Title:
SOUND SOURCE SIGNAL ESTIMATION DEVICE, SOUND SOURCE SIGNAL ESTIMATION METHOD, AND PROGRAM
Document Type and Number:
WIPO Patent Application WO/2021/100094
Kind Code:
A1
Abstract:
Provided is a sound source signal estimation technique capable of suppressing deterioration in sound quality by removing a crosstalk component in consideration of both an amplitude component and a phase component. βm, m'(ω) (1 ≤ m ≤ M, 1 ≤ m' ≤ M, m' ≠ m) is used as a coefficient that indicates the degree that a crosstalk component of a signal from an m'-th sound source is mixed in an m-th separation sound source signal ^Sm(f, ω). A crosstalk component removal unit includes: a coefficient calculation unit that, by solving a prescribed optimization problem with respect to a pair of m and m' which satisfy 1 ≤ m ≤ M, 1 ≤ m' ≤ M, and m' ≠ m, calculates the coefficient βm, m'(ω); and a crosstalk component removal signal calculation unit that, from the m-th separation sound source signal ^Sm(f, ω) (m = 1, ..., M), and using the coefficient βm, m'(ω), calculates an m-th estimation sound source signal ~Sm(f, ω) (m = 1, ..., M).
More Like This:
Inventors:
EMURA SATORU (JP)
Application Number:
PCT/JP2019/045120
Publication Date:
May 27, 2021
Filing Date:
November 18, 2019
Export Citation:
Assignee:
NIPPON TELEGRAPH & TELEPHONE (JP)
International Classes:
H04R1/40; G10L21/028; G10L21/0308; H04R3/00
Other References:
MUKAI, RYO ET AL.: "REMOVAL OF RESIDUAL CROSS-TALK COMPONENTS IN BLIND SOURCE SEPARATION USING TIME- DELAYED SUBTRACTION, in Proc.", ICASSP, 2 May 2002 (2002-05-02), pages 1789 - 1792, Retrieved from the Internet [retrieved on 20200406]
SAWADA, HIROSHI ET AL.: "MLSP 2007 DATA ANALYSIS COMPETITION: FREQUENCY-DOMAIN BLIND SOURCE SEPARATION FOR CONVOLUTIVE MIXTURES OF SPEECH/AUDIO SIGNALS", IEEE INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING(MLSP 2007, August 2007 (2007-08-01), pages 45 - 50, XP031199060, Retrieved from the Internet [retrieved on 20200406]
AICHNER R., ZOURUB M., BUCHNER H., KELLERMANN W.: "POST-PROCESSING FOR CONVOLUTIVE BLIND SOURCE SEPARATION", PROC. ICASSP, 5 May 2006 (2006-05-05), pages 37 - 41, XP010931283, Retrieved from the Internet [retrieved on 20200406]
SAWADA, HIROSHI ET AL.: "MLSP 2007 DATA ANALYSIS COMPETITION: FREQUENCY-DOMAIN BLIND SOURCE SEPARATION FOR CONVOLUTIVE MIXTURES OF SPEECH/AUDIO SIGNALS", IEEE INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING(MLSP 2007, August 2007 (2007-08-01), pages 45 - 50, XP031199060, Retrieved from the Internet
AICHNER R., ZOURUB M., BUCHNER H., KELLERMANN W.: "POST-PROCESSING FOR CONVOLUTIVE BLIND SOURCE SEPARATION", PROC. ICASSP, 5 May 2006 (2006-05-05), pages 37 - 41, XP010931283, Retrieved from the Internet
Attorney, Agent or Firm:
NAKAO, Naoki et al. (JP)
Download PDF: