Image Sharpener
Apply a 3×3 unsharp-mask convolution to make a soft photo crisp. Adjustable strength.
Quick answer: Image Sharpener: apply a 3×3 unsharp-mask convolution to make a soft photo crisp. Adjustable strength. Runs entirely in your browser, free, no signup.
Last updated
Good to know
Sharpening works by boosting the contrast at edges that already exist — an unsharp mask compares each pixel to its neighbors and exaggerates the difference — so it can make a slightly soft photo look crisp, but it cannot invent detail the sensor never captured. The best time to use it is after downscaling a large image, since shrinking always blurs edges a little, or to rescue mildly out-of-focus phone shots and screenshots that look soft on high-DPI displays.
Strength is the whole game. A setting of 0.5–1.0 adds natural-looking clarity; push past 2.0 and you start to see haloing — bright and dark fringes tracing every edge — which looks harsh on portraits but is sometimes acceptable on technical diagrams or text where crisp edges matter more than realism. If skin or sky starts to look gritty, you've gone too far; grain is amplified along with edges.
The alpha channel is preserved, so logos and icons keep their transparency after sharpening. But don't over-sharpen images destined for JPG — the compressor will exaggerate the halos into visible artifacts, so sharpen gently or export as PNG.
If the image is genuinely low-resolution rather than just soft, the upscale tool reconstructs detail far better than sharpening ever can. And when a photo is both noisy and soft, run the denoiser first, then sharpen lightly — the same order Lightroom uses.
Frequently asked questions
- When is sharpening helpful?
- After downscaling a large photo, after camera shake, on out-of-focus phone shots and on screenshots that look slightly soft on Retina displays. Don't sharpen images that are already crisp — it just adds grain.
- Will sharpening recover detail that wasn't captured?
- No. Sharpening boosts the edge contrast that's already there; it cannot invent texture. For low-resolution photos, our Upscale Image tool gives better results.
- What strength should I pick?
- Start at 0.5–1.0 for natural-looking sharpening. Above 2.0 you'll see haloing around edges, which is usually too much for portraits but fine for technical diagrams.
- Are my images uploaded?
- No — the convolution runs in your browser via the canvas API. Even big photos stay on your device.
- Does it work on transparent PNGs?
- Yes — the alpha channel is preserved, so logos and icons keep their transparency after sharpening.