Skin tone test pattern

I'm unsure of what you are trying to accomplish. Are you talking about the Flesh Tone control on Samsungs? If so, I've measured the differences between the minimum and maximum settings and not much shows up when measuring 100% saturations patterns. Now that 75% pattern measurement is enabled in ChromaPure perhaps there will be something to see. Adjusting the control does make a small noticeable visual difference but I would rather get the colors correct in the CMS rather than doctor the finished product with anything other than the Color control.

As to your patch, everyone's skin is different and there is no way that one single color point can represent it. Why not use known material to check for visual accuracy? I use everything from the ISF/Monster disc (3 women) to the Qtec High Definition Reference Disc (Color Girls 1 & 2) to a Blu-Ray that I cut myself. If flesh tones are wrong there is never any doubt.

I don't think there is anything about the Flesh Tone control or your pattern that is worth measuring. When you are finished calibrating normally, and if using the control makes your skin tone references look better visually, then use it. My experience is that it is not required for Samsung 7000 and 8000 series LCD and PDP displays but that may not be true for your particular TV.

What is the difference between the color control and flesh tone control?

great thinking! wrong skintones are THE most annoying thing and easy to see with a human eye.

:thumbup::thumbup::thumbup:

Flesh tone discretely increases or decreases red saturation in very minute steps, helped on my previous Samsung plasma, D8000 model.
Believe Zoyd also used this control.

On a few shows lowering it too much makes people look like they have lipstick on.Raising it adds some nice extra color pop on some shows although overatursaturates .

Just trying some combinations of white balance and flesh tone.Raising flesh tone control up and raising color temperature up. I have no cms on samsung e450.

There is no standard for "skin tone." The color of human skin varies from one person to the next and even on the same person. Emotional states affect the color of the complexion. So do physical exercise, general state of health, diet, sun exposure, wind exposure, external temperature, etc., etc.

How human complexion appears in video programs is a constantly moving target. Variables include the above, and also technical issues, such as: set lighting, wardrobe, makeup, scene changes, reflections from nearby colored surfaces, final tweaks in mastering by the colorist, etc., etc.

The control on Samsung's TVs is not a universally used feature across the industry. This control should be adjusted for least effect possible, ideally zero, and left there. The color performance of any video display should simply be aligned with general video standards. That is the only way to insure the faithful reproduction of a program. Fiddling with color controls by eye to achieve a more individually "pleasing" result is simply guessing and personal preference. Individual viewer preference has no common place in display calibration discussions. We calibrate to standards, not every audience member's whim, fancy, mood, or notion. The fundamental principles of imaging science are universally based upon human visual perception, physics, and electrical theory. Forum members frequently go astray in their presumptions about video and display devices due to an insufficient understanding of these fundamentals. Forming a logical construction, launched from a faulty premise, inevitably ends in unnecessary complexity, confusion, and failure.

Best regards and beautiful pictures,
G. Alan Brown, President
CinemaQuest, Inc.
A Lion AV Consultants affiliate

"Advancing the art and science of electronic imaging"
Edited by GeorgeAB - 1/20/13 at 9:50pm

This control on the samsung moves the y-coordinate a small amount for colors in the vicinity of the Munsell color checker flesh tone patterns. One can use it to reduce errors for these patterns but the improvement is quite small compared to it's zero setting assuming you have properly aligned the primary/secondary colors already.

Too much red / burgundy can make skin tones look horrible on SOME shows like they have been living under a sun lamp.

fleshtone control has more saturation effects on sd vs hd on the sammy e450 here. Anything higher than +5 on sd saturates skin tones too much ( higher number on hd).Anything too low is undersaturated and faded looking on sd(on hd lower than about -8 starts the lipstick lips look).Probably different numbers on different tvs though.

Skin tone is one of the most recognizable features in color correction and human eye, unfortunately each of us think of a "neutral" skin tone and it is different from each other.
Perception, social places, cultures, all are part of what we understand as "skin tone" for a picture.

I grew up in Italy and there i started to do color correction, now I'm in California and the color I'm asked as neutral is really different than the one we are talking about back in Rome!!!

I think that as a general idea a single flat field skin color is a bad idea in general being the display (any display) easily fooled by a flat field, on top of it if you are using any sort of math/LUT to fix your color space you will incur in interpolation errors.

If you do not have an image that is good and represent a true color skin: which one do you like, white? cucasian? japanese? asian? african? native? tan? pale? Porcelan? joung? old? what is a color skin? Redneck? office guru? farmer? It is better to me to try a different approach...

assuming that the color meter you have can average a particular area, why we cannot envision a random pattern made of 5-10 colors that represent the average of skin tones? I think that should be not too difficult to do.

another approach is to go with a chroma meter around the daylight and get real measurement of bellies at the beach (fun experiment) and average some of the XYZ measurement to come out with a cloud of points that represent the real skin tones....

Even the most common source of large skin tones (.... porn .....) is made on different lights conditions sometimes to emphasize some features: soft light for glamor, bad light for "amateur" (even if they have 1000s of lighting and cameras on the set), red/candelight because it looks better...

go outside, put your belly under the sun and you see which color is your skin tones!!!

-Walter
(Digital intermediate Colorist at Fotokem...)

The problem of a RGB absolute value is that only have meaning in the original display that has been created for, even when you transform that in XYZ you implicitly embed the color space you use for the RGB->XYZ formula.....

Sometimes I like to raise the colors up so you got cartoon characters.Not technically right,but better than a dark colorless screen sometimes.

I'd guess a lot of skin tones look yellow toned in the sun.
Edited by Vic12345 - 2/1/13 at 10:19pm

Let's all just guess.

Every channel/ program probably would look best with a different setting.It also depends on what color temperature your using(lower temp would need less red, higher more red possibly) .i think my tv MIGHT look better with a higher color temperature so the skin tones are more white than 6500.


If Im guessing I would say a - number on hd and +1-2 on sd.


The adjusting continues.
Edited by Vic12345 - 2/2/13 at 5:20pm

Nada ,on the Sam e450 plasma flesh control affects yellows oranges reds etc. .. I think turning it up adds a good rich saturated look to the whole picture and skintone..it does not affect grayscale much if at all..and I'm also using native colorspace which cranks up the greens,yellows and others possibly.

Is this for a blu- ray source or cable? Native is not accurate to use for color space , cranking up the greens and yellows seems like it's oversaturating the colors and making them too bright. This is more preference and not accuracy. Accurate is Auto if you don't have a meter. If you do have a meter then just use your tv's cms to dail in the colors. If this is for a source that is say cable then it's very hard to get accuracy since every channel will be different.

also, there is no CMS on lower end Samsungs like the E450 plasma... so auto is the right choice even if you have a meter

I've tried various techniques of matching The tv to an iPad which is around 7000k and It never looks that good(too dull).The out of box warm2 setting looked overly red grayscale and skin tones had problems.

My latest Technique is doing the grayscale first( with scrolling burn protection screen) then flesh control ( this control works good for me mostly with yellows) too add rich color and balance the over saturated greens ,then Mostly for skin tones the tint and white balance and last color control.Then watch it for a few days and see which colors are too much or too little and adjust it.the greens are excellent but a bit oversaturated.Yellow shows most improvement.I'm trying to get as much color as possible,without making the skin tones look too oversaturated.Gamma+2.


I only watch cable
Edited by Vic12345 - 2/17/13 at 5:29pm

the proper way to set white balance and gamma is with a meter and it's best to leave settings like flesh control at their default setting (zero on the Samsungs) regardless

you can also use a meter for color/tint... which is better than methods involving color filters or even color isolation modes

This whole discussion is really a waste of time. My "fleshtone" is different in the winter and summer. Which one are you going to use? What if you pick a "fleshtone" "reference" that makes Asian, Latin, African, or middle eastern fleshtones look wrong? Is your fleshtone reference for fleshtones inside under incandescent light? Or outdoors? If it is outdoors... was it a sunny day or cloudy day (you'll get a very different reference either way)? Were there a lot of trees (adding a lot of extra green to the light)? Were the walls painted blue (which will add blue to the fleshtone reference)? What was the time of day you decided on for your fleshtone reference? Because any other time of day will make the fleshtone different (the sun moves and the shadows will move). There are a bazillion variables... which means ZERO "references".

The only way you're ever going to get images to look righ is via calibration with a meter and software and a reference video source... either do it yourself or hire a pro. Otherwise everything you do is a TOTAL guess and is about 99% certain to fail your objective. What if you make your reference fleshtone look (what you think is) right, but stop signs then look too orange. You think that won't bother you? Or what if your perfect "fleshtone" adjustment makes grass look too blue? You'll be happy with that? That's the kind of thing that happens when you muck around doing things without the guidance of a meter, software and an accurate pattern source. When you get red, green, blue, yellow, cyan, and magenta as accurate as your TV can make them, and when you get the grayscale and gamma as accurate as your TV can make them, THEN the fleshtones (all of them from every corner of the globe. every time of year, every lighting condition) will be as accurate as they can be. And that is the ONLY time ALL fleshtones will be that accurate at the same time.

tbh sounds like another gimmicky post processing setting? I'd turn it off.

Also no way of knowing unless the movie is of someone you personally know, say if you're friends with an actor so you know what his/her skin tone is like, ask them to stand next to your TV and calibrate the image on screen to the person, and the lighting is spot on with no colour filters whilst it was filmed.

I can only identify wrong grayscale when I first look at tv. If I look too long it does not work..i look, adjust it then check it the next evening.

Isn't it best to get all the colors to appear on the tv as if the midday sun is shining on them?

Clearly not, because TV contents are made under every light condition.

Did you not just read the second post ahead of yours? I can't fathom how Doug could have made it any more clear! This thread is hopeless utter nonsense if reading comprehension has reached such a low state among the readership. What an awful waste of time! A dead horse, subsequently beaten, is not any more dead, regardless of how much the assailant insists otherwise.