Uplink call noise reduction method for a detachable microphone headset system and the headset system

By setting up multiple ambient microphones and corresponding noise reduction filters in the headset system, the problem of decreased noise reduction effect after the external call microphone is removed is solved, achieving stable uplink call noise reduction and call quality, and avoiding interruptions and quality jumps caused by repeated filter convergence.

CN123123778APending Publication Date: 2026-11-17SHENZHEN SEVENSTAR TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611618848.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-20
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

When the external microphone of a detachable microphone headset system is removed, the uplink call noise reduction loses the reference of ambient noise, resulting in a decrease or even interruption of the noise reduction effect, and possible a jump in call voice quality.

Method used

One or more ambient microphones are set up in the headset system, independent of the call microphone, to provide a stable ambient reference signal. First and second ambient noise reduction filters are configured for the on and off states of the external call microphone, respectively. The coefficients of the two filters are continuously updated and optimized when the microphone is on, and the updated filters are directly called for noise reduction processing when the state changes.

Benefits of technology

It achieves stable uplink call noise reduction before and after the external microphone is plugged in and out, avoiding interruption of noise reduction effect and abrupt changes in call voice quality, and ensuring call continuity and clarity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123123778A_ABST
    Figure CN123123778A_ABST
Patent Text Reader

Abstract

This invention discloses an uplink call noise reduction method and a headset system for a detachable microphone headset system, relating to the field of headset technology. It addresses the technical problem that noise reduction performance decreases or even ceases after the external microphone is removed, and that call voice quality may abruptly change. The method executes the following steps: monitoring the connection status of the external microphone; obtaining an environmental reference signal based on an environmental microphone; obtaining a second environmental noise estimate using a second environmental denoising filter; when connected: fusing the output signals of the external and internal microphones to obtain a synthesized call signal, obtaining a first environmental noise estimate using a first environmental denoising filter, and updating the coefficients of the first environmental denoising filter; updating the coefficients of the second environmental denoising filter based on the output signal of the internal microphone and the second environmental noise estimate; when disconnected: updating the coefficients of the second environmental denoising filter based on the output signal of the internal microphone and the second environmental noise estimate.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of headphone technology, and more particularly to an uplink call noise reduction method and headphone system for headphone systems with detachable microphones. Background Technology

[0002] Headphone systems such as over-ear and neckband headphones often include a detachable microphone that connects to the headset body. This allows users to choose whether to install the microphone based on their usage scenario, or to replace the microphone separately if it fails, without having to replace the entire headset. It also improves the signal-to-noise ratio (SNR) of the call (since external microphones are usually located on the extension arm and facing the user's mouth). In scenarios such as gaming, calls, commuting, and office work, headphone systems with detachable microphones generally face problems with environmental noise interference and insufficient call clarity. To improve call quality, uplink noise reduction designs are typically implemented to filter out noise other than speech from the microphone's output signal.

[0003] In existing noise cancellation designs, the earphone itself has a built-in microphone. When an external microphone is plugged in, the external microphone is used for call pickup, and the output signal of the built-in microphone is used as an ambient noise reference to filter out ambient noise in the output signal of the external microphone. However, when the external microphone is unplugged, it usually switches to single-microphone noise cancellation. That is, the built-in microphone needs to switch from the role of ambient noise reference to the role of call pickup and can no longer provide an independent ambient noise reference. The uplink call noise cancellation effect decreases or even stops, and the switching of the built-in microphone's working mode may cause a jump in call voice quality.

[0004] In the process of realizing this invention, the inventors discovered at least the following problems in the prior art: When the external microphone of a detachable microphone headset system is removed, the uplink call noise reduction loses the reference of ambient noise, resulting in a decrease or even interruption of the noise reduction effect, and possible a jump in call voice quality. Summary of the Invention

[0005] The purpose of this invention is to provide an uplink call noise reduction method and a headset system for a detachable microphone headset system, to solve the technical problems existing in the prior art where, after the external microphone is removed, the uplink call noise reduction loses its environmental noise reference, resulting in a decrease or even interruption of the noise reduction effect, and possible abrupt changes in call voice quality. The various technical effects of the preferred solutions among the many technical solutions provided by this invention are detailed below.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides an uplink call noise reduction method for a headset system with a detachable microphone. The headset system includes a headset body, an extension arm detachably connected to the headset body, an external call microphone mounted on the extension arm, a built-in call microphone disposed on the headset body, and one or more ambient microphones for picking up ambient noise. The method periodically executes: monitoring the access status of the external call microphone; obtaining an environmental reference signal based on the output signals of one or more ambient microphones; processing the environmental reference signal using a second environmental noise reduction filter to obtain a second environmental noise estimate; when the access status is "on": fusing the output signals of the external call microphone and the built-in call microphone to obtain a synthesized call signal; and using a first environmental noise reduction filter. The filter processes the environmental reference signal to obtain a first environmental noise estimate; based on the synthesized call signal and the first environmental noise estimate, an environmentally denoised uplink call signal is obtained, and the coefficients of the first environmental denoising filter are updated using the environmentally denoised uplink call signal as the first residual environmental signal; based on the output signal of the built-in call microphone and the second environmental noise estimate, a second residual environmental signal is obtained, and the coefficients of the second environmental denoising filter are updated using the second residual environmental signal; when the access state is disconnected: based on the output signal of the built-in call microphone and the second environmental noise estimate, an environmentally denoised uplink call signal is obtained, and the coefficients of the second environmental denoising filter are updated using the environmentally denoised uplink call signal as the second residual environmental signal.

[0007] Preferably, when the access state switch is detected, within a preset future time period, the uplink call signal after environmental denoising at the previous discrete time point and the uplink call signal after environmental denoising at the current discrete time point are synthesized using a cross-fading method to obtain the uplink call signal after environmental denoising and fading processing, and the uplink call signal is output.

[0008] Preferably, when the access state is detected to switch from connected to disconnected, the coefficients of the first environmental denoising filter are frozen. When the access state is detected to switch from disconnected back to connected, the environmental reference signal is processed using the first environmental denoising filter with frozen coefficients to obtain the first environmental noise estimate.

[0009] Preferably, after the coefficients of the first environmental denoising filter and the second environmental denoising filter converge, the coefficients of the first environmental denoising filter and the second environmental denoising filter are frozen; after the step of obtaining an environmental reference signal based on the output signal of one or more environmental microphones, the following steps are performed instead: when the access state is detected to switch from connected to disconnected: the environmental reference signal is processed using the second environmental denoising filter with frozen coefficients to obtain a second environmental noise estimate; based on the output signal of the built-in call microphone and the second environmental noise estimate, an environmentally denoised uplink call signal is obtained; when the access state is detected to switch from disconnected to connected: the output signal of the external call microphone and the output signal of the built-in call microphone are fused to obtain a synthesized call signal; the environmental reference signal is processed using the first environmental denoising filter with frozen coefficients to obtain a first environmental noise estimate; based on the synthesized call signal and the first environmental noise estimate, an environmentally denoised uplink call signal is obtained.

[0010] Preferably, obtaining the environmental reference signal based on the output signals of one or more environmental microphones includes: synthesizing the output signals of one or more environmental microphones to obtain an original environmental reference signal; when the access state is on, performing pre-filtering processing on the original environmental reference signal using a first-level path estimation model to obtain the environmental reference signal; when the access state is off, performing pre-filtering processing on the original environmental reference signal using a second-level path estimation model to obtain the environmental reference signal.

[0011] Preferably, a left ambient microphone and a right ambient microphone are respectively provided on the left and right sides of the earphone body; obtaining an ambient reference signal based on the output signals of one or more ambient microphones includes: estimating the spatial direction of the sound source based on the arrival time difference between any two output signals from the left ambient microphone, the right ambient microphone, and the built-in call microphone; if the estimated spatial direction of the sound source matches the direction of the built-in call microphone toward the wearer's mouth, then the signal components corresponding to the spatial direction of the sound source in the output signals of the left and right ambient microphones are attenuated or removed, and the signal components corresponding to the direction components that do not match the direction of the built-in call microphone toward the wearer's mouth are retained; the ambient reference signal is obtained based on the attenuated or removed output signals of the left and right ambient microphones.

[0012] Preferably, the earphone body is provided with a speaker driving circuit and one or more speakers. The method further includes: using the speaker driving signal output by the speaker driving circuit as an echo reference signal; processing the echo reference signal using a second echo filter to obtain a second echo estimate; when the access state is on: processing the echo reference signal using a first echo filter to obtain a first echo estimate; obtaining an echo-cancelled uplink call signal based on the first residual environmental signal and the first echo estimate, and updating the coefficients of the first echo filter using the echo-cancelled uplink call signal as the first residual echo signal; obtaining a second residual echo signal based on the second residual environmental signal and the second echo estimate; updating the coefficients of the second echo filter based on the second residual echo signal; when the access state is off: obtaining an echo-cancelled uplink call signal based on the second residual environmental signal and the second echo estimate, and updating the coefficients of the second echo filter using the echo-cancelled uplink call signal as the second residual echo signal.

[0013] Preferably, the method further includes: performing voice activity detection on the output signal of the built-in microphone to obtain a wearer voice activity detection result; performing dual-talk detection based on the wearer voice activity detection result and the speaker drive signal to obtain a dual-talk detection result; if the wearer voice activity detection result indicates the presence of wearer voice, and / or the dual-talk detection result indicates the simultaneous presence of wearer voice and speaker-played voice, then freezing or slowing down the coefficient updates of the first echo filter and / or the second echo filter.

[0014] Preferably, the method further includes: performing call noise reduction processing on the echo-cancelled uplink call signal to obtain a call noise-reduced uplink call signal.

[0015] The present invention also provides an earphone system comprising: a headphone body with a left earcup and a right earcup, wherein the left earcup or the right earcup is provided with a detachable interface and a built-in microphone, and the earphone body is provided with one or more ambient microphones for picking up ambient noise; a speaker is provided in the left earcup and / or the right earcup; an extension arm detachably connected to the detachable interface, wherein an external microphone is provided on the extension arm; a plug-in detection unit located on the earphone body for detecting the connection status of the external microphone; an audio processor and a speaker driver circuit located within the earphone body, wherein the audio processor is electrically connected to the plug-in detection unit and the speaker driver circuit respectively, and the speaker driver circuit is electrically connected to the speaker; the audio processor executes the steps of the uplink call noise reduction method for the earphone system with a detachable microphone provided by the present invention.

[0016] Implementing one of the above-described technical solutions of the present invention has the following advantages or beneficial effects: This invention independently sets up one or more environmental microphones separate from the call microphone (external call microphone and internal call microphone) to ensure that the environmental reference signal is stably provided whether the external call microphone is on or off, while the internal call microphone is fixedly used to pick up the call signal. A first environmental noise reduction filter is configured for the on state of the external call microphone, and a second environmental noise reduction filter is configured for the off state of the external call microphone. The coefficients of the two environmental noise reduction filters are continuously updated and optimized when the microphone is on, and the coefficients of the second environmental noise reduction filter are continuously updated and optimized when the microphone is off. Therefore, when the external call microphone switches from on to off, the second environmental noise reduction filter with continuously updated parameters is directly called to perform environmental noise reduction on the output signal of the internal call microphone; when switching back from off to on, the first environmental noise reduction filter with continuously updated parameters is directly called to perform environmental noise reduction on the synthesized call signal. As can be seen, before and after the external microphone is plugged in or unplugged (before and after the access state is switched), the present invention does not require re-optimizing the coefficients of the first or second environmental noise reduction filter. The two environmental noise reduction filters with continuously updated parameters can combine with a stable environmental noise reference signal to provide stable environmental noise reduction capability, achieve uninterrupted and non-degraded uplink call noise reduction, and avoid the problem of call voice quality jump caused by the switching of the built-in microphone role or the reconvergence of the coefficients of the environmental noise reduction filter. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart illustrating the uplink call noise reduction method for a headphone system with a detachable microphone provided in Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of the headphone system in an example of Embodiment 1 of the present invention; Figure 3 This is a system block diagram of the headphone system in Embodiment 4 of the present invention; Figure 4 This is a front view of the headphone system in Embodiment 4 of the present invention; Figure 5 This is a rear view of the headphone system in Embodiment 4 of the present invention; Figure 6 This is a schematic diagram of a plug-in / plug-out detection unit in Embodiment 4 of the present invention.

[0018] In the image: 1. Headphone body; 101. Right earcup; 102. Left earcup; 2. Extension arm; 3. External call microphone; 4. Internal call microphone; 5. Right ambient microphone; 6. Left ambient microphone. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the present invention clearer, various exemplary embodiments described below will be referenced to the accompanying drawings, which form part of the exemplary embodiments, illustrating various exemplary embodiments that may be used to implement the present invention. Unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the disclosure of the present invention. It should be understood that they are merely examples of processes, methods, and apparatuses consistent with some aspects of the present invention disclosed as detailed in the appended claims, and other embodiments may be used, or structural and functional modifications may be made to the embodiments listed herein without departing from the scope and spirit of the present invention.

[0020] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," etc., indicate the orientation or positional relationship based on the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the referred element must have a specific orientation, or be constructed and operated in a specific orientation. The terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. The term "multiple" means two or more. The terms "connected" and "linked" should be interpreted broadly, for example, they can be fixed connections, pluggable connections, integral connections, mechanical connections, electrical connections, communication connections, direct connections, indirect connections through an intermediate medium, and can be the internal connection of two elements or the interaction relationship between two elements. The term "and / or" includes any and all combinations of one or more of the related listed items. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0021] To illustrate the technical solution described in this invention, specific embodiments are described below, showing only the parts related to the embodiments of this invention.

[0022] Example 1: This embodiment provides an uplink call noise reduction method for a headset system with a detachable microphone. The headset system is as follows: Figure 2 As shown, the device includes an earphone body 1, an extension arm 2 detachably connected to the earphone body 1, an external microphone 3 mounted on the extension arm 2, a built-in microphone 4 disposed on the earphone body 1, and one or more ambient microphones for picking up ambient noise. For example, Figure 2 In the headset body 1, there is one or more ambient microphones, including a left ambient microphone 6 and a right ambient microphone 5 on the left and right sides respectively.

[0023] In this embodiment, the earphone body 1 can be any of the following: over-ear, neckband, ear-hook, or helmet-style. Regardless of the form, the ambient microphone is located on the outward-facing side of the earphone body 1 to pick up external stereo ambient noise. The built-in call microphone 4 is fixed to the earphone body 1 and faces the wearer's mouth. The external call microphone 3 is positioned near the mouth via the extension arm 2 and is connected to the standardized pluggable interface of the earphone body 1 via a standardized connector. The output signals of the ambient microphone and the built-in call microphone 4 are transmitted to the execution body of the aforementioned uplink call noise reduction method via the inner conductor of the earphone body 1. The extension arm 2 has a fixed geometry and does not hinge or rotate after deployment; its configuration is changed only by plugging and unplugging. The output signal of the external call microphone 3 is transmitted to the execution body of the aforementioned uplink call noise reduction method via the inner conductor of the extension arm 2 and the connector.

[0024] In this embodiment, uplink call noise reduction refers to reducing the noise of the uplink voice signal (i.e., the signal output by the external call microphone 3 and the built-in call microphone 4) generated by the wearer's speech, so as to suppress environmental noise, echoes and other non-target sounds and improve the voice clarity sent to the host computer communication link.

[0025] This embodiment provides an uplink call noise reduction method for a headset system with a detachable microphone. The execution entity of this method is preferably an audio processor within the headset system. The audio processor includes one or more processors, which may be at least one of a microcontroller (MCU), digital signal processor (DSP), system-on-a-chip (SoC), or field-programmable gate array (FPGA). Figure 1 As shown, the method periodically executes steps S1 to S8, where periodic execution can be performed at preset intervals of a number of work cycles, such as 1. Alternatively, steps S1 to S8 can be executed once for each frame of call signal (output signal from external microphone 3 and internal microphone 4). After each cycle of processing steps is completed (executes to step S7 in the connected state and to step S8 in the disconnected state), it is determined whether the process has ended. If it has not ended, steps S1 to S8 continue to be executed in the next execution cycle. The process is considered complete when a command to power off the headset system, disable the wearer's voice call function, or any other stop command is received.

[0026] Step S1: Monitor the access status of the external microphone 3.

[0027] In this embodiment, the method execution subject can obtain the access status of the external microphone 3 by whether a communication link can be successfully established with it. When the communication link is successfully established, the access status of the external microphone 3 is considered to be connected; when the communication link fails to be established, the access status of the external microphone 3 is considered to be disconnected. The communication link refers to the interface link between the method execution subject and the external microphone 3 used for transmitting microphone audio signals and / or control signals.

[0028] Step S2: Obtain an environmental reference signal based on the output signal of one or more environmental microphones.

[0029] In this embodiment, the output signals of one or more environmental microphones are preprocessed separately. When there is only one environmental microphone, its preprocessed output signal is used as the environmental reference signal. When there are two or more environmental microphones, the preprocessed output signals of the two or more environmental microphones are weighted and summed to obtain the environmental reference signal. The preprocessing includes at least one of DC removal, high-pass filtering, and gain calibration. Gain calibration is used to ensure that the sensitivity of all environmental microphones is consistent, avoiding bias of the environmental reference signal towards one path due to sensitivity differences.

[0030] Specifically, in the above weighted summation process, the weight of each ambient microphone is not limited to being set empirically. Preferably, the weight is determined based on the signal-to-noise ratio (SNR) of the preprocessed output signal of the ambient microphone. For example, a left ambient microphone 6 and a right ambient microphone 5 are set, and the SNR of the preprocessed output signal of the left ambient microphone 6 is estimated. And the signal-to-noise ratio of the preprocessed output signal of the right environmental microphone 5. The weights of the left ambient microphone 6 and the right ambient microphone 5 are respectively , , , Follow It increases with the increase of, and with As it increases, it decreases. , Normalization , The obtained environmental reference signal is represented as follows: ; The index of the signal acquisition point in the output signals of the environmental reference signal, the left environmental microphone 6, and the right environmental microphone 5 is also called the discrete time point index and is a positive integer. , These represent the preprocessed output signals of the left ambient microphone 6 and the right ambient microphone 5, respectively.

[0031] Step S3: Process the environmental reference signal using the second environmental denoising filter to obtain the second environmental noise estimate.

[0032] The second environmental denoising filter is not limited to a finite impulse response (FIR) filter or an infinite impulse response (IIR) filter. Specifically, the second environmental denoising filter is of order [order missing]. For a finite impulse response filter, the vector of its filter coefficients (i.e., the filter coefficient vector) is denoted as: ; in, Indicates the discrete-time index; order The sampling rate and the acoustic path length of the earphone body cavity 1 can be determined, for example, from 64 to 256. , , These represent the second environmental denoising filter at discrete time points. The 0th order filter coefficients, the 1st order filter coefficients, and the 2nd order filter coefficients 1st order filter coefficients.

[0033] In step S3, the second environmental denoising filter is based on the environmental reference signal. and its current filter coefficient vector Calculate the second environmental noise estimate : ; This indicates the transpose of the orientation variable; , Indicates environmental reference signal The delay vector, , These represent environmental reference signals. The previous discrete time point, the previous Environmental reference signal at discrete time points.

[0034] When the external microphone 3 is in the connected state, perform the following steps S4 to S7: Step S4: Combine the output signal of the external microphone 3 and the output signal of the built-in microphone 4 to obtain a synthesized call signal.

[0035] The existing delay-superposition beamforming method can be used to fuse the output signals of the external microphone 3 and the internal microphone 4 to obtain a synthesized call signal. In this way, the speech in the direction of the mouth is enhanced due to in-phase superposition, and the lateral environmental noise is partially canceled out during superposition due to the different directions and different time delays of arrival at the two microphones, thereby improving the signal-to-noise ratio of the synthesized call signal.

[0036] Step S5: Process the environmental reference signal using the first environmental denoising filter to obtain the first environmental noise estimate.

[0037] The first environmental denoising filter is not limited to a finite impulse response (FIR) filter or an infinite impulse response (IIR) filter. Specifically, the first environmental denoising filter is a finite impulse response filter, and its order can be the same as or different from that of the second environmental denoising filter. Preferably, both filters use the same filter order. This facilitates dual-state hot backup and filter coefficient switching between connected and disconnected states, as well as hardware resource reuse and unified software management.

[0038] The vector composed of the filter coefficients of the first environmental denoising filter (i.e., the filter coefficient vector) is denoted as: ; in, , , These represent the first environmental denoising filter at discrete time points. The 0th order filter coefficients, the 1st order filter coefficients, and the 2nd order filter coefficients 1st order filter coefficients.

[0039] In step S5, the first environmental denoising filter is based on the environmental reference signal. and its current filter coefficient vector That is, at the current discrete time point The filter coefficient vector is used to calculate the first environmental noise estimate. : .

[0040] Step S6, based on the synthesized call signal and first environmental noise estimation The uplink call signal after environmental denoising is obtained, and the uplink call signal after environmental denoising is used as the first residual environmental signal. Update the coefficients of the first environmental denoising filter.

[0041] Specifically, the synthesized call signal Subtract the first environmental noise estimate The first residual environmental signal was obtained. : . It is used both to update the parameters of the first environmental denoising filter and as the output of the uplink call signal after environmental denoising when the access status is connected.

[0042] Utilizing the first residual environmental signal and environmental reference signals delay vector The filter coefficient vector of the first environmental denoising filter is updated according to the following weight update formula: ; in, The step size factor of the first environmental denoising filter is used to control the convergence speed and steady-state error. Its value ranges from 0.001 to 0.01, and the specific value can be determined through experimental debugging. This weight update formula makes the first residual environmental signal... The mean square value is minimized, even though the first ambient noise estimate continuously approaches the real ambient noise component in the output signal of the external microphone 3. This indicates that the first environmental denoising filter at the next discrete time point (i.e., discrete time point) The filter coefficient vector.

[0043] Step S7, based on the output signal of the built-in microphone 4 Second environmental noise estimation The second residual environmental signal is obtained, and the coefficients of the second environmental denoising filter are updated using the second residual environmental signal.

[0044] Specifically, the output signal of the built-in microphone 4 Subtract the second environmental noise estimate The second residual environmental signal was obtained. : . It is used to update the parameters of the second environmental denoising filter, and in the subsequent step S8, it is also used as the uplink call signal output after environmental denoising when the access state is disconnected.

[0045] Utilizing the second residual environmental signal and environmental reference signals delay vector The filter coefficient vector of the second environmental denoising filter is updated according to the following weight update formula: ; in, The step size factor of the second environmental noise reduction filter is used to control the convergence speed and steady-state error. Its value ranges from 0.001 to 0.01, and the specific value can be determined through experimental debugging. This weight update formula minimizes the mean square value of the second residual environmental signal, so that the second environmental noise estimate continuously approaches the real environmental noise component in the output signal of the built-in microphone 4. This indicates that the second environmental denoising filter is at the next discrete time point (i.e., discrete time point). The filter coefficient vector.

[0046] When the access status is disconnected, execute the following step S8: Step S8: Based on the output signal of the built-in microphone 4 and the second environmental noise estimation, obtain the uplink call signal after environmental denoising. Use the uplink call signal after environmental denoising as the second residual environmental signal to update the coefficients of the second environmental denoising filter. The specific processing procedure is the same as step S7 above, and will not be repeated here.

[0047] As an optional implementation, when an access state switch is detected, including switching from connected to disconnected and switching from disconnected to connected, the uplink call signal after environmental denoising at the previous discrete time point and the uplink call signal after environmental denoising at the current discrete time point are synthesized using a cross-fading method within a preset future time period to obtain the uplink call signal after environmental denoising and fading processing, and the uplink call signal is output.

[0048] Specifically, let the discrete time point before the access state switch be . The discrete time points after the switch are The preset future time period includes T moments, and the first moment in the preset future time period is... The uplink call signal after environmental denoising and fade-out processing at each discrete time point is represented as follows: ; in, , These represent the first and second weights, respectively, both of which change over time. ,and , along with It increases and then decreases until it reaches 0. along with It increases and then decreases until it reaches 1. . Indicates according to discrete time points before the switch The uplink call signal after environmental noise reduction is obtained from the above steps S6 or S8, corresponding to the access status of the external call microphone 3. Indicates that after the switch, according to discrete time points The uplink call signal after environmental noise reduction is obtained from the above steps S6 or S8, corresponding to the access status of the external call microphone 3.

[0049] This embodiment enables the output uplink call signal to be noise-reduced and faded. The uplink call signal smoothly transitions from the noise-reduced environment before the switch to the noise-reduced environment after the switch, avoiding sudden changes in the voice waveform, clicking sounds, or call interruptions before and after the external microphone is plugged in or out, thus ensuring the continuity and clarity of the uplink call.

[0050] As an optional implementation, when the access state is detected to switch from connected to disconnected, the coefficients of the first environmental denoising filter are frozen. When the access state is detected to switch back from disconnected to connected, the environmental reference signal is processed using the first environmental denoising filter with frozen coefficients to obtain the first environmental noise estimate.

[0051] In this embodiment, when the access state changes from connected to disconnected, the coefficients (i.e., the filter coefficient vector) of the first environmental noise reduction filter are frozen. This allows for direct recall when the access state changes from disconnected back to connected, restoring the environmental noise reduction performance at the time of connection without the need for re-convergence. The first environmental noise reduction filter and the second environmental noise reduction filter form a dual-state hot backup, achieving zero-repetition convergence and no jumps during plugging and unplugging, ensuring that uplink call noise reduction is uninterrupted and does not degrade before and after plugging and unplugging the external call microphone 3.

[0052] As an optional implementation, after the coefficients of the first environmental denoising filter and the second environmental denoising filter converge, the coefficients of the first environmental denoising filter and the second environmental denoising filter are frozen; at this time, after step S2, that is, after the step of obtaining the environmental reference signal based on the output signal of one or more environmental microphones, steps S3 to S8 are replaced by performing: When the access status is detected to switch from connected to disconnected: the environmental reference signal is processed by the second environmental noise reduction filter with coefficient freezing to obtain the second environmental noise estimate; based on the output signal of the built-in call microphone 4 and the second environmental noise estimate, specifically by subtracting the two, the uplink call signal after environmental noise reduction is obtained. When the access status is detected to switch from disconnected to connected: the output signal of the external call microphone 3 and the output signal of the built-in call microphone 4 are fused to obtain a synthesized call signal; the environmental reference signal is processed by the first environmental denoising filter with coefficient freezing to obtain a first environmental noise estimate; based on the synthesized call signal and the first environmental noise estimate, specifically by subtracting the two, the uplink call signal after environmental denoising is obtained.

[0053] In this embodiment, after the coefficients of the first environmental noise reduction filter and the second environmental noise reduction filter have converged and been frozen, when the external microphone 3 is plugged in and out, there is no need to re-converge the coefficients of the environmental noise reduction filter. This achieves seamless environmental noise reduction with zero repetition and no jumps, ensuring that the uplink call noise reduction is not interrupted or degraded before and after the external microphone 3 is plugged in and out.

[0054] As an optional implementation, in step S2 above, obtaining an environmental reference signal based on the output signals of one or more environmental microphones includes: Step S21a: Synthesize the output signals of one or more environmental microphones to obtain the original environmental reference signal, which is the synthesized signal of the output signals of one or more environmental microphones.

[0055] Step S22a: When the access status is connected, the original environmental reference signal is pre-filtered using the first-level path estimation model to obtain the environmental reference signal. When the access status is disconnected, the original environmental reference signal is pre-filtered using the second-level path estimation model to obtain the environmental reference signal.

[0056] In this embodiment, the first-level path refers to the electroacoustic path from the speaker of the earphone system, the earphone body 1 cavity, the air to the external microphone 3 and the internal microphone 4, and the second-level path refers to the electroacoustic path from the speaker of the earphone system, the earphone body 1 cavity, the air to the internal microphone 4. Phase and amplitude distortions occur in the electroacoustic paths, causing the first and second environmental denoising filters to easily diverge or even howl under non-stationary noise. Therefore, a first-level path estimation model and a second-level path estimation model are set to pre-filter the original environmental reference signal. The environmental reference signal is used to drive the first and second environmental denoising filters to update until stable convergence, compensating for electroacoustic path distortion.

[0057] In this embodiment, both the first-level path estimation model and the second-level path estimation model can be implemented using existing secondary path estimation models, such as finite impulse response (FIR) filter models or infinite impulse response (IIR) filter models. Regarding the acquisition of the coefficients for the first-level and second-level path estimation models, they can be predetermined and fixed offline, or updated online using adaptive algorithms such as least mean square (LMS) or normalized least mean square (NLMS).

[0058] In this embodiment, the external microphone 3 sets corresponding secondary path estimation models in both connected and disconnected access states, so that the pre-filtered environmental reference signal can accurately match the actual electroacoustic path, compensate for the phase and amplitude distortion caused by the electroacoustic path, and enable the first environmental noise reduction filter and the second environmental noise reduction filter to still converge stably, without diverging or howling under non-stationary noise.

[0059] As an optional implementation, a left ambient microphone 6 and a right ambient microphone 5 are respectively provided on the left and right sides of the earphone body 1; in the above step S2, obtaining an ambient reference signal based on the output signal of one or more ambient microphones includes: Step S21b: Estimate the spatial direction of the sound source based on the arrival time difference between any two of the output signals from the left ambient microphone 6, the right ambient microphone 5, and the built-in intercom microphone 4.

[0060] The arrival time difference between any two output signals can be solved using the existing Generalized Cross-Correlation with Phase Transform (GCC-PHAT) algorithm. The specific process is as follows: For any two microphone signals, perform frame-by-frame windowing and Fast Fourier Transform (FFT) respectively, calculate their cross-power spectrum and apply phase transform weighting, then perform an inverse FFT to obtain the generalized cross-correlation function, and search for the peak position of this function to obtain the arrival time difference between the two output signals.

[0061] Three pairs of microphones were selected: left ambient microphone 6 and right ambient microphone 5, left ambient microphone 6 and built-in call microphone 4, and right ambient microphone 5 and built-in call microphone 4. The arrival time differences of each pair of microphones were obtained. Based on the arrival time differences of the three pairs of microphones, and the known positional relationships and spacing between the left ambient microphone 6, right ambient microphone 5, and built-in call microphone 4, the projection relationship of the sound source direction onto the line connecting the three pairs of microphones was determined in the coordinate system of the earphone body 1. The projection relationships of the three pairs of microphones were combined to solve for the direction vector of the sound source in the coordinate system of the earphone body 1. After normalizing the direction vector, it was converted into azimuth and pitch angles relative to the coordinate system of the earphone body 1, which were used as the spatial direction of the sound source.

[0062] In step S22b, if the estimated spatial direction of the sound source matches the direction of the built-in microphone 4 toward the wearer's mouth, the signal components corresponding to the spatial direction of the sound source in the output signals of the left ambient microphone 6 and the right ambient microphone 5 are attenuated or removed, while retaining the signal components corresponding to the direction components that do not match the direction of the built-in microphone 4 toward the wearer's mouth. Based on the attenuated or removed output signals of the left ambient microphone 6 and the right ambient microphone 5, an environmental reference signal is obtained. The method of obtaining the signal is not limited to the aforementioned weighted summation or the processing procedures of steps S21a and S22a, which will not be described in detail here.

[0063] If the spatial direction of the sound source does not match the direction of the built-in microphone 4 toward the wearer's mouth, it indicates that the sound source mainly comes from the external environment rather than the wearer's own voice. In this case, the output signal of the left environmental microphone 6 and the output signal of the right environmental microphone 5 are directly synthesized to obtain the environmental reference signal. The synthesis method is not limited to the aforementioned weighted summation or the processing procedures of the aforementioned steps S21a and S22a, which will not be elaborated here.

[0064] In this embodiment, the built-in microphone 4 is fixedly mounted on the earphone body 1 and faces the wearer's mouth; its orientation is known and fixed. Directional matching refers to the absolute value of the angle between the spatial direction of the sound source and the direction in which the built-in microphone 4 faces the wearer's mouth falling within a preset angle range. The preset angle range is... , The upper limit of the included angle, The value ranges from 15° to 30°. Directional mismatch refers to the absolute value of the angle between the spatial direction of the sound source and the direction in which the built-in microphone 4 is directed toward the wearer's mouth, exceeding the upper limit of the angle. .

[0065] In this embodiment, when the spatial direction of the sound source matches the direction of the built-in microphone 4 towards the wearer's mouth, it indicates that the wearer's own voice is mixed into the output signals of the left ambient microphone 6 and the right ambient microphone 5. The signal component corresponding to the spatial direction of the sound source in the output signals of the left ambient microphone 6 and the right ambient microphone 5 is obtained. This can be achieved using existing spatial filtering or time-frequency masking methods. For example, using spatial filtering involves constructing a beamformer pointing in the estimated spatial direction of the sound source, and then using the beamformer to filter the output signals of the left ambient microphone 6 and the right ambient microphone 5 respectively to obtain the signal component corresponding to the spatial direction of the sound source.

[0066] In this embodiment, the attenuation or removal method includes multiplying the signal component corresponding to the spatial direction of the acquired sound source by an attenuation coefficient less than 1, such as 0 to 0.5. As a result, the speech component from the wearer's mouth direction in the output signals of the left environmental microphone 6 and the right environmental microphone 5 is weakened or removed, while the environmental noise component (not the wearer's speech) is retained. This prevents the wearer's own speech from leaking into the environmental reference signal and prevents subsequent environmental noise reduction from misinterpreting their own speech as environmental noise and canceling it out.

[0067] Since the built-in microphone 4 faces the wearer's mouth, by comparing the estimated spatial direction of the sound source with the direction of the built-in microphone 4 facing the mouth, the sound source can be determined to be the wearer's own voice when the direction is matched. The signal component corresponding to that direction in the output signal of the left ambient microphone 6 and the right ambient microphone 5 is attenuated or removed, thereby preventing the wearer's own voice from leaking into the environmental reference signal and preventing subsequent environmental noise reduction from misjudging the wearer's own voice as environmental noise and canceling it out.

[0068] Example 2: This embodiment provides an uplink call noise reduction method for a headset system with a detachable microphone. The difference from Embodiment 1 is that, after obtaining the uplink call signal with environmental noise reduction in Embodiment 1, echo cancellation processing is further performed on the uplink call signal with environmental noise reduction obtained in Embodiment 1, i.e., two-stage noise reduction is achieved. Specifically, the headset body 1 is equipped with a speaker driver circuit and one or more speakers. The uplink call noise reduction method in this embodiment further includes: Step S9: Use the loudspeaker driving signal output by the loudspeaker driving circuit as the echo reference signal.

[0069] Specifically, the speaker driver circuit drives the speaker to play audio via a speaker drive signal. The sound played by the speaker is picked up by the two microphones through the cavity of the earphone body 1 and the air, forming an echo. Discrete time points The speaker drive signal is denoted as It can accurately represent the sound that the speaker is about to emit.

[0070] Step S10: Process the echo reference signal using the second echo filter to obtain the second echo estimate.

[0071] Specifically, the second echo filter is used to estimate the echo path when the call pickup originates from the built-in microphone 4. The second echo filter is preferably a finite impulse response filter, and its coefficients are updated online using the Normalized Least Mean Square (NLMS) algorithm. The filter coefficient vector of the second echo filter is denoted as... : ; This represents the order of the second echo filter, and is a positive integer. , , These represent the second echo filter at discrete time points. The 0th order filter coefficients, the 1st order filter coefficients, and the 2nd order filter coefficients 1st order filter coefficients.

[0072] echo reference signal The delay vector is denoted as : , , These represent the echo reference signals. The previous discrete time point, the previous Echo reference signal at discrete time points.

[0073] The second echo filter calculates the second echo estimate based on the echo reference signal and the current filter coefficient vector. : Step S10 is executed both when the access state is on and off, so that the second echo filter continues to run and update, forming a hot backup when disconnected.

[0074] When the access status is "connected", execute the following steps S11 to S13: Step S11: Process the echo reference signal using the first echo filter to obtain the first echo estimate.

[0075] Specifically, the first echo filter is used to estimate the echo paths when the call pickup originates from the external microphone 3 and the internal microphone 4. The first echo filter is preferably a finite impulse response filter, and its coefficients are updated online using the Normalized Least Mean Square (NLMS) algorithm. The first and second echo filters can have the same or different orders. Preferably, they have the same order to simplify processing. The filter coefficient vector of the first echo filter is denoted as... : ; , , These represent the first echo filter at discrete time points. The 0th order filter coefficients, the 1st order filter coefficients, and the 2nd order filter coefficients 1st order filter coefficients.

[0076] The first echo filter calculates the first echo estimate based on the echo reference signal and the current filter coefficient vector. : .

[0077] Step S12: Based on the first residual environmental signal and the first echo estimation, obtain the echo-cancelled uplink call signal, update the coefficients of the first echo filter with the echo-cancelled uplink call signal as the first residual echo signal, and output the echo-cancelled uplink call signal, that is, output the first residual echo signal.

[0078] Specifically, the first residual environmental signal Subtract the first echo estimate The uplink communication signal after echo cancellation is obtained and output, i.e., the first residual echo signal. , .

[0079] Using the first residual echo signal and echo reference signal delay vector The coefficients (i.e., the filter coefficient vector) of the first echo filter are updated according to the following weight update formula: ; in, This represents the step size factor of the first echo filter; This represents the small regularization constant of the first echo filter to avoid a denominator of 0; Represents the delay vector The square of the L2 norm. This indicates that the first echo filter is at the next discrete time point (i.e., discrete time point). The filter coefficient vector.

[0080] Step S13: Based on the second residual environmental signal and the second echo estimation, obtain the second residual echo signal; update the coefficients of the second echo filter based on the second residual echo signal.

[0081] Specifically, the second residual environmental signal Subtract the second echo estimate The uplink communication signal after echo cancellation is obtained, i.e., the second residual echo signal. , Using the second residual echo signal and echo reference signal delay vector The coefficients (i.e., the filter coefficient vector) of the second echo filter are updated according to the following weight update formula: ; in, This represents the step size factor of the second echo filter; This represents a small regularization constant for the second echo filter to avoid a denominator of 0; This indicates that the second echo filter is at the next discrete time point (i.e., discrete time point). The filter coefficient vector.

[0082] When the access status is disconnected, proceed to step S14: Step S14: Based on the second residual environmental signal and the second echo estimation, obtain the echo-cancelled uplink call signal. Update the coefficients (i.e., filter coefficient vector) of the second echo filter using the echo-cancelled uplink call signal as the second residual echo signal, and output the echo-cancelled uplink call signal at this time. For the specific processing procedure, please refer to step S13 above, which will not be repeated here.

[0083] The insertion and removal of the external microphone 3 alters the call pickup source and the echo path from the speaker to the microphone. When connected, the call signal is a combination of the external microphone 3 and the internal microphone 4, and the echo path includes the paths of both the external microphone 3 and the internal microphone 4. When disconnected, the call signal comes only from the internal microphone 4, and the echo path includes only the path of the internal microphone 4. If only a single echo filter is used, the filter coefficients will mismatch with the actual echo path after the external microphone 3 is inserted or removed, requiring re-convergence. During re-convergence, echo and howling residue remain, leading to a decrease in call quality. In this embodiment, a first echo filter and a second echo filter are set for connected and disconnected states, respectively. In the connected state, the second echo filter is continuously updated and kept converging in the disconnected state, forming a dual-state hot backup. When the external microphone 3 is removed, the second echo filter is directly invoked; when connected again, the first echo filter is directly invoked. This achieves zero-repetition convergence and no jumps during the insertion and removal of the external microphone 3, ensuring uninterrupted and non-degraded echo cancellation.

[0084] As an optional implementation, when an access state switch is detected, the state switch includes switching from connected to disconnected and switching from disconnected to connected. Within a preset future time period, the uplink call signal after echo cancellation at the previous discrete time point and the uplink call signal after echo cancellation at the current discrete time point are synthesized using a cross-fading method to obtain a faded uplink call signal after echo cancellation, and the uplink call signal is output.

[0085] The crossfading method in this embodiment can refer to the crossfading process in Example 1, and will not be repeated here. Through crossfading, the call quality is avoided from being affected by incomplete echo cancellation before and after the external microphone 3 is plugged in and out.

[0086] As an optional implementation, when the access state is detected to switch from connected to disconnected, the filter coefficient vector of the first echo filter is frozen. When the access state is detected to switch back from disconnected to connected, the echo reference signal is processed using the first echo filter with frozen coefficients to obtain the first echo estimate.

[0087] In this embodiment, when the access state changes from connected to disconnected, the filter coefficient vector of the first echo filter is frozen. This allows for direct recall when the access state changes from disconnected back to connected, restoring the echo cancellation function at the time of connection without the need for re-convergence. Zero repetition convergence and no jumps are observed during plugging and unplugging, ensuring that the uplink call echo cancellation is uninterrupted and does not degrade before and after the external microphone 3 is plugged and unplugged.

[0088] As an optional implementation, after the coefficients of the first echo filter and the second echo filter converge, the coefficients of the first echo filter and the second echo filter are frozen. In this embodiment, steps S10 to S14 in the uplink call noise reduction method are replaced by executing: When the access status is detected to switch from connected to disconnected: the echo reference signal is processed by the second echo filter with coefficient freezing to obtain the second echo estimate; based on the second residual environmental signal and the second echo estimate (subtracting the two), the uplink call signal after echo cancellation is obtained; When the access status is detected to switch from disconnected to connected: the echo reference signal is processed by the first echo filter with coefficient freezing to obtain the first echo estimate; based on the first residual environmental signal and the first echo estimate (the two are subtracted), the uplink call signal after echo cancellation is obtained.

[0089] In this embodiment, after the coefficients of the first echo filter and the second echo filter have converged and been frozen, when the external microphone 3 is plugged in or out, there is no need to re-converge the coefficients of the echo filter. This achieves seamless echo cancellation with zero repetition and no jumps, ensuring that the uplink call echo cancellation is uninterrupted and does not degrade before and after the external microphone 3 is plugged in or out.

[0090] As an optional implementation method, the uplink call noise reduction method also includes: Step S15: Perform voice activity detection on the output signal of the built-in microphone 4 to obtain the wearer's voice activity detection result. For example, the wearer's voice activity detection result is 0 or 1, where 1 indicates that the wearer has voice activity and 0 indicates that the wearer has no voice activity; the probability value between 0 and 1 for the wearer's voice activity detection result represents the probability that the wearer has voice activity.

[0091] Step S16: Based on the wearer's voice activity detection result and the speaker drive signal, perform dual-talk detection to obtain the dual-talk detection result. For example, the dual-talk detection result is 0 or 1, where 1 indicates that the wearer's voice and the speaker's voice are present simultaneously, and 0 indicates that the wearer's voice and the speaker's voice are not present simultaneously; the probability value between 0 and 1 for the dual-talk detection result represents the probability that the wearer's voice activity and the speaker's sound will occur simultaneously.

[0092] Step S17: If the wearer's voice activity detection result indicates the presence of the wearer's voice (i.e., 1 or a probability value greater than a preset probability threshold, such as 0.55), and / or the dual-talk detection result indicates that the wearer's voice and the speaker's voice are present simultaneously, then freeze or slow down the coefficient updates of the first echo filter and / or the second echo filter. Specifically, this can be done by updating once every multiple working cycles, or by stopping the coefficient updates.

[0093] When the wearer is engaged in speech activity, and / or when the wearer's speech and the speaker's voice are present simultaneously, it indicates that the wearer's voice is mixed into the microphone signal. If the echo filter coefficients are continuously updated based on the residual echo signal at this time, the echo filter will misjudge the near-end speech as a change in the echo path, resulting in coefficient divergence or damage to the near-end speech. Therefore, the coefficient updates of the first echo filter and / or the second echo filter are frozen or slowed down to keep the echo filter stable during dual-talk. The updates are resumed after the wearer stops speaking to avoid near-end speech distortion and improve call clarity and echo cancellation stability.

[0094] Example 3: This embodiment discloses an uplink call noise reduction method for a headphone system with a detachable microphone. The difference from Embodiment 2 is that after obtaining the echo-cancelled uplink call signal, the echo-cancelled uplink call signal is further subjected to call noise reduction processing to obtain the call noise-reduced uplink call signal, thereby achieving three-level noise reduction processing and ensuring the quality of uplink call voice.

[0095] In this embodiment, the call noise reduction process includes: Step S18 involves performing frame-by-frame windowing processing on the echo-cancelled uplink call signal obtained in Example 2. This process divides the long-term non-stationary signal into short-term stationary signals, facilitating subsequent frequency domain analysis. For example, the frame length is 20ms to 32ms, the inter-frame overlap can be 50%, and the window function used is the Hanning window.

[0096] Step S19: Perform a Fast Fourier Transform (FFT) on the windowed time-domain signal of each frame to obtain the frequency-domain signal. , This is the frequency index. The number of points in the FFT can be the same as the frame length, for example, 512 points.

[0097] Step S20, for the frequency domain signal Noise power spectrum estimation is performed to identify the noise power spectrum and the speech power spectrum. The ratio of the speech power spectrum to the noise power spectrum at each frequency point in the current frame is calculated to obtain the prior signal-to-noise ratio. .

[0098] Specifically, it is not limited to using the existing Minima Controlled Recursive Averaging (MCRA) algorithm for power noise estimation. At each frequency point, it tracks the minimum value of the signal power spectrum, uses the minimum power value as a coarse estimate of the noise power spectrum, and then smoothly updates the coarse noise power spectrum through recursive averaging.

[0099] Step S21, based on the prior signal-to-noise ratio Construct a gain function for the frequency domain signal. Perform frequency domain filtering.

[0100] Specifically, the gain function can be either Wiener gain or Minimum Mean Square Error (MMSE) gain. The expression for Wiener gain is: ; When the prior signal-to-noise ratio When it is very large, Approaching 1 indicates a frequency point Primarily for speech, retain the signal at that frequency; when the prior signal-to-noise ratio... When I was very young, Approaching 0 indicates a frequency point The signal at this frequency point is suppressed, primarily due to noise.

[0101] Gain function With frequency domain signals Multiply to obtain the frequency-domain filtered signal. : .

[0102] Step S22 involves performing an inverse fast Fourier transform on the frequency-domain filtered signal to recover the time-domain signal, which is then superimposed and added to output the noise-reduced uplink call signal, completing the three-stage noise reduction of the uplink call signal. Since step S18 employs frame-by-frame windowing and 50% overlap, it is necessary to superimpose and add the time-domain signals of adjacent frames to reconstruct a continuous uplink call signal. During superposition and addition, a synthesis window compensation can be applied to the windowed signal to eliminate amplitude modulation caused by the window function. Finally, the noise-reduced uplink call signal is output, which is the clear voice signal and can be transmitted back to the host via communication links such as Bluetooth, 2.4GHz proprietary wireless, or USB (Universal Serial Bus) interface.

[0103] In this embodiment, the call noise reduction processing can further suppress residual steady-state background noise after echo cancellation, such as fan noise, air conditioner noise, and keyboard noise, thereby improving the clarity and naturalness of uplink voice. Combining the environmental noise reduction of Embodiment 1 and the echo cancellation of Embodiment 2, this embodiment achieves three-level noise reduction processing: environmental noise reduction, echo cancellation, and call noise reduction, ensuring stable and clear uplink voice quality before and after the external call microphone 3 is plugged in and out.

[0104] Example 4: This embodiment discloses a headphone system. Figure 3 This is a system block diagram of the headphone system. Figure 4 A front view of an example headphone system. Figure 5 for Figure 4 Rear view of the headphone system. The headphone system includes: The headset body 1 has a left earcup 102 and a right earcup 101. The left earcup 102 or the right earcup 101 is provided with a detachable interface and a built-in microphone 4. The headset body 1 is provided with one or more ambient microphones for picking up ambient noise. The left earcup 102 and / or the right earcup 101 are provided with speakers. An extension arm 2 is pluggably connected to a pluggable interface, and an external microphone 3 is provided on the extension arm 2. The plug-in / plug-out detection unit, located on the earphone body 1, is used to detect the connection status of the external call microphone 3; An audio processor and a speaker driver circuit are located inside the headphone body 1. The audio processor is electrically connected to the plug-in detection unit and the speaker driver circuit, respectively, and the speaker driver circuit is electrically connected to the speaker. The audio processor performs the steps of the uplink call noise reduction method for a headphone system with a detachable microphone provided in Embodiment 1, Embodiment 2, or Embodiment 3.

[0105] In this embodiment, the earphone body 1 is preferably a headband type. The built-in microphone 4 can be disposed on the left earcup 102 or the right earcup 101. The pluggable interface and the built-in microphone 4 are disposed on the same earcup, so that the external microphone 3 and the built-in microphone 4 are arranged on the same earcup, which facilitates the centralized pickup of the wearer's voice. When worn, the built-in microphone 4 faces the wearer's mouth. Preferably, one or more environmental microphones include a left environmental microphone 6 installed on the left earcup 102 and a right environmental microphone 5 installed on the right earcup 101. Both the left environmental microphone 6 and the right environmental microphone 5 face outward to pick up ambient sound.

[0106] In this embodiment, the extension arm 2 refers to a slender support member extending from the earcup portion of the earphone body 1, used to support the external microphone 3 and position the acoustic inlet of the external microphone 3 near the wearer's mouth. It generally adopts a rigid or semi-rigid rod, tube, strip, or shaped bent rod structure, and can internally house wires, shielding layers, flexible circuit boards, etc., for transmitting the output signal of the external microphone. A microphone mount is provided at the end of the extension arm 2 for mounting the external microphone 3. A connector is provided at the base of the extension arm 2 to connect with a pluggable interface on the earphone body 1. The connector is not limited to existing spring-loaded contact connectors, magnetic contact connectors, or snap-on contact connectors. The extension arm 2 preferably has a fixed geometry, meaning that its length, bending shape, and angle relative to the earphone body 1 are fixed after deployment, and the position of the external microphone 3 is not changed by hinge rotation during use. The extension arm 2 can only change its configuration by plugging and unplugging; that is, the external microphone 3 or the entire extension arm 2 can be detached from or plugged back into the earphone body 1.

[0107] In this embodiment, in one example, the result of the insertion / removal detection unit is as follows: Figure 6 As shown, the device includes a first resistor R1 and a metal spring K. The first end of the metal spring K is fixedly mounted (e.g., soldered) to a signal contact within a pluggable interface on the headphone body 1, while the second end of the metal spring K is suspended. The first end of the metal spring K is also electrically connected to the first end of the first resistor R1 and the first input / output pin (GPIO1) of the audio processor. The second end of the first resistor R1 is electrically connected to the power supply terminal VCC. When the extension arm 2 is not inserted, the second end of the metal spring K is suspended. Under the pull-up action of the first resistor R1, the audio processor detects a high voltage level at the first input / output pin (GPIO1). When the extension arm 2 is inserted into the pluggable interface on the headphone body 1, the second end of the metal spring K is pushed towards the ground contact until it is electrically connected to the ground contact. Figure 6The metal spring K (located at the dotted line position) causes the voltage of the first input / output pin (GPIO1) to switch from high to low. The audio processor detects this level change and determines that the external microphone's connection status has changed from off to on. Conversely, when the extension arm 2 is pulled out of the pluggable interface on the headset body 1, the metal spring K, under the action of its elasticity, disconnects from the ground contact and returns to its floating position, causing the voltage of the first input / output pin (GPIO1) to switch from low to high. The audio processor detects this level change and determines that the external microphone's connection status has changed from on to off. This hardware pin detection method for connection status monitoring is low-cost and reliable.

[0108] In this embodiment, the insertion / removal detection unit may also be a strain sensor (to identify the connection status by sensing the pressure of the root connector of the extension arm 2), a position switch, etc.

[0109] The embodiment is merely a specific example and does not indicate that this is the only way to implement the present invention.

[0110] The above description is merely a preferred embodiment of the present invention. Those skilled in the art will understand that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the present invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. An uplink call noise reduction method for a headset system with a detachable microphone, characterized in that, The headset system includes a headset body, an extension arm detachably connected to the headset body, an external microphone mounted on the extension arm, a built-in microphone on the headset body, and one or more ambient microphones for picking up ambient noise; the method is performed periodically: Monitor the access status of the external microphone; An environmental reference signal is obtained based on the output signals of one or more environmental microphones. The environmental reference signal is processed using a second environmental denoising filter to obtain a second environmental noise estimate; When the access status is connected: the output signal of the external microphone and the output signal of the built-in microphone are combined to obtain a synthesized call signal; The environmental reference signal is processed using a first environmental denoising filter to obtain a first environmental noise estimate; based on the synthesized call signal and the first environmental noise estimate, an environmentally denoised uplink call signal is obtained, and the coefficients of the first environmental denoising filter are updated using the environmentally denoised uplink call signal as the first residual environmental signal; Based on the output signal of the built-in microphone and the second environmental noise estimation, a second residual environmental signal is obtained, and the coefficients of the second environmental denoising filter are updated using the second residual environmental signal. When the access status is disconnected: Based on the output signal of the built-in call microphone and the second environmental noise estimation, the uplink call signal after environmental denoising is obtained, and the coefficients of the second environmental denoising filter are updated with the uplink call signal after environmental denoising as the second residual environmental signal.

2. The uplink call noise reduction method for a headphone system with a detachable microphone according to claim 1, characterized in that, When the access state switch is detected, within a preset future time period, the uplink call signal after environmental denoising at the previous discrete time point and the uplink call signal after environmental denoising at the current discrete time point are synthesized using a cross-fading method to obtain the uplink call signal after environmental denoising and fading processing, and the uplink call signal is output.

3. The uplink call noise reduction method for a headphone system with a detachable microphone according to claim 1, characterized in that, When the access state is detected to switch from connected to disconnected, the coefficients of the first environmental noise reduction filter are frozen. When the access state is detected to switch back from disconnected to connected, the environmental reference signal is processed using the first environmental noise reduction filter with frozen coefficients to obtain the first environmental noise estimate.

4. The uplink call noise reduction method for a headphone system with a detachable microphone according to claim 1, characterized in that, Once the coefficients of the first environmental denoising filter and the second environmental denoising filter converge, the coefficients of the first environmental denoising filter and the second environmental denoising filter are frozen; after the step of obtaining the environmental reference signal based on the output signal of one or more environmental microphones, it is replaced by performing: When the access state is detected to switch from connected to disconnected: the environmental reference signal is processed using a second environmental noise reduction filter with coefficient freezing to obtain a second environmental noise estimate; based on the output signal of the built-in call microphone and the second environmental noise estimate, the uplink call signal after environmental noise reduction is obtained; When the access status is detected to switch from disconnected to connected: the output signal of the external microphone and the output signal of the built-in microphone are merged to obtain a synthesized call signal; The environmental reference signal is processed using a first environmental denoising filter with coefficient freezing to obtain a first environmental noise estimate; based on the synthesized call signal and the first environmental noise estimate, an environmentally denoised uplink call signal is obtained.

5. The uplink call noise reduction method for a headphone system with a detachable microphone according to claim 1, characterized in that, The process of obtaining an environmental reference signal based on the output signals of one or more environmental microphones includes: The output signals of one or more environmental microphones are synthesized to obtain the original environmental reference signal; When the access status is connected, the original environmental reference signal is pre-filtered using the first-level path estimation model to obtain the environmental reference signal. When the access status is disconnected, the original environmental reference signal is pre-filtered using the second-level path estimation model to obtain the environmental reference signal.

6. The uplink call noise reduction method for a headphone system with a detachable microphone according to claim 1, characterized in that, The left and right ambient microphones are respectively provided on the left and right sides of the earphone body; The process of obtaining an environmental reference signal based on the output signals of one or more environmental microphones includes: Based on the time difference of arrival between any two of the output signals from the left ambient microphone, the right ambient microphone, and the built-in communication microphone, the spatial direction of the sound source is estimated. If the estimated spatial direction of the sound source matches the direction of the built-in microphone toward the wearer's mouth, then the signal components corresponding to the spatial direction of the sound source in the output signals of the left and right ambient microphones are attenuated or removed, while the signal components corresponding to the direction components that do not match the direction of the built-in microphone toward the wearer's mouth are retained. Based on the attenuated or removed output signals of the left and right ambient microphones, the environmental reference signal is obtained.

7. The uplink call noise reduction method for a headset system with a detachable microphone according to any one of claims 1-3, 5, and 6, characterized in that, The earphone body is provided with a speaker driver circuit and one or more speakers, and the method further includes: The loudspeaker driving signal output by the loudspeaker driving circuit is used as the echo reference signal; The echo reference signal is processed using a second echo filter to obtain a second echo estimate; When the access status is "connected": the echo reference signal is processed using a first echo filter to obtain a first echo estimate; based on the first residual environmental signal and the first echo estimate, an echo-cancelled uplink call signal is obtained, and the coefficients of the first echo filter are updated using the echo-cancelled uplink call signal as the first residual echo signal; based on the second residual environmental signal and the second echo estimate, a second residual echo signal is obtained; and the coefficients of the second echo filter are updated based on the second residual echo signal. When the access status is disconnected: Based on the second residual environmental signal and the second echo estimation, the uplink call signal after echo cancellation is obtained, and the coefficients of the second echo filter are updated using the uplink call signal after echo cancellation as the second residual echo signal.

8. The uplink call noise reduction method for a headphone system with a detachable microphone according to claim 7, characterized in that, The method further includes: The output signal of the built-in microphone is subjected to voice activity detection to obtain the wearer's voice activity detection result. Based on the wearer's voice activity detection results and the speaker drive signal, dual-talk detection is performed to obtain dual-talk detection results; If the wearer's voice activity detection result indicates the presence of the wearer's voice, and / or the dual-talk detection result indicates the simultaneous presence of the wearer's voice and the voice played by the speaker, then the coefficient updates of the first echo filter and / or the second echo filter are frozen or slowed down.

9. The uplink call noise reduction method for a headphone system with a detachable microphone according to claim 7, characterized in that, The method further includes: performing call noise reduction processing on the echo-cancelled uplink call signal to obtain a call noise-reduced uplink call signal.

10. A headphone system, characterized in that, The headphone system includes: The headset body is provided with a left earcup and a right earcup. The left earcup or the right earcup is provided with a detachable interface and a built-in microphone for communication. The headset body is provided with one or more ambient microphones for picking up ambient noise. The left earcup and / or the right earcup is provided with a speaker. An extension arm that can be plugged into the pluggable interface, wherein an external microphone is provided on the extension arm; A plug-in / plug-out detection unit, located on the earphone body, is used to detect the connection status of the external call microphone; An audio processor and a speaker driving circuit are located inside the headphone body. The audio processor is electrically connected to the plug-in detection unit and the speaker driving circuit, respectively, and the speaker driving circuit is electrically connected to the speaker. The audio processor performs the steps of the uplink call noise reduction method for a headphone system with a detachable microphone as described in any one of claims 1-9.