Auditory room size perception for modeled and measured rooms

Although there are many studies investigating auditory spatial impression in rooms, there are few that directly investigate the seemingly more basic question of auditory room size perception. In this study, subjective experiments using the method of paired comparisons were conducted to obtain room size ratings using binaurally presented stimuli. In the first experiment, binaural impulse responses from a computer-modelled room were used to auralize an anechoic speech sample. Room volume, source-receiver distance and reverberation time were investigated as parameters. The second experiment used binaural recordings of speech made in a real room of fixed size (for the same anechoic speech sample as Experiment I), with source-receiver distance and reverberation time as experiment parameters. The final experiment used binaural impulse responses of a concert auditorium convolved with anechoic music - so that both the room volume and reverberation time were constant. Results show that reverberation time strongly affects room size perception (much more so than the physical room volume). In a room of fixed volume but variable absorption and source-receiver distance, clarity index can be a good predictor of perceived room size. A comparison of Experiments I and II shows little or no difference between results for auralizations of computer modelled rooms and binaural reproductions of a real room. Results from the second and third experiments were compared with results from previous studies (of auditory distance perception, speech quality and spatial impression) which used identical stimuli. Auditory room size perception is not closely related to auditory distance perception (Experiment II), and is related to auditory intimacy (music stimuli, Experiment III) and speech quality (speech stimuli, Experiment II).