Sound source modelling by nonnegative matrix factorization for virtual reality applications

Abstract

Auralization is a suitable method for subjective noise evaluation of virtual prototypes. A basic requirement is the accurate modelling of the sound sources. This includes a dynamic and parametric description at multiple operating conditions. In the case of wave propagation including flow, such as aircraft or vehicle noise, aeroacoustics or fluid dynamics simulations are practically limited to the acoustic near field due to high computational costs. Especially challenging are simulations of rotating systems, such as fan noise radiation. For better applicability, the proposed method is based on in-situ recordings of flyovers. The processing chain compensates for source position and movement as well as atmospheric and soil damping effects on recorded data. The compensated source signal is decomposed into partial sources in spectro-temporal domain with nonnegative matrix factorization (NMF) and can optionally be enhanced by physically-based source information. The format of the source model obtained is ready to use for dynamic sound synthesis in real-time virtual reality applications.

Interpolation of scheduled simulation results for real-time auralization of moving sources

“…The input of the auralization chain is driven by a synthesized aircraft source signal based on a fan [24] and jet noise model [25]. The synthesis is done offline and the signal is included using a wav-file to introduce minimum computational load.…”

Linking atmospheric and urban auralization models

“…Nevertheless, the general trend of the TFs is not changed when adjusting the propagation delays. To further quantify those differences, the two sets of TFs are convolved with an aircraft source signal based on the model by Dreier and Vorländer [29]. Using a 2-second steady-state excerpt of each output signal, the psychoacoustic parameters loudness, specific loudness and sharpness are evaluated using the Matlab routines based on DIN 45631:1991 [30] and DIN 45692:2009 [31], respectively.…”

Real-world study cases for auralization validation: selection, measurements, and methods

“…In particular, the auralization of aircraft noise has found important advancements: recent sound propagation models are able to emulate the effect of an inhomogeneous atmosphere on the sound path geometry and consequently, the properties of the sound itself [5,6]. Moreover, inverse modelling sound source synthesis [7], as well as state-of-the-art models for sound reflection, diffraction, and scattering [8] are promising advancements for perceptually accurate noise auralizations in urban environments.…”