Help understanding Secondary bounces

[ene.xis]

well not as scientific as my memory recall it...but it might help:

http://www.chaosgroup.com/forum/phpB...ondary+bounces

[skogzort]

Wow – That post did have some good info. Thank you for adding it. The color bleed answer and RErender’s illustration have shown me why I would want to have secondary bounces turned on in VRay. This is good enough for me. I’m grateful for the info and ready to move on.

For the sake of explaining my original confusion it occurs to me after reading through all of these posts again that I should have worded my questions differently. I have always understood what primary and secondary bounces are. What I didn’t understand is the practical reason why Vray has a separate multiplier for each.
Consider:
If you increase the Primary Bounce multiplier, then obviously the intensity of secondary bounces will be increased. If you decrease the Primary Bounce Multiplier, then obviously the intensity of secondary bounces will be decreased. Since the primary bounce multiplier effectively controls both, I could not understand what practical reason you would use a separate secondary bounce multiplier for. It’s like having two volume dials on a stereo. It seems like it would make more sense to have a simple on/off check box for secondary bounces.

Much thanks again for everyone who participated.

[ene.xis]

ur absolutely right...from a scientific point of view...
Nevertheless the ability of tweaking this 2 values provides gr8ter control to achieve the look u need...instead of working ur math untill u get to it!
freedom!
u gotta LUV Vray!

[joconnell]

Another reason for the primary / secondary bounce settings is that the primary or first bounce of light as mentioned earlier contributes a hell of a lot to the image and in certain cases such as light hitting off a bright surface, it can actually be quite detailed and sharp so you need a really high quality method of calculating it to get an accurate result. For secondary bounces, the light rays have already bounced off 2 surfaces (the direct light hit, then the first GI bounce) and unless you're in a perfect environment where everything is a mirror like surface or flawlessly smooth white objects, the light will generally start to spread out and diffuse - the net results of this is that there aren't as many defined details in the lighting as there were in the first bounce or the direct light.



As you can see in the picture above, theres a distinct line at the corners of the white box where the light is bouncing off it so there's quite fine detail that we need to get in our orimary bounce.

Now on to vray itself, we have a choice of 4 different gi methods, each with different benefits. QMC and PPT are both very high quality, raytracing type solutions where each pixel in the image is sampled so you get extremely detailed but unfortunately grainy images because of the level of detail being returned in the global illumination. Irmap and lightcache are approximated methods so they aren't as realistic as QMC or PPT, they do on the other hand have the benefit of speed seeing as they both take shortcuts - they aren't sampling each pixel individually, they're taking a broader view of the GI solution.

So what does that mean in terms of primary and secondary? Well for primary we want to capture all of the detail in the lighting we can - seeing as the primary bounce contains most of the light in the GI solution we want to capture this as accurately as possible so we need a high quality method such as QMC or irradiance mapping with its settings turned up quite high. For the secondaries, we dont need anywhere near the same quality the main reason being that the quality and detail probably doesnt exist in the light in the first place - the more light bounces around, the more its chances of hitting an uneven surface which will scatter the light and diffuse it - the secondary bounce tends to be a lot softer and more blurry and if we were to use a high quality method like QMC then we'd possibly introduce grain due to the way it works and we'd en up turning up our samples to end up with a clean soft secondary bounce. The alternative to that is to use an approximating method like light cache or photon mapping which will do a less accurate lighting solution and take shortcuts - the good thing about that is there isn't any sharp details in the secondary bounces anyway so the lower quality methods blurry results actually suit our purpose perfectly. Since they end up using blurring techniques to smooth out their results its giving us a result that we want without having to resort to really high samples as we'd need for something like qmc.

Vray is pretty much giving you the option of splitting your GI methods so that you get really high quality in the primary where it's needed and speedy, clean results in the secondaries where you can cheat a little to bring render times down.

[RErender]

Ooooh, nice arrows joconnel!

[ene.xis]

nahhh....!they r not even close to urs!

[cpnichols]

Some additional reading:
http://www.spot3d.com/vray/help/150R1/gimethods.htm

[Morbid Angel]

lol those are spectra waves For some reason he only has R and B no G

[percydaman]

lol those are spectra waves For some reason he only has R and B no G

[RErender]

Still no donations?

I only want the Cintiq 21UX, come on get those donations rolling in!

Thanks!

[RErender]

lol those are spectra waves For some reason he only has R and B no G

Actually the green is right there where the red and blue overlap denoting the transitional zone between primary and secondary...

[mmikee]

LOL

[glyph]

my 0.5 cents worth:
that's why primary can't go above 1.0 - you'd get light feedback
but secondary can go higher than 1.0