On Sun, Jun 7, 2009 at 10:12 PM, W. Michael Petullo<mike@xxxxxxxx> wrote: >> Do you see the issue using tvtime? This will help isolate whether it's >> an application compatibility issue or whether it's related to endianness >> (and I do almost all my testing with tvtime). > > Tvtime works like a charm. The colors look fine. This is the first I've > seen tvtime and it seems great. Now we may be getting off the topic, but > does anyone know why xawtv, streamer and GStreamer's v4l2src would muck up > the colors? Ok, that's good to know. Perhaps it's some difference in the way the v4l stream is read and the frames are dequeued (mmap versus read(), etc). >> You indicated that you had reason to believe it's a PowerPC issue. Is >> there any reason that you came to that conclusion other than that you're >> running on ppc? I'm not discounting the possibility, but it would be >> good to know if you have other information that supports your theory. > > It was a hypothesis, but based on experience in "seeing" endian bugs in > video code and "hearing" endian bugs in audio code. After using PowerPC > long enough, you learn to jump to the endian conclusion pretty quickly. I > was wrong! Ok, well that's good to know. I did look at the code and couldn't see how it could possibly be an endianness bug. Bear in mind that the analog support for the 950q is still relatively new, and its entirely possible there are some application specific bugs to be worked out as there is more testing. Could you please describe in more detail the *exact* configuration you are attempting, including the versions of the applications you are using and command line arguments you are passing. If I can reproduce the issue here then I can probably debug it much faster. Thanks, Devin -- Devin J. Heitmueller - Kernel Labs http://www.kernellabs.com -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
