Image Sharpen (Unsharp Mask)
Image
Sharpening is the classic 'unsharp mask' algorithm — the same one Photoshop, Lightroom, and GIMP use under the hood. It works by blurring the image, subtracting the blur from the original (which leaves only the edges and fine detail), then adding that edge map back to the original at a controllable strength. The result: edges get crisper without amplifying smooth areas. Three knobs control the effect — amount (how strongly to add the edges back, 0-300%), radius (the blur kernel size in pixels; smaller = fine detail, larger = chunky structure), and threshold (only enhance differences larger than this, useful for skin and noise reduction). Four presets cover most cases: subtle for portraits, standard for general use, strong for screenshots and small text, extreme for heavy stylisation. The before/after split view lets you compare directly. Everything runs in your browser via canvas — the image is never uploaded.
How to use
- Drop an image onto the upload zone or click to choose a file.
- Start with the 'standard' preset, then tune amount and radius. Smaller radius = sharper edges with less halo; larger radius = boost local contrast.
- Raise threshold if you see noise getting amplified (skin pores, JPEG artifacts in flat areas). Threshold 0 sharpens everything; threshold 10-20 spares low-contrast regions.
- Toggle split view to compare before/after side by side.
- Click Download to save the sharpened PNG.
Frequently asked questions
- What does each slider do?
- Amount controls how much of the edge mask is added back (100% means add the edges at full strength; 50% half, 200% double-strength = aggressive). Radius is the blur kernel size used to compute the mask — smaller radius targets fine detail (texture, hair, text), larger radius targets local-contrast boost (gives that 'HDR' or 'pop' look). Threshold suppresses sharpening in flat areas: a pixel difference below threshold is left untouched, which keeps skin and gradients smooth while still sharpening edges.
- Why does heavy sharpening look halo-y?
- Unsharp mask overshoots near edges by design — that's what makes things look sharp. At low strength the overshoot is invisible; at amount >150% with radius >2px you start seeing bright/dark halos along high-contrast borders. To fix: lower amount, lower radius, or raise threshold. Halos are also more visible in flat sky/skin regions, which is where threshold helps most.
- Is this lossless?
- The math runs at 8-bit precision per channel (the canvas storage format), so each sharpening pass can lose a small amount of detail in clipped pixels. One pass on a normal photo has no visible degradation. Multiple passes (downloading, re-uploading, re-sharpening) compound the loss — always work from the original if you can.
- What about full-resolution images?
- The tool downscales anything over 1400px on the longer side before processing, to keep the preview responsive. The downloaded PNG is the displayed size — if you need full resolution, sharpen at the source (RAW developer, Photoshop) or keep your originals under 1400px. For most web use, 1400px is plenty.
- How does this compare to image-blur?
- They're inverse operations conceptually. Blur smooths everything; unsharp mask enhances edge contrast by adding (original - blur) back. You can chain them: blur a photo to soften noise, then sharpen the result with a small radius to bring back edge crispness — a poor man's edge-preserving filter.
Related tools
Image Crop (Aspect Ratio)
Crop any image to a standard aspect ratio (1:1, 16:9, 4:3, etc.) — centered automatically.
Image Rotate & Flip
Rotate an image 90° at a time or flip it horizontally / vertically, then download the result.
Image Resize
Resize an image to any dimensions and export as PNG, JPEG, or WebP — all in your browser.
Image to Base64
Convert an image to a Base64 data URL you can paste anywhere.
QR Code Generator
Create custom QR codes from any text or URL and download them.
Image Compressor
Compress and resize JPEG, PNG, and WebP images in your browser.