New filter: Fix Telecined Fades

[KingLir]

Originally Posted by jackoneill

You'll have to convert back to some integer format before output.

Can you give me the line for the vs script ?

Originally Posted by jackoneill

Or nag feisty2 to add support for integer clips, because there's no reason not to.

feisty2 - you are officially naged

[feisty2]

Originally Posted by KingLir

feisty2 - you are officially naged

I don't think that would be a good idea simply because integer formats are evil..

with floating point formats (half, single or double if supported someday), black is always [0.0, 0.0, 0.0]
with integer formats, black could either be
full range YUV: [0, 1 << (bits - 1), 1 << (bits - 1)]
or
tv range YUV: [16 << (bits - 8), 1 << (bits - 1), 1 << (bits - 1)]
or
RGB: [0, 0, 0]
which is already super confusing to newbies and the base color might not be black... and that's gonna get even more complicated...

I'd say just convert whatever you have to floating point format and save all that crap, floating point formats are much more beautiful than integer stuff from the mathematical point of view.

[KingLir]

Originally Posted by feisty2

I'd say just convert whatever you have to floating point format and save all that crap, floating point formats are much more beautiful than integer stuff from the mathematical point of view.

I see, thanks. What line should I use in vs script ?

[ChaosKing]

Maybe I found a bug:

a2 = core.ftf.FixFades(a2, mode=0, threshold=0.0002, color=[0.56, 0.39, 0.22], opt=True)

I found this broken frame https://imgur.com/a/hCXq7

I know the threshold is lower (I was experimenting). But it does not happen with pure white/black color values + same thresh.