能動音響制御を利用した音韻性と韻律性の変更による音声マスキング

抄録

Sound masking is a technique used to make conversations less audible to others. This technique can be used to protect speech privacy in open spaces such as hospitals or shared offices. On the other hand, a conventional sound masking technique can be noisy after masking because it adds masking sound to the target sound. In this study, we investigate sound masking using active noise control, which is effective for noise reduction. We focused on the phonological and prosodic characteristics of speech and proposed a low-noise sound masking method by using active noise control to change both features. In the previous study, the masking effect was good for monosyllables, but the masking effect decreased when continuous sounds were controlled. In contrast, the proposed method considered prosodic properties, which was not taken into account in the previous study, in addition to phonological properties, and as a result, better masking effect was obtained even for continuous sounds. Furthermore, it was confirmed that the proposed method can reduce the sound volume after masking.