Sound source separating device and sound source separating method

• US 9,583,119 B2
• Filed: 06/13/2016
• Issued: 02/28/2017
• Est. Priority Date: 06/18/2015
• Status: Active Grant

• Current Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Original Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Inventors: Mizumoto, Takeshi, Nakadai, Kazuhiro, Nakamura, Keisuke
• Primary Examiner(s): Neway, Samuel G
• Application Number:
  US15/180,210
• Publication Number:
  US 20160372129A1
• Time in Patent Office: 260 Days
• US Class Current: 1/1
• CPC Class Codes: G10L 15/00 Speech recognition G10L17/00 takes precedence
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 2021/02166 Microphone arrays; Beamforming
• CPC Class Codes: G10L 21/028 using properties of sound source
• CPC Class Codes: G10L 25/84 for discriminating voice from noise

Abstract

A sound source separating device includes: a sound collecting unit configured to collect sound signals using a plurality of microphones; a sound source localizing unit configured to localize a sound source on the basis of the sound signals collected by the sound collecting unit; a sound source separating unit configured to separate the sound signals collected by the sound collecting unit into sound sources corresponding to the number of fixed sound sources and fixed sound source directions without using the sound source localization result from the sound source localizing unit; and a speech extracting unit configured to extract speech signals of the fixed sound sources on the basis of the sound source localization result from the sound source localizing unit and the sound source separation result from the sound source separating unit.

Claims

1. A sound source separating device comprising:

a sound collecting unit configured to collect sound signals using a plurality of microphones;

a sound source localizing unit configured to localize a sound source on the basis of the sound signals collected by the sound collecting unit;

a sound source separating unit configured to separate the sound signals collected by the sound collecting unit into sound sources corresponding to a number of fixed sound sources and fixed sound source directions without using the sound source localization result from the sound source localizing unit; and

a speech extracting unit configured to extract speech signals of the fixed sound sources on the basis of the sound source localization result from the sound source localizing unit and the sound source separation result from the sound source separating unit.

2. The sound source separating device according to claim 1, wherein the sound source localizing unit detects a speech section for each sound source, and

the speech extracting unit extracts speech signals in a period in which the fixed sound sources produce speech on the basis of the sound source localization result and the detected speech section from the sound source localizing unit and the sound source separation result from the sound source separating unit.

3. The sound source separating device according to claim 1, wherein the sound source separating unit separates the sound signals into the number of the fixed sound sources using a separation matrix and updates the separation matrix for every predetermined time even in a non-utterance period.

4. The sound source separating device according to claim 1, wherein the sound source separating unit separates the sound signals into the number of sound sources using a separation matrix and does not reset the separation matrix when speech is detected.

5. The sound source separating device according to claim 1, wherein the fixed sound sources are a driver in a driver seat and a passenger in a passenger seat of a vehicle.

6. The sound source separating device according to claim 5, wherein a sound source other than the driver and the passenger is additionally allocated to the fixed sound sources.

7. The sound source separating device according to claim 1, wherein the sound collecting unit is disposed between a driver seat and a passenger seat of a vehicle.

8. The sound source separating device according to claim 1, wherein the sound collecting unit is disposed between a map lamp and an instrument panel.

9. The sound source separating device according to claim 1, further comprising a sound source information storage unit configured to store information on the fixed sound sources,

wherein the sound source separating unit separates the sound signals into a number of the fixed sound sources using the information stored in the sound source information storage unit.

10. A sound source separating method comprising:

a sound collecting step of collecting sound signals using a plurality of microphones;

a sound source localizing step of localizing a sound source on the basis of the sound signals collected in the sound collecting step;

a sound source separating step of separating the sound signals collected in the sound collecting step into sound sources corresponding to the number of fixed sound sources and fixed sound source directions without using the sound source localization result in the sound source localizing step; and

a speech extracting step of extracting speech signals of the fixed sound sources on the basis of the sound source localization result in the sound source localizing step and the sound source separation result in the sound source separating step.