Currently the colors are more or less random, based on the material properties of the layer such that layers with equal material properties have the same color. A better solution should have the following additional properties:
Small changes to the material shouldn't lead to a totally different color
Restrict the possible color space to something that looks good
First idea: Define a starting color c0 and two directions in the color space, c1 and c2. Then, define normalized material properties m1 = E/E_ref and m2 = rho/rho_ref. Calculate the layer colors as c = c0 + m1*c1 + m2*c2. Unclear question: How should the reference values be chosen? Are there values that work for most real world applications?
Currently the colors are more or less random, based on the material properties of the layer such that layers with equal material properties have the same color. A better solution should have the following additional properties:
First idea: Define a starting color
c0and two directions in the color space,c1andc2. Then, define normalized material propertiesm1 = E/E_refandm2 = rho/rho_ref. Calculate the layer colors asc = c0 + m1*c1 + m2*c2. Unclear question: How should the reference values be chosen? Are there values that work for most real world applications?