POV-Ray : Newsgroups : povray.binaries.images : Red channel greater than 1.0 Server Time
22 Aug 2025 17:57:20 EDT (-0400)
  Red channel greater than 1.0 (Message 1 to 1 of 1)  
From: Cousin Ricky
Subject: Red channel greater than 1.0
Date: 19 Aug 2025 20:19:40
Message: <68a5149c@news.povray.org>
After Clarence1898 said he would have to try metallic reflection with
values greater than 1--forbidden fruit effect, except no one gets
punished--I figured I would try it myself.  As I figured would happen,
the mutual reflections saturated the red to the extent that it's
brighter than the original color; but otherwise, it's not too bad.

Some reddening of mutual orange reflections is inevitable with 3-channel
RGB rendering, but having the red channel greater than 1 really steps up
the effect.  The spectrally rendered image (36 channels) shows very
little hue drift in the reflections.

But I still remembered an overly bright render with runaway reflections
and a necessity to limit max_trace_level to 5, so I tried transparent
textures to see if that was what I remembered.  The left sphere in each
image uses filtered transparency, and the right sphere uses interior
fading.  Still not bad.

Turns out what I remembered was M_Glass2 from textures.inc.  Don't use
M_Glass2.  M_Glass2 is horrible.  In fact, don't use any of the glass
materials in textures.inc; roll your own using Fresnel reflection.

Other notes:
 - The assumed_gamma is 1.0.
 - For consistency with spectral rendering, the finish diffuse is set
   to 1.0.  To compensate, all the object colors are multiplied by
   Clarence's diffuse, 0.6.  Thus, 2.0 becomes 1.2.
 - You may notice that the RGB glass renders have a higher green channel
   than the spectral render; this is to mitigate the primary color bias
   (in this case, red) that always occurs with 3-channel rendering of
   colored transparency.  This bias is also seen in the metallic mutual
   reflections, and to a lesser extent in the radiosity of the non-
   metallic object, but the bias there cannot be corrected without
   changing the color of the unreflected surface.
 - For the record, using my SPD generator (see MakeSpectrum in p.object-
   collection), I can get up to rgb <2.2557, 0.5639, 0> at diffuse 0.6
   before the color IRL starts reflecting more red than it receives.


Post a reply to this message


Attachments:
Download 'hyper_orange-096.jpg' (43 KB) Download 'hyper_orange-120.jpg' (45 KB) Download 'hyper_orange-120sr.jpg' (44 KB) Download 'hyper_orange-glass-096.jpg' (34 KB) Download 'hyper_orange-glass-120.jpg' (36 KB) Download 'hyper_orange-glass-120sr.jpg' (34 KB)

Preview of image 'hyper_orange-096.jpg'
hyper_orange-096.jpg

Preview of image 'hyper_orange-120.jpg'
hyper_orange-120.jpg

Preview of image 'hyper_orange-120sr.jpg'
hyper_orange-120sr.jpg

Preview of image 'hyper_orange-glass-096.jpg'
hyper_orange-glass-096.jpg

Preview of image 'hyper_orange-glass-120.jpg'
hyper_orange-glass-120.jpg

Preview of image 'hyper_orange-glass-120sr.jpg'
hyper_orange-glass-120sr.jpg


 

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.