Why do you look strange in photos?
― You're used to seeing your mirror-image self
You meet the you in the mirror every day. Yet look at a photo and you might think, "did I really look like that?" Selfies make it worse. The photo isn't distorted, and the mirror isn't lying to you. There are two causes, and both come down to a simple ratio. One is how often you're used to seeing each version; the other is the camera's distance.
Have you ever looked at a photo a friend took of you and thought, "that's a bad shot"? Yet the same friend usually says, "no, that came out great."
Or the reverse: you show someone a photo you think looks great, and their reaction is lukewarm.
Neither of you has something wrong with your eyes. You and your friend are simply used to seeing different versions of your face.
What you're used to is the face in the mirror; what your friend is used to is your actual face. The two are mirror images of each other.
Yourself in the mirror versus yourself in a photo. Count the occasions, and the gap comes to roughly 70-fold. Whichever you're more used to feels like "the real you."
At selfie distance, your nose photographs about 30% larger than your ears. That's not the camera's fault ― it's determined by distance alone.
A single division makes both clear. Let's go through them in order.
Reason 1 ― A huge gap in how often you're used to seeing each version
No one's face is perfectly symmetrical. Eye height, the tilt of your mouth corners, the shape of your eyebrows ― look closely, and something is always a little different on each side.
So flip left and right, and you get a different-looking face. To the person themselves, the difference is obvious.
The question is which one you're more used to. You see your mirror face many times a day, every day. You see your photographed face far less often. Count it up, and the gap comes to roughly 70-fold (the calculation is in the final collapsible section).
And people are thought to have a tendency: the more you see something, the more natural and likeable it feels. So the mirror face feels like "the real you," and the photo face looks "off."
What's interesting is that for your friend, this is exactly reversed. What your friend is used to seeing is a face oriented the same way as in the photo. So the photo you think looks bad, your friend finds perfectly natural. Neither of you is wrong ― you're just used to different things.
It's about which one you've seen more.
What a mirror actually flips is not left-right but front-back. The you in the mirror isn't a different person facing the same way as you ― it's you, turned inside out.
It feels like a "left-right flip" because, it's thought, when we look at a person, we habitually orient by left and right rather than up and down.
We covered this in detail in our article on mirrors. Written as an equation, it's a one-line story.
Reason 2 ― Shooting up close actually changes the shape of your face
There's a second cause, and this time it's really in the photo. This isn't a trick of perception ― the shape genuinely changes.
A face has depth. The nose is close to the camera; the ears are farther away. Things closer to the camera photograph larger, so the nose photographs bigger than the ears.
How much bigger comes down to one division. Taking the depth of a face as 10 cm:
Nose: 30, ear: 40. Ratio is 40 ÷ 30 ≒ 1.33. The nose photographs 30% bigger.
Nose: 200, ear: 210. Ratio is 210 ÷ 200 ＝ 1.05. Almost no difference.
Feeling that your nose looks "big" or your face looks "round" in a selfie is not just in your head. That's really how it photographs.
Here's one important point: this isn't the lens's fault. People say "wide-angle lenses distort faces," but that's not quite right. Distance alone does the distorting.
Trying to fill the frame with your face using a wide-angle lens inevitably means getting closer. So the distortion follows as a result. The lens looks like the culprit only because it's what pulls you closer. Shot from the same distance, any lens renders the shape of a face the same way.
Common selfie advice turns out to have a real reason behind it.
- Fully extend your arm: just going from 30 cm to 60 cm drops the ratio from 1.33 to 1.17
- Use a selfie stick: get further still, and it lands around 1.10
- Have someone else shoot it: at 2 m, it's 1.05 ― barely any distortion
- Crop the face smaller after shooting: shooting from farther away and enlarging later looks more natural
The reason ID-photo machines shoot from a slightly greater distance is the same. You could say "a flattering shot" is really "a shot with a ratio close to 1."
Something you can check for yourself
- Pick a photo of yourself and flip it left-right using your phone's editing tool
- Look at the original and the flipped version side by side
- The flipped one should feel like "the you you're used to." That's because it matches your mirror orientation
- Next, ask a family member or friend which one looks more natural to them
- They'll usually pick the original. That's the one they're used to seeing
Steps 4 and 5 are the heart of this observation. Once you confirm that you and the other person give different answers, you'll see that the question "which one is real" doesn't really make sense. One more thing: try photographing the same person from 30 cm and from 2 m (scale the face size to match afterward). The difference is so large it barely looks like the same person.
Summary
Your photo self looks odd because you're used to a mirror-flipped face, and because shooting up close makes your nose look bigger. The first comes down to a ratio of viewing counts; the second, to a ratio of distances. Both are a single division.
The photo isn't distorted.
You're just seeing yourself from a side you're not used to.
The same thing happens with voices. Your own voice, too, comes out on a recording differently from the "version" only you normally hear. For details, see why your recorded voice sounds like someone else.
For those who want to know more ― terms, numbers, and links to textbooksFrom middle-school science to topics still under research, each level is clearly labeled
- Middle schoolCovered in middle-school science and math
- High schoolCovered in high-school "Math I/A" and "Basic Physics"
- High school+Covered in high-school "Physics," or treated as advanced/sidebar material in textbooks
- UniversityNot covered in high school ― university-level specialist subjects (cognitive science, optics)
- ResearchNot even settled as "established theory" at university ― topics researchers are actively investigating
Middle schoolTerms: vocabulary around how we see things
- Mirror image: The image reflected in a mirror. A figure flipped front-to-back.
- Perspective: Things nearby look bigger, things far away look smaller. Photos follow this same rule.
- Focal length: A length that describes a lens's properties. It determines how wide a range is captured.
- Angle of view: The width of the captured range. The shorter the focal length, the wider the view.
- Mere-exposure effect: The tendency to find something more likeable the more often you've seen it.
High schoolCheck it with equations: both reasons come down to a ratio
What makes this article interesting is that two completely different reasons are both settled by a single "ratio." Let's go through them in order.
ratio ＝ times seen in mirror ÷ times seen in photos
| Times you check the mirror | assume 10 times/day |
| Times you see yourself in photos | assume 1 time/week |
| Period | assume 20 years |
| Mirror views over 20 years | 10 × 365 × 20 ＝ 73000 times |
| Photo views over 20 years | 52 × 20 ＝ 1040 times |
| Ratio | 73000 ÷ 1040 ≒ 70-fold |
You see your mirror face about 70 times more often than your photo face. That's what decides which one feels like "your real face."
Change the assumptions and the conclusion still holds. Even with the mirror at 3 times a day and photos at once a day, it's still 3-fold. For the ratio to approach 1, you'd need to see photos of yourself about as often as the mirror. In ordinary life, that doesn't happen.
※ These counts are illustrative estimates. But however you set them, the direction ― "the mirror wins" ― doesn't change.
distortion ＝ (distance ＋ face depth) ÷ distance
| Distance | length from camera to nose [cm] |
| Face depth | front-to-back length from nose to ear. Assume about 10 cm |
| Distortion | closer to 1 means more natural; larger means the nose photographs bigger |
| 30 cm (selfie, elbow bent) | 40 ÷ 30 ≒ 1.33 |
| 60 cm (arm fully extended) | 70 ÷ 60 ≒ 1.17 |
| 100 cm (selfie stick) | 110 ÷ 100 ＝ 1.10 |
| 200 cm (someone else shoots it) | 210 ÷ 200 ＝ 1.05 |
30% at 30 cm, 5% at 2 m. The gap is over sixfold. Just extending your arm drops the ratio from 1.33 to 1.17, so even that alone helps.
This equation has no lens in it. No focal length, no camera type, no price. Only distance and face depth appear. So the claim "the lens distorts things" turns out not to be quite accurate.
People say "85mm or longer is best for portraits." If the equation in ② has no lens in it, is that just a myth? No. The lens determines the distance.
To photograph the face at the same size, doubling the focal length requires doubling the distance too. They're proportional.
| Distance to fill the frame with a 24 mm lens | assume about 0.5 m |
| How many times 85 mm is 24 mm | 85 ÷ 24 ≒ 3.5x |
| Required distance | 0.5 × 3.5 ≒ 1.8 m |
Compare this with the table in ②. At 0.5 m the ratio is 1.20; at 1.8 m it's about 1.06. A telephoto lens looks natural not because the lens fixes the shape, but because getting the same framing forces you to back away.
The lens isn't the cause ― it acts through distance. When the equation has no lens in it, yet changing the lens changes the result, that's a sign to look for what's sitting in between.
If flipping left-right changes the impression, that means your face differs left to right. You can measure how much from a single photo.
- Prepare a photo of yourself shot from the front
- Draw a center line through the midpoint of both eyes
- Measure the distance from that line to the left mouth corner and the distance to the right mouth corner
- Divide the longer by the shorter
For example, 31 mm and 29 mm gives 31 ÷ 29 ≒ 1.07. Just 7%. Yet swap left and right, and it reads as "a different face."
This also shows just how fine-grained our ability to recognize faces is. We routinely pick up on a 7% difference.
High school+Photos are built from "central projection"
The image a camera forms is a solid object projected onto a plane along straight lines passing through a single point (the center of the lens). Under this projection, length ratios in the original solid are not preserved as-is. The distortion worked out in ② comes from exactly this.
By contrast, shooting from a very great distance with a telephoto lens makes the light rays nearly parallel, approaching a projection that does preserve ratios. This is why certain drafting projections, and photos taken from far away, look "flattened."
In other words, a "photo that looks natural" is, in terms of how it's projected, actually a shift away from reality. The naked eye also uses central projection when viewing a face, but we're not in the habit of bringing our face within 30 cm of someone's when looking at them. That's why we're not used to how things look at that distance.
UniversityHow far does "familiarity breeds fondness" really hold?
The tendency to like things you've encountered repeatedly has long been studied in psychology. For faces specifically, reports from the 1970s found that people prefer their own mirror image, while friends prefer the photo orientation ― this is the basis for the explanation in this article.
Some caution is needed, though. The strength of this effect varies by condition, and it doesn't always produce the same result. For stimuli that are disliked to begin with, repeated exposure has been reported not to increase liking, and sometimes to decrease it.
It has also been shown that the brain processes recognizing your own face somewhat differently from recognizing other people's faces. "Your own face" may receive special treatment, not just be a face you happen to see a lot.
ResearchWhat's still unresolved
- The finding that "people prefer their own mirror image" doesn't always replicate. The original study is well known, but follow-up attempts have reported mixed results depending on conditions. Reproducibility itself is a major issue across psychology as a field, and this topic is no exception.
- The mechanism by which we recognize our own face as "ours" is still being worked out. Is it matching facial features, or does it also use motion and context? Reports don't fully agree on which brain regions are involved, or how.
- Judgments of "attractiveness" vary by culture and era. Some studies report a preference for symmetrical faces, but the size of the effect varies across studies, and generalizing calls for caution.
- Facial recognition technology doesn't necessarily use the same cues people do. Machines can identify people with high accuracy, but explaining exactly what they're basing that on is difficult. Machines sometimes excel at conditions humans struggle with, and vice versa.
Sections ② and ③ of this article are geometry, so they're solid. Point ① ― "familiarity breeds fondness" ― is supported in direction, but how strong the effect is depends on conditions. Even within a single article, the degree of certainty varies.
Links to textbooks (by level)
| Level | Subject/unit | Where in this article |
|---|---|---|
| Middle school | Math: ratio and proportion / Science: reflection of light | Distance ratio, mirror images |
| High school | Math I: figures and similarity | Calculating distortion from the distance ratio |
| High school | Basic Physics: lenses and images | Proportionality of focal length and distance |
| High school+ | Physics: geometric optics / drafting projections | Central projection vs. parallel projection |
| University | Cognitive psychology: face perception | Mere-exposure effect, self-face processing |
| Research | Psychology / computer vision (unresolved) | Reproducibility, self-face recognition, machine decision basis |
| ― | Everyday life | A flattering shot is one with a ratio close to 1 |
- Mita, T. H., Dermer, M. & Knight, J., Reversed facial images and the mere-exposure hypothesis, J. Pers. Soc. Psychol. 35, 1977.
- Zajonc, R. B., Attitudinal effects of mere exposure, J. Pers. Soc. Psychol. 9, 1968 (the original mere-exposure effect paper).
- Open Science Collaboration, Estimating the reproducibility of psychological science, Science 349, 2015 (on the reproducibility problem).
- Hecht, E., Optics (a standard textbook on geometric optics and projection).
- General reference material on photographic technique (focal length and shooting distance in portraiture).
※ Face depth, mirror-viewing frequency, and shooting distance are all illustrative representative values. They vary by person and situation.
※This article is a general-audience science explainer. Statements about psychological effects describe reported tendencies and vary by individual. The figures given are approximations meant to illustrate the underlying mechanism.