I’m trying to emulate the Fujifilm Film simulation presets for use with Rawtherapee. To achieve this goal, I took shot a ColorChecker in raw+jpeg fine, going through each color preset from my Fujifilm X-T1 (Provia, Velvia, Astia, Classic Chrome, PRO Neg. Hi and Std, and the different monochrome versions). Then I can load the raw file with neutral preset, check “Tone curve” and “Look Table” From the DCP profile in color Management, and work with L* curve, LH, HH, SH and VH equalizers to tweak the levels, contrast and colors.
I tried first to emulate the Classic Chrome preset, and I think I got something pretty close to what the in-camera development produces. Classic_Chrome.pp3 (11.4 KB)
My question is, to share this work as presets with the community, should I just share pp3 files, or make HaldCLUTs?
I tried to make a 16-bit PNG HaldCLUT from within Rawtherapee, but the resulting PNG is huge (about 12MB). How can I reduce the size of this HaldCLUT? Or should I make a TIFF and use ImageMagick to convert to a lighter PNG?
Awesome! Not sure if you’re aware but @hanatos recently got himself a fancy new Fuji camera as well - and he couldn’t help but start playing with mapping the film emulations as well (in darktable of course). Perhaps he could chime in with some hints or tips?
Why not both? If you get me the resulting Hald CLUT files I can pass them onto @David_Tschumperle who can compress them quite nicely to include them as part of the packaged CLUTs in G’MIC (and eventually RT I would imagine - as soon as he let’s the rest of us know what he’s up to).
Either is fine. We don’t have a HaldCLUT folder for emulating camera effects, but neither do we have that for PP3 files, so we could make one. PP3s would be smaller. HaldCLUTs would be “cleaner” if you know what I mean.
Have you used the 12-level identity file, as described in RawPedia? http://rawpedia.rawtherapee.com/Film_Simulation
Save it as an 8-bit PNG with maximum compression. Don’t worry about it being 8-bit, RT’s HaldCLUT algorithm interpolates the missing values and I found no difference between using an 8-bit vs 16-bit file, other than 8-bit being smaller and faster.
If I remember correctly, the ImageMagick command I wrote in RawPedia and the identify file I bundled in the zip handle the gamma issue, so your HaldCLUT should just work if you use either the command or the sample identity file.
Indeed pp3 is lighter, but HaldCLUT has the davantage of leaving all curves and equalizers flat.
So I’ll make both pp3 and HaldCLUT files and let people decide what they want to use. It will take me some time though…
For work such as this, wouldn’t it be nice if RT had an eye dropper/pipette that we could lock onto a certain spot, so that we could follow the change of values while dragging a curve?
If you get me the resulting Hald CLUT files I can pass them onto 