Sharpen Image Online
A real unsharp mask with separate amount, radius and threshold, so you can sharpen genuine detail while leaving skies, skin and noise completely untouched. Live preview, drag-to-compare, batch and ZIP — and nothing ever leaves your browser.
Drop images here
or click to browse · or paste with Ctrl+V
JPG · PNG · WebP · GIF · BMP · AVIF
How strongly detected detail is amplified.
Areas varying by less than 3 levels are left completely untouched.
Presets
Your images are never uploaded. The mask is computed inside this browser tab.
People also use
Everything this tool does
A real unsharp mask
A gaussian blur is subtracted from the original to isolate detail, then that difference is added back amplified. The same method professional software uses, not a fixed convolution kernel.
Threshold that truly protects
Pixels differing from their surroundings by less than the threshold are skipped entirely rather than sharpened weakly, so a clear sky stays exactly as it was.
Independent radius
Radius decides what counts as detail. Small values sharpen texture; large values with a low amount produce local contrast, which is a genuinely different effect.
Separable blur
The two-dimensional gaussian is split into a horizontal and a vertical pass, which at radius 8 cuts the work per pixel from 289 operations to 34.
Drag-to-compare preview
A split handle wipes between original and sharpened, which is the only reliable way to judge an effect your eye adapts to within seconds.
Purpose-built presets
Separate starting points for web resizing, portraits, landscapes, text scans, soft focus and clarity, because those genuinely need different settings.
Luminance-based detection
Detail is measured on perceived brightness rather than a single channel, so a coloured edge is detected as reliably as a grey one.
Batch with one ZIP
Apply one setting across a whole shoot at the same output size and download everything as a single archive.
Transparency preserved
The alpha channel is untouched for PNG and WebP output, so only visible pixels are affected.
Nothing is uploaded
No server, no queue, nothing stored to delete afterwards. Close the tab and every trace goes with it.
Settings by purpose
Sharpening is output-dependent, so there is no universal setting. These are the starting points that work for each job, with the reasoning behind the threshold choice in particular.
| Purpose | Amount | Radius | Threshold | Why |
|---|---|---|---|---|
| After resizing for web | 70-90 | 1 | 3 | Downscaling softens everything; this restores what averaging removed. |
| Portraits | 80-100 | 2 | 10-15 | High threshold keeps skin texture and pores from being amplified. |
| Landscapes | 110-130 | 1 | 5-8 | Fine detail benefits, and the threshold keeps skies free of noise. |
| Text and document scans | 140-170 | 1 | 0 | Hard black-on-white edges, and there is no smooth area to protect. |
| Soft-focus rescue | 160-200 | 3 | 4 | A wider radius catches the broader softness, though focus misses cannot be fixed. |
| Clarity / local contrast | 40-60 | 6-8 | 2 | Large radius with low amount reads as depth rather than as sharpening. |
| Already noisy high-ISO | 50-70 | 1 | 15-20 | Sharpen only real edges; denoise first for a better result. |
3
independent controls
300%
max amount
0
files uploaded
8 px
max radius
ZIP
batch download
Free
no account needed
FlipMyFormat vs other sharpen tools
| Capability | FlipMyFormat | Typical free tool |
|---|---|---|
| Files stay in your browser | ||
| Adjustable radius, not a fixed kernel | ||
| Threshold to protect smooth areas | ||
| Local contrast (clarity) mode | ||
| Drag-to-compare before and after | ||
| Batch ZIP download without paying | ||
| No watermark on the output |
How to sharpen without halos
Two facts explain nearly every bad sharpening result. The first is that sharpening cannot create detail, it can only increase contrast at edges that are already recorded, which means a photo that missed focus is not a candidate for this tool at all — pushing the amount up on a soft image amplifies the noise and the compression blocks instead, and the halos arrive long before the sharpness does. What sharpening genuinely fixes is the mild softness that comes from a lens, from a camera's own processing, and above all from downscaling, because averaging pixels together to make an image smaller destroys exactly the fine transitions that read as crispness. That is also why order matters: resize to your final dimensions first and sharpen afterwards, because sharpening before a downscale simply feeds your new detail into the averaging. The second fact is that a photograph contains two kinds of variation and only one of them should be amplified. The edge of a branch against sky is detail; the faint mottling within the sky itself is noise and sensor grain, and a sharpening tool with no threshold has no way to tell them apart, so it makes both louder and your clean sky turns speckly. Threshold is the control that separates them, and it is worth being generous with it on anything containing skin or sky — a value around 10 to 15 leaves those areas genuinely untouched rather than lightly sharpened. Finally, judge the result at 100 percent on a high-contrast edge, not on the scaled preview: oversharpening looks appealing when an image is displayed small, which is precisely the trap.
- 1
Resize first, then add
Get the image to its final output size before sharpening, because a later downscale throws the work away.
- 2
Set radius, amount, threshold
Start from a preset. Radius sets the scale, amount sets the strength, threshold decides what is left alone.
- 3
Check for halos, then save
Look at a high-contrast edge for light or dark rims. Back off if you see them, then download or take the batch as one ZIP.
How the mask works
Subtracting a blur to find detail
Blurring an image removes fine detail while leaving broad shapes intact, so the difference between the original and its blurred copy is precisely the detail that was removed. Adding that difference back, multiplied by your amount, exaggerates every transition: the bright side of an edge gets brighter and the dark side darker, which your eye reads as sharpness. Radius controls how much blur is used, and therefore what scale of feature counts as detail rather than as shape.
What threshold does differently
Before any pixel is changed, its brightness difference from the blurred version is measured. If that difference is smaller than the threshold, the pixel is skipped entirely — not sharpened by a reduced amount, but left exactly as it was. That distinction matters: a scaled-down effect still amplifies noise, just less. Skipping means a flat sky comes through the operation bit for bit identical, which is the behaviour you actually want.
What people use it for
After resizing for the web
The single most common and most justified use. Downscaling softens; a light sharpen puts the crispness back.
Product photography
Fabric weave, stitching and surface texture read much better with a modest sharpen at a small radius.
Document and receipt scans
Hard edges and no smooth areas mean you can sharpen aggressively with threshold at zero.
Screenshots for documentation
Text and UI edges tighten up noticeably, which matters when a screenshot is displayed smaller than captured.
Photos from older phones
Heavy in-camera noise reduction leaves images soft. A small radius with a high threshold recovers detail without the noise.
Landscape clarity
A large radius with a low amount adds depth and separation without looking like a sharpening filter.
Preparing images for print
Ink spread on paper softens fine detail, so print files usually need slightly more sharpening than screen files.
Video thumbnails
Crisp edges survive platform re-compression better and stand out more at small sizes.
Six tips worth knowing
Always resize before sharpening. Sharpening then downscaling wastes the work and can leave artefacts that the resize multiplies.
Raise threshold to 10 or more on anything with skin or sky. Those areas contain noise, not detail, and sharpening them is always a loss.
Judge halos at 100 percent zoom on a high-contrast edge. Oversharpening genuinely looks good on a small preview, which is the trap.
Denoise before sharpening, never after. Sharpening noise makes it permanent and much harder to remove.
For local contrast rather than crispness, use radius 6 to 8 with amount around 50. It looks like depth, not like a filter.
If you need more than about 200 percent amount to see a difference, the detail is not in the file. Sharpening cannot recover missed focus.
Troubleshooting
White outlines around dark objects
Classic halos from too much amount or too large a radius. Reduce amount first, then radius. Halos cannot be removed once saved.
The sky went speckly
Threshold is too low, so sensor noise is being amplified along with real detail. Raise it to 10 or above.
Nothing seems to change
Either amount is very low, or your threshold is high enough to skip everything. Set threshold to 0 briefly to confirm the tool is working.
The photo still looks blurry
If focus was missed, the detail was never recorded and sharpening cannot invent it. That is a limitation of the physics, not the tool.
Frequently asked questions
How do I sharpen an image online?
Drop your photo onto the tool and set amount, radius and threshold. The preview updates as you drag and the compare handle wipes between original and sharpened. Press Apply and save. Nothing is uploaded — the whole operation runs in your browser.
What is unsharp mask and why is it called that?
The name comes from darkroom practice: a blurred (unsharp) copy of a negative was used as a mask to increase local contrast in the print. Digitally it does the same thing. The tool makes a blurred copy, subtracts it from the original to find where detail lives, and adds that difference back amplified. Confusingly, it is the standard method for sharpening despite the name.
What do amount, radius and threshold each control?
Amount is how strongly the detected detail is amplified. Radius is how far from an edge the effect reaches, which effectively sets what counts as detail rather than as shape. Threshold is the minimum difference a pixel must show before it is touched at all, so it decides what gets left alone. Amount is the volume, radius is the scale, threshold is the gate.
What does threshold actually protect?
Smooth areas. A clear sky or a patch of skin contains tiny variations that are noise rather than detail, and sharpening them makes the noise louder. In our own testing, a lightly noisy flat area sharpened at threshold zero changed measurably, while the same area at threshold 8 was left completely untouched — not reduced, untouched. That is why threshold matters more than most people realise.
Can sharpening recover a genuinely out-of-focus photo?
No, and it is worth being direct about that. Sharpening increases contrast at edges that already exist in the file. If focus missed, the edge information was never recorded and no amount of amplification invents it — you get halos and amplified noise instead of a sharp picture. What sharpening does fix is the mild softness introduced by resizing, by a lens, or by a camera's own processing.
Why do I see white or dark outlines around objects?
Those are halos, and they mean amount or radius is too high. The mask has pushed pixels on the bright side of an edge past white and the dark side past black, leaving a visible rim. Reduce amount first, then radius. Halos are the single clearest sign of oversharpening and they cannot be removed later.
What radius should I use?
For most photographs at full resolution, 1 to 2 pixels. Small radii sharpen fine texture; larger radii affect broader shapes and start to look like a contrast boost rather than sharpening. A large radius with a low amount — try the clarity preset — produces local contrast enhancement, which is a genuinely different and often more flattering effect.
Should I sharpen before or after resizing?
After, always. Downscaling averages pixels together and softens everything, so sharpening beforehand is thrown away and can leave artefacts that the resize then multiplies. Resize to your final dimensions first, then sharpen for that size. An image sharpened for print will look overdone on the web.
Are my photos uploaded to a server?
No. The blur and the mask are computed in your browser using the canvas API. Nothing is sent anywhere, and the tool keeps working if your connection drops after the page has loaded.
Why is it fast even with a large radius?
Because the blur is separable. A two-dimensional gaussian can be split into a horizontal pass followed by a vertical one, which turns the cost per pixel from radius squared into roughly twice the radius. At radius 8 that is the difference between 289 operations per pixel and 34.
Does sharpening make the file larger?
Usually yes. Compression works by finding smooth, predictable regions, and sharpening deliberately adds high-frequency variation. Expect a modest increase in JPEG and WebP output. Raising the quality setting a little often produces a better result than letting the encoder smooth your new detail back away.
Can I sharpen a batch of photos at once?
Yes. Add as many as you like and identical settings are applied to every one, which suits a set shot with the same lens at the same size. Download individually or take the whole batch as a single ZIP.
Which file formats are supported?
Any format your browser can decode: JPG, PNG, WebP, GIF (first frame), BMP, and AVIF or HEIC where supported. You can keep the original format or export as PNG, JPG or WebP.
Is transparency preserved?
Yes for PNG and WebP output. The alpha channel is untouched and only visible pixels are sharpened. JPG has no alpha, so transparent areas are filled with the colour you choose first.
How much sharpening is too much?
Judge by halos, not by how crisp it looks. Zoom to 100 percent and look at a high-contrast edge such as a branch against sky. If you can see a light rim on one side and a dark rim on the other, back off. On screen at reduced size oversharpening often looks appealing, which is exactly why people overdo it.
Does sharpening reduce image quality?
It amplifies whatever is already there, including noise and JPEG artefacts, and any pixel pushed past 0 or 255 is clipped permanently. Starting from a high-quality original and using threshold to protect flat areas both keep those costs low.
Does this remove EXIF data?
Yes. Re-encoding from the canvas drops camera model, GPS coordinates, timestamps and other metadata. If you want to strip metadata without changing the picture, use the remove EXIF tool instead.
Does the tool add a watermark or need an account?
No to both. No sign-up, no email wall, no daily limit and no watermark. The output is your image with the mask applied and nothing else added.