Exposing skin - spot vs incident

Tyler Clark · June 30, 2015

Is it safer to rely on spot or incident readings for skin?

I had someone sit in for a camera test recently and metered the key as an incident and found that the face looked over exposed. I spot metered her face from then on and found it was about 1 stop over the incident and went off of that.

Was wondering if it's a preference or is there a specific standard way to handle that.

Michael LaVoie · June 30, 2015

Skin tones can appear to either reflect or absorb light depending on the person. Someone with an extremely fair ivory skin tone may read way over someone who is extremely dark. Your incident reading is going to completely ignore this skin factor.

No idea what your situation is but for simplicity sake, in a sit down interview you could fix this a number of ways. You could use a single or double net on a C-stand at a distance from the unit and carefully lower the quantity of light on the face only so that it appears to read the same as the overall incident reading you're exposing for.

There's other ways but that would be the easiest if you had a normal gear package. Usually you wind up doing this for someone who is wearing a white shirt. The shirt may be blowing out on you so you may use a single net to take it down a bit being careful not to change the level on the face. It gets harder if the subject is moving around a lot on you. You may just want to split the difference on your exposure in camera at that point. But I'd always prioritize skin tones over anything else. Those should look correct.

Edited June 30, 2015 by Michael LaVoie

Dan Muchnik · June 30, 2015

Your incident reading won't lie. Typically, brown skin reads a stop under, and caucasian skin a stop over when taking a spot reading relative to your incident reading. Pacific islanders' skin is often dead on with the incident reading, hence labs in the olden days compared the exposure of film to a "China Girl", even using a densiometer. https://en.wikipedia.org/wiki/China_Girl_%28filmmaking%29.

Are you exposing with film? With digital, false color is probably your most valuable exposure tool.

John E Clark · July 1, 2015

Is it safer to rely on spot or incident readings for skin?

I had someone sit in for a camera test recently and metered the key as an incident and found that the face looked over exposed. I spot metered her face from then on and found it was about 1 stop over the incident and went off of that.

Was wondering if it's a preference or is there a specific standard way to handle that.

The general rule is Caucasian skin is about a stop over an 18% grey card, Mediterranean skin is about 18% and African skin is 1 stop under. These are generalities, and so could vary considerably... like 'deep tanned' Caucasians are more Mediterranean...

There is also an issue with calibrating your meter, incident or reflective to your camera's ISO values... Camera manufacturers for digital cameras are not limited to certain specific methods of determining ISO values for their cameras, and one 'method' is essentially 'what looks good to the engineers'...

I did a calibration process for my Blackmagic Pocket Cinema Camera for the BMDFilm mode, a 'log' representation, and used 38% IRE for an 18% grey card, to calibrate my ISO/Camera/meter settings.

I'd recommend doing that for your camera and meter.

In order to understand what the meter reading means... for incident... 'if there is an element that has 18% reflectivity... it will be represented with an 38% IRE level (with my criteria for calibration...)'. But there may not be anything in the scene that happens to be 18% reflective... think of Black velvet with only highlights outlining the shape of the cloth... which is an old exercise to learn how to expose for 'dark' while still getting texture and form.

With a spot meter then one only reads values in terms of 18% -> 38% IRE levels, and if something is known to be more reflective or less reflective, an adjustment needs to be made, such as the plus one, none, or minus one stop for skin tone.

Edited July 1, 2015 by John E Clark

David Mullen ASC · July 1, 2015

Is it safer to rely on spot or incident readings for skin?

I had someone sit in for a camera test recently and metered the key as an incident and found that the face looked over exposed. I spot metered her face from then on and found it was about 1 stop over the incident and went off of that.

Was wondering if it's a preference or is there a specific standard way to handle that.

You need to tell us more information, like was this a live video feed from a digital camera to a monitor? Was this a work print or video dailies of something shot on film?

If this were a live feed, then it also depends on how your camera likes to set-up the signal for Rec.709 output to a monitor.

But is also just may be that this person was on the pale side...

Tyler Clark · July 1, 2015

Thanks a lot for all the answers guys! I think this helps answer a lot of questions I've been having. I was doing the test actually just on the onboard monitor of a Canon 6D which I believe operates in REC709 if I'm not mistaken. I took the incident and exposed for that (f4 I believe). Then spotted on my talent, an ivory skinned, blonde, blue-eyed, southern bell if you will, which read a stop over.

I guess my question now is, should I make it a habit to always expose for accurately spot-read skin tones? Or is this a preference matter per the chosen camera?

David Mullen ASC · July 2, 2015

If the video camera has limited latitude, then often for light-toned faces you want to expose them a little down to hold detail in the hotter spots on the skin. Otherwise an incident meter would normally be fine but the other issue is whether the ASA / shutter speed setting on an Canon 6D is necessarily accurate, which is why you test. Your scene profile might also be boosting highlights to make them more poppy.