Getting The Best Skin Tone

T

tommygdawg

Well-known member
Hey all,

This may seem like a very broad question, but I was wondering if you all have any tips for getting the best skin tones. Which color temperatures tend to look best against different types of skin? One thing I've noticed is that I sometimes have to choose different colors and intensities of light for people of different ethnicities. Any tips for that? Basically, I don't want that really color corrected feel, and I really don't want my scenes to feel lit (though I suppose that's another topic). I want that really nice skin tone that you see in higher end productions, and I'm not really sure how to get there. Thoughts?

Thanks!
- Tommy
 
You should throw some money after "Lighting Set Techs Handbook"....

There's no one answer for your question. It all depends on the ambient light, the light you're using to light the talent(s) and what skin type they have.
Though I hear again and again that tungsten might be the best to render skintones, while HMI and Flos give of a slight greenish temp which you have to correct. LEDs are bad... they are good for hit & run setups etc because they make you efficient, but they do nothing for skin.

Read the book... it's very good.
 
legrevedotcom said:
You should throw some money after "Lighting Set Techs Handbook"....

There's no one answer for your question. It all depends on the ambient light, the light you're using to light the talent(s) and what skin type they have.
Though I hear again and again that tungsten might be the best to render skintones, while HMI and Flos give of a slight greenish temp which you have to correct. LEDs are bad... they are good for hit & run setups etc because they make you efficient, but they do nothing for skin.

Read the book... it's very good.
Click to expand...

Will check out the book. What about LED makes them bad on skin tone? Are they still bad even if you gel them down to the color temp of tungsten, or is it the actual difference in light type between LED and tungsten?
 
I prefer tungsten lighting units for many reasons, but this is one of the biggest!

HMI's give beautiful light as well, but I feel it has a sheen on the skin, a metalic feel, whereas tungsten just absorbs into the skin. Of course I have lit subjects with HMI lighting, so this isn't something that will smack the viewer in the face or look terrible. But on a very small aesthetic and subtle level, the HMI lighting tends to look different on the skin than tungsten. Maybe it's just me.

And this is not because of the possible Green/magenta spikes you can get, which is an issue all of it's own.
 
Ryan Patrick O'Hara said:
I prefer tungsten lighting units for many reasons, but this is one of the biggest!

HMI's give beautiful light as well, but I feel it has a sheen on the skin, a metalic feel, whereas tungsten just absorbs into the skin. Of course I have lit subjects with HMI lighting, so this isn't something that will smack the viewer in the face or look terrible. But on a very small aesthetic and subtle level, the HMI lighting tends to look different on the skin than tungsten. Maybe it's just me.

And this is not because of the possible Green/magenta spikes you can get, which is an issue all of it's own.
Click to expand...

That's strange.
Could it be that the HMI lamps,generally being more powerful, simply create more noticeable specular reflections?
 
Shooter said:
Post up some grabs of what you like...and start a discussion.

Its a great topic to get into.
Click to expand...

Great idea... I remember seeing Collateral and thinking that the sickening skintones in the car was pretty damn good and carried the mood well. I guess this is one case where my idea of LED is actually put to shame since the light served the mood of the scene pretty damn well.
They used Rosco Litepads amongst others in that one btw.

OP, you have the chance to help pick some brains here. Get busy ;)
 
gonzo_entertainment said:
On the "everything has a purpose" front. LED looks terrible, BUT there are times when that is fine and works with the story.
Click to expand...

In the most recent American Cinematographer re: "Prometheus".

http://www.theasc.com/ac_magazine/July2012/Prometheus/page1.php
--------
The LEDs and fluorescents had variations in color. Wolski states that both daylight sources had green spikes, but rather than try to correct the color temperature, he chose to, within limits, use the greenish and bluish tints as realistic industrial lighting. “We decided to play the lights for what they were. Some of the cool white LEDs are greener, while some are bluer. We just let them play as the sources on the ship.” Indeed, the ship’s interior is distinguished by a rich variety of blues, greens and yellows that were heightened by Wolski’s choice of a consistent 4,000°K setting on the Epic, which also yielded slightly cool exteriors.
--------
 
Yeah LEDs can be extremely ugly on skin.
I try to avoid them, whenever possible.
I even prefer tungsten over Kino tubes.
Shine a PAR can through a 4 x 4, looks better than any expensive Diva IMHO.

The CRI is not an absolute figure.
A light with 95 CRI can be ugly on skintones and an other light with just 80 can be pretty perfect.

Frank
 
cpc said:
That's strange.
Could it be that the HMI lamps,generally being more powerful, simply create more noticeable specular reflections?
Click to expand...


It might be a skin pigment thing. I don't know.

HMI's and Tungsten units do work in very different ways. HMI's are small arcs and are not true continuous lighting. They are small and quick bursts of arcing electricity. They also work around daylight color temperature. Tungsten is a pure and continuous light that works around 3200* kelvin. I'm going to assume that what I see is likely flesh tones, the actual skin and pigments, better responding to warmer light, even when the camera is correctly balanced for 3200 or 5600 when shooting HMI.

As I've said, this is on a very subtle level. I would not refuse to shoot a close up if my entire camera package was HMI's. Just an observation. Could be all made up in my mind like those voices in my head that tell me what to do. :)

The way a face looks in close up is a challenge. You don't want it to look unreal, such as having makeup go overboard with the shiny powder. Also the reason I refuse to shoot with variable ND's. They can remove the natural sheen the skin has, even the sheen you get with tungsten, and it makes the face texture look strange. I use real ND's. If I'm doing something that is run and gun, like a documentary, I try to use a camera with built in ND, like C300, F3, HPX-170, etc.

I once was working as a 2nd AC for a couple of days on a film Phedon Papamichael, ASC was doing. The one thing I remember from the whole experience was that he said to never use a pola in close up. Makes sense to me for it confirmed my observations stated above. It can do some unnatural altering of skin tones.

Who knows. Skin tones are so important and realistic good looking skin texture is as well. It's an art, that is for sure.
 
Last edited:
Just on the tungsten vs LED topic, i have seen a graph before showing the amount of light across the spectrum that Tungsten lights give out as opposed to LED, although i can't find it now. As i understand it, Tungsten lights project a fairly uniform amount of each colour of the visible light spectrum (with the obvious bump at the yellow/orange section) where as LED's have massive spikes in in the green(predominantly) and magenta portions of the spectrum. The document that i was reading at the time was giving this as a reason as to why skin tones always look better under tungsten, regardless of how you gel the lights, as the original light isn't "full spectrum" with LED's. I think it was a Kodak presentation actually. Anyway, this is all fairly useless without the graph, but it was an interesting point to make and does seem somewhat logical. Fluorescents were somewhere between the two on the graph by the way.

Basically: Large tungsten sources.
 
Ryan Patrick O'Hara said:
It might be a skin pigment thing. I don't know.

HMI's and Tungsten units do work in very different ways.
Click to expand...

Tungsten still has quite a bit of 'red' spectrum. HMI has less 'red' spectrum, but also does produce light in the UV portion of the spectrum, but a 'glass' cover, lens, whatever, should filter that... but perhaps there's still more UV content than with Tungsten. I think the effect would be minimal in ordinary film making... but human skin does have florescent property under UV light... but back when I use to be in houses that had lots of black light posters... I can't remember much... but I think people's skin was pretty 'wyrd'...
 
j1clark@ucsd.edu said:
Tungsten still has quite a bit of 'red' spectrum. HMI has less 'red' spectrum, but also does produce light in the UV portion of the spectrum, but a 'glass' cover, lens, whatever, should filter that... but perhaps there's still more UV content than with Tungsten. I think the effect would be minimal in ordinary film making... but human skin does have florescent property under UV light... but back when I use to be in houses that had lots of black light posters... I can't remember much... but I think people's skin was pretty 'wyrd'...
Click to expand...

excellent points!
 
A few random thoughts.

First, part of what gives peoples' skin its color is the underlying blood. Blood red is blood red, to the point that there's a "skin tone line" on vectorscopes. Line up a skin tone on that line, and it looks pretty good. Ethnicity doesn't much enter into it. That said, people can take that to the extreme, which may in part explain your "really color corrected feel", IDK. We'd know more if you were to post some examples of what you like / don't like. Right now we're just guessing.

Lights have color spectrums. One of the distinguishing features of a tungsten light spectrum is that it doesn't have much in the way of gaps from end to end. Its spectrum is considered to be continuous. Another feature is that it shoulders on the blue end of the spectrum, so it has more red than blue. If you can think of it this way, that's sort of like putting a minus blue filter in front of a white source. And the opposite of blue is orange. Which explains the sort of orangey glow that skin takes on under tungsten lights. And just coincidentally, the tungsten spectrum bears a striking resemblance to sunlight during the "golden hour". For whatever reason, people really like the look of other people under these lighting conditions.

The problem with non-continuous light sources (fluorescents, HMI's, and LEDs all fall in this category) is that their specrums show gaps. Also spikes. And this makes them significantly different that sunlight, which is what our eyes evolved to respond to. So we tend to see these discrepancies and not like them. Fluorescents in particular have a green spike that corresponds to the mercury they use to arc, which creates the UV used to make the phosphor coating(s) inside the tube fluoresce. And our eyes are really sensitive to green, so we see even a small green spike. Part of what can be done to fight this is to use more phosphors. I've seen tubes with as many as seven different phosphors, and the results are really quite good. Almost like sunlight. All this is doing is allowing the designer to bring down the big green spike while adding other spikes in other frequencies. This in turn results in fewer gaps and smaller spikes. Good things.

LEDs are actually fluorescent sources too. And it turns out that they have some similar problems with their larger and less efficient cousins. As the technology matures, the light quality improves. But it can be pretty ugly yet.

HMIs are arc sources. These too have spectrums that show some gaps and spikes. But modern HMIs have been tamed a fair amount and tend to spike more on the blue side where we aren't as sensitive. Still, modern HMIs have a ragged enough spectrum that they can't get a CRI much out of the 80s. And this does effect skin tones.

The big hope (well, for me at least) is plasma. The little plasma sources have a more or less continuous spectrum with not much in the way of gaps or spikes. But it's a young technology and therefore still very expensive. And you can't get the power levels that HMI (or even tungsten) can bring.

As to that "high end production skin tone look", at least part of it is coming from the lighting. But part of it is coming from either film, or now more likely high end digital CODECs that give you 4:4:4 chroma subsampling. This is important because of the gradients. Well lit faces are full of gradients as the light and shadow shape the face. The ability to see and record this as perfectly smooth changes in hue, saturation, and value is part of the magic that we loose when we use CODECs that compress more. AVCHD, for example, uses 4:2:0 chroma subsampling. It can't compete with a high end CODEC for skin tones, but it's several orders of magnitude more efficient.

Bottom line if skin tones are your game, lean toward tungsten lighting and when you are buying / renting cameras, look for CODECs that give you higher chroma subsampling if you can. But taking advantage of what the lighting and camera can do requires skill, so practice, practice, practice. But you already knew that.
 
Has anyone here worked black and white in a dark room?

Lets talk black and white for a moment and then come to colour.

You know the trick with a red filter, doing landscapes, makes for a black sky (with fluffy clouds very white) the red filter has filtered out the blue sky rendering black
http://www.flickr.com/photos/chris10585/6783186496/

OK now in B+W putting a blue filter on the lens is called the 'old man filter' it creates a grizzly look

What is occuring here is that the skin is red, and imperfections in the skin are darker red, the blue filter is taking out the red, so the dark reds go really dark
This creates the grizzled old man look

Conversly beauty portraits of women were typically (in the 1950s) done under tungsten lights.

Now the woman is lit with red.

So if her light skin might be 88-55-55 (RGB/100) and her imperfections might be 88-22-22, shoot that under red filtation and all her skin renders as 88-00-00 and her imperfections render at 88-00-00 too

Magic, all the imperfections are the same colour as the skin so dissapear !

Of course shooting a pure red filter would make the eyes and hair render 00-00-00 (black) so shooting with a pure red filter is a no go, but some red bias smooths the skin

Bingo - shooting tungsten makes for great skin tones

Now we move the image to the camera sensor, and the WB bings the colour back to 'normal' but because at some stage in the chain 88-55-55 and 88-33-33 became say 88-22-22 once the chanels are amplified back to say 88-88-88 the imperfections have melded into the skin

The magic of tungsten!

MV5BMTQxOTk1NTQ1M15BMl5BanBnXkFtZTcwMjI1MDAwNA@@._V1._SY314_CR19,0,214,314_.jpg
 
Last edited:
remoteman said:
Just on the tungsten vs LED topic, i have seen a graph before showing the amount of light across the spectrum that Tungsten lights give out as opposed to LED, although i can't find it now. As i understand it, Tungsten lights project a fairly uniform amount of each colour of the visible light spectrum (with the obvious bump at the yellow/orange section) where as LED's have massive spikes in in the green(predominantly) and magenta portions of the spectrum. The document that i was reading at the time was giving this as a reason as to why skin tones always look better under tungsten, regardless of how you gel the lights, as the original light isn't "full spectrum" with LED's. I think it was a Kodak presentation actually. Anyway, this is all fairly useless without the graph, but it was an interesting point to make and does seem somewhat logical. Fluorescents were somewhere between the two on the graph by the way.

Basically: Large tungsten sources.
Click to expand...

spikes in green AND magenta? that must be a ***** to correct.
 
Ryan Patrick O'Hara said:
It might be a skin pigment thing. I don't know.

HMI's and Tungsten units do work in very different ways. HMI's are small arcs and are not true continuous lighting. They are small and quick bursts of arcing electricity. They also work around daylight color temperature. Tungsten is a pure and continuous light that works around 3200* kelvin. I'm going to assume that what I see is likely flesh tones, the actual skin and pigments, better responding to warmer light, even when the camera is correctly balanced for 3200 or 5600 when shooting HMI.

As I've said, this is on a very subtle level. I would not refuse to shoot a close up if my entire camera package was HMI's. Just an observation. Could be all made up in my mind like those voices in my head that tell me what to do. :)
Click to expand...

I have seen some wonderful close-ups lit with HMI lights (here's one out of the top of my head, just to add spice to the thread: http://i629.photobucket.com/albums/uu15/fadingsapphire/Anton.jpg ), that''s why I find your observations even more interesting.
Noticeable sheen is, after all, not dependent on light color, rather light size and intensity. And, of course, on the reflective properties of the reflecting material.


Ryan Patrick O'Hara said:
The way a face looks in close up is a challenge. You don't want it to look unreal, such as having makeup go overboard with the shiny powder. Also the reason I refuse to shoot with variable ND's. They can remove the natural sheen the skin has, even the sheen you get with tungsten, and it makes the face texture look strange. I use real ND's. If I'm doing something that is run and gun, like a documentary, I try to use a camera with built in ND, like C300, F3, HPX-170, etc.

I once was working as a 2nd AC for a couple of days on a film Phedon Papamichael, ASC was doing. The one thing I remember from the whole experience was that he said to never use a pola in close up. Makes sense to me for it confirmed my observations stated above. It can do some unnatural altering of skin tones.

Who knows. Skin tones are so important and realistic good looking skin texture is as well. It's an art, that is for sure.
Click to expand...

Yes, it is common knowledge amongst photographers to avoid shooting people with polarizers. Obviously, under skylight (day exterior) you are losing polarized light from the sky getting reflected by skin. But what people often miss is that skin itself acts as a partial polarizer (polarization occurs on the boundary between two dielectric materials - air/water, air/glass, air/skin, etc - with different refractive indices). As a result, lots of the subsurface scattered light is rejected by the filter. Incidentally, exactly this light is responsible for the glowing quality of skin. So polarizers deny both the natural sheen (reflection) and the natural glow (subsurface scatter) of skin. Same with vari-NDs.
 
You must log in or register to reply here.
Back
Top