Scalable Machine Learning Approach to Classifying Transportation Noise at Two Urban Sites in Greater Boston, Massachusetts

The goal of this study was to characterize transportation noise by vehicle class in two urban communities, to inform studies of transport noise and ultra-fine particulates. Data were collected from April to September 2016 (150 days) of continuous recording in each urban community using high-resolution microphones. Training data was created for airplanes, trucks/buses, and train events by manual listening and extraction of audio files. Digital signal processing using STFT and Hanning windowing was performed in MATLAB, creating audio spectrograms with varying frequency: log vs linear frequency scales, and 4K vs 20K max frequency. For each of the four spectrogram sets, a neural net model using PyTorch was trained via a compute cluster. Initial results for a multi-class model provide an accuracy of 85%. Comparison between a selection of frequency scales and expanding to longer time periods is ongoing. Validation with airport transport logs and local bus and train schedules will be presented.

Shooting noise annoyance in communities around German military training areas

Shooting noise is characterized as impulsive, intermittent sound with high energy and low frequencies. Studies have shown that for given average sound levels shooting noise is regarded as more annoying than transportation noise, particularly road traffic noise. In comparison to transportation noise, responses to shooting noise are less frequently studied. The latest published German studies on community responses to shooting noise were conducted in the 1980ies and 1990ies. The study presented in this contribution aims to provide new data on shooting noise responses in communities around military training areas. Annoyance responses were collected using a survey with 1043 residents living around three military training sites in Germany. For the address of each resident, on the basis of shooting training in the year 2019 the average continuous sound levels and the sound exposure levels for day and night-time with the frequency weightings A, C, and Z was estimated for grid cells of 250 x 250 m. Results on the exposure-response relationship between these noise metrics and the percentage of highly annoyed persons (%HA) are presented. Among others, the results indicate, that non-acoustic factors, particularly attitudes related to the source have a strong impact on the annoyance.