← Products / VantaClear
Product 01 — Algorithm

VantaClear

An open-source sharpening algorithm that finds real edges before it sharpens them — so detail comes up without the pale halo that naive sharpening leaves behind.

License
MIT
Core
Python
Formats
PNG · JPEG · TIFF
Status
Available now
Demo

See the difference.

Drag the handle to compare input and output. Both use the same resolution test chart — the standard instrument for measuring sharpening performance.

Before
After — VantaClear
How it works

Three passes, in order.

Each step depends on the output of the one before it. The sequence is what makes the result artifact-free.

01

Find the edge

VantaClear separates real edges from sensor noise before touching a single pixel, so it never mistakes grain for detail.

02

Sharpen locally

Each region gets its own kernel strength, scaled to its local contrast — flat skies stay flat, fine texture gets pushed harder.

03

Clamp the overshoot

The result is capped against its neighborhood, which is what removes the pale halo that classic unsharp masking leaves along strong edges.

Built in the open, from the first release.

VantaClear is free to use, fork, and audit. Contributions welcome.