Listeners' ability to identify road-traffic, aircraft, or train sounds in environmental sound recordings was studied in a psychoacoustical experiment involving 16 participants. In free-labeling identification, excerpt traffic sounds were described in terms of "object" (sound-producing source) rather than in terms of perceptual attribute. The main sounds identified were traffic sounds, but a few references were also made to machine-related or water-related sources. Sounds from aircraft were easier to identify than the sounds from trains, which in turn were easier to identify than the sounds from road-traffic. This identification order was confirmed in multiple-choice and dominant-source identification tasks. Compared to free-labeling, multiple-choice identifications produced considerably more false alarms, i.e., identification of a sound source not present. For multiple-choice, several sound sources were particularly identified in the excerpt of road-traffic and train sounds although the (recorded) sound was typically clearly discerned in the joint dominant-source identification task. A comparison of the acoustic properties of the traffic sounds suggested that spectral rather than temporal cues were used in sound-source identification.
