Low Distortion Target Speech Extraction Method Using Two-Channel Microphone System
Takaaki Ishibashi, Kosuke Inoue, Chiharu Okuma, Kei Eguchi · 2018
This paper proposes a low distortion target speech signal extraction method. The authors have been proposed some target speech extraction methods using two-channel microphone system. The methods are based on a variable arbitrary directional pattern for multiple sound signals. And the pattern is directed to the target signal. However, when the directivity becomes sharp, the musical noise occurs in the estimated signal. Therefore, a new method makes a null pattern for the noise. The distortion of the estimated signal with the proposed method is low since the influence of the target sound is low. By several simulations, it is found that the proposed method can extract the target speech signals from multi speaker's utterance sounds under a noise.