On 1/27/07, Ricardo Lugo <[EMAIL PROTECTED]> wrote: > On Jan 26, 2007, at 5:31 AM, Martin van Es wrote: > > > Hi, > > > > I've recently upgraded my pvr-150 to a pvr-500 and am seeing > > occasional (+/- 5 minutes) mpg artefacts (glitches) in the > > recorded/shown mpg stream (mythtv). I started out with kernel 2.6.16 / > > ivtv 0.6.6, but it turned out the new Samsung tuner wasn't recognized > > so I upgraded to 2.6.19 / ivtv 0.9.1. This works fine (I can watch TV > > now) except for the fact that I now have the forementioned occasional > > glitches/artefacts in the lower part of my screen. > > I get these too. They are unrelated to signal quality or hard drive > activity as far as I can tell. And they happen frequently. > > My setup is a SMP PPC machine with ivtv-0.9.1 and a PVR-350. I'm not > sure what kind of processor architecture you're running, but it would > be interesting to hear if you are using a x86, PPC, etc and the > endianness of it.
My setup is a Via Epia ML-6000 (i386 based) but that isn't relevant anymore since I think I solved it (so far so good). It turns out recent ivtv versions are seriously braindamaged regarding vbi. I saw ivtv_enc_vbi taking up around 10% CPU time for each tuner (compare 0-1% for ivtv_enc alone). A little bit of searching for ivtv_enc_vbi revealed the borken vbi code and disabling it completely solved my problem. What I did was set VBI from 'Teletekst' to 'None' in mythtv-setup and rmmod ivtv; modprobe ivtv (you need to do this, because the vbi setting is persistant in the module). Furthermore I gathered from the discussion that they're reworking the ivtv/vbi code as we speak, so if you need vbi (teletekst/closed captions) you need to wait for 0.10.x release of ivtv. For the time being you should turn it off however. Hope this helps, everybody thanks for listening and my apologies for the FUD on the this dev-list! Regards, Martin _______________________________________________ ivtv-devel mailing list [email protected] http://ivtvdriver.org/mailman/listinfo/ivtv-devel
