Message ID | 20230223222513.3759-2-dheitmueller@ltnglobal.com |
---|---|
State | New |
Headers | show |
Series | Minor decklink output fixes | expand |
Context | Check | Description |
---|---|---|
yinshiyou/make_loongarch64 | success | Make finished |
yinshiyou/make_fate_loongarch64 | success | Make fate finished |
andriy/make_x86 | success | Make finished |
andriy/make_fate_x86 | success | Make fate finished |
On Thu, 23 Feb 2023, Devin Heitmueller wrote: > The existing code assumed that the first frame received by the decklink > output would always be PTS zero. However if running in other timing > modes than the default of CBR, items such as frame dropping at the > beginning may result in starting at a non-zero PTS. > > For example, in our setup because we discard probing data and run > with "-vsync 2" the first video frame scheduled to the decklink > output will have a PTS around 170. Scheduling frames too far into > the future will either fail or cause a backlog of frames scheduled > far enough into the future that the entire pipeline will stall. > > Issue can be reproduced with the following command-line: > > ./ffmpeg -copyts -i foo.ts -f decklink -vcodec v210 -ac 2 'DeckLink Duo (4)' > > Keep track of the PTS of the first frame received, so that when > we enable start playback we can provide that value to the decklink > driver. > > Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com> > --- > libavdevice/decklink_common.h | 1 + > libavdevice/decklink_enc.cpp | 7 +++++-- > 2 files changed, 6 insertions(+), 2 deletions(-) > > diff --git a/libavdevice/decklink_common.h b/libavdevice/decklink_common.h > index 79d6ac5b38..088e165ee7 100644 > --- a/libavdevice/decklink_common.h > +++ b/libavdevice/decklink_common.h > @@ -118,6 +118,7 @@ struct decklink_ctx { > > /* Status */ > int playback_started; > + int64_t first_pts; > int64_t last_pts; > unsigned long frameCount; > unsigned int dropped; > diff --git a/libavdevice/decklink_enc.cpp b/libavdevice/decklink_enc.cpp > index fb686b9032..c3dc2c0cac 100644 > --- a/libavdevice/decklink_enc.cpp > +++ b/libavdevice/decklink_enc.cpp > @@ -486,6 +486,9 @@ static int decklink_write_video_packet(AVFormatContext *avctx, AVPacket *pkt) > ctx->frames_buffer_available_spots--; > pthread_mutex_unlock(&ctx->mutex); > > + if (ctx->first_pts == 0) > + ctx->first_pts = pkt->pts; And what if first packet pts is 0? Then the second packet pts will be assigned to first pts? Maybe you should use AV_NOPTS_VALUE for the default first_pts value and check for that instead. Regards, Marton > + > /* Schedule frame for playback. */ > hr = ctx->dlo->ScheduleVideoFrame((class IDeckLinkVideoFrame *) frame, > pkt->pts * ctx->bmd_tb_num, > @@ -505,14 +508,14 @@ static int decklink_write_video_packet(AVFormatContext *avctx, AVPacket *pkt) > " Video may misbehave!\n"); > > /* Preroll video frames. */ > - if (!ctx->playback_started && pkt->pts > ctx->frames_preroll) { > + if (!ctx->playback_started && pkt->pts > (ctx->first_pts + ctx->frames_preroll)) { > av_log(avctx, AV_LOG_DEBUG, "Ending audio preroll.\n"); > if (ctx->audio && ctx->dlo->EndAudioPreroll() != S_OK) { > av_log(avctx, AV_LOG_ERROR, "Could not end audio preroll!\n"); > return AVERROR(EIO); > } > av_log(avctx, AV_LOG_DEBUG, "Starting scheduled playback.\n"); > - if (ctx->dlo->StartScheduledPlayback(0, ctx->bmd_tb_den, 1.0) != S_OK) { > + if (ctx->dlo->StartScheduledPlayback(ctx->first_pts * ctx->bmd_tb_num, ctx->bmd_tb_den, 1.0) != S_OK) { > av_log(avctx, AV_LOG_ERROR, "Could not start scheduled playback!\n"); > return AVERROR(EIO); > } > -- > 2.35.1.655.ga68dfadae5 > > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe". >
On Tue, Feb 28, 2023 at 3:46 PM Marton Balint <cus@passwd.hu> wrote: > And what if first packet pts is 0? Then the second packet pts will be > assigned to first pts? Maybe you should use AV_NOPTS_VALUE for the default > first_pts value and check for that instead. That's a good point. I never noticed that, but noticing that the first frame got dropped is probably not something that would have been immediately obvious. In any case, I prefer your approach and will resubmit. Thanks for the feedback, Devin
diff --git a/libavdevice/decklink_common.h b/libavdevice/decklink_common.h index 79d6ac5b38..088e165ee7 100644 --- a/libavdevice/decklink_common.h +++ b/libavdevice/decklink_common.h @@ -118,6 +118,7 @@ struct decklink_ctx { /* Status */ int playback_started; + int64_t first_pts; int64_t last_pts; unsigned long frameCount; unsigned int dropped; diff --git a/libavdevice/decklink_enc.cpp b/libavdevice/decklink_enc.cpp index fb686b9032..c3dc2c0cac 100644 --- a/libavdevice/decklink_enc.cpp +++ b/libavdevice/decklink_enc.cpp @@ -486,6 +486,9 @@ static int decklink_write_video_packet(AVFormatContext *avctx, AVPacket *pkt) ctx->frames_buffer_available_spots--; pthread_mutex_unlock(&ctx->mutex); + if (ctx->first_pts == 0) + ctx->first_pts = pkt->pts; + /* Schedule frame for playback. */ hr = ctx->dlo->ScheduleVideoFrame((class IDeckLinkVideoFrame *) frame, pkt->pts * ctx->bmd_tb_num, @@ -505,14 +508,14 @@ static int decklink_write_video_packet(AVFormatContext *avctx, AVPacket *pkt) " Video may misbehave!\n"); /* Preroll video frames. */ - if (!ctx->playback_started && pkt->pts > ctx->frames_preroll) { + if (!ctx->playback_started && pkt->pts > (ctx->first_pts + ctx->frames_preroll)) { av_log(avctx, AV_LOG_DEBUG, "Ending audio preroll.\n"); if (ctx->audio && ctx->dlo->EndAudioPreroll() != S_OK) { av_log(avctx, AV_LOG_ERROR, "Could not end audio preroll!\n"); return AVERROR(EIO); } av_log(avctx, AV_LOG_DEBUG, "Starting scheduled playback.\n"); - if (ctx->dlo->StartScheduledPlayback(0, ctx->bmd_tb_den, 1.0) != S_OK) { + if (ctx->dlo->StartScheduledPlayback(ctx->first_pts * ctx->bmd_tb_num, ctx->bmd_tb_den, 1.0) != S_OK) { av_log(avctx, AV_LOG_ERROR, "Could not start scheduled playback!\n"); return AVERROR(EIO); }
The existing code assumed that the first frame received by the decklink output would always be PTS zero. However if running in other timing modes than the default of CBR, items such as frame dropping at the beginning may result in starting at a non-zero PTS. For example, in our setup because we discard probing data and run with "-vsync 2" the first video frame scheduled to the decklink output will have a PTS around 170. Scheduling frames too far into the future will either fail or cause a backlog of frames scheduled far enough into the future that the entire pipeline will stall. Issue can be reproduced with the following command-line: ./ffmpeg -copyts -i foo.ts -f decklink -vcodec v210 -ac 2 'DeckLink Duo (4)' Keep track of the PTS of the first frame received, so that when we enable start playback we can provide that value to the decklink driver. Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com> --- libavdevice/decklink_common.h | 1 + libavdevice/decklink_enc.cpp | 7 +++++-- 2 files changed, 6 insertions(+), 2 deletions(-)