JP2022038610A5

JP2022038610A5 -

Info

Publication number: JP2022038610A5
Application number: JP2020143203A
Authority: JP
Filing date: 2020-08-27
Publication date: 2023-08-18

Description

環境音を取得するための第一のマイクと、ノイズ源からの音を取得するための第二のマイクと、前記第一のマイクからの音声信号をフーリエ変換し、周波数領域の第一の音声信号を生成する第一の変換手段と、前記第二のマイクからの音声信号をフーリエ変換し、周波数領域の第二の音声信号を生成する第二の変換手段と、前記第二の音声信号と、前記ノイズ源のノイズに係る第一のパラメータとを演算し、ノイズデータを生成する生成手段と、前記ノイズデータを用いて、前記第一の音声信号に含まれる前記ノイズ源からのノイズを低減し、ノイズが低減された音声信号を出力するノイズ低減手段と、前記ノイズ低減手段から出力された前記ノイズ低減された周波数領域の音声信号を逆フーリエ変換し、ノイズが低減された時間領域の音声信号を出力する第三の変換手段と、前記第一の音声信号と前記第二の音声信号とを用いて、前記ノイズ源のノイズに係るパラメータを新たに生成し、新たに生成したパラメータを用いて前記パラメータを更新する更新手段と、を有し、前記生成手段は、前記更新手段により前記第一のパラメータが更新された場合、前記更新手段により更新された前記第一のパラメータと、前記第二の音声信号とを演算し、前記ノイズデータを生成する。
a first microphone for acquiring environmental sound; a second microphone for acquiring sound from a noise source; a first transformation means for generating a signal; a second transformation means for Fourier transforming an audio signal from said second microphone to generate a frequency domain second audio signal; and said second audio signal. and a generating means for generating noise data by calculating a first parameter related to noise of said noise source, and reducing noise from said noise source contained in said first audio signal using said noise data. a noise reduction means for outputting a noise-reduced audio signal; and an inverse Fourier transform of the noise-reduced frequency-domain audio signal output from the noise reduction means to obtain noise-reduced time-domain audio. Using a third conversion means for outputting a signal, the first audio signal and the second audio signal, newly generating a parameter related to the noise of the noise source, and using the newly generated parameter and updating means for updating the parameter by means of the updating means, and the generating means, when the first parameter is updated by the updating means, the first parameter updated by the updating means and the first The noise data is generated by computing the two audio signals.

Claims

a first microphone for capturing ambient sound;
a second microphone for capturing sound from a noise source;
a first transformation means for Fourier transforming the audio signal from the first microphone to generate a first audio signal in the frequency domain ;
a second transformation means for Fourier transforming the audio signal from the second microphone to generate a second audio signal in the frequency domain ;
generating means for generating noise data by computing the second audio signal and a first parameter related to the noise of the noise source;
noise reduction means for reducing noise from the noise source contained in the first audio signal using the noise data and outputting a noise-reduced audio signal;
third transforming means for performing an inverse Fourier transform on the noise-reduced frequency-domain audio signal output from the noise reduction means , and outputting a noise-reduced time-domain audio signal ;
Updating of generating new parameters related to noise of the noise source using the first audio signal and the second audio signal, and updating the first parameters using the newly generated parameters means and
has
When the first parameter is updated by the updating means, the generating means calculates the first parameter updated by the updating means and the second audio signal to generate the noise data. A voice processing device characterized by:

The updating means does not update the parameter when the value of the newly generated parameter is not smaller than the value of the first parameter.
2. The audio processing device according to claim 1, wherein:

When the amplitude of the environmental sound when the newly generated parameter is generated is smaller than the amplitude of the environmental sound when the first parameter is generated, the updating means updates the newly generated parameter to 3. The speech processing device according to claim 1, wherein the first parameter is updated using

the first parameter has a plurality of frequency spectrum values corresponding to the frequency spectrum of the second audio signal;
The speech processing apparatus according to any one of claims 1 to 3, wherein the updating means updates the first parameter for each frequency spectrum based on the newly generated parameter.

Having driving means as the noise source,
The updating means generates a parameter related to the noise of the noise source using the first audio signal and the second audio signal while the driving means is driving. The audio processing device according to any one of claims 1 to 4.

When the update means generates a plurality of parameters while the driving means is driving, the update means uses, among the plurality of parameters, a parameter for which the amplitude of the environmental sound when generated is small to update the first 6. The speech processing device according to claim 5, wherein processing for updating parameters is executed.

further comprising imaging means;
7. The sound processing apparatus according to claim 5, wherein the driving means drives when the image pickup means picks up an image.

8. The sound according to claim 7, wherein the updating means executes the updating after the power of the sound processing device is turned on and before recording of the image captured by the imaging means is started. processing equipment.

9. The apparatus according to any one of claims 5 to 8 , wherein said updating means changes said first parameter updated by said updating means to an initial value when said driving means is replaced. audio processor.

further comprising recording means;
10. The first parameter updated by the updating means is retained by the recording means even when the power of the speech processing device is turned off. 3. The audio processing device according to .

13. The audio processing apparatus according to any one of claims 1 to 12, wherein said first parameter is a ratio of amplitudes of said first audio signal and said second audio signal.

a first microphone for capturing ambient sound;
a second microphone for capturing sound from a noise source; and a control method for an audio processing device comprising:
Fourier transforming an audio signal from the first microphone to generate a first audio signal in the frequency domain ;
Fourier transforming the audio signal from the second microphone to generate a second audio signal in the frequency domain ;
a generation step of generating noise data by computing the second audio signal and a first parameter related to the noise of the noise source;
a noise reduction step of using the noise data to reduce noise from the noise source contained in the first audio signal and outputting a noise-reduced frequency domain audio signal;
performing an inverse Fourier transform on the noise-reduced frequency-domain audio signal output by the noise reduction step , and outputting a noise-reduced time-domain audio signal ;
Updating of generating new parameters related to noise of the noise source using the first audio signal and the second audio signal, and updating the first parameters using the newly generated parameters and
In the generating step, when the first noise parameter is updated in the updating step, the first noise parameter updated in the updating step and the second audio signal are calculated, and the noise A control method characterized by generating data.

A computer-readable program for causing a computer to function as each means of the speech processing apparatus according to any one of claims 1 to 11 .