A device that transcripts multiple voices from multiple sound sources for guaranteed access to information for the deaf

Abstract

We investigated the usefulness of a new voice transcription device (VUEVO; Pixie Dust Technologies, Inc.,) that can simultaneously transcribe multiple voice inputs from multiple directions by localizing each sound source. In conference meetings and lectures, VUEVO was able to localize the voice, and the results of the transcription were independently shown on the screen. When a large screen was available for closed captioning, VUEVO was able to place each localized text near the corresponding position, making it easier to understand who said what. This interface was perceived positively by the deaf audience compared to the conventional caption system that showed all transcriptions at the same place on the screen. Thus, this new device shows promise for more efficient information assurance to help deaf and hard-of-hearing people.