A real unsharp mask with Amount, Radius and Threshold. Luminance-only sharpening, zoom to inspect for halos, presets and batch mode. Nothing leaves your device.
Unsharp Mask
Amount, Radius, Threshold
No Fringing
Luminance-only Mode
Zoom + Pan
Inspect for Halos
Private
No Uploads
Drop your image here
JPG, PNG, WebP, GIF, BMP, AVIF, HEIC or TIFF
Your file never leaves this device
The same three controls used in professional photo editing software, built to actually let you dial in the result.
Amount, Radius, Threshold
Independent control over sharpening strength, how wide an edge it affects, and how much contrast counts as an edge at all.
Luminance-only sharpening
Sharpens brightness without touching color, avoiding the faint color fringing per-channel sharpening can cause at edges.
Zoom and pan to inspect
Halos are often invisible at normal size. Zoom in and drag to check any part of the image at true detail before downloading.
Content-specific presets
Portrait, Landscape, Text and Screenshot, and more, each tuned to how that kind of image actually responds to sharpening.
Before and after toggle
Switch instantly between the original and the sharpened version on the same preview to judge the effect honestly.
Batch mode
Drop an entire shoot, dial in one set of settings, and export every image sharpened as a single ZIP download.
Recover a slightly soft phone photo
A photo that is just barely out of perfect focus, or was smoothed by in-camera noise reduction, often reads as noticeably crisper after a moderate sharpen.
Sharpen after resizing down
Shrinking an image softens fine detail. Sharpening should be the last step in an edit, applied after resizing, not before.
Clean up text and screenshots
Screenshots compressed for sharing often lose crisp text edges. The Text and Screenshot preset restores legibility without a halo.
Finish a product photo
A light, precise sharpen is a standard last step for e-commerce photography, making edges and texture read clearly at thumbnail size.
Prepare an image for print
Print output generally wants a stronger sharpen than the same image would need on screen, since printing itself softens fine detail.
Amount
How strong the effect is. Higher values make edges more pronounced, and eventually visible as halos.
Radius
How many pixels out from an edge the effect reaches. Small for fine detail like hair or text, larger for broad shapes like a landscape.
Threshold
How much local contrast is needed before a pixel counts as an edge at all. Higher values protect smooth areas like sky or skin from picking up noise.
An unsharp mask does not detect edges the way a person would. It creates a deliberately blurred copy of the image, then subtracts that blurred copy from the original. Wherever the image was already smooth, the original and the blurred version are nearly identical, so the subtraction produces almost nothing. Wherever there was a real edge, the blur smeared it out, so the original and blurred versions differ sharply right at that boundary, and the subtraction isolates exactly that boundary. That isolated difference is then scaled by the Amount and added back onto the original, which is why it is called an unsharp mask, the mask itself comes from a blur.
This is also exactly why sharpening can create a visible halo. The technique works by increasing contrast right at an edge, a thin dark line just outside the edge and a thin bright line just inside it, or the reverse depending on which side is lighter. At a reasonable Amount this reads as crispness. Push it too far, or set the Radius wider than the detail actually calls for, and those thin lines become visible on their own, especially against a plain background like sky or a wall. Halos are frequently only visible once you zoom in past normal viewing size, which is the entire reason this tool includes real zoom and pan on the preview rather than only a fit-to-screen view.
Resizing an image down averages neighboring pixels together, which softens fine detail as a direct side effect, even with a good resizing algorithm. If an image is sharpened before being resized, a lot of that sharpening gets smoothed right back out during the resize, wasting the effort and sometimes leaving behind faint sharpening artifacts that the resize did not fully erase. Doing it the other way, resizing first with the resize tool and sharpening last, means the sharpening is calculated against the actual final pixel grid the image will be viewed at, which is what the edge detection in an unsharp mask is built to respond to.
Sharpening each color channel independently can shift color slightly right at an edge, since red, green and blue rarely have identical contrast at exactly the same location. The fix used by most professional software is to convert to a luminance and color space, sharpen only the brightness channel, and leave color untouched. This tool reaches the same result more directly: the sharpening amount is calculated from luminance, and that same amount is added equally to red, green and blue, which produces an identical result to the full conversion without needing the extra step. It is on by default because there is rarely a reason to prefer the alternative.
No. Sharpening cannot recover detail that was never captured, it works by increasing the contrast right along existing edges, which is what makes an image read as crisp to the eye. A genuinely out of focus or heavily compressed photo will not become sharp, but a slightly soft photo, or one that lost a bit of edge contrast during resizing, often looks noticeably better.
That is usually a halo, a thin bright or dark outline that unsharp masking can create along strong edges when Amount or Radius is pushed too high. Halos are frequently invisible at normal viewing size and only show up once you are zoomed in close, which is exactly why this tool includes zoom and pan on the preview. If you see fringing at high zoom, lower the Amount or Radius slightly until it disappears.
Adjust Image's Sharpen slider is a fixed convolution kernel with a single intensity value, useful for a quick touch up. This tool is a full unsharp mask with independent control over Amount, Radius and Threshold, which is the same technique and the same three controls used in professional editing software. Use Adjust Image for a fast tweak alongside other adjustments, and this tool when the sharpening itself needs to be dialed in precisely.
Text and screenshots respond differently to sharpening than photos, since their edges are already hard and high contrast. A small Radius with a low Threshold tends to work best, which is exactly what the Text and Screenshot preset is tuned for. A large Radius on text usually looks worse, not better, since it exaggerates a much wider halo around every letter.
With it on, the sharpening amount is calculated from brightness alone and applied equally to red, green and blue, which leaves the underlying color completely unchanged and avoids the faint colored fringing that can appear at high-contrast edges when each color channel is sharpened on its own. It is on by default because it is very rarely the wrong choice, but turning it off is available for cases where per-channel sharpening is specifically wanted.
No. The unsharp mask, the preview, and the final export all run entirely in your browser using the Canvas API. Your image never leaves your device. You can disconnect from the internet after loading the page and the tool still works completely.
