公开(公告)号：US09414157B2

公开(公告)日：2016-08-09

申请号：US14411651

申请日：2013-12-12

Applicant: Goertek Inc.

Inventor： Shasha Lou ,  Bo Li ,  Qiuchen Huang

IPC: H04B3/20 ,  H04R3/00 ,  G10L21/02 ,  G10L21/0216 ,  G10L21/0208

CPC classification number: H04R3/002 ,  G10L21/02 ,  G10L2021/02082 ,  G10L2021/02165 ,  H04R3/005 ,  H04R2225/43 ,  H04R2227/009

Abstract: The invention discloses a method and a device for reducing voice reverberation based on double microphones. The method comprises the steps of calculating a transfer function h(t) from a secondary microphone to a primary microphone according to an input signal x2(t) of the primary microphone and an input signal x1(t) of the secondary microphone; judging the strength of reverberation according to h(t) and calculating a regulatory factor β of a gain function by taking a tail section hr(t) of the h(t); obtaining a late reverberation estimation signal {circumflex over (r)}(t) of x2(t) with the convolution of x1(t) and hr(t); calculating the gain function according to the frequency spectrum of x2(t), β and frequency spectrum of {circumflex over (r)}(t); obtaining the reverberation removed frequency spectrum of x2(t) by multiplying the frequency spectrum of x2(t) by the gain function; and obtaining a late reverberation removed time-domain signal of x2(t) by frequency-time conversion. Thus, the late reverberation can be removed from the input signal of the primary microphone, early reverberation can be preserved, processed voice is not caused to be thin, and the voice quality is improved. Meanwhile, spectral subtraction intensity is adjusted according to the strength of the reverberation so as to ensure that the voice is not damaged on the condition that the reverberation is weak and the voice intelligibility is originally high. Accurate estimation of DOA of direct sound is not needed, and therefore the microphones are not required to have high consistency.

Abstract translation: 本发明公开了一种基于双麦克风来减少语音混响的方法和装置。 该方法包括根据主麦克风的输入信号x2（t）和次麦克风的输入信号x1（t）计算从次麦克风到主麦克风的传递函数h（t）的步骤; 根据h（t）判断混响的强度，并通过采用h（t）的尾段hr（t）计算增益函数的调节因子β; 以x1（t）和hr（t）的卷积获得x2（t）的后期混响估计信号（r（r）}（t））。 根据x2（t），β的频谱和{r（r）}（t）的频谱计算增益函数; 通过将x2（t）的频谱乘以增益函数获得混响去除x2（t）的频谱; 并通过频率 - 时间转换获得后期混响去除x2（t）的时域信号。 因此，可以从主麦克风的输入信号中去除后期混响，可以保持早期混响，处理的声音不会变薄，语音质量得到改善。 同时，根据混响的强度来调整频谱减法强度，以确保在混响弱并且语音清晰度最初高的条件下，声音不被损坏。 直接声音的DOA的精确估计是不需要的，因此麦克风不需要具有高的一致性。

公开(公告)号：US20150189431A1

公开(公告)日：2015-07-02

申请号：US14411651

申请日：2013-12-12

Applicant: Goertek Inc.

Inventor： Shasha Lou ,  Bo Li ,  Qiuchen Huang

IPC: H04R3/00

CPC classification number: H04R3/002 ,  G10L21/02 ,  G10L2021/02082 ,  G10L2021/02165 ,  H04R3/005 ,  H04R2225/43 ,  H04R2227/009

Abstract: The invention discloses a method and a device for reducing voice reverberation based on double microphones. The method comprises the steps of calculating a transfer function h(t) from a secondary microphone to a primary microphone according to an input signal x2(t) of the primary microphone and an input signal x1(t) of the secondary microphone; judging the strength of reverberation according to h(t) and calculating a regulatory factor β of a gain function by taking a tail section hr(t) of the h(t); obtaining a late reverberation estimation signal {circumflex over (r)}(t) of x2(t) with the convolution of x1(t) and hr(t); calculating the gain function according to the frequency spectrum of x2(t), β and frequency spectrum of {circumflex over (r)}(t); obtaining the reverberation removed frequency spectrum of x2(t) by multiplying the frequency spectrum of x2(t) by the gain function; and obtaining a late reverberation removed time-domain signal of x2(t) by frequency-time conversion. Thus, the late reverberation can be removed from the input signal of the primary microphone, early reverberation can be preserved, processed voice is not caused to be thin, and the voice quality is improved. Meanwhile, spectral subtraction intensity is adjusted according to the strength of the reverberation so as to ensure that the voice is not damaged on the condition that the reverberation is weak and the voice intelligibility is originally high. Accurate estimation of DOA of direct sound is not needed, and therefore the microphones are not required to have high consistency.

Abstract translation: 本发明公开了一种基于双麦克风来减少语音混响的方法和装置。 该方法包括根据主麦克风的输入信号x2（t）和次麦克风的输入信号x1（t）计算从次麦克风到主麦克风的传递函数h（t）的步骤; 根据h（t）判断混响强度，计算调节因子bgr; 通过获取h（t）的尾部hr（t）来获得增益函数; 以x1（t）和hr（t）的卷积获得x2（t）的后期混响估计信号（r（r）}（t））。 根据x2（t），＆bgr的频谱计算增益函数; （r）}（t）的频谱; 通过将x2（t）的频谱乘以增益函数获得混响去除x2（t）的频谱; 并通过频率 - 时间转换获得后期混响去除x2（t）的时域信号。 因此，可以从主麦克风的输入信号中去除后期混响，可以保持早期混响，处理的声音不会变薄，语音质量得到改善。 同时，根据混响的强度来调整频谱减法强度，以确保在混响弱并且语音清晰度最初高的条件下，声音不被损坏。 直接声音的DOA的精确估计是不需要的，因此麦克风不需要具有高的一致性。