Basic Comping of a Vray render

[rpc212]

Nice question trixian I would like a answer to that too!

So cpnichols we should use linear color mapping?

What about working in Linear color space? Use throbs method to achieve this?

What if I am using the workflow gijs has suggested with gamma correction color mapping?

[cpnichols]

Nice question trixian I would like a answer to that too!

So cpnichols we should use linear color mapping?

What about working in Linear color space? Use throbs method to achieve this?

What if I am using the workflow gijs has suggested with gamma correction color mapping?

'
Yes, I did work in the exact linear space as Throb outlines.

The key is monitor calibration. If you get your monitor calibrated to an sRGB curve. Consider it a "standard."

The other thing is that "smart" compositing packages, including Nuke, Combustion (I think), and even Photoshop CS2, should understand that the image is linear and view it (not convert it) correctly. In Photoshop you select your monitor profile. In Nuke you tell it sRGB.

It is harder for me to comment on Combustion since I am not as familiar with it. Trying to ask my wife who is an Inferno Artist, but she just confuses me. That is why she gets paid the big bucks.

And yes... don't use the color mapping (keep in linear). Throb says this like it is going out of style, and from what I remember in Gijs's tutorial I don't think he says to convert it to gamma space.

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.

[rpc212]

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?

[cpnichols]

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.

Think of it conceptually like this:
1 - Trobs method would be like adding Adjustement layer in photoshop. Shows you the result with the correct gamma but does not change the original.
2 - Doing the color mapping, would be like doing a color adjutement in photoshop burning in the change into the image.

In fact what I do when I paint texture in photoshop is to have an adjustment layer at the top which adds a gamma of 2.2. I paint like that, then when I save the texture, I make sure that the adjustement layer is disable.

This is the big reason why throb is super against using color mapping at all.

The only reason you would want to use it, is if you want to have your image saved out of Vray in 8bit color with the LUT burned in. Only works of you are a person that does nto work in float and don't care about g-buffers.

[dlparisi]

Tried it with linear mapping 1/1-still no luck so I'm unsure how to proceed.

Has anyone else gotten this workflow to produce identical results with the comp'd version and the RGB_Color version? Especially in combustion? I'm considering an investment in combustion but would like to get this resolved.

Anyone interested in trying with my .exr files please PM me and I'll email them.

David

[dlparisi]

Also, although it's not designed for it, is there a way to layer up the passes in Photoshop rather than a compositing program? For stills (95% of my work) it would seem like this is way easier. Two problems I see right off the bat are an inability to have more than one layer in a float images and also no "ADD" blending mode. I think "lighten" is similar but have never been able to confirm this.

I changed the 32 bit .exr images to 16 bit so I could layer them up in photoshop and got results that are similar (although not exact) to my test in combustion. Has anyone else given this a try?

David

[Dynedain]

You can do an ADD mode in photoshop by going to adjustments->calculations

[dlparisi]

You can do an ADD mode in photoshop by going to adjustments->calculations

Yeah, I've seen that before but it's not very useful in that it doesn't really allow for working in the normal layers box (i.e. stacking up layers). I never understood why "Add" is only available in the image calcs rather than as a blending mode.

[Dynedain]

You can do an ADD mode in photoshop by going to adjustments->calculations

Yeah, I've seen that before but it's not very useful in that it doesn't really allow for working in the normal layers box (i.e. stacking up layers). I never understood why "Add" is only available in the image calcs rather than as a blending mode.

Because photoshop sucks like that

[cpnichols]

You can do an ADD mode in photoshop by going to adjustments->calculations

Yeah, I've seen that before but it's not very useful in that it doesn't really allow for working in the normal layers box (i.e. stacking up layers). I never understood why "Add" is only available in the image calcs rather than as a blending mode.

There is an add in After Effects... keep that in mind.

[Da_elf]

There is ADD in photoshop since version 7 only that is called Linear Dodge so it won't get confused with the one in After Effects which deals with alphas.

can someone comment on this?

[trixian]

Well heres my comment.
Adding pixel values to other pixel values works universaly. I'm having trouble figuring out why they would name an add function "linear dodge" in photoshop, and calling an add funtion in AE "add", but just for alphas.
Adding is adding regardless of what channel one is working in. Same goes for the rest of the modes i guess.

Anyway. I'd like to second dlparisis issue with getting a result in combustion 4 by replicating said workflow. I know the workflsow is correct, but i cant get my head around whats going wrong. Guess ill try with 8 bit renders. If that doesnt work, there has to be some secret issue with how combustion reads files. Nobody on this forum have any experience doing this in combustion?

[rpc212]

I'm trying with Combustion 4 and it is not working for me either. Don't know what is wrong? Probably something simple I am missing as I am not an experienced Combustion user.

[Giovanni]

Cool tut, thank you.

I am getting everything pretty much as expected, except the background behind some glass.

What g-buffer channel has that information?
The reflection channel does not, nor does the diffuse, nor the background...

Also, everybody, remember that g-buffer works only with v-ray materials...


Thank you again,

regards

gio

[Gijs]

There is ADD in photoshop since version 7 only that is called Linear Dodge so it won't get confused with the one in After Effects which deals with alphas.

can someone comment on this?

linear dodge=add

only they wanted to keep it secret