How accurate is voice biometric technology?

Computers are much better at analyzing and identifying the voice than people. You might not know with certainty who is on the other end of a phone call, even when calling people you know well. Voice biometric technology can usually tell you with a less than one percent error rate. (It’s even lower if the speaker is saying a predetermined passphrase.)  Voice biometric technologies are similar to fingerprint biometrics in terms of accuracy.

Accuracy is defined as a combination of false positives and missed identifications (false negatives). Accuracy depends on several factors, including:

  • The analysis and voiceprint technology used
  • The length of the recording being analyzed– accuracy increases with duration
  • The quality of the audio recording
  • Whether the analysis includes predetermined text (a passphrase or sentence)

The charts below show the probability and accuracy of the AGNITIO voice identification engine, using both a pass phrase (text-dependent) and free-form speech.