Filmic changes bright yellow to pink

Are you positive they’re not clipped? I think they are, and their green component hasn’t been scaled to the display white like the other two channels.

I’m not that familiar with @anon41087856’s filmic module yet, but it sounds like the desaturation setting (??) needs to be adjusted to allow the green clipped data to join it’s red and green brothers in the “desaturation oblivion” to display white. I see this all the time in my sunrise images; if I expose toward the shadows, the blown portions have to be desaturated to remove the magenta cast…

Edit: posting your raw file would tell all, if you’re so inclined…

I can confirm your observation. It is true filmic does cast yellow color on 5500K sun-lit white/grey objects. And not just yellow, other colors too. This artefact made me stop using filmic. It’s good for lightness, but not for chroma.
My guess yellow-ish cast is caused by the nature of filmic itself. When ‘chrominance preservation’ is on, it remaps original RGB values on modified lightness, which of course gets colors out of the line.
Did I get it right @anon41087856 ?

Yeah, thank you!
Base curve is off, of course.

Sure I am. It happens not with only very light objects but down the line, too. Yet not on every shot, but on quite some. And it may explain the greenish tint on light sky but not how the yellows become pink.

Just avoid chroma preservation mode until the issue is solved. Overall I’m quite happy with the filmic module, for it allows me to to things harldly achievable with my usual workflow. For instance, I’ve Nikon 50f1.4D lens – a good workhorse from the film age. But because it was designed in 1995 and because Nikon is so lazy so it keeps selling it with all the original flaws instead of making mk II (that would be quite easy), it has a quite average coating (comparable with current Chinese chip lens) that introduces grayish veil that makes images look dirty. If I use base curve (that comes very early in a pixelpipe AFAIK), I could hardly do something about it. Saturating did not help, producing oversaturated yet still dirty colors))) It’s beyond my understanding, but that’s what I see. Otherwise it’s an outstanding lens that gives a very live, 3D picture – much more interesting that its successors, so I have virtually no alternatives to this lens.
In this situation filmic helps a lot, or I’d have to compress dynamic range with multiple hipass instances, or use shadows/highlight that produces halos, or in another much more complicated way.

I shoot mostly landscapes and color preservation is important for me. While delivering very nice luminance, filmic loses a lot on color part. Resuced chrominance values if on do not help to remove the color cast per se. So I switched to the tone-curve only workflow. Maybe filmic in this part needs some upgrade, but I just assume this is how it is supposed to work.

I am not a fan of chrominance (or whatever else it is called) preservation. It always seems to have trade offs in dt and phf. Not saying it can’t be fixed; just not very reliable. I would use color balance in tandem with filmic.

I’m also convinced that the math behind what is called “chroma preservation” is not correct, or at least the current math is not providing what AFAIK is the intended goal, that is to preserve the original chroma and hue of the colours.

I have tried to explain this in the past, and proposed a possible solution. You can see the beginning of the discussion here.
@paulmiller also arrived to a similar conclusion (see here and the following comments from him).

Not sure, but I think this has been fixed.
I did a few tests and this over saturation effect doesn’t show anymore.
image

EDIT: Something has definitely changed. Here’s the same raw @Carmelo_DrRaw has used in the post he referred to above, with the same settings he has used that produced a very saturated image:

Okay, this is good news, however what I suffered from wasn’t oversaturation but actually a color shift. Yet it might be all in the same basket (like oversaturation of a red channel, for instance). I think I shall try a git version and find out how it behaves. If the issue’s gone, I’ll have to stick to a git version until it is released… which is not I’d love to do but absolutely fair and fully acceptable. Will probably report more issues then)))

@Carmelo_DrRaw Thanks for the links, but I feel it’s a little too complex for me now, for I don’t have a clear understanding how the Linux CM works. However, I noticed one strange thing while trying to comprehend what was written. How I understood the discussiion, filmic tries to involve an output ICC profile in its math – that’s what I don’t understand. It’s not its business, is it? My understanding is that output profile is applied at the very end of the pixel pipe, and none of the other modules shall be aware of it at all? Am I wrong?

1 Like

This implementation of filmic is applied on the RGB model. But there are many RGBs! I guess that is why it requires an ICC profile.

Built a git version. Overall I liked the interface much more, however I wouldn’t say the color issue with filmic has gone.


This is a crop showing yellow cast without chrome preservation (look at the stone and faces)


This one with chrome pres on (faces are unnaturally red, stone has a pink cast)


This one is with the regular workflow: faces are of normal color (giving the light was yellow), stone looks like it was in the scene). Not to say that filmic destroys the sky colors in both modes (not shown on the crops).

This quick comparsion is not quite accurate, for I just copied the last stack from another shot so white balance and some other things differ, but it still gives the whole idea what’s going on with filmic color-wise.

Here is the RAW:
20170626_0028.NEF (28.0 MB)

1 Like

There shouldn’t be errors in the math, that’s due to the hue preservation. Even Rawtherapee shows the same behaviour using the hue preserving curves (film-like)

Standard rgb

Film-like

As I have mentioned in the other thread, there will always be interactions among channels. Therefore, any attempt to manipulate tones will inevitably mess up the colours. When you factor in colour space and models, and assumptions that you have to make in how to separate colourfulness from luminance, you are in for a lot of variation. The unsatisfactory answer is that we do our best but it depends on what I just wrote and it depends on the input image.

That is why people like to dream about and create images using machine learning. In that setting, you have a lot more data on your hands to determine what would constitute the right colours but at the same time it won’t output radiometrically accurate results, just expected values.

2 Likes

I wouldn’t call it “hue preserving” because it obviously is not such. The Standard RGB one you provided is much more accurate in color rendition as you can see. And it is much better than both of my filmic attempts. Maybe I can get decent results by shifting color temp toward blue with chroma preservation off, but this would take more fiddling than not using filmic at all (and most certainly will spoil other colors). What a pity – initially I liked filmic and was glad it showed up. Well, maybe some solution will be worked up with the color rendition at some point…

What’s wrong with Lab space then? Why not using Lab?

The light in your examples looks extremely yellow… To me it makes sense that they turn yellow/red. If you white balance for that color or light, I think it’d get better.

Sure it is. The sun went bright yellow under the atmospheric conditions. That’s what the shot is about. And that’s why it’s difficult to develop. (Actually there was also great double rainbow at the moment I saw the scene, but crossing two roads to get to a shooting position took few minutes and the rainbow had almost gone). But the unusual light was still there, so two attractions turned one.

Twp things I’ve noticed about filmic: you need proper white balance and a good exposure. If you’re set to daylight for white balance with preserve chroma on, these are the results I’d expect. The light is yellow, so everthing gets saturated yellow. The direct sun puts lots of those tones in the lighter part of the tone curve, so when preserve chroma is turned on, it goes BAM! Colors.

Yeah, probably this particular shot is where filmic would fail, while I played with it a few days and sometimes got better results than when using a standard basecurve workflow. That’s what makes me think there’s something wrong with the math (or rather, with the approach to color processing). And at some conditions that error is better pronounced than in others. What I can say now is that for me it gives inconsistent output color-wise from shot to shot. That’s not the module I would want to put in a “standard” preset, though.
UPD: I usually underexpose shots in difficult condition, because pulling up shadows is possible while blown-out highlights are blown forever. That’s maybe the reason I get inconsistent results with filmic.

I also shoot to preserve the full tonal range, which generally means preserving highlights. If nothing is clipped, then yes, lift the shadows, which I find quite easy with filmic. If shadows are clipped, I make bracketed shots, which I generally then process with the base curve.

When I have an unclipped shot in one file, filmic is great.

There is another quite long filmic thread where @anon41087856 explained a lot of the math. If you are maths inclined, I’d suggest you look at that thread to see if you can spot the error you’re speaking about. If the error isn’t in that math, then perhaps there is an error in the code; but you’d have to read the code for that. The thread is: Introducing the filmic module in darktable

I think I shall read this tread, thank you. Hopefully I find some clues and how-tos regarding the filmic there. However I quit programming more than 25 years ago, so I obviously can read the code but there’s little chance I would understand it))) A lot of time… Now I can only helplessly waving hands crying “Possible error! Possible error!” And I may easily be wrong with that. Sorry.