How to better judge skintones accuracy with the help of Photoshop

[sudopera]

I learned this technique some 6 or 7 years ago in some short Photoshop course and totally forgot about it.

There are different combinations of CMYK percentages that represent common variations of proper skintones from various races.

Of course, no skintone is exactly the same so this is used as a guideline and you adjust the percentages just a little bit for your actual subject.

Then you can throw that corrected still into your editing suite and use it as a reference.

 

Here is tutorial link:

 

http://www.graphicconnectionkc.com/skin-tone-correction.html

 

At the bottom of that page are most common skintones with CMYK percentages.

[Axel]

For videographers, the whole article is quite confusing. To judge skintones in a video clip, you should of course not use Photoshop, let alone CMYK colors.

 

Those 'typical' ranges of skintones can very easily be found on and around the so-called fleshline a.k.a. skintone line a.k.a. in-phase-indicator in the CC vectorscope, see it here on the site of Stu Maschwitz' blog.

 

Maschwitz is one of the skintone dogmatists (older article: 'Memory Colors'), who always said: Preserve the skintones! In the text above he carefully backpedales, as it seems.

 

In his Ripple training for Resolve 9 (that happens to be the one I know), the author Alexis Van Hurkman (also author of the Resolve manual!) never explains about skintones. The demo footage features heavy casts on the skintones, i.e. orange from a sunset in the desert. He does little to counteract this. At one point, he makes a light remark about that skintone line. If you want to know more, he says, follow me on my blog Thinking Aloud.

I did.

 

 

 

To clarify, I would never and have never suggested that this line is a strict guideline for human hue. In my “color correction handbook†I wrote and illustrated more pages then my editor may have wished about the subtle variations of human skin tone, color interactions between a subject and the illuminant of a scene, and how the in-phase indicator under discussion is merely a general signpost. Like speed limits, nobody follows them exactly, but they let you know you’re around what you ought to be doing.

 

As a consequence, skintone accuracy in your headline suggests a completely wrong approach. Know where to quickly check it, keep in mind that conditions make deviations appropriate. Then forget it.

[sudopera]

For videographers, the whole article is quite confusing. To judge skintones in a video clip, you should of course not use Photoshop, let alone CMYK colors.

 

Those 'typical' ranges of skintones can very easily be found on and around the so-called fleshline a.k.a. skintone line a.k.a. in-phase-indicator in the CC vectorscope, see it here on the site of Stu Maschwitz' blog.

 

Maschwitz is one of the skintone dogmatists (older article: 'Memory Colors'), who always said: Preserve the skintones! In the text above he carefully backpedales, as it seems.

 

In his Ripple training for Resolve 9 (that happens to be the one I know), the author Alexis Van Hurkman (also author of the Resolve manual!) never explains about skintones. The demo footage features heavy casts on the skintones, i.e. orange from a sunset in the desert. He does little to counteract this. At one point, he makes a light remark about that skintone line. If you want to know more, he says, follow me on my blog Thinking Aloud.

I did.

 

 

As a consequence, skintone accuracy in your headline suggests a completely wrong approach. Know where to quickly check it, keep in mind that conditions make deviations appropriate. Then forget it.

 

This post was intended more as a way to see does the camera have accurate (as we see it in reality) skintones and not for grading, because grading gives stylized looks according to the mood and your preference. Regarding CMYK, in this technique it's used by photographers because they have separate control over blacks opposed to RGB. So when you have correct CMYK values in Photoshop, you can then export that jpeg or tiff in sRGB color space which is very similar to Rec709. Then you can put that jpeg side by side with your corrected but not graded footage in your editing suite and see the difference. I think it could be helpful for tweaking the camera color settings.

Of course this would probably only work for daylight balanced shots.

 

It wasn't my intention to suggest that everyone should now use Photoshop and change regular workflow for video, but finding some kind of a way to judge color accuracy objectively(with the help of some mathematical measurements), not in the subjective "I like Canon colors, you like Nikon colors, that one likes Sony etc." way. This was just an idea, so if someone finds it helpful great, if not nevermind.

[Axel]

I think it could be helpful for tweaking the camera color settings.

(...)

It wasn't my intention to suggest that everyone should now use Photoshop and change regular workflow for video, but finding some kind of a way to judge color accuracy objectively(with the help of some mathematical measurements), not in the subjective "I like Canon colors, you like Nikon colors, that one likes Sony etc." way. This was just an idea, so if someone finds it helpful great, if not nevermind.

 

And it wasn't my intention to be "right". 

 

You know that once you start grading in earnest, it suddenly is about very subtle differences in HSL. If you never did that before, it probably never occurred to you that you were on a very low level of accuracy. As soon as it does, it's hell. 

 

I totally agree with you, but I don't see how this article can help me find a path towards reliable skintones for video.

[maxotics]

All scientific and artistic advancement is built on a pile of discarded and seemingly useless ideas and findings.  One man's garbage becomes another's treasure and the other way too :)