5:50 PM - 6:10 PM
[2O6-OS-2b-02] Can people hear the physical between from the interlocutor's voice called upon?
Toward demonstration of the existence of "the sense of distance", widely known in the theatrical world
Keywords:proxemics, voice, speech synthesis, human interface
In recent years, the proliferation of AI speakers has increased opportunities to interact with AI voices indoors and outdoors. In addition, with the development of deep learning, speech synthesis has emerged that is so natural that the reading voice is not easily distinguishable from a human. There is active research on speech synthesis of emotions to make AI speech more human-friendly. However, in the theatre world, the concept of 'distance' from the dialogue partner is widely known as one of the most critical factors in acting. Since there is no scientific basis for this "sense of distance," we conducted an experiment as a first step toward demonstrating "physical distance" to investigate whether people can tell when they are being called at different distances. We gathered four subjects, one playing the role of speaker and three playing the roles of listeners, and we placed the three listeners at different distances with their backs to the speaker based on the concept of personal space and asked the speaker to call out them in some set words. As a result, the listeners were generally able to guess who was being called.
Authentication for paper PDF access
A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.