↑↓ to move ↵ to open Esc to close Browse all tools

Sharpen Image

Make a slightly soft photo look crisper with an unsharp mask, compare it with the original at real pixel size, and download it in full resolution. Nothing is uploaded.

Sharpen an image

Your photo is sharpened by your own browser. Nothing is uploaded.

Free printableVision Board TemplateA grid of 6 to 12 tiles for your own photos and words, printed as a PDF. Hand-checked listFree Stock Photo SitesPhoto sites and open archives where you can download images for free, with the license terms checked on each site.

Sharpen Image by Freebiefy, free online tools · Privacy · Legal notice

Embed this tool on your site

Paste this code into your page. The tool runs in the visitor's browser, free and without signup. The height adjusts to the content. All tools you can embed

What sharpening does, and what it cannot do

A photo looks sharp when its edges change quickly from dark to light. In a soft photo the same edge is spread over a few pixels. Sharpening makes that change steeper again: the dark side of an edge gets a bit darker and the light side a bit lighter. The eye reads the extra contrast as crispness, and that is what this image sharpener adds.

That is all it does. Sharpening works with the edges that are already in the picture and cannot add detail the camera never recorded. A face that is a smooth blur stays a smooth blur, and text that is unreadable stays unreadable. Recovering lost detail is a different and much harder problem, and no setting here claims to do it. What sharpening does well is fix the slight softness of most digital photos, and of every photo that was just made smaller.

How an unsharp mask works

The method is older than digital photos. In the darkroom, a blurred positive copy was made from the negative and printed together with it, and that blurred "unsharp" copy gave the method its name. This tool does the same with numbers, in three steps for every pixel:

  1. Blur a copy of the photo with a Gaussian blur of the chosen radius.
  2. Subtract the blurred copy from the original. The difference is close to zero in flat areas and large along edges.
  3. Add the difference back, multiplied by the amount: new value = original + amount × (original − blurred). Where the difference is not larger than the threshold, the pixel stays as it is.

Results below 0 or above 255 are clipped, so nothing wraps around. The blur is done in two passes, along rows and then along columns, which gives the same result as a full two-dimensional Gaussian with far less work. Transparency is never changed.

A worked example

Take one row of six gray pixels with an edge in the middle, and for this example a small blur that averages each pixel with its neighbors in the ratio 1 : 2 : 1:

Pixel123456
Original505050200200200
Blurred505087.5162.5200200
Difference00−37.537.500
Amount 100%505013238200200
Amount 200%50500255200200

Pixel 3 is (50 + 2 × 50 + 200) ÷ 4 = 87.5 after the blur, so its difference is 50 − 87.5 = −37.5, and at 100% it drops to 12.5, rounded to 13. The step from 50 to 200 is now a step from 13 to 238, which looks sharper. At 200% the values would be −25 and 275, so they are clipped to 0 and 255. A threshold of 40 would leave this edge alone, because a difference of 37.5 is not larger than 40.

Amount, radius and threshold

  • Amount sets how strong the effect is. Values from 50% to 150% are a common range for photos.
  • Radius is the standard deviation of the Gaussian blur in pixels. It decides how wide the edges are that get sharpened. Values from 0.5 to 2 pixels are common; larger values give a broader, more contrasty look and wider halos.
  • Threshold is the smallest difference, in levels from 0 to 255, that gets sharpened. Raising it protects smooth areas such as skin, sky and water, and keeps fine noise from turning into grit.

The presets are starting points chosen for this tool: Subtle (50%, 0.8 px, 3), Standard (100%, 1.0 px, 2), Strong (150%, 1.6 px, 1) and For web after resizing (80%, 0.5 px, 0).

Halos and noise

The dark and light bands that sharpening puts along an edge are what makes it look crisp, but past a point they become visible as halos: a light outline around a dark roof line, or a dark rim around a bright window. A high amount makes them stronger, a large radius makes them wider. Sharpening also strengthens noise and the block edges of a heavily compressed JPEG, because to the filter they look like edges too. Lightness only avoids colored fringes by sharpening brightness, not the three color channels on their own. Always judge the result at 100% or 200%, which is why the preview shows real pixels instead of a shrunken view.

When to sharpen

Sharpen at the final size, as the last step before saving. Resizing changes how many pixels an edge covers, so sharpening before it is lost or blown up into wide halos. A good order is: crop, adjust, resize with the image resizer, sharpen here, then shrink the file with the image compressor. To soften a picture or hide parts of it, the opposite of this tool, use blur image, and for a gray version, the black and white image converter.

Large photos and saving

The preview sharpens only the part you see, so it reacts at once. The download sharpens every pixel of the full image in the background and shows its progress, and you can cancel it. A photo of 12 megapixels takes a few seconds on a laptop and longer on a phone. Images larger than the browser can draw, for example above 16.7 megapixels in Safari on iPhones before iOS 18, are scaled down, and the tool says so. JPEG has no transparency, so transparent areas turn white; PNG keeps them.

How to sharpen an image

  1. 1Drop a photo on the box or tap it to choose one. HEIC photos from an iPhone work too.
  2. 2Pick a preset, then adjust amount, radius and threshold.
  3. 3Tap the small overview to look at an important part, such as eyes or text, and drag across the picture to compare.
  4. 4Choose JPEG or PNG and press Download image.

Sources

Frequently asked questions

Is my photo uploaded?

No. Your browser opens the photo, sharpens it and saves the new file on your own device. Nothing is sent to a server.

Can this fix a blurry photo?

Only a slightly soft one. Sharpening raises the contrast along edges that are already in the picture, so it looks crisper. It cannot bring back detail the camera never recorded. A photo that is badly out of focus or smeared by motion stays blurry, and strong settings then only add halos and noise. No setting here invents new detail.

What do amount, radius and threshold do?

Amount is how strong the effect is: at 100% the difference between the photo and its blurred copy is added once. Radius is the size of the blur in pixels, so it sets how wide the edges are that get sharpened and how wide any halo becomes. Threshold is the smallest difference, in levels from 0 to 255, that gets sharpened at all, so smooth areas such as skin and sky stay smooth.

Why do I see light and dark lines along edges?

Those are halos. Unsharp masking darkens the dark side of an edge and lightens the light side. In small doses that reads as sharpness; with a high amount or a large radius the bands become visible. Lower the amount or the radius, and check the result at 100% or 200% zoom.

Should I sharpen before or after resizing?

After. Resizing changes how many pixels an edge covers, so sharpening done before it is either lost when the photo gets smaller or enlarged into wide halos when it gets bigger. Resize first, sharpen last, then save. The For web after resizing preset uses a small radius made for photos that were just made smaller.

What is the difference between Lightness only and Each color channel?

Lightness only works out one difference from the brightness of each pixel and adds it to red, green and blue alike, so edges get crisper without colored fringes. Each color channel sharpens red, green and blue on their own, which can be a little crisper on colored edges but can add color fringes and color noise.

Does it keep the full resolution?

Yes. The preview shows a part of the photo at 100% or 200%, so you see real pixels, and the download sharpens the whole image at its full width and height. Very large images are scaled down only when the browser cannot draw them at full size, and the tool tells you when that happens.