RESOLVED FIXED 190683
[GStreamer][WebRTC] Do not try to handle framerate modulation in the encoder 
https://bugs.webkit.org/show_bug.cgi?id=190683
Summary [GStreamer][WebRTC] Do not try to handle framerate modulation in the encoder
Thibault Saunier
Reported 2018-10-17 12:57:51 PDT
[GStreamer][WebRTC] Do not try to handle framerate modulation in the encoder
Attachments
Patch (2.94 KB, patch)
2018-10-17 12:58 PDT, Thibault Saunier 		no flags
DetailsFormatted DiffDiff
Patch (2.96 KB, patch)
2018-10-17 12:59 PDT, Thibault Saunier 		no flags
DetailsFormatted DiffDiff
Patch (2.96 KB, patch)
2018-11-05 11:05 PST, Thibault Saunier 		no flags
DetailsFormatted DiffDiff
Patch (2.96 KB, patch)
2018-11-06 04:40 PST, Thibault Saunier 		no flags
DetailsFormatted DiffDiff
Show Obsolete (3) View All Add attachment proposed patch, testcase, etc.
Thibault Saunier
Comment 1 2018-10-17 12:58:07 PDT
Created attachment 352620 [details] Patch
Thibault Saunier
Comment 2 2018-10-17 12:59:18 PDT
Created attachment 352622 [details] Patch
Thibault Saunier
Comment 3 2018-11-05 11:05:06 PST
Created attachment 353888 [details] Patch
Xabier Rodríguez Calvar
Comment 4 2018-11-05 22:56:44 PST
Comment on attachment 353888 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=353888&action=review > Source/WebCore/ChangeLog:8 > + This has to already be handled in capturing pipeline or in libwebrtc itself Period at the end. > Source/WebCore/ChangeLog:10 > + No other encoder implementation try to do that, and libwebrtc is not happy with encoder that do not output the exact number of frames that have been passed in try -> tried, "with encoders that" and period at the end.
Xabier Rodríguez Calvar
Comment 5 2018-11-05 22:58:45 PST
Comment on attachment 353888 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=353888&action=review > Source/WebCore/ChangeLog:12 > + No regressions detected and libwebrtc is happier this way, less warning output and no more frame corruption in H264 streams found. My only concern is we should have a test for this, I'll let Phil decide.
Thibault Saunier
Comment 6 2018-11-06 04:15:43 PST
(In reply to Xabier Rodríguez Calvar from comment #5) > Comment on attachment 353888 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=353888&action=review > > > Source/WebCore/ChangeLog:12 > > + No regressions detected and libwebrtc is happier this way, less warning output and no more frame corruption in H264 streams found. > > My only concern is we should have a test for this, I'll let Phil decide. Well, we basically have "test", libwebrtc is not throwing WARNINGS anymore, I do not see how we could have a dedicated test at our level.
Thibault Saunier
Comment 7 2018-11-06 04:40:20 PST
Created attachment 353962 [details] Patch
WebKit Commit Bot
Comment 8 2018-11-06 06:23:32 PST
Comment on attachment 353962 [details] Patch Clearing flags on attachment: 353962 Committed r237860: <https://trac.webkit.org/changeset/237860>
WebKit Commit Bot
Comment 9 2018-11-06 06:23:33 PST
All reviewed patches have been landed. Closing bug.
Radar WebKit Bug Importer
Comment 10 2018-11-06 06:24:22 PST
<rdar://problem/45841276>
Note You need to log in before you can comment on or make changes to this bug.