Application of artificial neural networks for editing measured acoustical data for simulations in virtual environments

Acoustic simulation tools are used and demanded by various groups of people. Architects and urban planners as well as product designers and engineers are interested in simulating the acoustical properties of buildings, machines or other products. Acoustic simulation techniques are continually evolving. The current trend is towards integrating forward-looking technologies like virtual reality (VR) into the simulation process. Common acoustical simulation tools, such as numerical methods, are computationally expensive and cannot be applied in real time. This, however, is a mandatory requirement for VR-applications. For that reason, techniques based on measured acoustical data are often used for acoustic simulations in VR. However, various disturbance variables, such as interfering noise, can distort measurement results immensely. In this paper an Artificial Neural Network (ANN) is described which can be used for the post-processing of measured data. A concept specifically for the noise cancellation in acoustic measurement data is outlined.


Citation style:
Could not load citation form.


License Holder: ©2017 - TU Ilmenau

Use and reproduction:
All rights reserved