Development of Games Using Real-Time Voice Input by Whisper and Unreal Engine

Abstract

While voice input is intuitive, it is not fully used in games that require real-time operation due to its inferior processing time and accuracy. Therefore, in order to investigate the current status and practicality of the game using realtime voice input, we are developing third-person shooter using Whisper, a machine learning model of speech recognition, and Unreal Engine, a widely used game engine. By reporting on the process and results of this development, we contribute to the widespread use of real-time voice input in games.