論文アブストラクト：We propose a novel method of spatial audio rendering using ultrasound. An ultrasonic phased array generates one or more focal points in air, and they act as point sources of audible sound when the ultrasound waves are modulated. Our sound-point loudspeaker has two major advantages over conventional ultrasound-based sound-beam (superdirectional) loudspeakers. The higher audience selectivity means that our sound-point loudspeaker can deliver sound to the ears of the target person, whereas a sound-beam loudspeaker delivers sound to not only the target person but also other persons standing in the same direction. The other advantage is lower exposure to ultrasound; while an audible sound beam travels along an ultrasonic beam in a soundbeam loudspeaker, audible sound can be heard along the direction perpendicular to the ultrasonic beam in our soundpoint loudspeaker. This paper reports the principles of our sound-point loudspeaker, prototype construction, evaluation, and applications.