An Efficient Audio Coding Scheme for Quantitative and Qualitative Large Scale Acoustic Monitoring Using the Sensor Grid Approach

Abstract

The spreading of urban areas and the growth of human population worldwide raise societal and environmental concerns. To better address these concerns, the monitoring of the acoustic environment in urban as well as rural or wilderness areas is an important matter. Building on the recent development of low cost hardware acoustic sensors, we propose in this paper to consider a sensor grid approach to tackle this issue. In this kind of approach, the crucial question is the nature of the data that are transmitted from the sensors to the processing and archival servers. To this end, we propose an efficient audio coding scheme based on third octave band spectral representation that allows: (1) the estimation of standard acoustic indicators; and (2) the recognition of acoustic events at state-of-the-art performance rate. The former is useful to provide quantitative information about the acoustic environment, while the latter is useful to gather qualitative information and build perceptually motivated indicators using for example the emergence of a given sound source. The coding scheme is also demonstrated to transmit spectrally encoded data that, reverted to the time domain using state-of-the-art techniques, are not intelligible, thus protecting the privacy of citizens.

Low-Cost Sensors for Urban Noise Monitoring Networks—A Literature Review

“…A third octave band analysis and an A-weighting are performed using digital filtering. In addition, the data are encoded to reduce the memory load [26]. The influence of the air temperature have been evaluated in a climatic chamber, showing a very small deviation.…”

“…Thus, the use of low-cost Sensor Networks (SN) can be a solution to the current limits of the noise observatories, mentioned above, by making it possible to reach a density of measurement points in a territory that is capable of providing a very rich acoustic information. The use of such a noise SN also opens up many additional interesting opportunities, such as to assess and manage road traffic noise [12][13][14], to enhance the traditional strategic noise mapping process [15][16][17][18][19], to produce dynamics noise maps [20] or to capture the sound sources of interest or acoustic events within the signal [21][22][23][24][25][26][27][28][29]. One can mention also that low-cost SN may be used in other fields in acoustics, such as sound source localisation [30] and biodiversity monitoring [31][32][33].…”

Estimation of the Perceived Time of Presence of Sources in Urban Acoustic Environments Using Deep Learning Techniques

“…Using such information-rich representations of the signal also allows the computation of commonly used monitoring acoustical indicators. Though, it may not be desirable in long-term monitoring applications where analysis is performed offlinen as it requires large transmission bandwidth and storage capabilities, and raises privacy concerns if intelligible speech can be retrieved [10]. Some monitoring applications [11,12,13] use third octave sound level fast (125ms) measurements to underline relevant spectral information.…”

“…Some monitoring applications [11,12,13] use third octave sound level fast (125ms) measurements to underline relevant spectral information. This representation is easier to store in the long term and reduces privacy concerns [10]. In [11], the authors successfully linked acoustical indicators derived from third octave sound levels to the perceived activity of specific sound sources in recordings of polyphonic sound scenes.…”

“…Typically, spectral representations are preferred to the raw audio waveform because of the regularities they underline in the signal. Here, the thirdoctave spectrum is considered as it is commonly used in acoustic monitoring applications[1,10]. Thirdoctave spectra are computed for 125 ms frames and 29 frequency bands in the 20 Hz − 12.5 kHz range as the input signal representation.…”