Wed Apr 16, 2025 2:41 pm
I shall look at Jamie’s posted video when I can, but I have some thoughts on what is a best exposure that might be helpful. Hint: don’t worry, be happy.
There is a false colour monitor option that allows you to judge what might be a best exposure in theory for skin. The normal guidance is the important part of skin should be pink for light skin and green for dark skin. Unless you’re shooting for the Sports Illustrated Swimsuit Issue, skin usually refers to a face. But a face is not a flat piece of 18% grey cardboard (which is helpful in determining pink or green exposure in false colour).
As you know a face is quite complex in its components that includes varying contours, colorations, textures, highlights and shadows, and so on. It’s important for you to decide what is important to guide your exposure. It may be the eyes, it may be a cheek, it may be a forehead. Depends on your light, angles, and your intentions of the shot.
Okay so you think you’ve nailed the exposure for your face, but your actor is wearing stark white clothing or is bald or has light grey hair, and your lighting is trying to simulate some stronger direction of a sun such as noon or late afternoon. You’ll find your perfectly exposed face might look like a disaster in post. Too late, you realize you should have concentrated more on how the various light levels could be enhanced with bounce or muted with scrims that respects the exposure limits of your camera and the dynamic range in your scene.
Your camera actually has a better tool than false colour to avoid this disaster. First let’s talk about clipping. With a false colour display, clipping occurs where the false colour is red, right? Wrong! The red colour certainly MAY indicate where you have clipping of the sensor, but it MAY NOT be telling you the truth. As of firmware release 7.3, you have three little boxes that are usually empty, but when a colour channel is clipped, you’ll see red and/or blue and/or green fill in the boxes. If any of those boxes are coloured, you’ve lost accurate colour for that channel; if all three are coloured, the sensor is clipped and we don’t usually want that unless you know you’re scene has specular highlights that might be best left to clip such as bright chrome on a car or the tops of a sunlit ocean wave. But if you see the box filled on the far cheek of your actor, you likely don’t want that.
Don’t worry, be happy because in post you can adjust for most lighting errors as long as you didn’t clip the sensor. I regularly underexpose faces, so light faces are green or even a little darker because I always have to deal with abruptly changing light levels and poor costume choices that are out of my control.
In photography, your task is easier in a way because you have one frame for one tiny fraction of a second to nail the best exposure. In video, it’s a nightmare because everything is in motion in a shot that can be several minutes with multiple bodies altering their distances from your lighting and the focus of your shot (the part of your frame you want the viewer to focus on) might even change that changes the skin in the shot.
As long as your exposure is on the low side of a proper exposure, you can easily and confidently bring it back in DaVinci Resolve; it’s in post where you determine the perfect exposure of that face, not in camera.
Better safe than sorry in camera. You may be boosting exposure but I recommend the opposite in moderation. Certainly don’t feed the sensor more light than it needs to get a perfect exposure in post.
Now what to do when you can’t use your 18% grey card or colour chart? I don’t use charts. I’ve found on the BMPCC4K and UM4.6K, if the face looks nice to your eye (not bright but rich in depth) while you can see detail throughout the frame where you need detail, then you have a perfect exposure. I don’t worry if things are pretty dark in the shadows; don’t boost your camera exposure to see in the dark. Embrace the dark, shadows are what gives dimension and depth to the light. In post you might boost shadows a little and be happy with that.
If you have a grading monitor in your edit suite:
1) Exposure in post for HDR ST2084 P3 colour:
- be sure Resolve Project Settings are using HDR settings for 4KTV or for any modern phone (iPhone 13) or tablet using P3
- begin with tuning the Camera raw setting if appropriate to alter temperature and tint.
- use Nodes to separately manage the adjustment to your image or effects beginning with a node for primary colour
- use the six colour wheels for primary adjustment of exposure (don’t use the three Lift, Gamma, Gain wheels or I shoot you while you are sleeping).
- on the Colour page, mostly use the Waveform panel set to HDR with occasional Histogram or other panel
-diffuse white levels used to always be set close to 100nits but more recent 4KTV can use 200 or 300nit levels for diffuse white (your highlights exceed this up to say 1000nits depending upon the settings you choose
- generate your HDR deliverables
2) Exposure in post for SDR Rec.709 colour
- bypass your P3 node by adding a new node for 709
- set Colour page monitor to use SDR 10-bit colour
- now you can grade for HDTV or web viewing using the three colour wheels, Lift, Gamma, Gain
- your diffuse white levels can be higher than for 4KTV, up to 300 or 500 nits if your monitor supports that
- generate SDR 709 HDTV deliverables
3) Exposure in post for SDR sRGB which are for web viewing on older monitors
- like Rec.709 but now sRGB and almost no one does this as 709 seems to be the default everywhere you’ll just your video
- again higher diffuse white levels
- generate sRGB deliverables.
So that’s the very brief guidance that you can use to determine perfect exposure. Taking the shot is only the first step.
Rick Lang
