Jarvis–Judice–Ninke dithering online

About Jarvis–Judice–Ninke dithering

Jarvis, Judice and Ninke published this algorithm in 1976, the same year as Floyd–Steinberg. Instead of four neighbours it spreads error across twelve pixels over three rows.

The wider kernel produces smoother gradients and far fewer directional artifacts, at the cost of slightly softer detail and more computation. Use it when your image has large skies, skin tones or other soft transitions.

Diffusion kernel

          *   7   5
  3   5   7   5   3
  1   3   5   3   1   (÷48)

“*” marks the current pixel; numbers show how much of its quantization error each neighbour receives.

Best for

• Smooth gradients and skies
• Large prints
• Reducing worm artifacts

How to dither an image

1. Add your image. Click Upload image, drag a file onto the preview, or paste with Ctrl/⌘+V. JPG, PNG, WebP and GIF work.
2. Pick an algorithm and palette. Error diffusion (Floyd–Steinberg, Atkinson…) for photos, Bayer for a retro crosshatch.
3. Tune it. Raise pixel size for chunkier pixels; adjust brightness and contrast to control which tones become dark or light.
4. Download. Export a lossless PNG at up to 8× with crisp nearest-neighbour scaling.

FAQ