Mark Wagner wrote:

> If my reading of the source is correct, #2 will do *almost* the right thing. 
> What should produce correct results is scaling the water slightly larger
> than the hollow of the glass:

Actually, I tried that first, but it reduced render speed so much
I thought there must be a lot of reflection or similar going on
between those two surfaces, so i tried to merge the internal
surface away (hence test #3).

I just ran a non-anti-aliased render of your suggestion which still
runs in bearable time. The result actually looks almost identical to
the classic "smaller" scaling, except for some differences with light
on the lemon and its reflection at the other side of the glass.

Note this picture uses the old "incorrect" finish without
energy conservation to allow comparison to bubbles-800.jpg
of my original post.

Post a reply to this message

Attachments:
Download 'bubbles-800-noaa-scaleplus-nomerge.jpg' (22 KB)

Preview of image 'bubbles-800-noaa-scaleplus-nomerge.jpg'