Sound-separator device allows users to select what they hear

People who are hard of hearing could benefit from a new listening system that allows them to pick out individual voices and reduce background noise.

The technology could be particularly useful in situations where groups of people are talking at the same time, such as in shopping centres, in restaurants or the workplace.

‘This system is about giving listeners the flexibility to create their own acoustic environment,’ said research leader Dr Banu Günel, who developed the technology while at Surrey University’s Centre for Communications Systems Research.

Currently, induction-loop systems and hearing aids can be used to improve the volume of ambient sounds, but cannot target and amplify specific sounds effectively.

In response to this, Günel and colleagues developed a ‘sound-separator’ device, which can be used at the front end of induction-loop systems or assisted-listening devices to isolate sounds and improve speech intelligibility.

It uses a freestanding processor to give users the ability to select which sounds to listen to and to balance volume across an unlimited number of target sounds.

‘The listener can control the amplitude for those sounds they are choosing to listen to. If the listener wants to hear two people talking, but one is naturally much quieter than the other, they can increase the volume of the quieter talker and reduce that of the other,’ said Günel.

Crucially, the technology works in real time, so the listener can hear the output in synchronicity with lip movements.

The prototype technology was tested by 40 members of The Royal National Institute for Deaf People (RNID). Volunteers listened on their hearing aids (set on the T-coil setting) to a series of recordings played through a loop system. Each recording had a series of rhyming words that the volunteers had to identify, with and without the intelligent microphone, for different levels of background noise.

Participants understood almost 50 per cent more speech with the intelligent microphone with average background noise. When the background noise was increased by six decibels, volunteers understood 400 per cent more speech.

There are also potential applications for the technology aside from speech, such as studio recording, where the system is capable of ‘cleaning’ very rich sounds, including those from musical instruments.

‘Producers could record an ensemble and reduce the sound of one particular player who is making mistakes,’ Günel said.