How to do it, step by step
- Start from the original file Edit the photo in your library, not a screenshot, a download from a chat app or a re-shared copy. Each of those has already been recompressed, and no amount of editing gets that back.
- Fix exposure first Lift shadows until faces are readable and pull highlights down if the sky has gone white. Judge the whole photo at this stage, not one area, and do not move on until the brightness is right.
- Correct white balance second Warm up a photo taken in shade, cool down one taken under indoor bulbs, and judge it on skin rather than on walls. This is the step that makes a photo look like a better camera took it.
- Crop and straighten third Level the horizon, remove dead space at the edges, and crop tighter if the photo is soft. A tighter frame on a soft photo reads sharper because sharpness is relative to display size.
- Add contrast and local detail gently A small amount of contrast, and a little structure or clarity. Stop well before it looks processed, because the point at which you notice is already past the point where other people notice.
- Sharpen last, and check on skin Apply sharpening at the end, at the size you will actually use, and zoom into a face rather than an edge. Skin is where over-sharpening shows first and worst.
The order that actually works
Most people start with a sharpening slider or an "enhance" button, which is the last step and the least effective one. Editing in the wrong order means every adjustment fights the previous one, and the result is the over-processed look everyone recognises.
Exposure first, because every other judgement depends on it. If the photo is a stop dark, the colours look wrong, the contrast looks wrong and the noise looks worse than it is. Lift shadows before touching anything else and recover highlights if the sky has gone white.
White balance second. This is the step that does the most for photos of people and the step almost nobody does. Indoor lighting turns skin orange, shade turns it blue, and mixed light turns it both at once. Correcting it makes a photo look like a better camera took it.
Crop third, once you can see what is actually in the frame. Straighten the horizon, cut the dead space, and remember that cropping in makes a soft photo look sharper because sharpness is relative to display size.
Contrast and local detail fourth, gently. Sharpening last, checked on skin rather than on edges, because skin is where over-sharpening first becomes obvious.
That sequence takes ninety seconds and beats every one-tap enhance button in the app store.
Doing it with what you already have
Every step above is available free on the phone in your pocket, and reaching for a downloaded app first is usually a mistake.
The built-in editor. On iPhone, Photos gives you Exposure, Brilliance, Shadows, Highlights, Contrast, Warmth, Tint, Sharpness and Definition, plus a crop with straightening. On Android, Google Photos gives the same set. Between them these cover the entire order above, they are non-destructive, and they are instant because nothing uploads.
Snapseed when you need more. Free, no subscription, Google-owned. Its Selective tool lets you brighten one face without brightening the whole photo, its Curves give real control over contrast, and its Healing patch removes small distractions without generative redrawing. The Structure slider in Details is the honest version of "enhance": it raises mid-scale detail from what is already in the file.
Lightroom Mobile if you shoot a lot. The free tier covers most editing including profiles and colour mixing, with masking behind the paid plan. It is slower to use for a one-off photo and better than anything else for a consistent look across a set.
Google Photos Photo Unblur for old scans and mildly soft shots, free to all users and more conservative than a full generative pass.
Where AI helps and where it lies
AI is excellent at some of these jobs and structurally unsuited to others. The distinction is whether the job needs new content invented or existing content adjusted.
| The job | Best method | Is AI adding content? | What it costs you |
|---|---|---|---|
| Brighten an underexposed photo | Exposure and shadow sliders | No | Nothing; this is arithmetic on real data |
| Fix orange or blue skin | White balance slider | No | Nothing, and it is reversible |
| Reduce noise in a night photo | Trained noise reduction | Slightly | Fine texture; skin can go waxy at high strength |
| Remove an object | Clean Up, Magic Eraser or generative fill | Yes, in one area | Little, if the fill is background rather than a subject |
| Sharpen a soft photo | Sharpening, then Structure | No | Noise gets exaggerated too |
| Upscale for a larger print | Generative super-resolution | Yes, everywhere | Every fine detail at the new size is a guess |
| Restore a damaged old print | Patch healing first, AI second | Yes, heavily on faces | The faces become partly invented; keep the scan |
| Make a photo look professional | Better light, not better software | No | Requires shooting it again |
Read the third column as a risk rating. Where it says no, use the tool freely. Where it says yes, decide first whether you would mind the result being wrong.
Quality for print against quality for screen
The two have almost nothing in common, and confusing them is why people upscale photos that did not need it.
For a screen, resolution stopped being the limit years ago. A modern phone photo has several times more pixels than any social platform will display, and the platform is going to recompress it anyway. What limits quality on screen is light, colour and composition. Upscaling a photo before posting it achieves nothing except adding invented detail that the platform then compresses.
For print, size and viewing distance decide everything. A phone photo prints beautifully at small and medium sizes from the original file with no upscaling at all. Beyond that, upscaling helps, but you should know that every detail added at the new size is generated, and that the larger the print the closer people stand and the more they see.
Three print habits worth having. Export from the original, never from a photo that has been sent through a messaging app, which recompresses heavily. Sharpen at the output size, not before, since sharpening for a phone screen and for an A3 print are different amounts. Check the crop against the paper ratio, because the default 4:3 photo does not fit standard print sizes and a lab will crop it for you, usually through somebody's head.
The biggest gain is at capture
Everything above is damage limitation. The gap between an average photo and a good one is almost entirely decided before the shutter fires, and four habits close most of it.
Find better light rather than better settings. Move the person towards a window, turn them so the light falls across their face rather than behind them, and get out of overhead sun. Soft directional light does more for quality than every slider on this page.
Clean the lens. A fingerprint smears highlights and reduces contrast across the whole frame, and it is the single most common cause of photos that look inexplicably soft.
Use the main camera. The ultra wide has a smaller sensor and a weaker lens, and digital zoom past the optical range is upscaling by another name. If you need to get closer, walk.
Take more frames than you need. The difference between a good photographer's output and yours is largely the number of frames they deleted. Shoot ten, keep one, and the keeper will be sharper, better timed and better composed than the single frame you would otherwise have taken.
Where ShutterAI fits
Naming our own app so you can weigh it accordingly: ShutterAI is built around the argument in this section, that quality is decided at capture. True to You corrects skin-tone white balance as you shoot rather than afterwards, Light Finder tells you when the good light is, Best Face picks the sharp frame out of a burst, and Glow retouches without erasing the texture that makes skin look real. Vault keeps every original, so any edit can be taken back.
It is not out yet. ShutterAI turns up on iPhone this autumn, free to begin with, and the early list is on the homepage. The five-step order above works today, on any phone.
Questions people ask
How to enhance photo quality?
Work in order: exposure, white balance, crop, contrast, then light sharpening. Those five steps in your phone's built-in editor improve most photos more than any one-tap AI enhance button, and they add nothing that was not there.
Does AI improve photo quality?
It improves how a photo looks while adding detail that was never recorded. That is a fair trade for social posts and the wrong trade for documents, identification, evidence or archival copies of family photographs.
What is the best free app to improve photo quality?
The editor already on your phone, then Snapseed for selective adjustments and structure, then Lightroom Mobile's free tier if you edit a lot. All three are free and none of them invent detail.
Why do my photos look worse after sharing them?
Messaging apps and social platforms recompress images heavily. Always export and edit from the original in your library rather than from a copy that has been sent through a chat.
Do I need to upscale a photo before printing it?
Not for small and medium prints, where a modern phone photo has plenty of pixels. For large prints upscaling helps, but the added detail is generated, and the bigger the print the closer people stand.