February 2022 - ffmpeg-cvslog

ffmpeg: flush delayed frames in codec copy scenarios
by James Almer 28 Feb '22

28 Feb '22

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Mon Feb 14 14:07:01 2022 -0300| [afe485ee6b3bbcd4b0e106eb9fc4dcf4846db1b8] | committer: James Almer ffmpeg: flush delayed frames in codec copy scenarios Bitstream filters inserted between the input and output were never drained, resulting in packets being lost if the bsf had any buffered. Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=afe485ee6b3bbcd4b… --- fftools/ffmpeg.c | 13 ++++++++----- fftools/ffmpeg.h | 1 + fftools/ffmpeg_filter.c | 1 + fftools/ffmpeg_opt.c | 4 ++++ 4 files changed, 14 insertions(+), 5 deletions(-) diff --git a/fftools/ffmpeg.c b/fftools/ffmpeg.c index 7beea11933..66e5ec8c8e 100644 --- a/fftools/ffmpeg.c +++ b/fftools/ffmpeg.c @@ -2001,7 +2001,7 @@ static int check_output_constraints(InputStream *ist, OutputStream *ost) if (ost->source_index != ist_index) return 0; - if (ost->finished) + if (ost->finished & MUXER_FINISHED) return 0; if (of->start_time != AV_NOPTS_VALUE && ist->pts < of->start_time) @@ -2733,7 +2733,9 @@ static int process_input_packet(InputStream *ist, const AVPacket *pkt, int no_eo } ist->pts = ist->dts; ist->next_pts = ist->next_dts; - } + } else if (!ist->decoding_needed) + eof_reached = 1; + for (i = 0; i < nb_output_streams; i++) { OutputStream *ost = output_streams[i]; @@ -4213,11 +4215,12 @@ static int process_input(int file_index) for (i = 0; i < ifile->nb_streams; i++) { ist = input_streams[ifile->ist_index + i]; avctx = ist->dec_ctx; - if (ist->decoding_needed) { + if (ist->processing_needed) { ret = process_input_packet(ist, NULL, 1); if (ret>0) return 0; - avcodec_flush_buffers(avctx); + if (ist->decoding_needed) + avcodec_flush_buffers(avctx); } } #if HAVE_THREADS @@ -4247,7 +4250,7 @@ static int process_input(int file_index) for (i = 0; i < ifile->nb_streams; i++) { ist = input_streams[ifile->ist_index + i]; - if (ist->decoding_needed) { + if (ist->processing_needed) { ret = process_input_packet(ist, NULL, 0); if (ret>0) return 0; diff --git a/fftools/ffmpeg.h b/fftools/ffmpeg.h index 81ec4d5970..1b8bbace3f 100644 --- a/fftools/ffmpeg.h +++ b/fftools/ffmpeg.h @@ -307,6 +307,7 @@ typedef struct InputStream { int decoding_needed; /* non zero if the packets must be decoded in 'raw_fifo', see DECODING_FOR_* */ #define DECODING_FOR_OST 1 #define DECODING_FOR_FILTER 2 + int processing_needed; /* non zero if the packets must be processed */ AVCodecContext *dec_ctx; const AVCodec *dec; diff --git a/fftools/ffmpeg_filter.c b/fftools/ffmpeg_filter.c index 2c3f21985f..b80d7189db 100644 --- a/fftools/ffmpeg_filter.c +++ b/fftools/ffmpeg_filter.c @@ -277,6 +277,7 @@ static void init_input_filter(FilterGraph *fg, AVFilterInOut *in) ist->discard = 0; ist->decoding_needed |= DECODING_FOR_FILTER; + ist->processing_needed = 1; ist->st->discard = AVDISCARD_NONE; ifilter = ALLOC_ARRAY_ELEM(fg->inputs, fg->nb_inputs); diff --git a/fftools/ffmpeg_opt.c b/fftools/ffmpeg_opt.c index 3102851885..bf1bbcae4c 100644 --- a/fftools/ffmpeg_opt.c +++ b/fftools/ffmpeg_opt.c @@ -2610,6 +2610,7 @@ loop_end: if (ost->encoding_needed && ost->source_index >= 0) { InputStream *ist = input_streams[ost->source_index]; ist->decoding_needed |= DECODING_FOR_OST; + ist->processing_needed = 1; if (ost->st->codecpar->codec_type == AVMEDIA_TYPE_VIDEO || ost->st->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) { @@ -2622,6 +2623,9 @@ loop_end: exit_program(1); } } + } else if (ost->stream_copy && ost->source_index >= 0) { + InputStream *ist = input_streams[ost->source_index]; + ist->processing_needed = 1; } /* set the filter output constraints */

1 0

tools/target_bsf_fuzzer: simplify the loop feeding packets to the filter
by James Almer 28 Feb '22

28 Feb '22

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Wed Feb 23 21:38:30 2022 -0300| [4f21a9ae6b86cc481af4f92249bef877bdf64fdc] | committer: James Almer tools/target_bsf_fuzzer: simplify the loop feeding packets to the filter And use a single AVPacket for the entire process. This more closely follows the suggested API usage in the doxy. Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=4f21a9ae6b86cc481… --- tools/target_bsf_fuzzer.c | 41 ++++++++++++++++------------------------- 1 file changed, 16 insertions(+), 25 deletions(-) diff --git a/tools/target_bsf_fuzzer.c b/tools/target_bsf_fuzzer.c index d6aaee3bd9..8ef9932690 100644 --- a/tools/target_bsf_fuzzer.c +++ b/tools/target_bsf_fuzzer.c @@ -43,7 +43,7 @@ int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { const uint8_t *last = data; const uint8_t *end = data + size; AVBSFContext *bsf = NULL; - AVPacket *in, *out; + AVPacket *pkt; uint64_t keyframes = 0; uint64_t flushpattern = -1; int res; @@ -118,9 +118,8 @@ int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { return 0; // Failure of av_bsf_init() does not imply that a issue was found } - in = av_packet_alloc(); - out = av_packet_alloc(); - if (!in || !out) + pkt = av_packet_alloc(); + if (!pkt) error("Failed memory allocation"); while (data < end) { @@ -133,11 +132,11 @@ int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { if (data + sizeof(fuzz_tag) > end) data = end; - res = av_new_packet(in, data - last); + res = av_new_packet(pkt, data - last); if (res < 0) error("Failed memory allocation"); - memcpy(in->data, last, data - last); - in->flags = (keyframes & 1) * AV_PKT_FLAG_DISCARD + (!!(keyframes & 2)) * AV_PKT_FLAG_KEY; + memcpy(pkt->data, last, data - last); + pkt->flags = (keyframes & 1) * AV_PKT_FLAG_DISCARD + (!!(keyframes & 2)) * AV_PKT_FLAG_KEY; keyframes = (keyframes >> 2) + (keyframes<<62); data += sizeof(fuzz_tag); last = data; @@ -146,28 +145,20 @@ int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { av_bsf_flush(bsf); flushpattern = (flushpattern >> 3) + (flushpattern << 61); - while (in->size) { - res = av_bsf_send_packet(bsf, in); - if (res < 0 && res != AVERROR(EAGAIN)) - break; - res = av_bsf_receive_packet(bsf, out); - if (res < 0) - break; - av_packet_unref(out); + res = av_bsf_send_packet(bsf, pkt); + if (res < 0) { + av_packet_unref(pkt); + continue; } - av_packet_unref(in); + while (av_bsf_receive_packet(bsf, pkt) >= 0) + av_packet_unref(pkt); } - res = av_bsf_send_packet(bsf, NULL); - while (!res) { - res = av_bsf_receive_packet(bsf, out); - if (res < 0) - break; - av_packet_unref(out); - } + av_bsf_send_packet(bsf, NULL); + while (av_bsf_receive_packet(bsf, pkt) >= 0) + av_packet_unref(pkt); - av_packet_free(&in); - av_packet_free(&out); + av_packet_free(&pkt); av_bsf_free(&bsf); return 0; }

1 0

avformat/av1: reduce the scope of some variables
by James Almer 28 Feb '22

28 Feb '22

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Sat Feb 26 20:43:59 2022 -0300| [7a4840a8cabccd1ae000ae2e41d62d45cb95f68d] | committer: James Almer avformat/av1: reduce the scope of some variables Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7a4840a8cabccd1ae… --- libavformat/av1.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/libavformat/av1.c b/libavformat/av1.c index 7771c09b4a..7caea0c377 100644 --- a/libavformat/av1.c +++ b/libavformat/av1.c @@ -33,8 +33,7 @@ static int av1_filter_obus(AVIOContext *pb, const uint8_t *buf, int size, int *offset) { const uint8_t *start = buf, *end = buf + size; - int64_t obu_size; - int off, start_pos, type, temporal_id, spatial_id; + int off; enum { START_NOT_FOUND, START_FOUND, @@ -44,6 +43,8 @@ static int av1_filter_obus(AVIOContext *pb, const uint8_t *buf, off = size = 0; while (buf < end) { + int64_t obu_size; + int start_pos, type, temporal_id, spatial_id; int len = parse_obu_header(buf, end - buf, &obu_size, &start_pos, &type, &temporal_id, &spatial_id); if (len < 0) @@ -333,8 +334,6 @@ static int parse_sequence_header(AV1SequenceParameters *seq_params, const uint8_ int ff_av1_parse_seq_header(AV1SequenceParameters *seq, const uint8_t *buf, int size) { - int64_t obu_size; - int start_pos, type, temporal_id, spatial_id; int is_av1c; if (size <= 0) @@ -373,6 +372,8 @@ int ff_av1_parse_seq_header(AV1SequenceParameters *seq, const uint8_t *buf, int } while (size > 0) { + int64_t obu_size; + int start_pos, type, temporal_id, spatial_id; int len = parse_obu_header(buf, size, &obu_size, &start_pos, &type, &temporal_id, &spatial_id); if (len < 0) @@ -401,8 +402,6 @@ int ff_isom_write_av1c(AVIOContext *pb, const uint8_t *buf, int size) PutBitContext pbc; uint8_t header[4], *meta; const uint8_t *seq; - int64_t obu_size; - int start_pos, type, temporal_id, spatial_id; int ret, nb_seq = 0, seq_size, meta_size; if (size <= 0) @@ -427,6 +426,8 @@ int ff_isom_write_av1c(AVIOContext *pb, const uint8_t *buf, int size) return ret; while (size > 0) { + int64_t obu_size; + int start_pos, type, temporal_id, spatial_id; int len = parse_obu_header(buf, size, &obu_size, &start_pos, &type, &temporal_id, &spatial_id); if (len < 0) {

1 0

avformat/av1: support av1C extradata in ff_av1_parse_seq_header()
by James Almer 28 Feb '22

28 Feb '22

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Sat Feb 26 20:30:45 2022 -0300| [76e10325fc65b816193b4acb9bfda174262ff6ba] | committer: James Almer avformat/av1: support av1C extradata in ff_av1_parse_seq_header() Fixes dash manifest creation for av1 streams with av1C formatted extradata. Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=76e10325fc65b8161… --- libavformat/av1.c | 35 ++++++++++++++++++++++++++++++++++- 1 file changed, 34 insertions(+), 1 deletion(-) diff --git a/libavformat/av1.c b/libavformat/av1.c index 1fcfac2356..7771c09b4a 100644 --- a/libavformat/av1.c +++ b/libavformat/av1.c @@ -335,10 +335,43 @@ int ff_av1_parse_seq_header(AV1SequenceParameters *seq, const uint8_t *buf, int { int64_t obu_size; int start_pos, type, temporal_id, spatial_id; + int is_av1c; if (size <= 0) return AVERROR_INVALIDDATA; + is_av1c = !!(buf[0] & 0x80); + if (is_av1c) { + GetBitContext gb; + int ret, version = buf[0] & 0x7F; + + if (version != 1 || size < 4) + return AVERROR_INVALIDDATA; + + ret = init_get_bits8(&gb, buf, 4); + if (ret < 0) + return ret; + + memset(seq, 0, sizeof(*seq)); + + skip_bits(&gb, 8); + seq->profile = get_bits(&gb, 3); + seq->level = get_bits(&gb, 5); + seq->tier = get_bits(&gb, 1); + seq->bitdepth = get_bits(&gb, 1) * 2 + 8; + seq->bitdepth += get_bits(&gb, 1) * 2; + seq->monochrome = get_bits(&gb, 1); + seq->chroma_subsampling_x = get_bits(&gb, 1); + seq->chroma_subsampling_y = get_bits(&gb, 1); + seq->chroma_sample_position = get_bits(&gb, 2); + seq->color_primaries = AVCOL_PRI_UNSPECIFIED; + seq->transfer_characteristics = AVCOL_TRC_UNSPECIFIED; + seq->matrix_coefficients = AVCOL_SPC_UNSPECIFIED; + + size -= 4; + buf += 4; + } + while (size > 0) { int len = parse_obu_header(buf, size, &obu_size, &start_pos, &type, &temporal_id, &spatial_id); @@ -358,7 +391,7 @@ int ff_av1_parse_seq_header(AV1SequenceParameters *seq, const uint8_t *buf, int buf += len; } - return AVERROR_INVALIDDATA; + return is_av1c ? 0 : AVERROR_INVALIDDATA; } int ff_isom_write_av1c(AVIOContext *pb, const uint8_t *buf, int size)

1 0

avformat: always compile isom_tags
by James Almer 28 Feb '22

28 Feb '22

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Sat Feb 26 11:32:41 2022 -0300| [b6c7f82db75d2c77df2312b92a048a9ae0c5e86a] | committer: James Almer avformat: always compile isom_tags It's needed for avformat_get_mov_video_tags() and avformat_get_mov_audio_tags(), both public symbols defined in avformat.h Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=b6c7f82db75d2c77d… --- libavformat/Makefile | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/libavformat/Makefile b/libavformat/Makefile index 6566e40cac..16d019df4e 100644 --- a/libavformat/Makefile +++ b/libavformat/Makefile @@ -13,6 +13,7 @@ OBJS = allformats.o \ format.o \ id3v1.o \ id3v2.o \ + isom_tags.o \ metadata.o \ mux.o \ options.o \ @@ -27,7 +28,7 @@ OBJS = allformats.o \ OBJS-$(HAVE_LIBC_MSVCRT) += file_open.o # subsystems -OBJS-$(CONFIG_ISO_MEDIA) += isom.o isom_tags.o +OBJS-$(CONFIG_ISO_MEDIA) += isom.o OBJS-$(CONFIG_NETWORK) += network.o OBJS-$(CONFIG_RIFFDEC) += riffdec.o OBJS-$(CONFIG_RIFFENC) += riffenc.o @@ -121,7 +122,7 @@ OBJS-$(CONFIG_AST_DEMUXER) += ast.o astdec.o OBJS-$(CONFIG_AST_MUXER) += ast.o astenc.o OBJS-$(CONFIG_AU_DEMUXER) += au.o pcm.o OBJS-$(CONFIG_AU_MUXER) += au.o rawenc.o -OBJS-$(CONFIG_AVI_DEMUXER) += avidec.o isom_tags.o +OBJS-$(CONFIG_AVI_DEMUXER) += avidec.o OBJS-$(CONFIG_AVI_MUXER) += avienc.o mpegtsenc.o avlanguage.o rawutils.o OBJS-$(CONFIG_AVM2_MUXER) += swfenc.o swf.o OBJS-$(CONFIG_AVR_DEMUXER) += avr.o pcm.o @@ -316,11 +317,11 @@ OBJS-$(CONFIG_LXF_DEMUXER) += lxfdec.o OBJS-$(CONFIG_M4V_DEMUXER) += m4vdec.o rawdec.o OBJS-$(CONFIG_M4V_MUXER) += rawenc.o OBJS-$(CONFIG_MATROSKA_DEMUXER) += matroskadec.o matroska.o \ - flac_picture.o isom_tags.o rmsipr.o \ + flac_picture.o rmsipr.o \ oggparsevorbis.o vorbiscomment.o \ qtpalette.o replaygain.o dovi_isom.o OBJS-$(CONFIG_MATROSKA_MUXER) += matroskaenc.o matroska.o \ - av1.o avc.o hevc.o isom_tags.o \ + av1.o avc.o hevc.o \ flacenc_header.o avlanguage.o \ vorbiscomment.o wv.o dovi_isom.o OBJS-$(CONFIG_MCA_DEMUXER) += mca.o @@ -382,7 +383,7 @@ OBJS-$(CONFIG_NISTSPHERE_DEMUXER) += nistspheredec.o pcm.o OBJS-$(CONFIG_NSP_DEMUXER) += nspdec.o pcm.o OBJS-$(CONFIG_NSV_DEMUXER) += nsvdec.o OBJS-$(CONFIG_NULL_MUXER) += nullenc.o -OBJS-$(CONFIG_NUT_DEMUXER) += nutdec.o nut.o isom_tags.o +OBJS-$(CONFIG_NUT_DEMUXER) += nutdec.o nut.o OBJS-$(CONFIG_NUT_MUXER) += nutenc.o nut.o OBJS-$(CONFIG_NUV_DEMUXER) += nuv.o OBJS-$(CONFIG_AV1_DEMUXER) += av1dec.o

1 0

qsvenc: avoid dereferencing the null pointer
by Tong Wu 28 Feb '22

28 Feb '22

ffmpeg | branch: master | Tong Wu <tong1.wu-at-intel.com(a)ffmpeg.org> | Thu Feb 24 10:27:40 2022 +0800| [a0a2ccd55d2038867b0677c31f70632a69f6e9e3] | committer: Haihao Xiang qsvenc: avoid dereferencing the null pointer The variable AVFrame *frame could be a null pointer, now add a null pointer check to avoid dereferencing the null pointer. Signed-off-by: Tong Wu <tong1.wu(a)intel.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a0a2ccd55d2038867… --- libavcodec/qsvenc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libavcodec/qsvenc.c b/libavcodec/qsvenc.c index 49f61ae6bb..9b71487666 100644 --- a/libavcodec/qsvenc.c +++ b/libavcodec/qsvenc.c @@ -1736,7 +1736,7 @@ static int encode_frame(AVCodecContext *avctx, QSVEncContext *q, goto free; } - if (ret == MFX_WRN_INCOMPATIBLE_VIDEO_PARAM && frame->interlaced_frame) + if (ret == MFX_WRN_INCOMPATIBLE_VIDEO_PARAM && frame && frame->interlaced_frame) print_interlace_msg(avctx, q); ret = 0;

1 0

qsv: add return value check for MFXQueryIMPL
by Tong Wu 28 Feb '22

28 Feb '22

ffmpeg | branch: master | Tong Wu <tong1.wu-at-intel.com(a)ffmpeg.org> | Thu Feb 24 10:27:39 2022 +0800| [d05ca3d779aed2df3d1cd74d1d295eb3729b7bf2] | committer: Haihao Xiang qsv: add return value check for MFXQueryIMPL add a return value check for function MFXQueryIMPL to handle the error message. Signed-off-by: Tong Wu <tong1.wu(a)intel.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d05ca3d779aed2df3… --- libavcodec/qsv.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/libavcodec/qsv.c b/libavcodec/qsv.c index b75877e698..67d0e3934a 100644 --- a/libavcodec/qsv.c +++ b/libavcodec/qsv.c @@ -424,7 +424,10 @@ int ff_qsv_init_internal_session(AVCodecContext *avctx, QSVSession *qs, return ret; } - MFXQueryIMPL(qs->session, &impl); + ret = MFXQueryIMPL(qs->session, &impl); + if (ret != MFX_ERR_NONE) + return ff_qsv_print_error(avctx, ret, + "Error querying the session attributes"); switch (MFX_IMPL_BASETYPE(impl)) { case MFX_IMPL_SOFTWARE:

1 0

libavcodec/qsvenc: add mbbrc to hevc_qsv
by Wenbin Chen 28 Feb '22

28 Feb '22

ffmpeg | branch: master | Wenbin Chen <wenbin.chen-at-intel.com(a)ffmpeg.org> | Mon Feb 21 16:56:56 2022 +0800| [342d4fb0563cafa4e444c83dc4ecfa0d69a71eda] | committer: Haihao Xiang libavcodec/qsvenc: add mbbrc to hevc_qsv Add mbbrc to hevc_qsv For detailed description, please see "mbbrc" part in: https://github.com/Intel-Media-SDK/MediaSDK/blob/master/doc/mediasdk-man.md… Signed-off-by: Wenbin Chen <wenbin.chen(a)intel.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=342d4fb0563cafa4e… --- doc/encoders.texi | 5 +++++ libavcodec/qsvenc.c | 5 +++-- 2 files changed, 8 insertions(+), 2 deletions(-) diff --git a/doc/encoders.texi b/doc/encoders.texi index 64e8c5d129..1bd38671ca 100644 --- a/doc/encoders.texi +++ b/doc/encoders.texi @@ -3395,6 +3395,11 @@ Maximum encoded frame size in bytes. @item @var{max_slice_size} Maximum encoded slice size in bytes. +@item @var{mbbrc} +Setting this flag enables macroblock level bitrate control that generally +improves subjective visual quality. Enabling this flag may have negative impact +on performance and objective visual quality metric. + @item @var{p_strategy} Enable P-pyramid: 0-default 1-simple 2-pyramid(bf need to be set to 0). diff --git a/libavcodec/qsvenc.c b/libavcodec/qsvenc.c index 3008f329e5..49f61ae6bb 100644 --- a/libavcodec/qsvenc.c +++ b/libavcodec/qsvenc.c @@ -878,8 +878,6 @@ static int init_video_param(AVCodecContext *avctx, QSVEncContext *q) if (avctx->codec_id == AV_CODEC_ID_H264) { if (q->bitrate_limit >= 0) q->extco2.BitrateLimit = q->bitrate_limit ? MFX_CODINGOPTION_ON : MFX_CODINGOPTION_OFF; - if (q->mbbrc >= 0) - q->extco2.MBBRC = q->mbbrc ? MFX_CODINGOPTION_ON : MFX_CODINGOPTION_OFF; #if QSV_HAVE_TRELLIS if (avctx->trellis >= 0) @@ -936,6 +934,9 @@ static int init_video_param(AVCodecContext *avctx, QSVEncContext *q) q->extco2.MaxQPP = q->extco2.MaxQPB = q->extco2.MaxQPI; } #endif + if (q->mbbrc >= 0) + q->extco2.MBBRC = q->mbbrc ? MFX_CODINGOPTION_ON : MFX_CODINGOPTION_OFF; + q->extco2.Header.BufferId = MFX_EXTBUFF_CODING_OPTION2; q->extco2.Header.BufferSz = sizeof(q->extco2);

1 0

libavcodec/qsvenc: add more ChromaFormat support to mjpeg_qsv
by Wenbin Chen 28 Feb '22

28 Feb '22

ffmpeg | branch: master | Wenbin Chen <wenbin.chen-at-intel.com(a)ffmpeg.org> | Mon Feb 21 15:28:33 2022 +0800| [4869ccb3f38a606401a1c3e3a31f18d35eb7dc39] | committer: Haihao Xiang libavcodec/qsvenc: add more ChromaFormat support to mjpeg_qsv ChromaForamt for mjpeg-qsv is always set to yuv420, and this will be wrong when encode other pixel format (for example yuyv422). ChromaFormat is changed to be adaptive to pix_fmt. Signed-off-by: Wenbin Chen <wenbin.chen(a)intel.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=4869ccb3f38a60640… --- libavcodec/qsvenc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/libavcodec/qsvenc.c b/libavcodec/qsvenc.c index 40d60cde3c..3008f329e5 100644 --- a/libavcodec/qsvenc.c +++ b/libavcodec/qsvenc.c @@ -616,7 +616,8 @@ static int init_video_param_jpeg(AVCodecContext *avctx, QSVEncContext *q) q->param.mfx.FrameInfo.CropH = avctx->height; q->param.mfx.FrameInfo.AspectRatioW = avctx->sample_aspect_ratio.num; q->param.mfx.FrameInfo.AspectRatioH = avctx->sample_aspect_ratio.den; - q->param.mfx.FrameInfo.ChromaFormat = MFX_CHROMAFORMAT_YUV420; + q->param.mfx.FrameInfo.ChromaFormat = MFX_CHROMAFORMAT_YUV420 + + !desc->log2_chroma_w + !desc->log2_chroma_h; q->param.mfx.FrameInfo.BitDepthLuma = desc->comp[0].depth; q->param.mfx.FrameInfo.BitDepthChroma = desc->comp[0].depth; q->param.mfx.FrameInfo.Shift = desc->comp[0].depth > 8;

1 0

libavcodec/vaapi_encode: Add async_depth to vaapi_encoder to increase performance
by Wenbin Chen 28 Feb '22

28 Feb '22

ffmpeg | branch: master | Wenbin Chen <wenbin.chen-at-intel.com(a)ffmpeg.org> | Fri Feb 18 11:07:47 2022 +0800| [d165ce22a4a7cc4ed60238ce8f3d5dcbbad3e266] | committer: Haihao Xiang libavcodec/vaapi_encode: Add async_depth to vaapi_encoder to increase performance Fix: #7706. After commit 5fdcf85bbffe7451c2, vaapi encoder's performance decrease. The reason is that vaRenderPicture() and vaSyncBuffer() are called at the same time (vaRenderPicture() always followed by a vaSyncBuffer()). Now I changed them to be called in a asynchronous way, which will make better use of hardware. Async_depth is added to increase encoder's performance. The frames that are sent to hardware are stored in a fifo. Encoder will sync output after async fifo is full. Signed-off-by: Wenbin Chen <wenbin.chen(a)intel.com> Signed-off-by: Haihao Xiang <haihao.xiang(a)intel.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d165ce22a4a7cc4ed… --- doc/encoders.texi | 6 +++++ libavcodec/vaapi_encode.c | 64 +++++++++++++++++++++++++++++++++++++---------- libavcodec/vaapi_encode.h | 16 ++++++++++-- 3 files changed, 71 insertions(+), 15 deletions(-) diff --git a/doc/encoders.texi b/doc/encoders.texi index 9afb920f69..64e8c5d129 100644 --- a/doc/encoders.texi +++ b/doc/encoders.texi @@ -3591,6 +3591,12 @@ will refer only to P- or I-frames. When set to greater values multiple layers of B-frames will be present, frames in each layer only referring to frames in higher layers. +@item async_depth +Maximum processing parallelism. Increase this to improve single channel +performance. This option doesn't work if driver doesn't implement vaSyncBuffer +function. Please make sure there are enough hw_frames allocated if a large +number of async_depth is used. + @item rc_mode Set the rate control mode to use. A given driver may only support a subset of modes. diff --git a/libavcodec/vaapi_encode.c b/libavcodec/vaapi_encode.c index 3f8c8ace2a..8c6e881702 100644 --- a/libavcodec/vaapi_encode.c +++ b/libavcodec/vaapi_encode.c @@ -965,8 +965,10 @@ static int vaapi_encode_pick_next(AVCodecContext *avctx, if (!pic && ctx->end_of_stream) { --b_counter; pic = ctx->pic_end; - if (pic->encode_issued) + if (pic->encode_complete) return AVERROR_EOF; + else if (pic->encode_issued) + return AVERROR(EAGAIN); } if (!pic) { @@ -1137,7 +1139,8 @@ static int vaapi_encode_send_frame(AVCodecContext *avctx, AVFrame *frame) if (ctx->input_order == ctx->decode_delay) ctx->dts_pts_diff = pic->pts - ctx->first_pts; if (ctx->output_delay > 0) - ctx->ts_ring[ctx->input_order % (3 * ctx->output_delay)] = pic->pts; + ctx->ts_ring[ctx->input_order % + (3 * ctx->output_delay + ctx->async_depth)] = pic->pts; pic->display_order = ctx->input_order; ++ctx->input_order; @@ -1191,18 +1194,47 @@ int ff_vaapi_encode_receive_packet(AVCodecContext *avctx, AVPacket *pkt) return AVERROR(EAGAIN); } - pic = NULL; - err = vaapi_encode_pick_next(avctx, &pic); - if (err < 0) - return err; - av_assert0(pic); + if (ctx->has_sync_buffer_func) { + pic = NULL; + + if (av_fifo_can_write(ctx->encode_fifo)) { + err = vaapi_encode_pick_next(avctx, &pic); + if (!err) { + av_assert0(pic); + pic->encode_order = ctx->encode_order + + av_fifo_can_read(ctx->encode_fifo); + err = vaapi_encode_issue(avctx, pic); + if (err < 0) { + av_log(avctx, AV_LOG_ERROR, "Encode failed: %d.\n", err); + return err; + } + av_fifo_write(ctx->encode_fifo, &pic, 1); + } + } - pic->encode_order = ctx->encode_order++; + if (!av_fifo_can_read(ctx->encode_fifo)) + return err; - err = vaapi_encode_issue(avctx, pic); - if (err < 0) { - av_log(avctx, AV_LOG_ERROR, "Encode failed: %d.\n", err); - return err; + // More frames can be buffered + if (av_fifo_can_write(ctx->encode_fifo) && !ctx->end_of_stream) + return AVERROR(EAGAIN); + + av_fifo_read(ctx->encode_fifo, &pic, 1); + ctx->encode_order = pic->encode_order + 1; + } else { + pic = NULL; + err = vaapi_encode_pick_next(avctx, &pic); + if (err < 0) + return err; + av_assert0(pic); + + pic->encode_order = ctx->encode_order++; + + err = vaapi_encode_issue(avctx, pic); + if (err < 0) { + av_log(avctx, AV_LOG_ERROR, "Encode failed: %d.\n", err); + return err; + } } err = vaapi_encode_output(avctx, pic, pkt); @@ -1220,7 +1252,7 @@ int ff_vaapi_encode_receive_packet(AVCodecContext *avctx, AVPacket *pkt) pkt->dts = ctx->ts_ring[pic->encode_order] - ctx->dts_pts_diff; } else { pkt->dts = ctx->ts_ring[(pic->encode_order - ctx->decode_delay) % - (3 * ctx->output_delay)]; + (3 * ctx->output_delay + ctx->async_depth)]; } av_log(avctx, AV_LOG_DEBUG, "Output packet: pts %"PRId64" dts %"PRId64".\n", pkt->pts, pkt->dts); @@ -2541,6 +2573,11 @@ av_cold int ff_vaapi_encode_init(AVCodecContext *avctx) vas = vaSyncBuffer(ctx->hwctx->display, VA_INVALID_ID, 0); if (vas != VA_STATUS_ERROR_UNIMPLEMENTED) { ctx->has_sync_buffer_func = 1; + ctx->encode_fifo = av_fifo_alloc2(ctx->async_depth, + sizeof(VAAPIEncodePicture *), + 0); + if (!ctx->encode_fifo) + return AVERROR(ENOMEM); } #endif @@ -2581,6 +2618,7 @@ av_cold int ff_vaapi_encode_close(AVCodecContext *avctx) av_freep(&ctx->codec_sequence_params); av_freep(&ctx->codec_picture_params); + av_fifo_freep2(&ctx->encode_fifo); av_buffer_unref(&ctx->recon_frames_ref); av_buffer_unref(&ctx->input_frames_ref); diff --git a/libavcodec/vaapi_encode.h b/libavcodec/vaapi_encode.h index 29d9e9b91c..1b40819c69 100644 --- a/libavcodec/vaapi_encode.h +++ b/libavcodec/vaapi_encode.h @@ -29,6 +29,7 @@ #include "libavutil/hwcontext.h" #include "libavutil/hwcontext_vaapi.h" +#include "libavutil/fifo.h" #include "avcodec.h" #include "hwconfig.h" @@ -47,6 +48,7 @@ enum { MAX_TILE_ROWS = 22, // A.4.1: table A.6 allows at most 20 tile columns for any level. MAX_TILE_COLS = 20, + MAX_ASYNC_DEPTH = 64, }; extern const AVCodecHWConfigInternal *const ff_vaapi_encode_hw_configs[]; @@ -297,7 +299,8 @@ typedef struct VAAPIEncodeContext { // Timestamp handling. int64_t first_pts; int64_t dts_pts_diff; - int64_t ts_ring[MAX_REORDER_DELAY * 3]; + int64_t ts_ring[MAX_REORDER_DELAY * 3 + + MAX_ASYNC_DEPTH]; // Slice structure. int slice_block_rows; @@ -348,6 +351,10 @@ typedef struct VAAPIEncodeContext { // Whether the driver support vaSyncBuffer int has_sync_buffer_func; + // Store buffered pic + AVFifo *encode_fifo; + // Max number of frame buffered in encoder. + int async_depth; } VAAPIEncodeContext; enum { @@ -458,7 +465,12 @@ int ff_vaapi_encode_close(AVCodecContext *avctx); { "b_depth", \ "Maximum B-frame reference depth", \ OFFSET(common.desired_b_depth), AV_OPT_TYPE_INT, \ - { .i64 = 1 }, 1, INT_MAX, FLAGS } + { .i64 = 1 }, 1, INT_MAX, FLAGS }, \ + { "async_depth", "Maximum processing parallelism. " \ + "Increase this to improve single channel performance. This option " \ + "doesn't work if driver doesn't implement vaSyncBuffer function.", \ + OFFSET(common.async_depth), AV_OPT_TYPE_INT, \ + { .i64 = 2 }, 1, MAX_ASYNC_DEPTH, FLAGS } #define VAAPI_ENCODE_RC_MODE(name, desc) \ { #name, desc, 0, AV_OPT_TYPE_CONST, { .i64 = RC_MODE_ ## name }, \

1 0