So, light is more or less magic when you get right down to it (Circular polarization works because light travels at different speeds along different axes? Not even going to try to understand that one, I can barley understand the most basic space and time relationships) and color is a phenomenon based off physical structures (sort of) which hurts my head. Iridescent surfaces are a great example of this, but I would still like to just think of everything as magic and let someone else deal with it.
In any case, different thicknesses of oxide layers create different colors, and the thickness in this case is dependent on the amount of heat delivered. 1/4″ Ti seems to work fine for getting consistent results, but if you go too thin the heat dissipation varies too much depending on how close you are to the edge. The spectrum is the same as that which you can get by anodizing the titanium.