LWF and highlight compression

[Nejc Kilar]

Hey guys,

I've been hitting my head with a hammer trying to understand PROPER linear workflow (ending up with a 32 bit file ready for compositing) and the difference in light distribution that it potentially brings. I am not inexperienced per say but to say I understand the real difference in LFW would be far fetched apparently.

So, my question is, if I send out a png image that is already tone mapped in the VFB  (highlight and exposure corrections) what would be the difference in the way light propagates in comparison to the same render exported as an OpenEXR with all the VFB settings at default (exposure and highlights) edited in PS / Nuke to match the png image?

Is there even a difference? My tests are inconclusive, especially so because LUTs work differently in 32bit compared to 8 bit, of course :) LFW to me was always about having extra breathing room in post and thats pretty much it.

I stumbled on the fact I don't have a clue about this while studying the difference between tone mapping options of V-Ray, Corona, F-Storm and Octane. Go figure right?

Answers greatly appreciated! Thanks!

[maru]

Hi, I am not sure if I understand the main question correctly. It is:

"if I send out a png image that is already tone mapped in the VFB  (highlight and exposure corrections) what would be the difference in the way light propagates in comparison to the same render exported as an OpenEXR with all the VFB settings at default (exposure and highlights) edited in PS / Nuke to match the png image?"

Right?

What exactly do you mean by "the difference in the way light propagates"? Do you mean simply the visual difference between an image which is tone mapped, and one which is not? In such case, you could simply save one image with the default settings (highlight compression at 1, etc), and another one with some extreme tweaks, and then compare them. Each operation in the "post" tab has some impact on the image. There are ways to measure this, for example this:
https://forum.corona-renderer.com/index.php/topic,10306.msg65725.html#msg65725

[valwizard]

I don't know if it is the right place to drop my question as well as I got it, but here it is:
- What is the right way of having an correct output 1/1/1 for further compositing? With all my respects to Corona but I could find only infinite debates all around this subject but never right in focus. With all the topics I got only lost in the mud instead of clarifying how to actually make a correct output exr32 in Corona in order to compose all passes. What should be the values in Tone mapping in order to have a correct output ( not washed, not guessed). For instances: in Vray you have separated colour mapping from the gamma. And you have the predictable result and a good starting point. In Corona instead - Tone mapping actually affects the gamma, for instances different exposure values will affect gamma output. Please help to understand this workflow, if possible. For now I am considering that the correct values should be Exposure = 0 Highlight compression= 1 Contrast =1, anything else should be left at 0. Is that correct?
- Off course I am considering that the right way to tweak is only the camera settings - not the sky or sun intensity values, is that correct too?

[Nejc Kilar]

Oh my Maru, I owe you a sincere apology. So sorry I never got back to you!

I had a brain fart type of a moment when I opened this thread - I was actually trying to grasp the different color mapping options (reinhard, linear etc...).

valwizard, there is actually a pretty cool video that might help you:

t=113s

Essentially you'll want to just disable the tone mapping options (Highlight compression at 1 etc...) and save as a 32 bit file.

:)

[valwizard]

Thanks, I've seen this video long time ago, but the question still remains, I am afraid. What is gonna happen, for example, if you check Photographic exposure instead? Will this affect the LWF?