[mythtv-users] Alternatives to lossless transcode

John Pilkington J.Pilk at tesco.net
Wed Nov 13 10:25:51 UTC 2013


On 13/11/13 09:43, mark wrote:
> On 12/11/13 20:42, Paul Gardiner wrote:
>
>> Do you mean running my script or just running mythtranscode?
>> Mythtranscode --fifodir run alone will stall eventually because the
>> fifos will fill up. It will also stall if something reads from them
>> unevenly. But, assuming you are running the script, you could easily
>> be correct that whatever upsets --mpeg2 is also upsetting --fifodir.
>>
>> Cheers,
>>      Paul.
>> _______________________________________________
>
> Sorry my mistake. I assumed that if I ran just mythtranscode --fifodir
> it would just output the two streams in their entirety to the nominated
> files. Trying your script now, for some reason my ffmpeg did not like
> the '-profile high' switch but have taken that out and it works.
>
> The audio was much more in sync than previously, but it still just seems
> a little off. I ran mythDVBcut over the same file (which uses ProjectX
> and no mythtranscode) and the file generated by your script had a delay
> of 80ms (from mediainfo) versus 112ms from the ProjectX version. I'm not
> sure that -hand on heart- I could pick a difference in sync of 32ms but
> it did seem just that little bit out whereas the ProjectX file just
> seemed spot on.

I think that delay will be specified by the -f 9 DVD-like profile in mplex

    INFO: [mplex] Setting specified specified data rate: 10080000
    INFO: [mplex] Run-in delay = 7200 Video delay = 7200 Audio delay = 10800

IIRC it's a 90 kHz MPEG clock; looks like 120 ms.  man mplex isn't 
entirely clear about it.  I think Project-X adjusts by single 
audio-frame steps of 24 ms.

>
> Again the question probably is whether this is due to errors in the
> original recording or something else.
>
> I'm updating my ffmpeg now and will re-try your script and see if that
> changes anything.
>



More information about the mythtv-users mailing list