Page 1 of 1

Keeping LUT at 16/32bit

Posted: Thu Dec 07, 2017 7:29 pm
by thanulee
Hi, it would be super useful, if we could have the image not clamped to 8bit when using LUTs.
Its common for the art director to do the final image inside octane 90%. But the high dynamic range should be preserved in export, is really important for fine tuning afterwards.
I found online most of the LUT collection that octane uses and i can do this in post but with one problem. Cant mimic the neutral response of any LUT. No blend operation in ae or nuke does this for me. Therefore i d like, either to know how i do this in post, or to have the image export at 16/32bit exr with lut on. I dont care about white clamp, but i care about keeping the grading on my shot. Plz consider this.

Cheers and ty

Re: Keeping LUT at 16/32bit

Posted: Thu Dec 07, 2017 10:39 pm
by aoktar
thanulee wrote:Hi, it would be super useful, if we could have the image not clamped to 8bit when using LUTs.
Its common for the art director to do the final image inside octane 90%. But the high dynamic range should be preserved in export, is really important for fine tuning afterwards.
I found online most of the LUT collection that octane uses and i can do this in post but with one problem. Cant mimic the neutral response of any LUT. No blend operation in ae or nuke does this for me. Therefore i d like, either to know how i do this in post, or to have the image export at 16/32bit exr with lut on. I dont care about white clamp, but i care about keeping the grading on my shot. Plz consider this.

Cheers and ty
It's not clamped to 8bit. It's clamped to between 0-1 for floatbuffers.

Re: Keeping LUT at 16/32bit

Posted: Thu Dec 07, 2017 10:47 pm
by thanulee
Ru sure? Cause when i export an EXR as tonemapped its practically 8bit even if it says 16 or 32. Try it, go in ae and use exposure and u ll see u got no dynamic range whats so ever. I ve tested this many times.
If i export it as linear/linear, EXR is fine all information there but i lose the LUT (and the highlight clamping from camera imager which is fine and expected).

Care to do a test see if that works for u? I ll post some images 2morrow to show u the issue. thanks a lot Ahmet

Re: Keeping LUT at 16/32bit

Posted: Fri Dec 08, 2017 1:20 am
by Terryvfx
You are having problems using LUTs in post because the LUTs you are using are most likely not calibrated to use on Linear images. Most of the LUTs or Film simulation solutions you will find online will only work well on non-linerize imagery.

You could set the gamma to 2.2 then LUT then gamma to .4545 to set it back to linear or use LUTs calibrated for linear space, you can get those here. That pack it's free and they also have a paid one with a different set.

Let me know if you find this useful.

Re: Keeping LUT at 16/32bit

Posted: Fri Dec 08, 2017 9:52 am
by thanulee
No i have 0 problems using luts in post :) I have problems mimic the neutral response of luts, nothing else.
But hey Ahmet is right, im confused cause at 3.06 the images were clamped. Im doing a render atm and is not!
To clarify : i render exr 16 bit at srgb tonemapped (so i can get the LUT information). The result in 3.06 was 8 bit, and in 3.07 is 16.

I need do some tests with building rgb from passes as well and let u know!! cheers

Re: Keeping LUT at 16/32bit

Posted: Fri Dec 08, 2017 11:07 pm
by thanulee
Sorry to be back at this again but it is as i initially wrote.
Do the following: do a render with some highlights, and export as EXR 16 srgb/tonemapped. Then as linear/linear. Put both in AE, make project 32bit and srgb/linearized on. Use exposure in an adjustment and go down like 5 stops. See dynamic range lost on the tonemapped. Even if the values are 0-1 the image is practically 8bit.