There’s a car coming? - Psychometric function for car pass-by in background noise based on simulated data

Paper in proceeding, 2015

To detect an approaching car in background noise is an important aspect of traffic safety. Therefore it is essential to understand the determinants that makes people recognize an approaching car especially when the cars become very quiet at low speeds as it is the case for electrical cars for instance. Most studies on the detection of passenger cars in background noise are based on recorded signals. This requires that suitable recordings are available both for the background sound as well as the test sound (i.e. the sound of the approaching vehicle). Due to the limited control of such situations the degrees of freedom to be varied in such experiments is limited as well. In the presented study a different approach has been employed. The utilized sounds are based on an auralization method that allows for simulating vehicle sounds including both tyre/road noise and propulsion noise. Single car events can thus be superposed to background sound with full control of all relevant parameters. The used auralization method has in earlier studies been validated for giving good perceptual ratings compared to recorded sounds. The method allows for evaluating the psychometric functions for single parameters and hopefully give a deeper understanding of the perceptual space for a car in background noise. In the present study the reaction time is measured for the detection of a car (test vehicle) passing by in the presence of background noise from a road with high traffic flow. The distance between the the path of the test vehicle and the highly trafficked road is varied. All other parameters (i.e. car-type, road surface, speed, etc.) are kept constant. The study shows that there is a logarithmic relation between the response times and the distance between the the track of the test car and the road with high traffic volume. At the same time there is a linear relation between reaction time and signal to noise ratio (i.e. the equivalent sound pressure level in relation to the background level).