Basic Comping of a Vray render

[Gijs]

So using gamma correct color mapping with Dark 1.0 and Light .45454 will screw up the g-buffers output (diffuse, specular and reflection) from Vray?

Isn't using this workflow (gamma correct color mapping with Dark 1.0 and Light .45454) the same as Throb's linear workflow tutorial?

aHA... that is a BIG and important no (and a good questions).

Using the linear workflow that Throb talks about keeps all the data linear and only alters the way the VFB displays the data to be in the correct space.

Appying the gamma correct color mapping, alters the data.

As far as I understood, this is only partially true. Gamma correction affects RGB, not realRGB data. When using gamma correction color mapping, when saving the realRGB to openExr, it is still linear. I don't know if the other channels do get affected by color mapping though(but it seems unlikely)

[Gijs]

If you DO use gamma space it will fail because, your rawGI, lawLight is in linear, and your diffuse, specular and reflection will be in gamma... bad bad bad. Keep it all linear.

ah, ok, I see it goes wrong with the other channels. Pretty strange that it is like that. As pointed out in another thread, gamma correct color mapping is quite helpful to get noise out the darker regions of your (linear) data.

[cpnichols]

If you DO use gamma space it will fail because, your rawGI, lawLight is in linear, and your diffuse, specular and reflection will be in gamma... bad bad bad. Keep it all linear.

ah, ok, I see it goes wrong with the other channels. Pretty strange that it is like that. As pointed out in another thread, gamma correct color mapping is quite helpful to get noise out the darker regions of your (linear) data.

You know, the more you point that out, the more it makes me think that this may be a "bug," or a needed feature. Rather than effect the render, it would be best to have a "gamma control" on the Irradiance map, or better yet, the QMC control, that could solve the issue without the risk of compromising the final rendered data. Just a thought.

[Gijs]

...or gamma correct color mapping affecting the channels is a bug. But your idea to have a 'gamma correction' for the qmc sampler isn't a bad idea.

Actually it is quite logical to take gamma correction into account, because in the very end your data is gamma corrected for viewing. Maybe we should put it on the wishlist?

[vlado]

In newer builds, color mapping can be enabled or disabled for the individual render elements. Further on, V-Ray has always taken color mapping in account when doing QMC calculations.

Best regards,
Vlado

[panthon]

I´ve got good results comping all the layers in combustion, working in linear mapping mode. WIth intensity exponential it gets all screwed.
There´s no secret in combustion to get it the right way, the trick is using nested layers. I could write a little tutorial if some people were interested, but I guess some more qualified people in this forum would do it better. I must warn: I don´t get exactly the same results, but it´s pretty close.

[Da_elf]

tutorials are fun to write. go ahead and write on

[Gijs]

In newer builds, color mapping can be enabled or disabled for the individual render elements. Further on, V-Ray has always taken color mapping in account when doing QMC calculations.

Best regards,
Vlado

That's cool. I know it affects QMC calculations, and therefore, it is great that you can disable it for render-elements, while keeping it for rgb.

Hey Chris, the wish is fulfilled before it has even been on the wishlist

[cpnichols]

Vlado... what about the gamma control in the QMC idea?

[Gijs]

Vlado... what about the gamma control in the QMC idea?

Isn't this exactly what is possible in newer builds? just use gamma correct color mapping and disable it for everything but the rgb channel?

[cpnichols]

Vlado... what about the gamma control in the QMC idea?

Isn't this exactly what is possible in newer builds? just use gamma correct color mapping and disable it for everything but the rgb channel?

Yes... but logically, I think as far as workflow and usability (this is speaking from the teacher in me), it would be logical to have that setting in the QMC. But hey, that is just me.

In either case, yes, now it is possible.... so cool, when the new build is out officially, I will do anothe tutorial. Or maybe save it for my next DVD...

[rpc212]

Panthon - That would be awesome if you could do a tutorial for Combustion! I am getting totally blown out results after following all the steps in cpnichols tutorial. I will post some images of what I am getting tomorrow.

[Dynedain]

I'm trying this in combustion as well, end everything looks great, but when I add together the (rawlight * diffuse) and the (rawgi * diffuse) I get a really bright blown out image. The reflection/refraction/specular passes seem to match and the end result seems right except that its blown out. I can't seem to figure out how to control the exposure to tune things down.

[rpc212]

Yep that's my problem too!

[cpnichols]

I'm trying this in combustion as well, end everything looks great, but when I add together the (rawlight * diffuse) and the (rawgi * diffuse) I get a really bright blown out image. The reflection/refraction/specular passes seem to match and the end result seems right except that its blown out. I can't seem to figure out how to control the exposure to tune things down.

Hmmm that seems odd. The math is correct

(RawGI * Diffuse) + (RawLight * Diffuse) + spec + reflection + refraction + ....

Try this

See if (RawGI * Diffuse) is the same as GI and that (RawLight * Diffuse) is the same as Lighting. Because it should be.

I am suspecting that is a problme with the way that combustion is reading the data and/or displaying it.