Tuesday, September 13, 2005

ivtv 0.3.8 and xdriver 0.10.6

Lots of Hauppauge PVR150/250/350/500/550 driver improvements recently...

I'd been working with Hans Verkuil to try and get the widescreen VBI switching to work in MythTV. We tested lots of configurations and he made a number of changes to VBI but he couldn't reproduce the problem and didn't have Myth to test against.

To compound the matter, the cheapo RGB->svideo converter I bought strips VBI completely so I can only test widescreen-switching when using a direct composite feed from my cablebox. :-(

Before 0.3.8, you could start watching a widescreen show and as long as you'd setup the ivtv driver to capture, passthrough and embed VBI WSS signals in advance:
# ivtvctl -w wss -x 1 -b wss
the WSS would get through to your TV and it would switch aspect ratios. However, as soon as you fast-forwarded or rewound or jumped back to the menu and jumped back in at a bookmark, widescreen signals (and NTSC closed-caption) would fail.

Finally, Hans figured out that it was due to a VBI byte-alignment problem and while initially he produced a patch for MythTV, he's now fixed it in the driver. Having said that, I switched my setup back to svideo a week ago and haven't been able to test that it now works ok...

During the same time, John Harvey was working on the Xv/YUV components and has managed to optimise the X driver and change the default interlacing mode within the ivtv driver. The former means that Xv playback consumes less CPU and generates less heat and noise on my frontend. The later means that the stuttering problems I was seeing when playing back certain types of material with xine are now gone! The combination of these updates means that Xv/YUV is now a real choice for me and I'm able to watch rips of my DVDs and other material on my Shuttle frontend without being distratced by noise or quality issues.

I still don't have digital audio-out working on the shuttle though, however now that the other pieces of the puzzle are in place, it's probably worth my while to debug the SPDIF problem further...


Post a Comment

<< Home