Be my guest, develop a better one. Your speech is basically saying that one technical solution should be perfect and have no trade-offs… I have never seen any perfect solution in the real world. Converting from large to small dynamic range will be destructive no matter what you do, and what you think it should be is cool but imcompatible with reality. The best we can do is damage control. Otherwise, we simply clip whites at 100% and be done with it.
You don’t get it. The raw capabilities don’t have a spirit of their own, they are worth nothing if the soft can’t use them. And your screen or printer have a quarter to an half of what the raw is capable of (color space gamut as well as dynamic range). Also the raw is full of metameric errors that don’t match human tristimulus and its colorimetry is entirely screwed. Shooting raw is a guaranty of nothing and you should open your church if you think it protects from anything.
Yeah, right. It’s not as if I had a full portfolio of darktable-edited pics that sacrifice absolutely nothing. Portfolio - Aurélien PIERRE, Photographe Also, that technical correctness has been embraced by Blender, Filmlight Baselight, Natron and other software targeted at people who actually know what they are doing (and went to school to learn it…). Make from that what you will.
A quick internet search shows that even Constable didn’t break the “rule of the sky”: it should be lighter than the ground. Which means the dynamic range allocated to the sky is pretty much bounded by the one taken by the ground. Which means you don’t have that much contrast available for cloud (typically 2.5 EV, over 8 EV of output DR). I’m not stuck in my head, I’m concerned about the validity of what photographers want before wondering how to do it.
Did you actually ever watch one of my videos ?
That’s actually the job of the color management system (CMS) and the purpose of the ICC… Those “perceptual”, “relative colorimetric” intents from the ICC are meant for just that. If you need to manually tweak colors and saturation for each medium, it only means the CMS failed and that’s something to report to its dev team.
I’m really starting to wonder if you people realize filmic is nothing but a tone curve. It does exactly what the base curve does, only with more options and in a generalized framework that can deal with scene-referred too. The main difference is in how you set up the curve, because the base curve GUI is really impractical when the dynamic range increases (otherwise, you get the log scaling of the graph… which filmic does internally).
At the end, both only produce a fucking S curve, which is what everybody does everywhere. So I don’t get where that anti-filmic/anti-correctness rage comes from. You can also use filmic in a display-referred framework (juste rework the dynamic range bounds).
As such, filmic holds the same drawback as any S curve : compressing highlights, hence local contrast. Nothing new. What’s new is you can set filmic straight from lightmeter readings done on scene.