Skip to content
GPU-assistedUpload required

What Reads as Professional Is Mostly Where the Camera Was

Hold a phone at arm's length and the lens sits about forty centimeters from your face. Your nose is four centimeters closer to it than your cheeks, which on that scale is a tenth of the whole distance, so it is rendered noticeably larger while your ears fall away and shrink. A portrait photographer standing two meters back with a longer lens has the same four centimeters spread across a fiftieth of the distance, and the face comes out with its real proportions. Nothing about skin, lighting, or wardrobe accounts for that difference. It is arithmetic about where the camera stood.

Use this without the search next time. Prathom Workbench puts Prathom's tools in your toolbar.

Add to Chrome — free

Drop an image here, or click to browse

Up to 25 MB. Uploaded temporarily, deleted after processing.

GPU-assisted — results vary slightly between runs.

What it does

  • Wide-angle face distortion corrected
  • Camera raised to eye level
  • Eyes placed on the upper third
  • Light sharpening pass after resampling

How to use AI Professional Headshot

  1. 1

    Note how the photo was taken

    An arm's-length selfie or a photo shot from chest height has geometry to fix. A photo already taken from a distance at eye level has little for this page to do.

  2. 2

    Run the correction

    The frame is rebuilt as though shot at eye level with a short telephoto, with the enlarged-nose and narrowed-ear signature of a close wide lens named as something to undo.

  3. 3

    Check the proportions against a mirror, not the source

    The original is the distorted one. Compare the result to how your face looks to you, since the point is to move away from the source, not towards it.

How it works

The upload is scaled to about one and a fifth megapixels and repainted at a moderate denoise. The instruction is unusually specific about optics: rebuild the frame as if it were shot at eye level with a short telephoto, undo the wide-angle enlargement of the nose and narrowing of the ears, and put the eyes on the upper third.

The negative prompt lists the same failures as physical descriptions — wide-angle distortion, enlarged nose, a low or high camera angle, a tilted horizon — so both prompts describe the same geometry from opposite ends.

After decoding, a light sharpening pass runs with a small radius, which recovers the edge contrast that the resample costs.

Distortion is a property of distance, not focal length

The common way to say this is that wide lenses distort faces, and it is close enough to be useful and wrong enough to be worth correcting.

A wide lens does not bend anything. What it does is let you stand close and still fit a head in the frame, and standing close is what produces the effect. Shoot from two meters with a wide lens and crop in hard, and the proportions are correct — the picture is just noisy.

This matters because it tells you what fixes the problem at source: step back and crop, rather than hunting for a different lens. It also explains why the phone camera is the culprit so consistently. Arm's length is not a choice about photography, it is the length of your arm.

The one case where changing your face is the honest option

Every other page in this family warns you off edits that move your features, because a headshot that is not recognizably you has failed at the only thing it is for.

This page asks for a change to your features on purpose, and the distinction is worth being precise about. The nose in an arm's-length selfie is not your nose; it is your nose plus a projection error of maybe fifteen percent. Removing that error is not flattery. It is the same class of correction as straightening a tilted horizon.

The line to hold is that the correction should stop when the geometry is right. A model asked for a professional portrait will happily continue past that point into slimming and smoothing, which is why the negative prompt names altered identity and plastic skin alongside the optical faults. Compare the ratio between the width of your nose and the width of your mouth in both images. Optical correction changes that ratio a little. Beautification keeps going.

When the source has nothing to correct

If your photograph was taken from a couple of meters away with the camera at your eye line, this page has almost no work to do, and running it anyway is a bad trade.

The pass still repaints the frame. You accept the full generative risk — drift in the face, invented fabric, a changed hairline — in exchange for correcting an error that was not there. In that situation the generator page, which is honest about replacing clothing and background, or the studio page, which puts you on a consistent backdrop, does something you actually need.

Look at the photograph before choosing the tool. If the nose looks large and you can see up the nostrils, you are in the right place.

Publication gate

This page ships once the workflow has been run against an arm's-length selfie shot below the chin, a photo taken from three meters at eye level, a wide-angle group crop, and a profile in glasses, with the nose-to-mouth width ratio measured against the source in every case to confirm that the correction stops at geometry.

Examples

Arm's-length phone selfie

selfie.jpg - 1200x1600, taken at about 40cm, camera slightly below the chin
prathom-ai-professional-headshot.png - 980x1307

The best case for this page. There is real geometric error to remove, and removing it makes the face look more like the person, not less.

Photo already shot from across a room

event.jpg - 2400x3200, shot on a 70mm lens from three meters
event-corrected.png - 1000x1333

Little to correct. The pass still repaints the frame, so you take on generation risk without a distortion problem to solve.

Frequently asked questions

Why does correcting distortion change the shape of my face?

Because the shape in the photograph was already changed, by the lens. A close wide lens enlarges whatever is nearest it and shrinks whatever is furthest, which for a face means a bigger nose and smaller ears than you have. Undoing that moves the picture towards your proportions, which is the one edit in this category that makes a photo more accurate rather than less.

Does camera height matter as much as distance?

It changes different things. Distance sets proportion; height sets what the viewer reads into it. A camera below the chin shows the underside of the jaw and the nostrils and reads as looming, and one held above reads as diminished. Eye level is neutral, which is why nearly every portrait convention lands there.

What is the sharpening step for?

Resampling an image softens it slightly, and a small sharpening pass at the end restores the edge contrast that was lost. It restores apparent crispness; it does not recover detail. If the source was out of focus, nothing here brings the focus back.

Can this rescue a genuinely bad photo?

No. It fixes geometry, and geometry is only one of the ways a photo fails. Motion blur, a face in deep shadow, or a subject too small in the frame are all missing information rather than misplaced information, and generation fills missing information with invention.

Is the AI Professional Headshot tool free, and do I need an account?

It is free and there is nothing to register for. No account is created, no address is collected, and no reduced-quality preview is withheld for a plan that does not exist. The constraints that do apply come from the workflow itself rather than from billing: 25 MB in, a shared GPU queue, and a thirty-minute window to collect the output.