Generating Sound Stimuli with Given Emergence Level and Low Frequency Content by Mixing Recordings

Abstract

Noise annoyance and other effects of noise are reasonably correlated with A-weighted sound levels. Currently, the influence of low-frequency content and of sound emergence level (i.e. the sound level difference between total noise and residual noise) on effects of noise are being assessed in laboratories using stimuli based on recordings. Standards intended for regulations often include penalties depending on these factors. The difference between C-and A-weighted sound levels is frequently used as a descriptor of low frequency content. This work proposes a method to optimize the search and combination of a subset of audio files from a large set of noise sources recordings in order to achieve an environmental noise stimulus with previously specified values of factors (i) sound emergence level, (ii) C-A-weighted levels difference, (iii) A-weighted equivalent continuous sound level and (iv) duration of stimulus. The method is implemented, and one hundred stimuli are generated following a full factorial experimental design varying these four factors. The resulting stimuli show small differences between specified and measured values. The mean difference for factors i-iii is 0.46 dB and the maximum is 2.3 dB.

Room acoustics: Idealized field and real field considerations

“…Methods for automatic signal generation that combine sound signals in order to obtain desired spectral and temporal characteristics were developed by Accolti and Miyara (2015) and methods that can also obtain desired signal to noise ratios depending on frequency were developed by Accolti et al (2017b). The required measures can be obtained with questionnaires and auralization techniques (Vorländer, 2008), ray acoustics models (Savioja and Svensson, 2015), or wave equation models (Kuttruff, 2000) according to the application and in turn to the particular definition of the diffuse field.…”

Room acoustics: Idealized field and real field considerations

“…These criteria can be generalized for certain applications by combining methods for automatic signal generation and methods for room response modeling. Methods for automatic signal generation that combine sound signals in order to obtain desired spectral and temporal characteristics were developed by Accolti and Miyara (2015) and methods that can also obtain desired signal to noise ratios depending on frequency were developed by Accolti et al (2017b). The required measures can be obtained with questionnaires and auralization techniques (Vorländer, 2008), ray acoustics models (Savioja and Svensson, 2015), or wave equation models (Kuttruff, 2000) according to the application and in turn to the particular definition of the diffuse field.…”