How QR Codes Work

The Technology Behind QR Codes Explained

QR codes look deceptively simple — just a grid of black and white squares. But beneath that visual simplicity lies a remarkably elegant encoding system designed for speed, reliability, and error tolerance. Understanding how QR codes work helps explain why they've become the universal standard for linking physical objects to digital content.

The Anatomy of a QR Code

Every QR code follows the ISO/IEC 18004 standard and is composed of several distinct functional regions. Each plays a specific role in the encoding and decoding process.

Finder Patterns

The three identical square patterns in the corners (top-left, top-right, and bottom-left) are called finder patterns. These allow any QR scanner to immediately identify the code, determine its boundaries, and establish the correct reading orientation — regardless of the angle or skew.

Alignment Patterns

In larger QR codes (version 2 and above), smaller alignment squares are added to help the scanner correct perspective distortion. This is why you can photograph a QR code on a curved surface or from an angle and still read it successfully.

Timing Patterns

Running horizontally and vertically between the finder patterns, timing patterns are alternating black and white modules that establish the coordinate system of the code. They tell the decoder exactly how large each data cell is.

Format Information

Surrounding the finder patterns is format information that encodes the error correction level and mask pattern used in the code. This is read first so the decoder knows how to interpret the rest of the data.

Data and Error Correction Modules

The remaining area of the QR code contains the actual encoded data along with error correction codewords, encoded using Reed-Solomon error correction.

Error Correction: Why QR Codes Are So Robust

One of the most powerful features of QR codes is their built-in error correction. Even if a portion of the code is damaged, obscured, or worn, it can still be decoded correctly. There are four levels of error correction:

• Level L — recovers up to 7% data loss
• Level M — recovers up to 15% data loss
• Level Q — recovers up to 25% data loss
• Level H — recovers up to 30% data loss

This is why QR codes with logos placed in the center still work — the logo damages up to 30% of the code, which is within the tolerance of Level H error correction. It's also why creative, styled QR codes with patterns and colors can function perfectly despite their non-standard appearance.

How Data is Encoded

The data encoding process follows several steps:

1. Data analysis — the encoder determines the most efficient encoding mode (numeric, alphanumeric, byte, or Kanji) for the input text
2. Data encoding — the data is converted into a binary bitstream according to the chosen mode
3. Error correction — Reed-Solomon error correction codewords are generated and appended to the data
4. Structuring — the combined bitstream is placed into the QR code matrix
5. Masking — a mask pattern is applied to ensure the code doesn't have large areas of the same color, which could confuse scanners
6. Format information — the error correction level and mask pattern are encoded into the format information regions

How Scanners Decode QR Codes

Modern smartphone cameras perform the entire decoding process in milliseconds using a combination of image processing and mathematical algorithms:

1. The camera captures an image and the software detects the three finder patterns
2. The code's boundary and orientation are established
3. Perspective correction is applied using the alignment patterns
4. The format information is read to determine the error correction level and mask
5. The mask is removed and the data modules are read
6. Error correction is applied to fix any damaged or misread modules
7. The binary data is decoded into the original text, URL, or other content

QR Code Versions and Capacity

QR codes come in 40 versions (sizes). Version 1 is the smallest at 21×21 modules, and Version 40 is the largest at 177×177 modules. Larger versions can store more data:

• Version 1: up to 41 numeric or 25 alphanumeric characters
• Version 10: up to 652 numeric or 395 alphanumeric characters
• Version 40: up to 7,089 numeric or 4,296 alphanumeric characters

For most everyday uses — like encoding a website URL — Version 3 or 4 is sufficient, which keeps the code compact and easy to scan.

Try It Yourself

Now that you understand the technology, why not create your own QR code with QR Forge? Our free generator handles all the encoding automatically, producing a clean, high-resolution QR code you can use anywhere.