You get a reference dot now for a color picker selection in the vectorscope. Maybe you could click that dot and have the option to set a reference line mapped to that point created by the color picker??
This guy made a simple and clear explanation about how to read and understand the vectorscope, it will be helpful. I think the skin line indicator must be something like a guide, others may prefer a range of where the skin tone should be. If users like to work with it, fine, if don’t, equally fine, but they will have the choice in their hands, not the lack of it. Even an option of isolate the colors to have more control.
Anyway if devs are working in improvements maybe there will be an option of show or hide the skin/range indicator among other visual utility indicators.
There’s a difference between photographers and color grading staff. The latter needs to deliver consistent colors over plenty of single frames - and keep in mind that skin tones are detected as skin tones even when using heavy hue shifts to achieve an artistic look. The „real“ skin tone usually doesn’t matter.
Most photographers care about real skin tones, not about a norm skin tone.
Well, once we have put our cameras down and sit in front of software, we are all color grading staff. =)
The photographer in you might not care, but the Master Printer in me does. Giving consistency over hundreds or thousands of frames is one part of successful visual storytelling - wether the images are static or moving.
It’s just that the moving pictures industry is way ahead of photographers here and has the more evolved tools already.
i doubt hundreds or thousands of portraits share the same skin tone.
In movies no one cares about the real skin tone of an actor; in portraits most clients will be disappointed, if the real skin tone isn’t properly matched.
As much as I agree to your statement - I am partly colourblind. So, I never touch colours when skin is in the picture. However, I would like to do it . Unfortunately, I can never be sure if the skin does not turn up greenish or otherwise unfortunate. Some kind of indicator if skin tones are acceptable would be highly appreciated.
…to their personal unrealistic ideal. Yes.
But I’m thinking more about practical problems when doing reportage work under changing light conditions. Often the skintones is the only good color reference because people still do not wear white or grey cards as clothes.
And those situations do not really cater to the “everyone stop, I have to take a greycard/colorchecker reference photo first” - and just as often it is not practical to do that after the action has moved on.
PS: Does anyone know what those 3 other lines represent in the vectorscopes? Or are they just for orientation with quadrants starting from the skin line?
Those lines are
Q from YIQ. The idea is that
I+ roughly coincides with skin tones because of the blood vessels in and underneath the skin. It is antiquated but people still use it. In the digital world, we use the chrominance of YCbCr and friends.
I am almost scared to add a comment here in fear of criticism so be gentle. While I appreciate that skin tones vary greatly they are also surprising similar. When I had a colour darkroom for film I had a densitometer. I printed a pleasing colour of a portrait and then set the dials on the densitometer to read this as skin tone. Then I could get another negative and adjust the printing filters based on the densitometer. Surprisingly if the person was Caucasian, Asian or had a very dark skin colour the result was in the ball park and only a small correction would be needed. Since I could not place a grey card in every shot when photographing people this skin tone setting was the best option. Something like that in the color calibration or white balance setting would be so helpful, but would require the ability to fine tune the colour or even have options for various skin tones. BTW, when I run my color printing lab it was known that Europeans preferred different colours to Americans and the Japanese different colours again. Not just talking skin tones but general colour of images. At least that is what Kodak, Konica and Agfa taught me.
And that’s where the differences come from:
- some work from what they consider a pleasing/acceptable skin tone after editing, and notice that a lot of images end up with vey similar skin tones, giving the “skin line” on the vector scope;
- some start from what’s there, and try to see and keep the unprocessed skin tones, they see a much larger variation in colour, beyond what can be represented by a single line on the scope.
So I have to agree with those that don’t want a fixed line on the vectorscope, claiming to represent “the” skin tones. That would in practice lead to either many users “fixing” their images to fit that line, or complain about that line being wrong/rubbish (or both, possible from the same users :P). Having to option to set a marker on the scope, otoh, could be very helpful to get consistency within a series of images.
If all you want is to get similar skin tones, darktable already has a color picker and you can use something like the tone curve in LAB mode to match them. Or really almost any tool that can manipluatle color should be fine.
@paperdigits , it’s not about being able to edit.
It’s about seeing the image data in a different way.
Vectorscope, waveform, histogram and so on are all visual analyzing tools.
The color picker is probably the most crude and primitive of all and in most cases the least helpful because it is way too fixated on single pixel locations. I have fought my way through hundreds of color precise stilllife image edits with just the color picker. I hate that tool with a very profound passion.
Distilling the thread to the core need for “indicator” would be: “need some way to have reference visible on vectorscope(and possibly other histogram related tools), just to be able to better match stuff - would be cool for eg harmonizing colors across the shot(s) or as a help for those with worse (color) vision”
Is this de-charged of any “painting by the numbers” or “one true skintone” or other unnecessary stuff?
Could you have a tool similar to area whitebalance that instead creates a line in the vectorscope based on the selected area? The ability to store and name the setting would then give a very flexible tool.
The line could also be an area/three lines etc.
if it’s time for crazy wishes then integrating this https://color.adobe.com/de/create/color-wheel might be the one-to-rule-them all solution
What about an angle instead of a single line?
That way we could get away from the pseudo-accuracy that NTSC¹ tried to convey.
And on the topic of skintones …
Would it be worth the effort to create a skintone angle from a survey?
The pixls.us community should be diverse enough to come up with all kinds of skintones.
¹) especially since NTSC is referred to as Never Twice the Same Color
To play devil’s advocate, should we really be criticising those who do like to “paint by numbers”? While I think it’s excellent advice to encourage people to edit their photos in a way that’s aesthetically pleasing rather than what the numbers are doing in the background, we also need to cater to those who either have some visual impairment or simply prefer to take a more technical approach to their editing.
The software’s philosophy should remain: to provide powerful tools to process photos the way users want; not to define their workflow or artistic vision.
no one is criticising those who wants to “paint by numbers”. They’re just criticised if they expect darktable changing its priorities
The phrase “painting by numbers” is being used to describe a non-artistic and inferior way of approaching photography processing. It has implied criticism in it. Have you ever seen it used on this forum in a positive way? I haven’t, but I’m happy to be proven wrong.
Digital images encode colors with numbers. Problem is many people don’t get that these numbers don’t hold meaning or truth in themselves, they just encode something that has meaning only once decoded. “Painting by numbers” applies to any person doing digital imagery who thinks that, because numbers are involved, the whole process is scientific and/or should target some sort of objective truth through the use of magic numbers.
In darktable, I’m one of the few maths guys who solve equations in their dreams and bend the linear algebra to turn algorithms into image processing tools. But when I’m done coding and do actual image processing, fuck numbers, color pickers and histograms… The very reason I did all these maths while building the tool is to be able to forget all about them while I’m retouching a picture. I’m doing maths at building time to buy myself the right to care about the look, the look only, and nothing else at editing time.
Get a calibrated screen, train your eyes, use the color-assessment mode for contrast reference, and enjoy doing some art.
It’s not so much about how you approach art, rather than the wrong belief that some numerical quantities have magic values that will somehow save your ass. They won’t.
But for color blind and disabled people, I have no idea. On a more philosophical level, I would just say embrace your disability and make us see how you saw it, but I don’t know if that’s possible.