iq-scm / tony-mc-mapface

A cool-headed display transform

🌈 Tony McMapface 🌈

Takes HDR Rec.709/sRGB stimulus, and maps it to LDR. It's tranquil and collected, and won't set your eyes ablaze.

About

Tony is a display transform intended for real-time applications such as games. It is intentionally boring, does not increase contrast or saturation, and stays close to the input stimulus where compression isn't necessary.

It ships as a LUT along with a bit of shader code for sampling. There's no tweakables; if you need a different look, do it before and/or after the display transform.

Please note that the shader output is linear, so if "gamma space" sRGB is desired, the sRGB OETF must be used afterwards.

There's also a basic OCIO config with sRGB output.

Disclaimer: Tony McMapface is phenomenological, and tuned to its author's visual system (with mixed success). If you need something mathematically pure, try AgX.

Ports 💖

• DaVinci Resolve port by Sergey Makeev

Derivation

The generator's source code will come at a later time. Meanwhile, here's a TL;DR:

Tony McMapface takes a maximalist approach to the image formation process, and explicitly models perceptual phenomena:

• Brightness-equivalent luminance of the input stimulus is compressed (accounting for the Helmholtz–Kohlrausch effect).
  • The non-linearity resembles Reinhard.
• Color hues are preserved during compression, except for a deliberate Bezold–Brücke shift.
• To avoid posterization, selective desaturation is employed, with care to avoid the Abney effect.

As can be expected, the models underlying this process are imperfect, thus a liberal amount of eyeballing and fine tuning was applied.

Examples

HDR images from https://github.com/sobotka/Testing_Imagery

Acknowledgments

Thanks to Chris Brejon, David Briggs, Jed Smith, Troy Sobotka, and Alex Tardif for their excellent write-ups about color.

Special thanks to Troy Sobotka for eye-opening discussions, and triggering my obsessions 🙈

License

This contribution is dual licensed under EITHER OF

• Apache License, Version 2.0, (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.