Disparity between: comp of VrayRenderElements vs RGB_color (in PS)

[TomasEsperanza]

Hi, so my query/problem is a disparity between, the results of a PS comp with VrayRenderElements, versus the RGB_color.

 

With the RGB_color element on top of the stack, turning it on/off reveals differences in the highlights, that have no obvious cause. Does anyone know why?

 

 

 

Elements in the comp are:

 

+ VRaySelfIllumination

+ VRaySpecular

+ VRayReflectionFilter x VRayRawReflection

+ VRayRefractionFilter x VRayRawRefraction

+ VRayRawGlobalIllumination x VRayDiffuseFilter (copy)

+ VRayRawLighting x VRayDiffuseFilter

 

I'm pretty sure my comp is correct, but perhaps this is a quirk of VRay or PS that requires some kind of workaround? (Maybe I've overlooked something :/ )

 

 

(The 3ds Max process shouldn't make any difference, because, so far as I know, the RGB_Color element is the "Beauty pass" and the correct combination of elements can be used to reconstruct this exactly. However, just to elucidate, the procedure was:

 

Linear workflow.

 

Lighting = 1 x VRayLight Dome, with small blurry HDRI (set to invisible)

Environment Map.jpg (in Environment and Effects)

V-Ray GI Environment (skylight) override is black

 

Window Glass is Excluded from DomeLight.

 

Exposure Control (in Environment and Effects) is left as .

No Effects, no Atmosphere.

 

No D-O-F on VRayPhysicalCamera. f8, 1/6, 1600

 

Elements were saved as 16bit EXRs, from the VRay frame buffer, using the "save all image channels" button.

 

Load Files to Stack: Folder added in PS.

 

ProEXR plugin for PS installed; gamma is fine.)

 

 

But don't let that sidetrack you; back to the question:

 

With the RGB_color element on top of the comp, turning it on/off reveals differences in the highlights, that have no obvious cause. Does anyone know why?

Edited April 9, 2016 by TomasEsperanza

[TomasEsperanza]

I'm currently of the thinking that it maybe to do with the way in which the 16bit EXRs are being handled; PS opens them 32bit mode when imported via the script. There are different avenues to explore in response to this, and at present I find the results confusing.

[Francisco Penaloza]

I usually don't go that deep with passes, usually if I use them I only go with regular VRay reflection, refraction, etc... not the RAW passes.

Having said this, when comping in Fusion, you get a closer match of the beauty pass, but still there is discrepancies on the refraction pass, and lower antialising on the specular. Not sure why VRay does this, but it supposed to be a bug that it was fixed on V3. but I still can't get a straight comp out of passes, I usually have to adjust intensities or use curves with them.

[TomasEsperanza]

but it supposed to be a bug that it was fixed on V3.

 

Okay, so it's not just me Thanks for your reply, Francisco.

[Corey Beaulieu]

A couple of things,

 

You don't need to filter your own Reflection and Refraction or lighting for that matter. Unless you plan on effecting that filter in post. Just use the basic Reflection/Refraction/Lighting pass (not the raw).

 

Second, if you are using any Reinhardt value at all, this gets baked into the RGB, as does any other VFB adjustment that will not be baked into the passes.

 

Lastly, and this is the most likely, if you use "Affect All Channels" anywhere, for glass in particular, this will cause you to not need the refraction pass. In fact, the refraction pass then double the brightness because what is behind the glass is also included in all other elements as well. I would recommend Color+Alpha so that you don't have to consider what is and what is not Affecting All Channels.

[TomasEsperanza]

A couple of things,

 

You don't need to filter your own Reflection and Refraction or lighting for that matter. Unless you plan on effecting that filter in post. Just use the basic Reflection/Refraction/Lighting pass (not the raw).

 

Second, if you are using any Reinhardt value at all, this gets baked into the RGB, as does any other VFB adjustment that will not be baked into the passes.

 

Lastly, and this is the most likely, if you use "Affect All Channels" anywhere, for glass in particular, this will cause you to not need the refraction pass. In fact, the refraction pass then double the brightness because what is behind the glass is also included in all other elements as well. I would recommend Color+Alpha so that you don't have to consider what is and what is not Affecting All Channels.

 

Nice one Corey, sounds like that's probably what it is. I take a closer look at those aspects. Cheers