VrayColor vs CoronaSolidTex

[Jacolath]

Hi Guys. I am new-ish to Corona and my brother and myself has been using is for a little while now. An interesting debate has arisen between us whether to use vraycolor or coronasolidtex with 2,2 gamma.

I told him that it was better to use the CoronaSolidtex but then he said he can not get proper black in his scenes. I wanted to prove to him with some tests but then I didn't understand the results.

I made 2 balls in a empty scene with pure white environment. One has CoronaSoldtex and one has VrayColor both with a RGB value of 10;10;10. AFter I render and I do colour pick from the VFB the Vraycolor ball reads 10;10;10 and the CoronaSolidTex reads 58;58;58 (please see attached image.

With vray we always used Vraycolor set to 2.2 Gamma to align the solid colour with the 2.2 Imput gamma of the bitmaps. Whether this workflow is correct or not it begs the question as to why the Vraycolor in this instance is gicing expected render results (10;10;10) with a neutral corona setup (Exposure: 0 Highlight Compression: 1.

I have read through the forums and could not find a lot of info regarding this. Am I doing something wrong? Is the use of Vraycolor the acceptable way? Is Corona built to not make dark blacks?

Your thoughts.
Jack

[Ludvik Koutny]

Corona does not have option to set gamma interpretation for SolidTex, this will be added in future. For the time being, you can simply feed CoronaSolidtex through colorcorrection node and and in advanced settings of CC node set gamma to 0.454. That will give you same result.

[Jacolath]

Thanks for the quick reply Rawalanche!

Is this the recommended method of working though? It doenst seem like a big problem to other people if they haven't implemented it yet so it gets me thinking I am wrong in using this method.
Maybe they (other Corona users) dont work with 2.2 Gamma or they use the Colorcorrection/Vray color workaround? Or they dont care to gamma correct solid colours? I just assumed we work in gamma 2.2 like vray... is this recommended?

Thanks

[Ludvik Koutny]

Thanks for the quick reply Rawalanche!

Is this the recommended method of working though? It doenst seem like a big problem to other people if they haven't implemented it yet so it gets me thinking I am wrong in using this method.
Maybe they (other Corona users) dont work with 2.2 Gamma or they use the Colorcorrection/Vray color workaround? Or they don't care to gamma correct solid colours? I just assumed we work in gamma 2.2 like vray... is this recommended?

Thanks

Everyone who uses Corona is (or at least should be) using Linear Workflow. Corona does not support non-linear workflow. But those colors that you see, and input in Linear worklow colorspace are the correct ones. The manual gamma correction of solid colors is adequate only when you have exact RGB values from a Client for example that you need  to match.

[Jacolath]

Ok! So it is recommended that I use the standard CoronaSolidTex and rely that the way i see it in my material editor is the way it will render. However...

My brother was not using VrayColor for RGB accurate colours, instead he was complaining that he could not get a dark enough defuse for some of his metals with CoronaSolidTex in a 2.2 envrironment. Within the scene discussed above CoronaSolidtex goes from a render sampled value of 0;0;0 to a sampled value of 21;21;21 when you increase the value from 0;0;0 - 1;1;1 in the colour picker inside the material editor. So increasing the "brightness" from 0-1 had the resulting increase from 0-21 in the render which means he lost those levels of dark and the darkest black above 0 was 21. I guess in that case where you want variations of blacks darker than 21 then it is best to use a gamma mapped CoronaSolidTex or in his case Vraycolor set to gamma 2.2.

Thanks again for your insight.

[Ludvik Koutny]

If you look at your picture, then you can see that what you get in material editor is nearly the same, so WYSIWYG workflow is indeed there. If you see brighter gray in Mat. Editor, then you get the same in render, and if you see black there, then you are getting the same as well. So you should rely more on your eyes rather than numbers.

I think this is matter of personal workflow rather than correctness. Wrong looking metal materials can be caused by many more factors than diffuse color, which i would expect to have quite small impact on something like metal material.

It would help a lot more if you posted picture of the specific material you are having trouble with as well as some reference (ideally a photo) of what you want to achieve, rather than breaking it down into shading elements ;) Then i will be able to help you a lot more :)

[Ondra]

Everyone who uses Corona is (or at least should be) using Linear Workflow. Corona does not support non-linear workflow.

There is no such thing as nonlinear workflow, there is only linear workflow and then bad workflow.

[Juraj]

The widespread use of VrayColor in mapped slots I attribute to confusion (and bro-science) not necessity in more cases than some would think. It's comfortable on one hand, because you can input your colors directly from Photoshop or RAL chart or any outside resource but
it doesn't change anything to your end result. Lot of people still think it does, and that makes me remember those days were people thought Linear workflow consisted of 10 steps including ridicoulous things like "de-gamma"-ing, setting gamma manually
through color-correction node,etc....Well, there are still some people on Chaos forum to tell you linear is just wrong. I don't know what to say in such times.

Anyway, metals should be just black. Like all-the-time. Not even 1.1.1, Black. 0. Pure metal is always specular only. If it's not pure-black, and indeed contains diffuse reflection (because it's dirty, or paint coated,etc..) a "metalness" mask should rather be used in blend material to lend between metal and diffuse parts of material, each with respective reflective curve. But not single material. Well it might not look bad or wrong with very low values but, it's not correct if that's what people are after.
Non-black metals is more confusing peevie than same map in specular and glossy slot (like it ever made sense to begin with.... ).

And last thing, could you Keymaster make test sometimes if "color-correction" node increases overhead in rendering ? On chaos forum I've noticed Vlado mention few times it can increase it for like 30perc, which would be a lot. I don't know, I've never done any test but I dislike CC anyway.

[Ondra]

I will just add gamma override to solid tex

[Jacolath]

Thanks for all the info guys and for the tips on Metal. I was using CoronaSolidTex all the time and am happy with the WYSIWYG approach. My brother was the one debating the merits of the VrayColor and he has some point at least for getting very dark levels in diffuse.

A gamma override will def fix this but for those who are interested I can show what I mean by the difference.

There is no particular material I am having a problem with (I will always find a away) it is just I think that the CoronaSolidTex was not gamma corrected in a 2.2 workflow which does not allow for some lower black levels in diffuse - hence the workaround of CC with .454 Multiplier. I just know now that if I want a raw diffuse colour lower than 21;21;21 I will need gamma corrected colour. (See attached) - this is  a blank scene with standard white background and no adjustments to exposure.

Increasing colour levels vs rendered result is as follows:

Diffuse Colour: 0;0;0      Rendered Result: 0;0;0
Diffuse Colour: 1;1;1      Rendered Result: 21;21;21
Diffuse Colour: 2;2;2      Rendered Result: 28;28;28

In the exact same scene the VrayColor (and Colour Correction CoronaSolidTex) with a 2.2 gamma value has following results:

Diffuse Colour: 0;0;0      Rendered Result: 0;0;0
Diffuse Colour: 1;1;1      Rendered Result: 1;1;1
Diffuse Colour: 2;2;2      Rendered Result: 2;2;2
etc

Thanks for the override button Keymaster!!!!!

I am pretty new though so dont have daily build but can use workaround till Alpha 7 or start getting some renders up ;)

[Ludvik Koutny]

Ok...  i tried the scenario with ColorCorrection node and it does what it should...  if i input 1, i get 1 in VFB. It is very easy to set up, and should not bother you much because as i repeat... you should do this only rarely, almost never. You should work in the perceptually correct color space (Linear Workflow) and set up your materials by how they look to your eye, not by mathematical values. :)

[Jacolath]

Thanks Rawalanche understood. IMO if Keymaster puts in gamma override button most people will use override to match what you did with the CC node. For me it just makes sense to gamma correct your colour map along with the bitmaps which are being corrected by the Max bitmap input override. Even when I was using Vray this was the only reason I found to use Vraycolor ... if not I could have just used the colour picker in the diffuse slot.

Thanks for clearing things up. At least now I know I am not working in the wrong direction :D