Noise-limited reading of bar codes

The ubiquitous supermarket checkout scanner is a well-engineered and effective device. Existing scanners rely on simple and low-cost signal processing to interpret bar-code signals, which imposes restrictions on the system noise power that they can tolerate. In this paper, the authors describe the relationships between engineering parameters of the system that limit reader performance. If the combined noise is Gaussian, they show that the reader error probability depends on a single parameter, which they call a timing signal-to-noise ratio.