Method for inhibiting nonlinear crosstalk at microphone end of built-in communication equipment

Abstract

The invention belongs to the technical field of audio noise reduction, and particularly relates to a method for suppressing nonlinear crosstalk at a microphone end of built-in communication equipment. The suppression method specifically comprises the following steps: step 1, updating the weight of a filter based on an NLMS (Non-Least Mean Square) adaptive filtering algorithm; step 2, in a real-time audio processing process, calculating crosstalk data based on the filtering matrix and earphone end voice data; and step 3, subtracting the crosstalk data from the audio data collected by the microphone end to complete suppression of the nonlinear crosstalk of the microphone end. The built-in communication equipment is based on the NLMS adaptive filtering algorithm, the defects of a traditional audio noise reduction algorithm are effectively overcome, nonlinear crosstalk at the microphone end of the built-in communication equipment can be suppressed, the voice state definition of the whole machine in the flight state is greatly improved, and the voice quality is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of audio noise reduction, relates to crosstalk suppression of an intra-machine communication device, and specifically relates to a method for suppressing nonlinear crosstalk at a microphone end of an intra-machine communication device. Background Art

[0002] The in-flight communication equipment is a device that enables communication between passengers inside the aircraft. It is also an electronic device for people on board to communicate with people outside the aircraft when the aircraft is on the apron.

[0003] The AD acquisition chip and DA acquisition chip of the existing inter-machine communication equipment are located close to each other. The acquisition end of the AD acquisition chip is a small signal (mV), and the output end of the DA acquisition chip is a large signal (V). Since the large signal will radiate to the small signal acquisition end, nonlinear crosstalk will be generated and the voice quality will be reduced during the call.

[0004] In view of this, the present invention is proposed. Summary of the invention

[0005] The object of the present invention is to overcome the above-mentioned shortcomings of the prior art and provide a method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device to ensure that the voice quality is not affected.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The method for suppressing nonlinear crosstalk at the microphone end of the inter-phone communication device specifically comprises the following steps:

[0008] Step 1: Based on the NLMS adaptive filtering algorithm, update the filter weights;

[0009] Step 2: In the process of real-time audio processing, the crosstalk data is calculated based on the filter matrix and the voice data of the earphone end;

[0010] Step 3: Subtract the crosstalk data from the audio data collected at the microphone end to suppress the nonlinear crosstalk at the microphone end.

[0011] Specifically, the digital signal output from the first signal terminal of the intercom device is defined as X, and the digital signal crosstalked to the second signal terminal is defined as X1, then X1=X*W, wherein W is the filter weight.

[0012] Specifically, the step 1 includes the following steps:

[0013] Step 1.1, initialize the filter weights;

[0014] Step 1.2, according to the input signal of the filter, and in combination with the reference signal, calculate the output signal of the filter;

[0015] Step 1.3, calculating the error signal between the output signal of the filter and the expected signal;

[0016] Step 1.4, according to the error signal and the NLMS adaptive filtering algorithm, adjusting the filter weight to minimize the error;

[0017] Step 1.5: Repeat steps 1.2 to 1.4 until all input signals are trained and updated filter weights are obtained.

[0018] Specifically, in step 1.4, the NLMS adaptive filtering algorithm is based on the LMS filtering algorithm, and the step size factor is normalized to control the misalignment; and a new convergence factor is introduced to increase the convergence speed of the LMS filtering algorithm.

[0019] Specifically, when calculating the energy of the input signal, the input signal data of the filter at the current moment and several input signal data at the previous moment are combined into one frame and transferred to the frequency domain, and the mean square sum of the frequency domain amplitude in the frequency band of 300Hz to 3400Hz is used as the energy of the input signal. Among them, the number of data depends on the number of Fourier transform points. In addition, the in-flight communication equipment is used for communication between pilots, and the voice frequency is mainly concentrated in 300Hz to 3400Hz, that is, the energy of the effective voice in the frequency domain of the audio data collected by the microphone is mainly concentrated in the frequency band of 300Hz to 3400Hz, and the signal outside this frequency band can be considered as a noise signal. Therefore, in the NLMS filter training process, only the frequency band of 300Hz to 3400Hz is concerned, and the mean square sum of the amplitude of each frequency point is used as the input signal energy N, thereby greatly reducing the influence of low-frequency and high-frequency non-voice factors on the filter training, making the convergence speed faster and the robustness of the filter parameters stronger.

[0020] Furthermore, the mean square sum of the frequency domain amplitudes in the frequency band of 300 Hz to 3400 Hz is obtained by Fourier transform, and its mean is calculated as the energy of the input signal.

[0021] Specifically, when the error signal and the input signal are used to update the filter weights, the step factor is dynamically updated according to the energy size and changes dynamically with the input signal energy: the step factor increases with the increase of the input signal energy and decreases with the decrease of the input signal energy.

[0022] Specifically, in step 1.1, the initial coefficients of the filter are determined, and the step size factor μ is determined, where μ∈(0, 1].

[0023] Compared with the prior art, the technical solution provided by the present invention includes the following beneficial effects: the suppression method provided by the present invention, based on the NLMS adaptive filtering algorithm in the in-flight communication equipment, effectively makes up for the shortcomings of the traditional audio noise reduction algorithm, and can suppress the nonlinear crosstalk at the microphone end of the in-flight communication equipment, thereby greatly improving the clarity of the voice state of the entire aircraft when in flight, and improving the voice quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings are incorporated in and constitute a part of this specification and, together with the description, serve to explain the principles of the present invention.

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0026] Figure 1 A flow chart of a method for suppressing nonlinear crosstalk at a microphone end of an inter-phone communication device provided by the present invention;

[0027] Figure 2 A schematic diagram of interference at the microphone end of the inter-machine communication device provided by the present invention;

[0028] Figure 3 A schematic diagram of the relationship between a digital signal X output from a first signal terminal and a digital signal X1 output from a second signal terminal provided by the present invention;

[0029] Figure 4 A flow chart of completing the update of filter weights based on the NLMS adaptive filtering algorithm provided by the present invention;

[0030] Figure 5 The present invention provides a flow chart for suppressing nonlinear crosstalk at the microphone end of the inter-phone communication device. DETAILED DESCRIPTION

[0031] Here, exemplary embodiments will be described in detail, and the embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are only examples consistent with some aspects of the present invention described in detail in the appended claims.

[0032] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0033] Example

[0034] See also Figure 1As shown, this embodiment provides a method for suppressing nonlinear crosstalk at a microphone end of an inter-phone communication device, which specifically includes the following steps:

[0035] Step 1: Based on the NLMS adaptive filtering algorithm, update the filter weights;

[0036] Step 2: In the process of real-time audio processing, the crosstalk data is calculated based on the filter matrix and the voice data of the earphone end;

[0037] Step 3: Subtract the crosstalk data from the audio data collected at the microphone end to suppress the nonlinear crosstalk at the microphone end. Figure 5 .

[0038] Further, see Figures 2-3 As shown, the digital signal output from the first signal terminal (large signal terminal) of the intercom device is defined as X, and the digital signal output from the second signal terminal (small signal terminal) crosstalked is defined as X1, then X1 = X*W, where W is the filter weight.

[0039] Further, see Figure 4 As shown, the specific implementation process of step 1 in this embodiment is as follows:

[0040] Step 1.1, initialize the filter weight: define the initial coefficient of the filter as w(0), and determine the step size factor μ∈(0, 1];

[0041] Step 1.2: Calculate the output signal of the filter based on the input signal of the filter and the reference signal (referring to the filter coefficient); wherein,

[0042] Set the input signal x(n) and the expected signal d(n), and calculate the output signal of the filter according to the current filter coefficient. The specific calculation formula is:

[0043] y(n)=w T (n)x(n) (1)

[0044] In formula (1), w T (n) is the filter coefficient vector, x(n) is the input signal vector, and y(n) is the output signal;

[0045] Step 1.3: Calculate the error signal e(n) between the filter output signal y(n) and the desired signal d(n). The specific calculation formula is:

[0046] e(n)=d(n)-y(n) (2)

[0047] Step 1.4, according to the error signal and the NLMS adaptive filtering algorithm, adjusting the filter weight to minimize the error; wherein:

[0048] The energy of the input signal is calculated as follows: N(n) = f(x(n))(3)

[0049] In formula (3), f(x(n)) represents the Fourier transform of x(n) and several previous numbers x(n-1), x(n-2), ... and takes the sum of the squares of the amplitudes of the frequency points in the range of 300Hz to 3400Hz, and calculates the average value; where N(n) is the energy of the input signal at the current moment;

[0050] Use the error signal and input signal to update the filter weights, and dynamically update the step size factor μ according to the energy size. The specific calculation formula is as follows:

[0051]

[0052] In formula (4), μ max is the maximum step size factor, N(n) is the energy of the input signal at the current moment, N max is the maximum energy of the n input signals before N(n); in formula (5), w(n+1) is the filter coefficient at the next moment, w(n) is the filter coefficient at the current moment, e(n) is the error signal between the filter output signal y(n) and the expected signal d(n), and x(n) is the input signal vector.

[0053] Step 1.5: Repeat steps 1.2 to 1.4 until all input signals are trained (i.e., convergence conditions are met or the maximum number of iterations is reached), and obtain updated filter weights.

[0054] In this embodiment, the calculation method of the input signal energy is optimized, which is mainly reflected in the optimization of the normalized energy in NLMS. The square of the traditional time domain value is no longer used as energy, but the current data and the previous data are combined into a frame (the number of data depends on the number of Fourier transform points) and transferred to the frequency domain, and the mean square sum of the frequency domain amplitude in the 300Hz-3400Hz frequency band of the frequency domain is used as energy. The specific reason is that the in-flight communication equipment is used for communication between pilots, and the voice frequency is mainly concentrated in the 300Hz-3400Hz frequency band, that is, the energy of the effective voice in the frequency domain of the audio data collected by the microphone is mainly concentrated in the 300Hz-3400Hz frequency band, and the signal outside this frequency band can be considered as a noise signal. Therefore, in the NLMS filter training process, only the 300Hz-3400Hz frequency band is concerned, and the mean square sum of the amplitude of each frequency point is used as the input signal energy N, thereby greatly reducing the influence of low-frequency and high-frequency non-voice factors on the filter training, making the convergence speed faster and the robustness of the filter parameters stronger.

[0055] In this embodiment, the step factor changes dynamically with the energy of the input signal: that is, the step factor increases with the increase of the energy of the input signal, and decreases with the decrease of the energy of the input signal. This is because: the fixed step factor has poor adaptability to the signal. When the statistical characteristics of the input signal change, the step factor cannot be adjusted in time, which may cause the filtering performance to deteriorate. Especially in a high-noise environment, a fixed large step factor may cause the filter coefficient to diverge and fail to converge to the ideal value. The use of a dynamic step factor has strong adaptability and good stability. The step factor is dynamically adjusted according to the input signal energy, which can achieve a better balance between convergence speed and stability and improve filtering performance.

[0056] The suppression method provided by the present invention is based on the NLMS adaptive filtering algorithm in the in-flight communication equipment, which effectively makes up for the shortcomings of the traditional audio noise reduction algorithm, can suppress the nonlinear crosstalk at the microphone end of the in-flight communication equipment, greatly improves the clarity of the voice state when the whole aircraft is in flight, and improves the voice quality.

[0057] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention.

[0058] It should be understood that the present invention is not limited to what has been described above and that various modifications and changes may be made without departing from its scope. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device, characterized in that: The specific steps include: Step 1: Based on the NLMS adaptive filtering algorithm, update the filter weights; Step 2: In the process of real-time audio processing, the crosstalk data is calculated based on the filter matrix and the voice data of the earphone end; Step 3: Subtract the crosstalk data from the audio data collected at the microphone end to suppress the nonlinear crosstalk at the microphone end.

2. The method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device according to claim 1, characterized in that: Define the digital signal output from the first signal terminal of the intercom device as X, define the digital signal crosstalked to the second signal terminal as X1, then X1 = X*W, where W is the filter weight.

3. The method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device according to claim 1, characterized in that: The specific steps of step 1 are as follows: Step 1.1, initialize the filter weights; Step 1.2, according to the input signal of the filter, and in combination with the reference signal, calculate the output signal of the filter; Step 1.3, calculating the error signal between the output signal of the filter and the expected signal; Step 1.4, according to the error signal and the NLMS adaptive filtering algorithm, adjusting the filter weight to minimize the error; Step 1.5: Repeat steps 1.2 to 1.4 until all input signals are trained and updated filter weights are obtained.

4. The method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device according to claim 3, characterized in that: In step 1.4, the NLMS adaptive filtering algorithm is based on the LMS filtering algorithm, and the step size factor is normalized to control the misalignment; and a new convergence factor is introduced to increase the convergence speed of the LMS filtering algorithm.

5. The method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device according to claim 3, characterized in that: When calculating the energy of the input signal, the input signal data of the filter at the current moment and several input signal data at the previous moment are combined into one frame and transferred to the frequency domain, and the mean square sum of the frequency domain amplitudes in the frequency band of 300Hz to 3400Hz is used as the energy of the input signal.

6. The method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device according to claim 5, characterized in that: The mean square sum of the frequency domain amplitudes in the frequency band of 300 Hz to 3400 Hz is obtained by Fourier transform, and its mean is calculated as the energy of the input signal.

7. The method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device according to claim 5, characterized in that: When the error signal and the input signal are used to update the filter weights, the step factor is dynamically updated according to the energy level and changes dynamically with the input signal energy: the step factor increases with the increase of the input signal energy and decreases with the decrease of the input signal energy.

8. The method for suppressing nonlinear crosstalk at the microphone end of an inter-phone communication device according to claim 3, characterized in that: In step 1.1, the initial coefficients of the filter are determined, and the step size factor μ is determined, where μ∈(0, 1].

Images