byte signals another byte
+ * must be read.
+ * Values > UINT_MAX are truncated, but all coded bits are read.
+ */
+static inline unsigned BS_FUNC(read_leb)(BSCTX *bc) {
+ int more, i = 0;
+ unsigned leb = 0;
+
+ do {
+ int byte = BS_FUNC(read)(bc, 8);
+ unsigned bits = byte & 0x7f;
+ more = byte & 0x80;
+ if (i <= 4)
+ leb |= bits << (i * 7);
+ if (++i == 8)
+ break;
+ } while (more);
+
+ return leb;
+}
+
#undef BSCTX
#undef BS_FUNC
#undef BS_JOIN3
diff --git a/libavcodec/get_bits.h b/libavcodec/get_bits.h
index cfcf97c021..9e19d2a439 100644
--- a/libavcodec/get_bits.h
+++ b/libavcodec/get_bits.h
@@ -94,6 +94,7 @@ typedef BitstreamContext GetBitContext;
#define align_get_bits bits_align
#define get_vlc2 bits_read_vlc
#define get_vlc_multi bits_read_vlc_multi
+#define get_leb bits_read_leb
#define init_get_bits8_le(s, buffer, byte_size) bits_init8_le((BitstreamContextLE*)s, buffer, byte_size)
#define get_bits_le(s, n) bits_read_le((BitstreamContextLE*)s, n)
@@ -710,6 +711,29 @@ static inline int skip_1stop_8data_bits(GetBitContext *gb)
return 0;
}
+/**
+ * Read a unsigned integer coded as a variable number of up to eight
+ * little-endian bytes, where the MSB in a byte signals another byte
+ * must be read.
+ * All coded bits are read, but values > UINT_MAX are truncated.
+ */
+static inline unsigned get_leb(GetBitContext *s) {
+ int more, i = 0;
+ unsigned leb = 0;
+
+ do {
+ int byte = get_bits(s, 8);
+ unsigned bits = byte & 0x7f;
+ more = byte & 0x80;
+ if (i <= 4)
+ leb |= bits << (i * 7);
+ if (++i == 8)
+ break;
+ } while (more);
+
+ return leb;
+}
+
#endif // CACHED_BITSTREAM_READER
#endif /* AVCODEC_GET_BITS_H */
--
2.43.0
_______________________________________________
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] 12+ messages in thread* [FFmpeg-devel] [PATCH 6/8] avformat/aviobuf: add ffio_read_leb() and ffio_write_leb()
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
` (4 preceding siblings ...)
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 5/8] avcodec/get_bits: add get_leb() James Almer
@ 2023-12-14 20:14 ` James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 7/8] avformat: Immersive Audio Model and Formats demuxer James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 8/8] avformat: Immersive Audio Model and Formats muxer James Almer
7 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Signed-off-by: James Almer <jamrial@gmail.com>
---
libavformat/avio_internal.h | 10 ++++++++++
libavformat/aviobuf.c | 33 +++++++++++++++++++++++++++++++++
2 files changed, 43 insertions(+)
diff --git a/libavformat/avio_internal.h b/libavformat/avio_internal.h
index bd58499b64..f2e4ff30cb 100644
--- a/libavformat/avio_internal.h
+++ b/libavformat/avio_internal.h
@@ -146,6 +146,16 @@ int ffio_rewind_with_probe_data(AVIOContext *s, unsigned char **buf, int buf_siz
uint64_t ffio_read_varlen(AVIOContext *bc);
+/**
+ * Read a unsigned integer coded as a variable number of up to eight
+ * little-endian bytes, where the MSB in a byte signals another byte
+ * must be read.
+ * All coded bytes are read, but values > UINT_MAX are truncated.
+ */
+unsigned int ffio_read_leb(AVIOContext *s);
+
+void ffio_write_leb(AVIOContext *s, unsigned val);
+
/**
* Read size bytes from AVIOContext into buf.
* Check that exactly size bytes have been read.
diff --git a/libavformat/aviobuf.c b/libavformat/aviobuf.c
index 2899c75521..5a329ce465 100644
--- a/libavformat/aviobuf.c
+++ b/libavformat/aviobuf.c
@@ -971,6 +971,39 @@ uint64_t ffio_read_varlen(AVIOContext *bc){
return val;
}
+unsigned int ffio_read_leb(AVIOContext *s) {
+ int more, i = 0;
+ unsigned leb = 0;
+
+ do {
+ int byte = avio_r8(s);
+ unsigned bits = byte & 0x7f;
+ more = byte & 0x80;
+ if (i <= 4)
+ leb |= bits << (i * 7);
+ if (++i == 8)
+ break;
+ } while (more);
+
+ return leb;
+}
+
+void ffio_write_leb(AVIOContext *s, unsigned val)
+{
+ int len;
+ uint8_t byte;
+
+ len = (av_log2(val) + 7) / 7;
+
+ for (int i = 0; i < len; i++) {
+ byte = val >> (7 * i) & 0x7f;
+ if (i < len - 1)
+ byte |= 0x80;
+
+ avio_w8(s, byte);
+ }
+}
+
int ffio_fdopen(AVIOContext **s, URLContext *h)
{
uint8_t *buffer = NULL;
--
2.43.0
_______________________________________________
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] 12+ messages in thread* [FFmpeg-devel] [PATCH 7/8] avformat: Immersive Audio Model and Formats demuxer
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
` (5 preceding siblings ...)
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 6/8] avformat/aviobuf: add ffio_read_leb() and ffio_write_leb() James Almer
@ 2023-12-14 20:14 ` James Almer
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 8/8] avformat: Immersive Audio Model and Formats muxer James Almer
7 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Signed-off-by: James Almer <jamrial@gmail.com>
---
libavformat/Makefile | 1 +
libavformat/allformats.c | 1 +
libavformat/iamf.c | 125 +++++
libavformat/iamf.h | 163 ++++++
libavformat/iamf_parse.c | 1106 ++++++++++++++++++++++++++++++++++++++
libavformat/iamf_parse.h | 38 ++
libavformat/iamfdec.c | 503 +++++++++++++++++
7 files changed, 1937 insertions(+)
create mode 100644 libavformat/iamf.c
create mode 100644 libavformat/iamf.h
create mode 100644 libavformat/iamf_parse.c
create mode 100644 libavformat/iamf_parse.h
create mode 100644 libavformat/iamfdec.c
diff --git a/libavformat/Makefile b/libavformat/Makefile
index 2db83aff81..f23c22792b 100644
--- a/libavformat/Makefile
+++ b/libavformat/Makefile
@@ -258,6 +258,7 @@ OBJS-$(CONFIG_EVC_MUXER) += rawenc.o
OBJS-$(CONFIG_HLS_DEMUXER) += hls.o hls_sample_encryption.o
OBJS-$(CONFIG_HLS_MUXER) += hlsenc.o hlsplaylist.o avc.o
OBJS-$(CONFIG_HNM_DEMUXER) += hnm.o
+OBJS-$(CONFIG_IAMF_DEMUXER) += iamfdec.o iamf_parse.o iamf.o
OBJS-$(CONFIG_ICO_DEMUXER) += icodec.o
OBJS-$(CONFIG_ICO_MUXER) += icoenc.o
OBJS-$(CONFIG_IDCIN_DEMUXER) += idcin.o
diff --git a/libavformat/allformats.c b/libavformat/allformats.c
index c8bb4e3866..6e520b78a6 100644
--- a/libavformat/allformats.c
+++ b/libavformat/allformats.c
@@ -212,6 +212,7 @@ extern const FFOutputFormat ff_hevc_muxer;
extern const AVInputFormat ff_hls_demuxer;
extern const FFOutputFormat ff_hls_muxer;
extern const AVInputFormat ff_hnm_demuxer;
+extern const AVInputFormat ff_iamf_demuxer;
extern const AVInputFormat ff_ico_demuxer;
extern const FFOutputFormat ff_ico_muxer;
extern const AVInputFormat ff_idcin_demuxer;
diff --git a/libavformat/iamf.c b/libavformat/iamf.c
new file mode 100644
index 0000000000..5de70dc082
--- /dev/null
+++ b/libavformat/iamf.c
@@ -0,0 +1,125 @@
+/*
+ * Immersive Audio Model and Formats common helpers and structs
+ * Copyright (c) 2023 James Almer <jamrial@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 "libavutil/channel_layout.h"
+#include "libavutil/iamf.h"
+#include "libavutil/mem.h"
+#include "iamf.h"
+
+const AVChannelLayout ff_iamf_scalable_ch_layouts[10] = {
+ AV_CHANNEL_LAYOUT_MONO,
+ AV_CHANNEL_LAYOUT_STEREO,
+ // "Loudspeaker configuration for Sound System B"
+ AV_CHANNEL_LAYOUT_5POINT1_BACK,
+ // "Loudspeaker configuration for Sound System C"
+ AV_CHANNEL_LAYOUT_5POINT1POINT2_BACK,
+ // "Loudspeaker configuration for Sound System D"
+ AV_CHANNEL_LAYOUT_5POINT1POINT4_BACK,
+ // "Loudspeaker configuration for Sound System I"
+ AV_CHANNEL_LAYOUT_7POINT1,
+ // "Loudspeaker configuration for Sound System I" + Ltf + Rtf
+ AV_CHANNEL_LAYOUT_7POINT1POINT2,
+ // "Loudspeaker configuration for Sound System J"
+ AV_CHANNEL_LAYOUT_7POINT1POINT4_BACK,
+ // Front subset of "Loudspeaker configuration for Sound System J"
+ AV_CHANNEL_LAYOUT_3POINT1POINT2,
+ // Binaural
+ AV_CHANNEL_LAYOUT_STEREO,
+};
+
+const struct IAMFSoundSystemMap ff_iamf_sound_system_map[13] = {
+ { SOUND_SYSTEM_A_0_2_0, AV_CHANNEL_LAYOUT_STEREO },
+ { SOUND_SYSTEM_B_0_5_0, AV_CHANNEL_LAYOUT_5POINT1_BACK },
+ { SOUND_SYSTEM_C_2_5_0, AV_CHANNEL_LAYOUT_5POINT1POINT2_BACK },
+ { SOUND_SYSTEM_D_4_5_0, AV_CHANNEL_LAYOUT_5POINT1POINT4_BACK },
+ { SOUND_SYSTEM_E_4_5_1,
+ {
+ .nb_channels = 11,
+ .order = AV_CHANNEL_ORDER_NATIVE,
+ .u.mask = AV_CH_LAYOUT_5POINT1POINT4_BACK | AV_CH_BOTTOM_FRONT_CENTER,
+ },
+ },
+ { SOUND_SYSTEM_F_3_7_0, AV_CHANNEL_LAYOUT_7POINT2POINT3 },
+ { SOUND_SYSTEM_G_4_9_0, AV_CHANNEL_LAYOUT_9POINT1POINT4_BACK },
+ { SOUND_SYSTEM_H_9_10_3, AV_CHANNEL_LAYOUT_22POINT2 },
+ { SOUND_SYSTEM_I_0_7_0, AV_CHANNEL_LAYOUT_7POINT1 },
+ { SOUND_SYSTEM_J_4_7_0, AV_CHANNEL_LAYOUT_7POINT1POINT4_BACK },
+ { SOUND_SYSTEM_10_2_7_0, AV_CHANNEL_LAYOUT_7POINT1POINT2 },
+ { SOUND_SYSTEM_11_2_3_0, AV_CHANNEL_LAYOUT_3POINT1POINT2 },
+ { SOUND_SYSTEM_12_0_1_0, AV_CHANNEL_LAYOUT_MONO },
+};
+
+void ff_iamf_free_audio_element(IAMFAudioElement **paudio_element)
+{
+ IAMFAudioElement *audio_element = *paudio_element;
+
+ if (!audio_element)
+ return;
+
+ for (int i = 0; i < audio_element->nb_substreams; i++)
+ avcodec_parameters_free(&audio_element->substreams[i].codecpar);
+ av_free(audio_element->substreams);
+ av_free(audio_element->layers);
+ av_iamf_audio_element_free(&audio_element->element);
+ av_freep(paudio_element);
+}
+
+void ff_iamf_free_mix_presentation(IAMFMixPresentation **pmix_presentation)
+{
+ IAMFMixPresentation *mix_presentation = *pmix_presentation;
+
+ if (!mix_presentation)
+ return;
+
+ for (int i = 0; i < mix_presentation->count_label; i++)
+ av_free(mix_presentation->language_label[i]);
+ av_free(mix_presentation->language_label);
+ av_iamf_mix_presentation_free(&mix_presentation->mix);
+ av_freep(pmix_presentation);
+}
+
+void ff_iamf_uninit_context(IAMFContext *c)
+{
+ if (!c)
+ return;
+
+ for (int i = 0; i < c->nb_codec_configs; i++) {
+ av_free(c->codec_configs[i]->extradata);
+ av_free(c->codec_configs[i]);
+ }
+ av_freep(&c->codec_configs);
+ c->nb_codec_configs = 0;
+
+ for (int i = 0; i < c->nb_audio_elements; i++)
+ ff_iamf_free_audio_element(&c->audio_elements[i]);
+ av_freep(&c->audio_elements);
+ c->nb_audio_elements = 0;
+
+ for (int i = 0; i < c->nb_mix_presentations; i++)
+ ff_iamf_free_mix_presentation(&c->mix_presentations[i]);
+ av_freep(&c->mix_presentations);
+ c->nb_mix_presentations = 0;
+
+ for (int i = 0; i < c->nb_param_definitions; i++)
+ av_free(c->param_definitions[i]);
+ av_freep(&c->param_definitions);
+ c->nb_param_definitions = 0;
+}
diff --git a/libavformat/iamf.h b/libavformat/iamf.h
new file mode 100644
index 0000000000..ce94cb5bc4
--- /dev/null
+++ b/libavformat/iamf.h
@@ -0,0 +1,163 @@
+/*
+ * Immersive Audio Model and Formats common helpers and structs
+ * Copyright (c) 2023 James Almer <jamrial@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
+ */
+
+#ifndef AVFORMAT_IAMF_H
+#define AVFORMAT_IAMF_H
+
+#include <stdint.h>
+
+#include "libavutil/channel_layout.h"
+#include "libavutil/iamf.h"
+#include "libavcodec/codec_id.h"
+#include "libavcodec/codec_par.h"
+#include "avformat.h"
+
+#define MAX_IAMF_OBU_HEADER_SIZE (1 + 8 * 3)
+
+// OBU types (section 3.2).
+enum IAMF_OBU_Type {
+ IAMF_OBU_IA_CODEC_CONFIG = 0,
+ IAMF_OBU_IA_AUDIO_ELEMENT = 1,
+ IAMF_OBU_IA_MIX_PRESENTATION = 2,
+ IAMF_OBU_IA_PARAMETER_BLOCK = 3,
+ IAMF_OBU_IA_TEMPORAL_DELIMITER = 4,
+ IAMF_OBU_IA_AUDIO_FRAME = 5,
+ IAMF_OBU_IA_AUDIO_FRAME_ID0 = 6,
+ IAMF_OBU_IA_AUDIO_FRAME_ID1 = 7,
+ IAMF_OBU_IA_AUDIO_FRAME_ID2 = 8,
+ IAMF_OBU_IA_AUDIO_FRAME_ID3 = 9,
+ IAMF_OBU_IA_AUDIO_FRAME_ID4 = 10,
+ IAMF_OBU_IA_AUDIO_FRAME_ID5 = 11,
+ IAMF_OBU_IA_AUDIO_FRAME_ID6 = 12,
+ IAMF_OBU_IA_AUDIO_FRAME_ID7 = 13,
+ IAMF_OBU_IA_AUDIO_FRAME_ID8 = 14,
+ IAMF_OBU_IA_AUDIO_FRAME_ID9 = 15,
+ IAMF_OBU_IA_AUDIO_FRAME_ID10 = 16,
+ IAMF_OBU_IA_AUDIO_FRAME_ID11 = 17,
+ IAMF_OBU_IA_AUDIO_FRAME_ID12 = 18,
+ IAMF_OBU_IA_AUDIO_FRAME_ID13 = 19,
+ IAMF_OBU_IA_AUDIO_FRAME_ID14 = 20,
+ IAMF_OBU_IA_AUDIO_FRAME_ID15 = 21,
+ IAMF_OBU_IA_AUDIO_FRAME_ID16 = 22,
+ IAMF_OBU_IA_AUDIO_FRAME_ID17 = 23,
+ // 24~30 reserved.
+ IAMF_OBU_IA_SEQUENCE_HEADER = 31,
+};
+
+typedef struct IAMFCodecConfig {
+ unsigned codec_config_id;
+ enum AVCodecID codec_id;
+ uint32_t codec_tag;
+ unsigned nb_samples;
+ int seek_preroll;
+ int sample_rate;
+ int extradata_size;
+ uint8_t *extradata;
+} IAMFCodecConfig;
+
+typedef struct IAMFLayer {
+ unsigned int substream_count;
+ unsigned int coupled_substream_count;
+} IAMFLayer;
+
+typedef struct IAMFSubStream {
+ unsigned int audio_substream_id;
+
+ // demux
+ AVCodecParameters *codecpar;
+} IAMFSubStream;
+
+typedef struct IAMFAudioElement {
+ AVIAMFAudioElement *element;
+ unsigned int audio_element_id;
+
+ IAMFSubStream *substreams;
+ unsigned int nb_substreams;
+
+ unsigned int codec_config_id;
+
+ // mux
+ IAMFLayer *layers;
+ unsigned int nb_layers;
+} IAMFAudioElement;
+
+typedef struct IAMFMixPresentation {
+ AVIAMFMixPresentation *mix;
+ unsigned int mix_presentation_id;
+
+ // demux
+ unsigned int count_label;
+ char **language_label;
+} IAMFMixPresentation;
+
+typedef struct IAMFParamDefinition {
+ const IAMFAudioElement *audio_element;
+ AVIAMFParamDefinition *param;
+ int mode;
+ size_t param_size;
+} IAMFParamDefinition;
+
+typedef struct IAMFContext {
+ IAMFCodecConfig **codec_configs;
+ int nb_codec_configs;
+ IAMFAudioElement **audio_elements;
+ int nb_audio_elements;
+ IAMFMixPresentation **mix_presentations;
+ int nb_mix_presentations;
+ IAMFParamDefinition **param_definitions;
+ int nb_param_definitions;
+} IAMFContext;
+
+enum IAMF_Anchor_Element {
+ IAMF_ANCHOR_ELEMENT_UNKNWONW,
+ IAMF_ANCHOR_ELEMENT_DIALOGUE,
+ IAMF_ANCHOR_ELEMENT_ALBUM,
+};
+
+enum IAMF_Sound_System {
+ SOUND_SYSTEM_A_0_2_0 = 0, // "Loudspeaker configuration for Sound System A"
+ SOUND_SYSTEM_B_0_5_0 = 1, // "Loudspeaker configuration for Sound System B"
+ SOUND_SYSTEM_C_2_5_0 = 2, // "Loudspeaker configuration for Sound System C"
+ SOUND_SYSTEM_D_4_5_0 = 3, // "Loudspeaker configuration for Sound System D"
+ SOUND_SYSTEM_E_4_5_1 = 4, // "Loudspeaker configuration for Sound System E"
+ SOUND_SYSTEM_F_3_7_0 = 5, // "Loudspeaker configuration for Sound System F"
+ SOUND_SYSTEM_G_4_9_0 = 6, // "Loudspeaker configuration for Sound System G"
+ SOUND_SYSTEM_H_9_10_3 = 7, // "Loudspeaker configuration for Sound System H"
+ SOUND_SYSTEM_I_0_7_0 = 8, // "Loudspeaker configuration for Sound System I"
+ SOUND_SYSTEM_J_4_7_0 = 9, // "Loudspeaker configuration for Sound System J"
+ SOUND_SYSTEM_10_2_7_0 = 10, // "Loudspeaker configuration for Sound System I" + Ltf + Rtf
+ SOUND_SYSTEM_11_2_3_0 = 11, // Front subset of "Loudspeaker configuration for Sound System J"
+ SOUND_SYSTEM_12_0_1_0 = 12, // Mono
+};
+
+struct IAMFSoundSystemMap {
+ enum IAMF_Sound_System id;
+ AVChannelLayout layout;
+};
+
+extern const AVChannelLayout ff_iamf_scalable_ch_layouts[10];
+extern const struct IAMFSoundSystemMap ff_iamf_sound_system_map[13];
+
+void ff_iamf_free_audio_element(IAMFAudioElement **paudio_element);
+void ff_iamf_free_mix_presentation(IAMFMixPresentation **pmix_presentation);
+void ff_iamf_uninit_context(IAMFContext *c);
+
+#endif /* AVFORMAT_IAMF_H */
diff --git a/libavformat/iamf_parse.c b/libavformat/iamf_parse.c
new file mode 100644
index 0000000000..60305743f9
--- /dev/null
+++ b/libavformat/iamf_parse.c
@@ -0,0 +1,1106 @@
+/*
+ * Immersive Audio Model and Formats parsing
+ * Copyright (c) 2023 James Almer <jamrial@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 "libavutil/avassert.h"
+#include "libavutil/common.h"
+#include "libavutil/iamf.h"
+#include "libavutil/intreadwrite.h"
+#include "libavutil/log.h"
+#include "libavcodec/get_bits.h"
+#include "libavcodec/flac.h"
+#include "libavcodec/mpeg4audio.h"
+#include "libavcodec/put_bits.h"
+#include "avio_internal.h"
+#include "iamf_parse.h"
+#include "isom.h"
+
+static int opus_decoder_config(IAMFCodecConfig *codec_config,
+ AVIOContext *pb, int len)
+{
+ int left = len - avio_tell(pb);
+
+ if (left < 11)
+ return AVERROR_INVALIDDATA;
+
+ codec_config->extradata = av_malloc(left + 8);
+ if (!codec_config->extradata)
+ return AVERROR(ENOMEM);
+
+ AV_WB32(codec_config->extradata, MKBETAG('O','p','u','s'));
+ AV_WB32(codec_config->extradata + 4, MKBETAG('H','e','a','d'));
+ codec_config->extradata_size = avio_read(pb, codec_config->extradata + 8, left);
+ if (codec_config->extradata_size < left)
+ return AVERROR_INVALIDDATA;
+
+ codec_config->extradata_size += 8;
+ codec_config->sample_rate = 48000;
+
+ return 0;
+}
+
+static int aac_decoder_config(IAMFCodecConfig *codec_config,
+ AVIOContext *pb, int len, void *logctx)
+{
+ MPEG4AudioConfig cfg = { 0 };
+ int object_type_id, codec_id, stream_type;
+ int ret, tag, left;
+
+ tag = avio_r8(pb);
+ if (tag != MP4DecConfigDescrTag)
+ return AVERROR_INVALIDDATA;
+
+ object_type_id = avio_r8(pb);
+ if (object_type_id != 0x40)
+ return AVERROR_INVALIDDATA;
+
+ stream_type = avio_r8(pb);
+ if (((stream_type >> 2) != 5) || ((stream_type >> 1) & 1))
+ return AVERROR_INVALIDDATA;
+
+ avio_skip(pb, 3); // buffer size db
+ avio_skip(pb, 4); // rc_max_rate
+ avio_skip(pb, 4); // avg bitrate
+
+ codec_id = ff_codec_get_id(ff_mp4_obj_type, object_type_id);
+ if (codec_id && codec_id != codec_config->codec_id)
+ return AVERROR_INVALIDDATA;
+
+ tag = avio_r8(pb);
+ if (tag != MP4DecSpecificDescrTag)
+ return AVERROR_INVALIDDATA;
+
+ left = len - avio_tell(pb);
+ if (left <= 0)
+ return AVERROR_INVALIDDATA;
+
+ codec_config->extradata = av_malloc(left);
+ if (!codec_config->extradata)
+ return AVERROR(ENOMEM);
+
+ codec_config->extradata_size = avio_read(pb, codec_config->extradata, left);
+ if (codec_config->extradata_size < left)
+ return AVERROR_INVALIDDATA;
+
+ ret = avpriv_mpeg4audio_get_config2(&cfg, codec_config->extradata,
+ codec_config->extradata_size, 1, logctx);
+ if (ret < 0)
+ return ret;
+
+ codec_config->sample_rate = cfg.sample_rate;
+
+ return 0;
+}
+
+static int flac_decoder_config(IAMFCodecConfig *codec_config,
+ AVIOContext *pb, int len)
+{
+ int left;
+
+ avio_skip(pb, 4); // METADATA_BLOCK_HEADER
+
+ left = len - avio_tell(pb);
+ if (left < FLAC_STREAMINFO_SIZE)
+ return AVERROR_INVALIDDATA;
+
+ codec_config->extradata = av_malloc(left);
+ if (!codec_config->extradata)
+ return AVERROR(ENOMEM);
+
+ codec_config->extradata_size = avio_read(pb, codec_config->extradata, left);
+ if (codec_config->extradata_size < left)
+ return AVERROR_INVALIDDATA;
+
+ codec_config->sample_rate = AV_RB24(codec_config->extradata + 10) >> 4;
+
+ return 0;
+}
+
+static int ipcm_decoder_config(IAMFCodecConfig *codec_config,
+ AVIOContext *pb, int len)
+{
+ static const enum AVSampleFormat sample_fmt[2][3] = {
+ { AV_CODEC_ID_PCM_S16BE, AV_CODEC_ID_PCM_S24BE, AV_CODEC_ID_PCM_S32BE },
+ { AV_CODEC_ID_PCM_S16LE, AV_CODEC_ID_PCM_S24LE, AV_CODEC_ID_PCM_S32LE },
+ };
+ int sample_format = avio_r8(pb); // 0 = BE, 1 = LE
+ int sample_size = (avio_r8(pb) / 8 - 2); // 16, 24, 32
+ if (sample_format > 1 || sample_size > 2)
+ return AVERROR_INVALIDDATA;
+
+ codec_config->codec_id = sample_fmt[sample_format][sample_size];
+ codec_config->sample_rate = avio_rb32(pb);
+
+ if (len - avio_tell(pb))
+ return AVERROR_INVALIDDATA;
+
+ return 0;
+}
+
+static int codec_config_obu(void *s, IAMFContext *c, AVIOContext *pb, int len)
+{
+ IAMFCodecConfig **tmp, *codec_config = NULL;
+ FFIOContext b;
+ AVIOContext *pbc;
+ uint8_t *buf;
+ enum AVCodecID avcodec_id;
+ unsigned codec_config_id, nb_samples, codec_id;
+ int16_t seek_preroll;
+ int ret;
+
+ buf = av_malloc(len);
+ if (!buf)
+ return AVERROR(ENOMEM);
+
+ ret = avio_read(pb, buf, len);
+ if (ret != len) {
+ if (ret >= 0)
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ ffio_init_context(&b, buf, len, 0, NULL, NULL, NULL, NULL);
+ pbc = &b.pub;
+
+ codec_config_id = ffio_read_leb(pbc);
+ codec_id = avio_rb32(pbc);
+ nb_samples = ffio_read_leb(pbc);
+ seek_preroll = avio_rb16(pbc);
+
+ switch(codec_id) {
+ case MKBETAG('O','p','u','s'):
+ avcodec_id = AV_CODEC_ID_OPUS;
+ break;
+ case MKBETAG('m','p','4','a'):
+ avcodec_id = AV_CODEC_ID_AAC;
+ break;
+ case MKBETAG('f','L','a','C'):
+ avcodec_id = AV_CODEC_ID_FLAC;
+ break;
+ default:
+ avcodec_id = AV_CODEC_ID_NONE;
+ break;
+ }
+
+ for (int i = 0; i < c->nb_codec_configs; i++)
+ if (c->codec_configs[i]->codec_config_id == codec_config_id) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ tmp = av_realloc_array(c->codec_configs, c->nb_codec_configs + 1, sizeof(*c->codec_configs));
+ if (!tmp) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ c->codec_configs = tmp;
+
+ codec_config = av_mallocz(sizeof(*codec_config));
+ if (!codec_config) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ codec_config->codec_config_id = codec_config_id;
+ codec_config->codec_id = avcodec_id;
+ codec_config->nb_samples = nb_samples;
+ codec_config->seek_preroll = seek_preroll;
+
+ switch(codec_id) {
+ case MKBETAG('O','p','u','s'):
+ ret = opus_decoder_config(codec_config, pbc, len);
+ break;
+ case MKBETAG('m','p','4','a'):
+ ret = aac_decoder_config(codec_config, pbc, len, s);
+ break;
+ case MKBETAG('f','L','a','C'):
+ ret = flac_decoder_config(codec_config, pbc, len);
+ break;
+ case MKBETAG('i','p','c','m'):
+ ret = ipcm_decoder_config(codec_config, pbc, len);
+ break;
+ default:
+ break;
+ }
+ if (ret < 0)
+ goto fail;
+
+ c->codec_configs[c->nb_codec_configs++] = codec_config;
+
+ len -= avio_tell(pbc);
+ if (len)
+ av_log(s, AV_LOG_WARNING, "Underread in codec_config_obu. %d bytes left at the end\n", len);
+
+ ret = 0;
+fail:
+ av_free(buf);
+ if (ret < 0) {
+ if (codec_config)
+ av_free(codec_config->extradata);
+ av_free(codec_config);
+ }
+ return ret;
+}
+
+static int update_extradata(AVCodecParameters *codecpar)
+{
+ GetBitContext gb;
+ PutBitContext pb;
+ int ret;
+
+ switch(codecpar->codec_id) {
+ case AV_CODEC_ID_OPUS:
+ AV_WB8(codecpar->extradata + 9, codecpar->ch_layout.nb_channels);
+ break;
+ case AV_CODEC_ID_AAC: {
+ uint8_t buf[5];
+
+ init_put_bits(&pb, buf, sizeof(buf));
+ ret = init_get_bits8(&gb, codecpar->extradata, codecpar->extradata_size);
+ if (ret < 0)
+ return ret;
+
+ ret = get_bits(&gb, 5);
+ put_bits(&pb, 5, ret);
+ if (ret == AOT_ESCAPE) // violates section 3.11.2, but better check for it
+ put_bits(&pb, 6, get_bits(&gb, 6));
+ ret = get_bits(&gb, 4);
+ put_bits(&pb, 4, ret);
+ if (ret == 0x0f)
+ put_bits(&pb, 24, get_bits(&gb, 24));
+
+ skip_bits(&gb, 4);
+ put_bits(&pb, 4, codecpar->ch_layout.nb_channels); // set channel config
+ ret = put_bits_left(&pb);
+ put_bits(&pb, ret, get_bits(&gb, ret));
+ flush_put_bits(&pb);
+
+ memcpy(codecpar->extradata, buf, sizeof(buf));
+ break;
+ }
+ case AV_CODEC_ID_FLAC: {
+ uint8_t buf[13];
+
+ init_put_bits(&pb, buf, sizeof(buf));
+ ret = init_get_bits8(&gb, codecpar->extradata, codecpar->extradata_size);
+ if (ret < 0)
+ return ret;
+
+ put_bits32(&pb, get_bits_long(&gb, 32)); // min/max blocksize
+ put_bits64(&pb, 48, get_bits64(&gb, 48)); // min/max framesize
+ put_bits(&pb, 20, get_bits(&gb, 20)); // samplerate
+ skip_bits(&gb, 3);
+ put_bits(&pb, 3, codecpar->ch_layout.nb_channels - 1);
+ ret = put_bits_left(&pb);
+ put_bits(&pb, ret, get_bits(&gb, ret));
+ flush_put_bits(&pb);
+
+ memcpy(codecpar->extradata, buf, sizeof(buf));
+ break;
+ }
+ }
+
+ return 0;
+}
+
+static int scalable_channel_layout_config(void *s, AVIOContext *pb,
+ IAMFAudioElement *audio_element,
+ const IAMFCodecConfig *codec_config)
+{
+ int nb_layers, k = 0;
+
+ nb_layers = avio_r8(pb) >> 5; // get_bits(&gb, 3);
+ // skip_bits(&gb, 5); //reserved
+
+ if (nb_layers > 6)
+ return AVERROR_INVALIDDATA;
+
+ for (int i = 0; i < nb_layers; i++) {
+ AVIAMFLayer *layer;
+ int loudspeaker_layout, output_gain_is_present_flag;
+ int substream_count, coupled_substream_count;
+ int ret, byte = avio_r8(pb);
+
+ layer = av_iamf_audio_element_add_layer(audio_element->element);
+ if (!layer)
+ return AVERROR(ENOMEM);
+
+ loudspeaker_layout = byte >> 4; // get_bits(&gb, 4);
+ output_gain_is_present_flag = (byte >> 3) & 1; //get_bits1(&gb);
+ if ((byte >> 2) & 1)
+ layer->flags |= AV_IAMF_LAYER_FLAG_RECON_GAIN;
+ substream_count = avio_r8(pb);
+ coupled_substream_count = avio_r8(pb);
+
+ if (output_gain_is_present_flag) {
+ layer->output_gain_flags = avio_r8(pb) >> 2; // get_bits(&gb, 6);
+ layer->output_gain = av_make_q(sign_extend(avio_rb16(pb), 16), 1 << 8);
+ }
+
+ if (loudspeaker_layout < 10)
+ av_channel_layout_copy(&layer->ch_layout, &ff_iamf_scalable_ch_layouts[loudspeaker_layout]);
+ else
+ layer->ch_layout = (AVChannelLayout){ .order = AV_CHANNEL_ORDER_UNSPEC,
+ .nb_channels = substream_count +
+ coupled_substream_count };
+
+ for (int j = 0; j < substream_count; j++) {
+ IAMFSubStream *substream = &audio_element->substreams[k++];
+
+ substream->codecpar->ch_layout = coupled_substream_count-- > 0 ? (AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO :
+ (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
+
+ ret = update_extradata(substream->codecpar);
+ if (ret < 0)
+ return ret;
+ }
+
+ }
+
+ return 0;
+}
+
+static int ambisonics_config(void *s, AVIOContext *pb,
+ IAMFAudioElement *audio_element,
+ const IAMFCodecConfig *codec_config)
+{
+ AVIAMFLayer *layer;
+ unsigned ambisonics_mode;
+ int output_channel_count, substream_count, order;
+ int ret;
+
+ ambisonics_mode = ffio_read_leb(pb);
+ if (ambisonics_mode > 1)
+ return 0;
+
+ output_channel_count = avio_r8(pb); // C
+ substream_count = avio_r8(pb); // N
+ if (audio_element->nb_substreams != substream_count)
+ return AVERROR_INVALIDDATA;
+
+ order = floor(sqrt(output_channel_count - 1));
+ /* incomplete order - some harmonics are missing */
+ if ((order + 1) * (order + 1) != output_channel_count)
+ return AVERROR_INVALIDDATA;
+
+ layer = av_iamf_audio_element_add_layer(audio_element->element);
+ if (!layer)
+ return AVERROR(ENOMEM);
+
+ layer->ambisonics_mode = ambisonics_mode;
+ if (ambisonics_mode == 0) {
+ for (int i = 0; i < substream_count; i++) {
+ IAMFSubStream *substream = &audio_element->substreams[i];
+
+ substream->codecpar->ch_layout = (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
+
+ ret = update_extradata(substream->codecpar);
+ if (ret < 0)
+ return ret;
+ }
+
+ layer->ch_layout.order = AV_CHANNEL_ORDER_CUSTOM;
+ layer->ch_layout.nb_channels = output_channel_count;
+ layer->ch_layout.u.map = av_calloc(output_channel_count, sizeof(*layer->ch_layout.u.map));
+ if (!layer->ch_layout.u.map)
+ return AVERROR(ENOMEM);
+
+ for (int i = 0; i < output_channel_count; i++)
+ layer->ch_layout.u.map[i].id = avio_r8(pb) + AV_CHAN_AMBISONIC_BASE;
+ } else {
+ int coupled_substream_count = avio_r8(pb); // M
+ int nb_demixing_matrix = substream_count + coupled_substream_count;
+ int demixing_matrix_size = nb_demixing_matrix * output_channel_count;
+
+ layer->ch_layout = (AVChannelLayout){ .order = AV_CHANNEL_ORDER_AMBISONIC, .nb_channels = output_channel_count };
+ layer->demixing_matrix = av_malloc_array(demixing_matrix_size, sizeof(*layer->demixing_matrix));
+ if (!layer->demixing_matrix)
+ return AVERROR(ENOMEM);
+
+ for (int i = 0; i < demixing_matrix_size; i++)
+ layer->demixing_matrix[i] = av_make_q(sign_extend(avio_rb16(pb), 16), 1 << 8);
+
+ for (int i = 0; i < substream_count; i++) {
+ IAMFSubStream *substream = &audio_element->substreams[i];
+
+ substream->codecpar->ch_layout = coupled_substream_count-- > 0 ? (AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO :
+ (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
+
+
+ ret = update_extradata(substream->codecpar);
+ if (ret < 0)
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
+static int param_parse(void *s, IAMFContext *c, AVIOContext *pb,
+ unsigned int type,
+ const IAMFAudioElement *audio_element,
+ AVIAMFParamDefinition **out_param_definition)
+{
+ IAMFParamDefinition *param_definition = NULL;
+ AVIAMFParamDefinition *param;
+ unsigned int parameter_id, parameter_rate, mode;
+ unsigned int duration = 0, constant_subblock_duration = 0, nb_subblocks = 0;
+ size_t param_size;
+
+ parameter_id = ffio_read_leb(pb);
+
+ for (int i = 0; i < c->nb_param_definitions; i++)
+ if (c->param_definitions[i]->param->parameter_id == parameter_id) {
+ param_definition = c->param_definitions[i];
+ break;
+ }
+
+ parameter_rate = ffio_read_leb(pb);
+ mode = avio_r8(pb) >> 7;
+
+ if (mode == 0) {
+ duration = ffio_read_leb(pb);
+ constant_subblock_duration = ffio_read_leb(pb);
+ if (constant_subblock_duration == 0)
+ nb_subblocks = ffio_read_leb(pb);
+ else
+ nb_subblocks = duration / constant_subblock_duration;
+ }
+
+ param = av_iamf_param_definition_alloc(type, nb_subblocks, ¶m_size);
+ if (!param)
+ return AVERROR(ENOMEM);
+
+ for (int i = 0; i < nb_subblocks; i++) {
+ void *subblock = av_iamf_param_definition_get_subblock(param, i);
+ unsigned int subblock_duration = constant_subblock_duration;
+
+ if (constant_subblock_duration == 0)
+ subblock_duration = ffio_read_leb(pb);
+
+ switch (type) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN: {
+ AVIAMFMixGain *mix = subblock;
+ mix->subblock_duration = subblock_duration;
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING: {
+ AVIAMFDemixingInfo *demix = subblock;
+ demix->subblock_duration = subblock_duration;
+ // DemixingInfoParameterData
+ demix->dmixp_mode = avio_r8(pb) >> 5;
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN: {
+ AVIAMFReconGain *recon = subblock;
+ recon->subblock_duration = subblock_duration;
+ break;
+ }
+ default:
+ av_free(param);
+ return AVERROR_INVALIDDATA;
+ }
+ }
+
+ param->parameter_id = parameter_id;
+ param->parameter_rate = parameter_rate;
+ param->duration = duration;
+ param->constant_subblock_duration = constant_subblock_duration;
+ param->nb_subblocks = nb_subblocks;
+
+ if (param_definition) {
+ if (param_definition->param_size != param_size || memcmp(param_definition->param, param, param_size)) {
+ av_log(s, AV_LOG_ERROR, "Incosistent parameters for parameter_id %u\n", parameter_id);
+ av_free(param);
+ return AVERROR_INVALIDDATA;
+ }
+ } else {
+ IAMFParamDefinition **tmp = av_realloc_array(c->param_definitions, c->nb_param_definitions + 1,
+ sizeof(*c->param_definitions));
+ if (!tmp) {
+ av_free(param);
+ return AVERROR(ENOMEM);
+ }
+ c->param_definitions = tmp;
+
+ param_definition = av_mallocz(sizeof(*param_definition));
+ if (!param_definition) {
+ av_free(param);
+ return AVERROR(ENOMEM);
+ }
+ param_definition->param = param;
+ param_definition->mode = !mode;
+ param_definition->param_size = param_size;
+ param_definition->audio_element = audio_element;
+
+ c->param_definitions[c->nb_param_definitions++] = param_definition;
+ }
+
+ av_assert0(out_param_definition);
+ *out_param_definition = param;
+
+ return 0;
+}
+
+static IAMFCodecConfig *get_codec_config(IAMFContext *c, unsigned int codec_config_id)
+{
+ for (int i = 0; i < c->nb_codec_configs; i++) {
+ if (c->codec_configs[i]->codec_config_id == codec_config_id)
+ return c->codec_configs[i];
+ }
+
+ return NULL;
+}
+
+static int audio_element_obu(void *s, IAMFContext *c, AVIOContext *pb, int len)
+{
+ const IAMFCodecConfig *codec_config;
+ AVIAMFAudioElement *element;
+ IAMFAudioElement **tmp, *audio_element = NULL;
+ FFIOContext b;
+ AVIOContext *pbc;
+ uint8_t *buf;
+ unsigned audio_element_id, codec_config_id, num_parameters;
+ int audio_element_type, ret;
+
+ buf = av_malloc(len);
+ if (!buf)
+ return AVERROR(ENOMEM);
+
+ ret = avio_read(pb, buf, len);
+ if (ret != len) {
+ if (ret >= 0)
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ ffio_init_context(&b, buf, len, 0, NULL, NULL, NULL, NULL);
+ pbc = &b.pub;
+
+ audio_element_id = ffio_read_leb(pbc);
+
+ for (int i = 0; i < c->nb_audio_elements; i++)
+ if (c->audio_elements[i]->audio_element_id == audio_element_id) {
+ av_log(s, AV_LOG_ERROR, "Duplicate audio_element_id %d\n", audio_element_id);
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ audio_element_type = avio_r8(pbc) >> 5;
+ codec_config_id = ffio_read_leb(pbc);
+
+ codec_config = get_codec_config(c, codec_config_id);
+ if (!codec_config) {
+ av_log(s, AV_LOG_ERROR, "Non existant codec config id %d referenced in an audio element\n", codec_config_id);
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ if (codec_config->codec_id == AV_CODEC_ID_NONE) {
+ av_log(s, AV_LOG_DEBUG, "Unknown codec id referenced in an audio element. Ignoring\n");
+ ret = 0;
+ goto fail;
+ }
+
+ tmp = av_realloc_array(c->audio_elements, c->nb_audio_elements + 1, sizeof(*c->audio_elements));
+ if (!tmp) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ c->audio_elements = tmp;
+
+ audio_element = av_mallocz(sizeof(*audio_element));
+ if (!audio_element) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ audio_element->nb_substreams = ffio_read_leb(pbc);
+ audio_element->codec_config_id = codec_config_id;
+ audio_element->audio_element_id = audio_element_id;
+ audio_element->substreams = av_calloc(audio_element->nb_substreams, sizeof(*audio_element->substreams));
+ if (!audio_element->substreams) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ element = audio_element->element = av_iamf_audio_element_alloc();
+ if (!element) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ element->audio_element_type = audio_element_type;
+
+ for (int i = 0; i < audio_element->nb_substreams; i++) {
+ IAMFSubStream *substream = &audio_element->substreams[i];
+
+ substream->codecpar = avcodec_parameters_alloc();
+ if (!substream->codecpar) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ substream->audio_substream_id = ffio_read_leb(pbc);
+
+ substream->codecpar->codec_type = AVMEDIA_TYPE_AUDIO;
+ substream->codecpar->codec_id = codec_config->codec_id;
+ substream->codecpar->frame_size = codec_config->nb_samples;
+ substream->codecpar->sample_rate = codec_config->sample_rate;
+ substream->codecpar->seek_preroll = codec_config->seek_preroll;
+
+ switch(substream->codecpar->codec_id) {
+ case AV_CODEC_ID_AAC:
+ case AV_CODEC_ID_FLAC:
+ case AV_CODEC_ID_OPUS:
+ substream->codecpar->extradata = av_malloc(codec_config->extradata_size + AV_INPUT_BUFFER_PADDING_SIZE);
+ if (!substream->codecpar->extradata) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ memcpy(substream->codecpar->extradata, codec_config->extradata, codec_config->extradata_size);
+ memset(substream->codecpar->extradata + codec_config->extradata_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
+ substream->codecpar->extradata_size = codec_config->extradata_size;
+ break;
+ }
+ }
+
+ num_parameters = ffio_read_leb(pbc);
+ if (num_parameters && audio_element_type != 0) {
+ av_log(s, AV_LOG_ERROR, "Audio Element parameter count %u is invalid"
+ " for Scene representations\n", num_parameters);
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ for (int i = 0; i < num_parameters; i++) {
+ unsigned type;
+
+ type = ffio_read_leb(pbc);
+ if (type == AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN) {
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ } else if (type == AV_IAMF_PARAMETER_DEFINITION_DEMIXING) {
+ ret = param_parse(s, c, pbc, type, audio_element, &element->demixing_info);
+ if (ret < 0)
+ goto fail;
+
+ element->default_w = avio_r8(pbc) >> 4;
+ } else if (type == AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN) {
+ ret = param_parse(s, c, pbc, type, audio_element, &element->recon_gain_info);
+ if (ret < 0)
+ goto fail;
+ } else {
+ unsigned param_definition_size = ffio_read_leb(pbc);
+ avio_skip(pbc, param_definition_size);
+ }
+ }
+
+ if (audio_element_type == AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL) {
+ ret = scalable_channel_layout_config(s, pbc, audio_element, codec_config);
+ if (ret < 0)
+ goto fail;
+ } else if (audio_element_type == AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE) {
+ ret = ambisonics_config(s, pbc, audio_element, codec_config);
+ if (ret < 0)
+ goto fail;
+ } else {
+ unsigned audio_element_config_size = ffio_read_leb(pbc);
+ avio_skip(pbc, audio_element_config_size);
+ }
+
+ c->audio_elements[c->nb_audio_elements++] = audio_element;
+
+ len -= avio_tell(pbc);
+ if (len)
+ av_log(s, AV_LOG_WARNING, "Underread in audio_element_obu. %d bytes left at the end\n", len);
+
+ ret = 0;
+fail:
+ av_free(buf);
+ if (ret < 0)
+ ff_iamf_free_audio_element(&audio_element);
+ return ret;
+}
+
+static int label_string(AVIOContext *pb, char **label)
+{
+ uint8_t buf[128];
+
+ avio_get_str(pb, sizeof(buf), buf, sizeof(buf));
+
+ if (pb->error)
+ return pb->error;
+ if (pb->eof_reached)
+ return AVERROR_INVALIDDATA;
+ *label = av_strdup(buf);
+ if (!*label)
+ return AVERROR(ENOMEM);
+
+ return 0;
+}
+
+static int mix_presentation_obu(void *s, IAMFContext *c, AVIOContext *pb, int len)
+{
+ AVIAMFMixPresentation *mix;
+ IAMFMixPresentation **tmp, *mix_presentation = NULL;
+ FFIOContext b;
+ AVIOContext *pbc;
+ uint8_t *buf;
+ unsigned mix_presentation_id;
+ int ret;
+
+ buf = av_malloc(len);
+ if (!buf)
+ return AVERROR(ENOMEM);
+
+ ret = avio_read(pb, buf, len);
+ if (ret != len) {
+ if (ret >= 0)
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ ffio_init_context(&b, buf, len, 0, NULL, NULL, NULL, NULL);
+ pbc = &b.pub;
+
+ mix_presentation_id = ffio_read_leb(pbc);
+
+ for (int i = 0; i < c->nb_mix_presentations; i++)
+ if (c->mix_presentations[i]->mix_presentation_id == mix_presentation_id) {
+ av_log(s, AV_LOG_ERROR, "Duplicate mix_presentation_id %d\n", mix_presentation_id);
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ tmp = av_realloc_array(c->mix_presentations, c->nb_mix_presentations + 1, sizeof(*c->mix_presentations));
+ if (!tmp) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+ c->mix_presentations = tmp;
+
+ mix_presentation = av_mallocz(sizeof(*mix_presentation));
+ if (!mix_presentation) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ mix_presentation->mix_presentation_id = mix_presentation_id;
+ mix = mix_presentation->mix = av_iamf_mix_presentation_alloc();
+ if (!mix) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ mix_presentation->count_label = ffio_read_leb(pbc);
+ mix_presentation->language_label = av_calloc(mix_presentation->count_label,
+ sizeof(*mix_presentation->language_label));
+ if (!mix_presentation->language_label) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ for (int i = 0; i < mix_presentation->count_label; i++) {
+ ret = label_string(pbc, &mix_presentation->language_label[i]);
+ if (ret < 0)
+ goto fail;
+ }
+
+ for (int i = 0; i < mix_presentation->count_label; i++) {
+ char *annotation = NULL;
+ ret = label_string(pbc, &annotation);
+ if (ret < 0)
+ goto fail;
+ ret = av_dict_set(&mix->annotations, mix_presentation->language_label[i], annotation,
+ AV_DICT_DONT_STRDUP_VAL | AV_DICT_DONT_OVERWRITE);
+ if (ret < 0)
+ goto fail;
+ }
+
+ mix->nb_submixes = ffio_read_leb(pbc);
+ mix->submixes = av_calloc(mix->nb_submixes, sizeof(*mix->submixes));
+ if (!mix->submixes) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ for (int i = 0; i < mix->nb_submixes; i++) {
+ AVIAMFSubmix *sub_mix;
+
+ sub_mix = mix->submixes[i] = av_mallocz(sizeof(*sub_mix));
+ if (!sub_mix) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ sub_mix->nb_elements = ffio_read_leb(pbc);
+ sub_mix->elements = av_calloc(sub_mix->nb_elements, sizeof(*sub_mix->elements));
+ if (!sub_mix->elements) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ for (int j = 0; j < sub_mix->nb_elements; j++) {
+ AVIAMFSubmixElement *submix_element;
+ IAMFAudioElement *audio_element = NULL;
+ unsigned int rendering_config_extension_size;
+
+ submix_element = sub_mix->elements[j] = av_mallocz(sizeof(*submix_element));
+ if (!submix_element) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ submix_element->audio_element_id = ffio_read_leb(pbc);
+
+ for (int k = 0; k < c->nb_audio_elements; k++)
+ if (c->audio_elements[k]->audio_element_id == submix_element->audio_element_id) {
+ audio_element = c->audio_elements[k];
+ break;
+ }
+
+ if (!audio_element) {
+ av_log(s, AV_LOG_ERROR, "Invalid Audio Element with id %u referenced by Mix Parameters %u\n",
+ submix_element->audio_element_id, mix_presentation_id);
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ for (int k = 0; k < mix_presentation->count_label; k++) {
+ char *annotation = NULL;
+ ret = label_string(pbc, &annotation);
+ if (ret < 0)
+ goto fail;
+ ret = av_dict_set(&submix_element->annotations, mix_presentation->language_label[k], annotation,
+ AV_DICT_DONT_STRDUP_VAL | AV_DICT_DONT_OVERWRITE);
+ if (ret < 0)
+ goto fail;
+ }
+
+ submix_element->headphones_rendering_mode = avio_r8(pbc) >> 6;
+
+ rendering_config_extension_size = ffio_read_leb(pbc);
+ avio_skip(pbc, rendering_config_extension_size);
+
+ ret = param_parse(s, c, pbc, AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN,
+ NULL,
+ &submix_element->element_mix_config);
+ if (ret < 0)
+ goto fail;
+ submix_element->default_mix_gain = av_make_q(sign_extend(avio_rb16(pbc), 16), 1 << 8);
+ }
+
+ ret = param_parse(s, c, pbc, AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN, NULL, &sub_mix->output_mix_config);
+ if (ret < 0)
+ goto fail;
+ sub_mix->default_mix_gain = av_make_q(sign_extend(avio_rb16(pbc), 16), 1 << 8);
+
+ sub_mix->nb_layouts = ffio_read_leb(pbc);
+ sub_mix->layouts = av_calloc(sub_mix->nb_layouts, sizeof(*sub_mix->layouts));
+ if (!sub_mix->layouts) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ for (int j = 0; j < sub_mix->nb_layouts; j++) {
+ AVIAMFSubmixLayout *submix_layout;
+ int info_type;
+ int byte = avio_r8(pbc);
+
+ submix_layout = sub_mix->layouts[j] = av_mallocz(sizeof(*submix_layout));
+ if (!submix_layout) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ submix_layout->layout_type = byte >> 6;
+ if (submix_layout->layout_type < AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS &&
+ submix_layout->layout_type > AV_IAMF_SUBMIX_LAYOUT_TYPE_BINAURAL) {
+ av_log(s, AV_LOG_ERROR, "Invalid Layout type %u in a submix from Mix Presentation %u\n",
+ submix_layout->layout_type, mix_presentation_id);
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+ if (submix_layout->layout_type == 2) {
+ int sound_system;
+ sound_system = (byte >> 2) & 0xF;
+ av_channel_layout_copy(&submix_layout->sound_system, &ff_iamf_sound_system_map[sound_system].layout);
+ }
+
+ info_type = avio_r8(pbc);
+ submix_layout->integrated_loudness = av_make_q(sign_extend(avio_rb16(pbc), 16), 1 << 8);
+ submix_layout->digital_peak = av_make_q(sign_extend(avio_rb16(pbc), 16), 1 << 8);
+
+ if (info_type & 1)
+ submix_layout->true_peak = av_make_q(sign_extend(avio_rb16(pbc), 16), 1 << 8);
+ if (info_type & 2) {
+ unsigned int num_anchored_loudness = avio_r8(pbc);
+
+ for (int k = 0; k < num_anchored_loudness; k++) {
+ unsigned int anchor_element = avio_r8(pbc);
+ AVRational anchored_loudness = av_make_q(sign_extend(avio_rb16(pbc), 16), 1 << 8);
+ if (anchor_element == IAMF_ANCHOR_ELEMENT_DIALOGUE)
+ submix_layout->dialogue_anchored_loudness = anchored_loudness;
+ else if (anchor_element <= IAMF_ANCHOR_ELEMENT_ALBUM)
+ submix_layout->album_anchored_loudness = anchored_loudness;
+ else
+ av_log(s, AV_LOG_DEBUG, "Unknown anchor_element. Ignoring\n");
+ }
+ }
+
+ if (info_type & 0xFC) {
+ unsigned int info_type_size = ffio_read_leb(pbc);
+ avio_skip(pbc, info_type_size);
+ }
+ }
+ }
+
+ c->mix_presentations[c->nb_mix_presentations++] = mix_presentation;
+
+ len -= avio_tell(pbc);
+ if (len)
+ av_log(s, AV_LOG_WARNING, "Underread in mix_presentation_obu. %d bytes left at the end\n", len);
+
+ ret = 0;
+fail:
+ av_free(buf);
+ if (ret < 0)
+ ff_iamf_free_mix_presentation(&mix_presentation);
+ return ret;
+}
+
+int ff_iamf_parse_obu_header(const uint8_t *buf, int buf_size,
+ unsigned *obu_size, int *start_pos, enum IAMF_OBU_Type *type,
+ unsigned *skip_samples, unsigned *discard_padding)
+{
+ GetBitContext gb;
+ int ret, extension_flag, trimming, start;
+ unsigned skip = 0, discard = 0;
+ unsigned size;
+
+ ret = init_get_bits8(&gb, buf, FFMIN(buf_size, MAX_IAMF_OBU_HEADER_SIZE));
+ if (ret < 0)
+ return ret;
+
+ *type = get_bits(&gb, 5);
+ /*redundant =*/ get_bits1(&gb);
+ trimming = get_bits1(&gb);
+ extension_flag = get_bits1(&gb);
+
+ *obu_size = get_leb(&gb);
+ if (*obu_size > INT_MAX)
+ return AVERROR_INVALIDDATA;
+
+ start = get_bits_count(&gb) / 8;
+
+ if (trimming) {
+ discard = get_leb(&gb); // num_samples_to_trim_at_end
+ skip = get_leb(&gb); // num_samples_to_trim_at_start
+ }
+
+ if (skip_samples)
+ *skip_samples = skip;
+ if (discard_padding)
+ *discard_padding = discard;
+
+ if (extension_flag) {
+ unsigned int extension_bytes;
+ extension_bytes = get_leb(&gb);
+ if (extension_bytes > INT_MAX / 8)
+ return AVERROR_INVALIDDATA;
+ skip_bits_long(&gb, extension_bytes * 8);
+ }
+
+ if (get_bits_left(&gb) < 0)
+ return AVERROR_INVALIDDATA;
+
+ size = *obu_size + start;
+ if (size > INT_MAX)
+ return AVERROR_INVALIDDATA;
+
+ *obu_size -= get_bits_count(&gb) / 8 - start;
+ *start_pos = size - *obu_size;
+
+ return size;
+}
+
+int ff_iamfdec_read_descriptors(IAMFContext *c, AVIOContext *pb,
+ int max_size, void *log_ctx)
+{
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE + AV_INPUT_BUFFER_PADDING_SIZE];
+ int ret;
+
+ while (1) {
+ unsigned obu_size;
+ enum IAMF_OBU_Type type;
+ int start_pos, len, size;
+
+ if ((ret = ffio_ensure_seekback(pb, FFMIN(MAX_IAMF_OBU_HEADER_SIZE, max_size))) < 0)
+ return ret;
+ size = avio_read(pb, header, FFMIN(MAX_IAMF_OBU_HEADER_SIZE, max_size));
+ if (size < 0)
+ return size;
+
+ len = ff_iamf_parse_obu_header(header, size, &obu_size, &start_pos, &type, NULL, NULL);
+ if (len < 0 || obu_size > max_size) {
+ av_log(log_ctx, AV_LOG_ERROR, "Failed to read obu header\n");
+ avio_seek(pb, -size, SEEK_CUR);
+ return len;
+ }
+
+ if (type >= IAMF_OBU_IA_PARAMETER_BLOCK && type < IAMF_OBU_IA_SEQUENCE_HEADER) {
+ avio_seek(pb, -size, SEEK_CUR);
+ break;
+ }
+
+ avio_seek(pb, -(size - start_pos), SEEK_CUR);
+ switch (type) {
+ case IAMF_OBU_IA_CODEC_CONFIG:
+ ret = codec_config_obu(log_ctx, c, pb, obu_size);
+ break;
+ case IAMF_OBU_IA_AUDIO_ELEMENT:
+ ret = audio_element_obu(log_ctx, c, pb, obu_size);
+ break;
+ case IAMF_OBU_IA_MIX_PRESENTATION:
+ ret = mix_presentation_obu(log_ctx, c, pb, obu_size);
+ break;
+ case IAMF_OBU_IA_TEMPORAL_DELIMITER:
+ break;
+ default: {
+ int64_t offset = avio_skip(pb, obu_size);
+ if (offset < 0)
+ ret = offset;
+ break;
+ }
+ }
+ if (ret < 0) {
+ av_log(log_ctx, AV_LOG_ERROR, "Failed to read obu type %d\n", type);
+ return ret;
+ }
+ max_size -= obu_size + start_pos;
+ if (max_size < 0)
+ return AVERROR_INVALIDDATA;
+ if (!max_size)
+ break;
+ }
+
+ return 0;
+}
diff --git a/libavformat/iamf_parse.h b/libavformat/iamf_parse.h
new file mode 100644
index 0000000000..f4f297ecd4
--- /dev/null
+++ b/libavformat/iamf_parse.h
@@ -0,0 +1,38 @@
+/*
+ * Immersive Audio Model and Formats parsing
+ * Copyright (c) 2023 James Almer <jamrial@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
+ */
+
+#ifndef AVFORMAT_IAMF_PARSE_H
+#define AVFORMAT_IAMF_PARSE_H
+
+#include <stdint.h>
+
+#include "libavutil/iamf.h"
+#include "avio.h"
+#include "iamf.h"
+
+int ff_iamf_parse_obu_header(const uint8_t *buf, int buf_size,
+ unsigned *obu_size, int *start_pos, enum IAMF_OBU_Type *type,
+ unsigned *skip_samples, unsigned *discard_padding);
+
+int ff_iamfdec_read_descriptors(IAMFContext *c, AVIOContext *pb,
+ int size, void *log_ctx);
+
+#endif /* AVFORMAT_IAMF_PARSE_H */
diff --git a/libavformat/iamfdec.c b/libavformat/iamfdec.c
new file mode 100644
index 0000000000..0374d0f241
--- /dev/null
+++ b/libavformat/iamfdec.c
@@ -0,0 +1,503 @@
+/*
+ * Immersive Audio Model and Formats demuxer
+ * Copyright (c) 2023 James Almer <jamrial@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 "config_components.h"
+
+#include "libavutil/avassert.h"
+#include "libavutil/iamf.h"
+#include "libavutil/intreadwrite.h"
+#include "libavutil/log.h"
+#include "libavcodec/mathops.h"
+#include "avformat.h"
+#include "avio_internal.h"
+#include "demux.h"
+#include "iamf.h"
+#include "iamf_parse.h"
+#include "internal.h"
+
+typedef struct IAMFDemuxContext {
+ IAMFContext iamf;
+
+ // Packet side data
+ AVIAMFParamDefinition *mix;
+ size_t mix_size;
+ AVIAMFParamDefinition *demix;
+ size_t demix_size;
+ AVIAMFParamDefinition *recon;
+ size_t recon_size;
+} IAMFDemuxContext;
+
+static AVStream *find_stream_by_id(AVFormatContext *s, int id)
+{
+ for (int i = 0; i < s->nb_streams; i++)
+ if (s->streams[i]->id == id)
+ return s->streams[i];
+
+ av_log(s, AV_LOG_ERROR, "Invalid stream id %d\n", id);
+ return NULL;
+}
+
+static int audio_frame_obu(AVFormatContext *s, AVPacket *pkt, int len,
+ enum IAMF_OBU_Type type,
+ unsigned skip_samples, unsigned discard_padding,
+ int id_in_bitstream)
+{
+ const IAMFDemuxContext *const c = s->priv_data;
+ AVStream *st;
+ int ret, audio_substream_id;
+
+ if (id_in_bitstream) {
+ unsigned explicit_audio_substream_id;
+ int64_t pos = avio_tell(s->pb);
+ explicit_audio_substream_id = ffio_read_leb(s->pb);
+ len -= avio_tell(s->pb) - pos;
+ audio_substream_id = explicit_audio_substream_id;
+ } else
+ audio_substream_id = type - IAMF_OBU_IA_AUDIO_FRAME_ID0;
+
+ st = find_stream_by_id(s, audio_substream_id);
+ if (!st)
+ return AVERROR_INVALIDDATA;
+
+ ret = av_get_packet(s->pb, pkt, len);
+ if (ret < 0)
+ return ret;
+ if (ret != len)
+ return AVERROR_INVALIDDATA;
+
+ if (skip_samples || discard_padding) {
+ uint8_t *side_data = av_packet_new_side_data(pkt, AV_PKT_DATA_SKIP_SAMPLES, 10);
+ if (!side_data)
+ return AVERROR(ENOMEM);
+ AV_WL32(side_data, skip_samples);
+ AV_WL32(side_data + 4, discard_padding);
+ }
+ if (c->mix) {
+ uint8_t *side_data = av_packet_new_side_data(pkt, AV_PKT_DATA_IAMF_MIX_GAIN_PARAM, c->mix_size);
+ if (!side_data)
+ return AVERROR(ENOMEM);
+ memcpy(side_data, c->mix, c->mix_size);
+ }
+ if (c->demix) {
+ uint8_t *side_data = av_packet_new_side_data(pkt, AV_PKT_DATA_IAMF_DEMIXING_INFO_PARAM, c->demix_size);
+ if (!side_data)
+ return AVERROR(ENOMEM);
+ memcpy(side_data, c->demix, c->demix_size);
+ }
+ if (c->recon) {
+ uint8_t *side_data = av_packet_new_side_data(pkt, AV_PKT_DATA_IAMF_RECON_GAIN_INFO_PARAM, c->recon_size);
+ if (!side_data)
+ return AVERROR(ENOMEM);
+ memcpy(side_data, c->recon, c->recon_size);
+ }
+
+ pkt->stream_index = st->index;
+ return 0;
+}
+
+static const IAMFParamDefinition *get_param_definition(AVFormatContext *s, unsigned int parameter_id)
+{
+ const IAMFDemuxContext *const c = s->priv_data;
+ const IAMFContext *const iamf = &c->iamf;
+ const IAMFParamDefinition *param_definition = NULL;
+
+ for (int i = 0; i < iamf->nb_param_definitions; i++)
+ if (iamf->param_definitions[i]->param->parameter_id == parameter_id) {
+ param_definition = iamf->param_definitions[i];
+ break;
+ }
+
+ return param_definition;
+}
+
+static int parameter_block_obu(AVFormatContext *s, int len)
+{
+ IAMFDemuxContext *const c = s->priv_data;
+ const IAMFParamDefinition *param_definition;
+ const AVIAMFParamDefinition *param;
+ AVIAMFParamDefinition *out_param = NULL;
+ FFIOContext b;
+ AVIOContext *pb;
+ uint8_t *buf;
+ unsigned int duration, constant_subblock_duration;
+ unsigned int nb_subblocks;
+ unsigned int parameter_id;
+ size_t out_param_size;
+ int ret;
+
+ buf = av_malloc(len);
+ if (!buf)
+ return AVERROR(ENOMEM);
+
+ ret = avio_read(s->pb, buf, len);
+ if (ret != len) {
+ if (ret >= 0)
+ ret = AVERROR_INVALIDDATA;
+ goto fail;
+ }
+
+ ffio_init_context(&b, buf, len, 0, NULL, NULL, NULL, NULL);
+ pb = &b.pub;
+
+ parameter_id = ffio_read_leb(pb);
+ param_definition = get_param_definition(s, parameter_id);
+ if (!param_definition) {
+ av_log(s, AV_LOG_VERBOSE, "Non existant parameter_id %d referenced in a parameter block. Ignoring\n",
+ parameter_id);
+ ret = 0;
+ goto fail;
+ }
+
+ param = param_definition->param;
+ if (!param_definition->mode) {
+ duration = ffio_read_leb(pb);
+ constant_subblock_duration = ffio_read_leb(pb);
+ if (constant_subblock_duration == 0)
+ nb_subblocks = ffio_read_leb(pb);
+ else
+ nb_subblocks = duration / constant_subblock_duration;
+ } else {
+ duration = param->duration;
+ constant_subblock_duration = param->constant_subblock_duration;
+ nb_subblocks = param->nb_subblocks;
+ if (!nb_subblocks)
+ nb_subblocks = duration / constant_subblock_duration;
+ }
+
+ out_param = av_iamf_param_definition_alloc(param->type, nb_subblocks, &out_param_size);
+ if (!out_param) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ out_param->parameter_id = param->parameter_id;
+ out_param->type = param->type;
+ out_param->parameter_rate = param->parameter_rate;
+ out_param->duration = duration;
+ out_param->constant_subblock_duration = constant_subblock_duration;
+ out_param->nb_subblocks = nb_subblocks;
+
+ for (int i = 0; i < nb_subblocks; i++) {
+ void *subblock = av_iamf_param_definition_get_subblock(out_param, i);
+ unsigned int subblock_duration = constant_subblock_duration;
+
+ if (!param_definition->mode && !constant_subblock_duration)
+ subblock_duration = ffio_read_leb(pb);
+
+ switch (param->type) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN: {
+ AVIAMFMixGain *mix = subblock;
+
+ mix->animation_type = ffio_read_leb(pb);
+ if (mix->animation_type > AV_IAMF_ANIMATION_TYPE_BEZIER) {
+ ret = 0;
+ av_free(out_param);
+ goto fail;
+ }
+
+ mix->start_point_value = av_make_q(sign_extend(avio_rb16(pb), 16), 1 << 8);
+ if (mix->animation_type >= AV_IAMF_ANIMATION_TYPE_LINEAR)
+ mix->end_point_value = av_make_q(sign_extend(avio_rb16(pb), 16), 1 << 8);
+ if (mix->animation_type == AV_IAMF_ANIMATION_TYPE_BEZIER) {
+ mix->control_point_value = av_make_q(sign_extend(avio_rb16(pb), 16), 1 << 8);
+ mix->control_point_relative_time = av_make_q(avio_r8(pb), 1 << 8);
+ }
+ mix->subblock_duration = subblock_duration;
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING: {
+ AVIAMFDemixingInfo *demix = subblock;
+
+ demix->dmixp_mode = avio_r8(pb) >> 5;
+ demix->subblock_duration = subblock_duration;
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN: {
+ AVIAMFReconGain *recon = subblock;
+ const IAMFAudioElement *audio_element = param_definition->audio_element;
+ const AVIAMFAudioElement *element = audio_element->element;
+
+ av_assert0(audio_element && element);
+ for (int i = 0; i < element->nb_layers; i++) {
+ const AVIAMFLayer *layer = element->layers[i];
+ if (layer->flags & AV_IAMF_LAYER_FLAG_RECON_GAIN) {
+ unsigned int recon_gain_flags = ffio_read_leb(pb);
+ unsigned int bitcount = 7 + 5 * !!(recon_gain_flags & 0x80);
+ recon_gain_flags = (recon_gain_flags & 0x7F) | ((recon_gain_flags & 0xFF00) >> 1);
+ for (int j = 0; j < bitcount; j++) {
+ if (recon_gain_flags & (1 << j))
+ recon->recon_gain[i][j] = avio_r8(pb);
+ }
+ }
+ }
+ recon->subblock_duration = subblock_duration;
+ break;
+ }
+ default:
+ av_assert0(0);
+ }
+ }
+
+ len -= avio_tell(pb);
+ if (len) {
+ int level = (s->error_recognition & AV_EF_EXPLODE) ? AV_LOG_ERROR : AV_LOG_WARNING;
+ av_log(s, level, "Underread in parameter_block_obu. %d bytes left at the end\n", len);
+ }
+
+ switch (param->type) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN:
+ av_free(c->mix);
+ c->mix = out_param;
+ c->mix_size = out_param_size;
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING:
+ av_free(c->demix);
+ c->demix = out_param;
+ c->demix_size = out_param_size;
+ break;
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN:
+ av_free(c->recon);
+ c->recon = out_param;
+ c->recon_size = out_param_size;
+ break;
+ default:
+ av_assert0(0);
+ }
+
+ ret = 0;
+fail:
+ if (ret < 0)
+ av_free(out_param);
+ av_free(buf);
+
+ return ret;
+}
+
+static int iamf_read_packet(AVFormatContext *s, AVPacket *pkt)
+{
+ IAMFDemuxContext *const c = s->priv_data;
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE + AV_INPUT_BUFFER_PADDING_SIZE];
+ unsigned obu_size;
+ int ret;
+
+ while (1) {
+ enum IAMF_OBU_Type type;
+ unsigned skip_samples, discard_padding;
+ int len, size, start_pos;
+
+ if ((ret = ffio_ensure_seekback(s->pb, MAX_IAMF_OBU_HEADER_SIZE)) < 0)
+ return ret;
+ size = avio_read(s->pb, header, MAX_IAMF_OBU_HEADER_SIZE);
+ if (size < 0)
+ return size;
+
+ len = ff_iamf_parse_obu_header(header, size, &obu_size, &start_pos, &type,
+ &skip_samples, &discard_padding);
+ if (len < 0) {
+ av_log(s, AV_LOG_ERROR, "Failed to read obu\n");
+ return len;
+ }
+ avio_seek(s->pb, -(size - start_pos), SEEK_CUR);
+
+ if (type >= IAMF_OBU_IA_AUDIO_FRAME && type <= IAMF_OBU_IA_AUDIO_FRAME_ID17)
+ return audio_frame_obu(s, pkt, obu_size, type,
+ skip_samples, discard_padding,
+ type == IAMF_OBU_IA_AUDIO_FRAME);
+ else if (type == IAMF_OBU_IA_PARAMETER_BLOCK) {
+ ret = parameter_block_obu(s, obu_size);
+ if (ret < 0)
+ return ret;
+ } else if (type == IAMF_OBU_IA_TEMPORAL_DELIMITER) {
+ av_freep(&c->mix);
+ c->mix_size = 0;
+ av_freep(&c->demix);
+ c->demix_size = 0;
+ av_freep(&c->recon);
+ c->recon_size = 0;
+ } else {
+ int64_t offset = avio_skip(s->pb, obu_size);
+ if (offset < 0) {
+ ret = offset;
+ break;
+ }
+ }
+ }
+
+ return ret;
+}
+
+//return < 0 if we need more data
+static int get_score(const uint8_t *buf, int buf_size, enum IAMF_OBU_Type type, int *seq)
+{
+ if (type == IAMF_OBU_IA_SEQUENCE_HEADER) {
+ if (buf_size < 4 || AV_RB32(buf) != MKBETAG('i','a','m','f'))
+ return 0;
+ *seq = 1;
+ return -1;
+ }
+ if (type >= IAMF_OBU_IA_CODEC_CONFIG && type <= IAMF_OBU_IA_TEMPORAL_DELIMITER)
+ return *seq ? -1 : 0;
+ if (type >= IAMF_OBU_IA_AUDIO_FRAME && type <= IAMF_OBU_IA_AUDIO_FRAME_ID17)
+ return *seq ? AVPROBE_SCORE_EXTENSION + 1 : 0;
+ return 0;
+}
+
+static int iamf_probe(const AVProbeData *p)
+{
+ unsigned obu_size;
+ enum IAMF_OBU_Type type;
+ int seq = 0, cnt = 0, start_pos;
+ int ret;
+
+ while (1) {
+ int size = ff_iamf_parse_obu_header(p->buf + cnt, p->buf_size - cnt,
+ &obu_size, &start_pos, &type,
+ NULL, NULL);
+ if (size < 0)
+ return 0;
+
+ ret = get_score(p->buf + cnt + start_pos,
+ p->buf_size - cnt - start_pos,
+ type, &seq);
+ if (ret >= 0)
+ return ret;
+
+ cnt += FFMIN(size, p->buf_size - cnt);
+ }
+ return 0;
+}
+
+static int iamf_read_header(AVFormatContext *s)
+{
+ IAMFDemuxContext *const c = s->priv_data;
+ IAMFContext *const iamf = &c->iamf;
+ int ret;
+
+ ret = ff_iamfdec_read_descriptors(iamf, s->pb, INT_MAX, s);
+ if (ret < 0)
+ return ret;
+
+ for (int i = 0; i < iamf->nb_audio_elements; i++) {
+ IAMFAudioElement *audio_element = iamf->audio_elements[i];
+ AVStreamGroup *stg = avformat_stream_group_create(s, AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT, NULL);
+
+ if (!stg)
+ return AVERROR(ENOMEM);
+
+ stg->id = audio_element->audio_element_id;
+ stg->params.iamf_audio_element = audio_element->element;
+
+ for (int j = 0; j < audio_element->nb_substreams; j++) {
+ IAMFSubStream *substream = &audio_element->substreams[j];
+ AVStream *st = avformat_new_stream(s, NULL);
+
+ if (!st)
+ return AVERROR(ENOMEM);
+
+ ret = avformat_stream_group_add_stream(stg, st);
+ if (ret < 0)
+ return ret;
+
+ ret = avcodec_parameters_copy(st->codecpar, substream->codecpar);
+ if (ret < 0)
+ return ret;
+
+ st->id = substream->audio_substream_id;
+ avpriv_set_pts_info(st, 64, 1, st->codecpar->sample_rate);
+ }
+ }
+
+ for (int i = 0; i < iamf->nb_mix_presentations; i++) {
+ IAMFMixPresentation *mix_presentation = iamf->mix_presentations[i];
+ AVStreamGroup *stg = avformat_stream_group_create(s, AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION, NULL);
+ const AVIAMFMixPresentation *mix = mix_presentation->mix;
+
+ if (!stg)
+ return AVERROR(ENOMEM);
+
+ stg->id = mix_presentation->mix_presentation_id;
+ stg->params.iamf_mix_presentation = mix_presentation->mix;
+
+ for (int j = 0; j < mix->nb_submixes; j++) {
+ AVIAMFSubmix *sub_mix = mix->submixes[j];
+
+ for (int k = 0; k < sub_mix->nb_elements; k++) {
+ AVIAMFSubmixElement *submix_element = sub_mix->elements[k];
+ AVStreamGroup *audio_element = NULL;
+
+ for (int l = 0; l < s->nb_stream_groups; l++)
+ if (s->stream_groups[l]->type == AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT &&
+ s->stream_groups[l]->id == submix_element->audio_element_id) {
+ audio_element = s->stream_groups[l];
+ break;
+ }
+ av_assert0(audio_element);
+
+ for (int l = 0; l < audio_element->nb_streams; l++) {
+ ret = avformat_stream_group_add_stream(stg, audio_element->streams[l]);
+ if (ret < 0 && ret != AVERROR(EEXIST))
+ return ret;
+ }
+ }
+ }
+ }
+
+ return 0;
+}
+
+static int iamf_read_close(AVFormatContext *s)
+{
+ IAMFDemuxContext *const c = s->priv_data;
+ IAMFContext *const iamf = &c->iamf;
+
+ for (int i = 0; i < iamf->nb_audio_elements; i++) {
+ IAMFAudioElement *audio_element = iamf->audio_elements[i];
+ audio_element->element = NULL;
+ }
+ for (int i = 0; i < iamf->nb_mix_presentations; i++) {
+ IAMFMixPresentation *mix_presentation = iamf->mix_presentations[i];
+ mix_presentation->mix = NULL;
+ }
+
+ ff_iamf_uninit_context(&c->iamf);
+
+ av_freep(&c->mix);
+ c->mix_size = 0;
+ av_freep(&c->demix);
+ c->demix_size = 0;
+ av_freep(&c->recon);
+ c->recon_size = 0;
+
+ return 0;
+}
+
+const AVInputFormat ff_iamf_demuxer = {
+ .name = "iamf",
+ .long_name = NULL_IF_CONFIG_SMALL("Raw Immersive Audio Model and Formats"),
+ .priv_data_size = sizeof(IAMFDemuxContext),
+ .flags_internal = FF_FMT_INIT_CLEANUP,
+ .read_probe = iamf_probe,
+ .read_header = iamf_read_header,
+ .read_packet = iamf_read_packet,
+ .read_close = iamf_read_close,
+ .extensions = "iamf",
+ .flags = AVFMT_GENERIC_INDEX | AVFMT_NO_BYTE_SEEK | AVFMT_NOTIMESTAMPS | AVFMT_SHOW_IDS,
+};
--
2.43.0
_______________________________________________
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] 12+ messages in thread* [FFmpeg-devel] [PATCH 8/8] avformat: Immersive Audio Model and Formats muxer
2023-12-14 20:14 [FFmpeg-devel] [PATCH v7 0/8] avformat: introduce AVStreamGroup James Almer
` (6 preceding siblings ...)
2023-12-14 20:14 ` [FFmpeg-devel] [PATCH 7/8] avformat: Immersive Audio Model and Formats demuxer James Almer
@ 2023-12-14 20:14 ` James Almer
7 siblings, 0 replies; 12+ messages in thread
From: James Almer @ 2023-12-14 20:14 UTC (permalink / raw)
To: ffmpeg-devel
Signed-off-by: James Almer <jamrial@gmail.com>
---
libavformat/Makefile | 1 +
libavformat/allformats.c | 1 +
libavformat/iamf_writer.c | 860 ++++++++++++++++++++++++++++++++++++++
libavformat/iamf_writer.h | 51 +++
libavformat/iamfenc.c | 387 +++++++++++++++++
5 files changed, 1300 insertions(+)
create mode 100644 libavformat/iamf_writer.c
create mode 100644 libavformat/iamf_writer.h
create mode 100644 libavformat/iamfenc.c
diff --git a/libavformat/Makefile b/libavformat/Makefile
index f23c22792b..581e378d95 100644
--- a/libavformat/Makefile
+++ b/libavformat/Makefile
@@ -259,6 +259,7 @@ OBJS-$(CONFIG_HLS_DEMUXER) += hls.o hls_sample_encryption.o
OBJS-$(CONFIG_HLS_MUXER) += hlsenc.o hlsplaylist.o avc.o
OBJS-$(CONFIG_HNM_DEMUXER) += hnm.o
OBJS-$(CONFIG_IAMF_DEMUXER) += iamfdec.o iamf_parse.o iamf.o
+OBJS-$(CONFIG_IAMF_MUXER) += iamfenc.o iamf_writer.o iamf.o
OBJS-$(CONFIG_ICO_DEMUXER) += icodec.o
OBJS-$(CONFIG_ICO_MUXER) += icoenc.o
OBJS-$(CONFIG_IDCIN_DEMUXER) += idcin.o
diff --git a/libavformat/allformats.c b/libavformat/allformats.c
index 6e520b78a6..ce6be5f04d 100644
--- a/libavformat/allformats.c
+++ b/libavformat/allformats.c
@@ -213,6 +213,7 @@ extern const AVInputFormat ff_hls_demuxer;
extern const FFOutputFormat ff_hls_muxer;
extern const AVInputFormat ff_hnm_demuxer;
extern const AVInputFormat ff_iamf_demuxer;
+extern const FFOutputFormat ff_iamf_muxer;
extern const AVInputFormat ff_ico_demuxer;
extern const FFOutputFormat ff_ico_muxer;
extern const AVInputFormat ff_idcin_demuxer;
diff --git a/libavformat/iamf_writer.c b/libavformat/iamf_writer.c
new file mode 100644
index 0000000000..9962845049
--- /dev/null
+++ b/libavformat/iamf_writer.c
@@ -0,0 +1,860 @@
+/*
+ * Immersive Audio Model and Formats muxing helpers and structs
+ * Copyright (c) 2023 James Almer <jamrial@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 "libavutil/channel_layout.h"
+#include "libavutil/intreadwrite.h"
+#include "libavutil/iamf.h"
+#include "libavutil/mem.h"
+#include "libavcodec/get_bits.h"
+#include "libavcodec/flac.h"
+#include "libavcodec/mpeg4audio.h"
+#include "libavcodec/put_bits.h"
+#include "avformat.h"
+#include "avio_internal.h"
+#include "iamf.h"
+#include "iamf_writer.h"
+
+
+static int update_extradata(IAMFCodecConfig *codec_config)
+{
+ GetBitContext gb;
+ PutBitContext pb;
+ int ret;
+
+ switch(codec_config->codec_id) {
+ case AV_CODEC_ID_OPUS:
+ if (codec_config->extradata_size < 19)
+ return AVERROR_INVALIDDATA;
+ codec_config->extradata_size -= 8;
+ memmove(codec_config->extradata, codec_config->extradata + 8, codec_config->extradata_size);
+ AV_WB8(codec_config->extradata + 1, 2); // set channels to stereo
+ break;
+ case AV_CODEC_ID_FLAC: {
+ uint8_t buf[13];
+
+ init_put_bits(&pb, buf, sizeof(buf));
+ ret = init_get_bits8(&gb, codec_config->extradata, codec_config->extradata_size);
+ if (ret < 0)
+ return ret;
+
+ put_bits32(&pb, get_bits_long(&gb, 32)); // min/max blocksize
+ put_bits64(&pb, 48, get_bits64(&gb, 48)); // min/max framesize
+ put_bits(&pb, 20, get_bits(&gb, 20)); // samplerate
+ skip_bits(&gb, 3);
+ put_bits(&pb, 3, 1); // set channels to stereo
+ ret = put_bits_left(&pb);
+ put_bits(&pb, ret, get_bits(&gb, ret));
+ flush_put_bits(&pb);
+
+ memcpy(codec_config->extradata, buf, sizeof(buf));
+ break;
+ }
+ default:
+ break;
+ }
+
+ return 0;
+}
+
+static int fill_codec_config(IAMFContext *iamf, const AVStreamGroup *stg,
+ IAMFCodecConfig *codec_config)
+{
+ const AVStream *st = stg->streams[0];
+ IAMFCodecConfig **tmp;
+ int j, ret = 0;
+
+ codec_config->codec_id = st->codecpar->codec_id;
+ codec_config->sample_rate = st->codecpar->sample_rate;
+ codec_config->codec_tag = st->codecpar->codec_tag;
+ codec_config->nb_samples = st->codecpar->frame_size;
+ codec_config->seek_preroll = st->codecpar->seek_preroll;
+ if (st->codecpar->extradata_size) {
+ codec_config->extradata = av_memdup(st->codecpar->extradata, st->codecpar->extradata_size);
+ if (!codec_config->extradata)
+ return AVERROR(ENOMEM);
+ codec_config->extradata_size = st->codecpar->extradata_size;
+ ret = update_extradata(codec_config);
+ if (ret < 0)
+ goto fail;
+ }
+
+ for (j = 0; j < iamf->nb_codec_configs; j++) {
+ if (!memcmp(iamf->codec_configs[j], codec_config, offsetof(IAMFCodecConfig, extradata)) &&
+ (!codec_config->extradata_size || !memcmp(iamf->codec_configs[j]->extradata,
+ codec_config->extradata, codec_config->extradata_size)))
+ break;
+ }
+
+ if (j < iamf->nb_codec_configs) {
+ av_free(iamf->codec_configs[j]->extradata);
+ av_free(iamf->codec_configs[j]);
+ iamf->codec_configs[j] = codec_config;
+ return j;
+ }
+
+ tmp = av_realloc_array(iamf->codec_configs, iamf->nb_codec_configs + 1, sizeof(*iamf->codec_configs));
+ if (!tmp) {
+ ret = AVERROR(ENOMEM);
+ goto fail;
+ }
+
+ iamf->codec_configs = tmp;
+ iamf->codec_configs[iamf->nb_codec_configs] = codec_config;
+ codec_config->codec_config_id = iamf->nb_codec_configs;
+
+ return iamf->nb_codec_configs++;
+
+fail:
+ av_freep(&codec_config->extradata);
+ return ret;
+}
+
+static IAMFParamDefinition *add_param_definition(IAMFContext *iamf, AVIAMFParamDefinition *param,
+ const IAMFAudioElement *audio_element, void *log_ctx)
+{
+ IAMFParamDefinition **tmp, *param_definition;
+ IAMFCodecConfig *codec_config = NULL;
+
+ tmp = av_realloc_array(iamf->param_definitions, iamf->nb_param_definitions + 1,
+ sizeof(*iamf->param_definitions));
+ if (!tmp)
+ return NULL;
+
+ iamf->param_definitions = tmp;
+
+ param_definition = av_mallocz(sizeof(*param_definition));
+ if (!param_definition)
+ return NULL;
+
+ if (audio_element)
+ codec_config = iamf->codec_configs[audio_element->codec_config_id];
+
+ if (!param->parameter_rate) {
+ if (!codec_config) {
+ av_log(log_ctx, AV_LOG_ERROR, "parameter_rate needed but not set for parameter_id %u\n",
+ param->parameter_id);
+ return NULL;
+ }
+ param->parameter_rate = codec_config->sample_rate;
+ }
+ if (codec_config) {
+ if (!param->duration)
+ param->duration = codec_config->nb_samples;
+ if (!param->constant_subblock_duration)
+ param->constant_subblock_duration = codec_config->nb_samples;
+ }
+
+ param_definition->mode = !!param->duration;
+ param_definition->param = param;
+ param_definition->audio_element = audio_element;
+ iamf->param_definitions[iamf->nb_param_definitions++] = param_definition;
+
+ return param_definition;
+}
+
+int ff_iamf_add_audio_element(IAMFContext *iamf, const AVStreamGroup *stg, void *log_ctx)
+{
+ const AVIAMFAudioElement *iamf_audio_element;
+ IAMFAudioElement **tmp, *audio_element;
+ IAMFCodecConfig *codec_config;
+ int ret;
+
+ if (stg->type != AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT)
+ return AVERROR(EINVAL);
+
+ iamf_audio_element = stg->params.iamf_audio_element;
+ if (iamf_audio_element->audio_element_type == AV_IAMF_AUDIO_ELEMENT_TYPE_SCENE) {
+ const AVIAMFLayer *layer = iamf_audio_element->layers[0];
+ if (iamf_audio_element->nb_layers != 1) {
+ av_log(log_ctx, AV_LOG_ERROR, "Invalid amount of layers for SCENE_BASED audio element. Must be 1\n");
+ return AVERROR(EINVAL);
+ }
+ if (layer->ch_layout.order != AV_CHANNEL_ORDER_CUSTOM &&
+ layer->ch_layout.order != AV_CHANNEL_ORDER_AMBISONIC) {
+ av_log(log_ctx, AV_LOG_ERROR, "Invalid channel layout for SCENE_BASED audio element\n");
+ return AVERROR(EINVAL);
+ }
+ if (layer->ambisonics_mode >= AV_IAMF_AMBISONICS_MODE_PROJECTION) {
+ av_log(log_ctx, AV_LOG_ERROR, "Unsuported ambisonics mode %d\n", layer->ambisonics_mode);
+ return AVERROR_PATCHWELCOME;
+ }
+ for (int i = 0; i < stg->nb_streams; i++) {
+ if (stg->streams[i]->codecpar->ch_layout.nb_channels > 1) {
+ av_log(log_ctx, AV_LOG_ERROR, "Invalid amount of channels in a stream for MONO mode ambisonics\n");
+ return AVERROR(EINVAL);
+ }
+ }
+ } else
+ for (int j, i = 0; i < iamf_audio_element->nb_layers; i++) {
+ const AVIAMFLayer *layer = iamf_audio_element->layers[i];
+ for (j = 0; j < FF_ARRAY_ELEMS(ff_iamf_scalable_ch_layouts); j++)
+ if (!av_channel_layout_compare(&layer->ch_layout, &ff_iamf_scalable_ch_layouts[j]))
+ break;
+
+ if (j >= FF_ARRAY_ELEMS(ff_iamf_scalable_ch_layouts)) {
+ av_log(log_ctx, AV_LOG_ERROR, "Unsupported channel layout in stream group #%d\n", i);
+ return AVERROR(EINVAL);
+ }
+ }
+
+ for (int i = 0; i < iamf->nb_audio_elements; i++) {
+ if (stg->id == iamf->audio_elements[i]->audio_element_id) {
+ av_log(log_ctx, AV_LOG_ERROR, "Duplicated Audio Element id %"PRId64"\n", stg->id);
+ return AVERROR(EINVAL);
+ }
+ }
+
+ codec_config = av_mallocz(sizeof(*codec_config));
+ if (!codec_config)
+ return AVERROR(ENOMEM);
+
+ ret = fill_codec_config(iamf, stg, codec_config);
+ if (ret < 0) {
+ av_free(codec_config);
+ return ret;
+ }
+
+ audio_element = av_mallocz(sizeof(*audio_element));
+ if (!audio_element)
+ return AVERROR(ENOMEM);
+
+ audio_element->element = stg->params.iamf_audio_element;
+ audio_element->audio_element_id = stg->id;
+ audio_element->codec_config_id = ret;
+
+ audio_element->substreams = av_calloc(stg->nb_streams, sizeof(*audio_element->substreams));
+ if (!audio_element->substreams)
+ return AVERROR(ENOMEM);
+ audio_element->nb_substreams = stg->nb_streams;
+
+ audio_element->layers = av_calloc(iamf_audio_element->nb_layers, sizeof(*audio_element->layers));
+ if (!audio_element->layers)
+ return AVERROR(ENOMEM);
+
+ for (int i = 0, j = 0; i < iamf_audio_element->nb_layers; i++) {
+ int nb_channels = iamf_audio_element->layers[i]->ch_layout.nb_channels;
+
+ IAMFLayer *layer = &audio_element->layers[i];
+ if (!layer)
+ return AVERROR(ENOMEM);
+ memset(layer, 0, sizeof(*layer));
+
+ if (i)
+ nb_channels -= iamf_audio_element->layers[i - 1]->ch_layout.nb_channels;
+ for (; nb_channels > 0 && j < stg->nb_streams; j++) {
+ const AVStream *st = stg->streams[j];
+ IAMFSubStream *substream = &audio_element->substreams[j];
+
+ substream->audio_substream_id = st->id;
+ layer->substream_count++;
+ layer->coupled_substream_count += st->codecpar->ch_layout.nb_channels == 2;
+ nb_channels -= st->codecpar->ch_layout.nb_channels;
+ }
+ if (nb_channels) {
+ av_log(log_ctx, AV_LOG_ERROR, "Invalid channel count across substreams in layer %u from stream group %u\n",
+ i, stg->index);
+ return AVERROR(EINVAL);
+ }
+ }
+
+ if (iamf_audio_element->demixing_info) {
+ AVIAMFParamDefinition *param = iamf_audio_element->demixing_info;
+ IAMFParamDefinition *param_definition = ff_iamf_get_param_definition(iamf, param->parameter_id);
+
+ if (param->nb_subblocks != 1) {
+ av_log(log_ctx, AV_LOG_ERROR, "nb_subblocks in demixing_info for stream group %u is not 1\n", stg->index);
+ return AVERROR(EINVAL);
+ }
+
+ if (!param_definition) {
+ param_definition = add_param_definition(iamf, param, audio_element, log_ctx);
+ if (!param_definition)
+ return AVERROR(ENOMEM);
+ }
+ }
+ if (iamf_audio_element->recon_gain_info) {
+ AVIAMFParamDefinition *param = iamf_audio_element->recon_gain_info;
+ IAMFParamDefinition *param_definition = ff_iamf_get_param_definition(iamf, param->parameter_id);
+
+ if (param->nb_subblocks != 1) {
+ av_log(log_ctx, AV_LOG_ERROR, "nb_subblocks in recon_gain_info for stream group %u is not 1\n", stg->index);
+ return AVERROR(EINVAL);
+ }
+
+ if (!param_definition) {
+ param_definition = add_param_definition(iamf, param, audio_element, log_ctx);
+ if (!param_definition)
+ return AVERROR(ENOMEM);
+ }
+ }
+
+ tmp = av_realloc_array(iamf->audio_elements, iamf->nb_audio_elements + 1, sizeof(*iamf->audio_elements));
+ if (!tmp)
+ return AVERROR(ENOMEM);
+
+ iamf->audio_elements = tmp;
+ iamf->audio_elements[iamf->nb_audio_elements++] = audio_element;
+
+ return 0;
+}
+
+int ff_iamf_add_mix_presentation(IAMFContext *iamf, const AVStreamGroup *stg, void *log_ctx)
+{
+ IAMFMixPresentation **tmp, *mix_presentation;
+
+ if (stg->type != AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION)
+ return AVERROR(EINVAL);
+
+ for (int i = 0; i < iamf->nb_mix_presentations; i++) {
+ if (stg->id == iamf->mix_presentations[i]->mix_presentation_id) {
+ av_log(log_ctx, AV_LOG_ERROR, "Duplicate Mix Presentation id %"PRId64"\n", stg->id);
+ return AVERROR(EINVAL);
+ }
+ }
+
+ mix_presentation = av_mallocz(sizeof(*mix_presentation));
+ if (!mix_presentation)
+ return AVERROR(ENOMEM);
+
+ mix_presentation->mix = stg->params.iamf_mix_presentation;
+ mix_presentation->mix_presentation_id = stg->id;
+
+ for (int i = 0; i < mix_presentation->mix->nb_submixes; i++) {
+ const AVIAMFSubmix *submix = mix_presentation->mix->submixes[i];
+ AVIAMFParamDefinition *param = submix->output_mix_config;
+ IAMFParamDefinition *param_definition;
+
+ if (!param) {
+ av_log(log_ctx, AV_LOG_ERROR, "output_mix_config is not present in submix %u from "
+ "Mix Presentation ID %"PRId64"\n", i, stg->id);
+ return AVERROR(EINVAL);
+ }
+
+ param_definition = ff_iamf_get_param_definition(iamf, param->parameter_id);
+ if (!param_definition) {
+ param_definition = add_param_definition(iamf, param, NULL, log_ctx);
+ if (!param_definition)
+ return AVERROR(ENOMEM);
+ }
+
+ for (int j = 0; j < submix->nb_elements; j++) {
+ const AVIAMFSubmixElement *element = submix->elements[j];
+ param = element->element_mix_config;
+
+ if (!param) {
+ av_log(log_ctx, AV_LOG_ERROR, "element_mix_config is not present for element %u in submix %u from "
+ "Mix Presentation ID %"PRId64"\n", j, i, stg->id);
+ return AVERROR(EINVAL);
+ }
+ param_definition = ff_iamf_get_param_definition(iamf, param->parameter_id);
+ if (!param_definition) {
+ param_definition = add_param_definition(iamf, param, NULL, log_ctx);
+ if (!param_definition)
+ return AVERROR(ENOMEM);
+ }
+ }
+ }
+
+ tmp = av_realloc_array(iamf->mix_presentations, iamf->nb_mix_presentations + 1, sizeof(*iamf->mix_presentations));
+ if (!tmp)
+ return AVERROR(ENOMEM);
+
+ iamf->mix_presentations = tmp;
+ iamf->mix_presentations[iamf->nb_mix_presentations++] = mix_presentation;
+
+ return 0;
+}
+
+static int iamf_write_codec_config(const IAMFContext *iamf,
+ const IAMFCodecConfig *codec_config,
+ AVIOContext *pb)
+{
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE];
+ AVIOContext *dyn_bc;
+ uint8_t *dyn_buf = NULL;
+ PutBitContext pbc;
+ int dyn_size;
+
+ int ret = avio_open_dyn_buf(&dyn_bc);
+ if (ret < 0)
+ return ret;
+
+ ffio_write_leb(dyn_bc, codec_config->codec_config_id);
+ avio_wl32(dyn_bc, codec_config->codec_tag);
+
+ ffio_write_leb(dyn_bc, codec_config->nb_samples);
+ avio_wb16(dyn_bc, codec_config->seek_preroll);
+
+ switch(codec_config->codec_id) {
+ case AV_CODEC_ID_OPUS:
+ avio_write(dyn_bc, codec_config->extradata, codec_config->extradata_size);
+ break;
+ case AV_CODEC_ID_AAC:
+ return AVERROR_PATCHWELCOME;
+ case AV_CODEC_ID_FLAC:
+ avio_w8(dyn_bc, 0x80);
+ avio_wb24(dyn_bc, codec_config->extradata_size);
+ avio_write(dyn_bc, codec_config->extradata, codec_config->extradata_size);
+ break;
+ case AV_CODEC_ID_PCM_S16LE:
+ avio_w8(dyn_bc, 0);
+ avio_w8(dyn_bc, 16);
+ avio_wb32(dyn_bc, codec_config->sample_rate);
+ break;
+ case AV_CODEC_ID_PCM_S24LE:
+ avio_w8(dyn_bc, 0);
+ avio_w8(dyn_bc, 24);
+ avio_wb32(dyn_bc, codec_config->sample_rate);
+ break;
+ case AV_CODEC_ID_PCM_S32LE:
+ avio_w8(dyn_bc, 0);
+ avio_w8(dyn_bc, 32);
+ avio_wb32(dyn_bc, codec_config->sample_rate);
+ break;
+ case AV_CODEC_ID_PCM_S16BE:
+ avio_w8(dyn_bc, 1);
+ avio_w8(dyn_bc, 16);
+ avio_wb32(dyn_bc, codec_config->sample_rate);
+ break;
+ case AV_CODEC_ID_PCM_S24BE:
+ avio_w8(dyn_bc, 1);
+ avio_w8(dyn_bc, 24);
+ avio_wb32(dyn_bc, codec_config->sample_rate);
+ break;
+ case AV_CODEC_ID_PCM_S32BE:
+ avio_w8(dyn_bc, 1);
+ avio_w8(dyn_bc, 32);
+ avio_wb32(dyn_bc, codec_config->sample_rate);
+ break;
+ default:
+ break;
+ }
+
+ init_put_bits(&pbc, header, sizeof(header));
+ put_bits(&pbc, 5, IAMF_OBU_IA_CODEC_CONFIG);
+ put_bits(&pbc, 3, 0);
+ flush_put_bits(&pbc);
+
+ dyn_size = avio_close_dyn_buf(dyn_bc, &dyn_buf);
+ avio_write(pb, header, put_bytes_count(&pbc, 1));
+ ffio_write_leb(pb, dyn_size);
+ avio_write(pb, dyn_buf, dyn_size);
+ av_free(dyn_buf);
+
+ return 0;
+}
+
+static inline int rescale_rational(AVRational q, int b)
+{
+ return av_clip_int16(av_rescale(q.num, b, q.den));
+}
+
+static int scalable_channel_layout_config(const IAMFAudioElement *audio_element,
+ AVIOContext *dyn_bc)
+{
+ const AVIAMFAudioElement *element = audio_element->element;
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE];
+ PutBitContext pb;
+
+ init_put_bits(&pb, header, sizeof(header));
+ put_bits(&pb, 3, element->nb_layers);
+ put_bits(&pb, 5, 0);
+ flush_put_bits(&pb);
+ avio_write(dyn_bc, header, put_bytes_count(&pb, 1));
+ for (int i = 0; i < element->nb_layers; i++) {
+ AVIAMFLayer *layer = element->layers[i];
+ int layout;
+ for (layout = 0; layout < FF_ARRAY_ELEMS(ff_iamf_scalable_ch_layouts); layout++) {
+ if (!av_channel_layout_compare(&layer->ch_layout, &ff_iamf_scalable_ch_layouts[layout]))
+ break;
+ }
+ init_put_bits(&pb, header, sizeof(header));
+ put_bits(&pb, 4, layout);
+ put_bits(&pb, 1, !!layer->output_gain_flags);
+ put_bits(&pb, 1, !!(layer->flags & AV_IAMF_LAYER_FLAG_RECON_GAIN));
+ put_bits(&pb, 2, 0); // reserved
+ put_bits(&pb, 8, audio_element->layers[i].substream_count);
+ put_bits(&pb, 8, audio_element->layers[i].coupled_substream_count);
+ if (layer->output_gain_flags) {
+ put_bits(&pb, 6, layer->output_gain_flags);
+ put_bits(&pb, 2, 0);
+ put_bits(&pb, 16, rescale_rational(layer->output_gain, 1 << 8));
+ }
+ flush_put_bits(&pb);
+ avio_write(dyn_bc, header, put_bytes_count(&pb, 1));
+ }
+
+ return 0;
+}
+
+static int ambisonics_config(const IAMFAudioElement *audio_element,
+ AVIOContext *dyn_bc)
+{
+ const AVIAMFAudioElement *element = audio_element->element;
+ AVIAMFLayer *layer = element->layers[0];
+
+ ffio_write_leb(dyn_bc, 0); // ambisonics_mode
+ ffio_write_leb(dyn_bc, layer->ch_layout.nb_channels); // output_channel_count
+ ffio_write_leb(dyn_bc, audio_element->nb_substreams); // substream_count
+
+ if (layer->ch_layout.order == AV_CHANNEL_ORDER_AMBISONIC)
+ for (int i = 0; i < layer->ch_layout.nb_channels; i++)
+ avio_w8(dyn_bc, i);
+ else
+ for (int i = 0; i < layer->ch_layout.nb_channels; i++)
+ avio_w8(dyn_bc, layer->ch_layout.u.map[i].id);
+
+ return 0;
+}
+
+static int param_definition(const IAMFContext *iamf,
+ const IAMFParamDefinition *param_def,
+ AVIOContext *dyn_bc, void *log_ctx)
+{
+ const AVIAMFParamDefinition *param = param_def->param;
+
+ ffio_write_leb(dyn_bc, param->parameter_id);
+ ffio_write_leb(dyn_bc, param->parameter_rate);
+ avio_w8(dyn_bc, param->duration ? 0 : 1 << 7);
+ if (param->duration) {
+ ffio_write_leb(dyn_bc, param->duration);
+ ffio_write_leb(dyn_bc, param->constant_subblock_duration);
+ if (param->constant_subblock_duration == 0) {
+ ffio_write_leb(dyn_bc, param->nb_subblocks);
+ for (int i = 0; i < param->nb_subblocks; i++) {
+ const void *subblock = av_iamf_param_definition_get_subblock(param, i);
+
+ switch (param->type) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN: {
+ const AVIAMFMixGain *mix = subblock;
+ ffio_write_leb(dyn_bc, mix->subblock_duration);
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING: {
+ const AVIAMFDemixingInfo *demix = subblock;
+ ffio_write_leb(dyn_bc, demix->subblock_duration);
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN: {
+ const AVIAMFReconGain *recon = subblock;
+ ffio_write_leb(dyn_bc, recon->subblock_duration);
+ break;
+ }
+ }
+ }
+ }
+ }
+
+ return 0;
+}
+
+static int iamf_write_audio_element(const IAMFContext *iamf,
+ const IAMFAudioElement *audio_element,
+ AVIOContext *pb, void *log_ctx)
+{
+ const AVIAMFAudioElement *element = audio_element->element;
+ const IAMFCodecConfig *codec_config = iamf->codec_configs[audio_element->codec_config_id];
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE];
+ AVIOContext *dyn_bc;
+ uint8_t *dyn_buf = NULL;
+ PutBitContext pbc;
+ int param_definition_types = AV_IAMF_PARAMETER_DEFINITION_DEMIXING, dyn_size;
+
+ int ret = avio_open_dyn_buf(&dyn_bc);
+ if (ret < 0)
+ return ret;
+
+ ffio_write_leb(dyn_bc, audio_element->audio_element_id);
+
+ init_put_bits(&pbc, header, sizeof(header));
+ put_bits(&pbc, 3, element->audio_element_type);
+ put_bits(&pbc, 5, 0);
+ flush_put_bits(&pbc);
+ avio_write(dyn_bc, header, put_bytes_count(&pbc, 1));
+
+ ffio_write_leb(dyn_bc, audio_element->codec_config_id);
+ ffio_write_leb(dyn_bc, audio_element->nb_substreams);
+
+ for (int i = 0; i < audio_element->nb_substreams; i++)
+ ffio_write_leb(dyn_bc, audio_element->substreams[i].audio_substream_id);
+
+ if (element->nb_layers == 1)
+ param_definition_types &= ~AV_IAMF_PARAMETER_DEFINITION_DEMIXING;
+ if (element->nb_layers > 1)
+ param_definition_types |= AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN;
+ if (codec_config->codec_tag == MKTAG('f','L','a','C') ||
+ codec_config->codec_tag == MKTAG('i','p','c','m'))
+ param_definition_types &= ~AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN;
+
+ ffio_write_leb(dyn_bc, av_popcount(param_definition_types)); // num_parameters
+
+ if (param_definition_types & 1) {
+ const AVIAMFParamDefinition *param = element->demixing_info;
+ const IAMFParamDefinition *param_def;
+ const AVIAMFDemixingInfo *demix;
+
+ if (!param) {
+ av_log(log_ctx, AV_LOG_ERROR, "demixing_info needed but not set in Stream Group #%u\n",
+ audio_element->audio_element_id);
+ return AVERROR(EINVAL);
+ }
+
+ demix = av_iamf_param_definition_get_subblock(param, 0);
+ ffio_write_leb(dyn_bc, AV_IAMF_PARAMETER_DEFINITION_DEMIXING); // type
+
+ param_def = ff_iamf_get_param_definition(iamf, param->parameter_id);
+ ret = param_definition(iamf, param_def, dyn_bc, log_ctx);
+ if (ret < 0)
+ return ret;
+
+ avio_w8(dyn_bc, demix->dmixp_mode << 5); // dmixp_mode
+ avio_w8(dyn_bc, element->default_w << 4); // default_w
+ }
+ if (param_definition_types & 2) {
+ const AVIAMFParamDefinition *param = element->recon_gain_info;
+ const IAMFParamDefinition *param_def;
+
+ if (!param) {
+ av_log(log_ctx, AV_LOG_ERROR, "recon_gain_info needed but not set in Stream Group #%u\n",
+ audio_element->audio_element_id);
+ return AVERROR(EINVAL);
+ }
+ ffio_write_leb(dyn_bc, AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN); // type
+
+ param_def = ff_iamf_get_param_definition(iamf, param->parameter_id);
+ ret = param_definition(iamf, param_def, dyn_bc, log_ctx);
+ if (ret < 0)
+ return ret;
+ }
+
+ if (element->audio_element_type == AV_IAMF_AUDIO_ELEMENT_TYPE_CHANNEL) {
+ ret = scalable_channel_layout_config(audio_element, dyn_bc);
+ if (ret < 0)
+ return ret;
+ } else {
+ ret = ambisonics_config(audio_element, dyn_bc);
+ if (ret < 0)
+ return ret;
+ }
+
+ init_put_bits(&pbc, header, sizeof(header));
+ put_bits(&pbc, 5, IAMF_OBU_IA_AUDIO_ELEMENT);
+ put_bits(&pbc, 3, 0);
+ flush_put_bits(&pbc);
+
+ dyn_size = avio_close_dyn_buf(dyn_bc, &dyn_buf);
+ avio_write(pb, header, put_bytes_count(&pbc, 1));
+ ffio_write_leb(pb, dyn_size);
+ avio_write(pb, dyn_buf, dyn_size);
+ av_free(dyn_buf);
+
+ return 0;
+}
+
+static int iamf_write_mixing_presentation(const IAMFContext *iamf,
+ const IAMFMixPresentation *mix_presentation,
+ AVIOContext *pb, void *log_ctx)
+{
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE];
+ const AVIAMFMixPresentation *mix = mix_presentation->mix;
+ const AVDictionaryEntry *tag = NULL;
+ PutBitContext pbc;
+ AVIOContext *dyn_bc;
+ uint8_t *dyn_buf = NULL;
+ int dyn_size;
+
+ int ret = avio_open_dyn_buf(&dyn_bc);
+ if (ret < 0)
+ return ret;
+
+ ffio_write_leb(dyn_bc, mix_presentation->mix_presentation_id); // mix_presentation_id
+ ffio_write_leb(dyn_bc, av_dict_count(mix->annotations)); // count_label
+
+ while ((tag = av_dict_iterate(mix->annotations, tag)))
+ avio_put_str(dyn_bc, tag->key);
+ while ((tag = av_dict_iterate(mix->annotations, tag)))
+ avio_put_str(dyn_bc, tag->value);
+
+ ffio_write_leb(dyn_bc, mix->nb_submixes);
+ for (int i = 0; i < mix->nb_submixes; i++) {
+ const AVIAMFSubmix *sub_mix = mix->submixes[i];
+ const IAMFParamDefinition *param_def;
+
+ ffio_write_leb(dyn_bc, sub_mix->nb_elements);
+ for (int j = 0; j < sub_mix->nb_elements; j++) {
+ const IAMFAudioElement *audio_element = NULL;
+ const AVIAMFSubmixElement *submix_element = sub_mix->elements[j];
+
+ for (int k = 0; k < iamf->nb_audio_elements; k++)
+ if (iamf->audio_elements[k]->audio_element_id == submix_element->audio_element_id) {
+ audio_element = iamf->audio_elements[k];
+ break;
+ }
+
+ av_assert0(audio_element);
+ ffio_write_leb(dyn_bc, submix_element->audio_element_id);
+
+ if (av_dict_count(submix_element->annotations) != av_dict_count(mix->annotations)) {
+ av_log(log_ctx, AV_LOG_ERROR, "Inconsistent amount of labels in submix %d from Mix Presentation id #%u\n",
+ j, audio_element->audio_element_id);
+ return AVERROR(EINVAL);
+ }
+ while ((tag = av_dict_iterate(submix_element->annotations, tag)))
+ avio_put_str(dyn_bc, tag->value);
+
+ init_put_bits(&pbc, header, sizeof(header));
+ put_bits(&pbc, 2, submix_element->headphones_rendering_mode);
+ put_bits(&pbc, 6, 0); // reserved
+ flush_put_bits(&pbc);
+ avio_write(dyn_bc, header, put_bytes_count(&pbc, 1));
+ ffio_write_leb(dyn_bc, 0); // rendering_config_extension_size
+
+ param_def = ff_iamf_get_param_definition(iamf, submix_element->element_mix_config->parameter_id);
+ ret = param_definition(iamf, param_def, dyn_bc, log_ctx);
+ if (ret < 0)
+ return ret;
+
+ avio_wb16(dyn_bc, rescale_rational(submix_element->default_mix_gain, 1 << 8));
+ }
+
+ param_def = ff_iamf_get_param_definition(iamf, sub_mix->output_mix_config->parameter_id);
+ ret = param_definition(iamf, param_def, dyn_bc, log_ctx);
+ if (ret < 0)
+ return ret;
+ avio_wb16(dyn_bc, rescale_rational(sub_mix->default_mix_gain, 1 << 8));
+
+ ffio_write_leb(dyn_bc, sub_mix->nb_layouts); // nb_layouts
+ for (int i = 0; i < sub_mix->nb_layouts; i++) {
+ const AVIAMFSubmixLayout *submix_layout = sub_mix->layouts[i];
+ int layout, info_type;
+ int dialogue = submix_layout->dialogue_anchored_loudness.num &&
+ submix_layout->dialogue_anchored_loudness.den;
+ int album = submix_layout->album_anchored_loudness.num &&
+ submix_layout->album_anchored_loudness.den;
+
+ if (layout == FF_ARRAY_ELEMS(ff_iamf_sound_system_map)) {
+ av_log(log_ctx, AV_LOG_ERROR, "Invalid Sound System value in a submix\n");
+ return AVERROR(EINVAL);
+ }
+
+ if (submix_layout->layout_type == AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS) {
+ for (layout = 0; layout < FF_ARRAY_ELEMS(ff_iamf_sound_system_map); layout++) {
+ if (!av_channel_layout_compare(&submix_layout->sound_system, &ff_iamf_sound_system_map[layout].layout))
+ break;
+ }
+ if (layout == FF_ARRAY_ELEMS(ff_iamf_sound_system_map)) {
+ av_log(log_ctx, AV_LOG_ERROR, "Invalid Sound System value in a submix\n");
+ return AVERROR(EINVAL);
+ }
+ }
+ init_put_bits(&pbc, header, sizeof(header));
+ put_bits(&pbc, 2, submix_layout->layout_type); // layout_type
+ if (submix_layout->layout_type == AV_IAMF_SUBMIX_LAYOUT_TYPE_LOUDSPEAKERS) {
+ put_bits(&pbc, 4, ff_iamf_sound_system_map[layout].id); // sound_system
+ put_bits(&pbc, 2, 0); // reserved
+ } else
+ put_bits(&pbc, 6, 0); // reserved
+ flush_put_bits(&pbc);
+ avio_write(dyn_bc, header, put_bytes_count(&pbc, 1));
+
+ info_type = (submix_layout->true_peak.num && submix_layout->true_peak.den);
+ info_type |= (dialogue || album) << 1;
+ avio_w8(dyn_bc, info_type);
+ avio_wb16(dyn_bc, rescale_rational(submix_layout->integrated_loudness, 1 << 8));
+ avio_wb16(dyn_bc, rescale_rational(submix_layout->digital_peak, 1 << 8));
+ if (info_type & 1)
+ avio_wb16(dyn_bc, rescale_rational(submix_layout->true_peak, 1 << 8));
+ if (info_type & 2) {
+ avio_w8(dyn_bc, dialogue + album); // num_anchored_loudness
+ if (dialogue) {
+ avio_w8(dyn_bc, IAMF_ANCHOR_ELEMENT_DIALOGUE);
+ avio_wb16(dyn_bc, rescale_rational(submix_layout->dialogue_anchored_loudness, 1 << 8));
+ }
+ if (album) {
+ avio_w8(dyn_bc, IAMF_ANCHOR_ELEMENT_ALBUM);
+ avio_wb16(dyn_bc, rescale_rational(submix_layout->album_anchored_loudness, 1 << 8));
+ }
+ }
+ }
+ }
+
+ init_put_bits(&pbc, header, sizeof(header));
+ put_bits(&pbc, 5, IAMF_OBU_IA_MIX_PRESENTATION);
+ put_bits(&pbc, 3, 0);
+ flush_put_bits(&pbc);
+
+ dyn_size = avio_close_dyn_buf(dyn_bc, &dyn_buf);
+ avio_write(pb, header, put_bytes_count(&pbc, 1));
+ ffio_write_leb(pb, dyn_size);
+ avio_write(pb, dyn_buf, dyn_size);
+ av_free(dyn_buf);
+
+ return 0;
+}
+
+int ff_iamf_write_descriptors(const IAMFContext *iamf, AVIOContext *pb, void *log_ctx)
+{
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE];
+ PutBitContext pbc;
+ AVIOContext *dyn_bc;
+ uint8_t *dyn_buf = NULL;
+ int dyn_size;
+
+ int ret = avio_open_dyn_buf(&dyn_bc);
+ if (ret < 0)
+ return ret;
+
+ // Sequence Header
+ init_put_bits(&pbc, header, sizeof(header));
+ put_bits(&pbc, 5, IAMF_OBU_IA_SEQUENCE_HEADER);
+ put_bits(&pbc, 3, 0);
+ flush_put_bits(&pbc);
+
+ avio_write(dyn_bc, header, put_bytes_count(&pbc, 1));
+ ffio_write_leb(dyn_bc, 6);
+ avio_wb32(dyn_bc, MKBETAG('i','a','m','f'));
+ avio_w8(dyn_bc, iamf->nb_audio_elements > 1); // primary_profile
+ avio_w8(dyn_bc, iamf->nb_audio_elements > 1); // additional_profile
+
+ dyn_size = avio_close_dyn_buf(dyn_bc, &dyn_buf);
+ avio_write(pb, dyn_buf, dyn_size);
+ av_free(dyn_buf);
+
+ for (int i = 0; i < iamf->nb_codec_configs; i++) {
+ ret = iamf_write_codec_config(iamf, iamf->codec_configs[i], pb);
+ if (ret < 0)
+ return ret;
+ }
+
+ for (int i = 0; i < iamf->nb_audio_elements; i++) {
+ ret = iamf_write_audio_element(iamf, iamf->audio_elements[i], pb, log_ctx);
+ if (ret < 0)
+ return ret;
+ }
+
+ for (int i = 0; i < iamf->nb_mix_presentations; i++) {
+ ret = iamf_write_mixing_presentation(iamf, iamf->mix_presentations[i], pb, log_ctx);
+ if (ret < 0)
+ return ret;
+ }
+
+ return 0;
+}
diff --git a/libavformat/iamf_writer.h b/libavformat/iamf_writer.h
new file mode 100644
index 0000000000..93354670b8
--- /dev/null
+++ b/libavformat/iamf_writer.h
@@ -0,0 +1,51 @@
+/*
+ * Immersive Audio Model and Formats muxing helpers and structs
+ * Copyright (c) 2023 James Almer <jamrial@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
+ */
+
+#ifndef AVFORMAT_IAMF_WRITER_H
+#define AVFORMAT_IAMF_WRITER_H
+
+#include <stdint.h>
+
+#include "libavutil/common.h"
+#include "avformat.h"
+#include "avio.h"
+#include "iamf.h"
+
+static inline IAMFParamDefinition *ff_iamf_get_param_definition(const IAMFContext *iamf,
+ unsigned int parameter_id)
+{
+ IAMFParamDefinition *param_definition = NULL;
+
+ for (int i = 0; i < iamf->nb_param_definitions; i++)
+ if (iamf->param_definitions[i]->param->parameter_id == parameter_id) {
+ param_definition = iamf->param_definitions[i];
+ break;
+ }
+
+ return param_definition;
+}
+
+int ff_iamf_add_audio_element(IAMFContext *iamf, const AVStreamGroup *stg, void *log_ctx);
+int ff_iamf_add_mix_presentation(IAMFContext *iamf, const AVStreamGroup *stg, void *log_ctx);
+
+int ff_iamf_write_descriptors(const IAMFContext *iamf, AVIOContext *pb, void *log_ctx);
+
+#endif /* AVFORMAT_IAMF_WRITER_H */
diff --git a/libavformat/iamfenc.c b/libavformat/iamfenc.c
new file mode 100644
index 0000000000..0a043ce3a0
--- /dev/null
+++ b/libavformat/iamfenc.c
@@ -0,0 +1,387 @@
+/*
+ * IAMF muxer
+ * Copyright (c) 2023 James Almer
+ *
+ * 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 <stdint.h>
+
+#include "libavutil/avassert.h"
+#include "libavutil/common.h"
+#include "libavutil/iamf.h"
+#include "libavcodec/get_bits.h"
+#include "libavcodec/put_bits.h"
+#include "avformat.h"
+#include "avio_internal.h"
+#include "iamf.h"
+#include "iamf_writer.h"
+#include "internal.h"
+#include "mux.h"
+
+typedef struct IAMFMuxContext {
+ IAMFContext iamf;
+
+ int first_stream_id;
+} IAMFMuxContext;
+
+static int iamf_init(AVFormatContext *s)
+{
+ IAMFMuxContext *const c = s->priv_data;
+ IAMFContext *const iamf = &c->iamf;
+ int nb_audio_elements = 0, nb_mix_presentations = 0;
+ int ret;
+
+ if (!s->nb_streams) {
+ av_log(s, AV_LOG_ERROR, "There must be at least one stream\n");
+ return AVERROR(EINVAL);
+ }
+
+ for (int i = 0; i < s->nb_streams; i++) {
+ if (s->streams[i]->codecpar->codec_type != AVMEDIA_TYPE_AUDIO ||
+ (s->streams[i]->codecpar->codec_tag != MKTAG('m','p','4','a') &&
+ s->streams[i]->codecpar->codec_tag != MKTAG('O','p','u','s') &&
+ s->streams[i]->codecpar->codec_tag != MKTAG('f','L','a','C') &&
+ s->streams[i]->codecpar->codec_tag != MKTAG('i','p','c','m'))) {
+ av_log(s, AV_LOG_ERROR, "Unsupported codec id %s\n",
+ avcodec_get_name(s->streams[i]->codecpar->codec_id));
+ return AVERROR(EINVAL);
+ }
+
+ if (s->streams[i]->codecpar->ch_layout.nb_channels > 2) {
+ av_log(s, AV_LOG_ERROR, "Unsupported channel layout on stream #%d\n", i);
+ return AVERROR(EINVAL);
+ }
+
+ for (int j = 0; j < i; j++) {
+ if (s->streams[i]->id == s->streams[j]->id) {
+ av_log(s, AV_LOG_ERROR, "Duplicated stream id %d\n", s->streams[j]->id);
+ return AVERROR(EINVAL);
+ }
+ }
+ }
+
+ if (!s->nb_stream_groups) {
+ av_log(s, AV_LOG_ERROR, "There must be at least two stream groups\n");
+ return AVERROR(EINVAL);
+ }
+
+ for (int i = 0; i < s->nb_stream_groups; i++) {
+ const AVStreamGroup *stg = s->stream_groups[i];
+
+ if (stg->type == AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT)
+ nb_audio_elements++;
+ if (stg->type == AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION)
+ nb_mix_presentations++;
+ }
+ if ((nb_audio_elements < 1 && nb_audio_elements > 2) || nb_mix_presentations < 1) {
+ av_log(s, AV_LOG_ERROR, "There must be >= 1 and <= 2 IAMF_AUDIO_ELEMENT and at least "
+ "one IAMF_MIX_PRESENTATION stream groups\n");
+ return AVERROR(EINVAL);
+ }
+
+ for (int i = 0; i < s->nb_stream_groups; i++) {
+ const AVStreamGroup *stg = s->stream_groups[i];
+ if (stg->type != AV_STREAM_GROUP_PARAMS_IAMF_AUDIO_ELEMENT)
+ continue;
+
+ ret = ff_iamf_add_audio_element(iamf, stg, s);
+ if (ret < 0)
+ return ret;
+ }
+
+ for (int i = 0; i < s->nb_stream_groups; i++) {
+ const AVStreamGroup *stg = s->stream_groups[i];
+ if (stg->type != AV_STREAM_GROUP_PARAMS_IAMF_MIX_PRESENTATION)
+ continue;
+
+ ret = ff_iamf_add_mix_presentation(iamf, stg, s);
+ if (ret < 0)
+ return ret;
+ }
+
+ c->first_stream_id = s->streams[0]->id;
+
+ return 0;
+}
+
+static int iamf_write_header(AVFormatContext *s)
+{
+ IAMFMuxContext *const c = s->priv_data;
+ IAMFContext *const iamf = &c->iamf;
+ int ret;
+
+ ret = ff_iamf_write_descriptors(iamf, s->pb, s);
+ if (ret < 0)
+ return ret;
+
+ c->first_stream_id = s->streams[0]->id;
+
+ return 0;
+}
+
+static inline int rescale_rational(AVRational q, int b)
+{
+ return av_clip_int16(av_rescale(q.num, b, q.den));
+}
+
+static int write_parameter_block(AVFormatContext *s, const AVIAMFParamDefinition *param)
+{
+ const IAMFMuxContext *const c = s->priv_data;
+ const IAMFContext *const iamf = &c->iamf;
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE];
+ IAMFParamDefinition *param_definition = ff_iamf_get_param_definition(iamf, param->parameter_id);
+ PutBitContext pb;
+ AVIOContext *dyn_bc;
+ uint8_t *dyn_buf = NULL;
+ int dyn_size, ret;
+
+ if (param->type > AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN) {
+ av_log(s, AV_LOG_DEBUG, "Ignoring side data with unknown type %u\n",
+ param->type);
+ return 0;
+ }
+
+ if (!param_definition) {
+ av_log(s, AV_LOG_ERROR, "Non-existent Parameter Definition with ID %u referenced by a packet\n",
+ param->parameter_id);
+ return AVERROR(EINVAL);
+ }
+
+ if (param->type != param_definition->param->type) {
+ av_log(s, AV_LOG_ERROR, "Inconsistent values for Parameter Definition "
+ "with ID %u in a packet\n",
+ param->parameter_id);
+ return AVERROR(EINVAL);
+ }
+
+ ret = avio_open_dyn_buf(&dyn_bc);
+ if (ret < 0)
+ return ret;
+
+ // Sequence Header
+ init_put_bits(&pb, header, sizeof(header));
+ put_bits(&pb, 5, IAMF_OBU_IA_PARAMETER_BLOCK);
+ put_bits(&pb, 3, 0);
+ flush_put_bits(&pb);
+ avio_write(s->pb, header, put_bytes_count(&pb, 1));
+
+ ffio_write_leb(dyn_bc, param->parameter_id);
+ if (!param_definition->mode) {
+ ffio_write_leb(dyn_bc, param->duration);
+ ffio_write_leb(dyn_bc, param->constant_subblock_duration);
+ if (param->constant_subblock_duration == 0)
+ ffio_write_leb(dyn_bc, param->nb_subblocks);
+ }
+
+ for (int i = 0; i < param->nb_subblocks; i++) {
+ const void *subblock = av_iamf_param_definition_get_subblock(param, i);
+
+ switch (param->type) {
+ case AV_IAMF_PARAMETER_DEFINITION_MIX_GAIN: {
+ const AVIAMFMixGain *mix = subblock;
+ if (!param_definition->mode && param->constant_subblock_duration == 0)
+ ffio_write_leb(dyn_bc, mix->subblock_duration);
+
+ ffio_write_leb(dyn_bc, mix->animation_type);
+
+ avio_wb16(dyn_bc, rescale_rational(mix->start_point_value, 1 << 8));
+ if (mix->animation_type >= AV_IAMF_ANIMATION_TYPE_LINEAR)
+ avio_wb16(dyn_bc, rescale_rational(mix->end_point_value, 1 << 8));
+ if (mix->animation_type == AV_IAMF_ANIMATION_TYPE_BEZIER) {
+ avio_wb16(dyn_bc, rescale_rational(mix->control_point_value, 1 << 8));
+ avio_w8(dyn_bc, av_clip_uint8(av_rescale(mix->control_point_relative_time.num, 1 << 8,
+ mix->control_point_relative_time.den)));
+ }
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_DEMIXING: {
+ const AVIAMFDemixingInfo *demix = subblock;
+ if (!param_definition->mode && param->constant_subblock_duration == 0)
+ ffio_write_leb(dyn_bc, demix->subblock_duration);
+
+ avio_w8(dyn_bc, demix->dmixp_mode << 5);
+ break;
+ }
+ case AV_IAMF_PARAMETER_DEFINITION_RECON_GAIN: {
+ const AVIAMFReconGain *recon = subblock;
+ const AVIAMFAudioElement *audio_element = param_definition->audio_element->element;
+
+ if (!param_definition->mode && param->constant_subblock_duration == 0)
+ ffio_write_leb(dyn_bc, recon->subblock_duration);
+
+ if (!audio_element) {
+ av_log(s, AV_LOG_ERROR, "Invalid Parameter Definition with ID %u referenced by a packet\n", param->parameter_id);
+ return AVERROR(EINVAL);
+ }
+
+ for (int j = 0; j < audio_element->nb_layers; j++) {
+ const AVIAMFLayer *layer = audio_element->layers[j];
+
+ if (layer->flags & AV_IAMF_LAYER_FLAG_RECON_GAIN) {
+ unsigned int recon_gain_flags = 0;
+ int k = 0;
+
+ for (; k < 7; k++)
+ recon_gain_flags |= (1 << k) * !!recon->recon_gain[j][k];
+ for (; k < 12; k++)
+ recon_gain_flags |= (2 << k) * !!recon->recon_gain[j][k];
+ if (recon_gain_flags >> 8)
+ recon_gain_flags |= (1 << k);
+
+ ffio_write_leb(dyn_bc, recon_gain_flags);
+ for (k = 0; k < 12; k++) {
+ if (recon->recon_gain[j][k])
+ avio_w8(dyn_bc, recon->recon_gain[j][k]);
+ }
+ }
+ }
+ break;
+ }
+ default:
+ av_assert0(0);
+ }
+ }
+
+ dyn_size = avio_close_dyn_buf(dyn_bc, &dyn_buf);
+ ffio_write_leb(s->pb, dyn_size);
+ avio_write(s->pb, dyn_buf, dyn_size);
+ av_free(dyn_buf);
+
+ return 0;
+}
+
+static int iamf_write_packet(AVFormatContext *s, AVPacket *pkt)
+{
+ const IAMFMuxContext *const c = s->priv_data;
+ AVStream *st = s->streams[pkt->stream_index];
+ uint8_t header[MAX_IAMF_OBU_HEADER_SIZE];
+ PutBitContext pb;
+ AVIOContext *dyn_bc;
+ uint8_t *side_data, *dyn_buf = NULL;
+ unsigned int skip_samples = 0, discard_padding = 0;
+ size_t side_data_size;
+ int dyn_size, type = st->id <= 17 ? st->id + IAMF_OBU_IA_AUDIO_FRAME_ID0 : IAMF_OBU_IA_AUDIO_FRAME;
+ int ret;
+
+ if (s->nb_stream_groups && st->id == c->first_stream_id) {
+ AVIAMFParamDefinition *mix =
+ (AVIAMFParamDefinition *)av_packet_get_side_data(pkt, AV_PKT_DATA_IAMF_MIX_GAIN_PARAM, NULL);
+ AVIAMFParamDefinition *demix =
+ (AVIAMFParamDefinition *)av_packet_get_side_data(pkt, AV_PKT_DATA_IAMF_DEMIXING_INFO_PARAM, NULL);
+ AVIAMFParamDefinition *recon =
+ (AVIAMFParamDefinition *)av_packet_get_side_data(pkt, AV_PKT_DATA_IAMF_RECON_GAIN_INFO_PARAM, NULL);
+
+ if (mix) {
+ ret = write_parameter_block(s, mix);
+ if (ret < 0)
+ return ret;
+ }
+ if (demix) {
+ ret = write_parameter_block(s, demix);
+ if (ret < 0)
+ return ret;
+ }
+ if (recon) {
+ ret = write_parameter_block(s, recon);
+ if (ret < 0)
+ return ret;
+ }
+ }
+ side_data = av_packet_get_side_data(pkt, AV_PKT_DATA_SKIP_SAMPLES,
+ &side_data_size);
+
+ if (side_data && side_data_size >= 10) {
+ skip_samples = AV_RL32(side_data);
+ discard_padding = AV_RL32(side_data + 4);
+ }
+
+ ret = avio_open_dyn_buf(&dyn_bc);
+ if (ret < 0)
+ return ret;
+
+ init_put_bits(&pb, header, sizeof(header));
+ put_bits(&pb, 5, type);
+ put_bits(&pb, 1, 0); // obu_redundant_copy
+ put_bits(&pb, 1, skip_samples || discard_padding);
+ put_bits(&pb, 1, 0); // obu_extension_flag
+ flush_put_bits(&pb);
+ avio_write(s->pb, header, put_bytes_count(&pb, 1));
+
+ if (skip_samples || discard_padding) {
+ ffio_write_leb(dyn_bc, discard_padding);
+ ffio_write_leb(dyn_bc, skip_samples);
+ }
+
+ if (st->id > 17)
+ ffio_write_leb(dyn_bc, st->id);
+
+ dyn_size = avio_close_dyn_buf(dyn_bc, &dyn_buf);
+ ffio_write_leb(s->pb, dyn_size + pkt->size);
+ avio_write(s->pb, dyn_buf, dyn_size);
+ av_free(dyn_buf);
+ avio_write(s->pb, pkt->data, pkt->size);
+
+ return 0;
+}
+
+static void iamf_deinit(AVFormatContext *s)
+{
+ IAMFMuxContext *const c = s->priv_data;
+ IAMFContext *const iamf = &c->iamf;
+
+ for (int i = 0; i < iamf->nb_audio_elements; i++) {
+ IAMFAudioElement *audio_element = iamf->audio_elements[i];
+ audio_element->element = NULL;
+ }
+
+ for (int i = 0; i < iamf->nb_mix_presentations; i++) {
+ IAMFMixPresentation *mix_presentation = iamf->mix_presentations[i];
+ mix_presentation->mix = NULL;
+ }
+
+ ff_iamf_uninit_context(iamf);
+
+ return;
+}
+
+static const AVCodecTag iamf_codec_tags[] = {
+ { AV_CODEC_ID_AAC, MKTAG('m','p','4','a') },
+ { AV_CODEC_ID_FLAC, MKTAG('f','L','a','C') },
+ { AV_CODEC_ID_OPUS, MKTAG('O','p','u','s') },
+ { AV_CODEC_ID_PCM_S16LE, MKTAG('i','p','c','m') },
+ { AV_CODEC_ID_PCM_S16BE, MKTAG('i','p','c','m') },
+ { AV_CODEC_ID_PCM_S24LE, MKTAG('i','p','c','m') },
+ { AV_CODEC_ID_PCM_S24BE, MKTAG('i','p','c','m') },
+ { AV_CODEC_ID_PCM_S32LE, MKTAG('i','p','c','m') },
+ { AV_CODEC_ID_PCM_S32BE, MKTAG('i','p','c','m') },
+ { AV_CODEC_ID_NONE, MKTAG('i','p','c','m') }
+};
+
+const FFOutputFormat ff_iamf_muxer = {
+ .p.name = "iamf",
+ .p.long_name = NULL_IF_CONFIG_SMALL("Raw Immersive Audio Model and Formats"),
+ .p.extensions = "iamf",
+ .priv_data_size = sizeof(IAMFMuxContext),
+ .p.audio_codec = AV_CODEC_ID_OPUS,
+ .init = iamf_init,
+ .deinit = iamf_deinit,
+ .write_header = iamf_write_header,
+ .write_packet = iamf_write_packet,
+ .p.codec_tag = (const AVCodecTag* const []){ iamf_codec_tags, NULL },
+ .p.flags = AVFMT_GLOBALHEADER | AVFMT_NOTIMESTAMPS,
+};
--
2.43.0
_______________________________________________
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] 12+ messages in thread