Amplitude-based speech enhancement with nonnegative matrix factorization for asynchronous distributed recording

Hironobu Chiba, Nobutaka Ono, Shigeki Miyabe, Yu Takahashi, Takeshi Yamada, Shoji Makino

Research output: Chapter in Book/Report/Conference proceedingConference contribution

24 Citations (Scopus)

Abstract

In this paper, we investigate amplitude-based speech enhancement for asynchronous distributed recording. In an ad-hoc microphone array context, it is supposed that different asynchronous devices record speech. As a result, the phase information is unreliable due to sampling frequency mismatch. For speech enhancement based on the amplitude information instead of the phase information, supervised nonnegative matrix factorization (NMF) is introduced in the time-channel domain. The basis vectors, which represents the gain of the transfer function from a source to each microphone, are trained in advance by using single source observation. The experimental evaluations show that this approach is well robust against the sampling frequency mismatch.

Original languageEnglish
Title of host publication2014 14th International Workshop on Acoustic Signal Enhancement, IWAENC 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages203-207
Number of pages5
ISBN (Electronic)9781479968084
DOIs
Publication statusPublished - 2014 Nov 11
Externally publishedYes
Event2014 14th International Workshop on Acoustic Signal Enhancement, IWAENC 2014 - Juan-les-Pins, France
Duration: 2014 Sep 82014 Sep 11

Publication series

Name2014 14th International Workshop on Acoustic Signal Enhancement, IWAENC 2014

Conference

Conference2014 14th International Workshop on Acoustic Signal Enhancement, IWAENC 2014
Country/TerritoryFrance
CityJuan-les-Pins
Period14/9/814/9/11

Keywords

  • ad-hoc microphone array
  • nonnegative matrix fac-torization
  • sampling frequency mismatch
  • Speech enhancement
  • time-frequency masking

ASJC Scopus subject areas

  • Signal Processing
  • Software
  • Acoustics and Ultrasonics
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Amplitude-based speech enhancement with nonnegative matrix factorization for asynchronous distributed recording'. Together they form a unique fingerprint.

Cite this