[mythtv] Re:Transcoding pvr-250 files

Isaac Richards ijr at po.cwru.edu
Fri Aug 29 17:51:34 EDT 2003


On Friday 29 August 2003 04:41 pm, Russ Southern wrote:
> On Fri, Aug 29, 2003 at 09:10:56AM -0500, Geoffrey Hausheer wrote:
> > On Fri, 29 Aug 2003 00:21:23 -0400, "Vinton Coffman
> > vintonc-at-redzone.com |mythtv/1.0-Allow|" <661whj4npx0t at sneakemail.com>
> >
> > said:
> > > I swapped in the new fix.  The total time is still wrong.  I transcoded
> > > another 1 hour show and the total time is 21:32.  I've been timing this
> > > show for a while and 20 minutes of real time translates into ~9
> > > minutes.  Sorry for the lack of precision.  If you need more accuracy I
> > > can retime the whole show tomorrow.
> >
> > Hmm.  I really can't explain this without sitting in front of my TV for a
> > while.  The patch I sent should have fixed it as far as I can tell.
> >
> > Oh well, the patch is right (even if it doesn't fix the problem), and
> > it's in CVS now.  I'll take a look at this sometime next week, I guess.
>
> Could this be related to mplayer's seek/time calc issue?  mplayer assumes
> CBR when calculating times/offsets (X bytes per frame), but MPEG is VBR.
> My experimentation showed almost exactly the skew Vinton is reporting (just
> less than 1/2).

Nope, not related.

> Does myth now try to calculate show length based on some number of bytes
> per frame or keyframe?

Calculates show time by the data that's present in the seektable.

Isaac


More information about the mythtv-dev mailing list