Abstract
Voices are highly individual, and this information may be used to recognize people. This chapter provides an overview of how speaker-specific voice information can be used to assist in recognizing unknown speakers in evidential audio recordings in order to assist in progressing criminal investigations or for evidential purposes. While the chapter is predominantly concerned with forensic applications of speaker recognition, it also considers commercially-orientated applications such as the voice-access systems increasingly used in banking, for example. The chapter distinguishes between biometric features of voice (i.e. aspects of speech that facilitate recognition of an individual) and biometric recognition (i.e. the process of using those features to undertake the recognition). Biometric features include a range of frequency- and time-domain characteristics that vary between speakers and are useable in human- based and automatic speaker-recognition processes, the principles of which the chapter explains and discusses. Finally, the chapter considers possible future scenarios in which biometric features of voice might be of use, should certain methodological obstacles be overcome.
| Original language | English |
|---|---|
| Title of host publication | The Oxford Handbook of Voice Perception |
| Editors | Sascha Frühholz , Pascal Belin |
| Place of Publication | Oxford |
| Publisher | Oxford University Press |
| Chapter | 36 |
| Pages | 777-798 |
| Number of pages | 21 |
| ISBN (Print) | 978-0-19874318-7 |
| Publication status | Published - 6 Dec 2018 |
Keywords
- forensics, voice identification, biometric feature, biometric recognition, automatic speaker recognition