Sharpening filter that subtracts a blurred version from the original — enhances edge contrast locally. More controlled than standard sharpening, results look natural.
You know the problem: After color correction or a resize, the image suddenly appears soft, the edges lose definition. Standard sharpening filters often amplify this digitally and uncontrollably – the result looks artificial, pixelated, with halos all around. This is where Unsharp Masking comes in – a process that is as elegant as it is counter-intuitive: you create an intentionally blurred copy of the original material, subtract it from the original, and thereby achieve edge enhancement without the typical sharpening artifacts.
The mechanics work like this: In the first step, a Gaussian blur is applied to the image – deliberately coarse, often with a radius between 0.5 and 2 pixels. Then, this soft version is subtracted from the original. What remains are the differences – precisely the high-frequency details and edges. These subtracted values are then combined with the original, which locally brightens the edges and slightly darkens the surrounding areas. This creates a three-dimensional, almost tactile effect – not through naive pixel boosting, but through local contrast enhancement.
On set, you notice this primarily in post. If you shot raw material with high ISO or are working with compression, every shot needs individual dosage. A typical Unsharp Mask in DaVinci Resolve or Nuke has three parameters: Radius (how coarse the blurred copy is), Amount (how strong the subtraction is), and Threshold (from which contrast value the mask engages – preventing noise from being overemphasized). With radius values below 1 pixel, you work extremely subtly; at 2–3 pixels, it becomes more dramatic but still looks more organic than a cheap unsharp sharpening plugin. For 4K material, I often use a radius of 1.2 and an amount between 80 and 150 – that's enough to regain resolution without it becoming blocky.
A practical tip: Combine Unsharp Masking with a Curve or Levels adjustment to selectively restrict it to midtones or highlights. This way, you avoid dark areas becoming too grainy. In VFX composites, Unsharp Masking is often used as a final finishing step before export – often as a node with soft falloff edges to seamlessly integrate synthetic elements into live-action material. The method also works excellently when you need to refine greenscreen mattes or after a resize – anywhere you need sharpness without artificiality.