Message ID | tencent_EE0E40DE9A569FE5AB454E6E700A2DA79A08@qq.com |
---|---|
State | New |
Headers | show |
Series | [FFmpeg-devel] avcodec/h264_mp4toannexb: fix missing SPS/PPS to IDR frames | 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, 2022-12-29 at 01:20 +0800, Zhao Zhili wrote: > From: Zhao Zhili <zhilizhao@tencent.com> > > If there is a single group of SPS/PPS before an IDR frame, but no > SPS/PPS after that, we will miss the chance to reset > idr_sps_seen/idr_pps_seen. The result is missing SPS/PPS for all > IDR frames except the first one. > > Signed-off-by: Zhao Zhili <zhilizhao@tencent.com> > --- > libavcodec/h264_mp4toannexb_bsf.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/libavcodec/h264_mp4toannexb_bsf.c > b/libavcodec/h264_mp4toannexb_bsf.c > index d11be455c2..bbdeca9c44 100644 > --- a/libavcodec/h264_mp4toannexb_bsf.c > +++ b/libavcodec/h264_mp4toannexb_bsf.c > @@ -259,7 +259,7 @@ static int h264_mp4toannexb_filter(AVBSFContext > *ctx, AVPacket *opkt) > > count_or_copy(&out, &out_size, buf, nal_size, > unit_type == H264_NAL_SPS || unit_type == > H264_NAL_PPS, j); > - if (!new_idr && unit_type == H264_NAL_SLICE) { > + if (unit_type == H264_NAL_SLICE) { > new_idr = 1; > sps_seen = 0; > pps_seen = 0; Ping for review.
Zhao Zhili: > From: Zhao Zhili <zhilizhao@tencent.com> > > If there is a single group of SPS/PPS before an IDR frame, but no > SPS/PPS after that, we will miss the chance to reset > idr_sps_seen/idr_pps_seen. The result is missing SPS/PPS for all > IDR frames except the first one. > > Signed-off-by: Zhao Zhili <zhilizhao@tencent.com> > --- > libavcodec/h264_mp4toannexb_bsf.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/libavcodec/h264_mp4toannexb_bsf.c b/libavcodec/h264_mp4toannexb_bsf.c > index d11be455c2..bbdeca9c44 100644 > --- a/libavcodec/h264_mp4toannexb_bsf.c > +++ b/libavcodec/h264_mp4toannexb_bsf.c > @@ -259,7 +259,7 @@ static int h264_mp4toannexb_filter(AVBSFContext *ctx, AVPacket *opkt) > > count_or_copy(&out, &out_size, buf, nal_size, > unit_type == H264_NAL_SPS || unit_type == H264_NAL_PPS, j); > - if (!new_idr && unit_type == H264_NAL_SLICE) { > + if (unit_type == H264_NAL_SLICE) { > new_idr = 1; > sps_seen = 0; > pps_seen = 0; If I understand you correctly, you want that to include the parameter sets from the extradata to every IDR frame without extradata if said IDR frame follows a non-IDR frame after in-band parameter sets. There is just one problem with this: Said in-band parameter sets might overwrite parameter sets from extradata and inserting the outdated parameter sets from extradata lateron might make the stream undecodable. - Andreas
On Tue, 2023-01-10 at 07:41 +0100, Andreas Rheinhardt wrote: > Zhao Zhili: > > From: Zhao Zhili <zhilizhao@tencent.com> > > > > If there is a single group of SPS/PPS before an IDR frame, but no > > SPS/PPS after that, we will miss the chance to reset > > idr_sps_seen/idr_pps_seen. The result is missing SPS/PPS for all > > IDR frames except the first one. > > > > Signed-off-by: Zhao Zhili <zhilizhao@tencent.com> > > --- > > libavcodec/h264_mp4toannexb_bsf.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/libavcodec/h264_mp4toannexb_bsf.c > > b/libavcodec/h264_mp4toannexb_bsf.c > > index d11be455c2..bbdeca9c44 100644 > > --- a/libavcodec/h264_mp4toannexb_bsf.c > > +++ b/libavcodec/h264_mp4toannexb_bsf.c > > @@ -259,7 +259,7 @@ static int h264_mp4toannexb_filter(AVBSFContext > > *ctx, AVPacket *opkt) > > > > count_or_copy(&out, &out_size, buf, nal_size, > > unit_type == H264_NAL_SPS || unit_type > > == H264_NAL_PPS, j); > > - if (!new_idr && unit_type == H264_NAL_SLICE) { > > + if (unit_type == H264_NAL_SLICE) { > > new_idr = 1; > > sps_seen = 0; > > pps_seen = 0; > > If I understand you correctly, you want that to include the parameter > sets from the extradata to every IDR frame without extradata if said > IDR > frame follows a non-IDR frame after in-band parameter sets. There is > just one problem with this: Said in-band parameter sets might > overwrite > parameter sets from extradata and inserting the outdated parameter > sets > from extradata lateron might make the stream undecodable. It's a good question. Current code works when an IDR has in-band parameter set, which is different than the out-of-band parameter set, and the decoder doesn't miss the in-band parameter set. But there is a high chance the decoder will miss the in-band parameter set for annexb bitstream, e.g., playback at the middle of a live mpegts streams. It's a serious issue of missing periodic in-band parameter set for annexb for a lot of usecases. Handle new parameter set isn't as common as the need of periodic in-band parameter set (I guess). It's not a new issue that this bitstream failed to update the parameter set, for example, it doesn't handle AV_PKT_DATA_NEW_EXTRADATA. I can add support for 1. handling AV_PKT_DATA_NEW_EXTRADATA 2. handling in-band sps/pps But it won't be soon. > > - Andreas > > _______________________________________________ > 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 Jan 10, 2023, at 23:31, Zhao Zhili <quinkblack@foxmail.com> wrote: > > On Tue, 2023-01-10 at 07:41 +0100, Andreas Rheinhardt wrote: >> Zhao Zhili: >>> From: Zhao Zhili <zhilizhao@tencent.com> >>> >>> If there is a single group of SPS/PPS before an IDR frame, but no >>> SPS/PPS after that, we will miss the chance to reset >>> idr_sps_seen/idr_pps_seen. The result is missing SPS/PPS for all >>> IDR frames except the first one. >>> >>> Signed-off-by: Zhao Zhili <zhilizhao@tencent.com> >>> --- >>> libavcodec/h264_mp4toannexb_bsf.c | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/libavcodec/h264_mp4toannexb_bsf.c >>> b/libavcodec/h264_mp4toannexb_bsf.c >>> index d11be455c2..bbdeca9c44 100644 >>> --- a/libavcodec/h264_mp4toannexb_bsf.c >>> +++ b/libavcodec/h264_mp4toannexb_bsf.c >>> @@ -259,7 +259,7 @@ static int h264_mp4toannexb_filter(AVBSFContext >>> *ctx, AVPacket *opkt) >>> >>> count_or_copy(&out, &out_size, buf, nal_size, >>> unit_type == H264_NAL_SPS || unit_type >>> == H264_NAL_PPS, j); >>> - if (!new_idr && unit_type == H264_NAL_SLICE) { >>> + if (unit_type == H264_NAL_SLICE) { >>> new_idr = 1; >>> sps_seen = 0; >>> pps_seen = 0; >> >> If I understand you correctly, you want that to include the parameter >> sets from the extradata to every IDR frame without extradata if said >> IDR >> frame follows a non-IDR frame after in-band parameter sets. There is >> just one problem with this: Said in-band parameter sets might >> overwrite >> parameter sets from extradata and inserting the outdated parameter >> sets >> from extradata lateron might make the stream undecodable. > > It's a good question. Current code works when an IDR has in-band > parameter set, which is different than the out-of-band parameter set, > and the decoder doesn't miss the in-band parameter set. But there is a > high chance the decoder will miss the in-band parameter set for annexb > bitstream, e.g., playback at the middle of a live mpegts streams. It's > a serious issue of missing periodic in-band parameter set for annexb > for a lot of usecases. Handle new parameter set isn't as common as the > need of periodic in-band parameter set (I guess). > > It's not a new issue that this bitstream failed to update the parameter > set, for example, it doesn't handle AV_PKT_DATA_NEW_EXTRADATA. I can > add support for > 1. handling AV_PKT_DATA_NEW_EXTRADATA > 2. handling in-band sps/pps > > But it won't be soon. Ping. > >> >> - Andreas >> >> _______________________________________________ >> 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". > > _______________________________________________ > 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".
diff --git a/libavcodec/h264_mp4toannexb_bsf.c b/libavcodec/h264_mp4toannexb_bsf.c index d11be455c2..bbdeca9c44 100644 --- a/libavcodec/h264_mp4toannexb_bsf.c +++ b/libavcodec/h264_mp4toannexb_bsf.c @@ -259,7 +259,7 @@ static int h264_mp4toannexb_filter(AVBSFContext *ctx, AVPacket *opkt) count_or_copy(&out, &out_size, buf, nal_size, unit_type == H264_NAL_SPS || unit_type == H264_NAL_PPS, j); - if (!new_idr && unit_type == H264_NAL_SLICE) { + if (unit_type == H264_NAL_SLICE) { new_idr = 1; sps_seen = 0; pps_seen = 0;