The grant from the Engineering and Physical Sciences Research Council will enable the Salford team work with global design and engineering firm, Arup, the University of Reading and BBC R&D. Together they hope to solve the problems associated with demonstrating to clients how building designs for large spaces like concert halls, sound once they've been built.
In practice, designers programme the designs of a building or room into a computer model which calculates how the structure affects the sound sources which inhabit it, such as musicians in an orchestra. The modified sound of the sources is then reproduced through a special loudspeaker system, allowing the client to hear what the building will sound like in advance of it being constructed.
Some of the acoustic modeling technologies behind this approach are currently not as accurate as they could be, because they treat sound as a beam rather than a wave and unusual acoustic features of the design can make the calculated sound unrealistic. Real time interactive alterations are also not possible - if a client requests a change to the design then the model has to be recomputed.
Current acoustic modelling technology either suffers from these problems or cannot be used to model mid and high pitch sounds due to excessive requirements for computational power. However the aim of the three year Salford project is to create a unified model which gives accuracy of for all conceivable scenarios without an unacceptable increase in processing requirements.
It is hoped that this will not only allow models to be improved, but also to be taken outside for use by town planners or into smaller buildings such as schools.
The new algorithm will be developed with future support for real-time interactive alterations in mind. In particular the team hopes that changes to small parts of the building will only require changes to small parts of the computer model, so they may be achieved with minimal processing. They will also develop a new hardware interface which will connect the loudspeaker system directly to the computer model.
Research Co-Investigator, Dr Jonathan Hargreaves from the University of Salford's School of Computing, Science and Engineering said, "This technology is currently used in large-scale projects to give the client an idea of what the finished building will sound like. We feel that with new algorithms we can improve this experience dramatically and make it much more realistic.
"In time, we hope that we are on a path where this modelling can be cost-effective enough to use on smaller buildings such as homes or schools, or accurate enough to be used for town planning purposes."
The project will conclude in September 2015 following the completion of theoretical and practical pilot studies.
(Jim Evans)
