Image Sharpen Tool
A real unsharp mask with a threshold control — so you sharpen the edges without amplifying the noise.
How strongly edges are boosted.
How wide an edge is considered. Keep it small.
Ignore differences smaller than this — protects smooth areas from noise.
Original
Sharpened
What the image Sharpen Tool does
Sharpening cannot add detail that was never captured. What it does is increase local contrast along edges, which your visual system reads as sharpness. The classic technique, and the one used here, is called unsharp masking: blur a copy of the image, subtract it from the original to isolate the edges, then add that difference back at an adjustable strength.
Three controls govern it, and they interact. Amount is how much of the edge difference is added back. Radius is how wide a transition counts as an edge — small values (around 1 pixel) sharpen fine texture, larger values affect broader transitions and quickly start to look artificial. Threshold is the one most tools hide and the one that matters most: differences smaller than the threshold are left alone entirely.
That threshold is what separates sharpening from noise amplification. A photograph's smooth areas — sky, skin, shadow — are not perfectly flat; they contain low-level sensor noise. With the threshold at zero, that noise is an “edge” as far as the algorithm is concerned, and sharpening makes it grainy. Raising the threshold to around 5–10 protects those areas while still sharpening genuine detail.
The failure mode to watch for is haloing: a bright line on the light side of a high-contrast edge and a dark line on the other. Once you can see it, it cannot be removed, and it is the clearest signal that an image has been over-processed. It appears first along roof lines against sky. If you can see halos in the preview, reduce the amount or the radius.
How to use it
- Load a soft image. The preview shows original and sharpened side by side at up to 900 pixels.
- Start with amount around 80%, radius 1.0–1.5 and threshold 5–10.
- Raise the amount until detail looks crisp, then look at a high-contrast edge for halos and back off.
- Download — the full-size image is sharpened at the same settings, not the preview.
A worked example
Typical starting points by subject:
| Subject | Amount | Radius | Threshold |
|---|---|---|---|
| Landscape, fine detail | 90% | 0.8 | 3 |
| Portrait | 50% | 1.2 | 10 |
| Screenshot or text | 120% | 0.6 | 0 |
| High-ISO night photo | 60% | 1.0 | 15 |
The portrait and night rows show the threshold doing its job. A portrait sharpened at threshold 0 emphasises every pore and blemish; at 10 it sharpens eyes and hair while leaving skin smooth. A high-ISO photo is mostly noise in the shadows, so the threshold has to go higher still. Screenshots are the opposite case: no noise at all, hard edges everywhere, so threshold 0 is correct.
Frequently asked questions
Can sharpening recover a blurred or out-of-focus photo?
No. Sharpening increases contrast at edges that already exist; it cannot reconstruct detail the lens never resolved. A genuinely out-of-focus photo will just look crunchy. Camera-shake blur is sometimes partly recoverable, but not by unsharp masking.
What does the threshold setting actually do?
It tells the algorithm to ignore differences below a certain size. Since sensor noise in smooth areas is a small difference, a threshold of 5–10 protects sky and skin from being sharpened into grain while leaving real edges untouched.
What are halos and how do I avoid them?
A bright fringe on one side of a high-contrast edge and a dark one on the other — most visible where a roof meets sky. They come from too much amount or too large a radius. They cannot be removed afterwards, so check the preview before downloading.
Should I sharpen before or after resizing?
After. Downscaling softens an image, so sharpening first and then resizing throws away the work. Resize to your final dimensions, then sharpen as the last step before saving.