US 12,131,747 B2
Voice signal processing apparatus and noise suppression method
Ryuichi Namba, Tokyo (JP); Seiji Miyama, Tokyo (JP); Yoshihiro Manabe, Tokyo (JP); and Yoshiaki Oikawa, Tokyo (JP)
Assigned to SONY CORPORATION, Tokyo (JP)
Appl. No. 17/283,398
Filed by SONY CORPORATION, Tokyo (JP)
PCT Filed Aug. 23, 2019, PCT No. PCT/JP2019/033029
§ 371(c)(1), (2) Date Apr. 7, 2021,
PCT Pub. No. WO2020/079957, PCT Pub. Date Apr. 23, 2020.
Claims priority of application No. 2018-194440 (JP), filed on Oct. 15, 2018.
Prior Publication US 2021/0343307 A1, Nov. 4, 2021
Int. Cl. G10L 21/00 (2013.01); G10L 21/0216 (2013.01); H04R 1/32 (2006.01)
CPC G10L 21/0216 (2013.01) [H04R 1/326 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A voice signal processing apparatus, comprising:
a Central Processing Unit (CPU) configured to:
obtain an input voice signal via a microphone arranged at a sound reception point;
acquire installation environment information that includes information regarding a type of noise and an orientation between the sound reception point and a noise source;
acquire noise dictionary data from a noise database based on the installation environment information,
wherein the noise dictionary data includes a directional characteristic of the noise source; and
perform a noise suppression processing on the input voice signal based on the noise dictionary data to output a voice signal as a noise-reduced output.
 
20. A noise suppression method comprising:
in a voice signal processing apparatus that includes a Central Processing Unit (CPU):
obtaining, by the CPU, an input voice signal via a microphone arranged at a sound reception point;
acquiring installation environment information that includes information regarding a type of noise and an orientation between the sound reception point and a noise source;
acquiring, by the CPU, noise dictionary data read out from a noise database based on the installation environment information;
wherein the noise dictionary data includes a directional characteristic of the noise source; and
performing a noise suppression processing on the input voice signal based on the noise dictionary data to output a voice signal as a noise-reduced output.