Sound source localization accuracy in car-cabins: A human perception perspective

The aim of this paper is to present a human-oriented approach to define the minimum accuracy required for source localization techniques when used in car interiors in combination with auralization strategies. A ray-tracing model of a car cabin is used as data provider. By assigning a random distribution of the source positions around the ideal one, the associated reflections are then consequently estimated from a raytracing model of the cavity. Source and reflections signals are convolved with corresponding Head Related Transfer Functions (HRTFs) for generating data aimed at filling a data base for subjective evaluation. A listening experiment was conducted in order to understand the influence of both source localization accuracy and different reverberant environments (due to reflections) on human perception. A Minimum Audible Angle MAA is also proposed as a parameter to assess the Just Noticeable Difference (JND) for sound source localization measurements inside a car