New TomsMoComp v 0.0.1.6

[trbarry]

I just posted a small update to my TomsMoComp filter on my site. It fixes the problem of always using SE=0 on YUY2 and also improves YV12 handling a number of places where it still had a hard coded ratio of 2 bits/pixel.

It now also supports the DBL_Resize function of SE=-1 on both YV12 and YUY2.

It did not take any other code changes but I can also verify the automatic top first option (TopFirst=-1) seems now to work reliably with MPEG2DEC3. I think Marc FD fixed that some months ago but I didn't know.

- Tom

[Piper]

<sigh> Another reason to skip out of work early. Thanks Tom!

[neuron2]

Tom,

Were you able to test the static scene handling with SE=1? I've never seen a version that leaves the frame alone as it should.

[trbarry]

Donald -

I can now get results for both YV12 and YUY2 that are fairly good but I can not make it leave the image totally unchanged. And I can not figure out why either. With your test file it should indeed be unchanged, but isn't.

But again, ignore the first frame (and the man behind the curtain). That is always done with SE=0 because there is no history yet. Though maybe I should cheat and look forward in that case ????

But for every pixel I step through in debug it finds 0 motion and uses the weave pixels. Yet the image is obviously changed a little. :(

BTW, if using my source, be sure to check the dates that I changed in the last 24 hours. There were a few places that had to have the BPP definition used, including one member that was in the zip but not even included in the project file, so I had not seen it previously.

- Tom

[neuron2]

@Tom

Thank you for the explanation. I'll do some debugging this weekend on it to see if I can find out why it is not performing completely as expected.