testing fixed blend

[BlessOd]

Hi. New Blend works very well. But reflections are cleaned from noise very quickly and highlights are still very noisy in complex Blend material((
simple scene (plane, teapot and 8 light)
1 picture - blend in blend  (create GGX BRDF) - 400 passes - very noisy highlight.
2 picture - a simple blend - less noise
3 picture - a simple mat - almost no noise
4 picture - blend in blend 1k passes - still noisy

render settings - default, image filter - off

[Ludvik Koutny]

It's not fixed. It's actually more broken than it was before ;)

[BlessOd]

It's not fixed. It's actually more broken than it was before ;)

yes, blends can not fix because it is broken at the base
but now reflection are cleaned from the noise much better

[romullus]

Rawalanche, could you elaborate more about it? Or is it better to keep from public?
I'm asking not because i'm overly curious, i just working on some scene where blending is quite heavily involved, so i'd like to know, should i postpone my experiments :]

[Ludvik Koutny]

Blending media (including SSS) does not work, blending bump (so each layer can have bump that does not affect the other layer) does not work, blended refraction does not work correctly... etc...

[romullus]

I see. So, basically, nothing changed, right? :]

[lacilaci]

well afaik recently was fixed only that it works now with hdcache... so... The question is: is there a corona layer/blend material in the works or not? Cause possibly it would be better solution than trying to fix how it works with max's standard blend i guess.

[Ondra]

Blend is completely independent of 3dsmax. CoronaBlendMtl would only add better UI (possibility to use multiple layers).

Only major problem is that all layers use bump mapping from single layer. Having multiple shading normals at a single surface point is just too brainfuckish to implement

Blending of media is not planned, because this setup is highly physically unplausible (you cannot have single volume with IOR 1.3 and IOR 1.5 at the same time. You can create materials with multiple refraction lobes, but for secondary rays exiting from the medium only single IOr will be used.

[romullus]

I hope this not means, that current blend implementation is feature complete. I largely miss shellac support.