# Material Properties

Dear list,

You might try Axel's Color picker for Radiance

Oooh, ouh, steady on now. The CP4R is still purely experimental. I've
not had enough feedback to release it to the big wide world. I have
actually been breeding over a rather serious question which might
potentially question the whole idea behind it. But do read on...

Let's just say that somebody has gone through the exercise of
measuring all RAL colours with a chroma meter, based on a set of
colour swatches. The readings are presented in XYZ format. Let's
further assume, just for the sake of the argument, that this somebody
has written a little script to convert those XYZ readings to RGB,
using Radiance's xyz_rgb.cal. Here is an example:

-- this guy's ---
#Radiance rgb values for RAL colors converted from X Y Z color coordinates
#X Y Z values measured with Minolta CM-2600D Illuminant D65 observer 10 degree
# RAL 8023 orangebraun
# X=18.9 Y=14.6 Z=3.2
# reflectance 0.1460
void plastic RAL8023
0
0
5 0.3017 0.0971 0.0150 0 0

-- from Radiance colour picker (data from the Internet, don't remember
where from) --
# RAL colour: 8023 Orange brown
# Reflectance: rho=0.476
void plastic identifier
0
0
5 0.676 0.426 0.176 0 0

-- from Radiance colour picker (converted from XYZ data) --
# Imported from XYZ: ( 18.9, 14.6, 3.2 )
# Reflectance: rho=0.404
void plastic identifier
0
0
5 0.644 0.335 0.126 0 0

Checked against this web site:
http://easyrgb.com/index.php?X=CALC
which gives pretty much the same results: (0.64380, 0.33552, 0.12613)
Also compare:
http://www.brucelindbloom.com/index.html?Eqn_XYZ_to_RGB.html

Your will appreciate that the measured values, converted to RGB are
wayyyyyy off mark.

that the conversion algo is taken from:
"Procedural Elements for Computer Graphics" by David Rodgers (McGraw Hill),
which is sitting right next to me.

I have not gone through it yet to check for typos in yxz_rgb.cal, but
I find it unlikely that Greg has made such a mistake. It's far more
likely that I'm missing something here. One possible explanation would
be the primaries: xyz_rgb.cal uses the standard CRT primaries, whereas
it's unclear as to what the others are assuming.

Any enlightenment on the issue would be greatly appreciated.

Thanks for your feedback on this issue

Axel

1 Like

Hi Axel,

Indeed, the RGB<->XYZ conversion in xyz_rgb.cal follows the derivation in Roger's book, and it's been checked pretty thoroughly. As you note, the actual results are determined by the choice of RGB primaries and white point, and this transform assumes you are going for absolute rather than perceptually white-adjusted values. The primaries chosen are the standard ones for Radiance, which are a near miss to the CCIR-709 recommendation used for the sRGB color space and most modern computer monitors, which is given in xyz_srgb.cal along with a D65 (versus equal-energy) white point.

What other people have used for their color conversions is anyone's guess.

By the way, did you mean to respond to the radiance-general posting on radiance-dev as you did? I'm not sure Nick and Steve are on this list....

Cheers,
-Greg

ยทยทยท

From: "Axel Jacobs" <[email protected]>
Date: June 10, 2008 2:20:45 PM PDT

Dear list,

You might try Axel's Color picker for Radiance

Oooh, ouh, steady on now. The CP4R is still purely experimental. I've
not had enough feedback to release it to the big wide world. I have
actually been breeding over a rather serious question which might
potentially question the whole idea behind it. But do read on...

Let's just say that somebody has gone through the exercise of
measuring all RAL colours with a chroma meter, based on a set of
colour swatches. The readings are presented in XYZ format. Let's
further assume, just for the sake of the argument, that this somebody
has written a little script to convert those XYZ readings to RGB,
using Radiance's xyz_rgb.cal. Here is an example:

-- this guy's ---
#Radiance rgb values for RAL colors converted from X Y Z color coordinates
#X Y Z values measured with Minolta CM-2600D Illuminant D65 observer 10 degree
# RAL 8023 orangebraun
# X=18.9 Y=14.6 Z=3.2
# reflectance 0.1460
void plastic RAL8023
0
5 0.3017 0.0971 0.0150 0 0

-- from Radiance colour picker (data from the Internet, don't remember
where from) --
# RAL colour: 8023 Orange brown
# Reflectance: rho=0.476
void plastic identifier
0
5 0.676 0.426 0.176 0 0

-- from Radiance colour picker (converted from XYZ data) --
# Imported from XYZ: ( 18.9, 14.6, 3.2 )
# Reflectance: rho=0.404
void plastic identifier
0
5 0.644 0.335 0.126 0 0

Checked against this web site:
http://easyrgb.com/index.php?X=CALC
which gives pretty much the same results: (0.64380, 0.33552, 0.12613)
Also compare:
http://www.brucelindbloom.com/index.html?Eqn_XYZ_to_RGB.html

Your will appreciate that the measured values, converted to RGB are
wayyyyyy off mark.

that the conversion algo is taken from:
"Procedural Elements for Computer Graphics" by David Rodgers (McGraw Hill),
which is sitting right next to me.

I have not gone through it yet to check for typos in yxz_rgb.cal, but
I find it unlikely that Greg has made such a mistake. It's far more
likely that I'm missing something here. One possible explanation would
be the primaries: xyz_rgb.cal uses the standard CRT primaries, whereas
it's unclear as to what the others are assuming.

Any enlightenment on the issue would be greatly appreciated.

Thanks for your feedback on this issue

Axel

Is there some kind of library out there which would have typical spec,
rough, trans, tspec for various common materials? If not, could this be
found on an individual basis other than testing in real life or trial and

I haven't determined the material I want to use yet, so perhaps that is the
first step, but for instance if I picked a specific type of cut
acrylic block to cover a light...

void trans acrylic

0

0
7 red green blue spec rough trans tspec

Thanks,
Nick

You might try Axel's Color picker for Radiance

regards
Steve

Is there some kind of library out there which would have typical spec, rough, trans, tspec for various common materials? If not, could this be found on an individual basis other than testing in real life or trial and error in Radiance?

I haven't determined the material I want to use yet, so perhaps that is the first step, but for instance if I picked a specific type of cut acrylic block to cover a light...

void trans acrylic
0
0
7 red green blue spec rough trans tspec

Thanks,
Nick

ยทยทยท

Date: Tue, 10 Jun 2008 13:09:10 -0400From: [email protected]: [email protected]: [Radiance-general] Material Properties
_________________________________________________________________
Try Chicktionary, a game that tests how many words you can form from the letters given. Find this and more puzzles at Live Search Games!
http://g.msn.ca/ca55/207