rgb2lab is returning unexpected results?

This is the code I'm running, nothing fancy :

>> rgb2lab([0 1 0],'ColorSpace','sRGB')

ans =

87.7347 -86.1827 83.1793

There must be an argument I'm missing, somewhere (?), because the result does not match all that well what other ICC-aware applications are leading me to believe about the "color appearance" of sRGB Green (RGB = 0 255 0) or [0 1 0], which is CIE Lab 88 -79 81.

87.7341 is perfect but -86.1827 and 83.1793 are not right.

I tried to use arguments like "Adapted White Point" but those are not defined in the function parameters, as far as I can humbly tell.

I tried with cform and applycform but the result is not directly interpretable in CIE Lab, so I can't tell for sure.

(I'll have to figure this one out one of these days, for my own curiosity: I guess the L is expressed on an 8bit 0-255 scale, so that's not the end of the world but the a* and b*? Hmmm. Something about subtracting or adding 128... But that's as far as I'm willing to go... for now)

[Image Analyst]

What white points are you using? I've let the developers know and they might answer. In the meantime you can

>> edit rgb2lab.m

and see what you can learn though most of the computations are done outside that in a DLL.

[Roger Breton]

That's the "key"!!!!!

To match Photoshop and all other "ICC-based" applications, the white point has to be D50.

Silly me :-)

Now the result is perfect :

>> rgb2lab([0 1 0],'WhitePoint','d50')

ans =

87.8184 -79.2871 80.9893

Thank YOU SO VERY MUCH!!!!!!

Answer 1

[Image Analyst]

Set the white point to be D50.

[Roger Breton]

Never thought of chromatic adaptation... Egg on my face...

[Image Analyst]

Hey, don't feel bad. Everyone makes mistakes. And I've talked to the worlds leading experts in person and even they'll admit they don't know everything. This stuff is confusing even for them - that's why research keeps going on. We don't know everything yet. Because we have physics and optics intertwined with physiology and psychology and anatomical/biological differences, color science is one of the most bewildering fields of science. I'm sure you know lots of things that I don't.