[mythtv-users] Digital Pegs CPU

[Rod Smith]

On Thursday 27 November 2008 11:12:20 am Paulin wrote:

> Well I ran your script (Thank for that) and got the following on the
> two different recording types.
>
> H1250 (Digital)
> size: 4669085696

[...much snippage below...]

> ID_VIDEO_FORMAT=0x10000002
> ID_VIDEO_BITRATE=6000000
> ID_VIDEO_WIDTH=720
> ID_VIDEO_HEIGHT=480
> ID_VIDEO_FPS=29.970
> ID_LENGTH=7197.39

> H250 (Analog)
> size: 2333061120
>
> ID_VIDEO_FORMAT=0x10000002
> ID_VIDEO_BITRATE=6000000
> ID_VIDEO_WIDTH=720
> ID_VIDEO_HEIGHT=480
> ID_VIDEO_FPS=29.970
> ID_LENGTH=3595.59

> Do any of the help?

Your two files are remarkably similar in settings -- the same bitrate (both by 
ID_VIDEO_BITRATE and by computation of size / ID_LENGTH), the same 
resolution, the same FPS, etc. In fact, they're similar enough that I suggest 
you double-check that they're actually recorded from two different devices 
(your HDR-1250 and your PVR-250).

If in fact they ARE recorded from two different devices, and yet one plays 
back OK and the other doesn't, I'm pretty much at a loss. They're similar 
enough in overall format that I don't know why one would do better than 
another -- but that's a reflection of my modest knowledge of video formats; I 
don't mean to imply that you're imagining things!

> Also I found out that MiniMyth (which is what I'm using on the VIA board)
> has XvMC built in support.  I had it enabled.  I also tried XvMC-VLD.  So
> based on earlier discussion I believe (but I could defiantly be wrong) that
> the XvMC support isn't the problem.

XvMC support requires both support in MythTV and support in your drivers, so 
even if you select XvMC in the front-end, it might not be used.

> Now I did have Deinterlace set to bobdeint.  There are a host to choose
> from.  Not sure what I should be using here.  Any recommendations.

Start with none. That's the least CPU-intensive option, but it may produce 
ugly streaks in video that features fast motion or panning. Use this option 
to test the video playback quality with a recording with known problems. If 
it's OK (aside from streaking, if this artifact appears), then try other 
options. I personally prefer kerndeint (kernel de-interlacing), but it might 
not work for you -- you'll just have to try each one in turn and test it on 
your hardware. Take notes on each -- both the quality of the de-interlacing 
and on how much stuttering you see, since that's the problem you're trying to 
solve. Try to find a recording with lots of fast horizontal motion or panning 
to test the de-interlacing, and play the same video segment for each 
de-interlacing setting.

-- 
Rod Smith