How to Crop Someone Into a Picture (2026 Guide)

Published October 9, 2026 · Daniel Tello · RevelRaw

You've got the right subject, the right background, and a composite that still looks off. The face may be cleanly cut out, but the person doesn't feel like they belong in the frame yet, and that's the part most people notice only after export.

The fix is rarely one thing. A convincing result comes from matching the mask, the geometry, the light, and the contact cues so the cutout reads like it was photographed there in the first place. This guide walks those four in order, then covers the opposite job, cropping someone out, and the export check that catches what the edit view hides.

Why Does the Composite Look Fake?

The first bad sign usually shows up before you can explain it. The file exports cleanly, the edges are technically acceptable, and still the person looks pasted on, like the background and subject never met the same camera.

That happens because the eye does not judge composites by cut quality alone. It checks whether the subject could physically occupy that scene, whether the lighting agrees, and whether the edges carry traces of the old background color or a halo around hair and shoulders. If any one of those cues is wrong, the whole edit feels staged.

The four usual failures

The problem areas are edges, lighting, color, and perspective. Halos around hair and fingers, a mismatch between skin tone and background warmth, or a subject placed at the wrong scale all read as a composite even when the mask itself is precise.

A technically clean cutout can still look wrong if the subject's shadow, height, or camera angle doesn't belong to the destination image.

A useful way to audit your own edit is to borrow the checklist that people use to catch other people's. The AI Image Detector blog's guide to forensic methods for image edits (a vendor post, but a clear one) tells readers to look at shadows before faces, to study edges for halos and blur that stops at a subject boundary, and to compare noise and compression between regions. Those are exactly the places your composite will give itself away, so check them yourself first.

The useful shift is this: cropping someone into a picture isn't just removal. It's placement, and placement has to satisfy the viewer's instinct for physical consistency.

How Do You Mask the Person Without Destroying the Original?

Start by keeping the source image intact. Erasing pixels may feel faster, but it boxes you into mistakes you can't easily undo, while a layer mask lets you revise edge quality, transparency, and feathering without destroying the original subject.

Place the subject above the destination background, then use automatic subject selection only as a starting point. A selection tool can get you close, but it won't solve hair, hands, motion blur, or transparent details on its own.

Refine the mask before you trust the composite

Convert the selection into a mask, then inspect it at 100% and 200%. That's where the mistakes show up, especially along high-frequency areas like hair, fingers, clothing seams, and motion-blurred borders.

Use grayscale on the mask, not guesswork. White reveals, black hides, and gray creates partial transparency, which matters when the edge should fade naturally instead of snapping into a hard outline.

A practical sequence works better than improvisation:

  1. Duplicate the source layer so the original stays untouched.
  2. Build the initial selection with your subject tool of choice.
  3. Add the mask instead of deleting pixels.
  4. Refine the edge with a brush, selection refinement, or channel contrast.
  5. Keep feathering minimal so you don't smear detail.
  6. Check against a light and a dark background to catch color contamination early.

Common failures are easy to spot once you know them: halos from retained background color, unnaturally hard hair edges, missing fine detail, and limbs that turn semi-transparent by accident. If the mask is weak, later corrections won't hide it.

Generative tools take a different route to a similar picture. Virtual try-on services such as Vtry AI for fashion ecommerce render a garment onto a model rather than masking one photo into another, and the company's own page warns that generated imagery can alter details around small graphics and layered edges and still needs a manual review. Different method, same review gate: zoom in on the edges before you trust the result.

How Do You Match Scale and Perspective?

Once the cutout is clean, judge whether the subject could stand convincingly in the scene. That means checking scale, horizon, and vanishing points before you touch color, because a perfectly lit person still looks wrong if their body size or camera angle clashes with the destination image.

Use recognizable references in the background. Doorframes, furniture, stairs, and other people give you a real-world sense of proportion that is harder to fake than eyeballing the canvas.

Align the camera view first

Compare the horizon line in both images. If one frame was shot from a low angle and the other from eye level, the body will feel misplaced in the scene unless you adjust perspective to match the destination view. Small corrections can be faked with a perspective or warp transform. Large ones cannot, so when you have a choice of source photos, pick the one shot from the same height and distance as the background.

Watch the feet and nearby architecture. The angle of the shoes, the tilt of a floor line, and the convergence of walls all tell the viewer whether the subject belongs there. If those lines disagree, the composite breaks even when the cutout itself is flawless.

Practical rule: match the subject to the destination scene, not the other way around. The background sets the geometry, and the subject has to obey it.

A good habit is to check placement, scale, perspective, and horizon in that order. If one of them is off, fix it before moving on, because later color work can't rescue a person who appears to be standing in the wrong space.

How Do You Match Color and Exposure?

Color is where a lot of composites go from almost right to obviously edited. The subject may be in the correct place with a believable mask, but if the exposure or white balance doesn't match the background, the eye catches the mismatch instantly.

Start with tone, not saturation. Bring the subject's exposure and contrast into the same range as the destination image first, then fine-tune warmth or coolness with white balance adjustments based on a neutral area or sampled background tone.

Adjust in layers so nothing gets locked in

Use Curves or Levels for exposure and contrast, then adjust color with Hue/Saturation or Photo Filter style tools as needed. Clip each adjustment layer to the subject layer so it affects only the person, and keep them as layers rather than baked edits so you can revisit the match without degrading the cutout.

A good color match usually has three parts:

You can also use a neutral reference or sampled background tones when the image gives you one. That keeps you from guessing at warmth by eye, which is where many edits drift into the “screen looked fine until export” problem.

If both source files are RAW, do the rough match before the composite: develop the subject and the background with the same look and white balance in your RAW editor, so the layer editor only has to close a small gap. For the thinking behind those tonal decisions, the article on color grading photography is a useful companion read. The main point is simple, though: color corrections should stay nondestructive so you can still revisit the mask and geometry if the composite starts to drift.

How Do You Ground the Subject With Shadows and Edges?

The smallest fix often sells the whole illusion. A subject without a grounded shadow looks like they're hovering, even if the edge work and color are otherwise strong.

Match the shadow to the light already present in the destination image. Its direction, softness, and opacity should follow the same logic as the scene, not the logic of the source portrait.

Build the shadow as part of the environment

Paint the shadow on its own layer beneath the subject, soft and restrained under the feet or any other contact point, and lower its opacity until it matches the other shadows in the scene. A hard black shape looks like an added graphic, while a shadow that fades naturally feels like it belongs to the floor, pavement, or surface under the subject.

Edge softness matters for the same reason. A person placed into a hazy or lower-resolution background can't stay razor sharp without looking disconnected, but over-blurring destroys texture and makes the cutout feel mushy. A touch of matching grain or noise over the whole composite often does more than blur, because it gives subject and background the same texture.

The final visual check should cover the same things every time:

A composite becomes believable when the subject appears to share the same air, light, and surface as the background. That's the difference between a cutout and a picture.

What If You Need to Crop Someone Out Instead?

The same search often means the opposite job. If the goal is to remove bystanders or reduce identity exposure, the decision changes, because you're no longer only shaping composition, you're deciding what information stays visible.

Cropping only protects what falls outside the frame. If a face, badge, license plate, location clue, or reflection still sits inside the retained area, the edit hasn't finished the privacy job.

Choose the right privacy action for the image

Cropping works well when the person outside the center of interest is accidental and the remaining frame still tells the story. Blurring can protect identity while keeping some context, and deletion may be the stronger option when the remaining scene still carries sensitive clues.

Privacy edits should be made at full resolution, not only on a thumbnail view.

The practical distinction is between accidental bystanders and identifiable subjects. If the person is the point of the image, you need to think harder about consent, metadata, and whether cropping is enough at all.

The file can also leak what the frame no longer shows. Capture time, device details and GPS travel with the export unless you review them, so metadata review tools belong in a privacy edit as much as the crop does. When the assignment is sensitive, local processing matters, and RevelRaw is one option that keeps editing on-device without cloud upload or telemetry, with a choice of which creator, caption, location and GPS fields leave with the file.

The trade-off is simple. Cropping may preserve context better, but stronger privacy often comes from more decisive edits. Choose the goal first, then edit to match it.

How Do You Check and Export the Composite?

Never trust the composite just because it looks good at edit size. Compression, resizing, and platform-specific crops can expose halos, softness, or color mismatch that stayed hidden while you were working.

Run one last pass over the physical cues, not just the pixels. Check whether the horizon still lines up, whether the shadow points the same way as the scene light, and whether the subject's saturation or white balance drifts away from the background.

Export for the destination, not for a generic file

Different placements need different framing. A 1080 × 1350 portrait crop contains 1,458,000 pixels, while a 1080 × 1080 square crop contains 1,166,400 pixels, so the portrait version provides 291,600 more pixels and is 25% taller at the same width. The 9:16 format at 1080 × 1920 is about 42% taller than the 4:5 version, which is why it fits full-screen vertical stories better. Decide the final crop before you place the person, so the subject does not end up on a crop line.

That's why one universal crop rarely works. Feed posts, stories, web placements, and print all punish sloppy framing in different ways, and a clean edit at one size can fail at another.

For export decisions, the Instagram export settings guide is a practical reference point, and photo editing for Instagram covers choosing a crop that gives faces and hands room. The rule that matters most is still the same, though: inspect the result at the size people will see, not just at the size you edited.

RevelRaw does not composite layers, so the cutout itself belongs in a layer editor. What it does well is the work on either side of that step: develop the subject and background RAWs with the same look and white balance, use up to three free mask layers to match local exposure and warmth, and export the finished composite as JPEG, PNG, TIFF or HEIF in sRGB, Display P3 or Adobe RGB with the metadata you choose. Try RevelRaw for that preparation and output work on your Mac. Processing and export run locally, photos are not uploaded, and it requires macOS 26 or later on a Mac with Apple M1 or later.

Frequently Asked Questions

What is the easiest way to crop someone into a picture?

Open both photos in a layer-based editor, put the person on a layer above the new background, select the person with the subject tool and turn that selection into a layer mask instead of erasing. Refine the mask at 100 percent around hair and hands, scale the layer so the person matches a doorway or other reference, then match exposure and white balance with adjustment layers and add a soft shadow under the feet.

Why does the person look pasted on even with a clean cutout?

Because the eye checks placement, not just edges. A clean mask still fails when the person is the wrong size for the scene, the horizon or camera height differs between the two photos, the light comes from a different direction, the skin tone is warmer or cooler than the background, or there is no contact shadow. Fix scale and perspective first, then light, then color.

How do you make the lighting match when cropping someone into a picture?

Pick a subject photo lit from the same side and at a similar height as the background light, because direction cannot be fixed afterwards. Then use Curves or Levels to bring the subject's shadows and highlights into the background's range, adjust white balance so neutral areas agree, and paint a soft shadow on its own layer that follows the scene's light.

Can you crop someone into a picture on a phone?

Yes, for simple cases. Phone editors can lift a subject with one tap and paste it into another photo, which is fine for a casual composite. Check the edges at full zoom, scale the person to the scene, and add a shadow if the app allows it. Hair, glasses and motion blur usually need a desktop editor with a proper mask and brush.

Can RevelRaw crop someone into a picture?

No. RevelRaw is a RAW photo editor, not a layer compositor, so the cutout and placement happen in an app such as Photoshop, Affinity Photo, Pixelmator Pro or GIMP. RevelRaw is useful before and after that step: develop the subject and background RAWs with the same look, use up to three free mask layers to match local exposure and warmth, then export JPEG, PNG, TIFF or HEIF at the size the destination needs. Processing runs on your Mac and photos are not uploaded.