Super voice recognisers, people with extraordinary abilities to remember and recognise human voices, are out there. And they could potentially transform counter-terrorism and surveillance operations, some suggest.
People with an exceptional knack for recognising faces, known as super-recognisers, seized the zeitgeist more than 10 years ago, when it was discovered that about 2 per cent of the population had such an ability. London鈥檚 Metropolitan Police Service later created a specialised unit of officers with this skill, some of whom helped to identify suspects caught on CCTV during the 2011 London riots.
Ryan Jenkins at the University of Greenwich in London wanted to find out whether these people also had superior abilities to recognise and match voices.
Advertisement
To do this, he put 529 super face recognisers through a series of voice-recognition tests. One of the tests consists of 80 pairs of audio clips of someone voicing a different syllable. Participants must decide if the audio clips belong to the same person or not.
Another test looked at whether people could recognise celebrity voices, and another was included to ensure that the ability was specific to human voices, and not to other types of sounds like the ringing of a bell.
Jenkins found that super face recognisers were more likely to have super voice recognition abilities than the average person, but only 22 participants 鈥� just over 4 per cent 鈥� met the criteria for super voice recognition. They scored in the top 15 per cent of results for at least two tests.
However, Jenkins notes that the tests he uses in the experiment are 鈥減rimarily designed to test phonagnosia [an inability to recognise voices] which is the complete other end of the spectrum鈥�. To accurately tell who a super voice recogniser is, more extensive tests need to be developed, he says.
Josh Davis, also at the University of Greenwich, speculates that people with this ability could be able to help in all types of police operations. Voice recognition software is used by the police to identify suspects in phone-tapping operations, but Davis says that 鈥渓ike face recognition algorithms, they鈥檙e helpful but they do make mistakes鈥�.
A recent European Union project tried to develop a voice-recognition algorithm that can help better identify terrorists and criminals in intercepted phone-calls. Jenkins says super voice recognisers could work in conjunction with computers, making these investigations more effective.
We still know very little about whether these abilities are innate or if they can be improved with practice, says David White at the University of New South Wales in Australia. 鈥淭here is a chicken and egg problem here. Do we become good at social perceptual abilities because we enjoy interaction, or is it the other way around?鈥� he says.
Anna Bobak at the University of Stirling, UK, says the results of the study need to be replicated. 鈥淲e ought to be very cautious in making any claims and policy decisions about deploying these people,鈥� she says. 鈥淭his is because it is unclear how reliably they perform across tasks. Say if someone does well at familiar voice recognition, it鈥檚 not a given that they would be equally good at deciding that voices in three separate recordings belong to the same person, or two different people.鈥�
Reference: PsyArXiv,
