QR Code Error Correction Levels Explained: L, M, Q, H (2026)

Error correction is redundancy built into the QR code. When a QR code is generated, extra data is added alongside your actual information. This extra data lets a scanner reconstruct the original content even if part of the code is unreadable.

The technology behind it: QR codes use Reed-Solomon error correction, the same family of mathematics used in CDs, DVDs, and data transmission. It works by encoding your data with enough redundancy that a certain percentage of the code can be lost or corrupted, and the scanner can still recover the full original message.

Why QR codes need it: QR codes live in the real world, where they get damaged. A code printed on packaging gets scratched. A code on a poster gets rained on. A code with a logo has part of its pattern covered. A code on a curved surface gets distorted. Without error correction, any of these would make the code unreadable. With it, the code keeps working.

The trade-off: Higher error correction means more redundant data, which means more modules (the small squares that make up the code) packed into the same space. More modules make the code denser and visually more complex. At small sizes or low print quality, a very dense code can actually be harder to scan. So the goal is choosing the lowest level that gives you the resilience you actually need.

This trade-off, resilience versus density, is the entire decision. The four levels give you four points along that spectrum.

QR codes offer exactly four error correction levels. Each recovers a different percentage of the code:

Level L (Low): ~7% recovery The code can lose about 7% of its data and still scan. This is the least redundancy, which makes the code the least dense (fewest modules). Best when the code will be clean, undamaged, and displayed digitally or printed at good quality with no logo.

Level M (Medium): ~15% recovery The code can lose about 15% and still scan. This is the most common default and a good general-purpose balance. Suitable for most standard printed codes that will not be heavily handled or logo-covered.

Level Q (Quartile): ~25% recovery The code can lose about 25% and still scan. More redundancy, denser code. Good for codes that will face some wear or environmental exposure, or that have a small logo.

Level H (High): ~30% recovery The code can lose about 30% and still scan. The maximum redundancy, the densest code. Essential when embedding a logo (which covers part of the code), or for codes in harsh environments where significant damage is likely.

The percentages (7, 15, 25, 30) are the share of the code that can be missing or damaged while the code still scans reliably.

The right level depends on three factors: whether you are adding a logo, where the code will live, and how small it will be.

The most important practical rule: For everyday business use with no logo, M is the sensible default. The moment you add a logo, jump to H. Those two rules cover the vast majority of real-world decisions. Q sits in between for codes that face wear but have no logo or only a tiny one.

Whichever level you choose, the fundamentals still apply: adequate size, strong contrast, a clear quiet zone, and testing before you commit. See our QR code size guide and design guide.

The single most important practical use of error correction is enabling logos. Understanding this relationship helps you brand QR codes without breaking them.

Why a logo needs high error correction: When you place a logo in the center of a QR code, the logo covers part of the code's pattern, effectively "damaging" it from the scanner's perspective. If the logo covers, say, 15% of the code, you need error correction that can recover at least that much. Level H (30% recovery) gives you the headroom to place a reasonable logo and still have the code scan reliably.

Why not just always use H? Because H makes the code denser (more modules in the same space). For a clean code with no logo, that extra density is wasted and can even make the code slightly harder to scan at very small sizes. Use the resilience where you need it, not everywhere.

A common mistake: Adding a logo to a code that uses level L or M, then wondering why it scans unreliably. The logo covers more than the low error correction can recover, so the code works sometimes and fails other times, the worst outcome. Always pair logos with H.

Most good QR platforms handle this automatically: when you add a logo, they switch to H error correction for you. But understanding why helps you troubleshoot if a branded code ever misbehaves. For a full guide on branded codes, see our best QR generators with logo support.

Here is a quick reference matching error correction levels to common real-world scenarios:

Digital-only QR code (screen, email, PDF), no logo: Level L or M. No damage risk, so keep it low and clean.

Business card, no logo: Level M. Standard printing, protected in a wallet.

Business card with a logo: Level H. The logo requires it.

Restaurant menu, no logo: Level M or Q. Some handling, but generally protected.

Product packaging: Level Q or H. Faces handling, scanning, and possible scuffing.

Outdoor signage or posters: Level Q or H. Weather and distance scanning benefit from extra resilience.

Apparel (t-shirts): Level H. Fabric distortion, folds, and washing all "damage" the code.

Industrial or construction environments: Level H. Harsh conditions, dust, and rough handling.

Any code with a logo, anywhere: Level H. No exceptions.

Stickers and labels that may wear: Level Q or H depending on expected wear.

The overarching guidance: When in doubt between two levels, the cost of going one level higher is minor (slightly denser code), while the cost of going too low is a code that fails when damaged. For anything printed or handled, lean toward more resilience. For clean digital use, keep it low. And always test the finished code in its real conditions before committing to a print run. See our 15 common QR code mistakes guide for related pitfalls.