Identification of Sounds with Multiple Timbres

Abstract

The present research examined identification of complex sounds created by simultaneously playing two or more component sounds in various combinations. Sixteen component sounds were used, created by imposing four distinct temporal patterns on four basic timbres, two musical timbres and two complex real-world timbres. In the present experiment, complex sounds were created by simultaneously playing one to four component sounds, each with a different timbre. Subjects heard a complex sound, followed by a second complex sound that always differed from the first by adding a component, deleting a component or substituting a component. Subjects indicated which component had been added, deleted, or substituted. Sound changes were identified with moderate accuracy (above 60 percent). The errors committed varied with temporal pattern, timbre, sound change and density. The analyses of identification confusions indicated that subjects identified the correct timbre of the sound change even when temporal patterning was confused. The finding that temporal patterns were confused largely within the sound category of the correct response limits the previous interpretation of other research, which found that similar temporal patterns are confusable even with differences in spectra. Results of the present investigation suggest that multiple, temporal patterns with varying timbres can be presented from a single physical location to convey a change in state or status of an informative sound source. Design contributions of the present research to auditory information systems such as virtual reality are discussed. For such an application, a combination of physical separation and multiple patterns with varying timbres could provide a coherent, yet informationally complex, auditory display.