More on this topic after further testing: comparing a pure composite in
feed shows different results too. Here's how that looks:
http://capnbry.net/~bmayland/fi/pvr150/pvr150-250-2.jpg
In the bottom image (the 150), you can see in areas like the comedy ham
that there's a good deal of noise which blurs the edges of the ham.
Looks like luma being interpreted as chroma. Disabling the comb filters
made things even worse, so I'm looking squarely at the notch filter. So
I took a look at some black and white source material:
http://capnbry.net/~bmayland/fi/pvr150/modecompare.png
Again, the 250 produces a sharper image than the 150. The bottom image
is the 150 too, except for I have disabled the notch filter altogether
(COMB_NOTCH_MODE=0). I think it looks the best out of the three.
So we've got some potential here. We might be able to increase the
error limit before the algorithm falls back to notch, or even just chose
a different notch mode. Something else, the saa7115 selects a
"resolution enhancement filter", which amps luma 5.1dB at 4.1MHz. This
is close to setting PEAK_SEL to +5 and turning on PEAK_EN on the cx25840.
Lots of avenues still left to explore, but the cx25840 does give use
more "sliders" to play with. I'm betting we can get the same quality if
not better out of these cards.
Bryan Mayland wrote:
I was under the impression that the cx25840 used in the PVR-150 was a
better digitizer than the sa7115 used in the PVR-250. If that's true,
then something is up because it looks the opposite to me:
http://capnbry.net/~bmayland/fi/pvr150/pvr150-250.jpg
(This is tuner input with dnr_mode=3,temporal=0,spatial=0. I've also
tried switching cables between the two cards)
-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
ivtv-devel mailing list
ivtv-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ivtv-devel
