[Webkit-unassigned] [Bug 182200] Samples in sourceBufferPrivateDidReceiveSample are added but not inserted in the decodeQueue

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Jan 29 22:24:26 PST 2018


https://bugs.webkit.org/show_bug.cgi?id=182200

--- Comment #5 from Jer Noble <jer.noble at apple.com> ---
(In reply to Eric Stobbart from comment #3)
> For additional context, I'm writing a platform implementation for embedded
> devices. I'm taking advantage of a platform sample's duration and reporting
> an ISOBMFF as an array of all the samples in a single platform sample & call
> back the SourceBufferClient once. I'm reading duration off the trun so the
> calculations work well as is. This speeds up the update cycle to the
> SourceBuffer and frees up the event loop back to JavaScript. I'm with you on
> it being a hack for sure, open to any ideas here, but I didn't want to rock
> the boat too much either.

Interesting. Well, as long as you implement "divisible" MediaSamples in case the client asks to remove() a range that occurs in the middle of your giant 'trun' sample, that should work fine. On iOS, that's basically how the audio MediaSamples work on Mac and iOS (since adding 44100 samples-per-second to a map structure is a bit overkill).

But now I see why you might be concerned about dropping lastEnqueuedDecodeEndTime for lastEnqueuedDecodeTime.

So is the issue due to poorly encoded content? I.e., are you seeing files where the 'trun's duration + the 'tfdt's' baseMediaDecodeTime != the next 'tfdt's baseMediaDecodeTime?

> Few ideas, could we round down the lastEnqueuedDecodeEndTime when decodeTime
> and duration get added?

I don't think there's much to round here, since both values should be in the sample's track's media's timebase.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20180130/a8bc21c9/attachment.html>


More information about the webkit-unassigned mailing list