1.5 DB 13-7-2016 - SSS Artifacts and slowness

[pokoy]

The latest 1.5 DB shows some problems when SSS is used, see attached images. When using SSS the material displays facets, this didn't happen in 1.4.
Also, it seems that SSS is slower, mostly noticeable using maps in diffuse/reflection etc channel. Displaying a simple material preview takes 4 seconds with some maps for example, with SSS enabled (setting Absorption distance to >1) the time it takes to display the preview doubles.

[pokoy]

Sorry, forgot to add that the facets only appear when using SSS with refraction, using fraction only works fine and doesn't show them, it's also rendering faster than with refraction.

[maru]

I don't think the faceting is something new, see:
https://coronarenderer.freshdesk.com/support/solutions/articles/12000012390-my-sss-objects-are-faceted-

But it would be great if you could check if there is any difference between the newest daily and any other builds (1.4? 1.3?) nonetheless.

[pokoy]

Never came across this, probably because I didn't care to open the material preview in a bigger window - my bad.

Still, I'm seeing slower update times if SSS is based on refraction... I'm quite certain this was not the case before.

[Ondra]

Can you provide a scene/material sample where a significant difference (either in quality or speed) happens?

[pokoy]

This happened on a fresh simple scene, with a sphere with an SSS material - nothing complicated really.
The difference was noticeable between the two SSS modes - fraction and refraction - where the refraction mode was the one that took longer to update and render.

[Ondra]

still though, without solid evidence, there is no bug. It is perfectly normal that the different modes render differently, I was under the impression that it was not the case before - that would indeed be a bug

[pokoy]

still though, without solid evidence, there is no bug. It is perfectly normal that the different modes render differently, I was under the impression that it was not the case before - that would indeed be a bug

I didn't test against 1.3, and as you say no hard evidence, more along the lines of 'this used to be faster'.
Maybe move to post to resolved and I'll keep an eye on this and report back once I have a more solid impression.

[Ondra]

yep... it is nothing personal, just my experience that "feeling this used to be faster" disappears when compared with older version ;)

[pokoy]

Yes, probably safe to assume it's in the same category ;)
Thanks!