Retroqr-codesdensohistorystandards
QR code corners use a pattern Denso's engineers struggled to find in ordinary print
Masahiro Hara's team surveyed printed matter to choose the 1:1:3:1:1 marker ratio. Finding the code quickly came before reading its contents.

Your phone has to find the QR code before it can read the menu somebody refused to print. Those conspicuous corner squares help it do that, and their proportions came from studying the surrounding visual clutter.
Denso Wave’s development history describes Masahiro Hara and one teammate surveying fliers, magazines, boxes, and other printed matter. A scanner could mistake a similar mark for a code’s position marker, so they needed a distinctive black-and-white sequence.
They chose widths in the ratio 1:1:3:1. Denso calls it the “least used ratio” in the printed material they examined. Recognizing that pattern also helped determine orientation when a code arrived at a different angle.
Denso announced QR Code in 1994. The company’s history says it made the specification public and declined to exercise its patent rights, helping other people adopt it. Getting the scanner to recognize the mark was one problem; getting everybody else to use it was another.
The corner pattern solves a search problem that comes before the data problem. A page can contain photos, text, boxes and other marks; the scanner needs a way to pick out the thing it should read. More storage inside the symbol would be little help if the reader kept selecting the decoration next to it.
Denso’s survey is a useful design lesson for anyone building a machine-readable label. The environment around the mark belongs in the test. A symbol that behaves well by itself on an empty page has not yet faced the flyer, package or screen on which people will actually encounter it. The rare ratio gave the code a way to distinguish itself from that surrounding material.
The four-module margin makes the same point in a printable form. Empty space can be a functional part of the design even when a layout editor treats it as available room. The printing guide’s worked example calculates the symbol and then adds the margin around it. Counting only the dark square would underestimate the space the finished code needs.
For a real invitation or label, a scan test should use the final printed size and surrounding artwork. A successful scan of the large source image on a monitor answers a different question. The remaining test is the version somebody will hold in front of a camera after the design has been cropped, placed and printed.
Reading a QR code with a pencil takes the mechanism apart from the other end. Before decoding any of those squares, though, the reader needs a clear target. The next invitation designer adding a logo has to leave the scanner enough of one.