Consider the image below:
The four grey patches in the centre of the four squares are phsyically the same. If seen in isolation (or against a common background) they would look the same. So are the grey colours the same or different?
Some people would argue the two greys are really different (this is convincing because after all they look different). However, others would argue that they are the same colour – it’s just an illusion that they look different because of the background and context. Which is right? Actually, I think both are correct. The former would be talking about perceptual colour and the latter would be talking about physical colour. Part of the reason that colour is so complex and that there seem to be so many ideas that clash is that colour is not very well defined. When people talk about colour they are sometimes talking about different things. I am wondering whether it would be really helpful to develop more formally the ideas of physical and perceptual colour.
In case on your screen the greys look more similar that they do on my screen I append another example of colour contrast below. 
In this example, there are only two physical colours (red and green) or three (if you include white). However, perceptually there are two reds and two greens.
“Physical colour” seems an unfortunate choice of words, and would certainly be misconstrued as supporting the objectivist viewpoint that has been “enriching the debate about colour” lately. I’m sure you would agree that all colours are perceptions, and that the idea that objects have a single physical colour irrespective of viewing conditions is an illusion. What about adopting the painter’s term “local colour” for this purpose? This term is generally understood to mean the colour of the isolated surface viewed under “white” light, as opposed to its appearance under the varying conditions of the scene. (Used scientifically, the precise viewing conditions would be implicit in the units that you use- Munsell, L*a*b*, etc).
David Briggs
Hi David
I like the term “local” colour. That’s interesting.
I agree that physical colour has certain implications that may not be desirable. You’re right – I do think that colour is perceptual and is no more a property of an object than is pain the property of a hot frying pan. Interestingly though, I am not sure that my view (and perhaps yours) is the prevailing view. A great number of people do take the view that objects have a physical colour.
Steve
As a graphic design student, I think the perceptual colour is the most important than the physical colour, the perceptual colour is what the person applying the colour has to consider when designing. They should consider the viewer, the conditions the colour will be viewed in as well as the other colours around or influencing the design. Each of these factors (and more) will influence how the colour is perceived and has to be taken into consideration with every design a person does.
I like a lot this kind of illusion, when you see two perceptual colours for a same physical one. For the last illusion, I saw that one a while ago and I found it spectacular : http://www.cuartoderecha.com/2823/
Romain Bachy
Steve,
have you ever heard of the EDGE COLOR? I read for the first time about it in Jan Koenderink’s new book COlor for the Sciences. I’m trying to figure out a way to implement it in Matlab, but it’s tough.
Thank you,
Matteo
OOOPS
Scratch that. I’m so used to work in Lab or L Ch and for once it was hard to think of a solution. But fter some thought, this one should be easy to implement directly in RGB space. I will post it on the Matlab file exchange when I manage to write the code. I will post an update here.
Matteo
Thanks Matteo – look forward to that!
Steve
Hi Steve
Here it is. Comments and ideas are welcome.
Cheers,
Matteo
http://www.mathworks.cn/matlabcentral/fileexchange/28982-perceptually-improved-colormaps
I also wanted to share link to this incredible collection of visual phenomena and optical illusions
http://www.michaelbach.de/ot/index.html
Have fun!
Matteo