Man I hate asking this but I find working in a vacuum or making too many assumptions = bad
no matter how much I think I understand something.
After using various approaches to LWF the time has swung about again for me to ask
a few questions to ensure sanity and at least my assumptions (yeah I know) are correct or not
So to start.
Normally most colors in the neighborhood are ok but I just upgraded to 2.30.1 from 2.1x (Max 2012 64bit) (which I know has probably nothing to do with this) and was rendering using VFB, added a curve and saved and was checking VFB vs Window image viewer vs photoshop
and the image viewer and photoshop look similar but neither had the crispness, contrast or color of the vfb, yup its full float but the
color/contrast being different really sticks out.
I typically use
max gamma 2.2
In 2.2 out 1
Affect selector and mat editor on
render 1/1/2.2 dont affect colors checked and use sRGB for preview and either save exr 1.0 or non-float adding 2.2
So the question is what is happening here is max or something else getting the gamma wrong? I had been using what I always considered a generally accepted method but I cannot ignore these results. Did some more testing with more simplistic scenes with the same results.
Also, in the Scenario A method does the LWF affect all map types? Vraycolor, colorcorrect, VRaysoftbox and so on or only bitmaps?
1/1/2.2 don’t affect colors checked and use sRGB for previewThat is applying a double gamma, isn’t it? I either have VRAY 2.2 or VFB sRGB, but never both. So, 1//1/1 and sRGB, or 1/1/2.2 and no sRGB. Or, I guess you can also have, 1/1/1 without VFB sRGB, but set MAX’s gamma to 2.2 out. I think the confusion comes because there is multiple ways to do the same thing. Unless, of coarse, I am dead wrong
Typically yes, I use sRGB for preview only. I guess it and the monitor profile are too far apart?
as I understand it when you check dont affect colors Vray calcs the light in the correct gamma space but does not affect the colors
so you have to add the 2.2 back to the image when it is used with sRGB preview to see it in VFB then of course whatever method you use
for post cc.
Aside from sRGB preview or not, saved or not even with an EXR no curve gamma 1 is appearing different between VFB and
photoshop unless I use the “Scenario A” from above, no max gamma, 1/1/2.2 lwf button ticked.
I will double check everything when I get to work to make sure I am not talking nonsense, make sure I am seeing what I think I am seeing.
If you set the max gamma out to 2.2, or if you use a gamma of 2.2 in vray’s colour mapping without turning on don’t affect colours, you burn in the gamma, which you don’t want for nuke or after effects. If you want linear for nuke comping, you use gamma out 1.0 in max, vray gamma 2.2 and don’t affect colours on.
yup just calibrated yesterday with the I1, it sounds like my process is ok but something I do not understand or most likely forgot about at some point.
going to do some tests now to verify what I think I see
I have seen some notes here and again about a MAX gamma bug, not familiar with it, could be older going to check that out too.
So have I been doing it wrong this entire time?
and anyway back to the original problem VFB and Photoshop matching (see attachment)
it also seems to happen no matter the method, no max gamma, max gamma etc
I seem to remember old dicussions about this but its been so long I really cannot remember
oh yeah here is the link to the video from the thread reply by, Svetlozar Draganov
Ok just made this test, hmm, not sure I understand what is going on
either something is wrong with making the image maps or how they are being read.
Max gamma is off in this test.
I should have put this under the “problems” section me thinks
you WILL get a difference between vfb and photoshop if you save your image out to a jpg, tiff, tga etc and then override to 2.2 (or have max gamma out to 2.2)
The problem here is you are burning in the gamma into the image for photoshop. The gamma you see in the vfb with srgb button ticked is purely a display thing.
You should be following as joconnel suggested and also then you HAVE to save out to linear exr or linear vrimg file. If you open the liner exr (which basically have gamma 1) then photoshop will automatically apply 2.2 to the image when you open it and your vfb and photoshop image WILL be 99.99% the same. (this will work if your lwf setup is like so: in max set gamma 2.2. Tick affect color and affect material. Input gamma 2.2, output gamma 1. In vray set linear colour mapping gamma 2.2 and tick dont affect colors. UNTICK linear workflow. In vfb tick srgb button. in vfb set to render raw image file either exr or vrimg file.
At the end of the day it doesnt matter how you get to the end result, as long as your images look good. Most clients dont know what is lwf, GI etc anyways
Hmm - seems that there’s a subtle difference between using the linear workflow checkbox to apply your 2.2 and using the max gamma / lut setup though - Svetlozar hinted that it was more accurate using the linear workflow button in his post too. Admittedly the linear workflow button won’t work on a lot of non vray materials so it’s only applicable for “pure” vray scenes, which is likely most of them these days.
Dman isn’t burning in his gamma, he’s just finding a slight difference between photoshop and the VFB if you use max gamma compared to the linear workflow checkbox method.
Correct and its not slight enough to go unnoticed by us or clients. Not talking about what it looks like in VFB but the VFB
and resulting image file, any file with or without gamma, float or not. Colors shifting from light, reflections whatever is not the issue.
But also I get it clearly that the sRGB is just a preview and does not match the Photoshop sRGB model, is it possible
to load the profile pshop uses into Vray VFB, adobe rgb for example? If its close that does not work then so that precludes using “dont affect colors”
for taking care of noise in dark areas. Than that makes more sense about what svetlovar is pointing out.
Also yes relativity is just that, I really do not care how anything gets there but what I want to do is to use in the neighborhood
of real world values, at least in lights to get a starting point to work from. If anything I am seeing is completely unreliable it is a problem.
We have many clients (like everyone else) that can be super anal about things. LWF is just another tool in the box of course
but when trying to mimic photography (yes I know unbiased etc) I need to at least start with some sort of known
to at least start from but I want to do it with VRay. Yes adding lights, color etc change materials, none of that is the issue here.
Then again maybe that helps answer my question as well.
I will try the method pointed out by Svetlovar and make sure its 100% vray scene and see what happens.
I cant render and get something a little different in post, it needs to be the same both ends.
It you tweak a mat in VRay to look perceptively correct then getting something slightly different in post
wont work for us, that also makes percetually scene set up sort of pointless.
ah thanks Vlado, ok so we can load it, is it possible to load the ones from photoshop, just updated to 2.3 and just noticed in last night.
need to look around for them.
Anyone know where I can d/l icc files for adobe rgb etc?
whoop, found this
and
ty much, I always feel like a complete imbecil at the end of these
and thank you everyone for jumping in, every little bit helps and contributes to my understanding, even if its only what I think is accurate
then gets corrected
cheers
I had a workflow that worked. I didn’t really understand how PSD-Manager handled the image so I inquired over at Cebas. I got a really good explanation and it was not at all what I was doing. Basically, if you are saving at 32 bit, this is how it should be set up:
Recommended typical setup for V-Ray:
V-ray Color mapping rollout:
Gamma: 2.2
Don’t Affect Colors: on
Linear workflow: off (this is an old option in V-ray, you have a true linear workflow without it)
3ds max Preferences > Gamma LUT:
Enable Gamma correction: on
Input Gamma & Display Gamma & Output Gamma: all set to 2.2
I put the whole conversation on my website, so I have a reference. Is it safe to say that the only role the VFB’s sRBG does is overlays a gamma curve on whatever you show below it. Let’s say you have V-RAY’s gamma at 1, because you are doing gamma in post, the sRGB would show you how the post would look. In essence the VFB is an image editor and you are applying the curve there. It would burn it in if you saved from the VFB, but not if you are saving in another way (“Render to V-Ray raw”, or PSD-Manager)
Dan, over at Cebas, implied that V-Ray will render faster if you have Gamma set to 2.2 and “Don’t affect colors” enabled.
You tell V-Ray with these options (V-ray Gamma 2.2 and Don’t affect colors ON) to take more time rendering very dark areas, because you will later look at it with a 2.2 gamma correction applied. So it will take longer to render. But if you were to leave the V-Ray gamma at 1.0 and use other options to raise the quality just to fix noise in these dark areas then this would probably take even longer to render.
My image is still opening up, in PS, with a double gamma. I’m saving an 32bit PSD, through PSD-Manager. I looked through PS settings and I am not seeing anything off. It’s not a great big deal, because I manually do an inverse gamma in PS (4545).
