How QR Codes Work: A Complete Visual Guide

Discover the anatomy of a QR code — finder patterns, timing patterns, data modules — and understand exactly how scanners decode them in milliseconds.

A QR code (Quick Response code) might look like a random mosaic of squares, but every element has a precise purpose. Once you understand the anatomy, you'll see why QR codes are so reliable — and why they can still be read even when damaged.

The Building Blocks of a QR Code

QR codes are built on a grid of dark and light modules (squares). A standard version-1 QR code uses a 21×21 grid; larger versions add 4 rows and columns per version number, up to version 40 (177×177). The more data you encode, the higher the version required.

Finder Patterns

The three large square targets in the top-left, top-right, and bottom-left corners are called finder patterns. They let the scanner detect the presence and orientation of the code at any angle and in any lighting condition. The scanner uses the ratio of dark:light:dark modules (1:1:3:1:1) to identify them reliably.

Alignment Patterns

Larger QR codes (version 2 and above) include smaller alignment patterns — miniature squares placed at specific coordinates. These help the scanner correct for distortion when a code is printed on a curved surface or photographed at an angle.

Timing Patterns

Running horizontally and vertically between the finder patterns are alternating dark and light modules called timing patterns. They tell the scanner how large each module is, making it possible to read codes at varying distances and resolutions.

Data Modules and Error Correction

The remaining modules encode the actual data using Reed-Solomon error correction. QR codes support four error-correction levels: L (7% recovery), M (15%), Q (25%), and H (30%). The higher the level, the more of the code can be damaged or obscured and still scan correctly — which is exactly why adding a logo to the center works.

How Scanning Works

When you point a camera at a QR code, the imaging software first locates the three finder patterns to determine the code's position and rotation. It then samples the grid of modules, decodes the binary data using the format information (which describes the error-correction level and mask pattern applied), applies error correction, and finally interprets the data as a URL, text, WiFi credentials, or other content — all in under 100 milliseconds on a modern device.

More Articles