[vlc-devel] MPEGTS: audio stops after PCR jump

[Steve Lhomme]

Hi,

I think you should post these extra information on the issue.
This mailing list is very low traffic. And when people will look at the 
issue they will not have these extra information.

Thanks.

On 2022-06-21 10:35, Kristof Vandenbussche wrote:
> Hi All,
> 
> I am looking for a solution to the problem as mentioned in 
> https://code.videolan.org/videolan/vlc/-/issues/26768 
> <https://code.videolan.org/videolan/vlc/-/issues/26768>.
> (vlc-3.0.17)
> I am focussing on why the audio stream stops playing as soon as there is 
> a (huge) jump in the PCR, because video keeps playing.
> 
> What I see is that the problem is that the PTS from the mpeg audio 
> stream does not get updated when a jump in the PCR happens, and still 
> uses the old base value, causing packages to be discarded
> 
>  >>[00007fbfc405ebf0] main decoder debug: discarded audio buffer
> 
> The problem is that the code  in ClockStreamToSystem starts returning 
> negative times.
> this is because cl->ref.i_stream > i_stream.
> 
> The cl->ref.i_stream was correctly updated on a PTS gap (clock reset), 
> but apparently the audio stream's clock is still the old one (and does 
> not correct after a few packages)
> 
> I have verified that the PTS of the MPEGTS demuxer tags the packets with 
> the correct pts/dts, and I also see the same timestamps being used by 
> the MPEG packetizer.
> However, the MPEG decoder still seems to get the buffers with the old 
> base clock.
> 
> The problem is that I don't know what could be still in between. A 
> stream filter? Or am I overlooking something?
> Any help appreciated.
> 
> Best Regards,
> 
> 
> Kristof Vandenbussche
> ** **
> 
> GPG fingerprint: 19F6 0638 A33B 563D ED71  F07F E5AE 2082 E07D 9A38
> 
> 
> 
> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> https://mailman.videolan.org/listinfo/vlc-devel