For years I'd noticed problems with dispersion and dispersion_samples, but  
I never put together a test suite until last June.  In October, C.Lipka  
posted some tests to p.b.i on what appears to be the same matter.  Since  
then, I rewrote my tests.  These are the results:

dispersion_view-362.png - A view through a prism, version 3.6.2.  Note  
that the number of samples is low by 2.  Version 3.7 beta 34 had this, as  
well as other problems.

dispersion_view-37b35.png - A view through a prism, version 3.7 beta 35.

dispersion_comp-37b34.png - A comparision of two glass hemispheres, the  
right with dispersion, the left without.  Note the color differences.   
This used version 3.7 beta 34, but version 3.6.2 had the same results.

dispersion_comp-37b35.png - The same comparison using version 3.7 beta 35.

I did other tests using photons, but this appears to need more work in the  
beta test versions, so I have not posted them.

A great big thanks to C.Lipka for all his hard work on this matter.


-- 
<Insert witty .sig here>

Post a reply to this message

Attachments:
Download 'dispersion_view-362.png' (61 KB) Download 'dispersion_view-37b35.png' (63 KB) Download 'dispersion_comp-37b34.png' (111 KB) Download 'dispersion_comp-37b35.png' (64 KB)

Preview of image 'dispersion_view-362.png'

Preview of image 'dispersion_view-37b35.png'

Preview of image 'dispersion_comp-37b34.png'

Preview of image 'dispersion_comp-37b35.png'