Photoshop (&GIMP) 32-bit depth painting, brush UI magenta explanation

Hello!

I am trying to understand how, exactly, the painting system in Photoshop works, in 32bit mode.I can sort of guess that with GIMP, the 0-100 slider means that at 100, the values in the pixel are set at the highest that a 32-bit pixel can be.

Specifically, the two most bothersome questions I have atm are:

  1. Why exactly does the colour choice palette turn magenta
  2. How is EV level for the brush calculated

For example, Photoshop enables a +20 value on a brush, but when an EXR has 256 EVs of range it just seems laughably small.

Naturally, since 32bit is mostly used in 3D I thought you guys might know of any literature, videos, articles, anything really, that could explain how (, precisely,) the EV UI in Photoshop should be read.

Thanks in advance!

PS is mainly focus on photographers and not 3D artist. They prefer use stop ranges as it is more friendly to camera users. As a 3D artist you just need to look to RGB values. If you’re texturing, it’s better to use Substance Painter, Mari or Blender, it’s more suitable for that range of colors. For matte paintings/projections you can keep at 16-bit since most of professionals cameras like RED, ARRI or Panavision won’t pass that. By testing, each stop increase values at this rate RGB * 2^stop

with gimp the 32 bit values are between 0 and 1

and NOT!!! the -32768 to +32768 ( the normal 32 bit depth )

Well I am kind of writing from a photographer’s point of view. :slight_smile:
So if I understand correctly in gimp a 32-bit value would produce a number like 0.51423, but not a negative one. Still, this must translate to floating point precision somehow to be saved and reinterpreted, or are values in an HDR format (like EXR for example) truly written in a 0-1 format? Maybe I’m overshooting with the question… anyway:

I found that GIMP’s documentation is very informative on the subject ( https://docs.gimp.org/2.10/en/glossary.html#channel-encoding ) but I still have to find something similar for Adobe.

Given that Photoshop’s colour picker only enables -20 to +20EV of range I wonder how that is translated into data that can be saved, and then interpreted from. Adobe set EV=0 to correspond with RGB=1, but then extension of RGB values to 20 would mean Photoshop clips everything above 20EV of range, which would still leave out about 4-8 stops that are accessible on Earth. It seems like an oversight, but surely they had a reason for it?

Furthermore, I just realized that the EV (intensity) scale in Photoshop is strange. In the below image you can see that at 6,38EV on the slider the RGB values are already at 20, but the slider can still be dragged further up to +20EV, yet the RGB values will stay at 20,0. RGB values will also clip to 20 with a warning whenever a higher number is input.

The plot thickens.

That stop range will be only used by a professional who needs pixel information to finish his work, most of time overexposed RAW images. Rarelly someone with take a 20+ stops overexposed photo, it’s better to adjust camera parameters. EXR can store high values but wasn’t totally meant to RGB numbers, is more for tech data like depth, UV, vector, deep.
Eventually you will have to tonemap or normalize you file to a delivery format for your client, yet those 20 stops is Photoshop range. Probably Adobe assume this is enough to photographers. The 0 to 1 format is to make math easier to see. If you have a RGB 0.5-0.5-0.5 background and another layer RGB 0-0.5-0 in multiply blend mode result will be RGB 0-0.25-0. Difference will be RGB 0.5-0-0.5. Add RGB 0.5-1-0.5.