Exposure Differences Between GI Engines

[marcdevon]

I don't know if anyone else here has previously identified an issue regarding exposure differences between the GI engines but I wanted to bring this issue to light. I noticed that a scene rendered with PPT seems to be exposed differently than a scene rendered with IM + LC GI. The images below illustrate this issue (btw, please excuse the quality). Both scenes are identical. The cameras are each set to f-number 16.0, shutter speed 1/15s, and ISO 200. However, notice how the PPT render (right) is significantly brighter than the IM + LC render (left).




The only conclusion that I have is that PPT image as rendered is exposed approximately two (2) f-stops higher than the IM + LC image. My conclusion is based on increasing the number of stops of the IM + LC renders by +2.00 (see below). Notice how the exposure of these images are nearly identical to those rendered in by PPT. However, I do not know which image is truly rendered at f-number 16.0, shutter speed 1/15s, and ISO 200.



Any thoughts? Any solutions? Has anyone else encountered the same conundrum?

[vlado]

I can't reproduce this on a simple test, but there may be different reasons for this. How many subdivs are you using for the light cache in the IM+LC case? Have you changed any GI multipliers (primary/secondary); any of the GI post-processing controls? Any unusual settings for the lights?

Best regards,
Vlado

[marcdevon]

In the above post I set the subdivisions for the light cache to 5000. The primary and secondary bounce multipliers are both set at 1. I checked the environment and effects to confirm that there is no exposure control. GI post processing controls are set as follows:

Saturation: 0.65
Contrast: 1.0
Contrast base: 0.5

I tried another render (see below) and I set the subdivisions for the light cache to 12000. As you can see, I am getting the same result.



FYI

I am using VRay Version 1.50 SP2 with 3ds Max Design 2009 Service Pack 1. I have rendered all images using LWF. I have enabled gamma correction in 3ds Max. Gamma is set to 2.2. Affect Color Selectors and Affect Material Editor is selected. For bitmap files, input gamma is set to 2.2 with the output gamma set to 1.0. All colors use colorcorrect with the gamma set to 2.2.

[Codi]

Are you using subpixel-mapping? If active it can produce different light intensities between bruteforce and a IRmap-LC renders.

[marcdevon]

Are you using subpixel-mapping? If active it can produce different light intensities between bruteforce and a IRmap-LC renders.

Subpixel mapping has been disabled in the renderings.

[marcdevon]

Has anyone with Chaosgroup identified the cause of this problem. I can send to Chaos Group the files I used for the renderings for potential bug identification.

[vlado]

The scene file will be useful, yes. I have not been able to reproduce this otherwise.

Best regards,
Vlado

[MoonDoggie]

It certainly doesn't look like this is the case (If it were the inverse would be true, the imap one being significantly brighter) but another cause for this issue is vray lights set to "store with irradiance map" which get ignored during (for example brute force,light cache).