Call noise reduction method and device and earphone
By acquiring and fusing the characteristics of microphone signals and using neural networks to perform call noise reduction processing, the problem of insufficient vocal resolution capabilities in traditional technology is solved, and the noise reduction effect is significantly improved.
Patent Information
- Application Number
- CN202310538786.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-12
- Publication Date
- 2024-11-12
AI Technical Summary
The traditional double-mill noise reduction solution has weak vocal resolution capabilities and cannot effectively reduce environmental noise and other vocals, resulting in limited noise reduction capabilities.
By acquiring the echo cancellation reference signal, the first noise reduction reference signal received by the first microphone and the call signal received by the second microphone, the features are extracted and fused, and the noise reduction processing is performed using complex convolutional networks and convolutional neural networks, combined with The prediction network performs frequency band weighting processing to obtain the noise-reduced call signal.
Effectively distinguishing human voices significantly improves the suppression ability of call noise reduction, has a good suppression effect on non-human noise noise, and improves the noise reduction effect.
Smart Images

Figure CN118945526A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of audio processing, and in particular to a call noise reduction method, device and headset. Background Art
[0002] In the traditional dual-microphone noise reduction solution, dual-microphone traditional beamforming + AEC (echo cancellation) + single-channel AI noise reduction is generally used. However, the traditional beam has a weak ability to distinguish human voices and cannot effectively reduce environmental noise and other human voices, resulting in the limited noise reduction ability of the existing noise reduction solution. Summary of the invention
[0003] The present application mainly provides a call noise reduction method, device and headset, which solves the problem of poor call noise reduction effect in the prior art.
[0004] To solve the above technical problems, the first aspect of the present application provides a call noise reduction method, including: obtaining an echo cancellation reference signal, a first noise reduction reference signal received by a first microphone, and a call signal received by a second microphone; extracting a first fusion feature of the first noise reduction reference signal and the call signal, and extracting the echo signal feature of the echo cancellation reference signal; fusing the first fusion feature with the echo signal feature to obtain a second fusion feature; and performing noise reduction processing on the call signal using the second fusion feature to obtain a noise-reduced call signal.
[0005] Optionally, extracting the first fusion feature of the first noise reduction reference signal and the call signal includes: using a first complex convolutional network to process the first noise reduction reference signal and the call signal to obtain the first noise reduction signal feature and the call signal feature, respectively.
[0006] Optionally, extracting the echo signal feature of the echo cancellation reference signal includes: processing the echo cancellation reference signal using a second complex convolutional network to obtain the echo signal feature.
[0007] Optionally, the fusing of the first noise reduction signal feature, the call signal feature and the echo signal feature to obtain a second fused feature includes: concatenating the first fused feature and the echo signal feature and performing a modulo operation to obtain the second fused feature, where the second fused feature is a real number feature.
[0008] Optionally, the use of the second fusion feature to perform noise reduction processing on the call signal includes: using a convolutional neural network to process the second fusion feature to obtain the second fusion feature after convolution processing; using a prediction network to process the second fusion feature after convolution processing to obtain probability results corresponding to each frequency band; converting the call signal into a frequency domain signal, taking the weighted sum of the frequency domain signals falling into each frequency band with the probability as the weight, and converting the sum result to the time domain to obtain the noise reduced call signal.
[0009] Optionally, the method also includes: obtaining a second noise reduction reference signal received by a third microphone, and extracting a second noise reduction signal feature of the second noise reduction reference signal; fusing the first fusion feature and the echo signal feature to obtain a second fusion feature, including: concatenating the first fusion feature, the echo signal feature and the second noise reduction signal feature and performing a modulo operation to obtain the second fusion feature.
[0010] Optionally, extracting the second denoised signal feature of the second denoised reference signal includes: processing the second denoised reference signal using a third complex convolutional network to obtain the second denoised signal feature.
[0011] Optionally, the first fusion feature includes phase difference information between the first noise reduction reference signal and the call signal and amplitude information corresponding to the first noise reduction reference signal and the call signal respectively.
[0012] To solve the above technical problems, the second aspect of the present application provides a call noise reduction device, comprising a processor and a memory coupled to each other; a computer program is stored in the memory, and the processor is used to execute the computer program to implement the call noise reduction method provided in the first aspect.
[0013] To solve the above technical problems, the third aspect of the present application provides an earphone, which includes a first microphone, a second microphone, and a processor; the first microphone is used to receive a first noise reduction reference signal, and the second microphone is used to receive a call signal; the processor is used to perform noise reduction processing on the call signal using the first noise reduction reference signal according to the call noise reduction method provided by the first aspect above.
[0014] The beneficial effect of the present application is: different from the prior art, the present application first obtains an echo cancellation reference signal, a first noise reduction reference signal received by a first microphone, and a call signal received by a second microphone, and then extracts a first fusion feature of the first noise reduction reference signal and the call signal, extracts the echo signal feature of the echo cancellation reference signal, fuses the first fusion feature with the echo signal feature to obtain a second fusion feature, and uses the second fusion feature to perform noise reduction processing on the call signal to obtain a noise-reduced call signal. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0016] Figure 1 This is a schematic flow chart of an embodiment of a method for reducing call noise in the present application;
[0017] Figure 2 This is a schematic flow chart of an embodiment of step S14 of the present application;
[0018] Figure 3 This is a schematic flow chart of another embodiment of the call noise reduction method of the present application;
[0019] Figure 4 This is a schematic structural block diagram of an embodiment of a call noise reduction device of the present application;
[0020] Figure 5 It is a structural schematic block diagram of another embodiment of the call noise reduction device of the present application;
[0021] Figure 6 It is a structural schematic block diagram of another embodiment of the call noise reduction device of the present application;
[0022] Figure 7 is a schematic structural block diagram of an embodiment of the earphone of the present application;
[0023] Figure 8 It is a schematic block diagram of the structure of an embodiment of a computer-readable storage medium of the present application. DETAILED DESCRIPTION
[0024] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0025] The terms "first" and "second" in this application are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first" and "second" can explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices.
[0026] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0027] See also Figure 1 , Figure 1 1 is a flowchart of an embodiment of the call noise reduction method of the present application. It should be noted that if there is substantially the same result, this embodiment does not take Figure 1 The process sequence shown is limited. This embodiment includes the following steps:
[0028] The methods disclosed in the various embodiments of the present application may be applied to a communication device including a speaker and a microphone, such as a headset.
[0029] Step S11: Acquire an echo cancellation reference signal, a first noise reduction reference signal received by the first microphone, and a call signal received by the second microphone.
[0030] The echo cancellation reference signal is generally referred to as an AEC reference signal, which is the sound from the speaker transmitted back to the other end of the call through the microphone. The echo cancellation reference signal can be generated by subtracting the audio of the speaker from the audio signal received by the microphone.
[0031] This embodiment is a dual-microphone embodiment, wherein the first microphone is an external microphone, mainly used to receive environmental sound signals, and the second microphone is a call microphone, mainly used to receive call sound signals.
[0032] Step S12: extracting a first fusion feature of the first noise reduction reference signal and the call signal, and extracting an echo signal feature of the echo cancellation reference signal.
[0033] In one embodiment, a first complex convolutional neural network may be used to process the first noise reduction reference signal and the call signal, and output a first fusion feature of the call signal and the first noise reduction reference signal.
[0034] The first fusion feature includes phase difference information between the first noise reduction reference signal and the call signal, and amplitude information corresponding to the first noise reduction reference signal and the call signal. Specifically, the output result of the first complex convolutional neural network is a complex number, and the phase difference information and amplitude information can be determined by the real part and imaginary part of the complex number.
[0035] The second complex convolutional neural network can be used to process the echo cancellation reference signal to obtain the echo signal feature.
[0036] Step S13: Fusing the first noise reduction signal feature, the call signal feature and the echo signal feature to obtain a second fusion feature.
[0037] Optionally, the first noise reduction signal feature and the call signal feature are concatenated and then modulo-operated to obtain a second fusion feature in real form. Specifically, the first noise reduction signal feature and the call signal feature obtained by processing with a complex convolutional network are both complex results and can be directly concatenated and modulo-operated.
[0038] This embodiment converts the signal into real form after complex convolution, so that the subsequent real signal integrates phase difference information and amplitude information. Compared with completing the algorithm in the complex domain, the operation memory usage and the amount of calculation are greatly reduced. Using this model with the same amount of calculation can greatly improve the noise reduction capability.
[0039] Step S14: Use the second fusion feature to perform noise reduction processing on the call signal to obtain a noise-reduced call signal.
[0040] See also Figure 2 , Figure 2 is a flowchart of an embodiment of step S14 of the present application. It should be noted that if there is substantially the same result, this embodiment does not necessarily Figure 2 The process sequence shown is limited. This embodiment includes the following steps:
[0041] S141: Process the second fusion feature using a convolutional neural network to obtain the second fusion feature after convolution processing.
[0042] In this step, the second fusion feature is processed to obtain a feature expression of the audio signal represented by the feature in multiple dimensions.
[0043] S142: Using the prediction network to process the second fused features after the convolution process, to obtain probability results corresponding to each frequency band.
[0044] The second fused features after convolution processing include feature representations of the audio signal in multiple dimensions. The second fused features after convolution are input into a pre-trained prediction network to output multiple probability results corresponding to multiple frequency bands.
[0045] Specifically, the call signal, the first noise reduction reference signal and the echo cancellation reference signal are all sampled according to a set length, for example, a signal sample is divided every 10 seconds, and the signal frequency band is predicted according to this step.
[0046] The frequency band division method can be set according to the needs and is not limited here. For example, the frequency around 1000 Hz is likely to be noise. If this band is set to a lower probability value, most of the noise can be removed, and the noise reduction effect is good.
[0047] Optionally, the prediction network is a gated recurrent neural network.
[0048] S143: Convert the call signal into a frequency domain signal, perform weighted summation on the frequency domain signals falling into each frequency band using probability as a weight, convert the summation result into the time domain, and obtain the call signal after noise reduction.
[0049] This step converts the call signal into a frequency domain signal to obtain the part of the call signal falling in each frequency band. The call signal falling in each frequency band is weighted and summed according to the corresponding probability and then converted to the time domain to obtain the noise-reduced call signal.
[0050] Different from the prior art, this embodiment uses a first complex convolutional neural network to extract the features of the first noise reduction reference signal and the call signal, uses a second complex convolutional neural network to extract the features of the echo cancellation signal, and uses a neural network to process the signal features to complete the noise reduction operation of the call signal. It uses an end-to-end neural network model for processing, has a simple structure, and does not require additional beamforming and echo cancellation schemes. Different complex convolutional neural networks are used to fuse the signals of different microphones, respectively, and the input signals are phase and amplitude encoded, which can effectively distinguish human voices and has a good suppression capability for non-human voice noise.
[0051] In another embodiment, three microphones may be included: a first microphone, a second microphone and a third microphone. The first microphone is an external microphone, mainly used to receive ambient sound signals; the second microphone is a call microphone, mainly used to receive call sound signals; the third microphone is an in-ear microphone, which is arranged inside the earphone and mainly used to receive audio signals inside the earphone. The received audio signals have the characteristics of high signal-to-noise ratio. Figure 3 , Figure 3 This is a schematic flow chart of another embodiment of the call noise reduction method of the present application. This embodiment may include the following steps:
[0052] S21: Acquire an echo cancellation reference signal, a first noise reduction reference signal received by the first microphone, and a call signal received by the second microphone, and acquire a second noise reduction reference signal received by the third microphone.
[0053] The third microphone is located inside the earphone, and the received audio signal consists mostly of human voice signals, with only a very small amount of ambient noise signals.
[0054] S22: Extracting a first fusion feature of the first noise reduction reference signal and the call signal, an echo signal feature of the echo cancellation reference signal, and a second noise reduction signal feature of the second noise reduction reference signal.
[0055] Optionally, a first complex convolutional neural network is used to process a first noise reduction reference signal and a call signal, and a first fusion feature of the call signal and the first noise reduction reference signal is output; a second complex convolutional neural network is used to process an echo cancellation reference signal to obtain an echo signal feature; and a third complex convolutional neural network is used to process a second noise reduction reference signal to obtain a second noise reduction signal feature.
[0056] S23: concatenate the first fusion feature, the echo signal feature, and the second noise reduction signal feature, and perform a modulo operation to obtain a second fusion feature.
[0057] The first fusion feature, echo signal feature and second noise reduction signal feature obtained by complex convolution network processing are all in complex form. After direct concatenation and modulo taking, the second fusion feature in real form can be obtained.
[0058] S24: Use the second fusion feature to perform noise reduction processing on the call signal to obtain a noise-reduced call signal.
[0059] The process is the same as step S14 and will not be described again here.
[0060] Different from the prior art, this application uses a neural network to process multiple audio signals from headphones, and through feature extraction and audio probability calculation, the original call signal is weighted by probability to complete the noise reduction operation. The model can be flexibly adjusted. When increasing or decreasing the number of microphones, it is only necessary to adjust the complex convolution method according to the characteristics of the microphone signal. The subsequent network structure remains basically unchanged, and the network structure is simple and easy to adjust.
[0061] See also Figure 4 , Figure 4 The structure schematic diagram of an embodiment of a call noise reduction device of the present application is shown in FIG. The call noise reduction device 100 comprises an acquisition module 110 , a feature extraction module 120 , a fusion module 130 and a noise reduction module 140 .
[0062] Among them, the acquisition module 110 is used to obtain the echo cancellation reference signal, the first noise reduction reference signal received by the first microphone, and the call signal received by the second microphone; the feature extraction module 120 is used to extract the first fusion feature of the first noise reduction reference signal and the call signal, and extract the echo signal feature of the echo cancellation reference signal; the fusion module 130 is used to fuse the first fusion feature with the echo signal feature to obtain the second fusion feature; the noise reduction module 140 is used to use the second fusion feature to perform noise reduction processing on the call signal to obtain a noise-reduced call signal.
[0063] Optionally, the feature extraction module 120 includes a first complex convolutional neural network 121 and a second complex convolutional neural network 122. The first complex convolutional neural network 121 is used to extract a first fusion feature of the first noise reduction reference signal and the call signal, and the second complex convolutional neural network 122 is used to extract the echo signal feature of the echo cancellation reference signal.
[0064] The noise reduction module 140 may include a convolutional neural network 141 and a gated recurrent neural network 142 connected to each other.
[0065] For the specific methods of each module regarding the processing and execution of each step, please refer to the description of each step of the method embodiment of steps S11 to S14 of the present application above, which will not be repeated here.
[0066] See also Figure 5 , Figure 5 The call noise reduction device 200 includes an acquisition module 210 , a feature extraction module 220 , a fusion module 230 and a noise reduction module 240 .
[0067] Among them, the acquisition module 210 is used to obtain the echo cancellation reference signal, the first noise reduction reference signal received by the first microphone, the call signal received by the second microphone, and the second noise reduction reference signal received by the third microphone; the feature extraction module 220 is used to extract the first fusion feature of the first noise reduction reference signal and the call signal, the echo signal feature of the echo cancellation reference signal, and the second noise reduction signal feature of the second noise reduction reference signal; the fusion module 230 is used to fuse the first fusion feature, the echo signal feature and the second noise reduction signal feature to obtain the second fusion feature; the noise reduction module 240 is used to use the second fusion feature to perform noise reduction processing on the call signal to obtain a noise-reduced call signal.
[0068] Optionally, the feature extraction module 220 includes a first complex convolutional neural network 221, a second complex convolutional neural network 222, and a third complex convolutional neural network 223. The first complex convolutional neural network 221 is used to extract a first fusion feature of the first noise reduction reference signal and the call signal, the second complex convolutional neural network 222 is used to extract the echo signal feature of the echo cancellation reference signal, and the third complex convolutional neural network 223 is used to extract the second noise reduction signal feature of the second noise reduction reference signal.
[0069] The noise reduction module 240 may include a convolutional neural network 241 and a gated recurrent neural network 242 connected to each other.
[0070] For the specific methods of each module regarding the processing and execution of each step, please refer to the description of each step of the method embodiment of steps S21 to S24 of the above-mentioned present application, which will not be repeated here.
[0071] See also Figure 6 , Figure 6 The structure schematic block diagram of another embodiment of the call noise reduction device of the present application. The call noise reduction device 300 includes a processor 310 and a memory 320 coupled to each other, the memory 320 stores a computer program, and the processor 310 is used to execute the computer program to implement the call noise reduction method described in the above embodiments.
[0072] For the description of each step of the processing execution, please refer to the description of each step of the embodiment of the call noise reduction method of the present application above, which will not be repeated here.
[0073] The memory 320 can be used to store program data and modules. The processor 310 executes various functional applications and data processing by running the program data and modules stored in the memory 320. The memory 320 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as an audio signal feature extraction function, an audio noise reduction function, etc.), etc.; the data storage area may store data created according to the use of the call noise reduction device 300 (such as a first noise reduction reference signal data, call signal data, etc.), etc. In addition, the memory 320 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 320 may also include a memory controller to provide the processor 310 with access to the memory 320.
[0074] In each embodiment of the present application, the disclosed method and device can be implemented in other ways. For example, the various embodiments of the call noise reduction device 300 described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.
[0075] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0076] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0077] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium.
[0078] See also Figure 7 , Figure 7 The headset 400 includes a first microphone 410 , a second microphone 420 and a processing unit 430 .
[0079] The first microphone 410 is used to receive the first noise reduction reference signal, and the second microphone 420 is used to receive the call signal. The first microphone 410 and the second microphone 420 may be connected to the processing unit 430 .
[0080] The processing unit 430 is used to perform noise reduction processing on the call signal using the first noise reduction reference signal according to the steps of each embodiment of the above-mentioned call noise reduction method.
[0081] Optionally, the earphone 400 may further include a third microphone (not shown), which is an in-ear microphone, which is arranged inside the earphone and is mainly used to receive audio signals in the earphone. The received signal has a high signal-to-noise ratio.
[0082] Optionally, the earphones 400 may include but are not limited to wired earphones, wireless wearable devices, such as wireless earphones (headphones, semi-in-ear earphones, in-ear earphones, etc.) and wireless audio glasses, etc. The earphones 400 can establish wired or wireless communication connections with smart devices such as mobile phones and tablets to realize remote calls through the earphones 400 and the smart devices.
[0083] For the specific methods of processing each step, please refer to the description of each step in the above-mentioned embodiment of the call noise reduction method of the present application, which will not be repeated here.
[0084] See also Figure 8 , Figure 8 This is a schematic block diagram of the structure of an embodiment of a computer-readable storage medium of the present application. The computer-readable storage medium 500 stores program data 510. When the program data 510 is executed, the steps of each embodiment of the call noise reduction method as described above are implemented.
[0085] For the description of each step of the processing execution, please refer to the description of each step of the embodiment of the call noise reduction method of the present application above, which will not be repeated here.
[0086] The computer-readable storage medium 500 may be a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc., which may store program codes.
[0087] The above descriptions are merely embodiments of the present application and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.
Claims
1. A method for reducing call noise, It is characterized in that The method comprises: Acquire an echo cancellation reference signal, a first noise reduction reference signal received by the first microphone, and a call signal received by the second microphone; Extracting a first fusion feature of the first noise reduction reference signal and the call signal, and extracting an echo signal feature of the echo cancellation reference signal; Fusing the first fusion feature with the echo signal feature to obtain a second fusion feature; The call signal is subjected to noise reduction processing by using the second fusion feature to obtain a noise-reduced call signal.
2. The method according to claim 1, It is characterized in that The extracting a first fusion feature of the first noise reduction reference signal and the call signal includes: The first noise reduction reference signal and the call signal are processed by a first complex convolutional network to perform complex convolution fusion on the first noise reduction reference signal and the call signal to obtain the first fusion feature, wherein the first fusion feature includes phase information and amplitude information corresponding to the first noise reduction reference signal and the call signal respectively.
3. The method according to claim 1, It is characterized in that The extracting the echo signal feature of the echo cancellation reference signal comprises: The echo cancellation reference signal is processed using a second complex convolutional network to obtain the echo signal feature.
4. The method according to claim 1, It is characterized in that The fusing the first fusion feature and the echo signal feature to obtain a second fusion feature includes: The first fusion feature and the echo signal feature are concatenated and then subjected to a modulo operation to obtain the second fusion feature, where the second fusion feature is a real number feature.
5. The method according to claim 1, It is characterized in that The performing noise reduction processing on the call signal by using the second fusion feature includes: Processing the second fused features using a convolutional neural network to obtain a second fused feature after convolution processing; Processing the second fused features after the convolution processing by using a prediction network to obtain probability results corresponding to each frequency band; The call signal is converted into a frequency domain signal, the frequency domain signals falling into each frequency band are weighted and summed using the probability as a weight, and the summation result is converted into the time domain to obtain the noise-reduced call signal.
6. The method according to claim 1, It is characterized in that The method further comprises: Acquire a second noise reduction reference signal received by the third microphone, and extract a second noise reduction signal feature of the second noise reduction reference signal; The fusing the first fusion feature and the echo signal feature to obtain the second fusion feature includes: performing a modulo operation on the first fusion feature, the echo signal feature and the second noise reduction signal feature after concatenating them to obtain the second fusion feature.
7. The method according to claim 6, It is characterized in that The extracting a second noise reduction signal feature of the second noise reduction reference signal comprises: The second denoised reference signal is processed using a third complex convolutional network to obtain the second denoised signal feature.
8. The method according to claim 1, It is characterized in that The first fusion feature includes phase difference information between the first noise reduction reference signal and the call signal and amplitude information corresponding to the first noise reduction reference signal and the call signal respectively.
9. A call noise reduction device, It is characterized in that The device comprises a processor and a memory coupled to each other; a computer program is stored in the memory, and the processor is used to execute the computer program to implement the steps of the method according to any one of claims 1 to 8.
10. A headset, It is characterized in that The headset includes a first microphone, a second microphone, and a processing unit; The first microphone is used to receive a first noise reduction reference signal, and the second microphone is used to receive a call signal; The processing unit is used to perform noise reduction processing on the call signal using the first noise reduction reference signal according to the steps of the method according to any one of claims 1 to 8.
