Dithered QR Codes

This is the image that will be dithered into the QR code

This is the text (usually a URL) that people will see when they scan the QR code

No pixel will be allowed to be darker than this colour (before diffusion). This is useful if your image has large areas of black or white, to avoid too much error being built up.

No pixel will be allowed to be brighter than this colour (before diffusion). This is useful if your image has large areas of black or white, to avoid too much error being built up.

QR codes come in 40 sizes. Smaller ones can't contain as much text, but you can use a large one to store less text if you need more space for your picture. Size 6 is the largest one that doesn't add any alignment pattern blocks in the middle of the image.

Makes a QR code designed for a black background.

Rotates the QR code before adding the image, which can help move the required patterns to less obtrusive places.

QR codes still work if you mirror them so who knows, this might put the data modules in a better arrangement for your image.

There are eight different XOR masks that can be applied to QR codes. The choice is fairly arbitrary but picking the right one might happen to give you a nicer pattern of data pixels to work with.

Uses error diffusion when adding the data squares — this makes the image worse and the only reason to turn this off is for demo purposes.

Use error diffusion to dither the image rather than simply thresholding it.

If disabled, just generate a regular QR code with no image in it.

If checked, the free points will be left transparent so you can fill in whatever you want there.

How many times to subdivide each pixel of the QR code. Eg, if this is 3 then each block of the QR code will be replaced by 9 (3²) pixels, the middle one of which contains the data. If it's an even number then the middle 4 pixels will contain the data.

If you have enough error correction in the image, you may be able to change a few data pixels without breaking the QR code. If this is set to 0.01 then we will change the worst 1% of data pixels. This can easily render the code unscannable but might help in some cases.