← Back to Diagnostic Tools

On-Device Face Tracking

Key Light Calculator

Not a LUX meter — an uncalibrated webcam cannot honestly produce one. What this measures is the brightness ratio between the left and right sides of your face, captured in the same frame under identical exposure, which is a genuinely reliable signal and exactly what "key light to fill light ratio" actually means. Real-time face tracking runs entirely on your device.

Camera off

Readout

Your Key-to-Fill Lighting Ratio

Click Start and sit facing the camera as you normally would for a call.

Brightness Ratio
  • Brighter side:
  • Verdict:

Share

Share Your Result

Notes

Reading your ratio

GradeRatioLook
Flat< 1.3:1Even, ring-light-style — clean but can look dimensionless.
Balanced1.3:1 – 2.2:1Classic key + fill — the range most guides recommend.
Noticeable2.2:1 – 4:1One-sided but workable; a fill light softens it.
Harsh> 4:1Shadow side likely losing real detail.

Why not just report LUX?

A webcam's auto-exposure and auto-gain are actively working to make a scene look reasonably bright regardless of the actual ambient light level — which means the raw pixel values a web page sees are already the product of an adjustment process the page has no visibility into or control over. That makes an absolute brightness figure fundamentally untrustworthy from this kind of sensor. A ratio between two regions in the same frame sidesteps the problem entirely, because both regions were captured under the exact same exposure settings — the comparison is valid even though neither side's absolute value is.

Step-by-Step

How to Run the Test

  1. Click "Start" and allow camera access when prompted.
  2. Sit in your normal call position under your usual lighting setup.
  3. Read your ratio and verdict, plus which side of your face is brighter.
  4. Adjust one light at a time and watch the ratio update live.

How It Works

How the Ratio Is Measured

The tool auto-detects your face, then samples the relative brightness of the lit side against the shadow side within that same region, in the same frame, under the same exposure. Because both samples come from one frame, there's no need to calibrate the sensor — whatever gain or exposure the camera happens to be using at that instant applies equally to both samples, and cancels out of the ratio.

That's the same reason this approach stays reliable even though the underlying sensor readings are not absolute — a relative comparison within one frame sidesteps the calibration problem entirely.

Platform Guide

Why Your Ratio Looks Different on a Call

Zoom, Teams, and Meet all apply their own auto-exposure and, on some settings, skin-smoothing before you ever see your own preview — both can visibly flatten or exaggerate contrast compared to what this tool measures directly from the raw feed.

For an accurate read that matches what this tool reports, turn off any "touch up my appearance" or beauty-filter setting before testing, and check your ratio here first, then treat the call platform's preview as a rough approximation rather than a precise match.

Setup

Getting a Reliable Key-to-Fill Ratio in Practice

Light intensity falls off with the square of distance — halving the distance between a lamp and your face roughly quadruples its effective brightness there, which makes small position changes surprisingly powerful for adjusting your ratio without buying anything new.

A diffuser — a softbox, umbrella, or even a bedsheet over a bare bulb — lowers the ratio by spreading light across a wider area, producing a flatter, lower-contrast look. A single light with no fill at all will read as a high ratio by design; that's not a fault, it's just a more dramatic, higher-contrast style.

Common Questions

Frequently Asked Questions

Why doesn't this tool report LUX or foot-candles?
An uncalibrated webcam sensor cannot produce a trustworthy absolute light-level reading — its exposure and gain are constantly auto-adjusting in ways a web page cannot see or control. What it CAN measure reliably is the ratio between two regions captured in the exact same frame under identical exposure, which is exactly what key-to-fill lighting ratio actually is.
What is a good lighting ratio for video calls?
Roughly 1.3:1 to 2.2:1 is the classic flattering range — enough shadow to add dimension without looking harsh. Under 1.3:1 tends to look flat, like a ring light straight-on. Beyond about 4:1 the shadow side is likely losing visible detail.
Does this send my face or video anywhere?
No. Face tracking runs entirely on your device using a downloaded AI model — no image or video data is ever transmitted.
How do I fix an uneven lighting ratio?
Add a second, dimmer light source on the darker side — even bounced light off a wall or a lamp works. Alternatively, move your main light further off-axis from the camera so it wraps around your face more, which naturally softens the contrast between the two sides.
Why does it say "no face detected"?
Make sure your face is clearly in frame and reasonably lit — the model needs a basic view of your face to locate it before it can compare brightness across the two sides.