Perceptively based design of new car horn sounds

Abstract

Due to the technologies used, there exist only a few different kinds of car horn sounds. We propose a method for the design of new sounds, based on a perceptual study of the actual sounds of car horns. Firstly we deal with recordings of existing car horns. We show that the different kinds of horn sounds can be divided into nine main families. Within these families we demonstrate secondly that the perception of timbre results from the integration of three elementary sensations. Thirdly, another experiment reveals that some sounds are better identified as car horns than others. A relationship between the perceived timbre of the sounds and their ability to be identified as car horns is established. Finally, we generalize our results to synthesized sounds. The synthesis method was designed to explore and enlarge the perceptual space. Studying these sounds confirms and generalizes the previous results. A model is proposed, which is able to synthesize sounds and predict their ability to be identified as car horns.