Emitter plane and backlit pusbuttons

[treddie]

I'm not going crazy then. I'm only getting .35Ms/sec on the render on a 560.

[Refracty]

point 2 : put in RGB color into the distribution as well

[Magog]

Perfect, now you can insert the material into LiveDB!

[treddie]

point 2 : put in RGB color into the distribution as well

I'm curious...What does the distribution do, if the light source does not use an image or an IES file for the pattern?

Perfect, now you can insert the material into LiveDB!

Yes, I think this could go into the DB. Just need to do some final tweaks.

[treddie]

Here are my final backlit pushbutton results and attached materials.

What I found was that in order to get enough light to come out of the pushbuttons, I found that the suggestions by Chris about putting a white reflector inside the button, and by Marcus for putting in a larger emitter were really essential in order for the noise to clear up in a reasonable time. I experimented and found that a diffuse white lamp housing was best and that the emitter had to be quite a bit larger than the tiny filament-sized emitter I was using for the lamp. My reasoning was that the tiny emitter would create a circular halo around the light source when viewed through the button. That part worked OK but noise took forever to clear up. So as you can see in the first image, the lamp is significantly larger and is not just a flat surface, but a domed mesh to get the light to bounce around better and the white interior walls of the lamp/button housing were a big help in that respect.

In addition, there was a suggestion to go with a single-walled mesh (no wall thickness) for the translucent plastic to also speed up render times. It worked, but the effect was less realistic and there were rendering artifacts that seemed a bit odd (see marked up image at bottom).

Below are three images. The first is with thick-walled pushbutton plastic, and the second and third are with the single-walled buttons:

Below, not as realistic as above, but fairly close and not without issues:

Below, the orange arrows point to weird render artifacts that don't seem to come from anywhere in the model. The red arrows point to odd behavior at the extreme edges of the buttons.

One note...Maybe because of the relative strength of different wavelengths of light, I found that as I worked my button plastic colors from short blue wavelengths down to long wavelengths (red), that emitter power had to be increased along the way. Red was such a problem that it had to be REALLY intense. Another problem with red, is that because the Octane lighting model is based on incandescence, that a red light tended to look too orange even in the best of conditions, and raising the power of the red light only made the problem worse. All of the button lamps are white, or off white except for the red pushbutton's lamp. That lamp had to be red. And instead of using RGBspectrum for the emission color, I used GuassianSpectrum and cut the bandwidth down to a really narrow limit. That made the majority of the orange and yellowing go away but also made the power of the lamp very weak. So that required raising the intensity of the red lamp even more...which brought back a little bit more of the orangishness, but not too badly, so long as I didn't raise the lamp intensity too high.


Finally, here are the materials in a zip:

Backlit Pushbutton Materials.zip

(9.71 KiB) Downloaded 180 times

[FooZe]

Very nice.
Looks like all the hard work and tweaking paid off!

For the red buttons have you tried beta 3.03? This may help as there have been changes made to the color system: http://render.otoy.com/forum/viewtopic.php?f=9&t=23669.

Cheers
Chris.

[treddie]

Oh cool! Thanks for letting me know. I'll install it and see what goes.

Yah...I never was able to get this in Maxwell...noise would never clear up.

OH! Forgot to mention that if anyone looks at the actual material nodes I uploaded, there is no sub-surface used. It is all just surface roughness. Seems counter-intuitive maybe, but adding sub-surface to the buttons does not add any advantage and even seems to work against the effect.