February 2021 - ffmpeg-devel - lists.ffmpeg.org

[PATCH 00/27] Metadata handling in CBS
by Mark Thompson 27 May '21

27 May '21

Happy New Year! The fun part here is: AVMasteringDisplayMetadata mdm = { ... }; AVContentLightMetadata clm = { ... }; ff_cbs_insert_metadata(cbc, fragment, CBS_METADATA_MASTERING_DISPLAY, &mdm); ff_cbs_insert_metadata(cbc, fragment, CBS_METADATA_CONTENT_LIGHT_LEVEL, &clm); which does whatever is needed to splice that metadata into the given bitstream fragment. It's implemented for H.264, H.265 and AV1 here. Extension to the common parts of H.266 is trivial (just add to the type table). The metadata bitstream filters are extended to support that as well from side data. That's only library-usable, though - there is no ffmpeg-the-program interface for including the side data in the first place. (From discussion on IRC a few days ago there was interest in adding ffmpeg support for that along with automatic BSF insertion?) Not particularly well tested, since much of this was written in the last few days (though some parts like the common BSF implementation have been sent before). - Mark Mark Thompson (27): cbs_h2645: Merge SEI messages in common between codecs h264_metadata_bsf: Move SEI user data parsing to init time cbs_h2645: Merge SEI message handling in common between codecs cbs_sei: Remove restrictions on MDCV values h264_metadata_bsf: Refactor the filter function into smaller parts h264_metadata_bsf: Improve interpretation of input display matrices cbs: Implement common parts of cbs-based bitstream filters separately h264_metadata_bsf: Use common cbs bsf implementation h265_metadata_bsf: Use common cbs bsf implementation mpeg2_metadata_bsf: Use common cbs bsf implementation vp9_metadata_bsf: Use common cbs bsf implementation av1_metadata_bsf: Use common cbs bsf implementation h264_redundant_pps_bsf: Use common cbs bsf implementation cbs: Add interface for manipulating metadata in fragments cbs_sei: Implement metadata manipulation for SEI codecs cbs_sei: Implement fill and extract for HDR SEI messages cbs_h264: Implement fill and extract for display orientation SEI message cbs_av1: Implement metadata manipulation cbs_av1: Support manipulating HDR CLL and MDCV metadata cbs: Add interface to allow BSFs easy access to metadata manipulation h264_metadata_bsf: Simplify display orientation handling h264_metadata_bsf: Support HDR metadata h265_metadata_bsf: Support HDR metadata av1_metadata_bsf: Support HDR metadata vaapi_encode: Unify SEI option mask enum vaapi_encode_h265: Ensure that HDR metadata is included with seek points vaapi_encode_h264: Support HDR metadata doc/bitstream_filters.texi | 51 +++ libavcodec/Makefile | 6 +- libavcodec/av1_metadata_bsf.c | 198 ++------- libavcodec/cbs_av1.c | 233 +++++++++++ libavcodec/cbs_bsf.c | 234 +++++++++++ libavcodec/cbs_bsf.h | 106 +++++ libavcodec/cbs_h264.h | 79 +--- libavcodec/cbs_h2645.c | 494 ++++++++++++++++------ libavcodec/cbs_h264_syntax_template.c | 271 +----------- libavcodec/cbs_h265.h | 66 +-- libavcodec/cbs_h265_syntax_template.c | 370 ++--------------- libavcodec/cbs_internal.h | 17 + libavcodec/cbs_metadata.c | 107 +++++ libavcodec/cbs_metadata.h | 94 +++++ libavcodec/cbs_sei.c | 469 +++++++++++++++++++++ libavcodec/cbs_sei.h | 350 ++++++++++++++++ libavcodec/cbs_sei_syntax_template.c | 307 ++++++++++++++ libavcodec/h264_metadata_bsf.c | 569 ++++++++++---------------- libavcodec/h264_redundant_pps_bsf.c | 99 +---- libavcodec/h265_metadata_bsf.c | 213 +++------- libavcodec/mpeg2_metadata_bsf.c | 94 +---- libavcodec/vaapi_encode.h | 8 + libavcodec/vaapi_encode_h264.c | 125 ++++-- libavcodec/vaapi_encode_h265.c | 123 ++---- libavcodec/vp9_metadata_bsf.c | 61 +-- 25 files changed, 2859 insertions(+), 1885 deletions(-) create mode 100644 libavcodec/cbs_bsf.c create mode 100644 libavcodec/cbs_bsf.h create mode 100644 libavcodec/cbs_metadata.c create mode 100644 libavcodec/cbs_metadata.h create mode 100644 libavcodec/cbs_sei.c create mode 100644 libavcodec/cbs_sei.h create mode 100644 libavcodec/cbs_sei_syntax_template.c -- 2.29.2

5 39

[PATCH 1/2] examples/avio_dir_cmd: drop support for move/delete operations
by Anton Khirnov 22 May '21

22 May '21

They use non-public functions, which is unacceptable for a public API example. Rename the example back to avio_list_dir. This effectively reverts c84d208c275d6a43b3c3421d38772179abf8acee and 767d780ec001167b2fd8f6cfe4ef78a3a8b1e34c. --- configure | 4 +- doc/examples/.gitignore | 2 +- doc/examples/Makefile | 2 +- doc/examples/Makefile.example | 2 +- .../{avio_dir_cmd.c => avio_list_dir.c} | 56 ++----------------- 5 files changed, 9 insertions(+), 57 deletions(-) rename doc/examples/{avio_dir_cmd.c => avio_list_dir.c} (69%) diff --git a/configure b/configure index 46f2038627..65a5a30f36 100755 --- a/configure +++ b/configure @@ -1664,7 +1664,7 @@ COMPONENT_LIST=" " EXAMPLE_LIST=" - avio_dir_cmd_example + avio_list_dir_example avio_reading_example decode_audio_example decode_video_example @@ -3599,7 +3599,7 @@ yadif_cuda_filter_deps="ffnvcodec" yadif_cuda_filter_deps_any="cuda_nvcc cuda_llvm" # examples -avio_dir_cmd_deps="avformat avutil" +avio_list_dir_deps="avformat avutil" avio_reading_deps="avformat avcodec avutil" decode_audio_example_deps="avcodec avutil" decode_video_example_deps="avcodec avutil" diff --git a/doc/examples/.gitignore b/doc/examples/.gitignore index 75152cb50b..44960e1de7 100644 --- a/doc/examples/.gitignore +++ b/doc/examples/.gitignore @@ -1,4 +1,4 @@ -/avio_dir_cmd +/avio_list_dir /avio_reading /decode_audio /decode_video diff --git a/doc/examples/Makefile b/doc/examples/Makefile index 2935424e54..81bfd34d5d 100644 --- a/doc/examples/Makefile +++ b/doc/examples/Makefile @@ -1,4 +1,4 @@ -EXAMPLES-$(CONFIG_AVIO_DIR_CMD_EXAMPLE) += avio_dir_cmd +EXAMPLES-$(CONFIG_AVIO_LIST_DIR_EXAMPLE) += avio_list_dir EXAMPLES-$(CONFIG_AVIO_READING_EXAMPLE) += avio_reading EXAMPLES-$(CONFIG_DECODE_AUDIO_EXAMPLE) += decode_audio EXAMPLES-$(CONFIG_DECODE_VIDEO_EXAMPLE) += decode_video diff --git a/doc/examples/Makefile.example b/doc/examples/Makefile.example index 6428154c51..a232d97f98 100644 --- a/doc/examples/Makefile.example +++ b/doc/examples/Makefile.example @@ -11,7 +11,7 @@ CFLAGS += -Wall -g CFLAGS := $(shell pkg-config --cflags $(FFMPEG_LIBS)) $(CFLAGS) LDLIBS := $(shell pkg-config --libs $(FFMPEG_LIBS)) $(LDLIBS) -EXAMPLES= avio_dir_cmd \ +EXAMPLES= avio_list_dir \ avio_reading \ decode_audio \ decode_video \ diff --git a/doc/examples/avio_dir_cmd.c b/doc/examples/avio_list_dir.c similarity index 69% rename from doc/examples/avio_dir_cmd.c rename to doc/examples/avio_list_dir.c index 0722bd9ab1..3073baaefa 100644 --- a/doc/examples/avio_dir_cmd.c +++ b/doc/examples/avio_list_dir.c @@ -102,38 +102,15 @@ static int list_op(const char *input_dir) return ret; } -static int del_op(const char *url) -{ - int ret = avpriv_io_delete(url); - if (ret < 0) - av_log(NULL, AV_LOG_ERROR, "Cannot delete '%s': %s.\n", url, av_err2str(ret)); - return ret; -} - -static int move_op(const char *src, const char *dst) -{ - int ret = avpriv_io_move(src, dst); - if (ret < 0) - av_log(NULL, AV_LOG_ERROR, "Cannot move '%s' into '%s': %s.\n", src, dst, av_err2str(ret)); - return ret; -} - - static void usage(const char *program_name) { - fprintf(stderr, "usage: %s OPERATION entry1 [entry2]\n" - "API example program to show how to manipulate resources " - "accessed through AVIOContext.\n" - "OPERATIONS:\n" - "list list content of the directory\n" - "move rename content in directory\n" - "del delete content in directory\n", - program_name); + fprintf(stderr, "usage: %s input_dir\n" + "API example program to show how to list files in directory " + "accessed through AVIOContext.\n", program_name); } int main(int argc, char *argv[]) { - const char *op = NULL; int ret; av_log_set_level(AV_LOG_DEBUG); @@ -145,32 +122,7 @@ int main(int argc, char *argv[]) avformat_network_init(); - op = argv[1]; - if (strcmp(op, "list") == 0) { - if (argc < 3) { - av_log(NULL, AV_LOG_INFO, "Missing argument for list operation.\n"); - ret = AVERROR(EINVAL); - } else { - ret = list_op(argv[2]); - } - } else if (strcmp(op, "del") == 0) { - if (argc < 3) { - av_log(NULL, AV_LOG_INFO, "Missing argument for del operation.\n"); - ret = AVERROR(EINVAL); - } else { - ret = del_op(argv[2]); - } - } else if (strcmp(op, "move") == 0) { - if (argc < 4) { - av_log(NULL, AV_LOG_INFO, "Missing argument for move operation.\n"); - ret = AVERROR(EINVAL); - } else { - ret = move_op(argv[2], argv[3]); - } - } else { - av_log(NULL, AV_LOG_INFO, "Invalid operation %s\n", op); - ret = AVERROR(EINVAL); - } + ret = list_op(argv[1]); avformat_network_deinit(); -- 2.24.1

4 11

[PATCH] fix some filter dependencies
by Matthias Troffaes 09 May '21

09 May '21

Currently, ffmpeg builds with --disable-avcodec --disable-avformat --enable-avfilter result in a broken avfilter library due to a few filters that require avcodec or avformat are mistakenly compiled in. For example, see: https://ci.appveyor.com/project/mcmtroffaes/ffmpeg-msvc-build/builds/376335… The attached patch fixes this.

3 4

[PATCH 01/30] avcodec/opustab: Make array static
by Andreas Rheinhardt 05 May '21

05 May '21

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt(a)gmail.com> --- libavcodec/opustab.c | 12 ++++++------ libavcodec/opustab.h | 1 - 2 files changed, 6 insertions(+), 7 deletions(-) diff --git a/libavcodec/opustab.c b/libavcodec/opustab.c index 64070f8299..7276dad793 100644 --- a/libavcodec/opustab.c +++ b/libavcodec/opustab.c @@ -950,7 +950,7 @@ const uint16_t ff_celt_qn_exp2[] = { 16384, 17866, 19483, 21247, 23170, 25267, 27554, 30048 }; -const uint32_t ff_celt_pvq_u[1272] = { +static const uint32_t celt_pvq_u[1272] = { /* N = 0, K = 0...176 */ 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, @@ -1153,9 +1153,9 @@ const float ff_celt_window2[120] = { }; const uint32_t * const ff_celt_pvq_u_row[15] = { - ff_celt_pvq_u + 0, ff_celt_pvq_u + 176, ff_celt_pvq_u + 351, - ff_celt_pvq_u + 525, ff_celt_pvq_u + 698, ff_celt_pvq_u + 870, - ff_celt_pvq_u + 1041, ff_celt_pvq_u + 1131, ff_celt_pvq_u + 1178, - ff_celt_pvq_u + 1207, ff_celt_pvq_u + 1226, ff_celt_pvq_u + 1240, - ff_celt_pvq_u + 1248, ff_celt_pvq_u + 1254, ff_celt_pvq_u + 1257 + celt_pvq_u + 0, celt_pvq_u + 176, celt_pvq_u + 351, + celt_pvq_u + 525, celt_pvq_u + 698, celt_pvq_u + 870, + celt_pvq_u + 1041, celt_pvq_u + 1131, celt_pvq_u + 1178, + celt_pvq_u + 1207, celt_pvq_u + 1226, celt_pvq_u + 1240, + celt_pvq_u + 1248, celt_pvq_u + 1254, celt_pvq_u + 1257 }; diff --git a/libavcodec/opustab.h b/libavcodec/opustab.h index 892126bb23..cdd0456e3c 100644 --- a/libavcodec/opustab.h +++ b/libavcodec/opustab.h @@ -152,7 +152,6 @@ extern const uint8_t ff_celt_bit_deinterleave[]; extern const uint8_t ff_celt_hadamard_order[]; extern const uint16_t ff_celt_qn_exp2[]; -extern const uint32_t ff_celt_pvq_u[1272]; extern const float ff_celt_postfilter_taps[3][3]; -- 2.25.1

9 49

[PATCH 1/4] avcodec/pnm_parser: Check av_image_get_buffer_size() for failure
by Michael Niedermayer 30 Apr '21

30 Apr '21

Fixes: out of array access Fixes: 30135/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_PBM_fuzzer-4997145650397184 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> --- libavcodec/pnm_parser.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/libavcodec/pnm_parser.c b/libavcodec/pnm_parser.c index d19dbfe98c..f3be6d640c 100644 --- a/libavcodec/pnm_parser.c +++ b/libavcodec/pnm_parser.c @@ -109,8 +109,10 @@ retry: if (next == END_NOT_FOUND) pnmpc->ascii_scan = sync - pnmctx.bytestream + skip; } else { - next = pnmctx.bytestream - pnmctx.bytestream_start + skip - + av_image_get_buffer_size(avctx->pix_fmt, avctx->width, avctx->height, 1); + int ret = av_image_get_buffer_size(avctx->pix_fmt, avctx->width, avctx->height, 1); + next = pnmctx.bytestream - pnmctx.bytestream_start + skip; + if (ret >= 0) + next += ret; } if (next != END_NOT_FOUND && pnmctx.bytestream_start != buf + skip) next -= pc->index; -- 2.17.1

2 12

[PATCH 01/45] avcodec/a64multienc: Fix memleak upon init failure
by Andreas Rheinhardt 30 Apr '21

30 Apr '21

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt(a)gmail.com> --- libavcodec/a64multienc.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/libavcodec/a64multienc.c b/libavcodec/a64multienc.c index 38f25020f5..5f4825d55d 100644 --- a/libavcodec/a64multienc.c +++ b/libavcodec/a64multienc.c @@ -407,6 +407,7 @@ AVCodec ff_a64multi_encoder = { .close = a64multi_close_encoder, .pix_fmts = (const enum AVPixelFormat[]) {AV_PIX_FMT_GRAY8, AV_PIX_FMT_NONE}, .capabilities = AV_CODEC_CAP_DELAY, + .caps_internal = FF_CODEC_CAP_INIT_CLEANUP, }; #endif #if CONFIG_A64MULTI5_ENCODER @@ -421,5 +422,6 @@ AVCodec ff_a64multi5_encoder = { .close = a64multi_close_encoder, .pix_fmts = (const enum AVPixelFormat[]) {AV_PIX_FMT_GRAY8, AV_PIX_FMT_NONE}, .capabilities = AV_CODEC_CAP_DELAY, + .caps_internal = FF_CODEC_CAP_INIT_CLEANUP, }; #endif -- 2.25.1

7 288

[PATCH] avcodec/libopusenc: reload packet loss at encode
by Philip-Dylan Gleonec 20 Apr '21

20 Apr '21

An estimation of packet loss is required by libopus to compute its FEC data. Currently, this estimation is constant, and can not be changed after configuration. This means an application using libopus through ffmpeg can not adapt the packet loss estimation when the network quality degrades. This patch makes the encoder reload the packet_loss AVOption before encoding samples, if fec is enabled. This way an application can modify the packet loss estimation by changing the AVOption. Typical use-case is a RTP stream, where packet loss can be estimated from RTCP packets. Signed-off-by: Philip-Dylan Gleonec <philip-dylan.gleonec(a)savoirfairelinux.com> --- libavcodec/libopusenc.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/libavcodec/libopusenc.c b/libavcodec/libopusenc.c index 70d17f802b..c18e8ae7fa 100644 --- a/libavcodec/libopusenc.c +++ b/libavcodec/libopusenc.c @@ -460,6 +460,15 @@ static int libopus_encode(AVCodecContext *avctx, AVPacket *avpkt, int ret; int discard_padding; + if (opus->opts.fec) { + ret = opus_multistream_encoder_ctl(opus->enc, + OPUS_SET_PACKET_LOSS_PERC(opus->opts.packet_loss)); + if (ret != OPUS_OK) + av_log(avctx, AV_LOG_WARNING, + "Unable to set expected packet loss percentage: %s\n", + opus_strerror(ret)); + } + if (frame) { ret = ff_af_queue_add(&opus->afq, frame); if (ret < 0) -- 2.25.1

2 2

[PATCH 00/39] Make mpegvideo encoders init-threadsafe
by Andreas Rheinhardt 17 Apr '21

17 Apr '21

This patchset makes all mpegvideo encoders (i.e. all those that use ff_mpv_encode_init()) init-threadsafe; some of the patches here supersede earlier patches (namely #214-216 from my earlier init-threadsafe patchset [1] as well as patches from the latter part of my init-vlc patchset (said part actually dealt with init-threadsafety), namely #181-183 [2]); also an error in [3] (which was not applied because of this) is corrected (the earlier patch ignored that the RLTables used by mpeg1/2 are initialized at two different places, so guarding one of them with an AVOnce is not enough). In order to make this patchset self-contained [4] from my earlier init-threadsafe patchset has been included here unchanged. Furthermore, there are also numerous fixes for RLTables interwoven in this patchset. Finally, the SpeedHQ encoder has also received attention. It called the wrong function, leading to potential segfaults. The mpegvideo decoders are my next target. [1]: https://ffmpeg.org/pipermail/ffmpeg-devel/2020-December/273287.html [2]: https://ffmpeg.org/pipermail/ffmpeg-devel/2020-November/272876.html [3]: https://ffmpeg.org/pipermail/ffmpeg-devel/2020-November/272873.html [4]: https://ffmpeg.org/pipermail/ffmpeg-devel/2020-December/273286.html Andreas Rheinhardt (39): avcodec/rl: Remove nonstatic storage of RL tables avcodec/mpeg12: Don't initialize encoder-only parts of RLTable avcodec/mpeg12: Make initializing VLCs thread-safe avcodec/mpeg12: Reindentation avcodec/mpeg12enc, speedhqenc: Avoid redundant copies of tables avcodec/speedhq: Remove cast avcodec/rl: Allow to create only a few VLC tables avcodec/h261dec: Don't initialize unused VLCs avcodec/msmpeg4dec: Don't initialize unused RL VLCs avcodec/mpeg4videodec: Don't initialize unused RL VLCs avcodec/ituh263dec: Don't initialize unused RL VLCs avcodec/msmpeg4dec: Reuse identical RL VLCs avcodec/msmpeg4dec: Reuse identical MV VLC avcodec/h261dec: Don't use too big max_depth in get_vlc2() avcodec/mpegvideo: Merge ff_mpv_decode_defaults into ff_mpv_decode_init avcodec/mpeg12dec: Remove redundant writes avcodec/me_cmp: Remove ff_check_alignment() avcodec/motion_est, mpegvideo: Make pointers to static storage const avcodec/mpegvideo_enc: Make mpv_encode_defaults thread-safe avcodec/h261dec: Don't initialize unused part of RLTable avcodec/h261dec: Make decoder init-threadsafe avcodec/h261dec: Reindentation avcodec/h261enc: Make encoder init-threadsafe avcodec/ituh263dec: Don't initialize unused parts of RLTable avcodec/ituh263enc: Remove unused function parameter avcodec/h263data, ituh263*: Make initializing RL inter table thread-safe avcodec/ituh263enc: Make static initializations thread-safe avcodec/mpeg4video: Making initializing RLTable thread-safe avcodec/mpeg4videoenc: Make encoder init-threadsafe avcodec/msmpeg4: Make ff_msmpeg4_common_init() thread-safe avcodec/msmpeg4: Reindentation avcodec/msmpeg4enc: Don't use code for static init that can fail avcodec/msmpeg4: Factor out common RLTable initialization code avcodec/msmpeg4enc: Make msmpeg4 encoders init-threadsafe avcodec/msmpeg4enc: Reindent after the last commit avcodec/speedhqenc: Call correct function avcodec/speedhq, speedhqenc: Make codecs init-threadsafe avcodec/speedhqenc: Hardcode table to save space configure, libavcodec/speedhq: Fix compiling SpeedHQ encoder configure | 1 + libavcodec/Makefile | 2 +- libavcodec/eamad.c | 1 + libavcodec/eatqi.c | 1 + libavcodec/flvenc.c | 2 +- libavcodec/h261.c | 13 ---- libavcodec/h261.h | 3 - libavcodec/h261dec.c | 45 ++++++------- libavcodec/h261enc.c | 16 +++-- libavcodec/h263.h | 2 + libavcodec/h263data.c | 16 ++++- libavcodec/h263data.h | 2 +- libavcodec/h263dec.c | 1 - libavcodec/ituh263dec.c | 12 ++-- libavcodec/ituh263enc.c | 39 +++++------ libavcodec/mdec.c | 1 + libavcodec/me_cmp.c | 23 ------- libavcodec/me_cmp.h | 2 - libavcodec/mjpegenc.c | 4 +- libavcodec/motion_est.c | 12 ++-- libavcodec/motion_est.h | 4 +- libavcodec/motion_est_template.c | 2 +- libavcodec/mpeg12.c | 67 +++++++++---------- libavcodec/mpeg12.h | 3 +- libavcodec/mpeg12dec.c | 17 ++--- libavcodec/mpeg12enc.c | 31 +++++---- libavcodec/mpeg4video.c | 14 +++- libavcodec/mpeg4video.h | 3 +- libavcodec/mpeg4videodec.c | 12 ++-- libavcodec/mpeg4videoenc.c | 26 ++++---- libavcodec/mpegvideo.c | 7 +- libavcodec/mpegvideo.h | 3 +- libavcodec/mpegvideo_enc.c | 34 +++++----- libavcodec/msmpeg4.c | 110 +++++++++++++++++-------------- libavcodec/msmpeg4.h | 2 +- libavcodec/msmpeg4data.c | 1 - libavcodec/msmpeg4data.h | 1 - libavcodec/msmpeg4dec.c | 23 +++---- libavcodec/msmpeg4enc.c | 63 ++++++++---------- libavcodec/rl.c | 53 ++++----------- libavcodec/rl.h | 19 +++--- libavcodec/rv10.c | 1 - libavcodec/rv10enc.c | 2 +- libavcodec/rv20enc.c | 2 +- libavcodec/rv34.c | 1 - libavcodec/speedhq.c | 9 +-- libavcodec/speedhqenc.c | 47 ++++--------- libavcodec/wmv2enc.c | 2 +- 48 files changed, 339 insertions(+), 418 deletions(-) -- 2.25.1

4 83

[PATCH 1/4] avformat/rtsp: set AV_OPT_FLAG_DEPRECATED on depracated options
by Andriy Gelman 10 Apr '21

10 Apr '21

From: Andriy Gelman <andriy.gelman(a)gmail.com> Signed-off-by: Andriy Gelman <andriy.gelman(a)gmail.com> --- libavformat/rtsp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libavformat/rtsp.c b/libavformat/rtsp.c index d9832bbf1f..2ef75f50e3 100644 --- a/libavformat/rtsp.c +++ b/libavformat/rtsp.c @@ -94,7 +94,7 @@ const AVOption ff_rtsp_options[] = { { "max_port", "set maximum local UDP port", OFFSET(rtp_port_max), AV_OPT_TYPE_INT, {.i64 = RTSP_RTP_PORT_MAX}, 0, 65535, DEC|ENC }, { "listen_timeout", "set maximum timeout (in seconds) to wait for incoming connections (-1 is infinite, imply flag listen)", OFFSET(initial_timeout), AV_OPT_TYPE_INT, {.i64 = -1}, INT_MIN, INT_MAX, DEC }, #if FF_API_OLD_RTSP_OPTIONS - { "timeout", "set maximum timeout (in seconds) to wait for incoming connections (-1 is infinite, imply flag listen) (deprecated, use listen_timeout)", OFFSET(initial_timeout), AV_OPT_TYPE_INT, {.i64 = -1}, INT_MIN, INT_MAX, DEC }, + { "timeout", "set maximum timeout (in seconds) to wait for incoming connections (-1 is infinite, imply flag listen) (deprecated, use listen_timeout)", OFFSET(initial_timeout), AV_OPT_TYPE_INT, {.i64 = -1}, INT_MIN, INT_MAX, DEC|AV_OPT_FLAG_DEPRECATED }, { "stimeout", "set timeout (in microseconds) of socket TCP I/O operations", OFFSET(stimeout), AV_OPT_TYPE_INT, {.i64 = 0}, INT_MIN, INT_MAX, DEC }, #else { "timeout", "set timeout (in microseconds) of socket TCP I/O operations", OFFSET(stimeout), AV_OPT_TYPE_INT, {.i64 = 0}, INT_MIN, INT_MAX, DEC }, @@ -102,7 +102,7 @@ const AVOption ff_rtsp_options[] = { COMMON_OPTS(), { "user_agent", "override User-Agent header", OFFSET(user_agent), AV_OPT_TYPE_STRING, {.str = LIBAVFORMAT_IDENT}, 0, 0, DEC }, #if FF_API_OLD_RTSP_OPTIONS - { "user-agent", "override User-Agent header (deprecated, use user_agent)", OFFSET(user_agent), AV_OPT_TYPE_STRING, {.str = LIBAVFORMAT_IDENT}, 0, 0, DEC }, + { "user-agent", "override User-Agent header (deprecated, use user_agent)", OFFSET(user_agent), AV_OPT_TYPE_STRING, {.str = LIBAVFORMAT_IDENT}, 0, 0, DEC|AV_OPT_FLAG_DEPRECATED }, #endif { NULL }, }; -- 2.28.0

3 12

[PATCH] set ulMaxDisplayDelay cuviddec parser option to zero if low_delay flag is on
by clime 09 Apr '21

09 Apr '21

From: Michal Novotny <michal.novotny(a)comprimato.com> * zero is recommended value in Nvidia coding samples for low latency use-cases Signed-off-by: Michal Novotny <michal.novotny(a)comprimato.com> --- libavcodec/cuviddec.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libavcodec/cuviddec.c b/libavcodec/cuviddec.c index 49775b5a09..ec57afdefe 100644 --- a/libavcodec/cuviddec.c +++ b/libavcodec/cuviddec.c @@ -999,7 +999,7 @@ static av_cold int cuvid_decode_init(AVCodecContext *avctx) } ctx->cuparseinfo.ulMaxNumDecodeSurfaces = ctx->nb_surfaces; - ctx->cuparseinfo.ulMaxDisplayDelay = 4; + ctx->cuparseinfo.ulMaxDisplayDelay = (avctx->flags & AV_CODEC_FLAG_LOW_DELAY) ? 0 : 4; ctx->cuparseinfo.pUserData = avctx; ctx->cuparseinfo.pfnSequenceCallback = cuvid_handle_video_sequence; ctx->cuparseinfo.pfnDecodePicture = cuvid_handle_picture_decode; -- 2.30.1

3 4