Anyone who has organised a family or team photo knows that somebody always blinks. People blink around 10 to 20 times a minute and each blink closes the eyes for roughly a quarter of a second, so at any instant each person has a small chance – about 4% – of being caught mid-blink. With one person that is easy to ignore, but the chances multiply: in a group of 20, the odds that everybody has open eyes in a single frame drop to around 40% in good light.
Australian physicists at CSIRO turned this into a simple model in 2006 (it later earned an Ig Nobel Prize). If x is the blink rate and t the time the eyes are closed, the chance a single person is fine is 1 − x·t, and the chance of a blink-free group photo is that number raised to the number of people. In poor light, people blink more and the shutter stays open longer, so the odds get worse.
The calculator gives the probability for one shot, the number of shots needed for your chosen confidence (99% by default), the expected number of shots, the CSIRO rule of thumb and a chart for groups of all sizes.
How to use the Blink-Free Photos Calculator
- Number of people in the photo
- Light
- Blink rate (blinks per minute)
- Blink duration (eyes closed)
- Exposure time (0 to ignore) — A slow shutter adds to the window in which a blink spoils the shot.
- Confidence you want
Everything is calculated in your browser, so results appear instantly and nothing you type is sent to a server.
How it works
The per-person chance of open eyes is one minus the blink rate (per second) times the closed-eye time plus the exposure time. Raising it to the number of people gives the chance of a blink-free frame; the number of shots for a target confidence follows from the probability that all of them fail.
Worked example
Five people in good light: 10 blinks per minute is 1/6 blink per second; with 0.25 s per blink the chance one person is blinking is 1/6 × 0.25 = 4.17%, so q = 0.9583.
A blink-free frame has p = 0.9583⁵ = 80.8%. For 99% confidence you need k = ln(0.01) ÷ ln(0.192) = 2.79 → 3 shots. For 20 people in bad light (15 blinks per minute), p = 0.9375²⁰ = 27.5% and you need 15 shots.
Shots needed for 99% confidence
| People | Good light | Bad light |
|---|---|---|
| 5 | 3 | 4 |
| 10 | 5 | 7 |
| 20 | 9 | 15 |
| 30 | 15 | 30 |
Formula
Chance one person is not blinking: q = 1 − x · (t_blink + t_exposure) · Blink-free shot: p = qⁿ · Shots for confidence C: k = ⌈ln(1 − C) ÷ ln(1 − p)⌉ · Expected shots until success = 1 ÷ p
Worked example
Inputs — Number of people in the photo: 5; Light: Good light (about 10 blinks per minute); Blink duration (eyes closed): 0.25 s; Exposure time (0 to ignore): 0 s; Confidence you want: 99 %.
Results — Chance of a blink-free photo = 80.83 %; Photos to take for your confidence = 3; Average number of photos until a good one = 1.24; Chance a single person has open eyes = 95.83 %; CSIRO rule of thumb (people ÷ 3, or ÷ 2 in bad light) = 2; Chance of at least one good shot = Photos taken Chance of at least one blink-free shot
─────────────────────────────────────────────────────
1 photo 80.83%
2 photos 96.33%
3 photos 99.3%
5 photos 99.97%
10 photos 100%
20 photos 100%.
Key terms
- Blink rate
- Number of blinks per minute – typically 10–20 at rest.
- Blink duration
- Time the eyelid is closed, about 0.1–0.4 s.
- Independent events
- Events whose probabilities multiply because one does not affect the other.
Frequently asked questions
How many photos should I take of 10 people?
In good light the chance of a blink-free shot is about 65%, so take 5 shots to be 99% sure of one good one.
Why is bad light worse?
People blink more in dim or harsh light, and slower shutter speeds make each blink more likely to overlap the exposure.
What is the CSIRO rule of thumb?
For groups under about 30, divide the number of people by 3 in good light (by 2 in bad light) to get the number of shots.
How can I reduce blinking?
Ask everyone to close their eyes and open them on "three", use a fast shutter and burst mode, and avoid looking into a bright sun or flash.
Does flash cause blinking?
Pre-flashes can trigger a blink just before the real exposure; turning off red-eye reduction helps.
Is the model exact?
No – it assumes independent random blinks. People often blink together when startled, which can make real results better or worse.