Psychoacoustic Assessment on off-road Vehicles

The noise produced by HVAC systems, especially on work vehicles like off-road vehicles where workers spend long periods of time, can be very harmful to the end user. The purpose of this study is to validate a linear prediction method capable of assessing the subjective acoustic annoyance of HVAC noise, in order to avoid subjective tests and have a single equation that defines the average perception of the end user. The research follows a systematic methodology that includes literature review, parameters translation, signal acquisition tests, psychoacoustic parameter analysis, subjective evaluations, and the development of a sound quality prediction model. The literature review on the topic helped to ensure the most accurate possible basic methodology, with particular attention to the psychoacoustic parameters used, subjective rating scales, and prediction methodologies applied. To ensure greater accuracy, a translation of the psychoacoustic parameters from English to Italian was carried out to be included in the questionnaires for the subjective tests, considering the nationality of most of the subjects tested. The translation involved two mother-tongue English speaker and an English Italian bilingual person. The translation methodology followed three steps, as explained in the article by Graniello et al. Objective measurements of the noise produced by HVAC systems were taken in the tractor cabin at the operator’s ear level using different types of microphones, such as an artificial head, an omnidirectional microphone, and a 19-channel microphone array. Subsequently, an analysis of the objective data was conducted to understand the sound characteristics of the noise under study. Parameters such as loudness, A-weighted sound pressure level (A-SPL), sound pressure level (SPL), sharpness, roughness, tonality, and fluctuation strength were extracted. Additionally, parameters directly related to the spectral shape of the sound were examined, such as Room Noise Criterion, Noise Rating, Preferred Noise Rating, and Noise Criterion. Knowing the acoustic characteristics of the noise, subjective tests were then conducted. Four different types of tests were performed, each with 15 different subjects. The first test was conducted on the tractor itself, the second in the Audio Space Lab at the Polytechnic University of Turin with a virtual reality headset displaying a 360° video of the tractor cabin and third-order ambisonic audio, the third with a virtual reality headset showing a 360° video of the tractor cabin and spatialized headphone audio, and the fourth with binaural headphone reproduction. The subjects, despite using different methodologies, listened to the same sounds and answered the same questions, presented with a 1-10 rating scale measuring annoyance and a semantic differential table with paired bipolar adjectives related to loudness, tonality, fluctuation strength, roughness, and sharpness. This also allowed for the evaluation of which reproduction method was the most accurate in terms of the accuracy of the prediction results. To avoid introducing too much randomness from the 60 different subjects, 8 subjects were selected to undergo the three tests conducted in the laboratory. The results are currently being analyzed.