Feature request: CIELAB L* tone curve

[abnormally-distributed]

Any chance the Lab L* transfer function can be added to LUTcalc?

The Lab OETF is basically like the sRGB OETF / tone curve/ whatever but uses a gamma of 3 spliced onto a small linear ramp instead of a 2.4. It is identical to the lightness formula used in the CIELAB color space.

If  (x <= 0.08000183/9.0331789 )
   9.0331789 * x    
else 
  ((x^(1/3)) - 0.13792419)/0.86207581)

[radioproektor]

This was requested long time ago here https://github.com/cameramanben/LUTCalc/issues/8

[radioproektor]

Hi there! Any progress on this? I use L gamma to invert film negative scans. It gives me the most "symmetrical" invert without strange gamma shifts. My LUT looks like this: g1.0 -> L -> Invert -> g1.0. I created it using grayscale 1024 samples pattern (maximum possible) in LUTbuddy plugin and apply transformations using Elle Stone's Well-Behaved ICC Profiles ICC profiles v4 in PhotoLine image editor in 32 bit depth. This is rather dumb method to create a LUT and it provide only 10bit precision. So i really would like to see L option in LUTCalc to create higher resolution g1.0 to L and L* to g1.0 LUTs.

[cameramanben]

Hi radioproektor,

I've just uploaded code to add CIE L* to the gamma list (it's with the other power function and linear options).

Just for fun, I've also added CIELAB as a gamut option. Considering that it is rather less useful than the L gamma in a LUT, it is a bit complicated so that it fits with the whole 'combine CIELAB gamut with L gamma to make CIELAB colourspace' workflow of LUTCalc. Still, it looks like it is working (ie it looks very odd in an RGB based system!) ;-)

It's in the main branch now, and I have it up for testing at lutcalc.net/testing

Ben

[nick-shaw]

lutcalc.net/testing does not work for me.

I can see it at https://cameramanben.github.io/LUTCalc/testing/

[cameramanben]

https://cameramanben.github.io/LUTCalc/testing/ Quite right Nick, oops!

[nick-shaw]

Oddly if I use your link it fails, but if I type in the URL it works. I think it may be https vs http.