Appearance, motion, and embodiment: unpacking avatars by finegrained communication analysis

摘要

Thanks to the maturity of motion tracking technologies, it became easy and inexpensive to use avatars thatrnreflect the user’s facial and body motions in real time. There is a possibility that those real-time avatars canrnserve as a substitute for videos in distant communication. We conducted two experiments to confirm positiverneffects of avatars on distant communication. In the first experiment, the subjects watched a conversationrnpartner’s video, avatar, photo, or nothing when speaking to the partner. Because the video and avatar deliveredrnthe partner’s motion while the video and photo delivered the partner’s appearance, we were able to observernthe effects of the motion and appearance separately. As a result, we found that presenting the motionrnvia a video or an avatar increased the degree of the smoothness of speaking to the partner, but the appearancerndid not significantly affect such an influence. In the second experiment, we compared the avatar andrnteleoperated robot to find how a physical embodiment influences the degree of the smoothness of speech.rnTeleoperated robots resemble avatars in transmitting the user’s body motions without disclosing the user’srnappearance but differ in reflecting these motions onto a robot that has a physical embodiment. We predictedrnthat the robot produces smoother speech because the physical embodiment might make the body motionsrnrecognizable. However, the robot decreased the degree of the smoothness of speech against our expectation.rnThe results of these experiments showed that real-time avatars that present only body motions are enough tornsmooth distant communication.