To provide a method and a system to estimate a clean speech value by using an alternative sensor signal received from a sensor other than an air conduction microphone.
In order to estimate a clean speech value, either the alternative sensor signal alone or the alternative sensor signal in conjunction with the air conduction microphone signal is used. The clean speech value is estimated without using a model trained from noisy training data collected from an air conduction microphone.Under one embodiment, correction vectors are added to a vector formed from the alternative sensor signal in order to form a filter, which is applied to the air conductive microphone signal to produce the clean speech estimate. In other embodiment, the pitch of a speech signal is determined from the alternative sensor signal and is used to decompose an air conduction microphone signal. The decomposed signal is used to determine a clean signal estimate.
| WO/2013/078474 | PROCESSING SIGNALS |
| JP4955676 | Acoustic beam forming device and method |
| WO/2006/127968 | RESTORING AUDIO SIGNALS CORRUPTED BY IMPULSIVE NOISE |
DROPPO JAMES G
DENG LI
SINCLAIR MICHAEL J
HUANG XUEDONG DAVID
ZHENG YANLI
ZHANG ZHENGYOU
LIU ZICHENG
| JPH0879868A | 1996-03-22 | |||
| JPH09284877A | 1997-10-31 | |||
| JPH04245720A | 1992-09-02 | |||
| JPH08214391A | 1996-08-20 | |||
| JPH08223677A | 1996-08-30 | |||
| JP2000261530A | 2000-09-22 | |||
| JP2000250577A | 2000-09-14 |
Kazuo Abe
Next Patent: CONTINUOUS FLEXIBLE BELT MOUNTING APPARATUS