## "Pink Elephant" Saturated Color Absorption Noise

Discussion related to the 'LuxBlend' Blender 2.4+ python exporter script.

Moderators: Ratow, SATtva, coordinators

### "Pink Elephant" Saturated Color Absorption Noise

Hi

Let's say i choose some RGB color from LuxBlend color picer and set abs at depth to 0.002.

It comes out for me as a news it's big no no if from 3 of the R, G and B values one value is 0 and another is 1.
Result is pink noise with material that looks like that:

Code: Select all
MakeNamedVolume "Blue" "clear"   "float fresnel" [1.458999991416931]   "color absorption" [34538.77475441 312.70161396 0.00000000]MakeNamedMaterial "Blue"   "bool architectural" ["false"]   "bool dispersion" ["true"]   "string type" ["glass2"]

Now i replace 0 with 0.05 and 1 with 0.9 or 0.95 and all is fine. Absorption looks like that now:
Code: Select all
   "color absorption" [802.50032607 209.59308668 52.68026857]

So my idea was in that we limit out 0 and 1 values from color picer.

I do get some color elephants even with one value beeing 0 or 1.

Meelis

Posts: 873
Joined: Sat Oct 17, 2009 2:16 am

### Re: "Pink Elephant" Saturated Color Absorption Noise

Well, when you think about it, it makes sense. You wind up with a volume that transmits some wavelengths perfectly, and others it blocks completely.

Unfortunately, I don't think we can define min/max values for color. Locking 1.0 would also prevent you from having a volume with 0 absorption, which I've found is useful sometimes, either with glass2 or using a homogeneous volume. (sometimes you just want scattering or IOR without absorption. Not really correct, but useful)
-Jason

J the Ninja

Posts: 2210
Joined: Wed May 19, 2010 9:54 pm
Location: Portland, USA

### Re: "Pink Elephant" Saturated Color Absorption Noise

J the Ninja wrote:Well, when you think about it, it makes sense. You wind up with a volume that transmits some wavelengths perfectly, and others it blocks completely.

Unfortunately, I don't think we can define min/max values for color. Locking 1.0 would also prevent you from having a volume with 0 absorption, which I've found is useful sometimes, either with glass2 or using a homogeneous volume. (sometimes you just want scattering or IOR without absorption. Not really correct, but useful)

Good point.

Perhaps we can add tiny itty pitty pinke elephant icon by 16x16 or 32x32 pix to color picer,
in case wnen 0 or 1 is choosen then pink elephant icon will appear MAGICALLY from nowhere
For newbies

Meelis

Posts: 873
Joined: Sat Oct 17, 2009 2:16 am