Well, This all started after SV’s post on the subject in Problems, and because recently I had some trouble with color spaces (I was the one who screwd it up, but was fun experimenting) I decided to do a little sunday mornin’ testin’ to see what’s the current situation considering color integrity.
Rendered a sample image with my usual settings, saved it as exr, screenprinted an image, ovelayed it in PS as difference, and got the expected full black frame. Thats where the curiosity kicked in. I’ve streched the histogram, to see, if there is any difference at all, an there was… pretty strange. Here are the images:
At first, I thought this is all related to my color settings in PS, but I’ve stretched the overlay image posted by SV in his op on the subject, and got the same results, or at least it was the same before the jpeg compression, but my point is visible:
This is not a major issue, but only an interesting thing to witness. So, what do you think? Where does this difference noise comes from? Is it PS? The processing of the exr into 8bit? Or something else? Obviously, this affects our everyday work ver little, so consider this a “just for fun” thread.
I’ve posted this in the original thread, but here too just to show that in my case it is visible without any comping in ps
on the right is how the EXR looks when saved… the effect is very different when displayed on different screens now that iv’e noticed this. on my old MAG19 it is less visible then on my new SyncMaster 226bw
Any chance screen profiles or color management issues are in play here? can’t se why though - does max take this things into account when saving files?
Sorry bakbek, I didn’ want to hijack your thread, and din’t quite know what’s going on with your settings, so this is why I opened another thread, and since it’s not really VRay related, it’s in the off topic section.
I’ve read trough your thread, but other than what the others said, I can’t add anything. Maybe that you shouldn’t use the max frame buffer at all, it has little to do with your output. Stick to VRay VFB and exrs. A "Clone VRay VFB feature has been promised, as far as I can remember.
hey, no problem… what i wanted to add here is that the file saved from the VFB is exactly as it shows when cloned to the Max Frame buffer - i’m not using it, just noticed that fact. i do use the VFB and save from it directly. maybe this can help in understanding why the saved file looks different then in the VFB
Interesting thing, is that I can’t bring the two images you provided to the same colors. Not with gamma, or simple transformations. But maybe this is all just because the screenshot is LDR.
Anywa, at first glance, I wanted to say, there is still a slight gamma mismatch, like the color mapping and MAX gamma setting would differ. But I really don’t know. You should be getting proper results if you’ve followed SV’s suggestions, so this is strange.
bakbek - what you’ve posted here is the Vray Frame Buffer which stores the images as linear HDR data, and your viewing it in sRGB colour space.
You’ve also duplicated this to the Max Frame buffer, which is totally different. Here, the Max frame buffer is displaying the colours (i think - someone please correct me if i’m wrong) with gamma 2.2 - this is different from what the Vray Frame Buffer is doing. Only subtle, but it is different. The Max frame buffer is also LDR i think.
If you are saving the .exr out from Max’s save - use the Vray save feature as i suggested to you - then it will save perfectly correctly.
yep, cs3 can. and the mismatch that aldaryn is describing here it’s much more evident in 32bit mode than 8bit (both in PS and AE), just did a couple of tests. might be the 8bit screen grab though, it may be that it just doesn’t match the full float output.
at first the max frame buffer clone looked just like the exr direct save, but further checking showed some difference. in the end the screen view with sRGB in the VFB was different from the saved EXR that i did directly from the VFB (or is that wrong)
Further more - VLADO’s response in that thread just made me even more confused now
people work with VRay for years now… how come there is no clear workflow on this subject?
There IS. Work linear, save linear, keep linear, display corrected. And here comes the hardest part: Force everyone you have to work with to do the same. This is where we usually fail, and just give in… hehe.
Sorry for confusing you further; basically there are two workflows: one where every image that you have is stored with gamma-correction burned in (which is what most people seem to imply when referring to the “linear workflow”), and one where no image is stored with burned gamma and everything is linear (which I find infinitely more simple to support).
In the first case, you have to take care that each program that you work with reverses the gamma correction of input images, and re-applies it on the output images, so that internally it works with the correct linear data. One of the problems here is that in addition to the gamma correction, there are many images using the sRGB color space instead, which is similar, but not quite the same. These must be converted from sRGB to gamma space by some image-conversion program.
Getting all this input/output conversion to be done properly may be tricky in some programs, which is where the confusion starts.
In the second case, since you are certain that you always work with linear images that do not require any sort of gamma correction or other color transformation, things are a lot less complicated.
Obviously, this affects our everyday work ver little, so consider this a “just for fun” thread.
Work linear, save linear, keep linear, display corrected. And here comes the hardest part: Force everyone you have to work with to do the same. This is where we usually fail, and just give in… hehe.