QR Code Error Correction Levels Explained
QR codes support four error correction levels. L, M, Q and H have approximate codeword restoration rates of 7%, 15%, 25% and 30%, respectively. QRSen uses M by default and H for logo uploads.
QR codes use Reed–Solomon error correction to recover some damaged codewords. The restoration rate describes a proportion of codewords. Damage to the quiet zone or finder patterns can prevent the reader from locating the symbol.
That redundancy isn't free, it takes up space in the grid alongside the actual data, so a higher error correction level means more of the pattern is dedicated to redundancy rather than content, which makes the code denser and, for a fixed amount of data, physically larger or made up of smaller, more tightly packed modules. Choosing L over H for the same content produces a visibly simpler, sparser-looking pattern.
L is reasonable for a code that will only ever be displayed cleanly on a screen with no risk of damage or dirt. M is a common general-purpose default. Q and H matter most anywhere the code faces real-world wear, outdoor signage exposed to weather, product packaging that gets scuffed in shipping, or anywhere a logo or design element sits on top of part of the pattern, since that overlay is effectively the same kind of damage the error correction needs to compensate for.
QRSen uses M for standard codes and automatically switches to H when you add a logo. Compare the finished design with a logo-free proof before printing.
It's worth remembering that error correction is a recovery mechanism, not a substitute for good print practices — a code sized correctly, printed with strong contrast, and given proper quiet-zone margin at level M will generally outperform a poorly sized, low-contrast code even at level H. Error correction buys tolerance for damage that happens after the code is otherwise well made; it doesn't compensate for basic sizing or contrast mistakes made at the design stage.