Gamma 2.2, LWF, exr - Elements not equalling RGB in Photoshop

[notamondayfan]

Hello guys!

I've never had a problem with LWF before until this morning. Basically when I composite the elements in Photoshop, the end result appears bleached and washed out. This is the first time I have done this with 3Ds Max 2012, and I noticed that the exr part of Max has been upgraded, but I doubt that is the problem.
My Max / Vray is set-up as follows -

3Ds Max, preferences, Gamma and LUT enabled. Tick Affect Colour Selectors. Tick Affect Material Editor. Input Gamma 2.2 Output Gamma 1.0.
Vray, Colour Mapping, Gamma 2.2. Tick Affect Background. Tick Don't Affect Colours.

Elements set to 16bit half float.

When the images are loaded into Photoshop as 32 bit, the elements composite just fine, but as soon as I change to 16 or 8 bit, the issues occur. I have never had this before. I accept the colours aren't going to match 100% when using 16 or 8 bit, but these are quite different.
If I composite the elements in after effects, everything is OK.

Does anyone have any suggestions as to what might be happening?

Thanks,

Dean

[Wobi]

but as soon as I change to 16 or 8 bit, the issues occur

When you change your document to 16 or 6 bit it is no longer a linear space document. This is normal. You can only composite them in photoshop in 32-bit. No way around.

[instinct]

Keep in mind that 16bit integer is not the same as 16bit half-float. It can still be linear, but int does not allow for negative values or values over 1.

Regards,
Thorsten

[notamondayfan]

Yes I understand that, and can accept that when dropping to 16bit I'm going to lose any info over 1, and below 0.

In the past (3ds max 2010) I have successfully composited images to look very similar to the RGB in 16bit photoshop, which I was happy with, but now the colours and tones are very different, and my work-flow hasn't changed, only the software and hardware.

[Wobi]

Keep in mind that 16bit integer is not the same as 16bit half-float. It can still be linear, but int does not allow for negative values or values over 1.

The only way to do this in photoshop is to have a profile applied in you r color settings that has a gamma of 1, even in a 8-bit image. then the colors will blend mostly as before in 32-bit. But Photoshop does not know 16-floats. Float images are always 32-bit in photoshop, even if you save them as 16-floats out of max. But, the main problem is that when you change your document to 8-bit in photoshop the rgb-profile thats set in your color settings gets applied to the image, and that is probably something like sRGB oder ECIRGB and these profile have some sort of gamma-curve in them, so the colors will change.
The only way to composite render elements ion photoshop the way they should look is to use 32-bit images. If you got "about" the same look in your workflow before then it was just kind of luck, sry.

[notamondayfan]

OK it's interesting that you mention colour profiles, as I have just recently upgrade my graphics card, could this have an affect on the way photoshop is handling the transition between 32 and 16 bit?

Also are colour profiles embedded into the render elements by max / vray?

Thanks, Dean