PDA

View Full Version : Bug in the masktools ?


Manao
11th November 2003, 18:06
I think there is a bug in the masktools. With a basic script like this :
result = source.MaskedMerge(source,mask,Y=3,U=3,V=3)
return resultI don't get the original input ( whatever the mask : for example, mask = source, or a 'real' mask ). The picture is "greenish". When I look at the documentation, I see that maskedmerge should return (255-mask)*source + mask*source, but looking in the code of MMerge_C, I see it returns (256-mask)*source + mask*source. However, the code of MMerge_MMX seems correct ( but I'm far from an assembler expert, especially since I've just discovered the different MMX operations ), and it should be this one which is used, since I have an athlon xp.

Can anyone confirm the bug, and recompile a version of the masktools at least with a correct MMerge_C function ?

And, by the way, why does the author of the masktools say at the end of his readme that development is stopped because users prefers ConditionnalFiltering : it isn't the same thing ( or I didn't understand how works ConditionnalFiltering ).

Richard Berg
13th November 2003, 09:00
Conditioning works by actually forking the filter chain. (Not a perfect explanation, but that's what it looks like to a user.)

Masks do all of their work in-place. Rest assured I still think they're a cool feature, and future versions of AVS will provide greater support.

sh0dan
13th November 2003, 11:48
MaskTools is a truly excellent filter pack, and I have no idea why Kuruso stopped development, and even took down his page.

I would be a great addition to AviSynth, but since it's unsupported, and still has a few quirks I don't feel it is right to include it.