KaraokeVUI: Utilizing Karaoke Subtitles for Voice User Interfaces to Navigate Users What They Would Say

Abstract

Voice user interfaces (VUIs) have the problems in discoverability and learnability. Conventional VUIs tried to solve these problems by listing voice commands for using features on the screen of devices in combination with GUIs. However, with these VUI help tools, it is difficult to know how accurately the commands must be spoken and to determine whether the target operation is being executed if the words are rephrased or spoken incorrectly during speaking. Therefore, users are too preoccupied with speaking phrases correctly, and the advantage of VUIs, which can be used without awareness of the operation method, has not been fully demonstrated. To address these issues, we propose KaraokeVUI, which is a VUI help tool for supporting voice operation with feedback by displaying spoken words through filling in the blanks or overlaying on phrases, just like on a karaoke screen. In this paper, we evaluate and verify KaraokeVUI usefulness and usability.