JP2022011892A - Noise suppression circuit - Google Patents

Abstract

To provide a noise suppression circuit that is used in a noise reduction circuit suppressing a high frequency component of remaining noise displeasing to ears and reducing a displeasing level of the remaining noise.SOLUTION: A noise suppression circuit 7 has: a suppression coefficient calculation unit 71 that calculates a suppression coefficient Gi(f) for suppressing a noise component; an average suppression coefficient calculation unit 72 that calculates an average suppression coefficient GM serving as an average value of a suppression coefficient Gi(f) ranging from a high frequency threshold frequency f1 to a maximum frequency f2 of a frequency f of an amplitude spectrum; a first suppression unit 73 that multiplies the suppression coefficient Gi(f) with a signal corresponding to the amplitude spectrum for a frequency band less than the high frequency threshold frequency f1, and calculates a post-subtraction processed amplitude spectrum Si(f); and a second suppression unit 74 that multiplies the average suppression coefficient GM with the signal corresponding to the amplitude spectrum for a frequency band equal to and more than the high frequency threshold frequency f1, and calculates a post-subtraction processed amplitude spectrum Si(f).SELECTED DRAWING: Figure 2

Description

The present invention relates to a noise suppression circuit, for example, a noise suppression circuit that can be used in a noise reduction circuit incorporated in a radio that transmits and receives high frequency signals.

A spectral subtraction method is known as a method for suppressing a noise component contained in an audio signal (see, for example, Patent Document 1 and Non-Patent Document 1).

Japanese Unexamined Patent Publication No. 4-238399

P. Scalart and J. Vieira Filho "Speech Enhancement Based on a Priori Signal to Noise Estimation", IEEE International Conference on. Acoustics, Speech, Signal Progressing, Atlanta, GA, USA, vol. 2, pp. 629-632, 1996

By the way, there is a problem that there is a high frequency component in the residual noise after suppressing the noise component by a spectrum subtraction method or the like, and it is easy to feel that it is jarring. This problem is particularly noticeable in the non-voiced section.

Therefore, an object of the present invention is to provide a noise suppression circuit capable of appropriately suppressing a high frequency component of residual noise attached to the ear to reduce the degree of residual noise that is offensive to the ear.

In order to solve the above problems, the invention according to claim 1 is the amplitude for suppressing a noise component included in a signal corresponding to an amplitude spectrum including an amplitude component in the frequency spectrum of the signal to be processed. The suppression coefficient calculation unit that calculates the suppression coefficient for each frequency of the spectrum, and the high frequency threshold frequency that divides the frequency band of the amplitude spectrum into the high frequency band and other frequency bands to the frequency of the amplitude spectrum. Regarding the range up to the maximum frequency, the average suppression coefficient calculation unit that calculates the average suppression coefficient, which is the average value of the suppression coefficients for each frequency of the amplitude spectrum, and the frequency band below the high frequency threshold frequency. Regarding the first suppression unit that multiplies the suppression coefficient and the signal corresponding to the amplitude spectrum to calculate the amplitude spectrum after subtraction processing for each frequency of the amplitude spectrum, and the frequency band above the high frequency threshold frequency. Noise characterized by having a second suppression unit for multiplying the average suppression coefficient and a signal corresponding to the amplitude spectrum to calculate the amplitude spectrum after subtraction processing for each frequency of the amplitude spectrum. It is a suppression circuit.

The invention according to claim 2 is characterized in that, in the noise suppression circuit according to claim 1, the high frequency threshold frequency is a value in the range of 1.2 to 1.8 kHz. ..

According to the first aspect of the present invention, it is possible to appropriately suppress the high frequency component of the residual noise that is attached to the ear, and it is possible to reduce the degree of the residual noise that is offensive to the ear.

According to the invention according to claim 2, the frequency for dividing the frequency band of the amplitude spectrum into the high frequency band and the other frequency band can be appropriately set, and the suppression of the high frequency component attached to the ear of the residual noise is more appropriate. It becomes possible to do it.

It is a functional block diagram which shows the schematic structure of the noise reduction circuit including the noise suppression circuit which concerns on embodiment of this invention. It is a functional block diagram which shows the schematic structure of the noise suppression circuit which concerns on embodiment.

Hereinafter, the present invention will be described based on the illustrated embodiment.

FIG. 1 is a functional block diagram showing a schematic configuration of a noise reduction circuit 1 including a noise suppression circuit 7 according to an embodiment of the present invention. FIG. 2 is a functional block diagram showing a schematic configuration of the noise suppression circuit 7 according to the embodiment.

The noise reduction circuit 1 is, for example, a circuit incorporated in a radio device that transmits and receives high-frequency signals to realize a spectrum subtraction method (Spectral Reduction), which is a method of suppressing noise components contained in an audio signal, and is mainly a pre. Emphasis circuit 2, window processing unit 3, time-frequency conversion unit 4, conversion result output unit 5, voice section detection / noise update unit 6, noise suppression circuit 7, synthesis unit 8, frequency-time conversion unit 9 and a de-emphasis circuit 10.

The pre-emphasis (PE) circuit 2 performs high-frequency enhancement processing for pre-amplifying the relative intensity of high-frequency components on an audio signal demodulated from a high-frequency signal received from an antenna, and after high-frequency enhancement processing. Output a signal.

The window processing unit 3 receives the input of the signal after the high frequency enhancement processing output from the pre-emphasis circuit 2, and extracts a frame having a predetermined time length from the input signal (for example, every 12.5 ms). A time waveform for 25 ms is extracted), and each frame is subjected to window processing by multiplying it by a window function such as a Hanning window. The window processing unit 3 outputs the frame after the window processing each time the window processing is performed on each frame.

The time-frequency conversion unit 4 receives the input of the window-processed frame output from the window processing unit 3, and each time the input of the frame is received, the time-frequency conversion unit 4 changes the signal in the time domain to the signal in the frequency domain with respect to the frame. The conversion process is performed, a frequency spectrum including an amplitude component and a phase component for each of a plurality of frequencies is calculated, and a signal of the frequency spectrum of a real number and an imaginary number is output. The time-frequency transform unit 4 executes a time-frequency transform and calculates a frequency spectrum by, for example, a discrete Fourier transform or a fast Fourier transform.

The conversion result output unit 5 receives the input of the signal of the frequency spectrum for each frame (for example, at intervals of about 12.5 ms) output from the time-frequency conversion unit 4, and for each frame, the input of the frequency spectrum. A signal corresponding to the amplitude spectrum including the amplitude component of each frequency is output to the noise suppression circuit 7, and a signal corresponding to the phase spectrum including the phase component of each frequency in the input frequency spectrum is output. Output to the synthesizer 8.

The voice section detection / noise update unit 6 receives an input of a signal corresponding to the amplitude spectrum for each frame output from the conversion result output unit 5 and branched, and uses the input signal corresponding to the amplitude spectrum to be used. Update the noise spectrum that represents the noise component for each frequency.

The voice section detection / noise update unit 6 first determines, for each frame, whether the input signal corresponding to the amplitude spectrum is a frame containing only a noise component or a frame containing a voice component. ..

The method of determining whether the frame to be processed contains only the noise component or the voice component in the voice section detection / noise update unit 6 is not limited to a specific procedure or method, and is not limited to a specific procedure or method, and is a conventional or new procedure. An appropriate procedure or method can be appropriately selected from the methods and methods.

As a method of determining whether the frame to be processed contains only a noise component or a noise component in the voice section detection / noise update unit 6, for example, focusing on the non-constancy of the voice, each frequency of the amplitude spectrum is used. When the average or dispersion value related to the magnitude of the amplitude is less than a predetermined threshold multiple times (for example, about 3 to 5 times) in a row in the latest frame, it is determined that the frame to be processed is only a noise component. , A method of determining that the frame to be processed has an audio component in cases other than the above, or processing when the value of the average or dispersion regarding the magnitude of the amplitude of each frequency of the amplitude spectrum is less than a predetermined threshold value. A method may be used in which it is determined that the target frame has only a noise component, and when the average or dispersion value is equal to or greater than the threshold value, it is determined that the target frame has an audio component.

The voice section detection / noise update unit 6 subsequently adds the amplitude spectrum of the current frame (in other words, the frame to be processed, the latest frame) to the noise spectrum representing the noise component for each frequency calculated in the past, that is, , The latest noise spectrum is updated by adding the input signal corresponding to the amplitude spectrum.

The method of updating the noise spectrum representing the noise component for each frequency in the voice section detection / noise update unit 6 is not limited to a specific procedure or method, and is appropriate from conventional or new procedures and methods. Procedures and methods can be selected as appropriate.

As a method for updating the noise spectrum representing the noise component for each frequency in the voice section detection / noise update unit 6, for example, using the input signal corresponding to the amplitude spectrum, the updated noise spectrum Ni (f). Can be calculated according to the following formula 1 or formula 2 depending on whether the frame to be processed contains only a noise component or a sound component. The subscript i in the following mathematical formulas is an ordinal number representing the order of the time series, and represents the order that is commonly applied to all the mathematical formulas. Further, f in the following mathematical formulas represents the frequency in the input amplitude spectrum.

Specifically, when the frame to be processed contains only noise components, the input signal corresponding to the amplitude spectrum is used as the input signal spectrum Yi (f), which is an abbreviation for IIR (Infinite Impulse Response; infinite impulse response). ) The updated noise spectrum Ni (f) is calculated for each frequency f according to the following equation 1 which is a filter. Ni-1 (f) in Equation 1 and Equation 2 represents the noise spectrum one frame before the update.

K in Equation 1 is an input signal spectrum Yi (f) which is an amplitude spectrum of the frame to be processed when the frame to be processed (in other words, the current frame and the latest frame) is a frame containing only a noise component. It is a constant that determines the weighting of the noise spectrum Ni-1 (f) one frame before the update. The constant K is not limited to a specific value as long as it is an integer of 0 or more, and specifically, for example, a range corresponding to about 0.06 to 0.20 seconds of the time constant of the IIR filter (for example, It is conceivable that the value is set to any of K = 5 to 16 in the frame interval of 12.5 ms, and in particular, the value corresponding to the time constant of the IIR filter of about 0.1 second (for example). It is conceivable that K = 8) is set at a frame interval of 12.5 ms.

When the frame to be processed contains an audio component, the updated noise spectrum Ni (f) is determined for each frequency f according to the following mathematical formula 2.

The voice section detection / noise update unit 6 outputs the updated signal corresponding to the noise spectrum Ni (f) for each frequency f to the noise suppression circuit 7 for each frame.

The noise suppression circuit 7 receives the input of the signal corresponding to the amplitude spectrum for each frame output from the conversion result output unit 5, and the updated noise spectrum for each frame output from the voice section detection / noise update unit 6. In response to the input of the signal corresponding to, for each frame, from the signal corresponding to the input amplitude spectrum, it corresponds to the updated noise spectrum input for each frequency (in other words, for each spectrum). Subtract the signal.

The synthesizing unit 8 receives the input of the signal corresponding to the phase spectrum for each frame output from the conversion result output unit 5, and the signal corresponding to the amplitude spectrum after the subtraction processing for each frame output from the noise suppression circuit 7. Is received, and for each frame, the input signal corresponding to the phase spectrum and the signal corresponding to the amplitude spectrum are combined to generate a frequency spectrum, and a signal having a frequency spectrum of a real number and an imaginary number is output. do.

The frequency-time conversion unit 9 receives the input of the frequency spectrum signal for each frame output from the synthesis unit 8, and for each frame, the signal in the frequency region to the signal in the time region with respect to the input signal of the frequency spectrum. The audio signal is output by performing the conversion process to, that is, the reverse conversion process of the conversion process in the time-frequency conversion unit 4. The frequency-time conversion unit 9 executes frequency-time conversion by, for example, inverse discrete Fourier transform or inverse fast Fourier transform to generate an audio signal.

The De-Emphasis (DE) circuit 10 receives the input of the audio signal output from the frequency-time conversion unit 9, and attenuates the relative intensity of the high-frequency component with respect to the input audio signal. That is, the attenuation processing by the inverse filter of the pre-emphasis circuit 2 is performed, and the audio signal after the high frequency attenuation processing is output.

Then, the noise suppression circuit 7 according to the embodiment suppresses the noise component included in the signal corresponding to the amplitude spectrum including the amplitude component in the frequency spectrum of the signal to be processed, for each frequency f of the amplitude spectrum. The suppression coefficient calculation unit 71 that calculates the suppression coefficient Gi (f) of The average suppression coefficient calculation unit 72 for calculating the average suppression coefficient _GM , which is the average value of the suppression coefficients Gi (f) for each frequency f of the amplitude spectrum in the range up to the maximum frequency f2, and the high frequency threshold frequency f1. First suppression unit 73 for calculating the subtraction-processed amplitude spectrum Si (f) for each frequency f of the amplitude spectrum by multiplying the suppression coefficient Gi (f) and the signal corresponding to the amplitude spectrum for the frequency band less than And, for the frequency band above the high frequency threshold frequency f1, the average suppression coefficient _GM and the signal corresponding to the amplitude spectrum are multiplied to calculate the subtraction-processed amplitude spectrum Si (f) for each frequency f of the amplitude spectrum. It has 2 suppression portions 74 and.

The noise suppression circuit 7 is a signal corresponding to the updated noise spectrum for each frame output from the voice section detection / noise update unit 6 from the signal corresponding to the amplitude spectrum for each frame output from the conversion result output unit 5. It has a suppression coefficient calculation unit 71, an average suppression coefficient calculation unit 72, a first suppression unit 73, and a second suppression unit 74.

As an example of the processing in the suppression coefficient calculation unit 71, the suppression coefficient Gi (f) can be calculated for each frequency f by using the following mathematical formula 3.

Xi (f) in the formula 3 is a prior SNR (Signal to Noise Ratio), and is calculated according to the following formula 4.

The variables in Equation 4 are as follows.
Gi-1 (f): The suppression coefficient one frame before the update of the current (in other words, the latest) suppression coefficient Gi (f) in the time series. The initial value of the suppression coefficient Gi (f) is set to, for example, 1.
Γi (f): Subsequent SNR, calculated according to Equation 5 below.
Γi-1 (f): Current (in other words, latest) posterior SNR in time series The posterior SNR one frame before the update of Γi (f).
α: Also called a forgetting coefficient, 0 ≦ α <1, and is set to a value of, for example, about 0.999.
max []: An operator that selects the maximum value among multiple numerical values in [].

The variables in Equation 5 are as follows.
Yi (f): A signal corresponding to the amplitude spectrum for each frame output from the conversion result output unit 5 as the input signal spectrum.
λDi (f): Dispersion of the updated noise spectrum Ni (f) output for each frame from the voice section detection / noise update unit 6 for each frequency (that is, not the dispersion in the frequency direction, but in the time direction / frame direction). Dispersion).

The suppression coefficient calculation unit 71 calculates the suppression coefficient Gi (f) each time a signal corresponding to the amplitude spectrum for each frame is input.

The average suppression coefficient calculation unit 72 calculates the average suppression coefficient _GM according to the following equation 6 using the suppression coefficient Gi (f) calculated by the suppression coefficient calculation unit 71. The average suppression coefficient _GM is, that is, the average value of the suppression coefficient Gi (f) in the frequency axis direction in the range of frequencies f1 to f2.

In Equation 6, f1 is the high frequency threshold frequency, f2 is the maximum frequency in the amplitude spectrum (that is, f1 <f2), and f1 and f2 are specifically the times in the time-frequency conversion unit 4. It is a value of frequency (frequency bin) in frequency conversion. Further, Fn represents the number of frequencies (frequency bins) in the range from the high frequency threshold frequency f1 to the maximum frequency f2. The number of frequencies Fn is, that is, the number of frequencies (frequency bins) in the range from frequency f1 to frequency f2 in the time-frequency conversion in the time-frequency conversion unit 4.

The high frequency threshold frequency f1 is a frequency value for separating a high frequency component that becomes annoying noise as residual noise and other components (in other words, dividing the frequency band of the amplitude spectrum into a high frequency band and another frequency band). The value is not limited to a specific value, and is appropriately set to a value (or a value considered to be classified) that appropriately classifies the high frequency component and other components that cause particularly jarring noise. The high frequency threshold frequency f1 may be set to any value in the range of, for example, 1.2 to 1.8 kHz, and may be set to about 1.5 kHz in particular.

The average suppression coefficient calculation unit 72 calculates the average suppression coefficient _GM using the suppression coefficient Gi (f) calculated by the suppression coefficient calculation unit 71 each time a signal corresponding to the amplitude spectrum for each frame is input. do.

Then, the signal corresponding to the amplitude spectrum input to the noise suppression circuit 7 is set as the input signal spectrum Yi (f), and the following <a> or <b> is used for each frequency f according to the value of the frequency f. Processing is done.

<A> For frequency bands below the high frequency threshold frequency (that is, f <f1)
The first suppression unit 73 calculates the amplitude spectrum Si (f) after the subtraction process according to the following mathematical formula 7, that is, by multiplying the suppression coefficient Gi (f) and the input signal spectrum Yi (f).

<A> For frequency bands above the high frequency threshold frequency (that is, f ≧ f1)
The second suppression unit 74 calculates the amplitude spectrum Si (f) after the subtraction process according to the following formula 8, that is, by multiplying the average suppression coefficient _GM by the input signal spectrum Yi (f).

That is, for the high frequency component of the amplitude spectrum, the average suppression coefficient GM, which is the average value of the suppression coefficient _Gi (f) in the frequency axis direction, is used instead of the suppression coefficient Gi (f). The high-frequency component of the residual noise attached to the ear is appropriately suppressed, and the degree of the residual noise that is offensive to the ear is reduced.

Each time the noise suppression circuit 7 receives an input of a signal corresponding to the amplitude spectrum for each frame, the noise suppression circuit 7 outputs a signal corresponding to the amplitude spectrum Si (f) for each frequency f after the subtraction process.

According to the noise suppression circuit 7 as described above, the high frequency component of the residual noise attached to the ear can be appropriately suppressed, and the degree of the residual noise that is offensive to the ear can be reduced.

Although the embodiment of the present invention has been described above, the specific configuration is not limited to the above-described embodiment, and even if there is a design change or the like within a range that does not deviate from the gist of the present invention. Included in the invention. For example, in the above embodiment, the case where the noise suppression circuit 7 according to the present invention is applied to the noise reduction circuit 1 whose schematic configuration is shown in FIG. 1 is described as an example, but the present invention is applied. The possible configuration of the noise reduction circuit is not limited to the example shown in FIG. Furthermore, the circuits to which the present invention can be applied are not limited to noise reduction circuits. That is, the present invention can be applied to various circuits in which noise components are required to be suppressed.

Further, the main point of the present invention is to multiply the suppression coefficient and the input signal spectrum for the frequency band below the high frequency threshold frequency when suppressing the noise component contained in the signal corresponding to the amplitude spectrum as the input signal spectrum, while the high frequency threshold. For the frequency band above the frequency, the average value of the suppression coefficient in the frequency axis direction is multiplied by the input signal spectrum, and the method for calculating the suppression coefficient is limited to the procedure / method in the above embodiment. Any procedure or method may be used as long as the suppression coefficient for each frequency f is calculated.

1 Noise reduction circuit 2 Pre-emphasis circuit 3 Window processing unit 4 Time frequency conversion unit 5 Conversion result output unit 6 Voice section detection / noise update unit 7 Noise suppression circuit 71 Suppression coefficient calculation unit 72 Average suppression coefficient calculation unit 73 First suppression Part 74 Second suppression part 8 Synthesis part 9 Frequency time conversion part 10 De-emphasis circuit

Claims (2)

A suppression coefficient calculation unit for calculating the suppression coefficient for each frequency of the amplitude spectrum for suppressing the noise component contained in the signal corresponding to the amplitude spectrum including the amplitude component of the frequency spectrum of the signal to be processed, and the suppression coefficient calculation unit.
The frequency of the amplitude spectrum is the range from the high frequency threshold frequency, which is a frequency that divides the frequency band of the amplitude spectrum into a high frequency band and other frequency bands, to the maximum frequency of the frequencies of the amplitude spectrum. An average suppression coefficient calculation unit that calculates the average suppression coefficient, which is the average value of the suppression coefficients for each frequency,
A first suppression unit that calculates the amplitude spectrum after subtraction processing for each frequency of the amplitude spectrum by multiplying the suppression coefficient and the signal corresponding to the amplitude spectrum for the frequency band lower than the high frequency threshold frequency.
With respect to the frequency band above the high frequency threshold frequency, a second suppression unit that calculates the amplitude spectrum after subtraction processing for each frequency of the amplitude spectrum by multiplying the average suppression coefficient and the signal corresponding to the amplitude spectrum. , Have,
A noise suppression circuit characterized by this. The high frequency threshold frequency is any value in the range of 1.2 to 1.8 kHz.
The noise suppression circuit according to claim 1.

Images