PDA

View Full Version : VRay: What are the differences between primary and secondary bounces?


h4di
03-03-2011, 06:34 PM
Hi,

I searched for this issue and I found out that one difference is the exact and approx methods and that each of these methods has it's own pros and cons for calculating GI and setting two steps for calculating GI (primary and secondary) would let to take the benefit of having a very adaptive and time-saving rendering.

Reading VRay documents I assume that in VRay this 2-step rendering solution has another functionality. I think primary bounces are bounces that actually are visible and are in front of the camera and since all GI solutions in VRay are "Gathering" methods and not "Shooting". so in order to calculate GI correctly we need another step to calculate bounces that are happening off the screen and bounces that are happening on the back side of the objects that are not visible.

is this assumption correct?

thanks for shedding some light on this for me :)

thev
03-03-2011, 08:30 PM
The primary bounces are those directly visible to the camera (or through reflections/refractions). The secondary bounces happen after a ray has already went through at least one GI bounce.

It is useful to separate the bounces in this way because they have different effects on the final image, and therefore different methods might be more appropriate for each type.

Since primary bounces are directly visible, they have a great impact on the image quality, so the methods that compute them should be able to represent them accurately. The requirements for secondary bounces are not so strict as they are not directly visible, but are only used to calculate the primary bounces.

For example, the irradiance map uses quite a sophisticated interpolation algorithm to compute the final result; it gives good results for as a primary engine, but would be too slow as a secondary one.

The light cache on the other hand is able to calculate secondary bounces very efficiently, but it is not very detailed and uses very simple (but fast) interpolation methods. It is good for a secondary engine, but not very useful as a primary one.

Best regards,
Vlado

h4di
03-04-2011, 04:48 AM
Vlado, Thank you for your answer. :)

So my assumption is wrong? is it?
So according to your definition I correct My though of that:
"Primary bounces are the first bounces that happen in the scene. they are very important and very effective because they have the biggest impact on overall illumination and color bleeding. but after several bounces light energy decreases and although they still bounce through the scene but not effective as the first bounces. so we group these and name them as secondary bounces."
how about this one? is it right?

Also I came to know that light cache and seems to be a "shooting" solutions. cause I set up a scene. I calculated light cache for camera 01 as you see below. I saved the solution in a file and reloaded it and when I rendered from opposite direction there were no blackness. GI was calculated but less accurate for parts that were not visible in camera 01.

http://i52.tinypic.com/mh3gw7.jpg

http://i56.tinypic.com/aeo1ly.jpg

CGTalk Moderation
03-04-2011, 04:48 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.