7
$\begingroup$

I have severe red/green protonopia and find the red and green to be the same when I use False Color mode to calibrate lighting exposure.

I'm wondering if there is a way to get numeric feedback on parts of a surface I point at or a way to change the colors used in False Color so that I may be able to see them better. Also if not I'm wondering where I could recommend these kind of features.

It would be nice to have a tooltip with a numeric representation of the exposure of where on a surface I am pointing, or shift the colors represented somehow.

$\endgroup$
  • 2
    $\begingroup$ The False Color view transform is defined in the filmic_false_color.spi3d and referenced by the OCIO configuration in config.ocio. It's entirely possible to replace this with a more suitable colormap, e.g. Viridis, Inferno or Turbo. $\endgroup$ – Robert Gützkow Feb 2 at 22:41
  • 2
    $\begingroup$ However I'll have to check how to auto-generate the LUT. The basic format is described here by Troy Sobotka (the developer of Filmic) $\endgroup$ – Robert Gützkow Feb 2 at 22:42
  • 2
    $\begingroup$ I've got a script ready, will likely write the answer tomorrow. $\endgroup$ – Robert Gützkow Feb 2 at 23:18
  • $\begingroup$ @Robert Gützkow Wow thank you! I look forward to it a lot. This is a level of technicality that is quite a bit above me right now looking at these links. Its always kind of strange to me how often red and green are used in things when they are in the highest frequency of color deficiency. This could probably be useful to many people who struggle more than they want to admit. $\endgroup$ – Sunlight Swift Feb 2 at 23:46
  • $\begingroup$ I hope my answer is helpful. Since I'm not affected by color vision deficiency, it's hard for me to judge whether the lookup tables provide a good solution for this particular use case. It might be necessary to create distinct ranges similar to the original false color view transform for a better solution. $\endgroup$ – Robert Gützkow Feb 3 at 15:56
9
$\begingroup$

The False Color view transform is defined in the filmic_false_color.spi3d and referenced by the OCIO configuration in config.ocio. In order to add custom 3D lookup tables as new view transforms, they have to be stored in the .spi3d format and the config.ocio has to be modified to references these new files.


Update 2020-02-17: The re-write of the tool is complete. It now includes a proper command line interface, it allows to create your own colormaps or to load colormaps from viscm. The release includes lookup tables that should be suitable for people with color vision deficiency. Turbo has been removed from the list of colormaps since it's not useful for people with achromatopsia and it would require to license all code under the Apache license. The matplotlib colormaps are still available.

Update 2020-02-05: The correct mapping of middle grey to the center of the colormap is applied and relative luminance is used for the mapping.

Update 2020-02-04: The lookup tables now have a distinct color for both the low and high clipping range.


I've implemented a lookup table generator that allows you to create your own false color view transforms. Detailed instructions for the installation, colormap creation, arguments for the command line interface and integration into Blender are provided in the README.md.

The first release also includes two false color view transforms, Ignis and Agnoscis, designed for people with color vision deficiency using viscm. Additionally they've been analyzed with Coblis. The generated .spi3d files and custom ocio.config can be downloaded here.


Examples

Test Scene Test Scene - View Transform: Filmic

Ignis False Color View Transform Test Scene - View Transform: False Color Ignis

Agnoscis False Color View Transform Test Scene - View Transform: False Color Agnoscis

Dante False Color View Transform Test Scene - View Transform: False Color Dante (based on Inferno)

Ignis Evaluation Ignis Evaluation

Agnoscis Evaluation Agnoscis Evaluation

Ignis Exposure Values to Colors Agnoscis Exposure Value to Colors

Exposure value to color, Ignis (l) and Agnoscis (r)


Installation

  1. Download the .spi3d files and custom ocio.config from the release page.
  2. Open Blender's directory and navigate to [2.8x]\datafiles\colormanagement (the 2.8x is a placeholder for your Blender version).
  3. Rename the original ocio.config in that directory, e.g. to ocio_old.config
  4. Copy the custom ocio.config into the [2.8x]\datafiles\colormanagement directory
  5. Create a directory/folder named custom in [2.8x]\datafiles\colormanagement
  6. Copy the .spi3d files into [2.8x]\datafiles\colormanagement\custom

Usage

The new view transforms should be available in Render Properties > Color Management > View Transforms after a successful installation.

enter image description here

| improve this answer | |
$\endgroup$
  • 2
    $\begingroup$ I wanna upvote this like a thousand times! You're awesome :). $\endgroup$ – Jachym Michal Feb 3 at 16:51
  • 1
    $\begingroup$ I look very much forward to test this out and provided feedback. I don't have time a whole lot for the next 48 hours but maybe after midnight I could test with it. For me the Magma and Turbo are very clear for finding more ideal exposure. I think I could switch to standard False color to find the extreme spots and then to Magma to see the mids. False Color all becomes just one color I'm almost entirely blind to it but I can see the very extreme white and red(I think). I will be tinkering with this a bit in the future and see if I can make a red protonopia specific solution. This helps a ton $\endgroup$ – Sunlight Swift Feb 3 at 18:09
  • 1
    $\begingroup$ @SunlightSwift Great! I would appreciate to get feedback. If this works really well, perhaps it could be included in Blender. Perhaps Troy has some suggestions as well. $\endgroup$ – Robert Gützkow Feb 3 at 18:21
  • 1
    $\begingroup$ Ok. This is excellent work as is typical of @RobertGützkow. Now take it further; write up a false colour with palettes specifically targeting tritanopia, deuteranopia, and protanopia. There are good formulas for deriving decent mixtures for each, and it should be feasible to implement a good series of increments to correspond with the EV increments under the traditional False Colour. Especially given that there's an audience to test the results. Excellent work. $\endgroup$ – troy_s Feb 4 at 22:50
  • 1
    $\begingroup$ I've pushed a repo that has a False Colour generator and the commands to GitHub. It is done using a Colour Ramp node, so should be more or less usable for a majority of people and only relies on the OCIO tool "ociolutimage". Good luck. github.com/sobotka/filmic-blender/tree/make-false-colour $\endgroup$ – troy_s Feb 4 at 23:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.