POV-Ray export questions

[Christoph Lipka]

I still don't think it's preferable. If you turn on radiosity, then the emission will light up neighboring objects. This is not realistic. You can get perfectly good looking results without using emission.

That is exactly what it suposed to do as far I understand it. It simulates the global scattering of lights you have in the real world whose lights seemingly come from all directions (== global or ambient lighting).

If you don't like it in the LDCad export you can always set it colorEmission to 0.0 or rewrite the material generation macro's (e.g. ldrawTexPlastic in above pov file)  as those are read from external files.

If I, as the designer of both POV-Ray's current gamma handling framework and the "emission" mechanism, may chime in:

• The "emission" mechanism was introduced specifically to allow for the "ambient" mechanism to be turned off in radiosity scenes, while still allowing to define materials that emit light all by themselves. You see, radiosity and "ambient" were essentially designed to model one and the same thing: Illumination due to light scattered off other diffuse objects, lighting up even areas that are in total shadow. "Ambient" is a quick but dirty approach adding a constant term, while radiosity is a more precise but costly approach essentialy computing this type of illumination for each point on the surface. Having two mechanisms enabled at the same time to model this effect is pointless, and hence "ambient" is now automatically suppressed if radiosity is switched on. The "emission" mechanism was added as a replacement for the cases where "ambient" was co-opted to model something entirely different, namely materials glowing all by themselves.
  If the "colorEmission" parameter is supposed to model global scattering of lights, then you should use "ambient". If instead it is supposed to model glow-in-the-dark materials, then "emission" is the mechanism to use.
  
• For realistic results (and who wouldn't want LDraw models rendered as realistically as possible!) "assumed_gamma 1.0" is a must, and this may also fix the issues you see with image-based textures appearing too dark. It is a known fact that using this mode increases the overall brightness in a scene and also reduces colour saturation, unless parameters are adjusted to compensate. You're already doing the most important thing: Using "srgb" instead of "rgb" in pigment definitions. However, it has recently been brought to attention that the finish parameters also needs some care: In general, any "ambient", "emission", "diffuse", "phong", "specular" or "reflection" setting lower than 1.0 should be set to an even lower values, with x^2.2 being a good starting point. It is impossible to give an exact formula, as the colour math in "assumed_gamma 2.2" mode is physically wrong, and thus the results in the physically accurate "assumed_gamma 1.0" mode will inevitably differ in one place or another.
  

I hope this helps. If you need more information or assistance, let me know.

As a side note, I'd like to point out that for optimal realism the index of refraction should be around 1.6, and that realism may benefit from fresnel-based variable reflection. Also, personally I'd recommend "specular" instead of "phong" for highlights, but that may just be personal preference.