Error correction and the quiet zone

QR codes repeat information so a scanner can rebuild the message when part of the symbol is damaged or covered. The four levels are L, M, Q, and H. Roughly, they can recover about 7%, 15%, 25%, and 30% of the codewords. The percentages are a property of the symbol, not a guarantee about a blurry photo.

Higher correction uses more modules for the same text, or refuses the text when it no longer fits. If QRNest says the content is too long, the library refused the payload at that level. Shorten the text, or drop from H toward M or L if the code will be printed cleanly and large.

M is the default here because it is a reasonable balance for a URL or a short note. Choose H when you add a logo or expect the print to be scratched. Choose L only for a long payload that must fit and will be displayed sharply.

The quiet zone is the empty margin around the modules, measured in module widths. The common recommendation is four. QRNest lets you set 0 through 10 and warns below 2. A code shoved against a dark photograph often fails even when the modules themselves are perfect.

Contrast matters as much as the level. Scanners expect dark modules on a light field. Inverting that can work on some phones and fail on others. The preview runs a decoder in the browser and says whether this image was readable. Test a printed copy anyway, especially on glossy paper and in low light.

Size is the pixel width of the downloaded image. 768 pixels is a practical default for a sheet of paper. A tiny code on a billboard is a contrast and resolution problem, not an error-correction problem. Give the symbol physical size before you raise the correction level.