上传完整项目文件

This commit is contained in:
xhgit committed 2026-06-21 14:26:15 +08:00
1 parent e63d58df8d
commit aa0dd45839
8663 files changed
+2393125

No files matched your search

+50
View File
@@ -0,0 +1,50 @@
clean::
$(RM) $(CLEANSUFFIXES:%=libavcodec/bsf/%)
OBJS-$(CONFIG_AAC_ADTSTOASC_BSF) += bsf/aac_adtstoasc.o
OBJS-$(CONFIG_AV1_FRAME_MERGE_BSF) += bsf/av1_frame_merge.o
OBJS-$(CONFIG_AV1_FRAME_SPLIT_BSF) += bsf/av1_frame_split.o
OBJS-$(CONFIG_AV1_METADATA_BSF) += bsf/av1_metadata.o
OBJS-$(CONFIG_CHOMP_BSF) += bsf/chomp.o
OBJS-$(CONFIG_DCA_CORE_BSF) += bsf/dca_core.o
OBJS-$(CONFIG_DTS2PTS_BSF) += bsf/dts2pts.o
OBJS-$(CONFIG_DUMP_EXTRADATA_BSF) += bsf/dump_extradata.o
OBJS-$(CONFIG_DV_ERROR_MARKER_BSF) += bsf/dv_error_marker.o
OBJS-$(CONFIG_EAC3_CORE_BSF) += bsf/eac3_core.o
OBJS-$(CONFIG_EVC_FRAME_MERGE_BSF) += bsf/evc_frame_merge.o
OBJS-$(CONFIG_EXTRACT_EXTRADATA_BSF) += bsf/extract_extradata.o
OBJS-$(CONFIG_FILTER_UNITS_BSF) += bsf/filter_units.o
OBJS-$(CONFIG_H264_METADATA_BSF) += bsf/h264_metadata.o
OBJS-$(CONFIG_H264_MP4TOANNEXB_BSF) += bsf/h264_mp4toannexb.o
OBJS-$(CONFIG_H264_REDUNDANT_PPS_BSF) += bsf/h264_redundant_pps.o
OBJS-$(CONFIG_HAPQA_EXTRACT_BSF) += bsf/hapqa_extract.o
OBJS-$(CONFIG_HEVC_METADATA_BSF) += bsf/h265_metadata.o
OBJS-$(CONFIG_DOVI_RPU_BSF) += bsf/dovi_rpu.o
OBJS-$(CONFIG_HEVC_MP4TOANNEXB_BSF) += bsf/hevc_mp4toannexb.o
OBJS-$(CONFIG_IMX_DUMP_HEADER_BSF) += bsf/imx_dump_header.o
OBJS-$(CONFIG_MEDIA100_TO_MJPEGB_BSF) += bsf/media100_to_mjpegb.o
OBJS-$(CONFIG_MJPEG2JPEG_BSF) += bsf/mjpeg2jpeg.o
OBJS-$(CONFIG_MJPEGA_DUMP_HEADER_BSF) += bsf/mjpega_dump_header.o
OBJS-$(CONFIG_MOV2TEXTSUB_BSF) += bsf/movsub.o
OBJS-$(CONFIG_MPEG2_METADATA_BSF) += bsf/mpeg2_metadata.o
OBJS-$(CONFIG_MPEG4_UNPACK_BFRAMES_BSF) += bsf/mpeg4_unpack_bframes.o
OBJS-$(CONFIG_NOISE_BSF) += bsf/noise.o
OBJS-$(CONFIG_NULL_BSF) += bsf/null.o
OBJS-$(CONFIG_OPUS_METADATA_BSF) += bsf/opus_metadata.o
OBJS-$(CONFIG_PCM_RECHUNK_BSF) += bsf/pcm_rechunk.o
OBJS-$(CONFIG_PGS_FRAME_MERGE_BSF) += bsf/pgs_frame_merge.o
OBJS-$(CONFIG_PRORES_METADATA_BSF) += bsf/prores_metadata.o
OBJS-$(CONFIG_REMOVE_EXTRADATA_BSF) += bsf/remove_extradata.o
OBJS-$(CONFIG_SETTS_BSF) += bsf/setts.o
OBJS-$(CONFIG_SHOWINFO_BSF) += bsf/showinfo.o
OBJS-$(CONFIG_TEXT2MOVSUB_BSF) += bsf/movsub.o
OBJS-$(CONFIG_TRACE_HEADERS_BSF) += bsf/trace_headers.o
OBJS-$(CONFIG_TRUEHD_CORE_BSF) += bsf/truehd_core.o
OBJS-$(CONFIG_VP9_METADATA_BSF) += bsf/vp9_metadata.o
OBJS-$(CONFIG_VP9_RAW_REORDER_BSF) += bsf/vp9_raw_reorder.o
OBJS-$(CONFIG_VP9_SUPERFRAME_BSF) += bsf/vp9_superframe.o
OBJS-$(CONFIG_VP9_SUPERFRAME_SPLIT_BSF) += bsf/vp9_superframe_split.o
OBJS-$(CONFIG_VVC_METADATA_BSF) += bsf/h266_metadata.o
OBJS-$(CONFIG_VVC_MP4TOANNEXB_BSF) += bsf/vvc_mp4toannexb.o
libavcodec/bsf/%.o: CPPFLAGS += -I$(SRC_PATH)/libavcodec/
+156
View File
@@ -0,0 +1,156 @@
/*
* MPEG-2/4 AAC ADTS to MPEG-4 Audio Specific Configuration bitstream filter
* Copyright (c) 2009 Alex Converse <alex.converse@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 "adts_header.h"
#include "adts_parser.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "put_bits.h"
#include "get_bits.h"
#include "mpeg4audio.h"
#include "mpeg4audio_copy_pce.h"
typedef struct AACBSFContext {
int first_frame_done;
} AACBSFContext;
/**
* This filter creates an MPEG-4 AudioSpecificConfig from an MPEG-2/4
* ADTS header and removes the ADTS header.
*/
static int aac_adtstoasc_filter(AVBSFContext *bsfc, AVPacket *pkt)
{
AACBSFContext *ctx = bsfc->priv_data;
PutBitContext pb;
AACADTSHeaderInfo hdr;
int ret;
ret = ff_bsf_get_packet_ref(bsfc, pkt);
if (ret < 0)
return ret;
if (bsfc->par_in->extradata && pkt->size >= 2 && (AV_RB16(pkt->data) >> 4) != 0xfff)
return 0;
if (pkt->size < AV_AAC_ADTS_HEADER_SIZE)
goto packet_too_small;
if (ff_adts_header_parse_buf(pkt->data, &hdr) < 0) {
av_log(bsfc, AV_LOG_ERROR, "Error parsing ADTS frame header!\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (!hdr.crc_absent && hdr.num_aac_frames > 1) {
avpriv_report_missing_feature(bsfc,
"Multiple RDBs per frame with CRC");
ret = AVERROR_PATCHWELCOME;
goto fail;
}
pkt->size -= AV_AAC_ADTS_HEADER_SIZE + 2 * !hdr.crc_absent;
if (pkt->size <= 0)
goto packet_too_small;
pkt->data += AV_AAC_ADTS_HEADER_SIZE + 2 * !hdr.crc_absent;
if (!ctx->first_frame_done) {
int pce_size = 0;
uint8_t pce_data[MAX_PCE_SIZE];
uint8_t *extradata;
if (!hdr.chan_config) {
GetBitContext gb;
init_get_bits(&gb, pkt->data, pkt->size * 8);
if (get_bits(&gb, 3) != 5) {
avpriv_report_missing_feature(bsfc,
"PCE-based channel configuration "
"without PCE as first syntax "
"element");
ret = AVERROR_PATCHWELCOME;
goto fail;
}
init_put_bits(&pb, pce_data, MAX_PCE_SIZE);
pce_size = ff_copy_pce_data(&pb, &gb) / 8;
flush_put_bits(&pb);
pkt->size -= get_bits_count(&gb)/8;
pkt->data += get_bits_count(&gb)/8;
}
extradata = av_packet_new_side_data(pkt, AV_PKT_DATA_NEW_EXTRADATA,
2 + pce_size);
if (!extradata) {
ret = AVERROR(ENOMEM);
goto fail;
}
init_put_bits(&pb, extradata, 2 + pce_size);
put_bits(&pb, 5, hdr.object_type);
put_bits(&pb, 4, hdr.sampling_index);
put_bits(&pb, 4, hdr.chan_config);
put_bits(&pb, 1, 0); //frame length - 1024 samples
put_bits(&pb, 1, 0); //does not depend on core coder
put_bits(&pb, 1, 0); //is not extension
flush_put_bits(&pb);
if (pce_size) {
memcpy(extradata + 2, pce_data, pce_size);
}
ctx->first_frame_done = 1;
}
return 0;
packet_too_small:
av_log(bsfc, AV_LOG_ERROR, "Input packet too small\n");
ret = AVERROR_INVALIDDATA;
fail:
av_packet_unref(pkt);
return ret;
}
static int aac_adtstoasc_init(AVBSFContext *ctx)
{
/* Validate the extradata if the stream is already MPEG-4 AudioSpecificConfig */
if (ctx->par_in->extradata) {
MPEG4AudioConfig mp4ac;
int ret = avpriv_mpeg4audio_get_config2(&mp4ac, ctx->par_in->extradata,
ctx->par_in->extradata_size, 1, ctx);
if (ret < 0) {
av_log(ctx, AV_LOG_ERROR, "Error parsing AudioSpecificConfig extradata!\n");
return ret;
}
}
return 0;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_AAC, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_aac_adtstoasc_bsf = {
.p.name = "aac_adtstoasc",
.p.codec_ids = codec_ids,
.priv_data_size = sizeof(AACBSFContext),
.init = aac_adtstoasc_init,
.filter = aac_adtstoasc_filter,
};
+167
View File
@@ -0,0 +1,167 @@
/*
* Copyright (c) 2019 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 "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_av1.h"
typedef struct AV1FMergeContext {
CodedBitstreamContext *input;
CodedBitstreamContext *output;
CodedBitstreamFragment frag[2];
AVPacket *pkt, *in;
int idx;
} AV1FMergeContext;
static void av1_frame_merge_flush(AVBSFContext *bsf)
{
AV1FMergeContext *ctx = bsf->priv_data;
ff_cbs_fragment_reset(&ctx->frag[0]);
ff_cbs_fragment_reset(&ctx->frag[1]);
av_packet_unref(ctx->in);
av_packet_unref(ctx->pkt);
}
static int av1_frame_merge_filter(AVBSFContext *bsf, AVPacket *out)
{
AV1FMergeContext *ctx = bsf->priv_data;
CodedBitstreamFragment *frag = &ctx->frag[ctx->idx], *tu = &ctx->frag[!ctx->idx];
AVPacket *in = ctx->in, *buffer_pkt = ctx->pkt;
int err, i;
err = ff_bsf_get_packet_ref(bsf, in);
if (err < 0) {
if (err == AVERROR_EOF && tu->nb_units > 0)
goto eof;
return err;
}
err = ff_cbs_read_packet(ctx->input, frag, in);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to read packet.\n");
goto fail;
}
if (frag->nb_units == 0) {
av_log(bsf, AV_LOG_ERROR, "No OBU in packet.\n");
err = AVERROR_INVALIDDATA;
goto fail;
}
if (tu->nb_units == 0 && frag->units[0].type != AV1_OBU_TEMPORAL_DELIMITER) {
av_log(bsf, AV_LOG_ERROR, "Missing Temporal Delimiter.\n");
err = AVERROR_INVALIDDATA;
goto fail;
}
for (i = 1; i < frag->nb_units; i++) {
if (frag->units[i].type == AV1_OBU_TEMPORAL_DELIMITER) {
av_log(bsf, AV_LOG_ERROR, "Temporal Delimiter in the middle of a packet.\n");
err = AVERROR_INVALIDDATA;
goto fail;
}
}
if (tu->nb_units > 0 && frag->units[0].type == AV1_OBU_TEMPORAL_DELIMITER) {
eof:
err = ff_cbs_write_packet(ctx->output, buffer_pkt, tu);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to write packet.\n");
goto fail;
}
av_packet_move_ref(out, buffer_pkt);
// Swap fragment index, to avoid copying fragment references.
ctx->idx = !ctx->idx;
} else {
for (i = 0; i < frag->nb_units; i++) {
err = ff_cbs_insert_unit_content(tu, -1, frag->units[i].type,
frag->units[i].content, frag->units[i].content_ref);
if (err < 0)
goto fail;
}
err = AVERROR(EAGAIN);
}
/* Buffer packets with timestamps (there should be at most one per TU)
* or any packet if buffer_pkt is empty. The latter is needed to
* passthrough positions in case there are no timestamps like with
* the raw OBU demuxer. */
if (!buffer_pkt->data ||
in->pts != AV_NOPTS_VALUE && buffer_pkt->pts == AV_NOPTS_VALUE) {
av_packet_unref(buffer_pkt);
av_packet_move_ref(buffer_pkt, in);
} else
av_packet_unref(in);
ff_cbs_fragment_reset(&ctx->frag[ctx->idx]);
fail:
if (err < 0 && err != AVERROR(EAGAIN))
av1_frame_merge_flush(bsf);
return err;
}
static int av1_frame_merge_init(AVBSFContext *bsf)
{
AV1FMergeContext *ctx = bsf->priv_data;
int err;
ctx->in = av_packet_alloc();
ctx->pkt = av_packet_alloc();
if (!ctx->in || !ctx->pkt)
return AVERROR(ENOMEM);
err = ff_cbs_init(&ctx->input, AV_CODEC_ID_AV1, bsf);
if (err < 0)
return err;
return ff_cbs_init(&ctx->output, AV_CODEC_ID_AV1, bsf);
}
static void av1_frame_merge_close(AVBSFContext *bsf)
{
AV1FMergeContext *ctx = bsf->priv_data;
ff_cbs_fragment_free(&ctx->frag[0]);
ff_cbs_fragment_free(&ctx->frag[1]);
av_packet_free(&ctx->in);
av_packet_free(&ctx->pkt);
ff_cbs_close(&ctx->input);
ff_cbs_close(&ctx->output);
}
static const enum AVCodecID av1_frame_merge_codec_ids[] = {
AV_CODEC_ID_AV1, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_av1_frame_merge_bsf = {
.p.name = "av1_frame_merge",
.p.codec_ids = av1_frame_merge_codec_ids,
.priv_data_size = sizeof(AV1FMergeContext),
.init = av1_frame_merge_init,
.flush = av1_frame_merge_flush,
.close = av1_frame_merge_close,
.filter = av1_frame_merge_filter,
};
+261
View File
@@ -0,0 +1,261 @@
/*
* Copyright (c) 2019 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
*/
/**
* @file
* This bitstream filter splits AV1 Temporal Units into packets containing
* just one frame, plus any leading and trailing OBUs that may be present at
* the beginning or end, respectively.
*
* Temporal Units already containing only one frame will be passed through
* unchanged. When splitting can't be performed, the Temporal Unit will be
* passed through containing only the remaining OBUs starting from the first
* one after the last successfully split frame.
*/
#include "libavutil/avassert.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_av1.h"
typedef struct AV1FSplitContext {
AVPacket *buffer_pkt;
CodedBitstreamContext *cbc;
CodedBitstreamFragment temporal_unit;
int nb_frames;
int cur_frame;
int cur_frame_idx;
int last_frame_idx;
} AV1FSplitContext;
static int av1_frame_split_filter(AVBSFContext *ctx, AVPacket *out)
{
AV1FSplitContext *s = ctx->priv_data;
CodedBitstreamFragment *td = &s->temporal_unit;
int i, ret;
int split = !!s->buffer_pkt->data;
if (!s->buffer_pkt->data) {
int nb_frames = 0;
ret = ff_bsf_get_packet_ref(ctx, s->buffer_pkt);
if (ret < 0)
return ret;
ret = ff_cbs_read_packet(s->cbc, td, s->buffer_pkt);
if (ret < 0) {
av_log(ctx, AV_LOG_WARNING, "Failed to parse temporal unit.\n");
goto passthrough;
}
for (i = 0; i < td->nb_units; i++) {
CodedBitstreamUnit *unit = &td->units[i];
if (unit->type == AV1_OBU_FRAME ||
unit->type == AV1_OBU_FRAME_HEADER)
nb_frames++;
else if (unit->type == AV1_OBU_TILE_LIST) {
av_log(ctx, AV_LOG_VERBOSE, "Large scale tiles are unsupported.\n");
goto passthrough;
}
}
if (nb_frames > 1) {
s->cur_frame = 0;
s->cur_frame_idx = s->last_frame_idx = 0;
s->nb_frames = nb_frames;
split = 1;
}
}
if (split) {
AV1RawFrameHeader *frame = NULL;
int cur_frame_type = -1, size = 0;
for (i = s->cur_frame_idx; i < td->nb_units; i++) {
CodedBitstreamUnit *unit = &td->units[i];
size += unit->data_size;
if (unit->type == AV1_OBU_FRAME) {
AV1RawOBU *obu = unit->content;
if (frame) {
av_log(ctx, AV_LOG_WARNING, "Frame OBU found when Tile data for a "
"previous frame was expected.\n");
goto passthrough;
}
frame = &obu->obu.frame.header;
cur_frame_type = obu->header.obu_type;
s->last_frame_idx = s->cur_frame_idx;
s->cur_frame_idx = i + 1;
s->cur_frame++;
// split here unless it's the last frame, in which case
// include every trailing OBU
if (s->cur_frame < s->nb_frames)
break;
} else if (unit->type == AV1_OBU_FRAME_HEADER) {
AV1RawOBU *obu = unit->content;
if (frame) {
av_log(ctx, AV_LOG_WARNING, "Frame Header OBU found when Tile data for a "
"previous frame was expected.\n");
goto passthrough;
}
frame = &obu->obu.frame_header;
cur_frame_type = obu->header.obu_type;
s->last_frame_idx = s->cur_frame_idx;
s->cur_frame++;
// split here if show_existing_frame unless it's the last
// frame, in which case include every trailing OBU
if (frame->show_existing_frame &&
s->cur_frame < s->nb_frames) {
s->cur_frame_idx = i + 1;
break;
}
} else if (unit->type == AV1_OBU_TILE_GROUP) {
AV1RawOBU *obu = unit->content;
AV1RawTileGroup *group = &obu->obu.tile_group;
if (!frame || cur_frame_type != AV1_OBU_FRAME_HEADER) {
av_log(ctx, AV_LOG_WARNING, "Unexpected Tile Group OBU found before a "
"Frame Header.\n");
goto passthrough;
}
if ((group->tg_end == (frame->tile_cols * frame->tile_rows) - 1) &&
// include every trailing OBU with the last frame
s->cur_frame < s->nb_frames) {
s->cur_frame_idx = i + 1;
break;
}
}
}
av_assert0(frame && s->cur_frame <= s->nb_frames);
ret = av_packet_ref(out, s->buffer_pkt);
if (ret < 0)
goto fail;
out->data = (uint8_t *)td->units[s->last_frame_idx].data;
out->size = size;
// skip the frame in the buffer packet if it's split successfully, so it's not present
// if the packet is passed through in case of failure when splitting another frame.
s->buffer_pkt->data += size;
s->buffer_pkt->size -= size;
if (!frame->show_existing_frame && !frame->show_frame)
out->pts = AV_NOPTS_VALUE;
if (s->cur_frame == s->nb_frames) {
av_packet_unref(s->buffer_pkt);
ff_cbs_fragment_reset(td);
}
return 0;
}
passthrough:
av_packet_move_ref(out, s->buffer_pkt);
ret = 0;
fail:
if (ret < 0) {
av_packet_unref(out);
av_packet_unref(s->buffer_pkt);
}
ff_cbs_fragment_reset(td);
return ret;
}
static const CodedBitstreamUnitType decompose_unit_types[] = {
AV1_OBU_TEMPORAL_DELIMITER,
AV1_OBU_SEQUENCE_HEADER,
AV1_OBU_FRAME_HEADER,
AV1_OBU_TILE_GROUP,
AV1_OBU_FRAME,
};
static int av1_frame_split_init(AVBSFContext *ctx)
{
AV1FSplitContext *s = ctx->priv_data;
CodedBitstreamFragment *td = &s->temporal_unit;
int ret;
s->buffer_pkt = av_packet_alloc();
if (!s->buffer_pkt)
return AVERROR(ENOMEM);
ret = ff_cbs_init(&s->cbc, AV_CODEC_ID_AV1, ctx);
if (ret < 0)
return ret;
s->cbc->decompose_unit_types = decompose_unit_types;
s->cbc->nb_decompose_unit_types = FF_ARRAY_ELEMS(decompose_unit_types);
if (!ctx->par_in->extradata_size)
return 0;
ret = ff_cbs_read_extradata(s->cbc, td, ctx->par_in);
if (ret < 0)
av_log(ctx, AV_LOG_WARNING, "Failed to parse extradata.\n");
ff_cbs_fragment_reset(td);
return 0;
}
static void av1_frame_split_flush(AVBSFContext *ctx)
{
AV1FSplitContext *s = ctx->priv_data;
av_packet_unref(s->buffer_pkt);
ff_cbs_fragment_reset(&s->temporal_unit);
}
static void av1_frame_split_close(AVBSFContext *ctx)
{
AV1FSplitContext *s = ctx->priv_data;
av_packet_free(&s->buffer_pkt);
ff_cbs_fragment_free(&s->temporal_unit);
ff_cbs_close(&s->cbc);
}
static const enum AVCodecID av1_frame_split_codec_ids[] = {
AV_CODEC_ID_AV1, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_av1_frame_split_bsf = {
.p.name = "av1_frame_split",
.p.codec_ids = av1_frame_split_codec_ids,
.priv_data_size = sizeof(AV1FSplitContext),
.init = av1_frame_split_init,
.flush = av1_frame_split_flush,
.close = av1_frame_split_close,
.filter = av1_frame_split_filter,
};
+231
View File
@@ -0,0 +1,231 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/common.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_av1.h"
typedef struct AV1MetadataContext {
CBSBSFContext common;
int td;
AV1RawOBU td_obu;
int color_primaries;
int transfer_characteristics;
int matrix_coefficients;
int color_range;
int chroma_sample_position;
AVRational tick_rate;
int num_ticks_per_picture;
int delete_padding;
} AV1MetadataContext;
static int av1_metadata_update_sequence_header(AVBSFContext *bsf,
AV1RawSequenceHeader *seq)
{
AV1MetadataContext *ctx = bsf->priv_data;
AV1RawColorConfig *clc = &seq->color_config;
AV1RawTimingInfo *tim = &seq->timing_info;
if (ctx->color_primaries >= 0 ||
ctx->transfer_characteristics >= 0 ||
ctx->matrix_coefficients >= 0) {
clc->color_description_present_flag = 1;
if (ctx->color_primaries >= 0)
clc->color_primaries = ctx->color_primaries;
if (ctx->transfer_characteristics >= 0)
clc->transfer_characteristics = ctx->transfer_characteristics;
if (ctx->matrix_coefficients >= 0)
clc->matrix_coefficients = ctx->matrix_coefficients;
}
if (ctx->color_range >= 0) {
if (clc->color_primaries == AVCOL_PRI_BT709 &&
clc->transfer_characteristics == AVCOL_TRC_IEC61966_2_1 &&
clc->matrix_coefficients == AVCOL_SPC_RGB) {
av_log(bsf, AV_LOG_WARNING, "Warning: color_range cannot be set "
"on RGB streams encoded in BT.709 sRGB.\n");
} else {
clc->color_range = ctx->color_range;
}
}
if (ctx->chroma_sample_position >= 0) {
if (clc->mono_chrome || !clc->subsampling_x || !clc->subsampling_y) {
av_log(bsf, AV_LOG_WARNING, "Warning: chroma_sample_position "
"can only be set for 4:2:0 streams.\n");
} else {
clc->chroma_sample_position = ctx->chroma_sample_position;
}
}
if (ctx->tick_rate.num && ctx->tick_rate.den) {
int num, den;
av_reduce(&num, &den, ctx->tick_rate.num, ctx->tick_rate.den,
UINT32_MAX > INT_MAX ? UINT32_MAX : INT_MAX);
tim->time_scale = num;
tim->num_units_in_display_tick = den;
seq->timing_info_present_flag = 1;
if (ctx->num_ticks_per_picture > 0) {
tim->equal_picture_interval = 1;
tim->num_ticks_per_picture_minus_1 =
ctx->num_ticks_per_picture - 1;
}
}
return 0;
}
static int av1_metadata_update_fragment(AVBSFContext *bsf, AVPacket *pkt,
CodedBitstreamFragment *frag)
{
AV1MetadataContext *ctx = bsf->priv_data;
int err, i;
for (i = 0; i < frag->nb_units; i++) {
if (frag->units[i].type == AV1_OBU_SEQUENCE_HEADER) {
AV1RawOBU *obu = frag->units[i].content;
err = av1_metadata_update_sequence_header(bsf, &obu->obu.sequence_header);
if (err < 0)
return err;
}
}
// If a Temporal Delimiter is present, it must be the first OBU.
if (frag->nb_units && frag->units[0].type == AV1_OBU_TEMPORAL_DELIMITER) {
if (ctx->td == BSF_ELEMENT_REMOVE)
ff_cbs_delete_unit(frag, 0);
} else if (pkt && ctx->td == BSF_ELEMENT_INSERT) {
err = ff_cbs_insert_unit_content(frag, 0, AV1_OBU_TEMPORAL_DELIMITER,
&ctx->td_obu, NULL);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to insert Temporal Delimiter.\n");
return err;
}
}
if (ctx->delete_padding) {
for (i = frag->nb_units - 1; i >= 0; i--) {
if (frag->units[i].type == AV1_OBU_PADDING)
ff_cbs_delete_unit(frag, i);
}
}
return 0;
}
static const CBSBSFType av1_metadata_type = {
.codec_id = AV_CODEC_ID_AV1,
.fragment_name = "temporal unit",
.unit_name = "OBU",
.update_fragment = &av1_metadata_update_fragment,
};
static int av1_metadata_init(AVBSFContext *bsf)
{
AV1MetadataContext *ctx = bsf->priv_data;
ctx->td_obu = (AV1RawOBU) {
.header.obu_type = AV1_OBU_TEMPORAL_DELIMITER,
};
return ff_cbs_bsf_generic_init(bsf, &av1_metadata_type);
}
#define OFFSET(x) offsetof(AV1MetadataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption av1_metadata_options[] = {
BSF_ELEMENT_OPTIONS_PIR("td", "Temporal Delimiter OBU",
td, FLAGS),
{ "color_primaries", "Set color primaries (section 6.4.2)",
OFFSET(color_primaries), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "transfer_characteristics", "Set transfer characteristics (section 6.4.2)",
OFFSET(transfer_characteristics), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "matrix_coefficients", "Set matrix coefficients (section 6.4.2)",
OFFSET(matrix_coefficients), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "color_range", "Set color range flag (section 6.4.2)",
OFFSET(color_range), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 1, FLAGS, .unit = "cr" },
{ "tv", "TV (limited) range", 0, AV_OPT_TYPE_CONST,
{ .i64 = 0 }, .flags = FLAGS, .unit = "cr" },
{ "pc", "PC (full) range", 0, AV_OPT_TYPE_CONST,
{ .i64 = 1 }, .flags = FLAGS, .unit = "cr" },
{ "chroma_sample_position", "Set chroma sample position (section 6.4.2)",
OFFSET(chroma_sample_position), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 3, FLAGS, .unit = "csp" },
{ "unknown", "Unknown chroma sample position", 0, AV_OPT_TYPE_CONST,
{ .i64 = AV1_CSP_UNKNOWN }, .flags = FLAGS, .unit = "csp" },
{ "vertical", "Left chroma sample position", 0, AV_OPT_TYPE_CONST,
{ .i64 = AV1_CSP_VERTICAL }, .flags = FLAGS, .unit = "csp" },
{ "colocated", "Top-left chroma sample position", 0, AV_OPT_TYPE_CONST,
{ .i64 = AV1_CSP_COLOCATED }, .flags = FLAGS, .unit = "csp" },
{ "tick_rate", "Set display tick rate (time_scale / num_units_in_display_tick)",
OFFSET(tick_rate), AV_OPT_TYPE_RATIONAL,
{ .dbl = 0.0 }, 0, UINT_MAX, FLAGS },
{ "num_ticks_per_picture", "Set display ticks per picture for CFR streams",
OFFSET(num_ticks_per_picture), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, INT_MAX, FLAGS },
{ "delete_padding", "Delete all Padding OBUs",
OFFSET(delete_padding), AV_OPT_TYPE_BOOL,
{ .i64 = 0 }, 0, 1, FLAGS},
{ NULL }
};
static const AVClass av1_metadata_class = {
.class_name = "av1_metadata_bsf",
.item_name = av_default_item_name,
.option = av1_metadata_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID av1_metadata_codec_ids[] = {
AV_CODEC_ID_AV1, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_av1_metadata_bsf = {
.p.name = "av1_metadata",
.p.codec_ids = av1_metadata_codec_ids,
.p.priv_class = &av1_metadata_class,
.priv_data_size = sizeof(AV1MetadataContext),
.init = &av1_metadata_init,
.close = &ff_cbs_bsf_generic_close,
.filter = &ff_cbs_bsf_generic_filter,
};
+45
View File
@@ -0,0 +1,45 @@
/*
* Chomp bitstream filter
* Copyright (c) 2010 Alex Converse <alex.converse@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 "bsf.h"
#include "bsf_internal.h"
static int chomp_filter(AVBSFContext *ctx, AVPacket *pkt)
{
int ret;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
while (pkt->size > 0 && !pkt->data[pkt->size - 1])
pkt->size--;
return 0;
}
/**
* This filter removes a string of NULL bytes from the end of a packet.
*/
const FFBitStreamFilter ff_chomp_bsf = {
.p.name = "chomp",
.filter = chomp_filter,
};
+61
View File
@@ -0,0 +1,61 @@
/*
* Copyright (c) 2016 Paul B Mahol
*
* 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 "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "dca_syncwords.h"
static int dca_core_filter(AVBSFContext *ctx, AVPacket *pkt)
{
GetByteContext gb;
uint32_t syncword;
int core_size = 0, ret;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
bytestream2_init(&gb, pkt->data, pkt->size);
syncword = bytestream2_get_be32(&gb);
bytestream2_skip(&gb, 1);
switch (syncword) {
case DCA_SYNCWORD_CORE_BE:
core_size = ((bytestream2_get_be24(&gb) >> 4) & 0x3fff) + 1;
break;
}
if (core_size > 0 && core_size <= pkt->size) {
pkt->size = core_size;
}
return 0;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_DTS, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_dca_core_bsf = {
.p.name = "dca_core",
.p.codec_ids = codec_ids,
.filter = dca_core_filter,
};
+285
View File
@@ -0,0 +1,285 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/common.h"
#include "libavutil/mem.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_av1.h"
#include "cbs_h265.h"
#include "dovi_rpu.h"
#include "h2645data.h"
#include "h265_profile_level.h"
#include "itut35.h"
#include "hevc/hevc.h"
typedef struct DoviRpuContext {
CBSBSFContext common;
DOVIContext dec;
DOVIContext enc;
int strip;
int compression;
} DoviRpuContext;
static int update_rpu(AVBSFContext *bsf, const AVPacket *pkt, int flags,
const uint8_t *rpu, size_t rpu_size,
uint8_t **out_rpu, int *out_size)
{
DoviRpuContext *s = bsf->priv_data;
AVDOVIMetadata *metadata = NULL;
int ret;
ret = ff_dovi_rpu_parse(&s->dec, rpu, rpu_size, 0);
if (ret < 0) {
ff_dovi_ctx_flush(&s->dec);
return ret;
}
ret = ff_dovi_get_metadata(&s->dec, &metadata);
if (ret == 0 /* no metadata */) {
*out_rpu = NULL;
*out_size = 0;
return 0;
} else if (ret < 0) {
ff_dovi_ctx_flush(&s->dec);
return ret;
}
if (pkt && !(pkt->flags & AV_PKT_FLAG_KEY))
flags |= FF_DOVI_COMPRESS_RPU;
ret = ff_dovi_rpu_generate(&s->enc, metadata, flags, out_rpu, out_size);
av_free(metadata);
if (ret < 0)
ff_dovi_ctx_flush(&s->enc);
return ret;
}
static int dovi_rpu_update_fragment_hevc(AVBSFContext *bsf, AVPacket *pkt,
CodedBitstreamFragment *au)
{
DoviRpuContext *s = bsf->priv_data;
CodedBitstreamUnit *nal = au->nb_units ? &au->units[au->nb_units - 1] : NULL;
uint8_t *rpu = NULL;
int rpu_size, ret;
if (!nal || nal->type != HEVC_NAL_UNSPEC62)
return 0;
if (s->strip) {
ff_cbs_delete_unit(au, au->nb_units - 1);
return 0;
}
ret = update_rpu(bsf, pkt, 0, nal->data + 2, nal->data_size - 2, &rpu, &rpu_size);
if (ret < 0)
return ret;
/* NAL unit header + NAL prefix */
if (rpu_size + 3 <= nal->data_size && av_buffer_is_writable(nal->data_ref)) {
memcpy(nal->data + 3, rpu, rpu_size);
av_free(rpu);
nal->data_size = rpu_size + 3;
} else {
AVBufferRef *ref = av_buffer_alloc(rpu_size + 3);
if (!ref) {
av_free(rpu);
return AVERROR(ENOMEM);
}
memcpy(ref->data, nal->data, 3);
memcpy(ref->data + 3, rpu, rpu_size);
av_buffer_unref(&nal->data_ref);
av_free(rpu);
nal->data = ref->data;
nal->data_size = rpu_size + 3;
nal->data_ref = ref;
nal->data_bit_padding = 0;
}
return 0;
}
static int dovi_rpu_update_fragment_av1(AVBSFContext *bsf, AVPacket *pkt,
CodedBitstreamFragment *frag)
{
DoviRpuContext *s = bsf->priv_data;
int provider_code, provider_oriented_code, rpu_size, ret;
AVBufferRef *ref;
uint8_t *rpu;
for (int i = 0; i < frag->nb_units; i++) {
AV1RawOBU *obu = frag->units[i].content;
AV1RawMetadataITUTT35 *t35 = &obu->obu.metadata.metadata.itut_t35;
if (frag->units[i].type != AV1_OBU_METADATA ||
obu->obu.metadata.metadata_type != AV1_METADATA_TYPE_ITUT_T35 ||
t35->itu_t_t35_country_code != ITU_T_T35_COUNTRY_CODE_US ||
t35->payload_size < 6)
continue;
provider_code = AV_RB16(t35->payload);
provider_oriented_code = AV_RB32(t35->payload + 2);
if (provider_code != ITU_T_T35_PROVIDER_CODE_DOLBY ||
provider_oriented_code != 0x800)
continue;
if (s->strip) {
ff_cbs_delete_unit(frag, i);
return 0;
}
ret = update_rpu(bsf, pkt, FF_DOVI_WRAP_T35,
t35->payload + 6, t35->payload_size - 6,
&rpu, &rpu_size);
if (ret < 0)
return ret;
ref = av_buffer_create(rpu, rpu_size, av_buffer_default_free, NULL, 0);
if (!ref) {
av_free(rpu);
return AVERROR(ENOMEM);
}
av_buffer_unref(&t35->payload_ref);
t35->payload_ref = ref;
t35->payload = rpu + 1; /* skip country code */
t35->payload_size = rpu_size - 1;
break; /* should be only one RPU per packet */
}
return 0;
}
static const CBSBSFType dovi_rpu_hevc_type = {
.codec_id = AV_CODEC_ID_HEVC,
.fragment_name = "access unit",
.unit_name = "NAL unit",
.update_fragment = &dovi_rpu_update_fragment_hevc,
};
static const CBSBSFType dovi_rpu_av1_type = {
.codec_id = AV_CODEC_ID_AV1,
.fragment_name = "temporal unit",
.unit_name = "OBU",
.update_fragment = &dovi_rpu_update_fragment_av1,
};
static int dovi_rpu_init(AVBSFContext *bsf)
{
int ret;
DoviRpuContext *s = bsf->priv_data;
s->dec.logctx = s->enc.logctx = bsf;
s->enc.enable = 1;
if (s->compression == AV_DOVI_COMPRESSION_RESERVED) {
av_log(bsf, AV_LOG_ERROR, "Invalid compression level: %d\n", s->compression);
return AVERROR(EINVAL);
}
if (s->strip) {
av_packet_side_data_remove(bsf->par_out->coded_side_data,
&bsf->par_out->nb_coded_side_data,
AV_PKT_DATA_DOVI_CONF);
} else {
const AVPacketSideData *sd;
sd = av_packet_side_data_get(bsf->par_out->coded_side_data,
bsf->par_out->nb_coded_side_data,
AV_PKT_DATA_DOVI_CONF);
if (sd) {
AVDOVIDecoderConfigurationRecord *cfg;
cfg = (AVDOVIDecoderConfigurationRecord *) sd->data;
s->dec.cfg = *cfg;
/* Update configuration record before setting to enc ctx */
cfg->dv_md_compression = s->compression;
if (s->compression && s->dec.cfg.dv_profile < 8) {
av_log(bsf, AV_LOG_ERROR, "Invalid compression level %d for "
"Dolby Vision profile %d.\n", s->compression, s->dec.cfg.dv_profile);
return AVERROR(EINVAL);
}
s->enc.cfg = *cfg;
} else {
av_log(bsf, AV_LOG_WARNING, "No Dolby Vision configuration record "
"found? Generating one, but results may be invalid.\n");
ret = ff_dovi_configure_ext(&s->enc, bsf->par_out, NULL, s->compression,
FF_COMPLIANCE_NORMAL);
if (ret < 0)
return ret;
/* Be conservative in accepting all compressed RPUs */
s->dec.cfg = s->enc.cfg;
s->dec.cfg.dv_md_compression = AV_DOVI_COMPRESSION_EXTENDED;
}
}
switch (bsf->par_in->codec_id) {
case AV_CODEC_ID_HEVC:
return ff_cbs_bsf_generic_init(bsf, &dovi_rpu_hevc_type);
case AV_CODEC_ID_AV1:
return ff_cbs_bsf_generic_init(bsf, &dovi_rpu_av1_type);
default:
return AVERROR_BUG;
}
}
static void dovi_rpu_close(AVBSFContext *bsf)
{
DoviRpuContext *s = bsf->priv_data;
ff_dovi_ctx_unref(&s->dec);
ff_dovi_ctx_unref(&s->enc);
ff_cbs_bsf_generic_close(bsf);
}
#define OFFSET(x) offsetof(DoviRpuContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption dovi_rpu_options[] = {
{ "strip", "Strip Dolby Vision metadata", OFFSET(strip), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, FLAGS },
{ "compression", "DV metadata compression mode", OFFSET(compression), AV_OPT_TYPE_INT, { .i64 = AV_DOVI_COMPRESSION_LIMITED }, 0, AV_DOVI_COMPRESSION_EXTENDED, FLAGS, .unit = "compression" },
{ "none", "Don't compress metadata", 0, AV_OPT_TYPE_CONST, {.i64 = 0}, .flags = FLAGS, .unit = "compression" },
{ "limited", "Limited metadata compression", 0, AV_OPT_TYPE_CONST, {.i64 = AV_DOVI_COMPRESSION_LIMITED}, .flags = FLAGS, .unit = "compression" },
{ "extended", "Extended metadata compression",0, AV_OPT_TYPE_CONST, {.i64 = AV_DOVI_COMPRESSION_EXTENDED}, .flags = FLAGS, .unit = "compression" },
{ NULL }
};
static const AVClass dovi_rpu_class = {
.class_name = "dovi_rpu_bsf",
.item_name = av_default_item_name,
.option = dovi_rpu_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID dovi_rpu_codec_ids[] = {
AV_CODEC_ID_HEVC, AV_CODEC_ID_AV1, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_dovi_rpu_bsf = {
.p.name = "dovi_rpu",
.p.codec_ids = dovi_rpu_codec_ids,
.p.priv_class = &dovi_rpu_class,
.priv_data_size = sizeof(DoviRpuContext),
.init = &dovi_rpu_init,
.close = &dovi_rpu_close,
.filter = &ff_cbs_bsf_generic_filter,
};
+550
View File
@@ -0,0 +1,550 @@
/*
* Copyright (c) 2022 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
*/
/**
* @file
* Derive PTS by reordering DTS from supported streams
*/
#include "libavutil/avassert.h"
#include "libavutil/fifo.h"
#include "libavutil/mem.h"
#include "libavutil/tree.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_h264.h"
#include "h264_parse.h"
#include "h264_ps.h"
#include "refstruct.h"
typedef struct DTS2PTSNode {
int64_t dts;
int64_t duration;
int poc;
int gop;
} DTS2PTSNode;
typedef struct DTS2PTSFrame {
AVPacket *pkt;
int poc;
int poc_diff;
int gop;
} DTS2PTSFrame;
typedef struct DTS2PTSH264Context {
H264POCContext poc;
SPS sps;
int poc_diff;
int last_poc;
int highest_poc;
int picture_structure;
} DTS2PTSH264Context;
typedef struct DTS2PTSContext {
struct AVTreeNode *root;
AVFifo *fifo;
FFRefStructPool *node_pool;
// Codec specific function pointers and constants
int (*init)(AVBSFContext *ctx);
int (*filter)(AVBSFContext *ctx);
void (*flush)(AVBSFContext *ctx);
size_t fifo_size;
CodedBitstreamContext *cbc;
CodedBitstreamFragment au;
union {
DTS2PTSH264Context h264;
} u;
int nb_frame;
int gop;
int eof;
} DTS2PTSContext;
// AVTreeNode callbacks
static int cmp_insert(const void *key, const void *node)
{
int ret = ((const DTS2PTSNode *)key)->poc - ((const DTS2PTSNode *)node)->poc;
if (!ret)
ret = ((const DTS2PTSNode *)key)->gop - ((const DTS2PTSNode *)node)->gop;
return ret;
}
static int cmp_find(const void *key, const void *node)
{
const DTS2PTSFrame * key1 = key;
const DTS2PTSNode *node1 = node;
int ret = FFDIFFSIGN(key1->poc, node1->poc);
if (!ret)
ret = key1->gop - node1->gop;
return ret;
}
static int dec_poc(void *opaque, void *elem)
{
DTS2PTSNode *node = elem;
int dec = *(int *)opaque;
node->poc -= dec;
return 0;
}
static int free_node(void *opaque, void *elem)
{
DTS2PTSNode *node = elem;
ff_refstruct_unref(&node);
return 0;
}
// Shared functions
static int alloc_and_insert_node(AVBSFContext *ctx, int64_t ts, int64_t duration,
int poc, int poc_diff, int gop)
{
DTS2PTSContext *s = ctx->priv_data;
for (int i = 0; i < poc_diff; i++) {
struct AVTreeNode *node = av_tree_node_alloc();
DTS2PTSNode *poc_node, *ret;
if (!node)
return AVERROR(ENOMEM);
poc_node = ff_refstruct_pool_get(s->node_pool);
if (!poc_node) {
av_free(node);
return AVERROR(ENOMEM);
}
if (i && ts != AV_NOPTS_VALUE)
ts += duration / poc_diff;
*poc_node = (DTS2PTSNode) { ts, duration, poc++, gop };
ret = av_tree_insert(&s->root, poc_node, cmp_insert, &node);
if (ret && ret != poc_node) {
*ret = *poc_node;
ff_refstruct_unref(&poc_node);
av_free(node);
}
}
return 0;
}
// H.264
static const CodedBitstreamUnitType h264_decompose_unit_types[] = {
H264_NAL_SPS,
H264_NAL_PPS,
H264_NAL_IDR_SLICE,
H264_NAL_SLICE,
};
static int h264_init(AVBSFContext *ctx)
{
DTS2PTSContext *s = ctx->priv_data;
DTS2PTSH264Context *h264 = &s->u.h264;
s->cbc->decompose_unit_types = h264_decompose_unit_types;
s->cbc->nb_decompose_unit_types = FF_ARRAY_ELEMS(h264_decompose_unit_types);
s->nb_frame = -(ctx->par_in->video_delay << 1);
h264->last_poc = h264->highest_poc = INT_MIN;
return 0;
}
static int get_mmco_reset(const H264RawSliceHeader *header)
{
if (header->nal_unit_header.nal_ref_idc == 0 ||
!header->adaptive_ref_pic_marking_mode_flag)
return 0;
for (int i = 0; i < H264_MAX_MMCO_COUNT; i++) {
if (header->mmco[i].memory_management_control_operation == 0)
return 0;
else if (header->mmco[i].memory_management_control_operation == 5)
return 1;
}
return 0;
}
static int h264_queue_frame(AVBSFContext *ctx, AVPacket *pkt, int poc, int *queued)
{
DTS2PTSContext *s = ctx->priv_data;
DTS2PTSH264Context *h264 = &s->u.h264;
DTS2PTSFrame frame;
int poc_diff, ret;
poc_diff = (h264->picture_structure == 3) + 1;
if (h264->sps.frame_mbs_only_flag && h264->poc_diff)
poc_diff = FFMIN(poc_diff, h264->poc_diff);
if (poc < 0) {
av_tree_enumerate(s->root, &poc_diff, NULL, dec_poc);
s->nb_frame -= poc_diff;
}
// Check if there was a POC reset (Like an IDR slice)
if (s->nb_frame > h264->highest_poc) {
s->nb_frame = 0;
s->gop = (s->gop + 1) % s->fifo_size;
h264->highest_poc = h264->last_poc;
}
ret = alloc_and_insert_node(ctx, pkt->dts, pkt->duration, s->nb_frame, poc_diff, s->gop);
if (ret < 0)
return ret;
av_log(ctx, AV_LOG_DEBUG, "Queueing frame with POC %d, GOP %d, dts %"PRId64"\n",
poc, s->gop, pkt->dts);
s->nb_frame += poc_diff;
// Add frame to output FIFO only once
if (*queued)
return 0;
frame = (DTS2PTSFrame) { pkt, poc, poc_diff, s->gop };
ret = av_fifo_write(s->fifo, &frame, 1);
av_assert2(ret >= 0);
*queued = 1;
return 0;
}
static int h264_filter(AVBSFContext *ctx)
{
DTS2PTSContext *s = ctx->priv_data;
DTS2PTSH264Context *h264 = &s->u.h264;
CodedBitstreamFragment *au = &s->au;
AVPacket *in;
int output_picture_number = INT_MIN;
int field_poc[2];
int queued = 0, ret;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
ret = ff_cbs_read_packet(s->cbc, au, in);
if (ret < 0) {
av_log(ctx, AV_LOG_WARNING, "Failed to parse access unit.\n");
goto fail;
}
for (int i = 0; i < au->nb_units; i++) {
CodedBitstreamUnit *unit = &au->units[i];
switch (unit->type) {
case H264_NAL_IDR_SLICE:
h264->poc.prev_frame_num = 0;
h264->poc.prev_frame_num_offset = 0;
h264->poc.prev_poc_msb =
h264->poc.prev_poc_lsb = 0;
// fall-through
case H264_NAL_SLICE: {
const H264RawSlice *slice = unit->content;
const H264RawSliceHeader *header = &slice->header;
const CodedBitstreamH264Context *cbs_h264 = s->cbc->priv_data;
const H264RawSPS *sps = cbs_h264->active_sps;
int got_reset;
if (!sps) {
av_log(ctx, AV_LOG_ERROR, "No active SPS for a slice\n");
goto fail;
}
// Initialize the SPS struct with the fields ff_h264_init_poc() cares about
h264->sps.frame_mbs_only_flag = sps->frame_mbs_only_flag;
h264->sps.log2_max_frame_num = sps->log2_max_frame_num_minus4 + 4;
h264->sps.poc_type = sps->pic_order_cnt_type;
h264->sps.log2_max_poc_lsb = sps->log2_max_pic_order_cnt_lsb_minus4 + 4;
h264->sps.offset_for_non_ref_pic = sps->offset_for_non_ref_pic;
h264->sps.offset_for_top_to_bottom_field = sps->offset_for_top_to_bottom_field;
h264->sps.poc_cycle_length = sps->num_ref_frames_in_pic_order_cnt_cycle;
for (int j = 0; j < h264->sps.poc_cycle_length; j++)
h264->sps.offset_for_ref_frame[j] = sps->offset_for_ref_frame[j];
h264->picture_structure = sps->frame_mbs_only_flag ? 3 :
(header->field_pic_flag ?
header->field_pic_flag + header->bottom_field_flag : 3);
h264->poc.frame_num = header->frame_num;
h264->poc.poc_lsb = header->pic_order_cnt_lsb;
h264->poc.delta_poc_bottom = header->delta_pic_order_cnt_bottom;
h264->poc.delta_poc[0] = header->delta_pic_order_cnt[0];
h264->poc.delta_poc[1] = header->delta_pic_order_cnt[1];
field_poc[0] = field_poc[1] = INT_MAX;
ret = ff_h264_init_poc(field_poc, &output_picture_number, &h264->sps,
&h264->poc, h264->picture_structure,
header->nal_unit_header.nal_ref_idc);
if (ret < 0) {
av_log(ctx, AV_LOG_ERROR, "ff_h264_init_poc() failure\n");
goto fail;
}
got_reset = get_mmco_reset(header);
h264->poc.prev_frame_num = got_reset ? 0 : h264->poc.frame_num;
h264->poc.prev_frame_num_offset = got_reset ? 0 : h264->poc.frame_num_offset;
if (header->nal_unit_header.nal_ref_idc != 0) {
h264->poc.prev_poc_msb = got_reset ? 0 : h264->poc.poc_msb;
if (got_reset)
h264->poc.prev_poc_lsb = h264->picture_structure == 2 ? 0 : field_poc[0];
else
h264->poc.prev_poc_lsb = h264->poc.poc_lsb;
}
if (output_picture_number != h264->last_poc) {
if (h264->last_poc != INT_MIN) {
int64_t diff = FFABS(h264->last_poc - (int64_t)output_picture_number);
if ((output_picture_number < 0) && !h264->last_poc)
h264->poc_diff = 0;
else if (FFABS((int64_t)output_picture_number) < h264->poc_diff) {
diff = FFABS(output_picture_number);
h264->poc_diff = 0;
}
if ((!h264->poc_diff || (h264->poc_diff > diff)) && diff <= INT_MAX) {
h264->poc_diff = diff;
if (h264->poc_diff == 1 && h264->sps.frame_mbs_only_flag) {
av_tree_enumerate(s->root, &h264->poc_diff, NULL, dec_poc);
s->nb_frame -= 2;
}
}
}
h264->last_poc = output_picture_number;
h264->highest_poc = FFMAX(h264->highest_poc, output_picture_number);
ret = h264_queue_frame(ctx, in, output_picture_number, &queued);
if (ret < 0)
goto fail;
}
break;
}
default:
break;
}
}
if (output_picture_number == INT_MIN) {
av_log(ctx, AV_LOG_ERROR, "No slices in access unit\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
ret = 0;
fail:
ff_cbs_fragment_reset(au);
if (!queued)
av_packet_free(&in);
return ret;
}
static void h264_flush(AVBSFContext *ctx)
{
DTS2PTSContext *s = ctx->priv_data;
DTS2PTSH264Context *h264 = &s->u.h264;
memset(&h264->sps, 0, sizeof(h264->sps));
memset(&h264->poc, 0, sizeof(h264->poc));
s->nb_frame = -(ctx->par_in->video_delay << 1);
h264->last_poc = h264->highest_poc = INT_MIN;
}
// Core functions
static const struct {
enum AVCodecID id;
int (*init)(AVBSFContext *ctx);
int (*filter)(AVBSFContext *ctx);
void (*flush)(AVBSFContext *ctx);
size_t fifo_size;
} func_tab[] = {
{ AV_CODEC_ID_H264, h264_init, h264_filter, h264_flush, H264_MAX_DPB_FRAMES * 2 * 2 },
};
static int dts2pts_init(AVBSFContext *ctx)
{
DTS2PTSContext *s = ctx->priv_data;
CodedBitstreamFragment *au = &s->au;
int i, ret;
for (i = 0; i < FF_ARRAY_ELEMS(func_tab); i++) {
if (func_tab[i].id == ctx->par_in->codec_id) {
s->init = func_tab[i].init;
s->filter = func_tab[i].filter;
s->flush = func_tab[i].flush;
s->fifo_size = func_tab[i].fifo_size;
break;
}
}
if (i == FF_ARRAY_ELEMS(func_tab))
return AVERROR_BUG;
av_assert0(s->filter && s->fifo_size);
s->fifo = av_fifo_alloc2(s->fifo_size, sizeof(DTS2PTSFrame), 0);
if (!s->fifo)
return AVERROR(ENOMEM);
s->node_pool = ff_refstruct_pool_alloc(sizeof(DTS2PTSNode),
FF_REFSTRUCT_POOL_FLAG_NO_ZEROING);
if (!s->node_pool)
return AVERROR(ENOMEM);
ret = ff_cbs_init(&s->cbc, ctx->par_in->codec_id, ctx);
if (ret < 0)
return ret;
if (s->init) {
ret = s->init(ctx);
if (ret < 0)
return ret;
}
if (!ctx->par_in->extradata_size)
return 0;
ret = ff_cbs_read_extradata(s->cbc, au, ctx->par_in);
if (ret < 0)
av_log(ctx, AV_LOG_WARNING, "Failed to parse extradata.\n");
ff_cbs_fragment_reset(au);
return 0;
}
static int dts2pts_filter(AVBSFContext *ctx, AVPacket *out)
{
DTS2PTSContext *s = ctx->priv_data;
DTS2PTSNode *poc_node = NULL, *next[2] = { NULL, NULL };
DTS2PTSFrame frame;
int ret;
// Fill up the FIFO and POC tree
while (!s->eof && av_fifo_can_write(s->fifo)) {
ret = s->filter(ctx);
if (ret < 0) {
if (ret != AVERROR_EOF)
return ret;
s->eof = 1;
}
}
if (!av_fifo_can_read(s->fifo))
return AVERROR_EOF;
// Fetch a packet from the FIFO
ret = av_fifo_read(s->fifo, &frame, 1);
av_assert2(ret >= 0);
av_packet_move_ref(out, frame.pkt);
av_packet_free(&frame.pkt);
// Search the timestamp for the requested POC and set PTS
poc_node = av_tree_find(s->root, &frame, cmp_find, (void **)next);
if (!poc_node) {
poc_node = next[1];
if (!poc_node || poc_node->poc != frame.poc)
poc_node = next[0];
}
if (poc_node && poc_node->poc == frame.poc) {
out->pts = poc_node->dts;
if (!s->eof) {
// Remove the found entry from the tree
DTS2PTSFrame dup = (DTS2PTSFrame) { NULL, frame.poc + 1, frame.poc_diff, frame.gop };
for (; dup.poc_diff > 0; dup.poc++, dup.poc_diff--) {
struct AVTreeNode *node = NULL;
if (!poc_node || poc_node->dts != out->pts)
continue;
av_tree_insert(&s->root, poc_node, cmp_insert, &node);
ff_refstruct_unref(&poc_node);
av_free(node);
poc_node = av_tree_find(s->root, &dup, cmp_find, NULL);
}
}
} else if (s->eof && frame.poc > INT_MIN) {
DTS2PTSFrame dup = (DTS2PTSFrame) { NULL, frame.poc - 1, frame.poc_diff, frame.gop };
poc_node = av_tree_find(s->root, &dup, cmp_find, NULL);
if (poc_node && poc_node->poc == dup.poc) {
out->pts = poc_node->dts;
if (out->pts != AV_NOPTS_VALUE)
out->pts += poc_node->duration;
ret = alloc_and_insert_node(ctx, out->pts, out->duration,
frame.poc, frame.poc_diff, frame.gop);
if (ret < 0) {
av_packet_unref(out);
return ret;
}
if (!ret)
av_log(ctx, AV_LOG_DEBUG, "Queueing frame for POC %d, GOP %d, dts %"PRId64", "
"generated from POC %d, GOP %d, dts %"PRId64", duration %"PRId64"\n",
frame.poc, frame.gop, out->pts,
poc_node->poc, poc_node->gop, poc_node->dts, poc_node->duration);
} else
av_log(ctx, AV_LOG_WARNING, "No timestamp for POC %d in tree\n", frame.poc);
} else
av_log(ctx, AV_LOG_WARNING, "No timestamp for POC %d in tree\n", frame.poc);
av_log(ctx, AV_LOG_DEBUG, "Returning frame for POC %d, GOP %d, dts %"PRId64", pts %"PRId64"\n",
frame.poc, frame.gop, out->dts, out->pts);
return 0;
}
static void dts2pts_flush(AVBSFContext *ctx)
{
DTS2PTSContext *s = ctx->priv_data;
DTS2PTSFrame frame;
if (s->flush)
s->flush(ctx);
s->eof = 0;
s->gop = 0;
while (s->fifo && av_fifo_read(s->fifo, &frame, 1) >= 0)
av_packet_free(&frame.pkt);
av_tree_enumerate(s->root, NULL, NULL, free_node);
av_tree_destroy(s->root);
s->root = NULL;
ff_cbs_fragment_reset(&s->au);
if (s->cbc)
ff_cbs_flush(s->cbc);
}
static void dts2pts_close(AVBSFContext *ctx)
{
DTS2PTSContext *s = ctx->priv_data;
dts2pts_flush(ctx);
av_fifo_freep2(&s->fifo);
ff_refstruct_pool_uninit(&s->node_pool);
ff_cbs_fragment_free(&s->au);
ff_cbs_close(&s->cbc);
}
static const enum AVCodecID dts2pts_codec_ids[] = {
AV_CODEC_ID_H264,
AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_dts2pts_bsf = {
.p.name = "dts2pts",
.p.codec_ids = dts2pts_codec_ids,
.priv_data_size = sizeof(DTS2PTSContext),
.init = dts2pts_init,
.flush = dts2pts_flush,
.close = dts2pts_close,
.filter = dts2pts_filter,
};
+106
View File
@@ -0,0 +1,106 @@
/*
* copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <string.h>
#include "bsf.h"
#include "bsf_internal.h"
#include "libavutil/log.h"
#include "libavutil/opt.h"
enum DumpFreq {
DUMP_FREQ_KEYFRAME,
DUMP_FREQ_ALL,
};
typedef struct DumpExtradataContext {
const AVClass *class;
AVPacket pkt;
int freq;
} DumpExtradataContext;
static int dump_extradata(AVBSFContext *ctx, AVPacket *out)
{
DumpExtradataContext *s = ctx->priv_data;
AVPacket *in = &s->pkt;
int ret = 0;
ret = ff_bsf_get_packet_ref(ctx, in);
if (ret < 0)
return ret;
if (ctx->par_in->extradata &&
(s->freq == DUMP_FREQ_ALL ||
(s->freq == DUMP_FREQ_KEYFRAME && in->flags & AV_PKT_FLAG_KEY)) &&
(in->size < ctx->par_in->extradata_size ||
memcmp(in->data, ctx->par_in->extradata, ctx->par_in->extradata_size))) {
if (in->size >= INT_MAX - ctx->par_in->extradata_size) {
ret = AVERROR(ERANGE);
goto fail;
}
ret = av_new_packet(out, in->size + ctx->par_in->extradata_size);
if (ret < 0)
goto fail;
ret = av_packet_copy_props(out, in);
if (ret < 0) {
av_packet_unref(out);
goto fail;
}
memcpy(out->data, ctx->par_in->extradata, ctx->par_in->extradata_size);
memcpy(out->data + ctx->par_in->extradata_size, in->data, in->size);
} else {
av_packet_move_ref(out, in);
}
fail:
av_packet_unref(in);
return ret;
}
#define OFFSET(x) offsetof(DumpExtradataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "freq", "When to dump extradata", OFFSET(freq), AV_OPT_TYPE_INT,
{ .i64 = DUMP_FREQ_KEYFRAME }, DUMP_FREQ_KEYFRAME, DUMP_FREQ_ALL, FLAGS, .unit = "freq" },
{ "k", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = DUMP_FREQ_KEYFRAME }, .flags = FLAGS, .unit = "freq" },
{ "keyframe", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = DUMP_FREQ_KEYFRAME }, .flags = FLAGS, .unit = "freq" },
{ "e", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = DUMP_FREQ_ALL }, .flags = FLAGS, .unit = "freq" },
{ "all", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = DUMP_FREQ_ALL }, .flags = FLAGS, .unit = "freq" },
{ NULL },
};
static const AVClass dump_extradata_class = {
.class_name = "dump_extradata bsf",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_dump_extradata_bsf = {
.p.name = "dump_extra",
.p.priv_class = &dump_extradata_class,
.priv_data_size = sizeof(DumpExtradataContext),
.filter = dump_extradata,
};
+138
View File
@@ -0,0 +1,138 @@
/*
* Copyright (c) 2022 Michael Niedermayer
*
* 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 "bsf.h"
#include "bsf_internal.h"
#include "libavutil/colorspace.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/opt.h"
typedef struct DVErrorMarkerContext {
const AVClass *class;
uint8_t color_rgba[4];
int sta;
uint8_t marked_block[76];
} DVErrorMarkerContext;
static void setdc(uint8_t *b, const uint8_t color_rgba[4], int cblocks, int y_step, int v_step, int u_step) {
for (int i=0; i<4; i++) {
b[0] = RGB_TO_Y_JPEG(color_rgba[0], color_rgba[1],color_rgba[2]) + 128;
b[1] = 0x06;
b += y_step;
}
for (int i=0; i<cblocks; i++) {
b[0] = RGB_TO_V_JPEG(color_rgba[0], color_rgba[1],color_rgba[2]) - 128;
b[1] = 0x16;
b += v_step;
}
for (int i=0; i<cblocks; i++) {
b[0] = RGB_TO_U_JPEG(color_rgba[0], color_rgba[1],color_rgba[2]) - 128;
b[1] = 0x16;
b += u_step;
}
}
static int dv_error_marker_init(AVBSFContext *ctx)
{
DVErrorMarkerContext *s = ctx->priv_data;
memset(s->marked_block, -1, 76);
setdc(s->marked_block, s->color_rgba, 1, 14, 10, 10);
setdc(s->marked_block, s->color_rgba, 2, 10, 10, 8);
return 0;
}
static int dv_error_marker_filter(AVBSFContext *ctx, AVPacket *pkt)
{
DVErrorMarkerContext *s = ctx->priv_data;
int ret = ff_bsf_get_packet_ref(ctx, pkt);
uint8_t *p;
int writable = 0;
int stamask = s->sta;
int match_count = 0;
if (ret < 0)
return ret;
p = pkt->data;
for(int i = 0; i < pkt->size - 79; i+=80) {
// see page 44-46 or section 5.5 of http://web.archive.org/web/20060927044735/http://www.smpte.org/smpte_store/standards/pdf/s314m.pdf.
if ((p[i] >> 4) == 9 && ((stamask >> (p[i+3] >> 4))&1)) {
if (!writable) {
ret = av_packet_make_writable(pkt);
if (ret < 0) {
av_packet_unref(pkt);
return ret;
}
writable = 1;
p = pkt->data;
}
memcpy(p+i+4, s->marked_block, 76);
match_count ++;
}
}
av_log(ctx, AV_LOG_DEBUG, "%8"PRId64": Replaced %5d blocks by color %X\n", pkt->pts, match_count, AV_RB32(s->color_rgba));
return 0;
}
#define OFFSET(x) offsetof(DVErrorMarkerContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "color" , "set color", OFFSET(color_rgba), AV_OPT_TYPE_COLOR, {.str = "yellow"}, 0, 0, FLAGS },
{ "sta" , "specify which error status value to match"
, OFFSET(sta ), AV_OPT_TYPE_FLAGS, {.i64 = 0xFFFE}, 0, 0xFFFF, FLAGS, .unit = "sta" },
{ "ok" , "No error, no concealment", 0, AV_OPT_TYPE_CONST, {.i64 = 0x0001}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "Aa" , "No error, concealment from previous frame type a",0, AV_OPT_TYPE_CONST, {.i64 = 0x0004}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "Ba" , "No error, concealment from next frame type a", 0, AV_OPT_TYPE_CONST, {.i64 = 0x0010}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "Ca" , "No error, unspecified concealment type a", 0, AV_OPT_TYPE_CONST, {.i64 = 0x0040}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "erri" , "Error with inserted code, No concealment", 0, AV_OPT_TYPE_CONST, {.i64 = 0x0080}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "erru" , "Error with unidentified pos, No concealment", 0, AV_OPT_TYPE_CONST, {.i64 = 0x8000}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "err" , "Error, No concealment", 0, AV_OPT_TYPE_CONST, {.i64 = 0x8080}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "Ab" , "No error, concealment from previous frame type b",0, AV_OPT_TYPE_CONST, {.i64 = 0x0400}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "Bb" , "No error, concealment from next frame type b", 0, AV_OPT_TYPE_CONST, {.i64 = 0x1000}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "Cb" , "No error, unspecified concealment type b", 0, AV_OPT_TYPE_CONST, {.i64 = 0x4000}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "A" , "No error, concealment from previous frame", 0, AV_OPT_TYPE_CONST, {.i64 = 0x0404}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "B" , "No error, concealment from next frame", 0, AV_OPT_TYPE_CONST, {.i64 = 0x1010}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "C" , "No error, unspecified concealment", 0, AV_OPT_TYPE_CONST, {.i64 = 0x4040}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "a" , "No error, concealment type a", 0, AV_OPT_TYPE_CONST, {.i64 = 0x0054}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "b" , "No error, concealment type b", 0, AV_OPT_TYPE_CONST, {.i64 = 0x5400}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "res" , "Reserved", 0, AV_OPT_TYPE_CONST, {.i64 = 0x2B2A}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "notok" , "Error or concealment", 0, AV_OPT_TYPE_CONST, {.i64 = 0xD4D4}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ "notres" , "Not reserved", 0, AV_OPT_TYPE_CONST, {.i64 = 0xD4D5}, 0, 0xFFFF, FLAGS, .unit = "sta"},
{ NULL },
};
static const AVClass dv_error_marker_class = {
.class_name = "dv_error_marker",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_dv_error_marker_bsf = {
.p.name = "dv_error_marker",
.p.codec_ids = (const enum AVCodecID []){ AV_CODEC_ID_DVVIDEO, AV_CODEC_ID_NONE },
.p.priv_class = &dv_error_marker_class,
.priv_data_size = sizeof(DVErrorMarkerContext),
.init = dv_error_marker_init,
.filter = dv_error_marker_filter,
};
+86
View File
@@ -0,0 +1,86 @@
/*
* Copyright (c) 2018 Paul B Mahol
*
* 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 "bsf.h"
#include "bsf_internal.h"
#include "get_bits.h"
#include "ac3_parser_internal.h"
static int eac3_core_filter(AVBSFContext *ctx, AVPacket *pkt)
{
AC3HeaderInfo hdr;
GetBitContext gbc;
int ret;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
ret = init_get_bits8(&gbc, pkt->data, pkt->size);
if (ret < 0)
goto fail;
ret = ff_ac3_parse_header(&gbc, &hdr);
if (ret < 0) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (hdr.frame_type == EAC3_FRAME_TYPE_INDEPENDENT ||
hdr.frame_type == EAC3_FRAME_TYPE_AC3_CONVERT) {
pkt->size = FFMIN(hdr.frame_size, pkt->size);
} else if (hdr.frame_type == EAC3_FRAME_TYPE_DEPENDENT && pkt->size > hdr.frame_size) {
AC3HeaderInfo hdr2;
ret = init_get_bits8(&gbc, pkt->data + hdr.frame_size, pkt->size - hdr.frame_size);
if (ret < 0)
goto fail;
ret = ff_ac3_parse_header(&gbc, &hdr2);
if (ret < 0) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (hdr2.frame_type == EAC3_FRAME_TYPE_INDEPENDENT ||
hdr2.frame_type == EAC3_FRAME_TYPE_AC3_CONVERT) {
pkt->size -= hdr.frame_size;
pkt->data += hdr.frame_size;
} else {
pkt->size = 0;
}
} else {
pkt->size = 0;
}
return 0;
fail:
av_packet_unref(pkt);
return ret;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_EAC3, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_eac3_core_bsf = {
.p.name = "eac3_core",
.p.codec_ids = codec_ids,
.filter = eac3_core_filter,
};
+290
View File
@@ -0,0 +1,290 @@
/*
* Copyright (c) 2019 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/mem.h"
#include "get_bits.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "evc.h"
#include "evc_parse.h"
#include "evc_ps.h"
// Access unit data
typedef struct AccessUnitBuffer {
uint8_t *data; // the data buffer
size_t data_size; // size of data in bytes
unsigned capacity; // buffer capacity
} AccessUnitBuffer;
typedef struct EVCFMergeContext {
AVPacket *in, *buffer_pkt;
EVCParamSets ps;
EVCParserPoc poc;
AccessUnitBuffer au_buffer;
} EVCFMergeContext;
static int end_of_access_unit_found(const EVCParamSets *ps, const EVCParserSliceHeader *sh,
const EVCParserPoc *poc, enum EVCNALUnitType nalu_type)
{
EVCParserPPS *pps = ps->pps[sh->slice_pic_parameter_set_id];
EVCParserSPS *sps = ps->sps[pps->pps_seq_parameter_set_id];
av_assert0(sps && pps);
if (sps->profile_idc == 0) { // BASELINE profile
if (nalu_type == EVC_NOIDR_NUT || nalu_type == EVC_IDR_NUT)
return 1;
} else { // MAIN profile
if (nalu_type == EVC_NOIDR_NUT) {
if (poc->PicOrderCntVal != poc->prevPicOrderCntVal)
return 1;
} else if (nalu_type == EVC_IDR_NUT)
return 1;
}
return 0;
}
static void evc_frame_merge_flush(AVBSFContext *bsf)
{
EVCFMergeContext *ctx = bsf->priv_data;
ff_evc_ps_free(&ctx->ps);
av_packet_unref(ctx->in);
av_packet_unref(ctx->buffer_pkt);
ctx->au_buffer.data_size = 0;
}
static int parse_nal_unit(AVBSFContext *bsf, const uint8_t *buf, int buf_size)
{
EVCFMergeContext *ctx = bsf->priv_data;
GetBitContext gb;
enum EVCNALUnitType nalu_type;
int tid, err;
err = init_get_bits8(&gb, buf, buf_size);
if (err < 0)
return err;
// @see ISO_IEC_23094-1_2020, 7.4.2.2 NAL unit header semantic (Table 4 - NAL unit type codes and NAL unit type classes)
// @see enum EVCNALUnitType in evc.h
if (get_bits1(&gb)) {// forbidden_zero_bit
av_log(bsf, AV_LOG_ERROR, "Invalid NAL unit header\n");
return AVERROR_INVALIDDATA;
}
nalu_type = get_bits(&gb, 6) - 1;
if (nalu_type < EVC_NOIDR_NUT || nalu_type > EVC_UNSPEC_NUT62) {
av_log(bsf, AV_LOG_ERROR, "Invalid NAL unit type: (%d)\n", nalu_type);
return AVERROR_INVALIDDATA;
}
tid = get_bits(&gb, 3);
skip_bits(&gb, 5); // nuh_reserved_zero_5bits
skip_bits1(&gb); // nuh_extension_flag
switch (nalu_type) {
case EVC_SPS_NUT:
err = ff_evc_parse_sps(&gb, &ctx->ps);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "SPS parsing error\n");
return err;
}
break;
case EVC_PPS_NUT:
err = ff_evc_parse_pps(&gb, &ctx->ps);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "PPS parsing error\n");
return err;
}
break;
case EVC_IDR_NUT: // Coded slice of a IDR or non-IDR picture
case EVC_NOIDR_NUT: {
EVCParserSliceHeader sh;
err = ff_evc_parse_slice_header(&gb, &sh, &ctx->ps, nalu_type);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Slice header parsing error\n");
return err;
}
// POC (picture order count of the current picture) derivation
// @see ISO/IEC 23094-1:2020(E) 8.3.1 Decoding process for picture order count
err = ff_evc_derive_poc(&ctx->ps, &sh, &ctx->poc, nalu_type, tid);
if (err < 0)
return err;
return end_of_access_unit_found(&ctx->ps, &sh, &ctx->poc, nalu_type);
break;
}
case EVC_SEI_NUT: // Supplemental Enhancement Information
case EVC_APS_NUT: // Adaptation parameter set
case EVC_FD_NUT: // Filler data
default:
break;
}
return 0;
}
static int evc_frame_merge_filter(AVBSFContext *bsf, AVPacket *out)
{
EVCFMergeContext *ctx = bsf->priv_data;
AVPacket *in = ctx->in, *buffer_pkt = ctx->buffer_pkt;
size_t data_size;
int au_end_found = 0, err;
while (!au_end_found) {
uint8_t *buffer;
uint32_t nalu_size;
if (!in->size) {
av_packet_unref(in);
err = ff_bsf_get_packet_ref(bsf, in);
if (err < 0) {
if (err == AVERROR_EOF && ctx->au_buffer.data_size > 0)
break;
return err;
}
/* Buffer packets with timestamps (there should be at most one per AU)
* or any packet if buffer_pkt is empty. The latter is needed to
* passthrough positions in case there are no timestamps like with
* the raw EVC demuxer. */
if (!buffer_pkt->data ||
in->pts != AV_NOPTS_VALUE && buffer_pkt->pts == AV_NOPTS_VALUE) {
err = av_packet_ref(buffer_pkt, in);
if (err < 0)
goto end;
}
}
// Buffer size is not enough for buffer to store NAL unit 4-bytes prefix (length)
if (in->size < EVC_NALU_LENGTH_PREFIX_SIZE)
return AVERROR_INVALIDDATA;
nalu_size = evc_read_nal_unit_length(in->data, EVC_NALU_LENGTH_PREFIX_SIZE, bsf);
if (!nalu_size || nalu_size > INT_MAX) {
av_log(bsf, AV_LOG_ERROR, "Invalid NAL unit size: (%u)\n", nalu_size);
err = AVERROR_INVALIDDATA;
goto end;
}
if (in->size < nalu_size + EVC_NALU_LENGTH_PREFIX_SIZE) {
err = AVERROR_INVALIDDATA;
goto end;
}
err = parse_nal_unit(bsf, in->data + EVC_NALU_LENGTH_PREFIX_SIZE, nalu_size);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Parsing of NAL unit failed\n");
goto end;
}
au_end_found = err;
nalu_size += EVC_NALU_LENGTH_PREFIX_SIZE;
data_size = ctx->au_buffer.data_size + nalu_size;
if (data_size > INT_MAX - AV_INPUT_BUFFER_PADDING_SIZE) {
av_log(bsf, AV_LOG_ERROR, "Assembled packet is too big\n");
err = AVERROR(ERANGE);
goto end;
}
buffer = av_fast_realloc(ctx->au_buffer.data, &ctx->au_buffer.capacity,
data_size);
if (!buffer) {
av_freep(&ctx->au_buffer.data);
err = AVERROR_INVALIDDATA;
goto end;
}
ctx->au_buffer.data = buffer;
memcpy(ctx->au_buffer.data + ctx->au_buffer.data_size, in->data, nalu_size);
ctx->au_buffer.data_size = data_size;
in->data += nalu_size;
in->size -= nalu_size;
}
av_packet_unref(in);
data_size = ctx->au_buffer.data_size;
ctx->au_buffer.data_size = 0;
// drop the data in buffer_pkt, if any, but keep the props
av_buffer_unref(&buffer_pkt->buf);
err = av_buffer_realloc(&buffer_pkt->buf, data_size + AV_INPUT_BUFFER_PADDING_SIZE);
if (err < 0)
goto end;
buffer_pkt->data = buffer_pkt->buf->data;
buffer_pkt->size = data_size;
av_packet_move_ref(out, buffer_pkt);
memcpy(out->data, ctx->au_buffer.data, data_size);
memset(out->data + data_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
err = 0;
end:
if (err < 0) {
av_packet_unref(in);
av_packet_unref(buffer_pkt);
ctx->au_buffer.data_size = 0;
}
return err;
}
static int evc_frame_merge_init(AVBSFContext *bsf)
{
EVCFMergeContext *ctx = bsf->priv_data;
ctx->in = av_packet_alloc();
ctx->buffer_pkt = av_packet_alloc();
if (!ctx->in || !ctx->buffer_pkt)
return AVERROR(ENOMEM);
return 0;
}
static void evc_frame_merge_close(AVBSFContext *bsf)
{
EVCFMergeContext *ctx = bsf->priv_data;
av_packet_free(&ctx->in);
av_packet_free(&ctx->buffer_pkt);
ff_evc_ps_free(&ctx->ps);
ctx->au_buffer.capacity = 0;
av_freep(&ctx->au_buffer.data);
ctx->au_buffer.data_size = 0;
}
static const enum AVCodecID evc_frame_merge_codec_ids[] = {
AV_CODEC_ID_EVC, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_evc_frame_merge_bsf = {
.p.name = "evc_frame_merge",
.p.codec_ids = evc_frame_merge_codec_ids,
.priv_data_size = sizeof(EVCFMergeContext),
.init = evc_frame_merge_init,
.flush = evc_frame_merge_flush,
.close = evc_frame_merge_close,
.filter = evc_frame_merge_filter,
};
+475
View File
@@ -0,0 +1,475 @@
/*
* 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/log.h"
#include "libavutil/mem.h"
#include "libavutil/opt.h"
#include "av1.h"
#include "av1_parse.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "h2645_parse.h"
#include "h264.h"
#include "startcode.h"
#include "vc1_common.h"
#include "vvc.h"
#include "hevc/hevc.h"
typedef struct ExtractExtradataContext {
const AVClass *class;
int (*extract)(AVBSFContext *ctx, AVPacket *pkt,
uint8_t **data, int *size);
/* AV1 specific fields */
AV1Packet av1_pkt;
/* H264/HEVC specific fields */
H2645Packet h2645_pkt;
/* AVOptions */
int remove;
} ExtractExtradataContext;
static int val_in_array(const int *arr, size_t len, int val)
{
for (size_t i = 0; i < len; i++)
if (arr[i] == val)
return 1;
return 0;
}
static int metadata_is_global(const AV1OBU *obu)
{
static const int metadata_obu_types[] = {
AV1_METADATA_TYPE_HDR_CLL, AV1_METADATA_TYPE_HDR_MDCV,
};
GetBitContext gb;
int metadata_type;
if (init_get_bits(&gb, obu->data, obu->size_bits) < 0)
return 0;
metadata_type = get_leb(&gb);
return val_in_array(metadata_obu_types, FF_ARRAY_ELEMS(metadata_obu_types),
metadata_type);
}
static int obu_is_global(const AV1OBU *obu)
{
static const int extradata_obu_types[] = {
AV1_OBU_SEQUENCE_HEADER, AV1_OBU_METADATA,
};
if (!val_in_array(extradata_obu_types, FF_ARRAY_ELEMS(extradata_obu_types),
obu->type))
return 0;
if (obu->type != AV1_OBU_METADATA)
return 1;
return metadata_is_global(obu);
}
static int extract_extradata_av1(AVBSFContext *ctx, AVPacket *pkt,
uint8_t **data, int *size)
{
ExtractExtradataContext *s = ctx->priv_data;
int extradata_size = 0, filtered_size = 0;
int i, has_seq = 0, ret = 0;
ret = ff_av1_packet_split(&s->av1_pkt, pkt->data, pkt->size, ctx);
if (ret < 0)
return ret;
for (i = 0; i < s->av1_pkt.nb_obus; i++) {
AV1OBU *obu = &s->av1_pkt.obus[i];
if (obu_is_global(obu)) {
extradata_size += obu->raw_size;
if (obu->type == AV1_OBU_SEQUENCE_HEADER)
has_seq = 1;
} else if (s->remove) {
filtered_size += obu->raw_size;
}
}
if (extradata_size && has_seq) {
AVBufferRef *filtered_buf = NULL;
PutByteContext pb_filtered_data, pb_extradata;
uint8_t *extradata;
if (s->remove) {
filtered_buf = av_buffer_alloc(filtered_size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!filtered_buf) {
return AVERROR(ENOMEM);
}
memset(filtered_buf->data + filtered_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
}
extradata = av_malloc(extradata_size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!extradata) {
av_buffer_unref(&filtered_buf);
return AVERROR(ENOMEM);
}
*data = extradata;
*size = extradata_size;
bytestream2_init_writer(&pb_extradata, extradata, extradata_size);
if (s->remove)
bytestream2_init_writer(&pb_filtered_data, filtered_buf->data, filtered_size);
for (i = 0; i < s->av1_pkt.nb_obus; i++) {
AV1OBU *obu = &s->av1_pkt.obus[i];
if (obu_is_global(obu)) {
bytestream2_put_bufferu(&pb_extradata, obu->raw_data, obu->raw_size);
} else if (s->remove) {
bytestream2_put_bufferu(&pb_filtered_data, obu->raw_data, obu->raw_size);
}
}
if (s->remove) {
av_buffer_unref(&pkt->buf);
pkt->buf = filtered_buf;
pkt->data = filtered_buf->data;
pkt->size = filtered_size;
}
}
return 0;
}
static int extract_extradata_h2645(AVBSFContext *ctx, AVPacket *pkt,
uint8_t **data, int *size)
{
static const int extradata_nal_types_vvc[] = {
VVC_VPS_NUT, VVC_SPS_NUT, VVC_PPS_NUT,
};
static const int extradata_nal_types_hevc[] = {
HEVC_NAL_VPS, HEVC_NAL_SPS, HEVC_NAL_PPS,
};
static const int extradata_nal_types_h264[] = {
H264_NAL_SPS, H264_NAL_PPS,
};
ExtractExtradataContext *s = ctx->priv_data;
int extradata_size = 0, filtered_size = 0;
const int *extradata_nal_types;
size_t nb_extradata_nal_types;
int i, has_sps = 0, has_vps = 0, ret = 0;
if (ctx->par_in->codec_id == AV_CODEC_ID_VVC) {
extradata_nal_types = extradata_nal_types_vvc;
nb_extradata_nal_types = FF_ARRAY_ELEMS(extradata_nal_types_vvc);
} else if (ctx->par_in->codec_id == AV_CODEC_ID_HEVC) {
extradata_nal_types = extradata_nal_types_hevc;
nb_extradata_nal_types = FF_ARRAY_ELEMS(extradata_nal_types_hevc);
} else {
extradata_nal_types = extradata_nal_types_h264;
nb_extradata_nal_types = FF_ARRAY_ELEMS(extradata_nal_types_h264);
}
ret = ff_h2645_packet_split(&s->h2645_pkt, pkt->data, pkt->size,
ctx, 0, ctx->par_in->codec_id, H2645_FLAG_SMALL_PADDING);
if (ret < 0)
return ret;
for (i = 0; i < s->h2645_pkt.nb_nals; i++) {
H2645NAL *nal = &s->h2645_pkt.nals[i];
if (val_in_array(extradata_nal_types, nb_extradata_nal_types, nal->type)) {
extradata_size += nal->raw_size + 3;
if (ctx->par_in->codec_id == AV_CODEC_ID_VVC) {
if (nal->type == VVC_SPS_NUT) has_sps = 1;
if (nal->type == VVC_VPS_NUT) has_vps = 1;
} else if (ctx->par_in->codec_id == AV_CODEC_ID_HEVC) {
if (nal->type == HEVC_NAL_SPS) has_sps = 1;
if (nal->type == HEVC_NAL_VPS) has_vps = 1;
} else {
if (nal->type == H264_NAL_SPS) has_sps = 1;
}
} else if (s->remove) {
filtered_size += nal->raw_size + 3;
}
}
if (extradata_size &&
((ctx->par_in->codec_id == AV_CODEC_ID_VVC && has_sps) ||
(ctx->par_in->codec_id == AV_CODEC_ID_HEVC && has_sps && has_vps) ||
(ctx->par_in->codec_id == AV_CODEC_ID_H264 && has_sps))) {
AVBufferRef *filtered_buf = NULL;
PutByteContext pb_filtered_data, pb_extradata;
uint8_t *extradata;
if (s->remove) {
filtered_buf = av_buffer_alloc(filtered_size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!filtered_buf) {
return AVERROR(ENOMEM);
}
memset(filtered_buf->data + filtered_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
}
extradata = av_malloc(extradata_size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!extradata) {
av_buffer_unref(&filtered_buf);
return AVERROR(ENOMEM);
}
*data = extradata;
*size = extradata_size;
bytestream2_init_writer(&pb_extradata, extradata, extradata_size);
if (s->remove)
bytestream2_init_writer(&pb_filtered_data, filtered_buf->data, filtered_size);
for (i = 0; i < s->h2645_pkt.nb_nals; i++) {
H2645NAL *nal = &s->h2645_pkt.nals[i];
if (val_in_array(extradata_nal_types, nb_extradata_nal_types,
nal->type)) {
bytestream2_put_be24u(&pb_extradata, 1); //startcode
bytestream2_put_bufferu(&pb_extradata, nal->raw_data, nal->raw_size);
} else if (s->remove) {
bytestream2_put_be24u(&pb_filtered_data, 1); // startcode
bytestream2_put_bufferu(&pb_filtered_data, nal->raw_data, nal->raw_size);
}
}
if (s->remove) {
av_buffer_unref(&pkt->buf);
pkt->buf = filtered_buf;
pkt->data = filtered_buf->data;
pkt->size = filtered_size;
}
}
return 0;
}
static int extract_extradata_vc1(AVBSFContext *ctx, AVPacket *pkt,
uint8_t **data, int *size)
{
ExtractExtradataContext *s = ctx->priv_data;
const uint8_t *ptr = pkt->data, *end = pkt->data + pkt->size;
uint32_t state = UINT32_MAX;
int has_extradata = 0, extradata_size = 0;
while (ptr < end) {
ptr = avpriv_find_start_code(ptr, end, &state);
if (state == VC1_CODE_SEQHDR || state == VC1_CODE_ENTRYPOINT) {
has_extradata = 1;
} else if (has_extradata && IS_MARKER(state)) {
extradata_size = ptr - 4 - pkt->data;
break;
}
}
if (extradata_size) {
*data = av_malloc(extradata_size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!*data)
return AVERROR(ENOMEM);
memcpy(*data, pkt->data, extradata_size);
*size = extradata_size;
if (s->remove) {
pkt->data += extradata_size;
pkt->size -= extradata_size;
}
}
return 0;
}
static int extract_extradata_mpeg12(AVBSFContext *ctx, AVPacket *pkt,
uint8_t **data, int *size)
{
ExtractExtradataContext *s = ctx->priv_data;
uint32_t state = UINT32_MAX;
int i, found = 0;
for (i = 0; i < pkt->size; i++) {
state = (state << 8) | pkt->data[i];
if (state == 0x1B3)
found = 1;
else if (found && state != 0x1B5 && state < 0x200 && state >= 0x100) {
*size = i - 3;
*data = av_malloc(*size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!*data)
return AVERROR(ENOMEM);
memcpy(*data, pkt->data, *size);
if (s->remove) {
pkt->data += *size;
pkt->size -= *size;
}
break;
}
}
return 0;
}
static int extract_extradata_mpeg4(AVBSFContext *ctx, AVPacket *pkt,
uint8_t **data, int *size)
{
ExtractExtradataContext *s = ctx->priv_data;
const uint8_t *ptr = pkt->data, *end = pkt->data + pkt->size;
uint32_t state = UINT32_MAX;
while (ptr < end) {
ptr = avpriv_find_start_code(ptr, end, &state);
if (state == 0x1B3 || state == 0x1B6) {
if (ptr - pkt->data > 4) {
*size = ptr - 4 - pkt->data;
*data = av_malloc(*size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!*data)
return AVERROR(ENOMEM);
memcpy(*data, pkt->data, *size);
if (s->remove) {
pkt->data += *size;
pkt->size -= *size;
}
}
break;
}
}
return 0;
}
static const struct {
enum AVCodecID id;
int (*extract)(AVBSFContext *ctx, AVPacket *pkt,
uint8_t **data, int *size);
} extract_tab[] = {
{ AV_CODEC_ID_AV1, extract_extradata_av1 },
{ AV_CODEC_ID_AVS2, extract_extradata_mpeg4 },
{ AV_CODEC_ID_AVS3, extract_extradata_mpeg4 },
{ AV_CODEC_ID_CAVS, extract_extradata_mpeg4 },
{ AV_CODEC_ID_H264, extract_extradata_h2645 },
{ AV_CODEC_ID_HEVC, extract_extradata_h2645 },
{ AV_CODEC_ID_MPEG1VIDEO, extract_extradata_mpeg12 },
{ AV_CODEC_ID_MPEG2VIDEO, extract_extradata_mpeg12 },
{ AV_CODEC_ID_MPEG4, extract_extradata_mpeg4 },
{ AV_CODEC_ID_VC1, extract_extradata_vc1 },
{ AV_CODEC_ID_VVC, extract_extradata_h2645 },
};
static int extract_extradata_init(AVBSFContext *ctx)
{
ExtractExtradataContext *s = ctx->priv_data;
int i;
for (i = 0; i < FF_ARRAY_ELEMS(extract_tab); i++) {
if (extract_tab[i].id == ctx->par_in->codec_id) {
s->extract = extract_tab[i].extract;
break;
}
}
if (!s->extract)
return AVERROR_BUG;
return 0;
}
static int extract_extradata_filter(AVBSFContext *ctx, AVPacket *pkt)
{
ExtractExtradataContext *s = ctx->priv_data;
uint8_t *extradata = NULL;
int extradata_size;
int ret = 0;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
ret = s->extract(ctx, pkt, &extradata, &extradata_size);
if (ret < 0)
goto fail;
if (extradata) {
memset(extradata + extradata_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
ret = av_packet_add_side_data(pkt, AV_PKT_DATA_NEW_EXTRADATA,
extradata, extradata_size);
if (ret < 0) {
av_freep(&extradata);
goto fail;
}
}
return 0;
fail:
av_packet_unref(pkt);
return ret;
}
static void extract_extradata_close(AVBSFContext *ctx)
{
ExtractExtradataContext *s = ctx->priv_data;
ff_av1_packet_uninit(&s->av1_pkt);
ff_h2645_packet_uninit(&s->h2645_pkt);
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_AV1,
AV_CODEC_ID_AVS2,
AV_CODEC_ID_AVS3,
AV_CODEC_ID_CAVS,
AV_CODEC_ID_H264,
AV_CODEC_ID_HEVC,
AV_CODEC_ID_MPEG1VIDEO,
AV_CODEC_ID_MPEG2VIDEO,
AV_CODEC_ID_MPEG4,
AV_CODEC_ID_VC1,
AV_CODEC_ID_VVC,
AV_CODEC_ID_NONE,
};
#define OFFSET(x) offsetof(ExtractExtradataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "remove", "remove the extradata from the bitstream", OFFSET(remove), AV_OPT_TYPE_INT,
{ .i64 = 0 }, 0, 1, FLAGS },
{ NULL },
};
static const AVClass extract_extradata_class = {
.class_name = "extract_extradata",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_extract_extradata_bsf = {
.p.name = "extract_extradata",
.p.codec_ids = codec_ids,
.p.priv_class = &extract_extradata_class,
.priv_data_size = sizeof(ExtractExtradataContext),
.init = extract_extradata_init,
.filter = extract_extradata_filter,
.close = extract_extradata_close,
};
+287
View File
@@ -0,0 +1,287 @@
/*
* 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 <stdbool.h>
#include <stdlib.h>
#include "libavutil/mem.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
typedef struct FilterUnitsContext {
const AVClass *class;
CodedBitstreamContext *cbc;
CodedBitstreamFragment fragment;
const char *pass_types;
const char *remove_types;
enum AVDiscard discard;
int discard_flags;
enum {
NOOP,
PASS,
REMOVE,
} mode;
CodedBitstreamUnitType *type_list;
int nb_types;
bool passthrough;
} FilterUnitsContext;
static int filter_units_make_type_list(const char *list_string,
CodedBitstreamUnitType **type_list,
int *nb_types)
{
CodedBitstreamUnitType *list = NULL;
int pass, count;
for (pass = 1; pass <= 2; pass++) {
long value, range_start, range_end;
const char *str;
char *value_end;
count = 0;
for (str = list_string; *str;) {
value = strtol(str, &value_end, 0);
if (str == value_end)
goto invalid;
str = (const char *)value_end;
if (*str == '-') {
++str;
range_start = value;
range_end = strtol(str, &value_end, 0);
if (str == value_end)
goto invalid;
for (value = range_start; value < range_end; value++) {
if (pass == 2)
list[count] = value;
++count;
}
} else {
if (pass == 2)
list[count] = value;
++count;
}
if (*str == '|')
++str;
}
if (pass == 1) {
list = av_malloc_array(count, sizeof(*list));
if (!list)
return AVERROR(ENOMEM);
}
}
*type_list = list;
*nb_types = count;
return 0;
invalid:
av_freep(&list);
return AVERROR(EINVAL);
}
static int filter_units_filter(AVBSFContext *bsf, AVPacket *pkt)
{
FilterUnitsContext *ctx = bsf->priv_data;
CodedBitstreamFragment *frag = &ctx->fragment;
int err, i, j;
err = ff_bsf_get_packet_ref(bsf, pkt);
if (err < 0)
return err;
if (ctx->passthrough)
return 0;
err = ff_cbs_read_packet(ctx->cbc, frag, pkt);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to read packet.\n");
goto fail;
}
ff_cbs_discard_units(ctx->cbc, frag, ctx->discard, ctx->discard_flags);
if (ctx->mode != NOOP) {
for (i = frag->nb_units - 1; i >= 0; i--) {
for (j = 0; j < ctx->nb_types; j++) {
if (frag->units[i].type == ctx->type_list[j])
break;
}
if (ctx->mode == REMOVE ? j < ctx->nb_types
: j >= ctx->nb_types)
ff_cbs_delete_unit(frag, i);
}
}
if (frag->nb_units == 0) {
// Don't return packets with nothing in them.
err = AVERROR(EAGAIN);
goto fail;
}
err = ff_cbs_write_packet(ctx->cbc, pkt, frag);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to write packet.\n");
goto fail;
}
fail:
if (err < 0)
av_packet_unref(pkt);
ff_cbs_fragment_reset(frag);
return err;
}
static int filter_units_init(AVBSFContext *bsf)
{
FilterUnitsContext *ctx = bsf->priv_data;
int err;
if (ctx->pass_types && ctx->remove_types) {
av_log(bsf, AV_LOG_ERROR, "Exactly one of pass_types or "
"remove_types is required.\n");
return AVERROR(EINVAL);
}
if (ctx->pass_types) {
ctx->mode = PASS;
err = filter_units_make_type_list(ctx->pass_types,
&ctx->type_list, &ctx->nb_types);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to parse pass_types.\n");
return err;
}
} else if (ctx->remove_types) {
ctx->mode = REMOVE;
err = filter_units_make_type_list(ctx->remove_types,
&ctx->type_list, &ctx->nb_types);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to parse remove_types.\n");
return err;
}
} else if (ctx->discard == AVDISCARD_NONE) {
ctx->passthrough = true;
return 0;
}
err = ff_cbs_init(&ctx->cbc, bsf->par_in->codec_id, bsf);
if (err < 0)
return err;
if (ctx->discard == AVDISCARD_NONE) {
// Don't actually decompose anything, we only want the unit data.
ctx->cbc->decompose_unit_types = ctx->type_list;
ctx->cbc->nb_decompose_unit_types = 0;
}
if (bsf->par_in->extradata) {
CodedBitstreamFragment *frag = &ctx->fragment;
err = ff_cbs_read_extradata(ctx->cbc, frag, bsf->par_in);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to read extradata.\n");
} else {
err = ff_cbs_write_extradata(ctx->cbc, bsf->par_out, frag);
if (err < 0)
av_log(bsf, AV_LOG_ERROR, "Failed to write extradata.\n");
}
ff_cbs_fragment_reset(frag);
}
return err;
}
static void filter_units_close(AVBSFContext *bsf)
{
FilterUnitsContext *ctx = bsf->priv_data;
av_freep(&ctx->type_list);
ff_cbs_fragment_free(&ctx->fragment);
ff_cbs_close(&ctx->cbc);
}
#define OFFSET(x) offsetof(FilterUnitsContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption filter_units_options[] = {
{ "pass_types", "List of unit types to pass through the filter.",
OFFSET(pass_types), AV_OPT_TYPE_STRING,
{ .str = NULL }, .flags = FLAGS },
{ "remove_types", "List of unit types to remove in the filter.",
OFFSET(remove_types), AV_OPT_TYPE_STRING,
{ .str = NULL }, .flags = FLAGS },
{ "discard", "Remove the selected frames",
OFFSET(discard), AV_OPT_TYPE_INT,
{ .i64 = AVDISCARD_NONE }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "none" , "discard none",
0, AV_OPT_TYPE_CONST,
{ .i64 = AVDISCARD_NONE }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "default" , "discard none, but can be changed after dynamically",
0, AV_OPT_TYPE_CONST,
{ .i64 = AVDISCARD_DEFAULT }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "nonref", "discard all non-reference frames",
0, AV_OPT_TYPE_CONST,
{ .i64 = AVDISCARD_NONREF }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "bidir", "discard all bidirectional frames",
0, AV_OPT_TYPE_CONST,
{ .i64 = AVDISCARD_BIDIR }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "nonintra", "discard all frames except I frames",
0, AV_OPT_TYPE_CONST,
{ .i64 = AVDISCARD_NONINTRA }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "nonkey", "discard all frames except keyframes",
0, AV_OPT_TYPE_CONST,
{ .i64 = AVDISCARD_NONKEY }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "all", "discard all frames",
0, AV_OPT_TYPE_CONST,
{ .i64 = AVDISCARD_ALL }, INT_MIN, INT_MAX, FLAGS, .unit = "discard"},
{ "discard_flags", "flags to control the discard frame behavior",
OFFSET(discard_flags), AV_OPT_TYPE_FLAGS,
{ .i64 = DISCARD_FLAG_NONE }, INT_MIN, INT_MAX, FLAGS, .unit = "discard_flags"},
{ "keep_non_vcl", "non-vcl units even if the picture has been dropped",
0, AV_OPT_TYPE_CONST,
{ .i64 = DISCARD_FLAG_KEEP_NON_VCL }, INT_MIN, INT_MAX, FLAGS, .unit = "discard_flags"},
{ NULL }
};
static const AVClass filter_units_class = {
.class_name = "filter_units",
.item_name = av_default_item_name,
.option = filter_units_options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_filter_units_bsf = {
.p.name = "filter_units",
.p.codec_ids = ff_cbs_all_codec_ids,
.p.priv_class = &filter_units_class,
.priv_data_size = sizeof(FilterUnitsContext),
.init = &filter_units_init,
.close = &filter_units_close,
.filter = &filter_units_filter,
};
+724
View File
@@ -0,0 +1,724 @@
/*
* 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/avstring.h"
#include "libavutil/display.h"
#include "libavutil/common.h"
#include "libavutil/mem.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_h264.h"
#include "cbs_sei.h"
#include "h264.h"
#include "h264_levels.h"
#include "h2645data.h"
#include "sei.h"
enum {
FLIP_HORIZONTAL = 1,
FLIP_VERTICAL = 2,
};
enum {
LEVEL_UNSET = -2,
LEVEL_AUTO = -1,
};
typedef struct H264MetadataContext {
CBSBSFContext common;
int done_first_au;
int aud;
H264RawAUD aud_nal;
AVRational sample_aspect_ratio;
int overscan_appropriate_flag;
int video_format;
int video_full_range_flag;
int colour_primaries;
int transfer_characteristics;
int matrix_coefficients;
int chroma_sample_loc_type;
AVRational tick_rate;
int fixed_frame_rate_flag;
int zero_new_constraint_set_flags;
int crop_left;
int crop_right;
int crop_top;
int crop_bottom;
const char *sei_user_data;
SEIRawUserDataUnregistered sei_user_data_payload;
int delete_filler;
int display_orientation;
double rotate;
int flip;
H264RawSEIDisplayOrientation display_orientation_payload;
int level;
} H264MetadataContext;
static int h264_metadata_insert_aud(AVBSFContext *bsf,
CodedBitstreamFragment *au)
{
H264MetadataContext *ctx = bsf->priv_data;
int primary_pic_type_mask = 0xff;
int err, i, j;
static const int primary_pic_type_table[] = {
0x084, // 2, 7
0x0a5, // 0, 2, 5, 7
0x0e7, // 0, 1, 2, 5, 6, 7
0x210, // 4, 9
0x318, // 3, 4, 8, 9
0x294, // 2, 4, 7, 9
0x3bd, // 0, 2, 3, 4, 5, 7, 8, 9
0x3ff, // 0, 1, 2, 3, 4, 5, 6, 7, 8, 9
};
for (i = 0; i < au->nb_units; i++) {
if (au->units[i].type == H264_NAL_SLICE ||
au->units[i].type == H264_NAL_IDR_SLICE) {
H264RawSlice *slice = au->units[i].content;
for (j = 0; j < FF_ARRAY_ELEMS(primary_pic_type_table); j++) {
if (!(primary_pic_type_table[j] &
(1 << slice->header.slice_type)))
primary_pic_type_mask &= ~(1 << j);
}
}
}
for (j = 0; j < FF_ARRAY_ELEMS(primary_pic_type_table); j++)
if (primary_pic_type_mask & (1 << j))
break;
if (j >= FF_ARRAY_ELEMS(primary_pic_type_table)) {
av_log(bsf, AV_LOG_ERROR, "No usable primary_pic_type: "
"invalid slice types?\n");
return AVERROR_INVALIDDATA;
}
ctx->aud_nal = (H264RawAUD) {
.nal_unit_header.nal_unit_type = H264_NAL_AUD,
.primary_pic_type = j,
};
err = ff_cbs_insert_unit_content(au, 0, H264_NAL_AUD,
&ctx->aud_nal, NULL);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to insert AUD.\n");
return err;
}
return 0;
}
static int h264_metadata_update_sps(AVBSFContext *bsf,
H264RawSPS *sps)
{
H264MetadataContext *ctx = bsf->priv_data;
int need_vui = 0;
int crop_unit_x, crop_unit_y;
if (ctx->sample_aspect_ratio.num && ctx->sample_aspect_ratio.den) {
int num, den, i;
av_reduce(&num, &den, ctx->sample_aspect_ratio.num,
ctx->sample_aspect_ratio.den, 65535);
for (i = 1; i < FF_ARRAY_ELEMS(ff_h2645_pixel_aspect); i++) {
if (num == ff_h2645_pixel_aspect[i].num &&
den == ff_h2645_pixel_aspect[i].den)
break;
}
if (i == FF_ARRAY_ELEMS(ff_h2645_pixel_aspect)) {
sps->vui.aspect_ratio_idc = 255;
sps->vui.sar_width = num;
sps->vui.sar_height = den;
} else {
sps->vui.aspect_ratio_idc = i;
}
sps->vui.aspect_ratio_info_present_flag = 1;
need_vui = 1;
}
#define SET_VUI_FIELD(field) do { \
if (ctx->field >= 0) { \
sps->vui.field = ctx->field; \
need_vui = 1; \
} \
} while (0)
if (ctx->overscan_appropriate_flag >= 0) {
SET_VUI_FIELD(overscan_appropriate_flag);
sps->vui.overscan_info_present_flag = 1;
}
if (ctx->video_format >= 0 ||
ctx->video_full_range_flag >= 0 ||
ctx->colour_primaries >= 0 ||
ctx->transfer_characteristics >= 0 ||
ctx->matrix_coefficients >= 0) {
SET_VUI_FIELD(video_format);
SET_VUI_FIELD(video_full_range_flag);
if (ctx->colour_primaries >= 0 ||
ctx->transfer_characteristics >= 0 ||
ctx->matrix_coefficients >= 0) {
SET_VUI_FIELD(colour_primaries);
SET_VUI_FIELD(transfer_characteristics);
SET_VUI_FIELD(matrix_coefficients);
sps->vui.colour_description_present_flag = 1;
}
sps->vui.video_signal_type_present_flag = 1;
}
if (ctx->chroma_sample_loc_type >= 0) {
sps->vui.chroma_sample_loc_type_top_field =
ctx->chroma_sample_loc_type;
sps->vui.chroma_sample_loc_type_bottom_field =
ctx->chroma_sample_loc_type;
sps->vui.chroma_loc_info_present_flag = 1;
need_vui = 1;
}
if (ctx->tick_rate.num && ctx->tick_rate.den) {
int num, den;
av_reduce(&num, &den, ctx->tick_rate.num, ctx->tick_rate.den,
UINT32_MAX > INT_MAX ? UINT32_MAX : INT_MAX);
sps->vui.time_scale = num;
sps->vui.num_units_in_tick = den;
sps->vui.timing_info_present_flag = 1;
need_vui = 1;
}
SET_VUI_FIELD(fixed_frame_rate_flag);
if (ctx->zero_new_constraint_set_flags) {
sps->constraint_set4_flag = 0;
sps->constraint_set5_flag = 0;
}
if (sps->separate_colour_plane_flag || sps->chroma_format_idc == 0) {
crop_unit_x = 1;
crop_unit_y = 2 - sps->frame_mbs_only_flag;
} else {
crop_unit_x = 1 + (sps->chroma_format_idc < 3);
crop_unit_y = (1 + (sps->chroma_format_idc < 2)) *
(2 - sps->frame_mbs_only_flag);
}
#define CROP(border, unit) do { \
if (ctx->crop_ ## border >= 0) { \
if (ctx->crop_ ## border % unit != 0) { \
av_log(bsf, AV_LOG_ERROR, "Invalid value for crop_%s: " \
"must be a multiple of %d.\n", #border, unit); \
return AVERROR(EINVAL); \
} \
sps->frame_crop_ ## border ## _offset = \
ctx->crop_ ## border / unit; \
sps->frame_cropping_flag = 1; \
} \
} while (0)
CROP(left, crop_unit_x);
CROP(right, crop_unit_x);
CROP(top, crop_unit_y);
CROP(bottom, crop_unit_y);
#undef CROP
if (ctx->level != LEVEL_UNSET) {
int level_idc;
if (ctx->level == LEVEL_AUTO) {
const H264LevelDescriptor *desc;
int64_t bit_rate;
int width, height, dpb_frames;
int framerate;
if (sps->vui.nal_hrd_parameters_present_flag) {
bit_rate = (sps->vui.nal_hrd_parameters.bit_rate_value_minus1[0] + 1) *
(INT64_C(1) << (sps->vui.nal_hrd_parameters.bit_rate_scale + 6));
} else if (sps->vui.vcl_hrd_parameters_present_flag) {
bit_rate = (sps->vui.vcl_hrd_parameters.bit_rate_value_minus1[0] + 1) *
(INT64_C(1) << (sps->vui.vcl_hrd_parameters.bit_rate_scale + 6));
// Adjust for VCL vs. NAL limits.
bit_rate = bit_rate * 6 / 5;
} else {
bit_rate = 0;
}
// Don't use max_dec_frame_buffering if it is only inferred.
dpb_frames = sps->vui.bitstream_restriction_flag ?
sps->vui.max_dec_frame_buffering : H264_MAX_DPB_FRAMES;
width = 16 * (sps->pic_width_in_mbs_minus1 + 1);
height = 16 * (sps->pic_height_in_map_units_minus1 + 1) *
(2 - sps->frame_mbs_only_flag);
if (sps->vui.timing_info_present_flag)
framerate = sps->vui.time_scale / sps->vui.num_units_in_tick / 2;
else
framerate = 0;
desc = ff_h264_guess_level(sps->profile_idc, bit_rate, framerate,
width, height, dpb_frames);
if (desc) {
level_idc = desc->level_idc;
} else {
av_log(bsf, AV_LOG_WARNING, "Stream does not appear to "
"conform to any level: using level 6.2.\n");
level_idc = 62;
}
} else {
level_idc = ctx->level;
}
if (level_idc == 9) {
if (sps->profile_idc == 66 ||
sps->profile_idc == 77 ||
sps->profile_idc == 88) {
sps->level_idc = 11;
sps->constraint_set3_flag = 1;
} else {
sps->level_idc = 9;
}
} else {
sps->level_idc = level_idc;
}
}
if (need_vui)
sps->vui_parameters_present_flag = 1;
return 0;
}
static int h264_metadata_handle_display_orientation(AVBSFContext *bsf,
AVPacket *pkt,
CodedBitstreamFragment *au,
int seek_point)
{
H264MetadataContext *ctx = bsf->priv_data;
SEIRawMessage *message;
int err;
message = NULL;
while (ff_cbs_sei_find_message(ctx->common.output, au,
SEI_TYPE_DISPLAY_ORIENTATION,
&message) == 0) {
H264RawSEIDisplayOrientation *disp = message->payload;
double angle = disp->anticlockwise_rotation * 180.0 / 65536.0;
int32_t *matrix;
matrix = av_malloc(9 * sizeof(int32_t));
if (!matrix)
return AVERROR(ENOMEM);
/* av_display_rotation_set() expects the angle in the clockwise
* direction, hence the first minus.
* The below code applies the flips after the rotation, yet
* the H.2645 specs require flipping to be applied first.
* Because of R O(phi) = O(-phi) R (where R is flipping around
* an arbitatry axis and O(phi) is the proper rotation by phi)
* we can create display matrices as desired by negating
* the degree once for every flip applied. */
angle = -angle * (1 - 2 * !!disp->hor_flip) * (1 - 2 * !!disp->ver_flip);
av_display_rotation_set(matrix, angle);
av_display_matrix_flip(matrix, disp->hor_flip, disp->ver_flip);
// If there are multiple display orientation messages in an
// access unit, then the last one added to the packet (i.e.
// the first one in the access unit) will prevail.
err = av_packet_add_side_data(pkt, AV_PKT_DATA_DISPLAYMATRIX,
(uint8_t*)matrix,
9 * sizeof(int32_t));
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to attach extracted "
"displaymatrix side data to packet.\n");
av_free(matrix);
return AVERROR(ENOMEM);
}
}
if (ctx->display_orientation == BSF_ELEMENT_REMOVE ||
ctx->display_orientation == BSF_ELEMENT_INSERT) {
ff_cbs_sei_delete_message_type(ctx->common.output, au,
SEI_TYPE_DISPLAY_ORIENTATION);
}
if (ctx->display_orientation == BSF_ELEMENT_INSERT) {
H264RawSEIDisplayOrientation *disp =
&ctx->display_orientation_payload;
uint8_t *data;
size_t size;
int write = 0;
data = av_packet_get_side_data(pkt, AV_PKT_DATA_DISPLAYMATRIX, &size);
if (data && size >= 9 * sizeof(int32_t)) {
int32_t matrix[9];
double dmatrix[9];
int hflip, vflip, i;
double scale_x, scale_y, angle;
memcpy(matrix, data, sizeof(matrix));
for (i = 0; i < 9; i++)
dmatrix[i] = matrix[i] / 65536.0;
// Extract scale factors.
scale_x = hypot(dmatrix[0], dmatrix[3]);
scale_y = hypot(dmatrix[1], dmatrix[4]);
// Select flips to make the main diagonal positive.
hflip = dmatrix[0] < 0.0;
vflip = dmatrix[4] < 0.0;
if (hflip)
scale_x = -scale_x;
if (vflip)
scale_y = -scale_y;
// Rescale.
for (i = 0; i < 9; i += 3) {
dmatrix[i] /= scale_x;
dmatrix[i + 1] /= scale_y;
}
// Extract rotation.
angle = atan2(dmatrix[3], dmatrix[0]);
if (!(angle >= -M_PI && angle <= M_PI) ||
matrix[2] != 0.0 || matrix[5] != 0.0 ||
matrix[6] != 0.0 || matrix[7] != 0.0) {
av_log(bsf, AV_LOG_WARNING, "Input display matrix is not "
"representable in H.264 parameters.\n");
} else {
disp->hor_flip = hflip;
disp->ver_flip = vflip;
disp->anticlockwise_rotation =
(uint16_t)rint((angle >= 0.0 ? angle
: angle + 2 * M_PI) *
32768.0 / M_PI);
write = 1;
}
}
if (seek_point) {
if (!isnan(ctx->rotate)) {
disp->anticlockwise_rotation =
(uint16_t)rint((ctx->rotate >= 0.0 ? ctx->rotate
: ctx->rotate + 360.0) *
65536.0 / 360.0);
write = 1;
}
if (ctx->flip) {
disp->hor_flip = !!(ctx->flip & FLIP_HORIZONTAL);
disp->ver_flip = !!(ctx->flip & FLIP_VERTICAL);
write = 1;
}
}
if (write) {
disp->display_orientation_repetition_period = 1;
err = ff_cbs_sei_add_message(ctx->common.output, au, 1,
SEI_TYPE_DISPLAY_ORIENTATION,
disp, NULL);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to add display orientation "
"SEI message to access unit.\n");
return err;
}
}
}
return 0;
}
static int h264_metadata_update_fragment(AVBSFContext *bsf, AVPacket *pkt,
CodedBitstreamFragment *au)
{
H264MetadataContext *ctx = bsf->priv_data;
int err, i, has_sps, seek_point;
if (ctx->aud == BSF_ELEMENT_REMOVE) {
for (i = au->nb_units - 1; i >= 0; i--) {
if (au->units[i].type == H264_NAL_AUD)
ff_cbs_delete_unit(au, i);
}
} else if (ctx->aud == BSF_ELEMENT_INSERT) {
if (pkt) {
err = h264_metadata_insert_aud(bsf, au);
if (err < 0)
return err;
}
}
has_sps = 0;
for (i = 0; i < au->nb_units; i++) {
if (au->units[i].type == H264_NAL_SPS) {
err = h264_metadata_update_sps(bsf, au->units[i].content);
if (err < 0)
return err;
has_sps = 1;
}
}
if (pkt) {
// The current packet should be treated as a seek point for metadata
// insertion if any of:
// - It is the first packet in the stream.
// - It contains an SPS, indicating that a sequence might start here.
// - It is marked as containing a key frame.
seek_point = !ctx->done_first_au || has_sps ||
(pkt->flags & AV_PKT_FLAG_KEY);
} else {
seek_point = 0;
}
if (ctx->sei_user_data && seek_point) {
err = ff_cbs_sei_add_message(ctx->common.output, au, 1,
SEI_TYPE_USER_DATA_UNREGISTERED,
&ctx->sei_user_data_payload, NULL);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to add user data SEI "
"message to access unit.\n");
return err;
}
}
if (ctx->delete_filler) {
for (i = au->nb_units - 1; i >= 0; i--) {
if (au->units[i].type == H264_NAL_FILLER_DATA) {
ff_cbs_delete_unit(au, i);
continue;
}
}
ff_cbs_sei_delete_message_type(ctx->common.output, au,
SEI_TYPE_FILLER_PAYLOAD);
}
if (pkt && ctx->display_orientation != BSF_ELEMENT_PASS) {
err = h264_metadata_handle_display_orientation(bsf, pkt, au,
seek_point);
if (err < 0)
return err;
}
if (pkt)
ctx->done_first_au = 1;
return 0;
}
static const CBSBSFType h264_metadata_type = {
.codec_id = AV_CODEC_ID_H264,
.fragment_name = "access unit",
.unit_name = "NAL unit",
.update_fragment = &h264_metadata_update_fragment,
};
static int h264_metadata_init(AVBSFContext *bsf)
{
H264MetadataContext *ctx = bsf->priv_data;
if (ctx->sei_user_data) {
SEIRawUserDataUnregistered *udu = &ctx->sei_user_data_payload;
int i, j;
// Parse UUID. It must be a hex string of length 32, possibly
// containing '-'s between hex digits (which we ignore).
for (i = j = 0; j < 32 && i < 64 && ctx->sei_user_data[i]; i++) {
int c, v;
c = ctx->sei_user_data[i];
if (c == '-') {
continue;
} else if (av_isxdigit(c)) {
c = av_tolower(c);
v = (c <= '9' ? c - '0' : c - 'a' + 10);
} else {
break;
}
if (j & 1)
udu->uuid_iso_iec_11578[j / 2] |= v;
else
udu->uuid_iso_iec_11578[j / 2] = v << 4;
++j;
}
if (j == 32 && ctx->sei_user_data[i] == '+') {
udu->data = (uint8_t*)ctx->sei_user_data + i + 1;
udu->data_length = strlen(udu->data) + 1;
} else {
av_log(bsf, AV_LOG_ERROR, "Invalid user data: "
"must be \"UUID+string\".\n");
return AVERROR(EINVAL);
}
}
return ff_cbs_bsf_generic_init(bsf, &h264_metadata_type);
}
#define OFFSET(x) offsetof(H264MetadataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption h264_metadata_options[] = {
BSF_ELEMENT_OPTIONS_PIR("aud", "Access Unit Delimiter NAL units",
aud, FLAGS),
{ "sample_aspect_ratio", "Set sample aspect ratio (table E-1)",
OFFSET(sample_aspect_ratio), AV_OPT_TYPE_RATIONAL,
{ .dbl = 0.0 }, 0, 65535, FLAGS },
{ "overscan_appropriate_flag", "Set VUI overscan appropriate flag",
OFFSET(overscan_appropriate_flag), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 1, FLAGS },
{ "video_format", "Set video format (table E-2)",
OFFSET(video_format), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 7, FLAGS},
{ "video_full_range_flag", "Set video full range flag",
OFFSET(video_full_range_flag), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 1, FLAGS },
{ "colour_primaries", "Set colour primaries (table E-3)",
OFFSET(colour_primaries), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "transfer_characteristics", "Set transfer characteristics (table E-4)",
OFFSET(transfer_characteristics), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "matrix_coefficients", "Set matrix coefficients (table E-5)",
OFFSET(matrix_coefficients), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "chroma_sample_loc_type", "Set chroma sample location type (figure E-1)",
OFFSET(chroma_sample_loc_type), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 5, FLAGS },
{ "tick_rate", "Set VUI tick rate (time_scale / num_units_in_tick)",
OFFSET(tick_rate), AV_OPT_TYPE_RATIONAL,
{ .dbl = 0.0 }, 0, UINT_MAX, FLAGS },
{ "fixed_frame_rate_flag", "Set VUI fixed frame rate flag",
OFFSET(fixed_frame_rate_flag), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 1, FLAGS },
{ "zero_new_constraint_set_flags", "Set constraint_set4_flag / constraint_set5_flag to zero",
OFFSET(zero_new_constraint_set_flags), AV_OPT_TYPE_BOOL,
{ .i64 = 0 }, 0, 1, FLAGS },
{ "crop_left", "Set left border crop offset",
OFFSET(crop_left), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, H264_MAX_WIDTH, FLAGS },
{ "crop_right", "Set right border crop offset",
OFFSET(crop_right), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, H264_MAX_WIDTH, FLAGS },
{ "crop_top", "Set top border crop offset",
OFFSET(crop_top), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, H264_MAX_HEIGHT, FLAGS },
{ "crop_bottom", "Set bottom border crop offset",
OFFSET(crop_bottom), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, H264_MAX_HEIGHT, FLAGS },
{ "sei_user_data", "Insert SEI user data (UUID+string)",
OFFSET(sei_user_data), AV_OPT_TYPE_STRING, { .str = NULL }, .flags = FLAGS },
{ "delete_filler", "Delete all filler (both NAL and SEI)",
OFFSET(delete_filler), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, FLAGS},
BSF_ELEMENT_OPTIONS_PIRE("display_orientation",
"Display orientation SEI",
display_orientation, FLAGS),
{ "rotate", "Set rotation in display orientation SEI (anticlockwise angle in degrees)",
OFFSET(rotate), AV_OPT_TYPE_DOUBLE,
{ .dbl = NAN }, -360.0, +360.0, FLAGS },
{ "flip", "Set flip in display orientation SEI",
OFFSET(flip), AV_OPT_TYPE_FLAGS,
{ .i64 = 0 }, 0, FLIP_HORIZONTAL | FLIP_VERTICAL, FLAGS, .unit = "flip" },
{ "horizontal", "Set hor_flip",
0, AV_OPT_TYPE_CONST,
{ .i64 = FLIP_HORIZONTAL }, .flags = FLAGS, .unit = "flip" },
{ "vertical", "Set ver_flip",
0, AV_OPT_TYPE_CONST,
{ .i64 = FLIP_VERTICAL }, .flags = FLAGS, .unit = "flip" },
{ "level", "Set level (table A-1)",
OFFSET(level), AV_OPT_TYPE_INT,
{ .i64 = LEVEL_UNSET }, LEVEL_UNSET, 0xff, FLAGS, .unit = "level" },
{ "auto", "Attempt to guess level from stream properties",
0, AV_OPT_TYPE_CONST,
{ .i64 = LEVEL_AUTO }, .flags = FLAGS, .unit = "level" },
#define LEVEL(name, value) name, NULL, 0, AV_OPT_TYPE_CONST, \
{ .i64 = value }, .flags = FLAGS, .unit = "level"
{ LEVEL("1", 10) },
{ LEVEL("1b", 9) },
{ LEVEL("1.1", 11) },
{ LEVEL("1.2", 12) },
{ LEVEL("1.3", 13) },
{ LEVEL("2", 20) },
{ LEVEL("2.1", 21) },
{ LEVEL("2.2", 22) },
{ LEVEL("3", 30) },
{ LEVEL("3.1", 31) },
{ LEVEL("3.2", 32) },
{ LEVEL("4", 40) },
{ LEVEL("4.1", 41) },
{ LEVEL("4.2", 42) },
{ LEVEL("5", 50) },
{ LEVEL("5.1", 51) },
{ LEVEL("5.2", 52) },
{ LEVEL("6", 60) },
{ LEVEL("6.1", 61) },
{ LEVEL("6.2", 62) },
#undef LEVEL
{ NULL }
};
static const AVClass h264_metadata_class = {
.class_name = "h264_metadata_bsf",
.item_name = av_default_item_name,
.option = h264_metadata_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID h264_metadata_codec_ids[] = {
AV_CODEC_ID_H264, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_h264_metadata_bsf = {
.p.name = "h264_metadata",
.p.codec_ids = h264_metadata_codec_ids,
.p.priv_class = &h264_metadata_class,
.priv_data_size = sizeof(H264MetadataContext),
.init = &h264_metadata_init,
.close = &ff_cbs_bsf_generic_close,
.filter = &ff_cbs_bsf_generic_filter,
};
+473
View File
@@ -0,0 +1,473 @@
/*
* H.264 MP4 to Annex B byte stream format filter
* Copyright (c) 2007 Benoit Fouet <benoit.fouet@free.fr>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <string.h>
#include "libavutil/avassert.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/mem.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "defs.h"
#include "h264.h"
#include "sei.h"
typedef struct H264BSFContext {
uint8_t *sps;
uint8_t *pps;
int sps_size;
int pps_size;
unsigned sps_buf_size;
unsigned pps_buf_size;
uint8_t length_size;
uint8_t new_idr;
uint8_t idr_sps_seen;
uint8_t idr_pps_seen;
int extradata_parsed;
} H264BSFContext;
enum PsSource {
PS_OUT_OF_BAND = -1,
PS_NONE = 0,
PS_IN_BAND = 1,
};
static void count_or_copy(uint8_t **out, uint64_t *out_size,
const uint8_t *in, int in_size, enum PsSource ps, int copy)
{
uint8_t start_code_size;
if (ps == PS_OUT_OF_BAND)
/* start code already present in out-of-band ps data, so don't need to
* add it manually again
*/
start_code_size = 0;
else if (ps == PS_IN_BAND || *out_size == 0)
start_code_size = 4;
else
start_code_size = 3;
if (copy) {
memcpy(*out + start_code_size, in, in_size);
if (start_code_size == 4) {
AV_WB32(*out, 1);
} else if (start_code_size) {
(*out)[0] =
(*out)[1] = 0;
(*out)[2] = 1;
}
*out += start_code_size + in_size;
}
*out_size += start_code_size + in_size;
}
static int h264_extradata_to_annexb(AVBSFContext *ctx,
uint8_t *extradata, int extradata_size)
{
H264BSFContext *s = ctx->priv_data;
GetByteContext ogb, *gb = &ogb;
uint16_t unit_size;
uint32_t total_size = 0;
uint8_t *out = NULL, unit_nb, sps_done = 0;
static const uint8_t nalu_header[4] = { 0, 0, 0, 1 };
const int padding = AV_INPUT_BUFFER_PADDING_SIZE;
int length_size, pps_offset = 0;
if (extradata_size < 7) {
av_log(ctx, AV_LOG_ERROR, "Invalid extradata size: %d\n", extradata_size);
return AVERROR_INVALIDDATA;
}
bytestream2_init(gb, extradata, extradata_size);
bytestream2_skipu(gb, 4);
/* retrieve length coded size */
length_size = (bytestream2_get_byteu(gb) & 0x3) + 1;
/* retrieve sps and pps unit(s) */
unit_nb = bytestream2_get_byteu(gb) & 0x1f; /* number of sps unit(s) */
if (!unit_nb) {
goto pps;
}
while (unit_nb--) {
int err;
/* possible overread ok due to padding */
unit_size = bytestream2_get_be16u(gb);
total_size += unit_size + 4;
av_assert1(total_size <= INT_MAX - padding);
if (bytestream2_get_bytes_left(gb) < unit_size + !sps_done) {
av_log(ctx, AV_LOG_ERROR, "Global extradata truncated, "
"corrupted stream or invalid MP4/AVCC bitstream\n");
av_free(out);
return AVERROR_INVALIDDATA;
}
if ((err = av_reallocp(&out, total_size + padding)) < 0)
return err;
memcpy(out + total_size - unit_size - 4, nalu_header, 4);
bytestream2_get_bufferu(gb, out + total_size - unit_size, unit_size);
pps:
if (!unit_nb && !sps_done++) {
unit_nb = bytestream2_get_byteu(gb); /* number of pps unit(s) */
pps_offset = total_size;
}
}
if (out)
memset(out + total_size, 0, padding);
if (pps_offset) {
uint8_t *sps;
s->sps_size = pps_offset;
sps = av_fast_realloc(s->sps, &s->sps_buf_size, s->sps_size);
if (!sps) {
av_free(out);
return AVERROR(ENOMEM);
}
s->sps = sps;
memcpy(s->sps, out, s->sps_size);
} else {
av_log(ctx, AV_LOG_WARNING,
"Warning: SPS NALU missing or invalid. "
"The resulting stream may not play.\n");
}
if (pps_offset < total_size) {
uint8_t *pps;
s->pps_size = total_size - pps_offset;
pps = av_fast_realloc(s->pps, &s->pps_buf_size, s->pps_size);
if (!pps) {
av_freep(&s->sps);
av_free(out);
return AVERROR(ENOMEM);
}
s->pps = pps;
memcpy(s->pps, out + pps_offset, s->pps_size);
} else {
av_log(ctx, AV_LOG_WARNING,
"Warning: PPS NALU missing or invalid. "
"The resulting stream may not play.\n");
}
av_freep(&ctx->par_out->extradata);
ctx->par_out->extradata = out;
ctx->par_out->extradata_size = total_size;
s->length_size = length_size;
s->new_idr = 1;
s->idr_sps_seen = 0;
s->idr_pps_seen = 0;
s->extradata_parsed = 1;
return 0;
}
static int h264_mp4toannexb_save_ps(uint8_t **dst, int *dst_size,
unsigned *dst_buf_size,
const uint8_t *nal, uint32_t nal_size,
int first)
{
static const uint8_t nalu_header[4] = { 0, 0, 0, 1 };
const int start_code_size = sizeof(nalu_header);
uint8_t *ptr;
uint32_t size;
if (first)
size = 0;
else
size = *dst_size;
ptr = av_fast_realloc(*dst, dst_buf_size, size + nal_size + start_code_size);
if (!ptr)
return AVERROR(ENOMEM);
memcpy(ptr + size, nalu_header, start_code_size);
size += start_code_size;
memcpy(ptr + size, nal, nal_size);
size += nal_size;
*dst = ptr;
*dst_size = size;
return 0;
}
static int h264_mp4toannexb_filter_ps(H264BSFContext *s,
const uint8_t *buf,
const uint8_t *buf_end)
{
int sps_count = 0;
int pps_count = 0;
uint8_t unit_type;
do {
uint32_t nal_size = 0;
/* possible overread ok due to padding */
for (int i = 0; i < s->length_size; i++)
nal_size = (nal_size << 8) | buf[i];
buf += s->length_size;
/* This check requires the cast as the right side might
* otherwise be promoted to an unsigned value. */
if ((int64_t)nal_size > buf_end - buf)
return AVERROR_INVALIDDATA;
if (!nal_size)
continue;
unit_type = *buf & 0x1f;
if (unit_type == H264_NAL_SPS) {
h264_mp4toannexb_save_ps(&s->sps, &s->sps_size, &s->sps_buf_size, buf,
nal_size, !sps_count);
sps_count++;
} else if (unit_type == H264_NAL_PPS) {
h264_mp4toannexb_save_ps(&s->pps, &s->pps_size, &s->pps_buf_size, buf,
nal_size, !pps_count);
pps_count++;
}
buf += nal_size;
} while (buf < buf_end);
return 0;
}
static int h264_mp4toannexb_init(AVBSFContext *ctx)
{
int extra_size = ctx->par_in->extradata_size;
/* retrieve sps and pps NAL units from extradata */
if (!extra_size ||
(extra_size >= 3 && AV_RB24(ctx->par_in->extradata) == 1) ||
(extra_size >= 4 && AV_RB32(ctx->par_in->extradata) == 1)) {
av_log(ctx, AV_LOG_VERBOSE,
"The input looks like it is Annex B already\n");
return 0;
}
return h264_extradata_to_annexb(ctx,
ctx->par_in->extradata,
ctx->par_in->extradata_size);
}
static int h264_mp4toannexb_filter(AVBSFContext *ctx, AVPacket *opkt)
{
H264BSFContext *s = ctx->priv_data;
AVPacket *in;
uint8_t unit_type, new_idr, sps_seen, pps_seen;
const uint8_t *buf;
const uint8_t *buf_end;
uint8_t *out;
uint64_t out_size;
int ret;
size_t extradata_size;
uint8_t *extradata;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
extradata = av_packet_get_side_data(in, AV_PKT_DATA_NEW_EXTRADATA,
&extradata_size);
if (extradata && extradata[0] == 1) {
ret = h264_extradata_to_annexb(ctx, extradata, extradata_size);
if (ret < 0)
goto fail;
av_packet_side_data_remove(in->side_data, &in->side_data_elems,
AV_PKT_DATA_NEW_EXTRADATA);
}
/* nothing to filter */
if (!s->extradata_parsed) {
av_packet_move_ref(opkt, in);
av_packet_free(&in);
return 0;
}
buf_end = in->data + in->size;
ret = h264_mp4toannexb_filter_ps(s, in->data, buf_end);
if (ret < 0)
goto fail;
#define LOG_ONCE(...) \
if (j) \
av_log(__VA_ARGS__)
for (int j = 0; j < 2; j++) {
buf = in->data;
new_idr = s->new_idr;
sps_seen = s->idr_sps_seen;
pps_seen = s->idr_pps_seen;
out_size = 0;
do {
uint32_t nal_size = 0;
enum PsSource ps;
/* possible overread ok due to padding */
for (int i = 0; i < s->length_size; i++)
nal_size = (nal_size << 8) | buf[i];
buf += s->length_size;
/* This check requires the cast as the right side might
* otherwise be promoted to an unsigned value. */
if ((int64_t)nal_size > buf_end - buf) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (!nal_size)
continue;
unit_type = *buf & 0x1f;
if (unit_type == H264_NAL_SPS) {
sps_seen = new_idr = 1;
} else if (unit_type == H264_NAL_PPS) {
pps_seen = new_idr = 1;
/* if SPS has not been seen yet, prepend the AVCC one to PPS */
if (!sps_seen) {
if (!s->sps_size) {
LOG_ONCE(ctx, AV_LOG_WARNING, "SPS not present in the stream, nor in AVCC, stream may be unreadable\n");
} else {
count_or_copy(&out, &out_size, s->sps, s->sps_size, PS_OUT_OF_BAND, j);
sps_seen = 1;
}
}
}
/* If this is a new IDR picture following an IDR picture, reset the idr flag.
* Just check first_mb_in_slice to be 0 as this is the simplest solution.
* This could be checking idr_pic_id instead, but would complexify the parsing. */
if (!new_idr && unit_type == H264_NAL_IDR_SLICE && (buf[1] & 0x80))
new_idr = 1;
/* If this is a buffering period SEI without a corresponding sps/pps
* then prepend any existing sps/pps before the SEI */
if (unit_type == H264_NAL_SEI && buf[1] == SEI_TYPE_BUFFERING_PERIOD &&
!sps_seen && !pps_seen) {
if (s->sps_size) {
count_or_copy(&out, &out_size, s->sps, s->sps_size, PS_OUT_OF_BAND, j);
sps_seen = 1;
}
if (s->pps_size) {
count_or_copy(&out, &out_size, s->pps, s->pps_size, PS_OUT_OF_BAND, j);
pps_seen = 1;
}
}
/* prepend only to the first type 5 NAL unit of an IDR picture, if no sps/pps are already present */
if (new_idr && unit_type == H264_NAL_IDR_SLICE && !sps_seen && !pps_seen) {
if (s->sps_size)
count_or_copy(&out, &out_size, s->sps, s->sps_size, PS_OUT_OF_BAND, j);
if (s->pps_size)
count_or_copy(&out, &out_size, s->pps, s->pps_size, PS_OUT_OF_BAND, j);
new_idr = 0;
/* if only SPS has been seen, also insert PPS */
} else if (new_idr && unit_type == H264_NAL_IDR_SLICE && sps_seen && !pps_seen) {
if (!s->pps_size) {
LOG_ONCE(ctx, AV_LOG_WARNING, "PPS not present in the stream, nor in AVCC, stream may be unreadable\n");
} else {
count_or_copy(&out, &out_size, s->pps, s->pps_size, PS_OUT_OF_BAND, j);
}
}
if (unit_type == H264_NAL_SPS || unit_type == H264_NAL_PPS)
ps = PS_IN_BAND;
else
ps = PS_NONE;
count_or_copy(&out, &out_size, buf, nal_size, ps, j);
if (unit_type == H264_NAL_SLICE) {
new_idr = 1;
sps_seen = 0;
pps_seen = 0;
}
buf += nal_size;
} while (buf < buf_end);
if (!j) {
if (out_size > INT_MAX - AV_INPUT_BUFFER_PADDING_SIZE) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
ret = av_new_packet(opkt, out_size);
if (ret < 0)
goto fail;
out = opkt->data;
}
}
#undef LOG_ONCE
av_assert1(out_size == opkt->size);
s->new_idr = new_idr;
s->idr_sps_seen = sps_seen;
s->idr_pps_seen = pps_seen;
ret = av_packet_copy_props(opkt, in);
if (ret < 0)
goto fail;
fail:
if (ret < 0)
av_packet_unref(opkt);
av_packet_free(&in);
return ret;
}
static void h264_mp4toannexb_close(AVBSFContext *ctx)
{
H264BSFContext *s = ctx->priv_data;
av_freep(&s->sps);
av_freep(&s->pps);
}
static void h264_mp4toannexb_flush(AVBSFContext *ctx)
{
H264BSFContext *s = ctx->priv_data;
s->idr_sps_seen = 0;
s->idr_pps_seen = 0;
s->new_idr = s->extradata_parsed;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_H264, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_h264_mp4toannexb_bsf = {
.p.name = "h264_mp4toannexb",
.p.codec_ids = codec_ids,
.priv_data_size = sizeof(H264BSFContext),
.init = h264_mp4toannexb_init,
.filter = h264_mp4toannexb_filter,
.close = h264_mp4toannexb_close,
.flush = h264_mp4toannexb_flush,
};
+126
View File
@@ -0,0 +1,126 @@
/*
* 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 <inttypes.h>
#include "libavutil/log.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_h264.h"
#include "codec_id.h"
#include "h264.h"
#include "packet.h"
#define NEW_GLOBAL_PIC_INIT_QP 26
typedef struct H264RedundantPPSContext {
CBSBSFContext common;
} H264RedundantPPSContext;
static int h264_redundant_pps_fixup_pps(H264RedundantPPSContext *ctx,
CodedBitstreamUnit *unit)
{
H264RawPPS *pps;
int err;
// The changes we are about to perform affect the parsing process,
// so we must make sure that the PPS is writable, otherwise the
// parsing of future slices will be incorrect and even raise errors.
err = ff_cbs_make_unit_writable(ctx->common.input, unit);
if (err < 0)
return err;
pps = unit->content;
// Overwrite pic_init_qp with the global value.
pps->pic_init_qp_minus26 = NEW_GLOBAL_PIC_INIT_QP - 26;
// Some PPSs have this set, so it must be set in all of them.
// (Slices which do not use such a PPS on input will still have
// *_weight_l*flag as zero and therefore write equivalently.)
pps->weighted_pred_flag = 1;
return 0;
}
static int h264_redundant_pps_fixup_slice(H264RedundantPPSContext *ctx,
H264RawSliceHeader *slice)
{
const CodedBitstreamH264Context *const in = ctx->common.input->priv_data;
const H264RawPPS *const pps = in->pps[slice->pic_parameter_set_id];
// We modified the PPS's qp value, now offset this by applying
// the negative offset to the slices.
slice->slice_qp_delta += pps->pic_init_qp_minus26
- (NEW_GLOBAL_PIC_INIT_QP - 26);
return 0;
}
static int h264_redundant_pps_update_fragment(AVBSFContext *bsf,
AVPacket *pkt,
CodedBitstreamFragment *au)
{
H264RedundantPPSContext *ctx = bsf->priv_data;
int err, i;
for (i = 0; i < au->nb_units; i++) {
CodedBitstreamUnit *nal = &au->units[i];
if (nal->type == H264_NAL_PPS) {
err = h264_redundant_pps_fixup_pps(ctx, nal);
if (err < 0)
return err;
}
if (nal->type == H264_NAL_SLICE ||
nal->type == H264_NAL_IDR_SLICE) {
H264RawSlice *slice = nal->content;
h264_redundant_pps_fixup_slice(ctx, &slice->header);
}
}
return 0;
}
static const CBSBSFType h264_redundant_pps_type = {
.codec_id = AV_CODEC_ID_H264,
.fragment_name = "access unit",
.unit_name = "NAL unit",
.update_fragment = &h264_redundant_pps_update_fragment,
};
static int h264_redundant_pps_init(AVBSFContext *bsf)
{
return ff_cbs_bsf_generic_init(bsf, &h264_redundant_pps_type);
}
static const enum AVCodecID h264_redundant_pps_codec_ids[] = {
AV_CODEC_ID_H264, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_h264_redundant_pps_bsf = {
.p.name = "h264_redundant_pps",
.p.codec_ids = h264_redundant_pps_codec_ids,
.priv_data_size = sizeof(H264RedundantPPSContext),
.init = &h264_redundant_pps_init,
.close = &ff_cbs_bsf_generic_close,
.filter = &ff_cbs_bsf_generic_filter,
};
+599
View File
@@ -0,0 +1,599 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/common.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_h265.h"
#include "h2645data.h"
#include "h265_profile_level.h"
#include "hevc/hevc.h"
enum {
LEVEL_UNSET = -2,
LEVEL_AUTO = -1,
};
typedef struct H265MetadataContext {
CBSBSFContext common;
H265RawAUD aud_nal;
int aud;
AVRational sample_aspect_ratio;
int video_format;
int video_full_range_flag;
int colour_primaries;
int transfer_characteristics;
int matrix_coefficients;
int chroma_sample_loc_type;
AVRational tick_rate;
int poc_proportional_to_timing_flag;
int num_ticks_poc_diff_one;
int crop_left;
int crop_right;
int crop_top;
int crop_bottom;
int width;
int height;
int level;
int level_guess;
int level_warned;
} H265MetadataContext;
static void h265_metadata_guess_level(AVBSFContext *bsf,
const CodedBitstreamFragment *au)
{
H265MetadataContext *ctx = bsf->priv_data;
const H265LevelDescriptor *desc;
const H265RawProfileTierLevel *ptl = NULL;
const H265RawHRDParameters *hrd = NULL;
int64_t bit_rate = 0;
int width = 0, height = 0;
int tile_cols = 0, tile_rows = 0;
int max_dec_pic_buffering = 0;
int i;
for (i = 0; i < au->nb_units; i++) {
const CodedBitstreamUnit *unit = &au->units[i];
if (unit->type == HEVC_NAL_VPS) {
const H265RawVPS *vps = unit->content;
ptl = &vps->profile_tier_level;
max_dec_pic_buffering = vps->vps_max_dec_pic_buffering_minus1[0] + 1;
if (vps->vps_num_hrd_parameters > 0)
hrd = &vps->hrd_parameters[0];
} else if (unit->type == HEVC_NAL_SPS) {
const H265RawSPS *sps = unit->content;
ptl = &sps->profile_tier_level;
max_dec_pic_buffering = sps->sps_max_dec_pic_buffering_minus1[0] + 1;
width = sps->pic_width_in_luma_samples;
height = sps->pic_height_in_luma_samples;
if (sps->vui.vui_hrd_parameters_present_flag)
hrd = &sps->vui.hrd_parameters;
} else if (unit->type == HEVC_NAL_PPS) {
const H265RawPPS *pps = unit->content;
if (pps->tiles_enabled_flag) {
tile_cols = pps->num_tile_columns_minus1 + 1;
tile_rows = pps->num_tile_rows_minus1 + 1;
}
}
}
if (hrd) {
if (hrd->nal_hrd_parameters_present_flag) {
bit_rate = (hrd->nal_sub_layer_hrd_parameters[0].bit_rate_value_minus1[0] + 1) *
(INT64_C(1) << hrd->bit_rate_scale + 6);
} else if (hrd->vcl_hrd_parameters_present_flag) {
bit_rate = (hrd->vcl_sub_layer_hrd_parameters[0].bit_rate_value_minus1[0] + 1) *
(INT64_C(1) << hrd->bit_rate_scale + 6);
// Adjust for VCL vs. NAL limits.
bit_rate = bit_rate * 11 / 10;
}
}
desc = ff_h265_guess_level(ptl, bit_rate, width, height,
0, tile_rows, tile_cols,
max_dec_pic_buffering);
if (desc) {
av_log(bsf, AV_LOG_DEBUG, "Stream appears to conform to "
"level %s.\n", desc->name);
ctx->level_guess = desc->level_idc;
}
}
static void h265_metadata_update_level(AVBSFContext *bsf,
uint8_t *level_idc)
{
H265MetadataContext *ctx = bsf->priv_data;
if (ctx->level != LEVEL_UNSET) {
if (ctx->level == LEVEL_AUTO) {
if (ctx->level_guess) {
*level_idc = ctx->level_guess;
} else {
if (!ctx->level_warned) {
av_log(bsf, AV_LOG_WARNING, "Unable to determine level "
"of stream: using level 8.5.\n");
ctx->level_warned = 1;
}
*level_idc = 255;
}
} else {
*level_idc = ctx->level;
}
}
}
static int h265_metadata_update_vps(AVBSFContext *bsf,
H265RawVPS *vps)
{
H265MetadataContext *ctx = bsf->priv_data;
if (ctx->tick_rate.num && ctx->tick_rate.den) {
int num, den;
av_reduce(&num, &den, ctx->tick_rate.num, ctx->tick_rate.den,
UINT32_MAX > INT_MAX ? UINT32_MAX : INT_MAX);
vps->vps_time_scale = num;
vps->vps_num_units_in_tick = den;
vps->vps_timing_info_present_flag = 1;
if (ctx->num_ticks_poc_diff_one > 0) {
vps->vps_num_ticks_poc_diff_one_minus1 =
ctx->num_ticks_poc_diff_one - 1;
vps->vps_poc_proportional_to_timing_flag = 1;
} else if (ctx->num_ticks_poc_diff_one == 0) {
vps->vps_poc_proportional_to_timing_flag = 0;
}
}
h265_metadata_update_level(bsf, &vps->profile_tier_level.general_level_idc);
return 0;
}
static int h265_metadata_deduce_crop(AVBSFContext *bsf, const H265RawSPS *sps,
int *crop_left, int *crop_right,
int *crop_top, int *crop_bottom)
{
const H265MetadataContext *ctx = bsf->priv_data;
int left = ctx->crop_left;
int right = ctx->crop_right;
int top = ctx->crop_top;
int bottom = ctx->crop_bottom;
if (ctx->width > 0) {
if (ctx->width > sps->pic_width_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"The width option value %d is larger than picture width %d\n",
ctx->width, sps->pic_width_in_luma_samples);
return AVERROR(EINVAL);
}
if (left < 0) {
if (right > 0)
left = sps->pic_width_in_luma_samples - ctx->width - right;
else
left = 0;
}
if (right < 0)
right = sps->pic_width_in_luma_samples - ctx->width - left;
if (left < 0 || right < 0 || (left + right + ctx->width) !=
sps->pic_width_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"Invalid value for crop_left %d, crop_right %d, width after "
"crop %d, with picture width %d\n",
ctx->crop_left, ctx->crop_right, ctx->width,
sps->pic_width_in_luma_samples);
return AVERROR(EINVAL);
}
}
if (ctx->height > 0) {
if (ctx->height > sps->pic_height_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"The height option value %d is larger than picture height %d\n",
ctx->height, sps->pic_height_in_luma_samples);
return AVERROR(EINVAL);
}
if (top < 0) {
if (bottom > 0)
top = sps->pic_height_in_luma_samples - ctx->height - bottom;
else
top = 0;
}
if (bottom < 0)
bottom = sps->pic_height_in_luma_samples - ctx->height - top;
if (top < 0 || bottom < 0 || (top + bottom + ctx->height) !=
sps->pic_height_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"Invalid value for crop_top %d, crop_bottom %d, height after "
"crop %d, with picture height %d\n",
ctx->crop_top, ctx->crop_bottom, ctx->height,
sps->pic_height_in_luma_samples);
return AVERROR(EINVAL);
}
}
*crop_left = left;
*crop_right = right;
*crop_top = top;
*crop_bottom = bottom;
return 0;
}
static int h265_metadata_update_sps(AVBSFContext *bsf,
H265RawSPS *sps)
{
H265MetadataContext *ctx = bsf->priv_data;
int need_vui = 0;
int crop_unit_x, crop_unit_y;
/* Use local variables to avoid modifying context fields in case of video
* resolution changed. Crop doesn't work well with resolution change, this
* is the best we can do.
*/
int crop_left, crop_right, crop_top, crop_bottom;
int ret;
if (ctx->sample_aspect_ratio.num && ctx->sample_aspect_ratio.den) {
int num, den, i;
av_reduce(&num, &den, ctx->sample_aspect_ratio.num,
ctx->sample_aspect_ratio.den, 65535);
for (i = 1; i < FF_ARRAY_ELEMS(ff_h2645_pixel_aspect); i++) {
if (num == ff_h2645_pixel_aspect[i].num &&
den == ff_h2645_pixel_aspect[i].den)
break;
}
if (i == FF_ARRAY_ELEMS(ff_h2645_pixel_aspect)) {
sps->vui.aspect_ratio_idc = 255;
sps->vui.sar_width = num;
sps->vui.sar_height = den;
} else {
sps->vui.aspect_ratio_idc = i;
}
sps->vui.aspect_ratio_info_present_flag = 1;
need_vui = 1;
}
#define SET_OR_INFER(field, value, present_flag, infer) do { \
if (value >= 0) { \
field = value; \
need_vui = 1; \
} else if (!present_flag) \
field = infer; \
} while (0)
if (ctx->video_format >= 0 ||
ctx->video_full_range_flag >= 0 ||
ctx->colour_primaries >= 0 ||
ctx->transfer_characteristics >= 0 ||
ctx->matrix_coefficients >= 0) {
SET_OR_INFER(sps->vui.video_format, ctx->video_format,
sps->vui.video_signal_type_present_flag, 5);
SET_OR_INFER(sps->vui.video_full_range_flag,
ctx->video_full_range_flag,
sps->vui.video_signal_type_present_flag, 0);
if (ctx->colour_primaries >= 0 ||
ctx->transfer_characteristics >= 0 ||
ctx->matrix_coefficients >= 0) {
SET_OR_INFER(sps->vui.colour_primaries,
ctx->colour_primaries,
sps->vui.colour_description_present_flag, 2);
SET_OR_INFER(sps->vui.transfer_characteristics,
ctx->transfer_characteristics,
sps->vui.colour_description_present_flag, 2);
SET_OR_INFER(sps->vui.matrix_coefficients,
ctx->matrix_coefficients,
sps->vui.colour_description_present_flag, 2);
sps->vui.colour_description_present_flag = 1;
}
sps->vui.video_signal_type_present_flag = 1;
need_vui = 1;
}
if (ctx->chroma_sample_loc_type >= 0) {
sps->vui.chroma_sample_loc_type_top_field =
ctx->chroma_sample_loc_type;
sps->vui.chroma_sample_loc_type_bottom_field =
ctx->chroma_sample_loc_type;
sps->vui.chroma_loc_info_present_flag = 1;
need_vui = 1;
}
if (ctx->tick_rate.num && ctx->tick_rate.den) {
int num, den;
av_reduce(&num, &den, ctx->tick_rate.num, ctx->tick_rate.den,
UINT32_MAX > INT_MAX ? UINT32_MAX : INT_MAX);
sps->vui.vui_time_scale = num;
sps->vui.vui_num_units_in_tick = den;
sps->vui.vui_timing_info_present_flag = 1;
need_vui = 1;
if (ctx->num_ticks_poc_diff_one > 0) {
sps->vui.vui_num_ticks_poc_diff_one_minus1 =
ctx->num_ticks_poc_diff_one - 1;
sps->vui.vui_poc_proportional_to_timing_flag = 1;
} else if (ctx->num_ticks_poc_diff_one == 0) {
sps->vui.vui_poc_proportional_to_timing_flag = 0;
}
}
ret = h265_metadata_deduce_crop(bsf, sps, &crop_left, &crop_right,
&crop_top, &crop_bottom);
if (ret < 0)
return ret;
if (sps->separate_colour_plane_flag || sps->chroma_format_idc == 0) {
crop_unit_x = 1;
crop_unit_y = 1;
} else {
crop_unit_x = 1 + (sps->chroma_format_idc < 3);
crop_unit_y = 1 + (sps->chroma_format_idc < 2);
}
#define CROP(border, unit) do { \
if (crop_ ## border >= 0) { \
if (crop_ ## border % unit != 0) { \
av_log(bsf, AV_LOG_ERROR, "Invalid value for crop_%s: " \
"must be a multiple of %d.\n", #border, unit); \
return AVERROR(EINVAL); \
} \
sps->conf_win_ ## border ## _offset = \
crop_ ## border / unit; \
sps->conformance_window_flag = 1; \
} \
} while (0)
CROP(left, crop_unit_x);
CROP(right, crop_unit_x);
CROP(top, crop_unit_y);
CROP(bottom, crop_unit_y);
#undef CROP
if (need_vui)
sps->vui_parameters_present_flag = 1;
h265_metadata_update_level(bsf, &sps->profile_tier_level.general_level_idc);
return 0;
}
static int h265_metadata_update_fragment(AVBSFContext *bsf, AVPacket *pkt,
CodedBitstreamFragment *au)
{
H265MetadataContext *ctx = bsf->priv_data;
int err, i;
// If an AUD is present, it must be the first NAL unit.
if (au->nb_units && au->units[0].type == HEVC_NAL_AUD) {
if (ctx->aud == BSF_ELEMENT_REMOVE)
ff_cbs_delete_unit(au, 0);
} else {
if (pkt && ctx->aud == BSF_ELEMENT_INSERT) {
H265RawAUD *aud = &ctx->aud_nal;
int pic_type = 0, temporal_id = 8, layer_id = 0;
for (i = 0; i < au->nb_units; i++) {
const H265RawNALUnitHeader *nal = au->units[i].content;
if (!nal)
continue;
if (nal->nuh_temporal_id_plus1 < temporal_id + 1)
temporal_id = nal->nuh_temporal_id_plus1 - 1;
if (au->units[i].type <= HEVC_NAL_RSV_VCL31) {
const H265RawSlice *slice = au->units[i].content;
layer_id = nal->nuh_layer_id;
if (slice->header.slice_type == HEVC_SLICE_B &&
pic_type < 2)
pic_type = 2;
if (slice->header.slice_type == HEVC_SLICE_P &&
pic_type < 1)
pic_type = 1;
}
}
aud->nal_unit_header = (H265RawNALUnitHeader) {
.nal_unit_type = HEVC_NAL_AUD,
.nuh_layer_id = layer_id,
.nuh_temporal_id_plus1 = temporal_id + 1,
};
aud->pic_type = pic_type;
err = ff_cbs_insert_unit_content(au, 0, HEVC_NAL_AUD, aud, NULL);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to insert AUD.\n");
return err;
}
}
}
if (ctx->level == LEVEL_AUTO && !ctx->level_guess)
h265_metadata_guess_level(bsf, au);
for (i = 0; i < au->nb_units; i++) {
if (au->units[i].type == HEVC_NAL_VPS) {
err = h265_metadata_update_vps(bsf, au->units[i].content);
if (err < 0)
return err;
}
if (au->units[i].type == HEVC_NAL_SPS) {
err = h265_metadata_update_sps(bsf, au->units[i].content);
if (err < 0)
return err;
}
}
return 0;
}
static const CBSBSFType h265_metadata_type = {
.codec_id = AV_CODEC_ID_HEVC,
.fragment_name = "access unit",
.unit_name = "NAL unit",
.update_fragment = &h265_metadata_update_fragment,
};
static int h265_metadata_init(AVBSFContext *bsf)
{
return ff_cbs_bsf_generic_init(bsf, &h265_metadata_type);
}
#define OFFSET(x) offsetof(H265MetadataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption h265_metadata_options[] = {
BSF_ELEMENT_OPTIONS_PIR("aud", "Access Unit Delimiter NAL units",
aud, FLAGS),
{ "sample_aspect_ratio", "Set sample aspect ratio (table E-1)",
OFFSET(sample_aspect_ratio), AV_OPT_TYPE_RATIONAL,
{ .dbl = 0.0 }, 0, 65535, FLAGS },
{ "video_format", "Set video format (table E-2)",
OFFSET(video_format), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 7, FLAGS },
{ "video_full_range_flag", "Set video full range flag",
OFFSET(video_full_range_flag), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 1, FLAGS },
{ "colour_primaries", "Set colour primaries (table E-3)",
OFFSET(colour_primaries), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "transfer_characteristics", "Set transfer characteristics (table E-4)",
OFFSET(transfer_characteristics), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "matrix_coefficients", "Set matrix coefficients (table E-5)",
OFFSET(matrix_coefficients), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "chroma_sample_loc_type", "Set chroma sample location type (figure E-1)",
OFFSET(chroma_sample_loc_type), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 5, FLAGS },
{ "tick_rate",
"Set VPS and VUI tick rate (time_scale / num_units_in_tick)",
OFFSET(tick_rate), AV_OPT_TYPE_RATIONAL,
{ .dbl = 0.0 }, 0, UINT_MAX, FLAGS },
{ "num_ticks_poc_diff_one",
"Set VPS and VUI number of ticks per POC increment",
OFFSET(num_ticks_poc_diff_one), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, INT_MAX, FLAGS },
{ "crop_left", "Set left border crop offset",
OFFSET(crop_left), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_WIDTH, FLAGS },
{ "crop_right", "Set right border crop offset",
OFFSET(crop_right), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_WIDTH, FLAGS },
{ "crop_top", "Set top border crop offset",
OFFSET(crop_top), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_HEIGHT, FLAGS },
{ "crop_bottom", "Set bottom border crop offset",
OFFSET(crop_bottom), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_HEIGHT, FLAGS },
{ "width", "Set width after crop",
OFFSET(width), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_WIDTH, FLAGS },
{ "height", "Set height after crop",
OFFSET(height), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_HEIGHT, FLAGS },
{ "level", "Set level (tables A.6 and A.7)",
OFFSET(level), AV_OPT_TYPE_INT,
{ .i64 = LEVEL_UNSET }, LEVEL_UNSET, 0xff, FLAGS, .unit = "level" },
{ "auto", "Attempt to guess level from stream properties",
0, AV_OPT_TYPE_CONST,
{ .i64 = LEVEL_AUTO }, .flags = FLAGS, .unit = "level" },
#define LEVEL(name, value) name, NULL, 0, AV_OPT_TYPE_CONST, \
{ .i64 = value }, .flags = FLAGS, .unit = "level"
{ LEVEL("1", 30) },
{ LEVEL("2", 60) },
{ LEVEL("2.1", 63) },
{ LEVEL("3", 90) },
{ LEVEL("3.1", 93) },
{ LEVEL("4", 120) },
{ LEVEL("4.1", 123) },
{ LEVEL("5", 150) },
{ LEVEL("5.1", 153) },
{ LEVEL("5.2", 156) },
{ LEVEL("6", 180) },
{ LEVEL("6.1", 183) },
{ LEVEL("6.2", 186) },
{ LEVEL("8.5", 255) },
#undef LEVEL
{ NULL }
};
static const AVClass h265_metadata_class = {
.class_name = "h265_metadata_bsf",
.item_name = av_default_item_name,
.option = h265_metadata_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID h265_metadata_codec_ids[] = {
AV_CODEC_ID_HEVC, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_hevc_metadata_bsf = {
.p.name = "hevc_metadata",
.p.codec_ids = h265_metadata_codec_ids,
.p.priv_class = &h265_metadata_class,
.priv_data_size = sizeof(H265MetadataContext),
.init = &h265_metadata_init,
.close = &ff_cbs_bsf_generic_close,
.filter = &ff_cbs_bsf_generic_filter,
};
+150
View File
@@ -0,0 +1,150 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/common.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_h266.h"
#include "vvc.h"
#define IS_H266_SLICE(nut) (nut <= VVC_RASL_NUT || (nut >= VVC_IDR_W_RADL && nut <= VVC_GDR_NUT))
typedef struct H266MetadataContext {
CBSBSFContext common;
H266RawAUD aud_nal;
int aud;
} H266MetadataContext;
static int h266_metadata_update_fragment(AVBSFContext *bsf, AVPacket *pkt,
CodedBitstreamFragment *pu)
{
H266MetadataContext *ctx = bsf->priv_data;
int err, i;
// If an AUD is present, it must be the first NAL unit.
if (pu->nb_units && pu->units[0].type == VVC_AUD_NUT) {
if (ctx->aud == BSF_ELEMENT_REMOVE)
ff_cbs_delete_unit(pu, 0);
} else if ( pkt && ctx->aud == BSF_ELEMENT_INSERT) {
const H266RawSlice *first_slice = NULL;
const H266RawPictureHeader *ph = NULL;
H266RawAUD *aud = &ctx->aud_nal;
int pic_type = 0, temporal_id = 8, layer_id = 0;
for (i = 0; i < pu->nb_units; i++) {
const H266RawNALUnitHeader *nal = pu->units[i].content;
if (!nal)
continue;
if (nal->nuh_temporal_id_plus1 < temporal_id + 1)
temporal_id = nal->nuh_temporal_id_plus1 - 1;
if ( nal->nal_unit_type == VVC_PH_NUT ) {
const H266RawPH *header = pu->units[i].content;
ph = &header->ph_picture_header;
} else if (IS_H266_SLICE(nal->nal_unit_type)) {
const H266RawSlice *slice = pu->units[i].content;
layer_id = nal->nuh_layer_id;
if (slice->header.sh_slice_type == VVC_SLICE_TYPE_B &&
pic_type < 2)
pic_type = 2;
if (slice->header.sh_slice_type == VVC_SLICE_TYPE_P &&
pic_type < 1)
pic_type = 1;
if (!first_slice) {
first_slice = slice;
if (first_slice->header.
sh_picture_header_in_slice_header_flag)
ph = &first_slice->header.sh_picture_header;
else if (!ph)
break;
}
}
}
if (!ph) {
av_log(bsf, AV_LOG_ERROR, "no avaliable picture header");
return AVERROR_INVALIDDATA;
}
aud->nal_unit_header = (H266RawNALUnitHeader) {
.nal_unit_type = VVC_AUD_NUT,
.nuh_layer_id = layer_id,
.nuh_temporal_id_plus1 = temporal_id + 1,
};
aud->aud_pic_type = pic_type;
aud->aud_irap_or_gdr_flag = ph->ph_gdr_or_irap_pic_flag;
err = ff_cbs_insert_unit_content(pu, 0, VVC_AUD_NUT, aud, NULL);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to insert AUD.\n");
return err;
}
}
/* TODO: implement more metadata parsing, like VUI, Levels etc. */
//for (i = 0; i < pu->nb_units; i++) {
// if (pu->units[i].type == VVC_SPS_NUT) {
// }
//}
return 0;
}
static const CBSBSFType h266_metadata_type = {
.codec_id = AV_CODEC_ID_VVC,
.fragment_name = "access unit",
.unit_name = "NAL unit",
.update_fragment = &h266_metadata_update_fragment,
};
static int h266_metadata_init(AVBSFContext *bsf)
{
return ff_cbs_bsf_generic_init(bsf, &h266_metadata_type);
}
#define OFFSET(x) offsetof(H266MetadataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption h266_metadata_options[] = {
BSF_ELEMENT_OPTIONS_PIR("aud", "Access Unit Delimiter NAL units",
aud, FLAGS),
{ NULL }
};
static const AVClass h266_metadata_class = {
.class_name = "h266_metadata_bsf",
.item_name = av_default_item_name,
.option = h266_metadata_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID h266_metadata_codec_ids[] = {
AV_CODEC_ID_VVC, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_vvc_metadata_bsf = {
.p.name = "vvc_metadata",
.p.codec_ids = h266_metadata_codec_ids,
.p.priv_class = &h266_metadata_class,
.priv_data_size = sizeof(H266MetadataContext),
.init = &h266_metadata_init,
.close = &ff_cbs_bsf_generic_close,
.filter = &ff_cbs_bsf_generic_filter,
};
+136
View File
@@ -0,0 +1,136 @@
/*
* HAPQA extract bitstream filter
* Copyright (c) 2017 Jokyo Images
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* HAPQA extract bitstream filter
* extract one of the two textures of the HAQA
*/
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "hap.h"
#include "libavutil/opt.h"
typedef struct HapqaExtractContext {
const AVClass *class;
int texture;/* index of the texture to keep (0 for rgb or 1 for alpha) */
} HapqaExtractContext;
static int check_texture(HapqaExtractContext *ctx, int section_type) {
if (((ctx->texture == 0)&&((section_type & 0x0F) == 0x0F)) || /* HapQ texture and rgb extract */
((ctx->texture == 1)&&((section_type & 0x0F) == 0x01))) /* HapAlphaOnly texture and alpha extract */
{
return 1; /* the texture is the one to keep */
} else {
return 0;
}
}
static int hapqa_extract(AVBSFContext *bsf, AVPacket *pkt)
{
HapqaExtractContext *ctx = bsf->priv_data;
GetByteContext gbc;
int section_size;
enum HapSectionType section_type;
int start_section_size;
int target_packet_size = 0;
int ret = 0;
ret = ff_bsf_get_packet_ref(bsf, pkt);
if (ret < 0)
return ret;
bytestream2_init(&gbc, pkt->data, pkt->size);
ret = ff_hap_parse_section_header(&gbc, &section_size, &section_type);
if (ret != 0)
goto fail;
if ((section_type & 0x0F) != 0x0D) {
av_log(bsf, AV_LOG_ERROR, "Invalid section type for HAPQA %#04x.\n", section_type & 0x0F);
ret = AVERROR_INVALIDDATA;
goto fail;
}
start_section_size = 4;
bytestream2_seek(&gbc, start_section_size, SEEK_SET);/* go to start of the first texture */
ret = ff_hap_parse_section_header(&gbc, &section_size, &section_type);
if (ret != 0)
goto fail;
target_packet_size = section_size + 4;
if (check_texture(ctx, section_type) == 0) { /* the texture is not the one to keep */
start_section_size += 4 + section_size;
bytestream2_seek(&gbc, start_section_size, SEEK_SET);/* go to start of the second texture */
ret = ff_hap_parse_section_header(&gbc, &section_size, &section_type);
if (ret != 0)
goto fail;
target_packet_size = section_size + 4;
if (check_texture(ctx, section_type) == 0){ /* the second texture is not the one to keep */
av_log(bsf, AV_LOG_ERROR, "No valid texture found.\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
}
pkt->data += start_section_size;
pkt->size = target_packet_size;
fail:
if (ret < 0)
av_packet_unref(pkt);
return ret;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_HAP, AV_CODEC_ID_NONE,
};
#define OFFSET(x) offsetof(HapqaExtractContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "texture", "texture to keep", OFFSET(texture), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, FLAGS, .unit = "texture" },
{ "color", "keep HapQ texture", 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, 0, 0, FLAGS, .unit = "texture" },
{ "alpha", "keep HapAlphaOnly texture", 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, 0, 0, FLAGS, .unit = "texture" },
{ NULL },
};
static const AVClass hapqa_extract_class = {
.class_name = "hapqa_extract_bsf",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_hapqa_extract_bsf = {
.p.name = "hapqa_extract",
.p.codec_ids = codec_ids,
.p.priv_class = &hapqa_extract_class,
.priv_data_size = sizeof(HapqaExtractContext),
.filter = hapqa_extract,
};
+208
View File
@@ -0,0 +1,208 @@
/*
* HEVC MP4 to Annex B byte stream format filter
* copyright (c) 2015 Anton Khirnov
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <string.h>
#include "libavutil/intreadwrite.h"
#include "libavutil/mem.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "defs.h"
#include "hevc/hevc.h"
#define MIN_HEVCC_LENGTH 23
typedef struct HEVCBSFContext {
uint8_t length_size;
int extradata_parsed;
} HEVCBSFContext;
static int hevc_extradata_to_annexb(AVBSFContext *ctx)
{
GetByteContext gb;
int length_size, num_arrays, i, j;
int ret = 0;
uint8_t *new_extradata = NULL;
size_t new_extradata_size = 0;
bytestream2_init(&gb, ctx->par_in->extradata, ctx->par_in->extradata_size);
bytestream2_skip(&gb, 21);
length_size = (bytestream2_get_byte(&gb) & 3) + 1;
num_arrays = bytestream2_get_byte(&gb);
for (i = 0; i < num_arrays; i++) {
int type = bytestream2_get_byte(&gb) & 0x3f;
int cnt = bytestream2_get_be16(&gb);
if (!(type == HEVC_NAL_VPS || type == HEVC_NAL_SPS || type == HEVC_NAL_PPS ||
type == HEVC_NAL_SEI_PREFIX || type == HEVC_NAL_SEI_SUFFIX)) {
av_log(ctx, AV_LOG_ERROR, "Invalid NAL unit type in extradata: %d\n",
type);
ret = AVERROR_INVALIDDATA;
goto fail;
}
for (j = 0; j < cnt; j++) {
const int nalu_len = bytestream2_get_be16(&gb);
if (!nalu_len ||
nalu_len > bytestream2_get_bytes_left(&gb) ||
4 + AV_INPUT_BUFFER_PADDING_SIZE + nalu_len > SIZE_MAX - new_extradata_size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
ret = av_reallocp(&new_extradata, new_extradata_size + nalu_len + 4 + AV_INPUT_BUFFER_PADDING_SIZE);
if (ret < 0)
goto fail;
AV_WB32(new_extradata + new_extradata_size, 1); // add the startcode
bytestream2_get_buffer(&gb, new_extradata + new_extradata_size + 4, nalu_len);
new_extradata_size += 4 + nalu_len;
memset(new_extradata + new_extradata_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
}
}
av_freep(&ctx->par_out->extradata);
ctx->par_out->extradata = new_extradata;
ctx->par_out->extradata_size = new_extradata_size;
if (!new_extradata_size)
av_log(ctx, AV_LOG_WARNING, "No parameter sets in the extradata\n");
return length_size;
fail:
av_freep(&new_extradata);
return ret;
}
static int hevc_mp4toannexb_init(AVBSFContext *ctx)
{
HEVCBSFContext *s = ctx->priv_data;
int ret;
if (ctx->par_in->extradata_size < MIN_HEVCC_LENGTH ||
AV_RB24(ctx->par_in->extradata) == 1 ||
AV_RB32(ctx->par_in->extradata) == 1) {
av_log(ctx, AV_LOG_VERBOSE,
"The input looks like it is Annex B already\n");
} else {
ret = hevc_extradata_to_annexb(ctx);
if (ret < 0)
return ret;
s->length_size = ret;
s->extradata_parsed = 1;
}
return 0;
}
static int hevc_mp4toannexb_filter(AVBSFContext *ctx, AVPacket *out)
{
HEVCBSFContext *s = ctx->priv_data;
AVPacket *in;
GetByteContext gb;
int got_irap = 0;
int i, ret = 0;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
if (!s->extradata_parsed) {
av_packet_move_ref(out, in);
av_packet_free(&in);
return 0;
}
bytestream2_init(&gb, in->data, in->size);
while (bytestream2_get_bytes_left(&gb)) {
uint32_t nalu_size = 0;
int nalu_type;
int is_irap, add_extradata, extra_size, prev_size;
if (bytestream2_get_bytes_left(&gb) < s->length_size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
for (i = 0; i < s->length_size; i++)
nalu_size = (nalu_size << 8) | bytestream2_get_byte(&gb);
if (nalu_size < 2 || nalu_size > bytestream2_get_bytes_left(&gb)) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
nalu_type = (bytestream2_peek_byte(&gb) >> 1) & 0x3f;
/* prepend extradata to IRAP frames */
is_irap = nalu_type >= HEVC_NAL_BLA_W_LP &&
nalu_type <= HEVC_NAL_RSV_IRAP_VCL23;
add_extradata = is_irap && !got_irap;
extra_size = add_extradata * ctx->par_out->extradata_size;
got_irap |= is_irap;
if (FFMIN(INT_MAX, SIZE_MAX) < 4ULL + nalu_size + extra_size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
prev_size = out->size;
ret = av_grow_packet(out, 4 + nalu_size + extra_size);
if (ret < 0)
goto fail;
if (extra_size)
memcpy(out->data + prev_size, ctx->par_out->extradata, extra_size);
AV_WB32(out->data + prev_size + extra_size, 1);
bytestream2_get_buffer(&gb, out->data + prev_size + 4 + extra_size, nalu_size);
}
ret = av_packet_copy_props(out, in);
if (ret < 0)
goto fail;
fail:
if (ret < 0)
av_packet_unref(out);
av_packet_free(&in);
return ret;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_HEVC, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_hevc_mp4toannexb_bsf = {
.p.name = "hevc_mp4toannexb",
.p.codec_ids = codec_ids,
.priv_data_size = sizeof(HEVCBSFContext),
.init = hevc_mp4toannexb_init,
.filter = hevc_mp4toannexb_filter,
};
+76
View File
@@ -0,0 +1,76 @@
/*
* imx dump header bitstream filter
* Copyright (c) 2007 Baptiste Coudurier
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* imx dump header bitstream filter
* modifies bitstream to fit in mov and be decoded by final cut pro decoder
*/
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
static int imx_dump_header(AVBSFContext *ctx, AVPacket *out)
{
/* MXF essence element key */
static const uint8_t imx_header[16] = { 0x06,0x0e,0x2b,0x34,0x01,0x02,0x01,0x01,0x0d,0x01,0x03,0x01,0x05,0x01,0x01,0x00 };
AVPacket *in;
int ret = 0;
uint8_t *out_buf;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
ret = av_new_packet(out, in->size + 20);
if (ret < 0)
goto fail;
out_buf = out->data;
bytestream_put_buffer(&out_buf, imx_header, 16);
bytestream_put_byte(&out_buf, 0x83); /* KLV BER long form */
bytestream_put_be24(&out_buf, in->size);
bytestream_put_buffer(&out_buf, in->data, in->size);
ret = av_packet_copy_props(out, in);
if (ret < 0)
goto fail;
fail:
if (ret < 0)
av_packet_unref(out);
av_packet_free(&in);
return ret;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_MPEG2VIDEO, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_imx_dump_header_bsf = {
.p.name = "imxdump",
.p.codec_ids = codec_ids,
.filter = imx_dump_header,
};
+169
View File
@@ -0,0 +1,169 @@
/*
* Media 100 to MJPEGB bitstream filter
* Copyright (c) 2023 Paul B Mahol
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* Media 100 to MJPEGB bitstream filter.
*/
#include "libavutil/intreadwrite.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
static av_cold int init(AVBSFContext *ctx)
{
ctx->par_out->codec_id = AV_CODEC_ID_MJPEGB;
return 0;
}
static int filter(AVBSFContext *ctx, AVPacket *out)
{
unsigned second_field_offset = 0;
unsigned next_field = 0;
unsigned dht_offset[2];
unsigned dqt_offset[2];
unsigned sod_offset[2];
unsigned sof_offset[2];
unsigned sos_offset[2];
unsigned field = 0;
GetByteContext gb;
PutByteContext pb;
AVPacket *in;
int ret;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
ret = av_new_packet(out, in->size + 1024);
if (ret < 0)
goto fail;
bytestream2_init(&gb, in->data, in->size);
bytestream2_init_writer(&pb, out->data, out->size);
second_field:
bytestream2_put_be32(&pb, 0);
bytestream2_put_be32(&pb, AV_RB32("mjpg"));
bytestream2_put_be32(&pb, 0);
bytestream2_put_be32(&pb, 0);
for (int i = 0; i < 6; i++)
bytestream2_put_be32(&pb, 0);
sof_offset[field] = bytestream2_tell_p(&pb);
bytestream2_put_be16(&pb, 17);
bytestream2_put_byte(&pb, 8);
bytestream2_put_be16(&pb, ctx->par_in->height / 2);
bytestream2_put_be16(&pb, ctx->par_in->width);
bytestream2_put_byte(&pb, 3);
bytestream2_put_byte(&pb, 1);
bytestream2_put_byte(&pb, 0x21);
bytestream2_put_byte(&pb, 0);
bytestream2_put_byte(&pb, 2);
bytestream2_put_byte(&pb, 0x11);
bytestream2_put_byte(&pb, 1);
bytestream2_put_byte(&pb, 3);
bytestream2_put_byte(&pb, 0x11);
bytestream2_put_byte(&pb, 1);
sos_offset[field] = bytestream2_tell_p(&pb);
bytestream2_put_be16(&pb, 12);
bytestream2_put_byte(&pb, 3);
bytestream2_put_byte(&pb, 1);
bytestream2_put_byte(&pb, 0);
bytestream2_put_byte(&pb, 2);
bytestream2_put_byte(&pb, 0x11);
bytestream2_put_byte(&pb, 3);
bytestream2_put_byte(&pb, 0x11);
bytestream2_put_byte(&pb, 0);
bytestream2_put_byte(&pb, 0);
bytestream2_put_byte(&pb, 0);
dqt_offset[field] = bytestream2_tell_p(&pb);
bytestream2_put_be16(&pb, 132);
bytestream2_put_byte(&pb, 0);
bytestream2_skip(&gb, 4);
for (int i = 0; i < 64; i++)
bytestream2_put_byte(&pb, bytestream2_get_be32(&gb));
bytestream2_put_byte(&pb, 1);
for (int i = 0; i < 64; i++)
bytestream2_put_byte(&pb, bytestream2_get_be32(&gb));
dht_offset[field] = 0;
sod_offset[field] = bytestream2_tell_p(&pb);
for (int i = bytestream2_tell(&gb) + 8; next_field == 0 && i < in->size - 4; i++) {
if (AV_RB32(in->data + i) == 0x00000001) {
next_field = i;
break;
}
}
bytestream2_skip(&gb, 8);
bytestream2_copy_buffer(&pb, &gb, next_field - bytestream2_tell(&gb));
bytestream2_put_be64(&pb, 0);
if (field == 0) {
field = 1;
second_field_offset = bytestream2_tell_p(&pb);
next_field = in->size;
goto second_field;
}
AV_WB32(out->data + 8, second_field_offset);
AV_WB32(out->data + 12, second_field_offset);
AV_WB32(out->data + 16, second_field_offset);
AV_WB32(out->data + 20, dqt_offset[0]);
AV_WB32(out->data + 24, dht_offset[0]);
AV_WB32(out->data + 28, sof_offset[0]);
AV_WB32(out->data + 32, sos_offset[0]);
AV_WB32(out->data + 36, sod_offset[0]);
AV_WB32(out->data + second_field_offset + 8, bytestream2_tell_p(&pb) - second_field_offset);
AV_WB32(out->data + second_field_offset + 12, bytestream2_tell_p(&pb) - second_field_offset);
AV_WB32(out->data + second_field_offset + 16, 0);
AV_WB32(out->data + second_field_offset + 20, dqt_offset[1] - second_field_offset);
AV_WB32(out->data + second_field_offset + 24, dht_offset[1]);
AV_WB32(out->data + second_field_offset + 28, sof_offset[1] - second_field_offset);
AV_WB32(out->data + second_field_offset + 32, sos_offset[1] - second_field_offset);
AV_WB32(out->data + second_field_offset + 36, sod_offset[1] - second_field_offset);
out->size = bytestream2_tell_p(&pb);
memset(out->data + out->size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
ret = av_packet_copy_props(out, in);
if (ret < 0)
goto fail;
fail:
if (ret < 0)
av_packet_unref(out);
av_packet_free(&in);
return ret;
}
const FFBitStreamFilter ff_media100_to_mjpegb_bsf = {
.p.name = "media100_to_mjpegb",
.p.codec_ids = (const enum AVCodecID []){ AV_CODEC_ID_MEDIA100, AV_CODEC_ID_NONE },
.init = init,
.filter = filter,
};
+142
View File
@@ -0,0 +1,142 @@
/*
* MJPEG/AVI1 to JPEG/JFIF bitstream format filter
* Copyright (c) 2010 Adrian Daerr and Nicolas George
*
* 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
*/
/*
* Adapted from mjpeg2jpeg.c, with original copyright:
* Paris 2010 Adrian Daerr, public domain
*/
#include <string.h>
#include "libavutil/error.h"
#include "libavutil/intreadwrite.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "jpegtables.h"
#include "mjpeg.h"
static const uint8_t jpeg_header[] = {
0xff, 0xd8, // SOI
0xff, 0xe0, // APP0
0x00, 0x10, // APP0 header size (including
// this field, but excluding preceding)
0x4a, 0x46, 0x49, 0x46, 0x00, // ID string 'JFIF\0'
0x01, 0x01, // version
0x00, // bits per type
0x00, 0x00, // X density
0x00, 0x00, // Y density
0x00, // X thumbnail size
0x00, // Y thumbnail size
};
static const int dht_segment_size = 420;
static const uint8_t dht_segment_head[] = { 0xFF, 0xC4, 0x01, 0xA2, 0x00 };
static const uint8_t dht_segment_frag[] = {
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09,
0x0a, 0x0b, 0x01, 0x00, 0x03, 0x01, 0x01, 0x01, 0x01, 0x01,
0x01, 0x01, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00,
};
static uint8_t *append(uint8_t *buf, const uint8_t *src, int size)
{
memcpy(buf, src, size);
return buf + size;
}
static uint8_t *append_dht_segment(uint8_t *buf)
{
buf = append(buf, dht_segment_head, sizeof(dht_segment_head));
buf = append(buf, ff_mjpeg_bits_dc_luminance + 1, 16);
buf = append(buf, dht_segment_frag, sizeof(dht_segment_frag));
buf = append(buf, ff_mjpeg_val_dc, 12);
*(buf++) = 0x10;
buf = append(buf, ff_mjpeg_bits_ac_luminance + 1, 16);
buf = append(buf, ff_mjpeg_val_ac_luminance, 162);
*(buf++) = 0x11;
buf = append(buf, ff_mjpeg_bits_ac_chrominance + 1, 16);
buf = append(buf, ff_mjpeg_val_ac_chrominance, 162);
return buf;
}
static int mjpeg2jpeg_filter(AVBSFContext *ctx, AVPacket *out)
{
AVPacket *in;
int ret = 0;
int input_skip, output_size;
uint8_t *output;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
if (in->size < 12) {
av_log(ctx, AV_LOG_ERROR, "input is truncated\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (AV_RB16(in->data) != 0xffd8) {
av_log(ctx, AV_LOG_ERROR, "input is not MJPEG\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (in->data[2] == 0xff && in->data[3] == APP0) {
input_skip = (in->data[4] << 8) + in->data[5] + 4;
} else {
input_skip = 2;
}
if (in->size < input_skip) {
av_log(ctx, AV_LOG_ERROR, "input is truncated\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
output_size = in->size - input_skip +
sizeof(jpeg_header) + dht_segment_size;
ret = av_new_packet(out, output_size);
if (ret < 0)
goto fail;
output = out->data;
output = append(output, jpeg_header, sizeof(jpeg_header));
output = append_dht_segment(output);
output = append(output, in->data + input_skip, in->size - input_skip);
ret = av_packet_copy_props(out, in);
if (ret < 0)
goto fail;
fail:
if (ret < 0)
av_packet_unref(out);
av_packet_free(&in);
return ret;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_MJPEG, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_mjpeg2jpeg_bsf = {
.p.name = "mjpeg2jpeg",
.p.codec_ids = codec_ids,
.filter = mjpeg2jpeg_filter,
};
+108
View File
@@ -0,0 +1,108 @@
/*
* MJPEG A dump header bitstream filter
* Copyright (c) 2006 Baptiste Coudurier
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* MJPEG A dump header bitstream filter
* modifies bitstream to be decoded by quicktime
*/
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "mjpeg.h"
static int mjpega_dump_header(AVBSFContext *ctx, AVPacket *out)
{
AVPacket *in;
uint8_t *out_buf;
unsigned dqt = 0, dht = 0, sof0 = 0;
int ret = 0, i;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
ret = av_new_packet(out, in->size + 44);
if (ret < 0)
goto fail;
ret = av_packet_copy_props(out, in);
if (ret < 0)
goto fail;
out_buf = out->data;
bytestream_put_byte(&out_buf, 0xff);
bytestream_put_byte(&out_buf, SOI);
bytestream_put_byte(&out_buf, 0xff);
bytestream_put_byte(&out_buf, APP1);
bytestream_put_be16(&out_buf, 42); /* size */
bytestream_put_be32(&out_buf, 0);
bytestream_put_buffer(&out_buf, "mjpg", 4);
bytestream_put_be32(&out_buf, in->size + 44); /* field size */
bytestream_put_be32(&out_buf, in->size + 44); /* pad field size */
bytestream_put_be32(&out_buf, 0); /* next ptr */
for (i = 0; i < in->size - 1; i++) {
if (in->data[i] == 0xff) {
switch (in->data[i + 1]) {
case DQT: dqt = i + 46; break;
case DHT: dht = i + 46; break;
case SOF0: sof0 = i + 46; break;
case SOS:
bytestream_put_be32(&out_buf, dqt); /* quant off */
bytestream_put_be32(&out_buf, dht); /* huff off */
bytestream_put_be32(&out_buf, sof0); /* image off */
bytestream_put_be32(&out_buf, i + 46); /* scan off */
bytestream_put_be32(&out_buf, i + 46 + AV_RB16(in->data + i + 2)); /* data off */
bytestream_put_buffer(&out_buf, in->data + 2, in->size - 2); /* skip already written SOI */
out->size = out_buf - out->data;
av_packet_free(&in);
return 0;
case APP1:
if (i + 8 < in->size && AV_RL32(in->data + i + 8) == AV_RL32("mjpg")) {
av_log(ctx, AV_LOG_ERROR, "bitstream already formatted\n");
av_packet_unref(out);
av_packet_move_ref(out, in);
av_packet_free(&in);
return 0;
}
}
}
}
av_log(ctx, AV_LOG_ERROR, "could not find SOS marker in bitstream\n");
fail:
av_packet_unref(out);
av_packet_free(&in);
return AVERROR_INVALIDDATA;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_MJPEG, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_mjpega_dump_header_bsf = {
.p.name = "mjpegadump",
.p.codec_ids = codec_ids,
.filter = mjpega_dump_header,
};
+87
View File
@@ -0,0 +1,87 @@
/*
* Copyright (c) 2008 Reimar Döffinger
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/common.h"
#include "libavutil/intreadwrite.h"
#include "bsf.h"
#include "bsf_internal.h"
static int text2movsub(AVBSFContext *ctx, AVPacket *out)
{
AVPacket *in;
int ret = 0;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
if (in->size > 0xffff) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
ret = av_new_packet(out, in->size + 2);
if (ret < 0) {
ret = AVERROR(ENOMEM);
goto fail;
}
ret = av_packet_copy_props(out, in);
if (ret < 0)
goto fail;
AV_WB16(out->data, in->size);
memcpy(out->data + 2, in->data, in->size);
fail:
if (ret < 0)
av_packet_unref(out);
av_packet_free(&in);
return ret;
}
const FFBitStreamFilter ff_text2movsub_bsf = {
.p.name = "text2movsub",
.filter = text2movsub,
};
static int mov2textsub(AVBSFContext *ctx, AVPacket *pkt)
{
int ret = 0;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
if (pkt->size < 2) {
av_packet_unref(pkt);
return AVERROR_INVALIDDATA;
}
pkt->size = FFMIN(pkt->size - 2, AV_RB16(pkt->data));
pkt->data += 2;
return 0;
}
const FFBitStreamFilter ff_mov2textsub_bsf = {
.p.name = "mov2textsub",
.filter = mov2textsub,
};
+243
View File
@@ -0,0 +1,243 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/common.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_mpeg2.h"
#include "mpeg12.h"
typedef struct MPEG2MetadataContext {
CBSBSFContext common;
MPEG2RawExtensionData sequence_display_extension;
AVRational display_aspect_ratio;
AVRational frame_rate;
int video_format;
int colour_primaries;
int transfer_characteristics;
int matrix_coefficients;
int mpeg1_warned;
} MPEG2MetadataContext;
static int mpeg2_metadata_update_fragment(AVBSFContext *bsf,
AVPacket *pkt,
CodedBitstreamFragment *frag)
{
MPEG2MetadataContext *ctx = bsf->priv_data;
MPEG2RawSequenceHeader *sh = NULL;
MPEG2RawSequenceExtension *se = NULL;
MPEG2RawSequenceDisplayExtension *sde = NULL;
int i, se_pos;
for (i = 0; i < frag->nb_units; i++) {
if (frag->units[i].type == MPEG2_START_SEQUENCE_HEADER) {
sh = frag->units[i].content;
} else if (frag->units[i].type == MPEG2_START_EXTENSION) {
MPEG2RawExtensionData *ext = frag->units[i].content;
if (ext->extension_start_code_identifier ==
MPEG2_EXTENSION_SEQUENCE) {
se = &ext->data.sequence;
se_pos = i;
} else if (ext->extension_start_code_identifier ==
MPEG2_EXTENSION_SEQUENCE_DISPLAY) {
sde = &ext->data.sequence_display;
}
}
}
if (!sh || !se) {
// No sequence header and sequence extension: not an MPEG-2 video
// sequence.
if (sh && !ctx->mpeg1_warned) {
av_log(bsf, AV_LOG_WARNING, "Stream contains a sequence "
"header but not a sequence extension: maybe it's "
"actually MPEG-1?\n");
ctx->mpeg1_warned = 1;
}
return 0;
}
if (ctx->display_aspect_ratio.num && ctx->display_aspect_ratio.den) {
int num, den;
av_reduce(&num, &den, ctx->display_aspect_ratio.num,
ctx->display_aspect_ratio.den, 65535);
if (num == 4 && den == 3)
sh->aspect_ratio_information = 2;
else if (num == 16 && den == 9)
sh->aspect_ratio_information = 3;
else if (num == 221 && den == 100)
sh->aspect_ratio_information = 4;
else
sh->aspect_ratio_information = 1;
}
if (ctx->frame_rate.num && ctx->frame_rate.den) {
int code, ext_n, ext_d;
ff_mpeg12_find_best_frame_rate(ctx->frame_rate,
&code, &ext_n, &ext_d, 0);
sh->frame_rate_code = code;
se->frame_rate_extension_n = ext_n;
se->frame_rate_extension_d = ext_d;
}
if (ctx->video_format >= 0 ||
ctx->colour_primaries >= 0 ||
ctx->transfer_characteristics >= 0 ||
ctx->matrix_coefficients >= 0) {
if (!sde) {
int err;
ctx->sequence_display_extension.extension_start_code =
MPEG2_START_EXTENSION;
ctx->sequence_display_extension.extension_start_code_identifier =
MPEG2_EXTENSION_SEQUENCE_DISPLAY;
sde = &ctx->sequence_display_extension.data.sequence_display;
*sde = (MPEG2RawSequenceDisplayExtension) {
.video_format = 5,
.colour_description = 0,
.colour_primaries = 2,
.transfer_characteristics = 2,
.matrix_coefficients = 2,
.display_horizontal_size =
se->horizontal_size_extension << 12 | sh->horizontal_size_value,
.display_vertical_size =
se->vertical_size_extension << 12 | sh->vertical_size_value,
};
err = ff_cbs_insert_unit_content(frag, se_pos + 1,
MPEG2_START_EXTENSION,
&ctx->sequence_display_extension,
NULL);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to insert new sequence "
"display extension.\n");
return err;
}
}
if (ctx->video_format >= 0)
sde->video_format = ctx->video_format;
if (ctx->colour_primaries >= 0 ||
ctx->transfer_characteristics >= 0 ||
ctx->matrix_coefficients >= 0) {
sde->colour_description = 1;
if (ctx->colour_primaries >= 0)
sde->colour_primaries = ctx->colour_primaries;
if (ctx->transfer_characteristics >= 0)
sde->transfer_characteristics = ctx->transfer_characteristics;
if (ctx->matrix_coefficients >= 0)
sde->matrix_coefficients = ctx->matrix_coefficients;
}
}
return 0;
}
static const CBSBSFType mpeg2_metadata_type = {
.codec_id = AV_CODEC_ID_MPEG2VIDEO,
.fragment_name = "frame",
.unit_name = "start code",
.update_fragment = &mpeg2_metadata_update_fragment,
};
static int mpeg2_metadata_init(AVBSFContext *bsf)
{
MPEG2MetadataContext *ctx = bsf->priv_data;
#define VALIDITY_CHECK(name) do { \
if (!ctx->name) { \
av_log(bsf, AV_LOG_ERROR, "The value 0 for %s is " \
"forbidden.\n", #name); \
return AVERROR(EINVAL); \
} \
} while (0)
VALIDITY_CHECK(colour_primaries);
VALIDITY_CHECK(transfer_characteristics);
VALIDITY_CHECK(matrix_coefficients);
#undef VALIDITY_CHECK
return ff_cbs_bsf_generic_init(bsf, &mpeg2_metadata_type);
}
#define OFFSET(x) offsetof(MPEG2MetadataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption mpeg2_metadata_options[] = {
{ "display_aspect_ratio", "Set display aspect ratio (table 6-3)",
OFFSET(display_aspect_ratio), AV_OPT_TYPE_RATIONAL,
{ .dbl = 0.0 }, 0, 65535, FLAGS },
{ "frame_rate", "Set frame rate",
OFFSET(frame_rate), AV_OPT_TYPE_RATIONAL,
{ .dbl = 0.0 }, 0, UINT_MAX, FLAGS },
{ "video_format", "Set video format (table 6-6)",
OFFSET(video_format), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 7, FLAGS },
{ "colour_primaries", "Set colour primaries (table 6-7)",
OFFSET(colour_primaries), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "transfer_characteristics", "Set transfer characteristics (table 6-8)",
OFFSET(transfer_characteristics), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ "matrix_coefficients", "Set matrix coefficients (table 6-9)",
OFFSET(matrix_coefficients), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 255, FLAGS },
{ NULL }
};
static const AVClass mpeg2_metadata_class = {
.class_name = "mpeg2_metadata_bsf",
.item_name = av_default_item_name,
.option = mpeg2_metadata_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID mpeg2_metadata_codec_ids[] = {
AV_CODEC_ID_MPEG2VIDEO, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_mpeg2_metadata_bsf = {
.p.name = "mpeg2_metadata",
.p.codec_ids = mpeg2_metadata_codec_ids,
.p.priv_class = &mpeg2_metadata_class,
.priv_data_size = sizeof(MPEG2MetadataContext),
.init = &mpeg2_metadata_init,
.close = &ff_cbs_bsf_generic_close,
.filter = &ff_cbs_bsf_generic_filter,
};
@@ -0,0 +1,166 @@
/*
* Bitstream filter for unpacking DivX-style packed B-frames in MPEG-4 (divx_packed)
* Copyright (c) 2015 Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.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 "bsf.h"
#include "bsf_internal.h"
#include "mpeg4videodefs.h"
#include "startcode.h"
typedef struct UnpackBFramesBSFContext {
AVBufferRef *b_frame_ref;
} UnpackBFramesBSFContext;
/* determine the position of the packed marker in the userdata,
* the number of VOPs and the position of the second VOP */
static void scan_buffer(const uint8_t *buf, int buf_size,
int *pos_p, int *nb_vop, int *pos_vop2) {
uint32_t startcode;
const uint8_t *end = buf + buf_size, *pos = buf;
while (pos < end) {
startcode = -1;
pos = avpriv_find_start_code(pos, end, &startcode);
if (startcode == USER_DATA_STARTCODE && pos_p) {
/* check if the (DivX) userdata string ends with 'p' (packed) */
for (int i = 0; i < 255 && pos + i + 1 < end; i++) {
if (pos[i] == 'p' && pos[i + 1] == '\0') {
*pos_p = pos + i - buf;
break;
}
}
} else if (startcode == VOP_STARTCODE && nb_vop) {
*nb_vop += 1;
if (*nb_vop == 2 && pos_vop2) {
*pos_vop2 = pos - buf - 4; /* subtract 4 bytes startcode */
}
}
}
}
static int mpeg4_unpack_bframes_filter(AVBSFContext *ctx, AVPacket *pkt)
{
UnpackBFramesBSFContext *s = ctx->priv_data;
int pos_p = -1, nb_vop = 0, pos_vop2 = -1, ret = 0;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
scan_buffer(pkt->data, pkt->size, &pos_p, &nb_vop, &pos_vop2);
av_log(ctx, AV_LOG_DEBUG, "Found %d VOP startcode(s) in this packet.\n", nb_vop);
if (pos_vop2 >= 0) {
if (s->b_frame_ref) {
av_log(ctx, AV_LOG_WARNING,
"Missing one N-VOP packet, discarding one B-frame.\n");
av_buffer_unref(&s->b_frame_ref);
}
/* store a reference to the packed B-frame's data in the BSFContext */
s->b_frame_ref = av_buffer_ref(pkt->buf);
if (!s->b_frame_ref) {
ret = AVERROR(ENOMEM);
goto fail;
}
s->b_frame_ref->data = pkt->data + pos_vop2;
s->b_frame_ref->size = pkt->size - pos_vop2;
}
if (nb_vop > 2) {
av_log(ctx, AV_LOG_WARNING,
"Found %d VOP headers in one packet, only unpacking one.\n", nb_vop);
}
if (nb_vop == 1 && s->b_frame_ref) {
AVBufferRef *tmp = pkt->buf;
/* make tmp accurately reflect the packet's data */
tmp->data = pkt->data;
tmp->size = pkt->size;
/* replace data in packet with stored data */
pkt->buf = s->b_frame_ref;
pkt->data = s->b_frame_ref->data;
pkt->size = s->b_frame_ref->size;
/* store reference to data into BSFContext */
s->b_frame_ref = tmp;
if (s->b_frame_ref->size <= MAX_NVOP_SIZE) {
/* N-VOP - discard stored data */
av_log(ctx, AV_LOG_DEBUG, "Skipping N-VOP.\n");
av_buffer_unref(&s->b_frame_ref);
}
} else if (nb_vop >= 2) {
/* use first frame of the packet */
pkt->size = pos_vop2;
} else if (pos_p >= 0) {
ret = av_packet_make_writable(pkt);
if (ret < 0)
goto fail;
av_log(ctx, AV_LOG_DEBUG, "Updating DivX userdata (remove trailing 'p').\n");
/* remove 'p' (packed) from the end of the (DivX) userdata string */
pkt->data[pos_p] = '\0';
} else {
/* use packet as is */
}
fail:
if (ret < 0)
av_packet_unref(pkt);
return ret;
}
static int mpeg4_unpack_bframes_init(AVBSFContext *ctx)
{
if (ctx->par_in->extradata) {
int pos_p_ext = -1;
scan_buffer(ctx->par_in->extradata, ctx->par_in->extradata_size, &pos_p_ext, NULL, NULL);
if (pos_p_ext >= 0) {
av_log(ctx, AV_LOG_DEBUG,
"Updating DivX userdata (remove trailing 'p') in extradata.\n");
ctx->par_out->extradata[pos_p_ext] = '\0';
}
}
return 0;
}
static void mpeg4_unpack_bframes_close_flush(AVBSFContext *bsfc)
{
UnpackBFramesBSFContext *ctx = bsfc->priv_data;
av_buffer_unref(&ctx->b_frame_ref);
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_MPEG4, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_mpeg4_unpack_bframes_bsf = {
.p.name = "mpeg4_unpack_bframes",
.p.codec_ids = codec_ids,
.priv_data_size = sizeof(UnpackBFramesBSFContext),
.init = mpeg4_unpack_bframes_init,
.filter = mpeg4_unpack_bframes_filter,
.flush = mpeg4_unpack_bframes_close_flush,
.close = mpeg4_unpack_bframes_close_flush,
};
+236
View File
@@ -0,0 +1,236 @@
/*
* copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
*
* 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 <stdlib.h>
#include "bsf.h"
#include "bsf_internal.h"
#include "libavutil/log.h"
#include "libavutil/mem.h"
#include "libavutil/opt.h"
#include "libavutil/eval.h"
static const char *const var_names[] = {
"n", ///< packet index, starting from zero
"tb", ///< timebase
"pts", ///< packet presentation timestamp
"dts", ///< packet decoding timestamp
"nopts", ///< AV_NOPTS_VALUE
"startpts", ///< first seen non-AV_NOPTS_VALUE packet timestamp
"startdts", ///< first seen non-AV_NOPTS_VALUE packet timestamp
"duration", "d", ///< packet duration
"pos", ///< original position of packet in its source
"size", ///< packet size
"key" , ///< packet keyframe flag
"state", ///< random-ish state
NULL
};
enum var_name {
VAR_N,
VAR_TB,
VAR_PTS,
VAR_DTS,
VAR_NOPTS,
VAR_STARTPTS,
VAR_STARTDTS,
VAR_DURATION, VAR_D,
VAR_POS,
VAR_SIZE,
VAR_KEY,
VAR_STATE,
VAR_VARS_NB
};
typedef struct NoiseContext {
const AVClass *class;
char *amount_str;
char *drop_str;
int dropamount;
AVExpr *amount_pexpr;
AVExpr *drop_pexpr;
double var_values[VAR_VARS_NB];
unsigned int state;
unsigned int pkt_idx;
} NoiseContext;
static int noise_init(AVBSFContext *ctx)
{
NoiseContext *s = ctx->priv_data;
int ret;
if (!s->amount_str) {
s->amount_str = (!s->drop_str && !s->dropamount) ? av_strdup("-1") : av_strdup("0");
if (!s->amount_str)
return AVERROR(ENOMEM);
}
if (ctx->par_in->codec_id == AV_CODEC_ID_WRAPPED_AVFRAME &&
strcmp(s->amount_str, "0")) {
av_log(ctx, AV_LOG_ERROR, "Wrapped AVFrame noising is unsupported\n");
return AVERROR_PATCHWELCOME;
}
ret = av_expr_parse(&s->amount_pexpr, s->amount_str,
var_names, NULL, NULL, NULL, NULL, 0, ctx);
if (ret < 0) {
av_log(ctx, AV_LOG_ERROR, "Error in parsing expr for amount: %s\n", s->amount_str);
return ret;
}
if (s->drop_str && s->dropamount) {
av_log(ctx, AV_LOG_WARNING, "Both drop '%s' and dropamount=%d set. Ignoring dropamount.\n",
s->drop_str, s->dropamount);
s->dropamount = 0;
}
if (s->drop_str) {
ret = av_expr_parse(&s->drop_pexpr, s->drop_str,
var_names, NULL, NULL, NULL, NULL, 0, ctx);
if (ret < 0) {
av_log(ctx, AV_LOG_ERROR, "Error in parsing expr for drop: %s\n", s->drop_str);
return ret;
}
}
s->var_values[VAR_TB] = ctx->time_base_out.den ? av_q2d(ctx->time_base_out) : 0;
s->var_values[VAR_NOPTS] = AV_NOPTS_VALUE;
s->var_values[VAR_STARTPTS] = AV_NOPTS_VALUE;
s->var_values[VAR_STARTDTS] = AV_NOPTS_VALUE;
s->var_values[VAR_STATE] = 0;
return 0;
}
static int noise(AVBSFContext *ctx, AVPacket *pkt)
{
NoiseContext *s = ctx->priv_data;
int i, ret, amount, drop = 0;
double res;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
s->var_values[VAR_N] = s->pkt_idx++;
s->var_values[VAR_PTS] = pkt->pts;
s->var_values[VAR_DTS] = pkt->dts;
s->var_values[VAR_DURATION] =
s->var_values[VAR_D] = pkt->duration;
s->var_values[VAR_SIZE] = pkt->size;
s->var_values[VAR_KEY] = !!(pkt->flags & AV_PKT_FLAG_KEY);
s->var_values[VAR_POS] = pkt->pos;
if (s->var_values[VAR_STARTPTS] == AV_NOPTS_VALUE)
s->var_values[VAR_STARTPTS] = pkt->pts;
if (s->var_values[VAR_STARTDTS] == AV_NOPTS_VALUE)
s->var_values[VAR_STARTDTS] = pkt->dts;
res = av_expr_eval(s->amount_pexpr, s->var_values, NULL);
if (isnan(res))
amount = 0;
else if (res < 0)
amount = (s->state % 10001 + 1);
else
amount = (int)res;
if (s->drop_str) {
res = av_expr_eval(s->drop_pexpr, s->var_values, NULL);
if (isnan(res))
drop = 0;
else if (res < 0)
drop = !(s->state % FFABS((int)res));
else
drop = !!res;
}
if(s->dropamount) {
drop = !(s->state % s->dropamount);
}
av_log(ctx, AV_LOG_VERBOSE, "Stream #%d packet %d pts %"PRId64" - amount %d drop %d\n",
pkt->stream_index, (unsigned int)s->var_values[VAR_N], pkt->pts, amount, drop);
if (drop) {
s->var_values[VAR_STATE] = ++s->state;
av_packet_unref(pkt);
return AVERROR(EAGAIN);
}
if (amount) {
ret = av_packet_make_writable(pkt);
if (ret < 0) {
av_packet_unref(pkt);
return ret;
}
}
for (i = 0; i < pkt->size; i++) {
s->state += pkt->data[i] + 1;
if (amount && s->state % amount == 0)
pkt->data[i] = s->state;
}
s->var_values[VAR_STATE] = s->state;
return 0;
}
static void noise_close(AVBSFContext *bsf)
{
NoiseContext *s = bsf->priv_data;
av_expr_free(s->amount_pexpr);
av_expr_free(s->drop_pexpr);
s->amount_pexpr = s->drop_pexpr = NULL;
}
#define OFFSET(x) offsetof(NoiseContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "amount", NULL, OFFSET(amount_str), AV_OPT_TYPE_STRING, { .str = NULL }, 0, 0, FLAGS },
{ "drop", NULL, OFFSET(drop_str), AV_OPT_TYPE_STRING, { .str = NULL }, 0, 0, FLAGS },
{ "dropamount", NULL, OFFSET(dropamount), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, FLAGS },
{ NULL },
};
static const AVClass noise_class = {
.class_name = "noise",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_noise_bsf = {
.p.name = "noise",
.p.priv_class = &noise_class,
.priv_data_size = sizeof(NoiseContext),
.init = noise_init,
.close = noise_close,
.filter = noise,
};
+29
View File
@@ -0,0 +1,29 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* Null bitstream filter -- pass the input through unchanged.
*/
#include "bsf_internal.h"
const FFBitStreamFilter ff_null_bsf = {
.p.name = "null",
.filter = ff_bsf_get_packet_ref,
};
+68
View File
@@ -0,0 +1,68 @@
/*
* 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 "bsf.h"
#include "bsf_internal.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/opt.h"
typedef struct OpusBSFContext {
const AVClass *class;
int gain;
} OpusBSFContext;
static int opus_metadata_init(AVBSFContext *bsfc)
{
OpusBSFContext *s = bsfc->priv_data;
if (bsfc->par_out->extradata_size < 19)
return AVERROR_INVALIDDATA;
AV_WL16(bsfc->par_out->extradata + 16, s->gain);
return 0;
}
#define OFFSET(x) offsetof(OpusBSFContext, x)
#define FLAGS (AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption opus_metadata_options[] = {
{ "gain", "Gain, actual amplification is pow(10, gain/(20.0*256))", OFFSET(gain),
AV_OPT_TYPE_INT, { .i64 = 0 }, -(INT16_MAX + 1), INT16_MAX, .flags = FLAGS },
{ NULL },
};
static const AVClass opus_metadata_class = {
.class_name = "opus_metadata_bsf",
.item_name = av_default_item_name,
.option = opus_metadata_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_OPUS, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_opus_metadata_bsf = {
.p.name = "opus_metadata",
.p.codec_ids = codec_ids,
.p.priv_class = &opus_metadata_class,
.priv_data_size = sizeof(OpusBSFContext),
.init = &opus_metadata_init,
.filter = &ff_bsf_get_packet_ref,
};
+241
View File
@@ -0,0 +1,241 @@
/*
* Copyright (c) 2020 Marton Balint
*
* 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 "bsf.h"
#include "bsf_internal.h"
#include "libavutil/avassert.h"
#include "libavutil/opt.h"
typedef struct PCMContext {
const AVClass *class;
int nb_out_samples;
int pad;
AVRational frame_rate;
AVPacket *in_pkt;
AVPacket *out_pkt;
int sample_size;
int64_t n;
} PCMContext;
static int init(AVBSFContext *ctx)
{
PCMContext *s = ctx->priv_data;
AVRational sr = av_make_q(ctx->par_in->sample_rate, 1);
int64_t min_samples;
if (ctx->par_in->ch_layout.nb_channels <= 0 || ctx->par_in->sample_rate <= 0)
return AVERROR(EINVAL);
ctx->time_base_out = av_inv_q(sr);
s->sample_size = ctx->par_in->ch_layout.nb_channels *
av_get_bits_per_sample(ctx->par_in->codec_id) / 8;
if (s->frame_rate.num) {
min_samples = av_rescale_q_rnd(1, sr, s->frame_rate, AV_ROUND_DOWN);
} else {
min_samples = s->nb_out_samples;
}
if (min_samples <= 0 || min_samples > INT_MAX / s->sample_size - 1)
return AVERROR(EINVAL);
s->in_pkt = av_packet_alloc();
s->out_pkt = av_packet_alloc();
if (!s->in_pkt || !s->out_pkt)
return AVERROR(ENOMEM);
return 0;
}
static void uninit(AVBSFContext *ctx)
{
PCMContext *s = ctx->priv_data;
av_packet_free(&s->in_pkt);
av_packet_free(&s->out_pkt);
}
static void flush(AVBSFContext *ctx)
{
PCMContext *s = ctx->priv_data;
av_packet_unref(s->in_pkt);
av_packet_unref(s->out_pkt);
s->n = 0;
}
static int send_packet(PCMContext *s, int nb_samples, AVPacket *pkt)
{
pkt->duration = nb_samples;
s->n++;
return 0;
}
static void drain_packet(AVPacket *pkt, int drain_data, int drain_samples)
{
pkt->size -= drain_data;
pkt->data += drain_data;
if (pkt->dts != AV_NOPTS_VALUE)
pkt->dts += drain_samples;
if (pkt->pts != AV_NOPTS_VALUE)
pkt->pts += drain_samples;
}
static int get_next_nb_samples(AVBSFContext *ctx)
{
PCMContext *s = ctx->priv_data;
if (s->frame_rate.num) {
AVRational sr = av_make_q(ctx->par_in->sample_rate, 1);
return av_rescale_q(s->n + 1, sr, s->frame_rate) - av_rescale_q(s->n, sr, s->frame_rate);
} else {
return s->nb_out_samples;
}
}
static void set_silence(AVCodecParameters *par, uint8_t *buf, size_t size)
{
int c = 0;
switch (par->codec_id) {
case AV_CODEC_ID_PCM_ALAW: c = 0xd5; break;
case AV_CODEC_ID_PCM_MULAW:
case AV_CODEC_ID_PCM_VIDC: c = 0xff; break;
case AV_CODEC_ID_PCM_U8: c = 0x80; break;
}
memset(buf, c, size);
}
static int rechunk_filter(AVBSFContext *ctx, AVPacket *pkt)
{
PCMContext *s = ctx->priv_data;
int nb_samples = get_next_nb_samples(ctx);
int data_size = nb_samples * s->sample_size;
int ret;
do {
if (s->in_pkt->size) {
if (s->out_pkt->size || s->in_pkt->size < data_size) {
int drain = FFMIN(s->in_pkt->size, data_size - s->out_pkt->size);
if (!s->out_pkt->size) {
ret = av_new_packet(s->out_pkt, data_size);
if (ret < 0)
return ret;
ret = av_packet_copy_props(s->out_pkt, s->in_pkt);
if (ret < 0) {
av_packet_unref(s->out_pkt);
return ret;
}
s->out_pkt->size = 0;
}
memcpy(s->out_pkt->data + s->out_pkt->size, s->in_pkt->data, drain);
s->out_pkt->size += drain;
drain_packet(s->in_pkt, drain, drain / s->sample_size);
if (!s->in_pkt->size)
av_packet_unref(s->in_pkt);
if (s->out_pkt->size == data_size) {
av_packet_move_ref(pkt, s->out_pkt);
return send_packet(s, nb_samples, pkt);
}
av_assert0(!s->in_pkt->size);
} else if (s->in_pkt->size > data_size) {
ret = av_packet_ref(pkt, s->in_pkt);
if (ret < 0)
return ret;
pkt->size = data_size;
drain_packet(s->in_pkt, data_size, nb_samples);
return send_packet(s, nb_samples, pkt);
} else {
av_assert0(s->in_pkt->size == data_size);
av_packet_move_ref(pkt, s->in_pkt);
return send_packet(s, nb_samples, pkt);
}
} else
av_packet_unref(s->in_pkt);
ret = ff_bsf_get_packet_ref(ctx, s->in_pkt);
if (ret == AVERROR_EOF && s->out_pkt->size) {
if (s->pad) {
set_silence(ctx->par_in, s->out_pkt->data + s->out_pkt->size, data_size - s->out_pkt->size);
s->out_pkt->size = data_size;
} else {
nb_samples = s->out_pkt->size / s->sample_size;
}
av_packet_move_ref(pkt, s->out_pkt);
return send_packet(s, nb_samples, pkt);
}
if (ret >= 0)
av_packet_rescale_ts(s->in_pkt, ctx->time_base_in, ctx->time_base_out);
} while (ret >= 0);
return ret;
}
#define OFFSET(x) offsetof(PCMContext, x)
#define FLAGS (AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "nb_out_samples", "set the number of per-packet output samples", OFFSET(nb_out_samples), AV_OPT_TYPE_INT, {.i64=1024}, 1, INT_MAX, FLAGS },
{ "n", "set the number of per-packet output samples", OFFSET(nb_out_samples), AV_OPT_TYPE_INT, {.i64=1024}, 1, INT_MAX, FLAGS },
{ "pad", "pad last packet with zeros", OFFSET(pad), AV_OPT_TYPE_BOOL, {.i64=1} , 0, 1, FLAGS },
{ "p", "pad last packet with zeros", OFFSET(pad), AV_OPT_TYPE_BOOL, {.i64=1} , 0, 1, FLAGS },
{ "frame_rate", "set number of packets per second", OFFSET(frame_rate), AV_OPT_TYPE_RATIONAL, {.dbl=0}, 0, INT_MAX, FLAGS },
{ "r", "set number of packets per second", OFFSET(frame_rate), AV_OPT_TYPE_RATIONAL, {.dbl=0}, 0, INT_MAX, FLAGS },
{ NULL },
};
static const AVClass pcm_rechunk_class = {
.class_name = "pcm_rechunk_bsf",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_PCM_ALAW,
AV_CODEC_ID_PCM_F16LE,
AV_CODEC_ID_PCM_F24LE,
AV_CODEC_ID_PCM_F32BE,
AV_CODEC_ID_PCM_F32LE,
AV_CODEC_ID_PCM_F64BE,
AV_CODEC_ID_PCM_F64LE,
AV_CODEC_ID_PCM_MULAW,
AV_CODEC_ID_PCM_S16BE,
AV_CODEC_ID_PCM_S16LE,
AV_CODEC_ID_PCM_S24BE,
AV_CODEC_ID_PCM_S24DAUD,
AV_CODEC_ID_PCM_S24LE,
AV_CODEC_ID_PCM_S32BE,
AV_CODEC_ID_PCM_S32LE,
AV_CODEC_ID_PCM_S64BE,
AV_CODEC_ID_PCM_S64LE,
AV_CODEC_ID_PCM_S8,
AV_CODEC_ID_PCM_SGA,
AV_CODEC_ID_PCM_U8,
AV_CODEC_ID_PCM_VIDC,
AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_pcm_rechunk_bsf = {
.p.name = "pcm_rechunk",
.p.codec_ids = codec_ids,
.p.priv_class = &pcm_rechunk_class,
.priv_data_size = sizeof(PCMContext),
.filter = rechunk_filter,
.init = init,
.flush = flush,
.close = uninit,
};
+191
View File
@@ -0,0 +1,191 @@
/*
* Copyright (c) 2020 John Stebbins <jstebbins.hb@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
*/
/**
* @file
* This bitstream filter merges PGS subtitle packets containing incomplete
* set of segments into a single packet
*
* Packets already containing a complete set of segments will be passed through
* unchanged.
*/
#include "libavutil/attributes.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/log.h"
#include "bsf.h"
#include "bsf_internal.h"
enum PGSSegmentType {
PALETTE_SEGMENT = 0x14,
OBJECT_SEGMENT = 0x15,
PRESENTATION_SEGMENT = 0x16,
WINDOW_SEGMENT = 0x17,
END_DISPLAY_SET_SEGMENT = 0x80,
};
typedef struct PGSMergeContext {
AVPacket *buffer_pkt, *in;
int presentation_found;
int pkt_flags;
} PGSMergeContext;
static av_cold void frame_merge_flush(AVBSFContext *bsf)
{
PGSMergeContext *ctx = bsf->priv_data;
ctx->presentation_found = ctx->pkt_flags = 0;
av_packet_unref(ctx->in);
av_packet_unref(ctx->buffer_pkt);
}
static int frame_merge_output(PGSMergeContext *ctx, AVPacket *dst, AVPacket *src)
{
if (!ctx->presentation_found)
ctx->pkt_flags |= AV_PKT_FLAG_CORRUPT;
ctx->presentation_found = 0;
src->flags |= ctx->pkt_flags;
ctx->pkt_flags = 0;
av_packet_move_ref(dst, src);
return 0;
}
static int frame_merge_filter(AVBSFContext *bsf, AVPacket *out)
{
PGSMergeContext *ctx = bsf->priv_data;
AVPacket *in = ctx->in, *pkt = ctx->buffer_pkt;
int ret, size, pos, display = 0, presentation = 0;
unsigned int i;
if (!in->data) {
ret = ff_bsf_get_packet_ref(bsf, in);
if (ret == AVERROR_EOF && pkt->data) {
// Output remaining data
ctx->pkt_flags |= AV_PKT_FLAG_CORRUPT;
return frame_merge_output(ctx, out, pkt);
}
if (ret < 0)
return ret;
}
if (!in->size) {
av_packet_unref(in);
return AVERROR(EAGAIN);
}
in->flags &= ~AV_PKT_FLAG_KEY; // Will be detected in the stream
// Validate packet data and find display_end segment
size = in->size;
i = 0;
while (i + 3 <= in->size) {
uint8_t segment_type = in->data[i];
int segment_len = AV_RB16(in->data + i + 1) + 3;
if (i + segment_len > in->size)
break; // Invalid, segments can't span packets
if (segment_type == PRESENTATION_SEGMENT && ctx->presentation_found)
break; // Invalid, there can be only one
if (segment_type == PRESENTATION_SEGMENT) {
uint8_t state;
if (segment_len < 11)
break; // Invalid presentation segment length
ctx->presentation_found = presentation = 1;
state = in->data[i + 10] & 0xc0;
if (state)
ctx->pkt_flags |= AV_PKT_FLAG_KEY;
else
ctx->pkt_flags &= ~AV_PKT_FLAG_KEY;
}
i += segment_len;
if (segment_type == END_DISPLAY_SET_SEGMENT) {
size = i;
display = 1;
break;
}
}
if (display && pkt->size == 0 && size == in->size) // passthrough
return frame_merge_output(ctx, out, in);
if (!display && i != in->size) {
av_log(bsf, AV_LOG_WARNING, "Failed to parse PGS segments.\n");
// force output what we have
size = in->size;
display = 1;
ctx->pkt_flags |= AV_PKT_FLAG_CORRUPT;
}
if (presentation) {
ret = av_packet_copy_props(pkt, in);
if (ret < 0)
goto fail;
}
pos = pkt->size;
ret = av_grow_packet(pkt, size);
if (ret < 0)
goto fail;
memcpy(pkt->data + pos, in->data, size);
if (size == in->size)
av_packet_unref(in);
else {
in->data += size;
in->size -= size;
}
if (display)
return frame_merge_output(ctx, out, pkt);
return AVERROR(EAGAIN);
fail:
frame_merge_flush(bsf);
return ret;
}
static av_cold int frame_merge_init(AVBSFContext *bsf)
{
PGSMergeContext *ctx = bsf->priv_data;
ctx->in = av_packet_alloc();
ctx->buffer_pkt = av_packet_alloc();
if (!ctx->in || !ctx->buffer_pkt)
return AVERROR(ENOMEM);
return 0;
}
static av_cold void frame_merge_close(AVBSFContext *bsf)
{
PGSMergeContext *ctx = bsf->priv_data;
av_packet_free(&ctx->in);
av_packet_free(&ctx->buffer_pkt);
}
static const enum AVCodecID frame_merge_codec_ids[] = {
AV_CODEC_ID_HDMV_PGS_SUBTITLE, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_pgs_frame_merge_bsf = {
.p.name = "pgs_frame_merge",
.p.codec_ids = frame_merge_codec_ids,
.priv_data_size = sizeof(PGSMergeContext),
.init = frame_merge_init,
.flush = frame_merge_flush,
.close = frame_merge_close,
.filter = frame_merge_filter,
};
+176
View File
@@ -0,0 +1,176 @@
/*
* Prores Metadata bitstream filter
* Copyright (c) 2018 Jokyo Images
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* Prores Metadata bitstream filter
* set frame colorspace property
*/
#include "libavutil/common.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
typedef struct ProresMetadataContext {
const AVClass *class;
int color_primaries;
int transfer_characteristics;
int matrix_coefficients;
} ProresMetadataContext;
static int prores_metadata(AVBSFContext *bsf, AVPacket *pkt)
{
ProresMetadataContext *ctx = bsf->priv_data;
int ret = 0;
int buf_size;
uint8_t *buf;
ret = ff_bsf_get_packet_ref(bsf, pkt);
if (ret < 0)
return ret;
ret = av_packet_make_writable(pkt);
if (ret < 0)
goto fail;
buf = pkt->data;
buf_size = pkt->size;
/* check start of the prores frame */
if (buf_size < 28) {
av_log(bsf, AV_LOG_ERROR, "not enough data in prores frame\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (AV_RL32(buf + 4) != AV_RL32("icpf")) {
av_log(bsf, AV_LOG_ERROR, "invalid frame header\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
if (AV_RB16(buf + 8) < 28) {
av_log(bsf, AV_LOG_ERROR, "invalid frame header size\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
/* set the new values */
if (ctx->color_primaries != -1)
buf[8+14] = ctx->color_primaries;
if (ctx->transfer_characteristics != -1)
buf[8+15] = ctx->transfer_characteristics;
if (ctx->matrix_coefficients != -1)
buf[8+16] = ctx->matrix_coefficients;
fail:
if (ret < 0)
av_packet_unref(pkt);
return ret;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_PRORES, AV_CODEC_ID_NONE,
};
static int prores_metadata_init(AVBSFContext *bsf)
{
ProresMetadataContext *ctx = bsf->priv_data;
/*! check options */
switch (ctx->color_primaries) {
case -1:
case 0:
case AVCOL_PRI_BT709:
case AVCOL_PRI_BT470BG:
case AVCOL_PRI_SMPTE170M:
case AVCOL_PRI_BT2020:
case AVCOL_PRI_SMPTE431:
case AVCOL_PRI_SMPTE432:
break;
default:
av_log(bsf, AV_LOG_ERROR, "Color primaries %d is not a valid value\n", ctx->color_primaries);
return AVERROR(EINVAL);
}
switch (ctx->matrix_coefficients) {
case -1:
case 0:
case AVCOL_SPC_BT709:
case AVCOL_SPC_SMPTE170M:
case AVCOL_SPC_BT2020_NCL:
break;
default:
av_log(bsf, AV_LOG_ERROR, "Colorspace %d is not a valid value\n", ctx->matrix_coefficients);
return AVERROR(EINVAL);
}
return 0;
}
#define OFFSET(x) offsetof(ProresMetadataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{"color_primaries", "select color primaries", OFFSET(color_primaries), AV_OPT_TYPE_INT, {.i64=-1}, -1, AVCOL_PRI_SMPTE432, FLAGS, .unit = "color_primaries"},
{"auto", "keep the same color primaries", 0, AV_OPT_TYPE_CONST, {.i64=-1}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"unknown", NULL, 0, AV_OPT_TYPE_CONST, {.i64=0}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"bt709", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_PRI_BT709}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"bt470bg", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_PRI_BT470BG}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"smpte170m", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_PRI_SMPTE170M}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"bt2020", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_PRI_BT2020}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"smpte431", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_PRI_SMPTE431}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"smpte432", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_PRI_SMPTE432}, INT_MIN, INT_MAX, FLAGS, .unit = "color_primaries"},
{"color_trc", "select color transfer", OFFSET(transfer_characteristics), AV_OPT_TYPE_INT, {.i64=-1}, -1, AVCOL_TRC_NB - 1, FLAGS, .unit = "color_trc"},
{"auto", "keep the same color transfer", 0, AV_OPT_TYPE_CONST, {.i64=-1}, INT_MIN, INT_MAX, FLAGS, .unit = "color_trc"},
{"unknown", NULL, 0, AV_OPT_TYPE_CONST, {.i64=0}, INT_MIN, INT_MAX, FLAGS, .unit = "color_trc"},
{"bt709", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_TRC_BT709}, INT_MIN, INT_MAX, FLAGS, .unit = "color_trc"},
{"smpte2084", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_TRC_SMPTE2084}, INT_MIN, INT_MAX, FLAGS, .unit = "color_trc"},
{"arib-std-b67", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_TRC_ARIB_STD_B67}, INT_MIN, INT_MAX, FLAGS, .unit = "color_trc"},
{"colorspace", "select colorspace", OFFSET(matrix_coefficients), AV_OPT_TYPE_INT, {.i64=-1}, -1, AVCOL_SPC_BT2020_NCL, FLAGS, .unit = "colorspace"},
{"auto", "keep the same colorspace", 0, AV_OPT_TYPE_CONST, {.i64=-1}, INT_MIN, INT_MAX, FLAGS, .unit = "colorspace"},
{"unknown", NULL, 0, AV_OPT_TYPE_CONST, {.i64=0}, INT_MIN, INT_MAX, FLAGS, .unit = "colorspace"},
{"bt709", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_SPC_BT709}, INT_MIN, INT_MAX, FLAGS, .unit = "colorspace"},
{"smpte170m", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_SPC_SMPTE170M}, INT_MIN, INT_MAX, FLAGS, .unit = "colorspace"},
{"bt2020nc", NULL, 0, AV_OPT_TYPE_CONST, {.i64=AVCOL_SPC_BT2020_NCL}, INT_MIN, INT_MAX, FLAGS, .unit = "colorspace"},
{ NULL },
};
static const AVClass prores_metadata_class = {
.class_name = "prores_metadata_bsf",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_prores_metadata_bsf = {
.p.name = "prores_metadata",
.p.codec_ids = codec_ids,
.p.priv_class = &prores_metadata_class,
.priv_data_size = sizeof(ProresMetadataContext),
.init = prores_metadata_init,
.filter = prores_metadata,
};
+254
View File
@@ -0,0 +1,254 @@
/*
* copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
*
* 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/log.h"
#include "libavutil/opt.h"
#include "av1_parse.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "h264.h"
#include "startcode.h"
#include "vc1_common.h"
#include "hevc/hevc.h"
enum RemoveFreq {
REMOVE_FREQ_KEYFRAME,
REMOVE_FREQ_ALL,
REMOVE_FREQ_NONKEYFRAME,
};
#define START_CODE 0x000001
typedef struct RemoveExtradataContext {
const AVClass *class;
int freq;
} RemoveExtradataContext;
static int av1_split(const uint8_t *buf, int buf_size, void *logctx)
{
AV1OBU obu;
const uint8_t *ptr = buf, *end = buf + buf_size;
while (ptr < end) {
int len = ff_av1_extract_obu(&obu, ptr, buf_size, logctx);
if (len < 0)
break;
if (obu.type == AV1_OBU_FRAME_HEADER ||
obu.type == AV1_OBU_FRAME) {
return ptr - buf;
}
ptr += len;
buf_size -= len;
}
return 0;
}
static int h264_split(const uint8_t *buf, int buf_size)
{
const uint8_t *ptr = buf, *end = buf + buf_size;
uint32_t state = -1;
int has_sps = 0;
int has_pps = 0;
int nalu_type;
while (ptr < end) {
ptr = avpriv_find_start_code(ptr, end, &state);
if ((state & 0xFFFFFF00) != 0x100)
break;
nalu_type = state & 0x1F;
if (nalu_type == H264_NAL_SPS) {
has_sps = 1;
} else if (nalu_type == H264_NAL_PPS)
has_pps = 1;
/* else if (nalu_type == 0x01 ||
* nalu_type == 0x02 ||
* nalu_type == 0x05) {
* }
*/
else if ((nalu_type != H264_NAL_SEI || has_pps) &&
nalu_type != H264_NAL_AUD && nalu_type != H264_NAL_SPS_EXT &&
nalu_type != 0x0f) {
if (has_sps) {
while (ptr - 4 > buf && ptr[-5] == 0)
ptr--;
return ptr - 4 - buf;
}
}
}
return 0;
}
// Split after the parameter sets at the beginning of the stream if they exist.
static int hevc_split(const uint8_t *buf, int buf_size)
{
const uint8_t *ptr = buf, *end = buf + buf_size;
uint32_t state = -1;
int has_vps = 0;
int has_sps = 0;
int has_pps = 0;
int nut;
while (ptr < end) {
ptr = avpriv_find_start_code(ptr, end, &state);
if ((state >> 8) != START_CODE)
break;
nut = (state >> 1) & 0x3F;
if (nut == HEVC_NAL_VPS)
has_vps = 1;
else if (nut == HEVC_NAL_SPS)
has_sps = 1;
else if (nut == HEVC_NAL_PPS)
has_pps = 1;
else if ((nut != HEVC_NAL_SEI_PREFIX || has_pps) &&
nut != HEVC_NAL_AUD) {
if (has_vps && has_sps) {
while (ptr - 4 > buf && ptr[-5] == 0)
ptr--;
return ptr - 4 - buf;
}
}
}
return 0;
}
static int mpegvideo_split(const uint8_t *buf, int buf_size)
{
uint32_t state = -1;
int found = 0;
for (int i = 0; i < buf_size; i++) {
state = (state << 8) | buf[i];
if (state == 0x1B3) {
found = 1;
} else if (found && state != 0x1B5 && state < 0x200 && state >= 0x100)
return i - 3;
}
return 0;
}
static int mpeg4video_split(const uint8_t *buf, int buf_size)
{
const uint8_t *ptr = buf, *end = buf + buf_size;
uint32_t state = -1;
while (ptr < end) {
ptr = avpriv_find_start_code(ptr, end, &state);
if (state == 0x1B3 || state == 0x1B6)
return ptr - 4 - buf;
}
return 0;
}
static int vc1_split(const uint8_t *buf, int buf_size)
{
const uint8_t *ptr = buf, *end = buf + buf_size;
uint32_t state = -1;
int charged = 0;
while (ptr < end) {
ptr = avpriv_find_start_code(ptr, end, &state);
if (state == VC1_CODE_SEQHDR || state == VC1_CODE_ENTRYPOINT) {
charged = 1;
} else if (charged && IS_MARKER(state))
return ptr - 4 - buf;
}
return 0;
}
static int remove_extradata(AVBSFContext *ctx, AVPacket *pkt)
{
RemoveExtradataContext *s = ctx->priv_data;
int ret;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
if (s->freq == REMOVE_FREQ_ALL ||
(s->freq == REMOVE_FREQ_NONKEYFRAME && !(pkt->flags & AV_PKT_FLAG_KEY)) ||
(s->freq == REMOVE_FREQ_KEYFRAME && pkt->flags & AV_PKT_FLAG_KEY)) {
int i;
switch (ctx->par_in->codec_id) {
case AV_CODEC_ID_AV1:
i = av1_split(pkt->data, pkt->size, ctx);
break;
case AV_CODEC_ID_AVS2:
case AV_CODEC_ID_AVS3:
case AV_CODEC_ID_CAVS:
case AV_CODEC_ID_MPEG4:
i = mpeg4video_split(pkt->data, pkt->size);
break;
case AV_CODEC_ID_H264:
i = h264_split(pkt->data, pkt->size);
break;
case AV_CODEC_ID_HEVC:
i = hevc_split(pkt->data, pkt->size);
break;
case AV_CODEC_ID_MPEG1VIDEO:
case AV_CODEC_ID_MPEG2VIDEO:
i = mpegvideo_split(pkt->data, pkt->size);
break;
case AV_CODEC_ID_VC1:
i = vc1_split(pkt->data, pkt->size);
break;
default:
i = 0;
}
pkt->data += i;
pkt->size -= i;
}
return 0;
}
#define OFFSET(x) offsetof(RemoveExtradataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "freq", NULL, OFFSET(freq), AV_OPT_TYPE_INT, { .i64 = REMOVE_FREQ_KEYFRAME }, REMOVE_FREQ_KEYFRAME, REMOVE_FREQ_NONKEYFRAME, FLAGS, .unit = "freq" },
{ "k", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = REMOVE_FREQ_NONKEYFRAME }, .flags = FLAGS, .unit = "freq" },
{ "keyframe", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = REMOVE_FREQ_KEYFRAME }, .flags = FLAGS, .unit = "freq" },
{ "e", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = REMOVE_FREQ_ALL }, .flags = FLAGS, .unit = "freq" },
{ "all", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = REMOVE_FREQ_ALL }, .flags = FLAGS, .unit = "freq" },
{ NULL },
};
static const AVClass remove_extradata_class = {
.class_name = "remove_extradata",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_remove_extradata_bsf = {
.p.name = "remove_extra",
.p.priv_class = &remove_extradata_class,
.priv_data_size = sizeof(RemoveExtradataContext),
.filter = remove_extradata,
};
+277
View File
@@ -0,0 +1,277 @@
/*
* Copyright (c) 2021 Paul B Mahol
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* Change the PTS/DTS timestamps.
*/
#include "libavutil/opt.h"
#include "libavutil/eval.h"
#include "bsf.h"
#include "bsf_internal.h"
static const char *const var_names[] = {
"N", ///< frame number (starting at zero)
"TS",
"POS", ///< original position in the file of the frame
"PREV_INPTS", ///< previous input PTS
"PREV_INDTS", ///< previous input DTS
"PREV_INDURATION", ///< previous input duration
"PREV_OUTPTS", ///< previous output PTS
"PREV_OUTDTS", ///< previous output DTS
"PREV_OUTDURATION", ///< previous output duration
"NEXT_PTS", ///< next input PTS
"NEXT_DTS", ///< next input DTS
"NEXT_DURATION", ///< next input duration
"PTS", ///< original PTS in the file of the frame
"DTS", ///< original DTS in the file of the frame
"DURATION", ///< original duration in the file of the frame
"STARTPTS", ///< PTS at start of movie
"STARTDTS", ///< DTS at start of movie
"TB", ///< input timebase of the stream
"TB_OUT", ///< output timebase of the stream
"SR", ///< sample rate of the stream
"NOPTS", ///< The AV_NOPTS_VALUE constant
NULL
};
enum var_name {
VAR_N,
VAR_TS,
VAR_POS,
VAR_PREV_INPTS,
VAR_PREV_INDTS,
VAR_PREV_INDUR,
VAR_PREV_OUTPTS,
VAR_PREV_OUTDTS,
VAR_PREV_OUTDUR,
VAR_NEXT_PTS,
VAR_NEXT_DTS,
VAR_NEXT_DUR,
VAR_PTS,
VAR_DTS,
VAR_DURATION,
VAR_STARTPTS,
VAR_STARTDTS,
VAR_TB,
VAR_TB_OUT,
VAR_SR,
VAR_NOPTS,
VAR_VARS_NB
};
typedef struct SetTSContext {
const AVClass *class;
char *ts_str;
char *pts_str;
char *dts_str;
char *duration_str;
AVRational time_base;
int64_t frame_number;
double var_values[VAR_VARS_NB];
AVExpr *ts_expr;
AVExpr *pts_expr;
AVExpr *dts_expr;
AVExpr *duration_expr;
AVPacket *prev_inpkt;
AVPacket *prev_outpkt;
AVPacket *cur_pkt;
} SetTSContext;
static int setts_init(AVBSFContext *ctx)
{
SetTSContext *s = ctx->priv_data;
int ret;
s->prev_inpkt = av_packet_alloc();
s->prev_outpkt = av_packet_alloc();
s->cur_pkt = av_packet_alloc();
if (!s->prev_inpkt || !s->prev_outpkt || !s->cur_pkt)
return AVERROR(ENOMEM);
if ((ret = av_expr_parse(&s->ts_expr, s->ts_str,
var_names, NULL, NULL, NULL, NULL, 0, ctx)) < 0) {
av_log(ctx, AV_LOG_ERROR, "Error while parsing ts expression '%s'\n", s->ts_str);
return ret;
}
if ((ret = av_expr_parse(&s->duration_expr, s->duration_str,
var_names, NULL, NULL, NULL, NULL, 0, ctx)) < 0) {
av_log(ctx, AV_LOG_ERROR, "Error while parsing duration expression '%s'\n", s->duration_str);
return ret;
}
if (s->pts_str) {
if ((ret = av_expr_parse(&s->pts_expr, s->pts_str,
var_names, NULL, NULL, NULL, NULL, 0, ctx)) < 0) {
av_log(ctx, AV_LOG_ERROR, "Error while parsing pts expression '%s'\n", s->pts_str);
return ret;
}
}
if (s->dts_str) {
if ((ret = av_expr_parse(&s->dts_expr, s->dts_str,
var_names, NULL, NULL, NULL, NULL, 0, ctx)) < 0) {
av_log(ctx, AV_LOG_ERROR, "Error while parsing dts expression '%s'\n", s->dts_str);
return ret;
}
}
if (s->time_base.num > 0 && s->time_base.den > 0)
ctx->time_base_out = s->time_base;
s->frame_number= 0;
s->var_values[VAR_STARTPTS] = AV_NOPTS_VALUE;
s->var_values[VAR_STARTDTS] = AV_NOPTS_VALUE;
s->var_values[VAR_NOPTS] = AV_NOPTS_VALUE;
s->var_values[VAR_TB] = ctx->time_base_in.den ? av_q2d(ctx->time_base_in) : 0;
s->var_values[VAR_TB_OUT]= ctx->time_base_out.den ? av_q2d(ctx->time_base_out) : 0;
s->var_values[VAR_SR] = ctx->par_in->sample_rate;
return 0;
}
static int setts_filter(AVBSFContext *ctx, AVPacket *pkt)
{
SetTSContext *s = ctx->priv_data;
int64_t new_ts, new_pts, new_dts, new_duration;
int ret;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0 && (ret != AVERROR_EOF || !s->cur_pkt->data))
return ret;
if (!s->cur_pkt->data) {
av_packet_move_ref(s->cur_pkt, pkt);
return AVERROR(EAGAIN);
}
if (s->var_values[VAR_STARTPTS] == AV_NOPTS_VALUE)
s->var_values[VAR_STARTPTS] = s->cur_pkt->pts;
if (s->var_values[VAR_STARTDTS] == AV_NOPTS_VALUE)
s->var_values[VAR_STARTDTS] = s->cur_pkt->dts;
s->var_values[VAR_N] = s->frame_number++;
s->var_values[VAR_TS] = s->cur_pkt->dts;
s->var_values[VAR_POS] = s->cur_pkt->pos;
s->var_values[VAR_PTS] = s->cur_pkt->pts;
s->var_values[VAR_DTS] = s->cur_pkt->dts;
s->var_values[VAR_DURATION] = s->cur_pkt->duration;
s->var_values[VAR_PREV_INPTS] = s->prev_inpkt->pts;
s->var_values[VAR_PREV_INDTS] = s->prev_inpkt->dts;
s->var_values[VAR_PREV_INDUR] = s->prev_inpkt->duration;
s->var_values[VAR_PREV_OUTPTS] = s->prev_outpkt->pts;
s->var_values[VAR_PREV_OUTDTS] = s->prev_outpkt->dts;
s->var_values[VAR_PREV_OUTDUR] = s->prev_outpkt->duration;
s->var_values[VAR_NEXT_PTS] = pkt->pts;
s->var_values[VAR_NEXT_DTS] = pkt->dts;
s->var_values[VAR_NEXT_DUR] = pkt->duration;
new_ts = llrint(av_expr_eval(s->ts_expr, s->var_values, NULL));
new_duration = llrint(av_expr_eval(s->duration_expr, s->var_values, NULL));
if (s->pts_str) {
s->var_values[VAR_TS] = s->cur_pkt->pts;
new_pts = llrint(av_expr_eval(s->pts_expr, s->var_values, NULL));
} else {
new_pts = new_ts;
}
if (s->dts_str) {
s->var_values[VAR_TS] = s->cur_pkt->dts;
new_dts = llrint(av_expr_eval(s->dts_expr, s->var_values, NULL));
} else {
new_dts = new_ts;
}
av_packet_unref(s->prev_inpkt);
av_packet_unref(s->prev_outpkt);
av_packet_move_ref(s->prev_inpkt, s->cur_pkt);
av_packet_move_ref(s->cur_pkt, pkt);
ret = av_packet_ref(pkt, s->prev_inpkt);
if (ret < 0)
return ret;
pkt->pts = new_pts;
pkt->dts = new_dts;
pkt->duration = new_duration;
ret = av_packet_ref(s->prev_outpkt, pkt);
if (ret < 0)
av_packet_unref(pkt);
return ret;
}
static void setts_close(AVBSFContext *bsf)
{
SetTSContext *s = bsf->priv_data;
av_packet_free(&s->prev_inpkt);
av_packet_free(&s->prev_outpkt);
av_packet_free(&s->cur_pkt);
av_expr_free(s->ts_expr);
s->ts_expr = NULL;
av_expr_free(s->pts_expr);
s->pts_expr = NULL;
av_expr_free(s->dts_expr);
s->dts_expr = NULL;
av_expr_free(s->duration_expr);
s->duration_expr = NULL;
}
#define OFFSET(x) offsetof(SetTSContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_SUBTITLE_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption options[] = {
{ "ts", "set expression for packet PTS and DTS", OFFSET(ts_str), AV_OPT_TYPE_STRING, {.str="TS"}, 0, 0, FLAGS },
{ "pts", "set expression for packet PTS", OFFSET(pts_str), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
{ "dts", "set expression for packet DTS", OFFSET(dts_str), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
{ "duration", "set expression for packet duration", OFFSET(duration_str), AV_OPT_TYPE_STRING, {.str="DURATION"}, 0, 0, FLAGS },
{ "time_base", "set output timebase", OFFSET(time_base), AV_OPT_TYPE_RATIONAL, {.dbl=0}, 0, INT_MAX, FLAGS },
{ NULL },
};
static const AVClass setts_class = {
.class_name = "setts_bsf",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFBitStreamFilter ff_setts_bsf = {
.p.name = "setts",
.p.priv_class = &setts_class,
.priv_data_size = sizeof(SetTSContext),
.init = setts_init,
.close = setts_close,
.filter = setts_filter,
};
+67
View File
@@ -0,0 +1,67 @@
/*
* Copyright (c) 2024 Anton Khirnov <anton@khirnov.net>
*
* 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 <inttypes.h>
#include "bsf.h"
#include "bsf_internal.h"
#include "libavutil/adler32.h"
#include "libavutil/log.h"
#include "libavutil/timestamp.h"
typedef struct ShowinfoContext {
uint64_t nb_packets;
} ShowinfoContext;
static int showinfo_filter(AVBSFContext *ctx, AVPacket *pkt)
{
ShowinfoContext *priv = ctx->priv_data;
uint32_t crc;
int ret;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
crc = av_adler32_update(0, pkt->data, pkt->size);
av_log(ctx, AV_LOG_INFO,
"n:%7"PRIu64" "
"size:%7d "
"pts:%s pt:%s "
"dts:%s dt:%s "
"ds:%"PRId64" d:%s "
"adler32:0x%08"PRIx32
"\n",
priv->nb_packets, pkt->size,
av_ts2str(pkt->pts), av_ts2timestr(pkt->pts, &ctx->time_base_in),
av_ts2str(pkt->dts), av_ts2timestr(pkt->dts, &ctx->time_base_in),
pkt->duration, av_ts2timestr(pkt->duration, &ctx->time_base_in), crc);
priv->nb_packets++;
return 0;
}
const FFBitStreamFilter ff_showinfo_bsf = {
.p.name = "showinfo",
.filter = showinfo_filter,
.priv_data_size = sizeof(ShowinfoContext),
};
+129
View File
@@ -0,0 +1,129 @@
/*
* 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 <inttypes.h>
#include "libavutil/avstring.h"
#include "libavutil/avutil.h"
#include "libavutil/log.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
typedef struct TraceHeadersContext {
CodedBitstreamContext *cbc;
CodedBitstreamFragment fragment;
} TraceHeadersContext;
static int trace_headers_init(AVBSFContext *bsf)
{
TraceHeadersContext *ctx = bsf->priv_data;
int err;
err = ff_cbs_init(&ctx->cbc, bsf->par_in->codec_id, bsf);
if (err < 0)
return err;
ctx->cbc->trace_enable = 1;
ctx->cbc->trace_level = AV_LOG_INFO;
ctx->cbc->trace_context = ctx->cbc;
ctx->cbc->trace_read_callback = ff_cbs_trace_read_log;
if (bsf->par_in->extradata) {
CodedBitstreamFragment *frag = &ctx->fragment;
av_log(bsf, AV_LOG_INFO, "Extradata\n");
err = ff_cbs_read_extradata(ctx->cbc, frag, bsf->par_in);
ff_cbs_fragment_reset(frag);
}
return err;
}
static void trace_headers_close(AVBSFContext *bsf)
{
TraceHeadersContext *ctx = bsf->priv_data;
ff_cbs_fragment_free(&ctx->fragment);
ff_cbs_close(&ctx->cbc);
}
static int trace_headers(AVBSFContext *bsf, AVPacket *pkt)
{
TraceHeadersContext *ctx = bsf->priv_data;
CodedBitstreamFragment *frag = &ctx->fragment;
char tmp[256] = { 0 };
int err;
err = ff_bsf_get_packet_ref(bsf, pkt);
if (err < 0)
return err;
if (pkt->flags & AV_PKT_FLAG_KEY)
av_strlcat(tmp, ", key frame", sizeof(tmp));
if (pkt->flags & AV_PKT_FLAG_CORRUPT)
av_strlcat(tmp, ", corrupt", sizeof(tmp));
if (pkt->pts != AV_NOPTS_VALUE)
av_strlcatf(tmp, sizeof(tmp), ", pts %"PRId64, pkt->pts);
else
av_strlcat(tmp, ", no pts", sizeof(tmp));
if (pkt->dts != AV_NOPTS_VALUE)
av_strlcatf(tmp, sizeof(tmp), ", dts %"PRId64, pkt->dts);
else
av_strlcat(tmp, ", no dts", sizeof(tmp));
if (pkt->duration > 0)
av_strlcatf(tmp, sizeof(tmp), ", duration %"PRId64, pkt->duration);
av_log(bsf, AV_LOG_INFO, "Packet: %d bytes%s.\n", pkt->size, tmp);
if (av_packet_get_side_data(pkt, AV_PKT_DATA_NEW_EXTRADATA, NULL)) {
av_log(bsf, AV_LOG_INFO, "Side data:\n");
err = ff_cbs_read_packet_side_data(ctx->cbc, frag, pkt);
ff_cbs_fragment_reset(frag);
if (err < 0) {
av_packet_unref(pkt);
return err;
}
av_log(bsf, AV_LOG_INFO, "Payload:\n");
}
err = ff_cbs_read_packet(ctx->cbc, frag, pkt);
ff_cbs_fragment_reset(frag);
if (err < 0)
av_packet_unref(pkt);
return err;
}
const FFBitStreamFilter ff_trace_headers_bsf = {
.p.name = "trace_headers",
.p.codec_ids = ff_cbs_all_codec_ids,
.priv_data_size = sizeof(TraceHeadersContext),
.init = &trace_headers_init,
.close = &trace_headers_close,
.filter = &trace_headers,
};
+182
View File
@@ -0,0 +1,182 @@
/*
* Copyright (c) 2018 Paul B Mahol
*
* 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 "bsf.h"
#include "bsf_internal.h"
#include "get_bits.h"
#include "mlp_parse.h"
#include "mlp.h"
typedef struct AccessUnit {
uint8_t bits[4];
uint16_t offset;
uint16_t optional;
} AccessUnit;
typedef struct TrueHDCoreContext {
MLPHeaderInfo hdr;
} TrueHDCoreContext;
static int truehd_core_filter(AVBSFContext *ctx, AVPacket *pkt)
{
TrueHDCoreContext *s = ctx->priv_data;
GetBitContext gbc;
AccessUnit units[MAX_SUBSTREAMS];
int ret, i, last_offset = 0;
int in_size, out_size;
int have_header = 0;
int substream_bytes = 0;
int end;
ret = ff_bsf_get_packet_ref(ctx, pkt);
if (ret < 0)
return ret;
if (pkt->size < 4) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
in_size = (AV_RB16(pkt->data) & 0xFFF) * 2;
if (in_size < 4 || in_size > pkt->size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
ret = init_get_bits8(&gbc, pkt->data + 4, pkt->size - 4);
if (ret < 0)
goto fail;
if (show_bits_long(&gbc, 32) == 0xf8726fba) {
if ((ret = ff_mlp_read_major_sync(ctx, &s->hdr, &gbc)) < 0)
goto fail;
have_header = 1;
}
if (s->hdr.num_substreams > MAX_SUBSTREAMS) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
for (i = 0; i < s->hdr.num_substreams; i++) {
for (int j = 0; j < 4; j++)
units[i].bits[j] = get_bits1(&gbc);
units[i].offset = get_bits(&gbc, 12);
if (i < 3) {
last_offset = units[i].offset * 2;
substream_bytes += 2;
}
if (units[i].bits[0]) {
units[i].optional = get_bits(&gbc, 16);
if (i < 3)
substream_bytes += 2;
}
}
end = get_bits_count(&gbc) >> 3;
out_size = end + 4 + last_offset;
if (out_size < in_size) {
int bpos = 0, reduce = end - have_header * 28 - substream_bytes;
uint16_t parity_nibble, dts = AV_RB16(pkt->data + 2);
uint16_t auheader;
uint8_t header[28];
av_assert1(reduce >= 0 && reduce % 2 == 0);
if (have_header) {
memcpy(header, pkt->data + 4, 28);
header[16] = (header[16] & 0x0c) | (FFMIN(s->hdr.num_substreams, 3) << 4);
header[17] &= 0x7f;
header[25] &= 0xfe;
AV_WL16(header + 26, ff_mlp_checksum16(header, 26));
}
pkt->data += reduce;
out_size -= reduce;
pkt->size = out_size;
ret = av_packet_make_writable(pkt);
if (ret < 0)
goto fail;
AV_WB16(pkt->data + 2, dts);
parity_nibble = dts;
parity_nibble ^= out_size / 2;
for (i = 0; i < FFMIN(s->hdr.num_substreams, 3); i++) {
uint16_t substr_hdr = 0;
substr_hdr |= (units[i].bits[0] << 15);
substr_hdr |= (units[i].bits[1] << 14);
substr_hdr |= (units[i].bits[2] << 13);
substr_hdr |= (units[i].bits[3] << 12);
substr_hdr |= units[i].offset;
AV_WB16(pkt->data + have_header * 28 + 4 + bpos, substr_hdr);
parity_nibble ^= substr_hdr;
bpos += 2;
if (units[i].bits[0]) {
AV_WB16(pkt->data + have_header * 28 + 4 + bpos, units[i].optional);
parity_nibble ^= units[i].optional;
bpos += 2;
}
}
parity_nibble ^= parity_nibble >> 8;
parity_nibble ^= parity_nibble >> 4;
parity_nibble &= 0xF;
auheader = (parity_nibble ^ 0xF) << 12;
auheader |= (out_size / 2) & 0x0fff;
AV_WB16(pkt->data, auheader);
if (have_header)
memcpy(pkt->data + 4, header, 28);
}
fail:
if (ret < 0)
av_packet_unref(pkt);
return ret;
}
static void truehd_core_flush(AVBSFContext *ctx)
{
TrueHDCoreContext *s = ctx->priv_data;
memset(&s->hdr, 0, sizeof(s->hdr));
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_TRUEHD, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_truehd_core_bsf = {
.p.name = "truehd_core",
.p.codec_ids = codec_ids,
.priv_data_size = sizeof(TrueHDCoreContext),
.filter = truehd_core_filter,
.flush = truehd_core_flush,
};
+146
View File
@@ -0,0 +1,146 @@
/*
* 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/log.h"
#include "libavutil/opt.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "cbs.h"
#include "cbs_bsf.h"
#include "cbs_vp9.h"
typedef struct VP9MetadataContext {
CBSBSFContext common;
int color_space;
int color_range;
int color_warnings;
} VP9MetadataContext;
static int vp9_metadata_update_fragment(AVBSFContext *bsf, AVPacket *pkt,
CodedBitstreamFragment *frag)
{
VP9MetadataContext *ctx = bsf->priv_data;
int i;
for (i = 0; i < frag->nb_units; i++) {
VP9RawFrame *frame = frag->units[i].content;
VP9RawFrameHeader *header = &frame->header;
int profile = (header->profile_high_bit << 1) + header->profile_low_bit;
if (header->frame_type == VP9_KEY_FRAME ||
header->intra_only && profile > 0) {
if (ctx->color_space >= 0) {
if (!(profile & 1) && ctx->color_space == VP9_CS_RGB) {
if (!(ctx->color_warnings & 2)) {
av_log(bsf, AV_LOG_WARNING, "Warning: RGB "
"incompatible with profiles 0 and 2.\n");
ctx->color_warnings |= 2;
}
} else
header->color_space = ctx->color_space;
}
if (ctx->color_range >= 0)
header->color_range = ctx->color_range;
if (header->color_space == VP9_CS_RGB) {
if (!(ctx->color_warnings & 1) && !header->color_range) {
av_log(bsf, AV_LOG_WARNING, "Warning: Color space RGB "
"implicitly sets color range to PC range.\n");
ctx->color_warnings |= 1;
}
header->color_range = 1;
}
} else if (!(ctx->color_warnings & 4) && header->intra_only && !profile &&
ctx->color_space >= 0 && ctx->color_space != VP9_CS_BT_601) {
av_log(bsf, AV_LOG_WARNING, "Warning: Intra-only frames in "
"profile 0 are automatically BT.601.\n");
ctx->color_warnings |= 4;
}
}
return 0;
}
static const CBSBSFType vp9_metadata_type = {
.codec_id = AV_CODEC_ID_VP9,
.fragment_name = "superframe",
.unit_name = "frame",
.update_fragment = &vp9_metadata_update_fragment,
};
static int vp9_metadata_init(AVBSFContext *bsf)
{
return ff_cbs_bsf_generic_init(bsf, &vp9_metadata_type);
}
#define OFFSET(x) offsetof(VP9MetadataContext, x)
#define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
static const AVOption vp9_metadata_options[] = {
{ "color_space", "Set colour space (section 7.2.2)",
OFFSET(color_space), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, VP9_CS_RGB, FLAGS, .unit = "cs" },
{ "unknown", "Unknown/unspecified", 0, AV_OPT_TYPE_CONST,
{ .i64 = VP9_CS_UNKNOWN }, .flags = FLAGS, .unit = "cs" },
{ "bt601", "ITU-R BT.601-7", 0, AV_OPT_TYPE_CONST,
{ .i64 = VP9_CS_BT_601 }, .flags = FLAGS, .unit = "cs" },
{ "bt709", "ITU-R BT.709-6", 0, AV_OPT_TYPE_CONST,
{ .i64 = VP9_CS_BT_709 }, .flags = FLAGS, .unit = "cs" },
{ "smpte170", "SMPTE-170", 0, AV_OPT_TYPE_CONST,
{ .i64 = VP9_CS_SMPTE_170 }, .flags = FLAGS, .unit = "cs" },
{ "smpte240", "SMPTE-240", 0, AV_OPT_TYPE_CONST,
{ .i64 = VP9_CS_SMPTE_240 }, .flags = FLAGS, .unit = "cs" },
{ "bt2020", "ITU-R BT.2020-2", 0, AV_OPT_TYPE_CONST,
{ .i64 = VP9_CS_BT_2020 }, .flags = FLAGS, .unit = "cs" },
{ "rgb", "sRGB / IEC 61966-2-1", 0, AV_OPT_TYPE_CONST,
{ .i64 = VP9_CS_RGB }, .flags = FLAGS, .unit = "cs" },
{ "color_range", "Set colour range (section 7.2.2)",
OFFSET(color_range), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, 1, FLAGS, .unit = "cr" },
{ "tv", "TV (limited) range", 0, AV_OPT_TYPE_CONST,
{ .i64 = 0 }, .flags = FLAGS, .unit = "cr" },
{ "pc", "PC (full) range", 0, AV_OPT_TYPE_CONST,
{ .i64 = 1 }, .flags = FLAGS, .unit = "cr" },
{ NULL }
};
static const AVClass vp9_metadata_class = {
.class_name = "vp9_metadata_bsf",
.item_name = av_default_item_name,
.option = vp9_metadata_options,
.version = LIBAVUTIL_VERSION_INT,
};
static const enum AVCodecID vp9_metadata_codec_ids[] = {
AV_CODEC_ID_VP9, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_vp9_metadata_bsf = {
.p.name = "vp9_metadata",
.p.codec_ids = vp9_metadata_codec_ids,
.p.priv_class = &vp9_metadata_class,
.priv_data_size = sizeof(VP9MetadataContext),
.init = &vp9_metadata_init,
.close = &ff_cbs_bsf_generic_close,
.filter = &ff_cbs_bsf_generic_filter,
};
+414
View File
@@ -0,0 +1,414 @@
/*
* 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/intmath.h"
#include "libavutil/log.h"
#include "libavutil/mem.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "get_bits.h"
#include "put_bits.h"
#define FRAME_SLOTS 8
typedef struct VP9RawReorderFrame {
AVPacket *packet;
int needs_output;
int needs_display;
int64_t pts;
int64_t sequence;
unsigned int slots;
unsigned int profile;
unsigned int show_existing_frame;
unsigned int frame_to_show;
unsigned int frame_type;
unsigned int show_frame;
unsigned int refresh_frame_flags;
} VP9RawReorderFrame;
typedef struct VP9RawReorderContext {
int64_t sequence;
VP9RawReorderFrame *slot[FRAME_SLOTS];
VP9RawReorderFrame *next_frame;
} VP9RawReorderContext;
static void vp9_raw_reorder_frame_free(VP9RawReorderFrame **frame)
{
if (*frame)
av_packet_free(&(*frame)->packet);
av_freep(frame);
}
static void vp9_raw_reorder_clear_slot(VP9RawReorderContext *ctx, int s)
{
if (ctx->slot[s]) {
ctx->slot[s]->slots &= ~(1 << s);
if (ctx->slot[s]->slots == 0)
vp9_raw_reorder_frame_free(&ctx->slot[s]);
else
ctx->slot[s] = NULL;
}
}
static int vp9_raw_reorder_frame_parse(AVBSFContext *bsf, VP9RawReorderFrame *frame)
{
GetBitContext bc;
int err;
unsigned int frame_marker;
unsigned int profile_low_bit, profile_high_bit, reserved_zero;
unsigned int error_resilient_mode;
unsigned int frame_sync_code;
err = init_get_bits(&bc, frame->packet->data, 8 * frame->packet->size);
if (err)
return err;
frame_marker = get_bits(&bc, 2);
if (frame_marker != 2) {
av_log(bsf, AV_LOG_ERROR, "Invalid frame marker: %u.\n",
frame_marker);
return AVERROR_INVALIDDATA;
}
profile_low_bit = get_bits1(&bc);
profile_high_bit = get_bits1(&bc);
frame->profile = (profile_high_bit << 1) | profile_low_bit;
if (frame->profile == 3) {
reserved_zero = get_bits1(&bc);
if (reserved_zero != 0) {
av_log(bsf, AV_LOG_ERROR, "Profile reserved_zero bit set: "
"unsupported profile or invalid bitstream.\n");
return AVERROR_INVALIDDATA;
}
}
frame->show_existing_frame = get_bits1(&bc);
if (frame->show_existing_frame) {
frame->frame_to_show = get_bits(&bc, 3);
return 0;
}
frame->frame_type = get_bits1(&bc);
frame->show_frame = get_bits1(&bc);
error_resilient_mode = get_bits1(&bc);
if (frame->frame_type == 0) {
frame_sync_code = get_bits(&bc, 24);
if (frame_sync_code != 0x498342) {
av_log(bsf, AV_LOG_ERROR, "Invalid frame sync code: %06x.\n",
frame_sync_code);
return AVERROR_INVALIDDATA;
}
frame->refresh_frame_flags = 0xff;
} else {
unsigned int intra_only;
if (frame->show_frame == 0)
intra_only = get_bits1(&bc);
else
intra_only = 0;
if (error_resilient_mode == 0) {
// reset_frame_context
skip_bits(&bc, 2);
}
if (intra_only) {
frame_sync_code = get_bits(&bc, 24);
if (frame_sync_code != 0x498342) {
av_log(bsf, AV_LOG_ERROR, "Invalid frame sync code: "
"%06x.\n", frame_sync_code);
return AVERROR_INVALIDDATA;
}
if (frame->profile > 0) {
unsigned int color_space;
if (frame->profile >= 2) {
// ten_or_twelve_bit
skip_bits(&bc, 1);
}
color_space = get_bits(&bc, 3);
if (color_space != 7 /* CS_RGB */) {
// color_range
skip_bits(&bc, 1);
if (frame->profile == 1 || frame->profile == 3) {
// subsampling
skip_bits(&bc, 3);
}
} else {
if (frame->profile == 1 || frame->profile == 3)
skip_bits(&bc, 1);
}
}
frame->refresh_frame_flags = get_bits(&bc, 8);
} else {
frame->refresh_frame_flags = get_bits(&bc, 8);
}
}
return 0;
}
static int vp9_raw_reorder_make_output(AVBSFContext *bsf,
AVPacket *out,
VP9RawReorderFrame *last_frame)
{
VP9RawReorderContext *ctx = bsf->priv_data;
VP9RawReorderFrame *next_output = last_frame,
*next_display = last_frame, *frame;
int s, err;
for (s = 0; s < FRAME_SLOTS; s++) {
frame = ctx->slot[s];
if (!frame)
continue;
if (frame->needs_output && (!next_output ||
frame->sequence < next_output->sequence))
next_output = frame;
if (frame->needs_display && (!next_display ||
frame->pts < next_display->pts))
next_display = frame;
}
if (!next_output && !next_display)
return AVERROR_EOF;
if (!next_display || (next_output &&
next_output->sequence < next_display->sequence))
frame = next_output;
else
frame = next_display;
if (frame->needs_output && frame->needs_display &&
next_output == next_display) {
av_log(bsf, AV_LOG_DEBUG, "Output and display frame "
"%"PRId64" (%"PRId64") in order.\n",
frame->sequence, frame->pts);
av_packet_move_ref(out, frame->packet);
frame->needs_output = frame->needs_display = 0;
} else if (frame->needs_output) {
if (frame->needs_display) {
av_log(bsf, AV_LOG_DEBUG, "Output frame %"PRId64" "
"(%"PRId64") for later display.\n",
frame->sequence, frame->pts);
} else {
av_log(bsf, AV_LOG_DEBUG, "Output unshown frame "
"%"PRId64" (%"PRId64") to keep order.\n",
frame->sequence, frame->pts);
}
av_packet_move_ref(out, frame->packet);
out->pts = out->dts;
frame->needs_output = 0;
} else {
PutBitContext pb;
av_assert0(!frame->needs_output && frame->needs_display);
if (frame->slots == 0) {
av_log(bsf, AV_LOG_ERROR, "Attempting to display frame "
"which is no longer available?\n");
frame->needs_display = 0;
return AVERROR_INVALIDDATA;
}
s = ff_ctz(frame->slots);
av_assert0(s < FRAME_SLOTS);
av_log(bsf, AV_LOG_DEBUG, "Display frame %"PRId64" "
"(%"PRId64") from slot %d.\n",
frame->sequence, frame->pts, s);
err = av_new_packet(out, 2);
if (err < 0)
return err;
init_put_bits(&pb, out->data, 2);
// frame_marker
put_bits(&pb, 2, 2);
// profile_low_bit
put_bits(&pb, 1, frame->profile & 1);
// profile_high_bit
put_bits(&pb, 1, (frame->profile >> 1) & 1);
if (frame->profile == 3) {
// reserved_zero
put_bits(&pb, 1, 0);
}
// show_existing_frame
put_bits(&pb, 1, 1);
// frame_to_show_map_idx
put_bits(&pb, 3, s);
while (put_bits_count(&pb) < 16)
put_bits(&pb, 1, 0);
flush_put_bits(&pb);
out->pts = out->dts = frame->pts;
frame->needs_display = 0;
}
return 0;
}
static int vp9_raw_reorder_filter(AVBSFContext *bsf, AVPacket *out)
{
VP9RawReorderContext *ctx = bsf->priv_data;
VP9RawReorderFrame *frame;
AVPacket *in;
int err, s;
if (ctx->next_frame) {
frame = ctx->next_frame;
} else {
err = ff_bsf_get_packet(bsf, &in);
if (err < 0) {
if (err == AVERROR_EOF)
return vp9_raw_reorder_make_output(bsf, out, NULL);
return err;
}
if (!in->size) {
av_packet_free(&in);
return AVERROR_INVALIDDATA;
}
if ((in->data[in->size - 1] & 0xe0) == 0xc0) {
av_log(bsf, AV_LOG_ERROR, "Input in superframes is not "
"supported.\n");
av_packet_free(&in);
return AVERROR(ENOSYS);
}
frame = av_mallocz(sizeof(*frame));
if (!frame) {
av_packet_free(&in);
return AVERROR(ENOMEM);
}
frame->packet = in;
frame->pts = in->pts;
frame->sequence = ++ctx->sequence;
err = vp9_raw_reorder_frame_parse(bsf, frame);
if (err) {
av_log(bsf, AV_LOG_ERROR, "Failed to parse input "
"frame: %d.\n", err);
goto fail;
}
frame->needs_output = 1;
frame->needs_display = frame->pts != AV_NOPTS_VALUE;
if (frame->show_existing_frame)
av_log(bsf, AV_LOG_DEBUG, "Show frame %"PRId64" "
"(%"PRId64"): show %u.\n", frame->sequence,
frame->pts, frame->frame_to_show);
else
av_log(bsf, AV_LOG_DEBUG, "New frame %"PRId64" "
"(%"PRId64"): type %u show %u refresh %02x.\n",
frame->sequence, frame->pts, frame->frame_type,
frame->show_frame, frame->refresh_frame_flags);
ctx->next_frame = frame;
}
for (s = 0; s < FRAME_SLOTS; s++) {
if (!(frame->refresh_frame_flags & (1 << s)))
continue;
if (ctx->slot[s] && ctx->slot[s]->needs_display &&
ctx->slot[s]->slots == (1 << s)) {
// We are overwriting this slot, which is last reference
// to the frame previously present in it. In order to be
// a valid stream, that frame must already have been
// displayed before the pts of the current frame.
err = vp9_raw_reorder_make_output(bsf, out, ctx->slot[s]);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to create "
"output overwriting slot %d: %d.\n",
s, err);
// Clear the slot anyway, so we don't end up
// in an infinite loop.
vp9_raw_reorder_clear_slot(ctx, s);
return AVERROR_INVALIDDATA;
}
return 0;
}
vp9_raw_reorder_clear_slot(ctx, s);
}
for (s = 0; s < FRAME_SLOTS; s++) {
if (!(frame->refresh_frame_flags & (1 << s)))
continue;
ctx->slot[s] = frame;
}
frame->slots = frame->refresh_frame_flags;
if (!frame->refresh_frame_flags) {
err = vp9_raw_reorder_make_output(bsf, out, frame);
if (err < 0) {
av_log(bsf, AV_LOG_ERROR, "Failed to create output "
"for transient frame.\n");
ctx->next_frame = NULL;
return AVERROR_INVALIDDATA;
}
if (!frame->needs_display) {
vp9_raw_reorder_frame_free(&frame);
ctx->next_frame = NULL;
}
return 0;
}
ctx->next_frame = NULL;
return AVERROR(EAGAIN);
fail:
vp9_raw_reorder_frame_free(&frame);
return err;
}
static av_cold void vp9_raw_reorder_flush_close(AVBSFContext *bsf)
{
VP9RawReorderContext *ctx = bsf->priv_data;
for (int s = 0; s < FRAME_SLOTS; s++)
vp9_raw_reorder_clear_slot(ctx, s);
vp9_raw_reorder_frame_free(&ctx->next_frame);
ctx->sequence = 0;
}
static const enum AVCodecID vp9_raw_reorder_codec_ids[] = {
AV_CODEC_ID_VP9, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_vp9_raw_reorder_bsf = {
.p.name = "vp9_raw_reorder",
.p.codec_ids = vp9_raw_reorder_codec_ids,
.priv_data_size = sizeof(VP9RawReorderContext),
.filter = &vp9_raw_reorder_filter,
.flush = &vp9_raw_reorder_flush_close,
.close = &vp9_raw_reorder_flush_close,
};
+231
View File
@@ -0,0 +1,231 @@
/*
* Vp9 invisible (alt-ref) frame to superframe merge bitstream filter
* Copyright (c) 2016 Ronald S. Bultje <rsbultje@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 "bsf.h"
#include "bsf_internal.h"
#include "get_bits.h"
#define MAX_CACHE 8
typedef struct VP9BSFContext {
int n_cache;
AVPacket *cache[MAX_CACHE];
} VP9BSFContext;
static void stats(AVPacket * const *in, int n_in,
unsigned *_max, unsigned *_sum)
{
int n;
unsigned max = 0, sum = 0;
for (n = 0; n < n_in; n++) {
unsigned sz = in[n]->size;
if (sz > max)
max = sz;
sum += sz;
}
*_max = max;
*_sum = sum;
}
static int merge_superframe(AVPacket * const *in, int n_in, AVPacket *out)
{
unsigned max, sum, mag, marker, n, sz;
uint8_t *ptr;
int res;
stats(in, n_in, &max, &sum);
mag = av_log2(max) >> 3;
marker = 0xC0 + (mag << 3) + (n_in - 1);
sz = sum + 2 + (mag + 1) * n_in;
res = av_new_packet(out, sz);
if (res < 0)
return res;
ptr = out->data;
for (n = 0; n < n_in; n++) {
memcpy(ptr, in[n]->data, in[n]->size);
ptr += in[n]->size;
}
#define wloop(mag, wr) \
do { \
for (n = 0; n < n_in; n++) { \
wr; \
ptr += mag + 1; \
} \
} while (0)
// write superframe with marker 110[mag:2][nframes:3]
*ptr++ = marker;
switch (mag) {
case 0:
wloop(mag, *ptr = in[n]->size);
break;
case 1:
wloop(mag, AV_WL16(ptr, in[n]->size));
break;
case 2:
wloop(mag, AV_WL24(ptr, in[n]->size));
break;
case 3:
wloop(mag, AV_WL32(ptr, in[n]->size));
break;
}
*ptr++ = marker;
av_assert0(ptr == &out->data[out->size]);
return 0;
}
static int vp9_superframe_filter(AVBSFContext *ctx, AVPacket *pkt)
{
GetBitContext gb;
VP9BSFContext *s = ctx->priv_data;
int res, invisible, profile, marker, uses_superframe_syntax = 0, n;
res = ff_bsf_get_packet_ref(ctx, pkt);
if (res < 0)
return res;
if (!pkt->size) {
/* In case the cache is empty we can pass side-data-only packets
* through unchanged. Otherwise, such a packet makes no sense. */
if (!s->n_cache)
return 0;
res = AVERROR_INVALIDDATA;
goto done;
}
marker = pkt->data[pkt->size - 1];
if ((marker & 0xe0) == 0xc0) {
int nbytes = 1 + ((marker >> 3) & 0x3);
int n_frames = 1 + (marker & 0x7), idx_sz = 2 + n_frames * nbytes;
uses_superframe_syntax = pkt->size >= idx_sz && pkt->data[pkt->size - idx_sz] == marker;
}
if ((res = init_get_bits8(&gb, pkt->data, pkt->size)) < 0)
goto done;
get_bits(&gb, 2); // frame marker
profile = get_bits1(&gb);
profile |= get_bits1(&gb) << 1;
if (profile == 3) profile += get_bits1(&gb);
if (get_bits1(&gb)) {
invisible = 0;
} else {
get_bits1(&gb); // keyframe
invisible = !get_bits1(&gb);
}
if (uses_superframe_syntax && s->n_cache > 0) {
av_log(ctx, AV_LOG_ERROR,
"Mixing of superframe syntax and naked VP9 frames not supported\n");
res = AVERROR(ENOSYS);
goto done;
} else if ((!invisible || uses_superframe_syntax) && !s->n_cache) {
// passthrough
return 0;
} else if (s->n_cache + 1 >= MAX_CACHE) {
av_log(ctx, AV_LOG_ERROR,
"Too many invisible frames\n");
res = AVERROR_INVALIDDATA;
goto done;
}
av_packet_move_ref(s->cache[s->n_cache++], pkt);
if (invisible) {
return AVERROR(EAGAIN);
}
av_assert0(s->n_cache > 0);
// build superframe
if ((res = merge_superframe(s->cache, s->n_cache, pkt)) < 0)
goto done;
res = av_packet_copy_props(pkt, s->cache[s->n_cache - 1]);
if (res < 0)
goto done;
for (n = 0; n < s->n_cache; n++)
av_packet_unref(s->cache[n]);
s->n_cache = 0;
done:
if (res < 0)
av_packet_unref(pkt);
return res;
}
static int vp9_superframe_init(AVBSFContext *ctx)
{
VP9BSFContext *s = ctx->priv_data;
int n;
// alloc cache packets
for (n = 0; n < MAX_CACHE; n++) {
s->cache[n] = av_packet_alloc();
if (!s->cache[n])
return AVERROR(ENOMEM);
}
return 0;
}
static void vp9_superframe_flush(AVBSFContext *ctx)
{
VP9BSFContext *s = ctx->priv_data;
int n;
// unref cached data
for (n = 0; n < s->n_cache; n++)
av_packet_unref(s->cache[n]);
s->n_cache = 0;
}
static void vp9_superframe_close(AVBSFContext *ctx)
{
VP9BSFContext *s = ctx->priv_data;
int n;
// free cached data
for (n = 0; n < MAX_CACHE; n++)
av_packet_free(&s->cache[n]);
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_VP9, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_vp9_superframe_bsf = {
.p.name = "vp9_superframe",
.p.codec_ids = codec_ids,
.priv_data_size = sizeof(VP9BSFContext),
.filter = vp9_superframe_filter,
.init = vp9_superframe_init,
.flush = vp9_superframe_flush,
.close = vp9_superframe_close,
};
@@ -0,0 +1,170 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* This bitstream filter splits VP9 superframes into packets containing
* just one frame.
*/
#include <stddef.h>
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "get_bits.h"
typedef struct VP9SFSplitContext {
AVPacket *buffer_pkt;
int nb_frames;
int next_frame;
size_t next_frame_offset;
int sizes[8];
} VP9SFSplitContext;
static int vp9_superframe_split_filter(AVBSFContext *ctx, AVPacket *out)
{
VP9SFSplitContext *s = ctx->priv_data;
AVPacket *in;
int i, j, ret, marker;
int is_superframe = !!s->buffer_pkt->data;
if (!s->buffer_pkt->data) {
ret = ff_bsf_get_packet_ref(ctx, s->buffer_pkt);
if (ret < 0)
return ret;
in = s->buffer_pkt;
if (!in->size)
goto passthrough;
marker = in->data[in->size - 1];
if ((marker & 0xe0) == 0xc0) {
int length_size = 1 + ((marker >> 3) & 0x3);
int nb_frames = 1 + (marker & 0x7);
int idx_size = 2 + nb_frames * length_size;
if (in->size >= idx_size && in->data[in->size - idx_size] == marker) {
GetByteContext bc;
int64_t total_size = 0;
bytestream2_init(&bc, in->data + in->size + 1 - idx_size,
nb_frames * length_size);
for (i = 0; i < nb_frames; i++) {
int frame_size = 0;
for (j = 0; j < length_size; j++)
frame_size |= bytestream2_get_byte(&bc) << (j * 8);
total_size += frame_size;
if (frame_size <= 0 || total_size > in->size - idx_size) {
av_log(ctx, AV_LOG_ERROR,
"Invalid frame size in a superframe: %d\n", frame_size);
ret = AVERROR(EINVAL);
goto fail;
}
s->sizes[i] = frame_size;
}
s->nb_frames = nb_frames;
s->next_frame = 0;
s->next_frame_offset = 0;
is_superframe = 1;
}
}
}
if (is_superframe) {
GetBitContext gb;
int profile, invisible = 0;
ret = av_packet_ref(out, s->buffer_pkt);
if (ret < 0)
goto fail;
out->data += s->next_frame_offset;
out->size = s->sizes[s->next_frame];
s->next_frame_offset += out->size;
s->next_frame++;
if (s->next_frame >= s->nb_frames)
av_packet_unref(s->buffer_pkt);
ret = init_get_bits8(&gb, out->data, out->size);
if (ret < 0)
goto fail;
get_bits(&gb, 2); // frame_marker
profile = get_bits1(&gb);
profile |= get_bits1(&gb) << 1;
if (profile == 3)
get_bits1(&gb);
if (!get_bits1(&gb)) {
get_bits1(&gb);
invisible = !get_bits1(&gb);
}
if (invisible)
out->pts = AV_NOPTS_VALUE;
} else {
passthrough:
av_packet_move_ref(out, s->buffer_pkt);
}
return 0;
fail:
if (ret < 0)
av_packet_unref(out);
av_packet_unref(s->buffer_pkt);
return ret;
}
static int vp9_superframe_split_init(AVBSFContext *ctx)
{
VP9SFSplitContext *s = ctx->priv_data;
s->buffer_pkt = av_packet_alloc();
if (!s->buffer_pkt)
return AVERROR(ENOMEM);
return 0;
}
static void vp9_superframe_split_flush(AVBSFContext *ctx)
{
VP9SFSplitContext *s = ctx->priv_data;
av_packet_unref(s->buffer_pkt);
}
static void vp9_superframe_split_uninit(AVBSFContext *ctx)
{
VP9SFSplitContext *s = ctx->priv_data;
av_packet_free(&s->buffer_pkt);
}
const FFBitStreamFilter ff_vp9_superframe_split_bsf = {
.p.name = "vp9_superframe_split",
.p.codec_ids = (const enum AVCodecID []){ AV_CODEC_ID_VP9, AV_CODEC_ID_NONE },
.priv_data_size = sizeof(VP9SFSplitContext),
.init = vp9_superframe_split_init,
.flush = vp9_superframe_split_flush,
.close = vp9_superframe_split_uninit,
.filter = vp9_superframe_split_filter,
};
+328
View File
@@ -0,0 +1,328 @@
/*
* H.266/VVC MP4 to Annex B byte stream format filter
* Copyright (c) 2022, Thomas Siedel
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <string.h>
#include "libavutil/intreadwrite.h"
#include "libavutil/mem.h"
#include "bsf.h"
#include "bsf_internal.h"
#include "bytestream.h"
#include "defs.h"
#include "vvc.h"
#define MIN_VVCC_LENGTH 23
typedef struct VVCBSFContext {
uint8_t length_size;
int extradata_parsed;
} VVCBSFContext;
static int vvc_extradata_to_annexb(AVBSFContext *ctx)
{
GetByteContext gb;
int length_size, num_arrays, i, j;
int ret = 0;
int temp = 0;
int ptl_present;
uint8_t *new_extradata = NULL;
size_t new_extradata_size = 0;
int max_picture_width = 0;
int max_picture_height = 0;
int avg_frame_rate = 0;
bytestream2_init(&gb, ctx->par_in->extradata, ctx->par_in->extradata_size);
temp = bytestream2_get_byte(&gb);
length_size = ((temp & 6) >> 1) + 1;
ptl_present = temp & 1;
if (ptl_present) {
int num_bytes_constraint_info;
int general_profile_idc;
int general_tier_flag;
int general_level_idc;
int ptl_frame_only_constraint_flag;
int ptl_multi_layer_enabled_flag;
int ptl_num_sub_profiles;
int temp3, temp4, temp5;
int temp2 = bytestream2_get_be16(&gb);
int ols_idx = (temp2 >> 7) & 0x1ff;
int num_sublayers = (temp2 >> 4) & 0x7;
int constant_frame_rate = (temp2 >> 2) & 0x3;
int chroma_format_idc = temp2 & 0x3;
int bit_depth_minus8 = (bytestream2_get_byte(&gb) >> 5) & 0x7;
av_log(ctx, AV_LOG_DEBUG,
"bit_depth_minus8 %d chroma_format_idc %d\n", bit_depth_minus8,
chroma_format_idc);
av_log(ctx, AV_LOG_DEBUG, "constant_frame_rate %d, ols_idx %d\n",
constant_frame_rate, ols_idx);
// VvcPTLRecord(num_sublayers) native_ptl
temp3 = bytestream2_get_byte(&gb);
num_bytes_constraint_info = (temp3) & 0x3f;
temp4 = bytestream2_get_byte(&gb);
general_profile_idc = (temp4 >> 1) & 0x7f;
general_tier_flag = (temp4) & 1;
general_level_idc = bytestream2_get_byte(&gb);
av_log(ctx, AV_LOG_DEBUG,
"general_profile_idc %d, general_tier_flag %d, general_level_idc %d, num_sublayers %d num_bytes_constraint_info %d\n",
general_profile_idc, general_tier_flag, general_level_idc,
num_sublayers, num_bytes_constraint_info);
temp5 = bytestream2_get_byte(&gb);
ptl_frame_only_constraint_flag = (temp5 >> 7) & 0x1;
ptl_multi_layer_enabled_flag = (temp5 >> 6) & 0x1;
for (i = 0; i < num_bytes_constraint_info - 1; i++) {
// unsigned int(8*num_bytes_constraint_info - 2) general_constraint_info;
bytestream2_get_byte(&gb);
}
av_log(ctx, AV_LOG_DEBUG,
"ptl_multi_layer_enabled_flag %d, ptl_frame_only_constraint_flag %d\n",
ptl_multi_layer_enabled_flag, ptl_frame_only_constraint_flag);
if (num_sublayers > 1) {
int temp6 = bytestream2_get_byte(&gb);
uint8_t ptl_sublayer_level_present_flag[8] = { 0 };
//uint8_t sublayer_level_idc[8] = {0};
for (i = num_sublayers - 2; i >= 0; i--) {
ptl_sublayer_level_present_flag[i] =
(temp6 >> (7 - (num_sublayers - 2 - i))) & 0x01;
}
// for (j=num_sublayers; j<=8 && num_sublayers > 1; j++)
// bit(1) ptl_reserved_zero_bit = 0;
for (i = num_sublayers - 2; i >= 0; i--) {
if (ptl_sublayer_level_present_flag[i]) {
//sublayer_level_idc[i] = bytestream2_get_byte(&gb);
}
}
}
ptl_num_sub_profiles = bytestream2_get_byte(&gb);
for (j = 0; j < ptl_num_sub_profiles; j++) {
// unsigned int(32) general_sub_profile_idc[j];
bytestream2_get_be16(&gb);
bytestream2_get_be16(&gb);
}
max_picture_width = bytestream2_get_be16(&gb); // unsigned_int(16) max_picture_width;
max_picture_height = bytestream2_get_be16(&gb); // unsigned_int(16) max_picture_height;
avg_frame_rate = bytestream2_get_be16(&gb); // unsigned int(16) avg_frame_rate; }
av_log(ctx, AV_LOG_DEBUG,
"max_picture_width %d, max_picture_height %d, avg_frame_rate %d\n",
max_picture_width, max_picture_height, avg_frame_rate);
}
num_arrays = bytestream2_get_byte(&gb);
for (i = 0; i < num_arrays; i++) {
int cnt;
int type = bytestream2_get_byte(&gb) & 0x1f;
if (type == VVC_OPI_NUT || type == VVC_DCI_NUT)
cnt = 1;
else
cnt = bytestream2_get_be16(&gb);
av_log(ctx, AV_LOG_DEBUG, "nalu_type %d cnt %d\n", type, cnt);
if (!(type == VVC_OPI_NUT || type == VVC_DCI_NUT ||
type == VVC_VPS_NUT || type == VVC_SPS_NUT || type == VVC_PPS_NUT
|| type == VVC_PREFIX_SEI_NUT || type == VVC_SUFFIX_SEI_NUT)) {
av_log(ctx, AV_LOG_ERROR,
"Invalid NAL unit type in extradata: %d\n", type);
ret = AVERROR_INVALIDDATA;
goto fail;
}
for (j = 0; j < cnt; j++) {
const int nalu_len = bytestream2_get_be16(&gb);
if (!nalu_len ||
nalu_len > bytestream2_get_bytes_left(&gb) ||
4 + AV_INPUT_BUFFER_PADDING_SIZE + nalu_len > SIZE_MAX - new_extradata_size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
ret = av_reallocp(&new_extradata, new_extradata_size + nalu_len + 4
+ AV_INPUT_BUFFER_PADDING_SIZE);
if (ret < 0)
goto fail;
AV_WB32(new_extradata + new_extradata_size, 1); // add the startcode
bytestream2_get_buffer(&gb, new_extradata + new_extradata_size + 4,
nalu_len);
new_extradata_size += 4 + nalu_len;
memset(new_extradata + new_extradata_size, 0,
AV_INPUT_BUFFER_PADDING_SIZE);
}
}
av_freep(&ctx->par_out->extradata);
ctx->par_out->extradata = new_extradata;
ctx->par_out->extradata_size = new_extradata_size;
if (!new_extradata_size)
av_log(ctx, AV_LOG_WARNING, "No parameter sets in the extradata\n");
return length_size;
fail:
av_freep(&new_extradata);
return ret;
}
static int vvc_mp4toannexb_init(AVBSFContext *ctx)
{
VVCBSFContext *s = ctx->priv_data;
int ret;
if (ctx->par_in->extradata_size < MIN_VVCC_LENGTH ||
AV_RB24(ctx->par_in->extradata) == 1 ||
AV_RB32(ctx->par_in->extradata) == 1) {
av_log(ctx, AV_LOG_VERBOSE,
"The input looks like it is Annex B already\n");
} else {
ret = vvc_extradata_to_annexb(ctx);
if (ret < 0)
return ret;
s->length_size = ret;
s->extradata_parsed = 1;
}
return 0;
}
static int vvc_mp4toannexb_filter(AVBSFContext *ctx, AVPacket *out)
{
VVCBSFContext *s = ctx->priv_data;
AVPacket *in;
GetByteContext gb;
int is_irap = 0;
int added_extra = 0;
int i, ret = 0;
ret = ff_bsf_get_packet(ctx, &in);
if (ret < 0)
return ret;
if (!s->extradata_parsed) {
av_packet_move_ref(out, in);
av_packet_free(&in);
return 0;
}
bytestream2_init(&gb, in->data, in->size);
/* check if this packet contains an IRAP. The extradata will need to be added before any potential PH_NUT */
while (bytestream2_get_bytes_left(&gb)) {
uint32_t nalu_size = 0;
int nalu_type;
if (bytestream2_get_bytes_left(&gb) < s->length_size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
for (i = 0; i < s->length_size; i++)
nalu_size = (nalu_size << 8) | bytestream2_get_byte(&gb);
if (nalu_size < 2 || nalu_size > bytestream2_get_bytes_left(&gb)) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
nalu_type = (bytestream2_peek_be16(&gb) >> 3) & 0x1f;
is_irap = nalu_type >= VVC_IDR_W_RADL && nalu_type <= VVC_RSV_IRAP_11;
if (is_irap) {
break;
}
bytestream2_seek(&gb, nalu_size, SEEK_CUR);
}
bytestream2_seek(&gb, 0, SEEK_SET);
while (bytestream2_get_bytes_left(&gb)) {
uint32_t nalu_size = 0;
int nalu_type;
int add_extradata, extra_size, prev_size;
if (bytestream2_get_bytes_left(&gb) < s->length_size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
for (i = 0; i < s->length_size; i++)
nalu_size = (nalu_size << 8) | bytestream2_get_byte(&gb);
if (nalu_size < 2 || nalu_size > bytestream2_get_bytes_left(&gb)) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
nalu_type = (bytestream2_peek_be16(&gb) >> 3) & 0x1f;
/* prepend extradata to IRAP frames */
add_extradata = is_irap && nalu_type != VVC_AUD_NUT && !added_extra;
extra_size = add_extradata * ctx->par_out->extradata_size;
added_extra |= add_extradata;
if (FFMIN(INT_MAX, SIZE_MAX) < 4ULL + nalu_size + extra_size) {
ret = AVERROR_INVALIDDATA;
goto fail;
}
prev_size = out->size;
ret = av_grow_packet(out, 4 + nalu_size + extra_size);
if (ret < 0)
goto fail;
if (extra_size)
memcpy(out->data + prev_size, ctx->par_out->extradata, extra_size);
AV_WB32(out->data + prev_size + extra_size, 1);
bytestream2_get_buffer(&gb, out->data + prev_size + 4 + extra_size,
nalu_size);
}
ret = av_packet_copy_props(out, in);
if (ret < 0)
goto fail;
fail:
if (ret < 0)
av_packet_unref(out);
av_packet_free(&in);
return ret;
}
static const enum AVCodecID codec_ids[] = {
AV_CODEC_ID_VVC, AV_CODEC_ID_NONE,
};
const FFBitStreamFilter ff_vvc_mp4toannexb_bsf = {
.p.name = "vvc_mp4toannexb",
.p.codec_ids = codec_ids,
.priv_data_size = sizeof(VVCBSFContext),
.init = vvc_mp4toannexb_init,
.filter = vvc_mp4toannexb_filter,
};