10-05-2006

An innovative technique that, for the first time, accurately measures exactly how sound behaves in “real-world” situation is now under development—and could improve acoustics in buildings ranging from concert halls to railway stations.

The potential impact of the technique, which could also assist in the development of more effective hearing aids, was described at this year’s BA Festival of Science in Norwich, UK.

The technique is designed to pinpoint precisely how indoor environments respond to music and speech while those areas are in everyday use. This opens up the prospect of basing acoustic design on more realistic information about the way sound behaves than has previously been possible. It may also contribute to the development of hearing aids that adapt the way they process sound according to the acoustic environment they are in, providing a much better listening experience for hearing aid users than is currently achievable.

The conventional way of measuring acoustics has been to make a short blast of noise (eg, gunshot), record it, and analyze how it dies away. The noise has to be very loud so that the environment’s effect on it can be assessed across the full range of sound, from very loud to very quiet—only in this way can comprehensive information on an environment’s acoustic performance be obtained. However, gunshot noise poses a risk to hearing and is unpleasant to listen to. This means that measurements taken in unoccupied areas are the norm even though these do not accurately indicate “real” acoustic performance—when people are present, moving, talking, etc.

Now, engineers at the University of Salford, Greater Manchester, UK, are exploring whether music played at an average level of audibility, or even the conversation of people in the indoor environment being tested, could be used instead of the loud, short blast of noise. The work is being funded by the Engineering and Physical Science Research Council (EPSRC).

Exploiting the major advances in computing power and sophistication achieved in the IT sector in recent years, the team is developing groundbreaking computer programs capable of isolating snippets or phrases from normal music or speech, analyzing their decay, and extrapolating this data so it provides an accurate indication of an environment’s effect on sound. Since loud test sounds are not required, this approach avoids the need to vacate the environment when testing takes place, enabling more realistic acoustic data to be gathered.

[SOURCE: EurekAlert!, a service of AAAS; Engineering and Physical Sciences Research Council, September 8, 2006]