Technical guide

Error correction adds resilience, but it also adds density

QR codes can reconstruct some missing or unreadable modules. The four correction levels trade compactness for recovery capacity, so the best choice depends on the artwork and its environment.

Written and reviewed by LaunchLabUpdated September 13, 2026About 7 minutes

What error correction actually does

A QR code stores redundant information so a scanner can recover data when part of the pattern is dirty, scratched, faded, or covered. Correction is not permission to crop the quiet zone, hide finder patterns, or use poor contrast. Those structural problems can prevent a scanner from locating the code before recovery begins.

LevelApproximate recoveryPractical use
LAbout 7%Clean, controlled digital or print use with short content
MAbout 15%Balanced default for ordinary printed codes
QAbout 25%Material exposed to handling, light wear, or modest branding
HAbout 30%Small center logo or environments where partial obstruction is plausible

The percentages are useful orientation, not a guarantee that any similarly sized area can be removed. Damage location, data layout, code version, print quality, and scanner software all influence the outcome.

Why higher is not automatically better

For the same payload, stronger correction generally requires more modules. That can create a denser pattern or a larger QR version. If the physical print size stays fixed, each module becomes smaller and harder to reproduce. A high-correction code containing a long URL may scan worse than a balanced code containing a short URL.

Good sequence: shorten unnecessary content first, choose correction for the environment second, then choose a print size that preserves clear modules.

How to choose a level

Use M for a clean default

Balanced correction works well for cards, flyers, menus, and signs printed with strong contrast on a flat surface. It provides resilience without making every code unnecessarily dense.

Move to Q when handling is expected

Consider Q for packaging, handouts, reusable table material, or other pieces likely to collect light scratches or stains. Still keep important design elements away from the QR pattern.

Use H deliberately for center branding

H is appropriate when a modest logo or label occupies the center. Keep the mark small, preserve the finder patterns and quiet zone, and test the final export. Correction cannot rescue an oversized logo that removes too much data.

Common mistakes

  • Using H to compensate for pale colors: correction does not restore missing contrast.
  • Covering a corner finder: these large squares help the scanner locate and orient the code.
  • Adding a white logo box without resizing: the box also covers modules and counts toward the obstruction.
  • Testing only at large preview size: density becomes a problem at the final printed dimensions.

A practical workflow

  1. Use the shortest stable destination that accurately serves the visitor.
  2. Generate at M with dark modules on a light background.
  3. If adding a center mark, move to H and keep the mark restrained.
  4. Export SVG for layout or a sufficiently large PNG.
  5. Print at actual size and test on multiple phones under realistic light.

Compare the settings visually

Generate the same content at different correction levels and test the actual production size.

Open the generator