Gamma Correct RGB Issues

[dalomar_cg]

So while testing the Vray Sun/Sky I ran into an issue with the reflections of the sky on the chrome sphere. Originally I thought the sky was not being gamma corrected in reflections, but it turned out to be a different issue. With Correct RGB enabled the color that was used for the reflection value was corrected as well, thus making my reflections darker than desired. I feel that the RGB values for certain values should be disregarded when it comes to correcting for Gamma. Any color that references intensity values should be disregarded for gamma correction as the color doesn't need to be corrected. I'm not sure if this effects other intensity values such as transparency, but I will test it. And this should probably extend to bitmaps as well.

Here are my examples

Reflections of Sky are fine and correct values


Reflections of Sky are darker due to gamma corrected reflection intensity

[Gijs]

Re: Gamma Correct RGB Issues

IMHO this is not a bug and a desired effect. If you want an object to be 50% reflective, it should be 50% reflective in gamma corrected mode. This means if you assign it a reflective color of RGB 128,128,128 and you apply gamma correction to the displayed image, the reflection will be more than 50%.
Vray needs to get all input offered at gamma=1 in order to create physically correct renderings, so all RGB values need to be corrected. I think the main reason that you want it not to be affected, is because you have become used to picking a certain value for a certain effect and you will probably have chosen darker colors intuitively to get the right output.
What value did you use for the reflective layer?

[Gijs]

Re: Gamma Correct RGB Issues

btw: in your first image, it appears to me that the reflections are lighter than the sky and ground itself, which is of course not realistic.

[dalomar_cg]

Re: Gamma Correct RGB Issues

I understand what your saying, but I was under the assumption that certain color values (ie reflection, transparency) used black and white color values as a means to express a value from 0 to 1. Therefore, inputing a value of 127,127,127 would actually be inputting a value of .5. In that case the color should not be the determining factor, but that value that it is intended to represent. This would explain why a parameter such as refraction glossiness is typically adjusted via a numerical value, but also has the capability to be mapped via a grayscale image.

The reflections appear lighter than the sky because it is reflecting the area where the sun is and the camera is looking at a darker part of the sky.

[Gijs]

Re: Gamma Correct RGB Issues

A value of 127,127,127 should indeed be a value of 0.5, this is exactly my point. If it is not corrected, the reflection will be too bright. For example, if you would make a rendering in linear space, and use a reflection value of 0.5, then a color like for example 255 0 0 would be reflected as 127,0,0. But when you convert this to gamma corrected space, the reflection would suddenly become too bright.

[dalomar_cg]

Re: Gamma Correct RGB Issues

I'm blown with this right now...I've been running tests, and I think that the Correct RGB may be affecting colors differently than the Correct LDR textures is, but now I'm not sure about it. I would map a white texture to the reflection layer and see for sure, but there's the bug where it renders the texture and not the reflections...Its something thats getting to me at the moment. Either way, 255,255,255 should be reflecting like a mirror, and both cases (the first too bright and the second too dark) look to be wrong... Maybe I'm just going mad.

[Gijs]

Re: Gamma Correct RGB Issues

Just as a side note, a value of 255,255,255 is the same in both linear and gamma correct space. On the other hand, if correct RGB is checked, it will of course affect all materials with an RGB color, so the reflections of those objects will still appear darker.

[dalomar_cg]

Re: Gamma Correct RGB Issues

Just as a side note, a value of 255,255,255 is the same in both linear and gamma correct space.

I know, but Correct RGB seams to be having an affect on white even though it shouldn't. Thats what looking like the "bug" per say. It may be possible that Correct RGB is shifting by a different curve, or only by a constant or something.

[Micha_cg]

Re: Gamma Correct RGB Issues

Yes, I wrote it at the beta 1 bug list. The gamma correction influence at white colors makes it difficult to work with colors and reflections. So, I stick at the old gamma workflow yet.

[dalomar_cg]

Re: Gamma Correct RGB Issues

Good...At least I wasn't seeing things due to staring at my computer screen for 18hrs straight

[Micha_cg]

Re: Gamma Correct RGB Issues

I know this effect.