Abstract

본 논문에서는 다양한 잡음 환경에서 음성향상을 위한 Minimum Statistics (MS) 잡음전력 추정 기법을 제시한다. 기존의 방법에서는 최소값 추적을 위해서 유한한 서치 (search)윈도우를 사용하여 최적으로 신호의 파워 스펙트럼을 수무딩하고 최소 확률을 적용하는 것을 기본으로 한다. 본 논문에서 제안된 알고리즘은 기존의 최소값 서치 윈도우가 다양한 잡음 환경에 상관없이 고정된 사이즈를 사용하는 것에 환경인식 정보를 적용하여 서치 윈도우 사이즈가 Gaussian mixture model(GMM)기반의 잡음 분류 알고리즘을 이용한 결과 값의 비교로 잡음 환경에 따라 변화 하도록 한다. 제안된 음성 향상 기법은 ITU-T P.862 perceptual evaluation of speech quality (PESQ)를 이용하여 평가하였고 기존의 MS방법보다 향상된 결과를 보였다. In this paper, we propose the improved noise power estimation in speech enhancement under various noise environments. The previous MS algorithm tracking the minimum value of finite search window uses the optimal power spectrum of signal for smoothing and adopts minimum probability. From the investigation of the previous MS-based methods it can be seen that a fixed size of the minimum search window is assumed regardless of the various environment. To achieve the different search window size, we use the noise classification algorithm based on the Gaussian mixture model (GMM). Performance of the proposed enhancement algorithm is evaluated by ITU-T P.862 perceptual evaluation of speech quality (PESQ) under various noise environments. Based on this, we show that the proposed algorithm yields better result compared to the conventional MS method.

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.