Synthesis of footstep sounds of crowd from single step sound based on cognitive property of footstep sounds

The crowd sound effect ("Gaya" in Japanese technical word) plays important role to create and perceive the atmosphere of the crowded scene of a movie, but the technique for authoring "Gaya" sound has not been scientifically described so far. In this study, footstep sounds of the crowd was focused and cognitive characteristics of footstep sounds was described through 3 psychophysical experiments, applying to the technique for authoring effective footstep sounds of crowd. In the experiment 1, the effect of the varying tempo of recorded footstep sounds to the reality of "walking" or "running" cognition of footstep sound was examined, revealing that there was most effective tempo for the reality of cognition of "walking" and "running" footstep sounds. In the experiment 2, the interaction effect between footstep duration and inter-step interval in the footstep sound to "walking" or "running" cognition of the footstep sound was examined, indicating the main contribution of inter-step interval to the discrimination between "walking" and "running" cognition. In the experiment 3, the perceived number of "walkers" from the hearing of the footstep sounds as a function of physical number of those was examined, indicating the non-monotonic increase and the underestimation of perceived number of "walkers" of footstep sounds. In conclusion, the footstep sound can be synthesized by merely lining recorded single step sound (element sound) alternating with appropriate duration of silence (inter-step interval) and the discrimination between "walking" cognition and "running" cognition was mainly depend on inter-step interval (silence duration). However, the footstep sounds of the crowd could not be synthesized merely by stacking footstep sound for one person because of the underestimation of the perceived number of "walkers" from the hearing of the footstep sounds in the experiment 3, suggesting that there are some kind of direct representation for crowd footstep sounds itself such as appropriately filtered noise. In fact, sound effect engineer\´s empirical technique to synthesize "Gaya" sound (crowd sound effect) is known that effective crowd footstep sounds consist of direct crowd sound (such as filtered noise) as background and a few footstep sounds as foreground. Thus, this study could explain those empirical knowledge.