You have the photo. The moment was right, the framing was right — and the picture is soft. Maybe your hand moved, maybe the camera focused on the background, maybe it is an old shot that was never sharp to begin with. So you search for a fix, and the internet answers with confidence: upload it here, tap once, watch it go sharp.
Here is the honest answer to the question “can you unblur a photo”: no tool can truly unblur one. A photo can often be improved, sometimes a lot. But improvement and restoration are not the same thing, and knowing the difference will save you from trusting a result you should not trust — and from deleting an original you should have kept.
Why a blurry photo cannot simply be reversed
Think of a sharp photo as a grid of tiny, distinct squares of colour. At a sharp edge — the boundary between a dark jacket and a bright sky, say — neighbouring squares are very different. That difference is what your eye reads as crispness.
Blur mixes neighbouring squares together. Light from one part of the scene spills into the pixels next to it, so a sharp boundary becomes a gradual slope. The distinct values are averaged away, and once averaged, the original values are gone. You can guess what they might have been, or make the slope steeper again so it looks sharper — but you cannot recover values that no longer exist in the file. Sharpening, deblurring, and AI can emphasise, estimate, or invent. They cannot resurrect.
There is one genuine exception: if a better version exists somewhere — the original camera file, an uncompressed copy, another shot taken a second later — that is a recovery. Finding the original is the only way to get detail the blurry copy lost.
First, work out what kind of blur you have
Not all blur is the same, and the kind of blur decides what kind of fix has any chance at all. Before you run any tool, look at the photo at full size and ask which description fits.
| Kind of blur | What it looks like | Chance of a useful improvement |
|---|---|---|
| Camera shake | The whole frame is smeared, often in one direction; worse in low light | Mild shake can improve; heavy shake rarely becomes truly sharp |
| Motion blur | The background is fine but a moving subject streaks | Light streaks can improve; fast movement usually stays soft |
| Missed focus | The subject is soft while something else is sharp | Mild softness improves well; a badly missed focus does not |
| Low resolution or heavy compression | Everything looks mushy or blocky when enlarged, rather than smeared | Small-display improvement only; enlarging adds no real detail |
| Noise plus blur | A grainy, soft low-light shot where sharpening also sharpens the grain | Improve with care; aggressive sharpening makes grain worse |
If the subject is still recognisable — eyes visible, edges present, text roughly readable — there is something for a tool to work with. If a face is a featureless smear or text is an unreadable blob, no honest tool will give you back the real face or the real words.
Three things sold as “unblur” — and what each actually does
1. Sharpening. Sharpening does not remove blur. It increases contrast along edges that are still in the photo, so the image reads as crisper. It works with detail that was actually recorded. It cannot create detail that was not.
2. Deblurring. Deblurring tries to model how the photo was blurred — for example, the direction and length of a camera shake — and then partially reverse that model. When the blur is simple and the model is right, this can recover structure sharpening cannot, because some original information is still present in the smeared pixels, just mixed together. But the reversal is never exact: noise gets amplified, and the result needs careful judgement. Deblurring estimates; it does not resurrect.
3. AI generation. Modern AI enhancers go further: given a blurry input, they generate detail that plausibly fits — hair strands, skin texture, clean edges — based on patterns learned from millions of other photos. The result can look impressively sharp. But the new detail is invented, not recovered. A sharpened-looking face from an AI tool is a likeness the model produced, not evidence of what the person actually looked like. That is fine for a social media post, not for evidence, identification, or any situation where exact detail matters. See this explainer on why AI enhancers generate rather than recover detail.
When a tool promises to “unblur” without saying which of these three it does, treat the promise with suspicion. The honest question is never “does it make it sharper?” but “is the new sharpness real, estimated, or invented?”
How sharpening really works: the unsharp mask in plain English
The classic sharpening method has a confusing name: the unsharp mask. Despite the name, it has nothing to do with making things unsharp. It works like this.
First, the tool makes a blurred copy of your photo. Then it compares the original with that blurred copy. Where the two differ a lot, there must be an edge or fine detail; where they are nearly the same, there is only smooth tone. Finally, it adds that difference back into the original, boosted by a chosen amount. The light side of each edge gets a little lighter, the dark side a little darker, and your eye reads the stronger contrast as sharpness.
Three controls govern the result:
- Amount — how strong the contrast boost is. More means crisper-looking edges, but also more artefacts.
- Radius — how wide a band around each edge is affected. A small radius sharpens fine detail; a large radius adds broader contrast.
- Threshold — how different two areas must be before they count as an edge. Raising it protects smooth areas — skies, skin, backgrounds — from turning crunchy.
The classic failure mode is overdoing it. Push the amount too high and you get halos: bright outlines around dark edges and dark outlines around bright ones. Skin turns waxy or gritty, text gets a ghostly double edge, and the image looks processed rather than sharp. Subtlety wins, and judgement has to happen at full size — a result that looks great as a small thumbnail can fall apart at 100% zoom.
What Manage My Tools’ sharpening tool does — and what it will not do
Manage My Tools’ Image Sharpening Tool is a sharpening tool in the first sense above. It strengthens edges using a convolution filter — the same edge-contrast idea as the unsharp mask — applied evenly across the whole image, at a strength you set with a slider. It is not an AI deblurrer, and it will not invent detail.
Here is how to use it well:
- Start from the best copy you have — the original camera file or the largest available version, not a screenshot or a copy re-saved several times. Every re-save loses a little more.
- Choose an image up to 10 MB and set the strength near the middle of the slider (the default is 50) for a first pass.
- Run it and compare at the size the photo will actually be shown. Small display sizes tolerate stronger sharpening; large images show halos sooner, so lower settings usually look more natural there. If halos appear, back the strength down — more is not always better.
- Keep the original. Work on a copy, so a worse pass can always be redone from the untouched file.
The tool runs entirely in your browser, so your photo is never sent to the site’s server.
What it will not do is fix a badly blurred photo. The tool page says this directly: sharpening cannot restore a blurred original, because the detail a blur removed is not hidden in the file waiting to be recovered. A mildly soft image improves noticeably; a blurred photo will still look blurred, just with harder edges.
How to judge a sharpened photo honestly
Whether you use this tool, an editor, or an AI enhancer, run the same checks before you trust the result:
- View at 100% zoom. Thumbnails flatter everything. Halos, plastic skin, and invented texture all show up at full size.
- Compare side by side with the original. Are edges cleaner, or just harsher? Is texture real, or repeated and artificial?
- Check the parts that matter. Faces, hands, and text are where invented detail is hardest to spot and most dangerous to trust. If text became readable, check it against another source before relying on it.
- Watch smooth areas and judge at the final size. Skies, walls, and skin should stay smooth; if they turned grainy, the sharpening is too strong. Sharpening is a finishing step, best applied once the image is at the size it will be shown — resizing afterwards softens it again or exaggerates artefacts (see how to reduce an image’s file size).
When to accept, when to use AI with care, and when to reshoot
Accept the sharpened result when the photo is mildly soft, the subject is recognisable, and the result will be shown small — a listing photo, a blog image, a social post. This is where sharpening earns its keep.
Use AI enhancement with care when the photo matters emotionally but not evidentially — an old family picture, a favourite snapshot. Enjoy the cleaner version, but keep the original alongside it, and do not present the AI version as what the scene truly looked like.
Do not rely on any enhancement when exact detail is the point: reading a plate, a sign, a document, or identifying a person. Say so plainly rather than trusting a plausible-looking reconstruction.
Reshoot, or find the original, when you can. A second shot with the camera steadied, focus tapped on the subject, and decent light will beat any amount of post-processing — and bright scenes let the camera use a faster shutter speed, which freezes movement. Before you spend an hour fixing a blurry copy, check whether the camera roll, a cloud backup, or the photographer still holds the original file. That five-minute search is the only true “unblur” that exists.
FAQ
Can you unblur a photo completely?
No. Sharpening makes a mildly soft photo look crisper, deblurring partially reverses simple blur, and AI generates plausible sharp detail — but none recovers the exact original detail blur destroyed. A fully sharp result from a badly blurred source is a reconstruction, not a recovery.
Why does my photo look worse after sharpening?
Usually the strength is too high, creating halos and amplifying noise in smooth areas. Lower the strength, and judge at 100% zoom and at the size it will actually be shown. Severely blurred photos only get harsher, not sharper.
Can AI unblur a photo better than a normal sharpener?
For visual appeal, often yes. But the detail is generated, not recovered. Use it for keepsakes and social posts, not where exact detail matters.
What is the best first step with a blurry photo?
Use the best copy you have — the original file, not a re-saved or screenshotted version — and try a moderate sharpening pass first. Compare with the original at full size. If the subject is still unrecognisable, look for another shot or the original file.
Try it on your own photo
If your photo is only mildly soft, try the Image Sharpening Tool — set the strength near the middle, compare the result with your original at full size, and keep the original safe. For more quick image fixes, browse the Image Editors collection or all free tools. And if you want to understand why some edits lose quality for good, read Lossy vs Lossless Image Compression: What Actually Changes — it explains the same “detail lost is detail gone” principle from the compression side.