Image Dither
Turn an image into a black-and-white dot pattern with Floyd–Steinberg or Bayer dithering. Compare the preview and save a full-size PNG locally.
This tool processes its input contents in your browser instead of sending them to a processing service. Page, asset and fixed usage-measurement requests are described on the Privacy page.
Image Dither turns a JPEG or still PNG into a black-and-white dot pattern using Floyd–Steinberg or ordered Bayer and saves a full-size PNG with alpha.
How to use it
- 1Choose a local JPEG or still PNG.
- 2Compare the original and dithered previews. Change the dot pattern if needed.
- 3Download the dithered PNG.
Example
- Input
- Four opaque gray pixels with RGB values 128, 128, 128 and 128. Select Floyd–Steinberg.
- Output
- A 4 × 1 PNG with white, black, white and black pixels.
Diffusing the quantization error lets nearby black and white dots represent mid-gray.
What happens to your data
Image decoding, dithering and PNG export run in your browser. The input image, filename and export are not sent to a processing server.
Last updated November 2026
Turn a photo, illustration or gradient into a two-tone pattern while keeping the decoded image dimensions. Choose a JPEG or still PNG to get an automatic Floyd–Steinberg preview beside the original. The alternate Bayer pattern produces a regular repeating grid. Download a full-size PNG directly from the result.
Dithering uses the distribution of light and dark dots to suggest shades between black and white. It can help when preparing retro artwork, making mockups for monochrome displays or experimenting with bold print graphics. A hard threshold is more suitable for sharply separating a flat background and a dark subject.
How it works
The tool computes a brightness value from the decoded sRGB red, green and blue components using weights of 0.2126, 0.7152 and 0.0722. Floyd–Steinberg scans left to right and distributes the difference between the current brightness and its black or white output to four future neighbors, with weights of 7/16, 3/16, 5/16 and 1/16. Errors at the outside boundary are discarded. The ordered method looks up a position-dependent threshold from a repeating 4 × 4 Bayer matrix.
The exported visible pixels use black or white and keep the original decoded alpha. Fully transparent pixels stay transparent and do not diffuse their invisible color. The smaller preview can blend dots when scaled down, so inspect the original-size PNG for exact pixels. The export keeps dimensions but does not carry over source metadata, ICC profiles or compression settings. It is a regular PNG with two visible colors and alpha, so check the file format required by your printer or display.
Use one still JPEG or a supported PNG through 8-bit, up to 20 MiB, 16 million pixels and 8,192 pixels per side. A file with animation or an unsupported container is rejected. Decoding and PNG encoding occur in the browser. Clear or choose another file to remove the previous preview and invalidate its download.
Worked example: Dither a four-pixel gray strip
- Input
- Four opaque gray pixels with RGB values 128, 128, 128 and 128. Select Floyd–Steinberg.
- Result
- A 4 × 1 PNG with white, black, white and black pixels.
Diffusing the quantization error lets nearby black and white dots represent mid-gray.
Common use cases
- Make retro black-and-white illustrations
- Test a dot pattern for a monochrome display
- Convert an image for a limited-color design prototype
- Compare regular Bayer dots against a scattered error-diffusion pattern
Frequently asked questions
Which image dither method should I choose?＋
Floyd–Steinberg spreads brightness error into neighboring dots and often gives a more scattered, natural-looking pattern. Ordered Bayer gives a repeated grid that makes the pattern more visible. Try both on the same source.
Will the dithering tool keep transparent areas?＋
Yes. The PNG keeps decoded alpha, including partial transparency. Fully transparent pixels remain transparent. The color channels of visible pixels become black or white.
Does the PNG have the original dimensions?＋
Yes, it uses the decoded image width and height. The before-and-after previews are scaled to fit the screen. The exported file does not preserve source metadata or the original compression.
Is this a 1-bit printer file?＋
The export is a standard PNG with black and white visible pixels and alpha. A specific printer, thermal label tool or embedded display may require its own packed 1-bit format, sizing or preprocessing.
Standards and limitations
Related tools
Image Grayscale
Make a photo grayscale with a clear brightness method and a transparency-preserving PNG.
Image Threshold
Convert an image to two-tone black and white, with a threshold and explicit alpha controls.
Image to ASCII
Turn a local image into character art with a live preview and script-free TXT, HTML and SVG exports.
Image Palette Extractor
Extract representative colors from a local photo and download a reusable CSS, JSON or SVG palette.