Three problems that look the same in a thumbnail
Every result for this is a tool, and every one of them makes the same promise: upload the picture and the blur comes off. The promise is not quite a lie, because some blur does come off. It is that the word blur is being used for three different things, and only some of them are reversible.
On a phone screen they are indistinguishable. Enlarged, they are not, and telling them apart takes about ten seconds and decides everything that follows. It also tells you when to stop paying for tools, which is the part the tool pages have no reason to mention.
The three things a soft picture can be suffering from
Each of these has a different cause and leaves a different fingerprint in the file, which is why they respond so differently to the same software.
- Movement. The camera or the subject shifted while the shutter was open, so every detail is smeared along the direction of travel. The detail was recorded, just spread across neighbouring pixels in a consistent direction.
- Missed focus. The lens was sharp on the wrong plane, so every point of light landed as a small disc rather than a point. Again the light arrived, and again it was distributed rather than lost, though this time evenly in all directions.
- Not enough pixels. Nothing was smeared. The picture was recorded at a size too small to hold the detail, or saved down until it was, and there is no fine detail anywhere in the file to redistribute. This is not blur at all, it only looks like it once the picture is enlarged.
How to tell which one you have
Open the file at full size rather than looking at the preview, and find a hard edge or a small bright point somewhere in the frame. A streetlight, a highlight on an eye, the corner of a window frame. What that point has turned into is the answer.
A short streak with a direction is movement, and every streak in the picture will lie along the same axis. A soft round disc with no direction is focus. Blocky steps that follow the pixel grid, or edges that look like a staircase, mean the file has run out of pixels rather than run out of sharpness.
One more check settles the focus case. Look for anything in the frame that is sharp. If some plane of the picture is crisp and your subject is not, the lens was focused, just not where you wanted, and the softness is genuine focus blur rather than anything to do with size.
What a model can put back, and where it stops
Movement and mild focus miss are the cases that respond. The detail is present in the file in a scrambled form, so a model that has learned what unscrambled versions of that kind of detail look like has something real to work from. Recovery is genuine here in the sense that the answer is constrained by what was recorded.
The constraint weakens fast as the softness increases. A slight camera shake has one plausible sharp version and the model will find it. A long exposure of somebody walking has an enormous number of plausible sharp versions, and what you get back is one of them chosen for looking convincing, not for being right.
The third case is a different job entirely. A picture with too few pixels is not being unblurred, it is being enlarged and drawn into, and where that guessing is safe and where it is not is the whole subject of the resolution guide rather than of this one.
Sharpening is not unblurring, and knowing that saves an afternoon
The slider in every photo app labelled sharpen does one thing: it increases contrast on either side of edges that already exist. Dark side darker, light side lighter. The picture reads as crisper because your eye takes edge contrast as a proxy for detail, and no detail was added.
It is genuinely useful on a picture that is already close, and it has a hard ceiling. Push it past that and you get bright outlines around every object, a grainy look in flat areas like sky and skin, and the original softness still sitting underneath all of it. If a sharpening slider has not fixed your picture by the time halos appear, the problem is not one sharpening can reach.
The cases where the answer is to go somewhere else
Some pictures do not come back, and the useful move is to stop working on the file and look for a better source of the same moment. This is unglamorous and it beats every tool.
- Look for another frame. Phones keep more than the one you kept, and burst sequences and live captures usually hold a sharper neighbour a fraction of a second away.
- Look for the original rather than a copy. A picture sent through a messaging app has been resized on the way, and the version still on the original device may have several times the detail with nothing else changed.
- Look for the physical object. If a print, a slide or a negative exists, it holds detail no phone snapshot of it has captured, and photographing or scanning it properly is the only route to detail that is really there.
- Ask what the picture has to survive. A face that has to identify somebody, or a sign that has to be read, is the case where an invented sharp answer is worse than an honest soft one, because a confident wrong face is believed and a blur is not.
Keep the original, whatever you do next
Every pass through any of these tools produces a version, and a version is a new file with decisions baked into it that cannot be taken back out. The soft original is the only thing that still holds the option of trying something else later, and better tools arrive.
Work on a copy, keep the copy beside the original rather than on top of it, and treat the enhanced file as one interpretation of the moment rather than as a replacement for the record of it.
More guides
Questions
Can a blurry photo really be unblurred?
Sometimes, and it depends entirely on why it is blurry. Camera shake and a slight focus miss can be substantially undone, because the detail is in the file in a scrambled form. A picture that was simply captured too small has no fine detail to recover, so anything sharp you get back was drawn rather than found.
Why does my photo still look soft after using an unblur tool?
Usually because the problem was pixel count rather than blur. Enlarging a small picture smooths between the pixels you already had and adds nothing, so the softness comes along at a larger size. Look at a hard edge at full size: staircase steps mean size, a directional streak means movement.
Does sharpening in a photo app fix a blurry picture?
It fixes a picture that is nearly there. Sharpening raises contrast at existing edges rather than recovering detail, so it helps slightly soft images and does nothing for badly blurred ones except add bright outlines and grain.