Collection of weak level sounds with multiple microphones

2021 Fiscal Year Final Research Report

• Project/Area Number: 19K12056
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 61010:Perceptual information processing-related
• Research Institution: Osaka Sangyo University
• Principal Investigator: Takahashi Toru 大阪産業大学, デザイン工学部, 教授 (30419494)
• Project Period (FY): 2019-04-01 – 2022-03-31
• Keywords: 音源追跡 / 音声強調 / 極小音 / マイクロホンアレイ / マルチマイクロホン / 分散マイクロホン / ビームフォーミング / 音源分離

Outline of Final Research Achievements

The objective of this research was to capture the sound produced by a single, extremely small sound source with multiple microphones. The research was based on the idea of emphasizing the sound source of interest by synchronously adding the extremely weak signal captured by each microphone in the time-frequency domain. Speech enhancement was performed by weighting according to the received sound level during synchronous addition. The extremely weak sound was difficult to distinguish from background sounds because it’s level was small even after the enhancement process. We aimed to realize a method to distinguish a very small sound from the background sound by taking advantage of the fact that when the target sound moves, the difference between the target sound and the background sound stands out. Finally, a method for tracking moving sound sources using multiple microphones, and a method for selecting microphones that can optimally collect sound as they move were developed.

Free Research Field

音声情報処理

Academic Significance and Societal Importance of the Research Achievements

極めて小さいレベルの音源が発する音を収音する技術は，マイクロホン素子の感度を改善することで行われてきた．本研究では，マイクロホン素子感度に依らず信号処理によって感度を改善しようという学術的に新しい試みであった．完全な時間同期が可能であれば，マイクロホン素子から得られる信号を強調することが理論的可能であった．現実には完全な時間同期は不可能であるため，時間周波数領域で不完全な時間同期でも強調する方法について検討した．本研究により単一のマイクロホン素子では収音が困難であった極めて小さいレベルの音の収音が可能になった．