Shoot photos that cut out perfectly
Here's a secret from the editing side: the quality of a background removal is mostly decided when you press the shutter, not when the tool runs.
Last updated: September 2026
Five habits at shooting time turn "pretty good" cutouts into flawless ones.
Full resolution. No watermark. Actually free. Some background removers only give you a small free preview and charge for the full-size file. This one gives you the full resolution free.
Related: Remove the Background From a Product Photo: Free, Full Size, remove.bg is shutting down on.
Your photo is deleted from the server the moment the cutout is sent back. Need the erase and restore brushes, or a coloured background behind the subject? Open the full editor.
1. Contrast is everything
A segmentation model finds your subject by looking for the boundary between subject and background. The clearer that boundary, the cleaner the cut. So the single most effective thing you can do costs nothing: put the subject in front of something that doesn't look like it.
- Dark subject → light background. Light subject → darker background.
- Watch colour too: a green plant against a hedge, a denim jacket against a blue wall, blonde hair against a beige curtain, all technically "different", all needlessly hard.
- Busy backgrounds aren't automatically bad, but a background containing things that look like the subject is: a person in a crowd, a mug on a shelf of mugs. The model may keep the wrong twin.
You don't need a studio backdrop: a plain wall, a large sheet of card, or simply three steps away from the clutter is enough. Distance alone helps: the further the background, the more it blurs, and blur is contrast's best friend.
2. Light evenly: shadows are edge-killers
Hard light from one direction throws a dense shadow that touches the subject's edge, and to a segmentation model, a shadow glued to a shoe looks like part of the shoe. The result is a cutout with a dark blob bleeding out of the outline.
- Best free light: an overcast sky, or a window without direct sun. Soft, even, no hard shadows.
- Indoors at night: don't aim a lamp at the subject; bounce it off a white wall or ceiling.
- Avoid backlight. A subject shot against a bright window becomes a dark silhouette whose edges dissolve into glow: hair detail is unrecoverable.
- Skip on-camera flash. It flattens the subject and paints a hard shadow directly behind it, exactly where the edge needs to be clean.
3. Sharp edges only: motion blur is fatal
Of everything on this list, motion blur is the one mistake that cannot be fixed afterwards. A blurred edge isn't an edge anymore: it's a gradient where subject and background are physically mixed in the pixels. No tool, and no person, can decide where one ends and the other begins, because the photo itself no longer contains that information.
- Brace your elbows, lean on something, or set the phone on a surface and use the timer.
- Photographing kids or pets? Take a burst and pick the sharpest frame: one frozen moment is all you need.
- In dim light, phones silently choose slow shutter speeds. If edges look soft, add light rather than trying again in the same gloom.
- Zoom to 100% and check the edge before leaving the scene. Sharp overall but soft on the moving hand? That hand will cut out soft, too.
4. Distance and framing
Give the subject room to be the subject:
- Fill most of the frame. A subject that occupies 10% of the photo gets only 10% of the pixels: edge detail like hair is described by a handful of pixels and comes out coarse. The same rule applies when you're photographing an old print with your phone instead of a person, PicReviver's guide to scanning old photos covers the framing, light and DPI that make that shot usable.
- Don't crop mid-limb. Keep the whole subject inside the frame with a little margin, so nothing important is amputated.
- Whole subject in focus. Strong portrait-mode blur on parts of the actual subject (an outstretched hand, the back of a jacket) makes those parts as hard to cut as motion blur. Step back slightly instead of shooting extremely close.
- One clear subject. The model keeps what looks most "foreground". If two people overlap, expect them to be cut as one: you can separate them afterwards with the erase brush, but a small gap at shooting time is easier.
5. Know what the tool genuinely finds hard
Even a perfect photo can contain subjects that are simply difficult. Knowing these in advance saves frustration:
- Glass and transparent objects: the background is visible through the subject, so there's no clean boundary. Shoot bottles and glasses against a plain, evenly lit background for the best chance.
- Thin structures: bicycle spokes, netting, chair legs, earbud wires. Sub-pixel details may vanish; get closer so they occupy more pixels.
- Reflective surfaces: chrome and mirrors reflect the old background, which stays visible inside the cutout even after a perfect cut.
- Wild hair against a busy scene: flyaway strands over a detailed background are the hardest case in the field. A plain background behind the head solves 90% of it (see why in our alpha-channel explainer).
- No clear subject at all: landscapes, textures, food flat-lays with ten items. Segmentation models are built to find a subject; if there isn't one, BGNinja will tell you rather than return an empty PNG.
Where people are when they land here
Our own logs for the 40 days from 13 July to 21 August 2026 split incoming visits into three kinds: 259 arrived via an outside link (a search result, a shared link, another site), 62 typed the address directly or came from a bookmark, and 20 moved here from another page on bgninja.com or one of our other tools. Most people who land on a guide page like this one, in other words, didn't stumble onto it while browsing around the site: they came looking for a specific fix, usually already holding a photo that didn't cut out the way they hoped.
That matches what this page is for: the best time to fix a bad edge is before you press the shutter, not after. If you're here because a cutout already came out rough, the fastest read is section 5 above: it lists exactly which of the five habits you probably missed, so you know what to change next time instead of fighting the same photo twice.
Shot something? Cut it out
Upload your photo and see the edges for yourself: free, no account, deleted immediately after.
Also need to log the receipts from the shoot for your own expense claim? ScanReviver's receipt scanner turns a phone photo of each one into one clean PDF.
Now bill the client for the shoot itself? ScanReviver's invoice generator turns your day rate and any expenses into a proper PDF invoice, no spreadsheet template needed.
Quick answers
Why did my photo cut out badly?
Almost always one of four things: low subject/background contrast, a hard shadow touching the subject, motion blur on the edge, or no single clear subject. All are prevented at shooting time.
Does a higher resolution photo give a better cutout?
Up to a point, yes: more pixels along the edge means finer detail like hair strands survives. Sharpness matters more than raw megapixels though: a sharp 12 MP photo beats a blurry 48 MP one every time.
Can I fix a bad edge after the cutout?
Often, yes: BGNinja's erase brush removes leftover background patches and the restore brush brings back anything the model dropped. What can't be fixed here is an edge that was blurred in the original photo: for a whole photo that's soft, not just one edge, PicReviver's blur fixer is the right tool, before you cut anything out.
Selling online?
There's a dedicated guide for marketplace sellers: product photos on a white background (requirements, shadows and a batch workflow).
Do I need to buy backdrops, or can I use free photography backgrounds?
You don't need to buy anything: a plain wall, a large sheet of card, or just stepping back from clutter are the best free photography backgrounds you'll find, because contrast at shooting time matters more than the material. Whatever's behind your subject gets replaced afterwards anyway: that's the photography edit background (or photography editing background) step, so the shooting-time background only needs to be plain and contrasting, not pretty.
What do AI background removers struggle with?
Transparent and reflective objects (glass, bottles), very thin structures (bicycle spokes, chair legs, netting), subjects that blend into the background, multiple overlapping subjects, and blurry edges. Knowing this in advance lets you shoot around it.
Does photography background colour matter, or just contrast?
Contrast matters more than any specific colour, but green causes its own problem: a photography background green (the classic green-screen choice) can bleed a faint colour cast onto the subject's edge, which some segmentation models keep. A plain grey, blue or white background avoids that entirely and works just as well without the risk.
By Sam Ridder, I build and run BGNinja on my own hardware, on my own. Who I am.