On Thursday 21 July 2005 18:41, Cory Papenfuss wrote: > > On Thursday 21 July 2005 17:09, Cory Papenfuss wrote: > >> The problem (as I've painfully discovered and discussed before) is > >> that the ivtv cards change the audio and video relative timestamps (PTS, > >> DTS, etc). That information is at the PS level, and avidemux blindly > >> strips PS into A and V. If the sync offset isn't constant, it gets off > >> as the clip progresses > > > > have you asked the ivtv bods if they could "fix" this? > > I haven't but it's not the fault of the driver. I believe that > the chip itself does it when it gets a bogus packet. Basically, since > everything is a convoluted mess of interleaved audio and video frames, > when a video frame comes up short (like an analog glitch in a VHS tape > capture for instance), it's much easier to just throw that frame away and > change a time offset. All of the other frames are still in buffers and > don't need to be messed with. > > ... or at least that's how I rationalize it. If the ivtv driver > were to look for this, it would basically have to do everything we're > talking about, but in the driver kernel space (not good). Basically, the > streams produced aren't incorrect, they're just a little more complicated > than the simplistic streams that 95% of the linux software mpeg utilities > out there expect
ok, but nevertheless, it might be worth asking them -- simon
_______________________________________________ mythtv-users mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
