How to Edit a Headshot: The Full Workflow, In the Order That Actually Matters

To edit a headshot, work in this order: choose the strongest frame, crop and straighten, correct exposure and white balance, set contrast without losing skin detail, neutralize color casts, remove only temporary blemishes, adjust eyes and teeth lightly, then sharpen and export at the right size for each destination. Order matters because every step changes the next.

This guide is tool-agnostic first. Every step below can be done in a free phone editor or in professional desktop software; where the tools genuinely differ, we say so. It also covers the two ways headshot edits go wrong, and an honest look at AI editing features, including the line between improving a photograph and manufacturing a new face.

The workflow at a glance

Most bad headshot edits come from doing things out of order: smoothing skin before fixing white balance, sharpening before resizing, or brightening eyes before the overall exposure is right. Global corrections come first, local corrections second, output steps last. Here is the sequence the rest of this article walks through.

1. 1Cull: pick the frame worth editing.
2. 2Crop and straighten.
3. 3Exposure and white balance.
4. 4Contrast and tonal range, watching shadows and highlights on skin.
5. 5Color casts from mixed light.
6. 6Skin: temporary blemishes only.
7. 7Eyes and teeth, with restraint.
8. 8Sharpening, resizing and export per destination.

Step 1: Culling is the most consequential edit

Choosing which frame to edit does more for the final result than any slider. A relaxed expression, open eyes, a clean hairline and a natural head angle cannot be added later, while exposure and color can. People skip this step because it feels like admin, then spend an hour trying to rescue a frame that was never right.

Look at your candidates at full size, not as thumbnails. A thumbnail hides soft focus, half-blinks and the tension around the mouth that makes a smile look held. Then look again at the size the photo will actually be seen: a LinkedIn avatar in a comment thread is tiny, and the frame with the clearest expression at small size often beats the technically sharper one.

• Focus on the near eye. If the eye closest to the camera is soft, discard the frame. Sharpening will not fix missed focus convincingly.
• Expression over polish. A slightly underexposed frame with a genuine expression is a better starting point than a perfect exposure with a stiff face.
• Check the edges. Stray hair across the eye, a collar folded under, a lapel mic cable. Note what is fixable and what is not.
• Pick one, then a backup. Editing five near-identical frames multiplies decisions without improving the result.

If you are still torn between two near-identical frames, ask someone who knows you which one looks more like you on a normal day. People who see your face regularly judge likeness better than you do, because you are used to your mirror image rather than the camera's view.

Step 2: Crop and straighten

Crop early because every later judgment, from exposure to how much background shows, depends on what stays in the frame. A standard professional headshot sits roughly from mid-chest to just above the head, with the eyes around the upper third. Straighten using vertical lines in the background, not the tilt of your head.

A natural head tilt is fine and often flattering; a tilted door frame behind you is not. Most editors have an auto-straighten or a level tool that snaps to a line you draw. Draw it along a wall edge or window frame. If there is no reliable line, leave the rotation alone rather than guessing.

• Crop a square master and a vertical master. Most profile platforms want square; websites, speaker bios and press kits often want 4:5 or 2:3. Cropping both from the original avoids re-cropping a crop.
• Leave margin for circles. LinkedIn and most chat apps display the photo in a circle, which cuts the corners. Keep the top of the head clear of the edge.
• Do not crop through joints. A frame edge right at the chin or elbow looks accidental.

Step 3: Exposure and white balance

Exposure and white balance are global corrections, and they come before anything local. Set exposure so the face reads as the brightest meaningful area without looking washed out. Then set white balance so skin looks like skin under neutral daylight, not like skin under a kitchen bulb or an overcast window.

The fastest reliable way to set white balance is a neutral reference: click the eyedropper on something that should be gray or white, such as a white shirt in shadow or a gray wall. Avoid bright highlights, which may be clipped, and avoid the skin itself, which is never neutral. If there is no neutral reference, adjust temperature and tint by eye while looking at the whites of the eyes and any white clothing.

Judge on a screen at moderate brightness. Phones and laptops at maximum brightness make almost every photo look brighter than it will on someone else's screen, so people under-expose to compensate. If you can, check the result on a second device before committing.

Step 4: Contrast, and the risk to skin detail at both ends

Contrast adds punch, but a contrast slider or an aggressive curve pushes tones toward the ends of the range, and faces are where that damage shows first. Dark skin loses texture and dimension when shadows are crushed to near-black. Light skin loses its modeling when highlights on the forehead, nose and cheeks blow out to flat white.

This is not a new problem. For decades, Kodak's color print calibration relied on reference cards known as Shirley cards, which for years featured only white models; that is widely documented as one reason darker skin was often rendered poorly. Some modern phone pipelines, such as Google's Real Tone work on Pixel cameras, were explicitly built to address that bias. Editing defaults and presets can still reintroduce it.

For deeper skin tones

• Lift shadows before adding contrast, and add contrast in the midtones with a gentle curve rather than the global slider.
• Watch the histogram: if a large block of the face is piled against the left edge, detail is gone.
• Resist the urge to brighten the whole face into a lighter tone. The goal is visible texture and dimension, not a different complexion.
• Check for a green or gray cast in shadows, which reads as dull, lifeless skin.

For lighter skin tones

• Pull highlights or whites down until the forehead and nose tip show texture again.
• Most editors have a clipping warning, often by holding a key while dragging the whites or exposure slider. Use it.
• Watch for red and magenta in cheeks and ears after warming the white balance; a small saturation cut on reds usually fixes it.

If the photo fights you at this stage, with half the face in deep shadow or a hot spot you cannot recover, the fix is usually in the lighting of the next shoot rather than the editor. Our plain-English guide to headshot lighting covers why that happens and how to set up light so there is less to correct.

Step 5: Color casts from mixed lighting

A single white balance setting cannot fix a photo lit by two different color temperatures. The classic home headshot has cool daylight from a window on one side of the face and warm tungsten or LED light on the other. Correct for one and the other side goes orange or blue. This needs a local fix, not a global one.

1. 1Set global white balance for the side of the face that carries the main light, usually the window side.
2. 2Select the other side with a brush or gradient mask, or an AI subject or skin mask if your editor has one.
3. 3Adjust temperature and tint within that mask until the two sides match.
4. 4Check the neck, ears and hands, which often keep the cast after the face is fixed.
5. 5Look at the background last. A slightly warm or cool background is fine; mismatched skin is not.

Colored light bouncing off a wall or a bright shirt causes a subtler version of this, typically a green or magenta tint under the chin. The same masking approach works. If you shoot at home regularly, the setup advice in our guide to professional headshots at home, especially turning off overhead lights when using a window, prevents most of it.

Step 6: Skin work, and the line between temporary and permanent

The single most useful rule for skin retouching is this: remove what will be gone in two weeks, keep what will still be there. A breakout, a shaving nick, a scratch or a stray hair is temporary. Moles, freckles, scars, birthmarks, laugh lines and pores are part of how you look, and removing them produces a photo of someone else.

Use a spot-healing or clone tool at a brush size just larger than the blemish, one spot at a time. Healing tools sample surrounding texture, so work from the outside of a larger area inward, and watch for repeated patterns where the tool has copied the same patch twice.

• Under-eye darkness: reduce, do not erase. Lower the opacity of the fix so some shadow remains; zero shadow under the eye looks unnatural.
• Shine: reduce highlights locally on the forehead or nose, or clone at low opacity. Do not smooth the whole area.
• Flyaway hair: remove individual strands that cross the face or stand out against the background. Leave the natural outline of the hairline.
• Global smoothing: avoid it, or keep it very low. Any smoothing that removes pore texture from the whole face is visible at full size.

Where exactly the line sits is partly a professional-norms question, and we cover it in depth in our article on how much headshot retouching is too much. This article is the how-to; that one is the judgment call. As a reference point, Getty Images has since October 2017 refused commercial creative submissions in which models' body shapes are retouched to look thinner or larger, while still accepting skin and blemish retouching.

Step 7: Eyes and teeth, and why the standard advice backfires

Almost every editing tutorial says to brighten the eyes and whiten the teeth. Done lightly, both help. Done to the level most tutorials demonstrate, they are the fastest way to the uncanny look, because real eye whites and real teeth are never pure white and never brighter than the skin's highlights around them.

The whites of the eyes contain small blood vessels and fall into shadow under the upper lid. When they are brightened uniformly, the eye loses its roundness and appears to glow. Teeth behave similarly: natural teeth are slightly warm and darker toward the back of the mouth. Whitening them to neutral gray-white makes them look like veneers or like a separate object pasted into the smile.

• Eyes: a small exposure lift on the iris, not the whites, restores life. If you touch the whites, reduce redness through a saturation cut rather than raising brightness.
• Catchlights: keep the natural ones. Adding or enlarging catchlights is a common tell.
• Teeth: reduce yellow saturation slightly and stop well before neutral. Never raise brightness of teeth above the brightest part of the skin.
• Test: toggle the adjustment off and on. If you can see the change at normal viewing size, it is probably too much.

Over-bright eyes and overly uniform teeth are also common tells in AI-generated portraits, which is part of why they read as artificial. Our list of signs an AI headshot looks fake is useful as a checklist for manually edited photos too.

Step 8: Sharpening, sizing and export

Sharpening and resizing interact, so treat them as one step at the very end. Resize to the output dimensions first, then apply output sharpening suited to that size. Sharpening a large file and then shrinking it wastes the effect; shrinking after heavy sharpening can leave halos around the hairline and shoulders.

Keep sharpening modest and mask it away from skin where possible, so it acts on eyes, eyelashes, hair and fabric texture rather than exaggerating pores. Many editors have a masking slider on the sharpening control that does exactly this. Then keep an unsharpened, full-resolution master file so you can produce new sizes later without stacking edits.

Export settings by destination

Each destination recompresses, crops or prints your file differently, so one export rarely serves them all. Produce each version from the master rather than from another export.

• LinkedIn profile photo: square JPG. LinkedIn recommends between 400 x 400 and 7680 x 4320 pixels, with a maximum file size of 8MB; exporting around 800 to 1000 pixels square gives headroom for high-density screens. Full specs, including the banner, are in our LinkedIn photo sizes reference.
• Other profiles (Slack, Teams, Google, GitHub): the same square file generally works. Platforms recompress uploads, so start from a clean, high-quality JPG rather than an already compressed copy.
• Website team page or speaker bio: match the site's aspect ratio exactly across everyone, typically 4:5 or square, and size to roughly twice the displayed width for sharp display on high-density screens.
• Print (business cards, programs, press): export from the full-resolution master at the physical size requested, commonly at 300 pixels per inch, and ask the printer whether they want a specific color profile.
• Email signature: small file, small dimensions. A multi-megabyte signature image is annoying for recipients.

The two failure modes: over-edited and under-edited

Almost every disappointing headshot edit falls into one of two categories. Over-editing produces a portrait that no longer looks like the person who will walk into the meeting. Under-editing leaves a distracting, easily fixed problem in the frame because the editor was afraid of doing too much. Both cost credibility, in different ways.

The over-edited portrait

The signs are consistent: poreless skin, glowing eyes, bright neutral teeth, a slimmed jaw, removed moles or lines, and often a color grade that makes skin look orange or porcelain. Each change might be small, but they compound. The practical cost is that a recruiter, client or patient meets you and notices the difference, which undermines trust at exactly the moment the photo was supposed to build it.

The fix is a before-and-after check at normal viewing size, ideally by someone who knows you. Ask them one question: does this look like you on a good day? If the answer is "it looks like a better-lit version of you," you are done. If it is "it looks like your sibling," roll back the local edits first.

The under-edited portrait

The opposite mistake is leaving in things a viewer's eye goes straight to: a crooked horizon, an orange cast on one cheek, a bright power socket behind your shoulder, a visible fresh blemish, or clothing lint on a dark jacket. None of these are part of your likeness, so removing them is not deceptive.

• Distracting objects in the background, especially bright ones near the head.
• Obvious color casts or mismatched skin tone across the face.
• Dust spots, lint and stray hairs.
• Crops that clip the head or leave awkward empty space.

If the background itself is the problem rather than one object in it, decide whether to tidy it, blur it or replace it before spending time on pixel-level cleanup. A plain wall that simply needs a lamp cloned out is a two-minute job; a cluttered room behind your head is usually better blurred or reshot.

Where free and paid editing tools actually differ

For the global steps, meaning crop, straighten, exposure, white balance, contrast and basic healing, free phone and browser editors are fully capable. Feature sets change frequently, so check current versions, but the meaningful differences between free and paid software tend to show up in local precision, raw file handling and batch work rather than in basic correction.

• Raw files: if your camera or phone shoots raw, dedicated desktop editors recover far more highlight and shadow detail than editing a JPG. That matters most for the skin-tone problems in Step 4.
• Masking: free tools typically offer brushes and simple gradients. Paid editors increasingly offer automatic selections for a subject, skin, eyes or teeth, which makes the mixed-light fix in Step 5 much faster.
• Healing quality: basic healing tools handle a single blemish well but struggle next to hairlines, glasses frames and patterned fabric, where more advanced or AI-assisted removal does better.
• Batch consistency: copying one set of adjustments across a team's photos is a paid-software strength and matters for company pages.
• Export control: desktop editors let you set exact pixel dimensions, quality, color profile and output sharpening; many phone apps choose some of these for you.

AI editing tools: what they do well, and where they show

AI features are now built into most editing software, from generative removal tools to one-tap skin smoothing and automatic background replacement. As a company that makes AI-generated headshots, we have an obvious interest here, so this section tries to be plain about both sides: these tools are genuinely useful, and they have recognizable failure patterns.

What they do well

• Generative removal of background distractions. Adobe's Lightroom Generative Remove, for example, lets you brush over an object and uses Adobe Firefly to generate a fill that blends with the rest of the frame. Removing a light switch, a sign or a person in the background is where this class of tool is strongest.
• Subject and skin masking. Automatic selections that once took ten minutes of brushing now take a second, and are usually accurate enough for color and exposure work.
• Background separation. Blurring or swapping a busy background works well when the edge between hair and background is clean.

Where viewers notice the artefacts

• Hair edges. Fine flyaways, curls and textured hair against a replaced background often turn into a soft halo or a cut-out outline.
• Skin smoothing. AI smoothing tends to remove texture uniformly, producing the waxy look even at low settings. It also sometimes softens moles or freckles you did not ask it to touch.
• Lighting mismatch. A replaced background with light coming from a different direction than the light on the face reads as fake even when people cannot say why.
• Invented detail. Generative fill near the face, glasses, ears or collars can create plausible but wrong detail: an earring that was not there, an extra collar point, asymmetrical frames.
• Metadata. Some tools attach Content Credentials, tamper-evident metadata based on the C2PA standard, that record AI-assisted edits. Adobe attaches them to photos edited with Generative Remove, and some platforms read and display them.

The ethical line: editing a photograph versus generating a likeness

There is a real difference between correcting a photograph of someone and generating new pixels that represent their face. Removing a background object, fixing color or healing a temporary spot all leave the person's actual features intact. Reshaping a jaw, changing eye shape or regenerating the face produces a likeness that the camera never recorded.

AI headshot generators, including ours, sit on the second side of that line by design: they create new portraits from reference selfies rather than editing one photograph. That is a legitimate use for a profile photo when the result still looks like you, and we explain what that means on our AI disclosure page. It is not a substitute for a real photograph where one is required, such as ID documents. Our comparison of AI headshots and a photographer covers where a human photographer still does a better job, including unusual features, specific brand looks and anything that must be documentary.

Whichever route you take, the quality of the input sets the ceiling. For a better starting photo from a phone, the camera-distance and light principles in how to take a good selfie prevent most of the problems this workflow corrects.

A two-minute final check before you publish

Before you upload, step away from the edit for a few minutes, then look at it fresh. Editing fatigue is real: after twenty minutes of looking at one face, the eye adapts to warm casts, heavy smoothing and over-bright eyes. A short break and a quick checklist catch most of what slipped through.

1. 1Toggle the original and the edit side by side at normal viewing size.
2. 2View it as a small circle, the way most people will see it.
3. 3Check it on a second screen or phone.
4. 4Confirm moles, scars, freckles and lines are still there.
5. 5Confirm nothing distracting remains near the head.
6. 6Export from the master, not from a previous export.

If the edited version looks like you on a well-lit, well-rested day, the edit has done its job. If it looks like a more polished stranger, go back to Step 6 and undo, because the photo's only real task is to be recognized when you walk into the room.