Speaker recognition systems use spectrograms to represent human voices.

Photo courtesy Richard Horne

Voiceprints

Your voice is unique because of the shape of your vocal cavities and the way you move your mouth when you speak. To enroll in a voiceprint system, you either say the exact words or phrases that it requires, or you give an extended sample of your speech so that the computer can identify you no matter which words you say.

When people think of voiceprints, they often think of the wave pattern they would see on an oscilloscope. But the data used in a voiceprint is a sound spectrogram, not a wave form. A spectrogram is basically a graph that shows a sound's frequency on the vertical axis and time on the horizontal axis. Different speech sounds create different shapes within the graph. Spectrograms also use colors or shades of grey to represent the acoustical qualities of sound. This tutorial has a lot more information on spectrograms and how to read them.

Some companies use voiceprint recognition so that people can gain access to information or give authorization without being physically present. Instead of stepping up to an iris scanner or hand geometry reader, someone can give authorization by making a phone call. Unfortunately, people can bypass some systems, particularly those that work by phone, with a simple recording of an authorized person's password. That's why some systems use several randomly-chosen voice passwords or use general voiceprints instead of prints for specific words. Others use technology that detects the artifacts created in recording and playback.

Other systems are more difficult to bypass. We'll look at some of those next.