The forensic voice expert

Martin Barry is a full-time forensic voice expert (article here). He is enlisted by the police to examine recordings of voices -- although some recordings are merely fragments and scraps against the crackling background noise of a poor-quality tape. In one case a singer claimed to be making a recording in the studio at the time of a murder, and produced the recording in question. Barry compared this with another recording from an earlier date and found that the two were identical in every way, suggesting that the singer had submitted the earlier recording as his alibi evidence.

 

You might think that a listener would notice whether two recordings were the same or not, but Barry says that people's memories of voices are notoriously unreliable. He cites one academic who did not even recognise his mother when presented with a short recording of her voice, without being told who it was. That is because when we are talking to someone on the phone we know who it is, therefore we never pay very close attention to individual sounds and pronunciation quirks.

 

Forensic phonetic evidence alone is unlikely to convict or acquit anyone, but it can often be a useful part of the evidence mix. Despite the advances being made in this field, many police forces still rely on old-fashioned crackly tapes for their interviews, although the Association of Chief Police Officers has recommended a switch to digital recording equipment.

 

The BBC article about Martin Barry, the voice expert, here. It mentions the infamous Yorkshire Ripper hoax, where police were misled by the hoaxer's accent. There's a bit more on that story in this old post of mine.

Comments

[marc leavitt]

Interesting post, but I have to disagree with the voice recognition. I just got a phone call yesterday from an old friend I haven't spoken to in months, and as soon a he said hello, I said his name.I agree it's not consistently true, but I wonder what the percentages are.

[Virtual Linguist]

Thanks, Marc. Not sure about percentages, but I expect that the academic's mother's voice was presented to him alongside a variety of other voices in a laboratory setting. Perhaps he didn't even know that she was among them, so didn't expect to hear her in that setting.

It's probably a similar situation to one I often experience as a teacher. I recognise my students provided that they are sitting in their usual places in the classroom, but if I bump into them in the supermarket, I never recognise them!

[Belstaff Outlets]

I like you on facebook and follow through google reader!