Why are moonlight lit interiors pretty well exposed?

[silvan schnelli]

I have been thinking about the differences of how I would handle lighting a moonlight-lit interior vs a sunlight-lit interior. As a result, I have also looked at a lot of references from movies and have noticed that the exposure levels on faces for night scenes tend to sometimes be just as bright as daylight scenes or perhaps just slightly less. 

I have also come across this reply from Sir Roger Deakins ASC BSC on lighting "Moon light for exteriors", and I think he has said something similar on the Deakins Podcast (https://www.rogerdeakins.com/forums/topic/moon-light-for-exterior/)

"There are photometric charts that you can study to find what light will give you the required foot candles and width of beam at any given distance. I would start by looking at the chart for an 18K HMI. To shoot at 2.8 and 800 ISO you need something like 12.5 foot candles depending on whether that is a back light or a front light."

In conclusion, it seems to me that from what I've seen, is that Cinematogaphers often still try to correctly expose for the key light, but achieve the nighttime or dark-feeling by having much stronger contrast ratios (such as through a weaker fill light) and that the scenes tend to be cooler or having a sort of blue-tint to them. Or do Cinematographers also slightly underexpose the key lights on the actors' faces for more "realism"? Attached below are examples from the movies Ad Astra, shot by Hoyte van Hoytema ASC, NSN, FSF and The Assassination of Jesse James West, shot by Roger Deakins, with contrasting day and night scenes that seem to have key lights of similar brightness.

[Doyle Smith]

I'm a film guy so I don't know how this comment translates to digital cameras, but here goes. The photometric charts the imminent Mr. Deakins references, would be useful if you are working up an equipment list for a rental house but in the real world you just lobby for biggest lights your budget permits and make do with what you end up with. 😉

The frames on the left hand side of your sample look like they have key lights of similar brightness because in the context of the sets, time of day, period sources, they are both low key situations.

I've done this type of shot many times. On a 3200 white balance interior, my moonlight would have been an 18K HMI's or Musco especially on exteriors. On HMI's I'd usually go a 1/2 CTO on 4x4 frames for what I term "half blue" ambiance. For an interior like in your examples, I would probably just cheat it with a smaller HMI fresnel (enough to cover the window opening) and far enough back that I don't get into inverse square problems as the actors move around the set.

Key light and key exposure are tricky terms and mean different things to different people. The key exposure on the face of the bearded gentleman with the vest (is this Jesse James?) would be less than the T-stop set on my lens assuming a Caucasian actor. Testing or experience would help you to determine how much latitude you could get away with. So let's say I set my lens for T 2.8. I'd want to see a 1.4 more or less on my incident lightmeter pointed to the camera. I would probably be working in foot-candles at this point but to keep it simple, I going to stay with T-stops for this example.

I could achieve this with a soft non directional fill. So my fill light could be considered the "key light" as it is frontally lighting the primary actor. That's the trap of thinking of key, fill, kicker. The day and night scene side light on our actor looks like it's losing a detail so that says to me it's over key exposure. Maybe a 4.0 more or less depending again on the stock or latitude of the imaging device.

I used to mentally approach lighting a scene with sort of a visually logical "hillbilly" zone system. I look at the time of day in the script, the colors of wardrobe and set, the practicals and windows and doors on the set, the mood or context of the scene, and then I think... what do I want the audience to see?

Is there a story point (exit door or say a firearm) that needs to be established? Maybe there's a something that shouldn't be seen at this point. Do we need to recognize the character immediately or do they move from shadow to recognition? Then I base the lighting relative to where these elements fall in my "hillbilly" zones.

This is hard for me to explain so maybe David M. can do better.

 

[David Mullen ASC]

He’s saying you need 12.5 fc to get a f/2.8 at 800 ISO — that’s not telling what he plans to do with that information, he might decide then to light to 6 fc at f/2.8 so that the face would be 1-stop underexposed.

How much you want to underexpose your moonlight depends on so many factors  — if it’s a long scene and moonlight is the only source, you can imagine that the characters have gotten used to the low level so they aren’t fighting to see anything so maybe it’s not too underexposed — whereas if it’s a scene with a flashlight or other source then the moonlight perhaps should look darker in comparison. Or it’s a brief moment in moonlight after someone switches off the lights and then the scene ends so it can be quite underexposed.

[silvan schnelli]

@David Mullen ASC Thank you that does make sense. The example with the flashlight source vs a long moonlight scene is something I have been thinking about a lot. Especially how a hard cut from a daytime scene to a long nighttime scene would render. Seeing as you would most likely want a scene that looks darker and have a night ambience, but where the audience is still able to see a lot, especially when it comes to the actors. 
 

I think this is where I have seen a lot of shots have bright key lights, but that are higher contrast and cooler.