Couple of precision issues here:
- Interpolation was rounding trunkcating colors, and because
of some precision issues value 254.999 became 254 leading
to troubles later.
Now color interpolaiton will do rounding to nearest int.
- Subtract was setting channels to something negative which
confused color management (especially negative alpha),
Now subtract effect will clamp channels to 0 from bottom
and also do some tricks to prevent small alpha which
could also confuse color management.
Originally issue was discovered when using stabilization and movie distortion
nodes, but in fact issue was caused by render layer node always doing nearest
interpolation. Now made it so this node will respect sampler passed to it's
executePixel function and do an interpolation.
Added two new functions to do bilinear/bicubic interpolation in float buffer
with variable number of components per element, so it could interpolate 1, 3
and 4 component vectors. This functions currently mostly duplicates the same
functions from imageprocess.c and it should actually be de-duplicated. Think
it's ok to leave a bit of time with such duplication, since functions should
be generalized one more time to support byte buffers, which could backfire on
readability.
Also removed mark as complex from stabilization node, which isn't needed sine
int fact this node is not complex.
