A method of mood induction according to feedback of voice sound modified emotional prosody

2016 Fiscal Year Final Research Report

• Project/Area Number: 26540068
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Cognitive science
• Research Institution: Saga University
• Principal Investigator: Oshima Chika 佐賀大学, 工学(系)研究科(研究院), 客員研究員 (10395147)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 感情の誤認識 / リアルタイム音声変換 / 音量 / 音高 / 認知症

Outline of Final Research Achievements

Our research aimed to construct a system that helps people to improve their mood naturally using a theory that emotions can be caused by the false recognition of physiological reactions. The system needs to modify a person's voice in real time. The first, we surveyed what kind of acoustic features are modified for emotion recognition in speech from related researches. There were dozens of kinds of acoustic features used for these researches. However, most of the acoustic features are not adequate to modify the voice in real time. Therefore, we developed the system that modifies pitches and volumes of real-time speech. We described 24 algorithms for converting sounds. Then, we conducted an experiment in which subjects read an essay while listening to their individual utterance converted by the system in real time. The results for the items calm, lazy, and comfortable show that the sounds converted like an equalizer were more agreeable than those converted the whole pitch.

Free Research Field

ヒューマンコンピュータインタラクション