Chapter 4

Single-Channel Speech Enhancement in the Time Domain

Abstract

In this chapter, we study the problem of noise reduction in the time domain with a single sensor. In the first part, we assume that the speech correlation matrix is rank deficient. With this assumption, we show the importance of the speech subspace in the optimal filtering matrices. In the second part of this chapter, we revisit the problem of noise reduction without the previous assumption. This approach also leads to some interesting results which are slightly different as compared to the ones in the first part.

Keywords

Time domain; Single channel; Orthogonal decomposition; Performance measures; Optimal filtering matrices

In this chapter, we study the problem of noise reduction ...

Get Speech Enhancement now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.