cinematography Question about maintaining skin tones.

I'm starting to notice this. An actor for example, is in a scene with tungsten lights, and the white balance is set to tungsten, to give it a bit of the yellow look. Then you want to do an office scene for example, and you want white fluorescent lights, and want them white, so you set the white balance to white fluorescent. Now since it's a different white balance, the actor's skin tones will be at a different saturation level from scene A compared to scene B. Will this be a problem in consistency of the look of the movie? Especially if it's in the same sequence, and the actor walks from a room with one light to a room with the other?

I could gel lights but in some locations I am forced to leave their natural lights on, such as street lights. Or will audiences find the skin tone change too distracting?
 
See my previous post - you can't reflect wavelengths that aren't there.

And again, gelling the lights won't work. Either add lights of your own (tungsten, fluorescent, HMI, LED… whatever you can get hold of) or make the orange tint part of the film.
L! :lol:
M! :lol:
A! :lol:
O! :lol:

Chilipie, maybe you should try to say that to him in English or some other language he understands. :lol:
 
Okay thanks. But why won't gelling the lights work? I was told to gel lights before to get them the color I want, and now I am told it won't work, that's what I don't get.
 
Last edited:
Okay thanks. But why won't gelling the lights work? I was told to gel lights before to get them the color I want, and now I am told it won't work, that's what I don't get.

Here's a graph from Olympus that shows the wavelengths emitted by different light sources.

lightsourcesfigure3.jpg


Sunlight and tungsten light cover the whole range of visible frequencies. The different wavelengths that make up sunlight are fairly evenly distributed, with a slight blue bias. Tungsten light is clearly biased towards the right of the graph - yellow, orange and red - something we can see for ourselves when taking a photograph of a room lit by both daylight and tungsten light.

Gels work by absorbing certain wavelengths of light. If we put a yellow gel on a tungsten lamp, it would only allow the "yellow" wavelengths of light (highlighted with the black box) to pass through, absorbing the others.

lightsourcesgel.png


But what if we were using a red barcode scanner and had put a yellow gel on that? How much light would get through?

Here's the same graph with the frequency response of a low pressure sodium lamp (which are often used for street lighting) composited in.

lightsourceswavelengthcomparison.png


You wanted to put a blue gel on that light, letting only the "blue" wavelengths through. Now can you see the problem?

(Any scientific inaccuracies are mine… simplicity seemed most important.)
 
Okay I see, thanks. Well I don't know if I can light the whole parking garage to be bright enough. The guy is really a camera operator more so, and if I can get a separate guy to light, that would be great. A couple of guys I found are good, but we don't have enough lights to light the whole parking garage, but I can look to see. It seems the only lights that could maybe work are 1000 watt halogens, but those might not be the right kelvin temperature for putting blue gels on maybe.

I will see what other really bright lights I can find. It's a shootout/chase scene, so I will be using a lot of wide shots of the parking garage, and not just closer up shots. So more light is in order.

Usually in movies though, their is a light turned on, so the audience knows where the source of light is coming from. You can see lamps turned on in bedrooms, lights turned on in offices, etc. Even though the source of light is coming from somewhere else mostly, to light the movie. Will the audience get suspiscous if they see that every natural light in the parking garage is off, and the sources of lights are coming from places that they cannot see?
 
Last edited:
In the parking garage, most likely, they're sodium lights -- orange... with no other colors in the spectrum. you won't get flesh toned flesh (not white) from those lights. If they're fluorescent, there's a green spike... I guarantee you've asked precisely these same questions in at least 3 other threads. Set up a camera and a friend, test shoot the lights as you'll use them on the day, then edit the footage together as you'll use them on the day... and see if you like it... if not, determine what you don't like about it and change it -- lather, rinse, repeat.

I love mixed lighting. White balance for the predominant light (or the light I don't have control over) so the flesh tones are as fleshy as possible. I only use tungsten, daylight and fluorescent presets, never custom white balance, then gel the lights I do have control over from there.

take the time to do test shoots. take the time to play with them. We CAN'T answer these questions for you. You have to play with them -- seriously. You'll come back with better questions.

Start with testing the things you question, then come here with more specific, more well informed questions.
 
Okay thanks. Well my question is similar to the ones as before but this time I am asking if it's okay to mix different skin tones from scene to scene. Like if an actor is outside in the sun, and it's balanced to daylight, is it okay, to walk into room with tungsten light, since the white balance will be different and the skin tone will be different.

I guess the question is, what's the limit as to how different the white balance is allowed to be before your movie looks too different from scene to scene, and it's jarring to the audience.

We have done some tests, but having rules on how far the skin tones are allowed to be pushed really helps, now that I learned this. I asked my friends, since I need some movie viewer input rather than just going with what I think is good, cause I could be disagreed with.

My friends like the pale skin in the parking garage, more, and one of them said it's because it gives the movie Saving Private Ryan skin tone look, which works for the thriller atmosphere of the script. They also said they like when the white balance is switched to shade, in the sunlight, cause it makes the sun golden yellow.

But will going from skin tones on shade, then switching to pale skin tones under white balanced sodium light, be too jarring to the audience...
 
Okay thanks. Well my question is similar to the ones as before but this time I am asking if it's okay to mix different skin tones from scene to scene. Like if an actor is outside in the sun, and it's balanced to daylight, is it okay, to walk into room with tungsten light, since the white balance will be different and the skin tone will be different.

If you shoot one scene in daylight with daylight white balance and another in tungsten light with tungsten white balance there should be no discernible colour difference: that's the whole point of white balancing.

But will going from skin tones on shade, then switching to pale skin tones under white balanced sodium light, be too jarring to the audience...

As knightly said…

take the time to do test shoots. take the time to play with them. We CAN'T answer these questions for you. You have to play with them -- seriously. You'll come back with better questions.

Start with testing the things you question, then come here with more specific, more well informed questions.

I'm sure you read the answers to your questions, but you never really seem to take them on board.

Here's a challenge for you that shouldn't take more than a few hours: go to your locations (the garage and an exterior lit by daylight); take photos (not video) of a friend using each preset white balance setting on your camera; post them in this thread (preferably without any excuses).
 
Great. Now cut together a sequence going from a scene lit by the sun to one lit by sodium lamps and see if you find the transition too jarring. Whether it is or not will be pretty subjective; it's your call.
 
You don't need to light the whole scene for this, just your actor... the scene is lit with sodium vapor. You just need to provide light frequencies to the little piece of space that your subject occupies. Let the rest be orange, the audience expects it... you can even use the sodium vapor as a rim light for your subject. You'll want to gel your key to reign in the orange of the background a little bit - then you'll white balance it bak in post (tungsten, halogen, LED or fluorescent sources will provide enough spectrum for you to play with in your color corrector). The edit bay is shifting the values of the frequencies recorded to other frequencies to "push" them warmer or cooler. The sodium vapor background is basically a black and white picture where the grays are replaced by oranges.

So you only need to light the subject. They are separate from the background... you can even over light them to drop the values of the background darker relative to the subject.
 
this time I am asking if it's okay to mix different skin tones from scene to scene.

Absolutely. Lighting is almost always different between exterior and interior. If they're separate scenes the skin tones don't need to match across the cut because it is a different scene with different light. Matching skin tones is mostly important within the same scene..

For example, if you shoot a bunch of footage outside on a sunny day, and then go back on a really cloudy day to shoot some pickup shots for that same scene, you'll need to do some work either with supplimentary lighting and/or color correction in post to make all of the shots within that scene appear to have occurred at the same time because cutting from a shot on the really sunny day to a shot on the really overcast day would be jarring.

If, on the other hand, they are shots for different days that happen to be the same location, the lighting could be slightly different and that would be ok.

The other scenario, a scene that takes place outside, followed by a scene that takes place inside.. No reason for color temperature to be the same, most people don't light their homes and offices with daylight bulbs. It's usually tungsten, fluorescent, or a mixture of them.

In post you'll want to do some color correction to push the colors around some so that things aren't completely out of whack and weird, but the reality is, lighting definitely differs from one location to the next, and there's no significant reason (usually) to try to overcome that.
 
Okay thanks. Sorry if I have been confusing before. I wasn't trying to turn the sodium lights blue. In that shot I turned the sodium lights white, and the mercury lights are blue because of it. That's what I want for my parking garage sequence, blue, and white lights on the actors. But they go pale because of it. As long as going from golden yellow sunlight balance outdoors, to going to that indoors, is okay. Cause that's kind of like going from the colorful golden yellow skin tones you see on 24, to the desaturated pale skin tones, you see on Saving Private Ryan, which is a big change. I will do more tests, but have to wait for a non-cloudy day too to get that golden yellow.
 
It makes sense if you know that skintones are more or less orange.

You have a DSLR, right?
Take a picture of someone's face (Caucasian!) in RAW.
Open it in Photoshop and play with the orange channel in the RAWeditor: that's the proof you need ;)
 
It might be helpful if you were to post a couple of stillframes from your problematic footage, then some other people could play with them and have much more useful information to provide you besides light control which it sounds like may not be entirely possible for you.
 
Okay thanks. But only stills are in raw, video is in H.264, so playing with orange turns the picture pink. So how will practicing with raw help?

I meant that you can actually see for yourself how much orange is present in human skintones.
It's not for practising grading, it's about getting insight through actual proof.
 
Back
Top