Noise reduction method based on sidetone, active noise reduction earphone and storage medium

Abstract

The invention provides a sidetone-based noise reduction method, an active noise reduction earphone and a storage medium, the sidetone-based noise reduction method is applied to a processor of the active noise reduction earphone, the active noise reduction earphone comprises a sidetone channel and a feedback noise reduction channel, and a feedback noise reduction filter is used for filtering a feedback audio signal; the noise reduction method based on sidetone comprises the following steps: in response to opening of the sidetone channel, controlling the sidetone channel to output a collected sidetone signal to the feedback noise reduction channel; under the condition that the feedback noise reduction channel receives the sidetone signal, controlling the feedback noise reduction channel to counteract the sidetone component in the feedback audio signal based on the sidetone signal; and controlling the feedback noise reduction filter to filter the feedback audio signal offsetting the sidetone component. According to the method, when active noise reduction is carried out on the earphone, a user can hear and feel speaking sound.

Description

Technical Field

[0001] The present application relates to the field of speech noise reduction, and in particular to a sidetone-based noise reduction method, an active noise reduction headset, and a storage medium. Background Art

[0002] Active noise-cancelling headphones have better noise reduction capabilities and can shield the sound outside the headphones to a certain extent.

[0003] When a person is speaking, if he cannot hear his own voice during the conversation, he will actively raise his voice. When wearing active noise reduction headphones for voice calls, he may not be able to hear his own voice and thus speak too loudly, affecting the listening experience of the other party during the call and the experience of other people in the wearer's location.

[0004] At present, some existing active noise reduction headphones provide a sidetone function. Sidetone refers to the phenomenon that the speaker hears his or her own voice through the receiver during a phone call. The existence of sidetone allows the speaker to perceive his or her own voice, thereby adjusting the volume and tone of the speech, providing a natural call experience. During a call, the audio signal collected by the feedback noise reduction channel of the active noise reduction headphones may include a sidetone signal component, so that the feedback noise reduction channel cancels the sidetone based on the sidetone signal component when filtering, so that the wearer cannot hear the sidetone. Summary of the invention

[0005] In view of this, the present application aims to provide a sidetone-based noise reduction method, an active noise reduction headset and a storage medium, so that when the headset is performing active noise reduction, the user can hear and feel the speaking sound, thereby making the call more natural.

[0006] In a first aspect, an embodiment of the present application provides a sidetone-based noise reduction method, which is applied to a processor of an active noise reduction headset, wherein the active noise reduction headset includes a sidetone channel and a feedback noise reduction channel, the sidetone channel includes a sidetone microphone for collecting a sidetone signal; the feedback noise reduction channel includes a feedback microphone and a feedback noise reduction filter, the feedback microphone is used to collect a feedback audio signal, and the feedback noise reduction filter is used to filter the feedback audio signal; the sidetone-based noise reduction method includes: in response to the activation of the sidetone channel, controlling the sidetone channel to output the collected sidetone signal to the feedback noise reduction channel; when the feedback noise reduction channel receives the sidetone signal, controlling the feedback noise reduction channel to cancel the sidetone component in the feedback audio signal based on the sidetone signal; and controlling the feedback noise reduction filter to filter the feedback audio signal that cancels the sidetone component.

[0007] In an embodiment of the present application, a sidetone channel is provided in the headset, which is different from the feedback noise reduction channel. The sidetone channel can collect the sidetone, thereby obtaining a sidetone signal that can be used to offset the sidetone component in the feedback audio signal collected by the feedback noise reduction channel, so that after the sidetone component in the feedback audio signal is offset, the feedback audio signal finally output by the feedback noise reduction channel will not filter the sidetone in the ear, thereby retaining the sidetone. The user can hear the sidetone and adjust the volume of his or her voice according to the sidetone, providing a more natural call experience.

[0008] In one embodiment, in response to the activation of the sidetone channel, the sidetone channel is controlled to output the collected sidetone signal to the feedback noise reduction channel before the method further includes: detecting whether the wearer of the active noise reduction headset is in a target state, the target state indicating that the wearer is in a call and / or speaking; if the wearer is in the target state, the sidetone channel is activated.

[0009] The sidetone channel provides a sidetone signal to the feedback noise reduction channel to offset the sidetone component in the feedback audio signal. However, when there is no need to listen to the sidetone, offsetting the sidetone component will reduce the noise reduction effect, thereby affecting the listening experience. Therefore, in an embodiment of the present application, it can be determined whether the wearer is in a target state where the sidetone needs to be listened to during speaking or talking. In the target state, the sidetone channel is turned on. As a result, the wearer can hear the sidetone in the target state, adjust the volume of his or her own speech, and provide a more natural call effect. In the non-target state, the sidetone channel is not turned on, the sidetone channel will not play the sidetone in the ear, and the ambient noise in the ear will not be increased due to the sidetone channel, so that the noise reduction effect is better.

[0010] In one embodiment, after detecting whether the wearer of the active noise reduction headset is in a target state, the method further includes: if the wearer is not in the target state, closing the sidetone channel.

[0011] In this embodiment, if the wearer is not in the target state, it indicates that the wearer does not need to listen to the sidetone. In this case, the sidetone channel is turned off so that the feedback noise reduction channel can be better filtered and the power consumption generated by the sidetone channel is reduced.

[0012] In one embodiment, the sidetone channel further includes multiple functional modules, a sidetone gain module and multiple filter circuits; the sidetone microphone and the multiple functional modules are connected in series; turning on the sidetone channel includes: putting the sidetone microphone and each of the multiple functional modules in a working state; and closing the sidetone channel includes: turning off the sidetone microphone and at least one of the multiple functional modules.

[0013] In this embodiment, the sidetone circuit is implemented by a hardware circuit, and the sidetone channel is in the transmission path of the audio signal to the speaker. Therefore, the processor does not need to process the sidetone audio signal through a software program, and then play the processed sidetone signal through the speaker, which reduces the delay of the sidetone channel and improves the listening experience of the sidetone signal. In addition, the hardware circuit has a higher response speed and processing efficiency, which effectively improves the efficiency of gain and filtering processing of the sidetone signal, and can also reduce the delay of the sidetone signal in gain and filtering processing. The sidetone microphone and each functional module in the sidetone channel are connected in series. After one of them is turned off, the sidetone channel cannot output the sidetone signal. Therefore, when the sidetone channel is turned off, some of the functional modules therein can be turned off. Therefore, when it is turned on, only part of the circuit that has been turned off needs to be turned on, without waiting for all circuits to be turned on, thereby improving the efficiency of turning on the sidetone channel and reducing the delay of sidetone signal processing.

[0014] In one embodiment, the sidetone channel also includes a sidetone gain module and multiple filter circuits; the sidetone microphone, the sidetone gain module and the multiple filter circuits are connected in series, and the sidetone gain module is arranged between the sidetone microphone and the filter circuit, or the sidetone gain module is arranged between any two of the filter circuits; the sidetone microphone, the sidetone gain module and each of the filter circuits are in working state; opening the sidetone channel includes: adjusting the gain of the sidetone gain module to be greater than or equal to a first preset value; closing the sidetone channel includes: adjusting the gain of the sidetone gain module to be less than or equal to a second preset value, or, making the gain of the sidetone gain module 0; the second preset value is less than the first preset value.

[0015] In an embodiment of the present application, when the gain of the sidetone gain module in the sidetone channel is too small (less than a preset value) or is 0, the sidetone channel will not be able to output a sidetone signal that can be used to offset the sidetone component, thereby achieving the closure of the sidetone channel. Correspondingly, when the gain of the sidetone gain module is greater than the first preset value, the sidetone channel can output the sidetone signal normally. In this embodiment, each module and circuit in the gain sidetone channel is always in a working state, and the sidetone channel can be turned on and off by only adjusting the gain. As a result, it is possible to switch between closing and opening without waiting for each module and circuit to start, effectively reducing the delay.

[0016] In one embodiment, the active noise reduction headset further includes an in-ear microphone, and the detecting whether the wearer of the active noise reduction headset is in the target state includes: acquiring an in-ear audio signal collected by the in-ear microphone; performing a wearer's voice activity detection based on the difference between the in-ear audio signal and the sidetone signal; and determining whether the wearer is in the target state of speaking based on the detection result of the wearer's voice activity detection.

[0017] When the wearer speaks, vibrations occur in the ear, causing changes in the phase and / or amplitude of the audio signal in the ear. Therefore, in an embodiment of the present application, the in-ear audio signal can be collected and compared with the sidetone signal. If the difference between the two is small, it indicates that the wearer is speaking. If the difference is large, it indicates that the wearer is not speaking, and it may be someone other than the wearer who is speaking. Therefore, it can be determined based on the difference between the in-ear audio signal and the sidetone signal whether the wearer is speaking, thereby determining whether the wearer is in the target state, so as to more accurately open or close the sidetone channel and improve the listening experience.

[0018] In one embodiment, the active noise cancelling headset is communicatively connected to a smart device; detecting whether the wearer of the active noise cancelling headset is in a target state includes: receiving data and / or instructions transmitted by the smart device; and determining whether the wearer is in the target state of the call process based on the data and / or instructions transmitted by the smart device.

[0019] When the wearer uses a smart device to talk with the headset, data and / or instructions are transmitted between the headset and the smart device. Therefore, in the embodiment of the present application, it is possible to determine whether the wearer is in the target state of the call process through the transmitted data and / or instructions, so as to improve the accuracy of target state recognition, and further improve the accuracy of opening or closing the sidetone channel.

[0020] In one embodiment, the sidetone microphones include multiple ones, and before the sidetone channel is controlled to output the collected sidetone signal to the feedback noise reduction channel, the method further includes: respectively acquiring the sidetone sub-signals collected by each of the sidetone microphones; performing beamforming processing in a preset direction on each of the sidetone sub-signals to obtain the sidetone signal, and the preset direction is used to characterize the direction of the active noise reduction headset pointing to the wearer's mouth.

[0021] In an embodiment of the present application, the sidetone signal is obtained by beamforming the sidetone sub-signals collected by multiple sidetone microphones, and the direction of the beamforming is the direction of the active noise reduction headphones pointing towards the wearer's mouth. As a result, the sidetone signal can more accurately represent the sound of the wearer's speech, thereby improving the accuracy of filtering based on the sidetone signal.

[0022] In one embodiment, the sidetone channel further includes a sidetone filter, the sidetone filter includes a bandpass filter, the bandwidth range of the bandpass filter is a preset bandwidth range, and the preset bandwidth range represents the bandwidth range of human voice; the sidetone signal is a signal filtered by the bandpass filter.

[0023] In the embodiment of the present application, the bandwidth range of the bandpass filter is the bandwidth range of the human voice. Therefore, the bandpass filter can more accurately retain the human voice and filter out other noise in the sidetone signal. Based on the sidetone signal, the sidetone component in the feedback audio signal can be better offset, so that the wearer can better hear the noise and provide a better noise reduction effect.

[0024] In one embodiment, the sidetone channel includes a sidetone filter, the sidetone filter includes a FIR adaptive filter, and the method further includes: acquiring the sidetone signal; converting the sidetone signal into a frequency domain signal; performing speech noise reduction on the sidetone signal in the frequency domain to obtain a frequency domain gain vector; converting the frequency domain gain vector in the time domain to obtain a time domain gain filter; the time domain gain filter includes filter coefficients for updating the FIR adaptive filter; and mapping the filter coefficients of the time domain gain filter to the filter coefficients of the FIR adaptive filter to update the filter coefficients of the FIR adaptive filter.

[0025] The filter coefficients of the FIR adaptive filter will be adjusted according to the changes in the environment in which the active noise reduction headphones are located, thereby providing a noise reduction effect that is more adapted to the environment. In the embodiment of the present application, the updated filter coefficients of the FIR adaptive filter are calculated by the processor, so that the filtering of the sidetone signal by the FIR adaptive filter and the calculation of the updated filter coefficients by the processor do not affect each other, effectively reducing the delay generated when filtering the sidetone component when calculating the updated filter coefficients.

[0026] In a second aspect, an embodiment of the present application provides an active noise reduction headset, comprising: a sidetone channel, including a sidetone microphone for collecting a sidetone signal; a feedback noise reduction channel, including a feedback microphone and a feedback noise reduction filter, the feedback microphone is used to collect a feedback audio signal, and the feedback noise reduction filter is used to filter the feedback audio signal; a processor, used to execute the sidetone-based noise reduction method as described in any one of the first aspects.

[0027] In a third aspect, an embodiment of the present application provides a readable storage medium, wherein the readable storage medium stores instructions, and the instructions can be executed by a processor to implement the sidetone-based noise reduction method as described in any one of the first aspects. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0029] Figure 1 A schematic diagram of the structure of an active noise reduction headset provided in one embodiment of the present application;

[0030] Figure 2 A schematic diagram of a noise reduction channel of an active noise reduction headset provided in an embodiment of the present application;

[0031] Figure 3 A schematic diagram of a sidetone filter provided in an embodiment of the present application;

[0032] Figure 4 A flowchart of a sidetone-based noise reduction method provided in an embodiment of the present application;

[0033] Figure 5 FIG. 1 is a schematic diagram of updating filter coefficients of a FIR adaptive filter provided by an embodiment of the present application;

[0034] Figure 6 A schematic diagram of a sidetone-based noise reduction device provided in one embodiment of the present application.

[0035] Icon: active noise reduction headset 100; processor 110; sidetone channel 120, sidetone microphone 121; sidetone filter 122; sidetone gain module 1221; IIR configurable filter 1222; FIR adaptive filter 1223; feedback noise reduction channel 130; feedback microphone 131; feedback noise reduction filter 132; feedforward active noise reduction channel 140; feedforward microphone 141; feedforward noise reduction filter 142; speaker 150; audio echo module 161; ADC 163; low-pass and downsampling filter 164; limiter 165; DAC 170; sidetone channel control module 510; feedback noise reduction channel control module 520; noise reduction control module 530. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solution and advantages of the embodiments of the present application more clear, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application and are not used to limit the present application.

[0037] See also Figure 1 , Figure 1 Schematic diagram of an active noise reduction headset provided in an embodiment of the present application. The present application provides an active noise reduction headset, which can be a semi-in-ear headset, an open-ear headset, or a head-mounted headset, without limitation. In an embodiment of the present application, the active noise reduction headset can be one of a left headset or a right headset, or can include both a left headset and a right headset.

[0038] The active noise reduction headset 100 may include: a processor 110 , a sidetone channel 120 , a feedback noise reduction channel 130 , and a speaker 150 . In some embodiments of the present application, the active noise reduction headset 100 may further include a feedforward noise reduction channel 140 .

[0039] The processor 110 is connected to the sidetone channel 120 and the feedback noise reduction channel 130 respectively, and is used to execute the sidetone-based noise reduction method provided in the present application. The sidetone-based noise reduction method will be expanded later and will not be repeated here.

[0040] In the embodiment of the present application, the sidetone channel 120, the feedforward noise reduction channel 140 and the feedback noise reduction channel 130 may all be referred to as noise reduction channels. Figure 2 , Figure 2 A schematic diagram of a noise reduction channel provided in an embodiment of the present application.

[0041] The sidetone channel 120 includes a sidetone microphone 121 and a sidetone filter 122. The sidetone microphone 121 is used to collect sidetone signals, and the sidetone filter 122 can filter and reduce noise on the sidetone signals to obtain a sidetone signal after filtering and noise reduction. The sidetone microphone 121 is arranged outside the earphone to collect sounds outside the ear, and is particularly used to collect sounds directed toward the wearer's mouth.

[0042] In some embodiments of the present application, the sidetone microphone 121 includes multiple sidetone microphones (not shown in the figure), and different sidetone microphones 121 are arranged at different positions of the active noise reduction headset 100. Different sidetone microphones 121 can collect sidetone sub-signals respectively, so that the processor 110 can perform beamforming processing according to the sidetone sub-signals collected by different sidetone microphones 121 to obtain a sidetone signal processed by beamforming.

[0043] In the embodiment of the present application, the sidetone channel 120 can collect the sidetone signal and output it to the DAC (digital-to-analog converter) 170 after filtering, and then the DAC 170 converts it into a module signal and plays it through the speaker 150.

[0044] In the embodiment of the present application, the sidetone filter 122 is connected after the sidetone microphone 121, see Figure 3 , Figure 3The schematic diagram of the structure of the sidetone filter 122 provided in one embodiment of the present application, the sidetone filter 122 includes a sidetone gain module 1221 and a plurality of filter circuits, and the plurality of filter circuits may include an IIR configurable filter 1222 circuit, a FIR adaptive filter 1223 circuit, etc. The sidetone microphone 121, the sidetone gain module 1221 and the plurality of filter circuits are connected in series, and the sidetone gain module 1221 is arranged between the sidetone microphone 121 and the filter circuit, or the sidetone gain module 1221 is arranged between any two filter circuits. That is, the sidetone gain module 1221 can be connected in series at any position in the sidetone filter 122, without limitation here.

[0045] The feedback noise reduction channel 130 includes a feedback microphone 131 and a feedback noise reduction filter 132. The feedback microphone 131 is arranged on the inner side of the earphone and is close to the ear canal when worn. It is used to collect the sound in the ear, that is, to collect the feedback audio signal. The feedback noise reduction filter 132 can perform noise reduction filtering on the feedback audio signal collected by the feedback microphone 131 to obtain a feedback audio signal for filtering the audio signal in the ear. After being converted into an analog signal by the DAC 170, it is played by the speaker 150. The played feedback audio signal can offset the noise in the ear, thereby performing filtering.

[0046] For ease of understanding, the working principle of the feedback noise reduction channel 130 is explained here. The feedback noise reduction channel 130 is used to collect the audio signal in the cavity between the earphone and the ear canal, and generate a corresponding feedback audio signal, so that the speaker 150 plays the feedback audio signal to offset the audio signal in the cavity, thereby achieving noise reduction. In the cavity, in addition to noise, the audio signal includes the audio component to be played of the audio to be played. The audio component to be played needs to be retained and heard by the user. Therefore, through the audio echo module 161, the audio to be played can be sent to the feedback noise reduction channel 130 to offset the audio component to be played in the feedback noise reduction signal through the audio to be played, so that the audio to be played in the cavity will not be offset based on the feedback noise reduction signal, thereby retaining the audio to be played.

[0047] Similarly, in an embodiment of the present application, the sidetone signal collected by the sidetone channel 120 is converted into an analog signal by the DAC 170 and then played by the speaker 150, and the feedback audio signal collected by the feedback microphone 131 may include a component of the sidetone signal, which may cause the feedback noise reduction channel 130 to cancel the sidetone signal through the feedback audio signal, thereby making it impossible for the user to hear the sidetone signal. Based on this, in an embodiment of the present application, the sidetone signal collected by the sidetone channel 120 can be output to the audio echo module 161, and output to the feedback noise reduction channel 130 through the audio echo module 161 to cancel the sidetone component in the feedback noise reduction signal, thereby causing the feedback noise reduction channel 130 to not cancel the sidetone signal in the cavity, so that the sidetone signal can be retained and heard by the user.

[0048] Accordingly, in an embodiment of the present application, the active noise reduction headset 100 may further include a feedforward noise reduction channel 140, the feedforward noise reduction channel 140 including a feedforward microphone 141 and a feedforward noise reduction filter 142, the feedforward microphone 141 being arranged outside the ear for collecting sounds outside the wearer's ear, and the feedforward noise reduction filter 142 being used for filtering the audio signal collected by the feedforward microphone 141. Through the feedforward noise reduction channel 140, the audio signal may be feedforward noise reduced, and the specific implementation method of the feedforward noise reduction may refer to the prior art and will not be expanded here. Through the feedforward noise reduction channel 140, the noise reduction effect of the active noise reduction headset 100 may be improved.

[0049] In the embodiment of the present application, the feedforward denoising filter 142 and the feedback denoising filter 132 may be adaptive filters or fixed filters, and may be filters with an IIR structure, filters with an FIR structure, or filters with a mixture of IIR and FIR structures.

[0050] In some embodiments of the present application, some feedforward microphones 141 and some sidetone microphones 121 can be reused, which is not limited here. For example, the active noise reduction headset 100 includes N1 feedforward microphones 141, where N2 feedforward microphones 141 can be configured as sidetone microphones 121, N2 and N1 are both positive integers, and N2 is less than N1.

[0051] In some other embodiments of the present application, the feedforward microphone 141 and the sidetone microphone 121 may also be relatively independent, which is not limited here.

[0052] It should be noted that the difference between the aforementioned sidetone microphone 121, feedforward microphone 141 and feedback microphone 131 is that they are set at different positions and have different functions in the active noise reduction headset 100, and the structures and types of the microphones used by each can be the same or different, which is not limited here.

[0053] In some other embodiments of the present application, the sidetone microphone 121 is an analog microphone, and the feedforward microphone 141 and the feedback microphone 131 are digital microphones. In this embodiment, the sidetone gain module and the ADC 163 may not be set in the feedforward noise reduction channel 140 and the feedback noise reduction channel 130.

[0054] In addition, if Figure 2 As shown, each noise reduction channel may include a side tone gain module, an ADC 163 (Analog to Digital Converter), and a low pass and down sampling filter 164. The side tone gain module of each noise reduction channel may amplify the signal collected by the channel, convert it into a digital signal through the ADC 163, and perform low pass and down sampling processing through the low pass and down sampling filter 164. The side tone gain module, the ADC 163, and the low pass and down sampling filter 164 may refer to the prior art and will not be expanded here.

[0055] In the embodiment of the present application, each of the sidetone channel 120 , the feedforward noise reduction channel 140 , and the feedback noise reduction channel 130 may include a limiter 165 , or may not include a limiter 165 , which is not limited herein.

[0056] The sidetone needs to be heard by the wearer. The sidetone is processed by the sidetone channel 120 and played through the speaker 150. In actual scenarios, the user's speaking voice may be heard by the user through the transmission of the physical space. Therefore, this requires that the user's speaking voice and the sidetone are as synchronized as possible, otherwise an echo may occur. This requires that the time delay between the output of the sidetone and the user's speaking voice is as low as possible.

[0057] Therefore, in the embodiment of the present application, the sidetone microphone 121, the sidetone filter 122 and other functional modules of the sidetone channel 120 can be implemented by hardware circuits, such as analog amplifier modules, ADC, low-pass and down-sampling filters, and sidetone filters are all implemented by corresponding circuits or devices. Compared with software programs, hardware circuits have lower latency, which can improve the timeliness of the output and playback of the sidetone signal, so as to synchronize it with the user's voice as much as possible.

[0058] Next, the sidetone-based noise reduction method provided in the embodiment of the present application will be described. The sidetone-based noise reduction method can be implemented as a software program and executed by the aforementioned processor 110. Figure 4 , Figure 4 A flow chart of a sidetone-based noise reduction method provided in an embodiment of the present application. The sidetone-based noise reduction method includes:

[0059] S110, in response to the sidetone channel being turned on, controlling the sidetone channel to output the collected sidetone signal to the feedback noise reduction channel.

[0060] In this embodiment, after the sidetone channel 120 is turned on, the sidetone microphone 121 can collect the sidetone, and after being processed by the sidetone filter 122 and other modules, it is played by the speaker 150, so that the wearer can hear the sidetone, and then the wearer can adjust the volume of his or her own speech according to the sidetone.

[0061] After the sidetone channel 120 is turned on, there is sidetone played by the speaker 150 in the cavity between the earphone and the ear canal. When the feedback noise reduction signal collects the feedback audio signal for filtering, the feedback noise reduction signal will carry the sidetone component, so that the feedback noise reduction signal will filter the sidetone in the cavity, so that the sidetone heard by the wearer is greatly attenuated by the feedback noise reduction, so as to reduce the listening experience of the sidetone. Therefore, in an embodiment of the present application, after the sidetone channel 120 is turned on, the sidetone channel 120 is controlled to output the collected sidetone signal to the feedback noise reduction channel 130, so as to offset the sidetone component in the feedback audio signal based on the sidetone signal.

[0062] In actual scenarios, listening to sidetone is not applicable to all scenarios. For example, when listening to music or watching videos, the wearer of the headphones would rather immerse himself in the sound of the audio corresponding to the music or video, and minimize the interference of external ambient sound. Therefore, in an embodiment of the present application, the conditions for opening the sidetone channel 120 can be set.

[0063] In one embodiment of the present application, before S110, it can be detected whether the wearer of the active noise reduction headset 100 is in the target state, and if the wearer is in the target state, the sidetone channel 120 is turned on. On the contrary, if the wearer is not in the target state, the sidetone channel 120 is turned off.

[0064] The sidetone is used to enable the wearer to hear the sound of his or her own speech. Therefore, the sidetone channel 120 can be turned on when the wearer is speaking. Correspondingly, the target state can represent that the wearer is speaking.

[0065] Similarly, the wearer is usually in the process of talking when speaking, so the target state can also represent that the wearer is in the process of talking. In the embodiments of the present application, the calling process includes but is not limited to dialing calls, voice functions of various applications, online meetings, etc., which are not limited here.

[0066] Thus, by judging whether the wearer is in the target state, the side tone can be turned on when the wearer needs to listen to the side tone, providing a more natural call effect, and turned off when the wearer does not need to listen to the side tone, so as to reduce the increase of the ambient noise in the ear due to the side tone channel playing the side tone in the ear, and at the same time, it can also effectively reduce the situation where the side tone channel 120 is mistakenly turned on. In addition, when the side tone channel 120 is turned off, the power consumption of the side tone channel 120 can be reduced.

[0067] The active noise reduction headset 100 also includes an in-ear microphone, which can be multiplexed with the feedback microphone 131, or the two can be independent of each other. In one embodiment of the present application, detecting whether the wearer is in the target state may include: obtaining an in-ear audio signal collected by the in-ear microphone; performing a wearer voice activity detection based on the difference between the in-ear audio signal and the sidetone signal; and determining whether the wearer is in the target state of speaking according to the detection result of the wearer voice activity detection.

[0068] When the wearer speaks, the wearer's ear canal vibrates, causing the phase and / or amplitude of the audio signal in the ear to change. Therefore, the audio signal in the ear can be collected to determine whether the phase and / or amplitude of the audio signal in the ear matches the sidetone signal (such as whether the similarity or difference meets the preset requirements), and then determine whether the user is speaking.

[0069] Therefore, in an embodiment of the present application, the wearer's voice activity detection can be performed based on the difference between the in-ear audio signal and the sidetone signal. For example, the difference can be to calculate the in-ear audio signal and the sidetone signal for correlation processing to obtain the correlation between the two, or to perform dot product processing to obtain the dot product result of the two, so as to determine the difference based on the correlation or the dot product result, and then judge whether the wearer is speaking. It can also be judged whether the difference value in phase and / or amplitude between the sidetone signal and the in-ear audio signal is less than a preset threshold, so as to judge whether the wearer is speaking. In addition, it is also possible to use a neural network model to detect the wearer's voice activity. The specific method of voice activity detection can use existing technology, which will not be expanded here.

[0070] It should be distinguished that the voice activity detection object of the present application is different from that of the prior art. The present application performs voice activity detection on the wearer rather than on the voice activity of others. Therefore, in the embodiments of the present application, the correlation (or difference) between the in-ear audio signal and the sidetone signal is used for voice detection to determine whether the wearer is speaking.

[0071] Based on the detection result of the wearer's voice activity detection, if the difference between the in-ear audio signal and the sidetone signal is determined to be less than a set threshold, it is determined that the in-ear audio signal and the sidetone signal are consistent, and it is highly likely that the wearer is speaking, and the sidetone channel 120 can be turned on at this time. On the contrary, if it is determined that the difference between the in-ear audio signal and the sidetone signal is greater than the preset threshold, it indicates that the difference between the in-ear audio signal and the sidetone signal is large, and the collected sidetone signal is highly likely to be the sound of other people or other sound sources at the wearer's location. Therefore, it can be determined that the user is not speaking, and the sidetone channel 120 is turned off.

[0072] The active noise reduction headset 100 is communicatively connected with the smart device to play control instructions, audio data, etc. sent by the smart device. In an embodiment of the present application, detecting whether the wearer is in a target state may include: receiving data and / or instructions transmitted by the smart device, and determining whether the wearer is in the target state of the call process based on the data and / or instructions transmitted by the smart device.

[0073] In an embodiment of the present application, the instruction transmitted by the smart device may include an instruction to instruct the transmission of audio data of a call. After receiving the instruction, the headset determines that it is in the target state of the call process, thereby opening the sidetone channel 120.

[0074] The data transmitted by the smart device may be call data, audio data, etc. There are differences between audio data of a call and audio data such as music, and the specifics may refer to the definition of audio data. Therefore, in other embodiments of the present application, the type of the transmitted data may be detected. If the transmitted data is audio data of a call, it is determined that the target state is in the call process, thereby opening the sidetone channel 120.

[0075] like Figure 2 As shown, the sidetone gain module, ADC 163, low-pass and down-sampling filter 164, and limiter 165 of the sidetone channel 120 are different functional modules of the sidetone channel 120, and the sidetone microphone 121 and the functional modules are connected in series. If the sidetone channel 120 needs to be turned on, the sidetone microphone 121 and the functional modules need to be in a working state. Therefore, turning on the sidetone channel 120 includes: making the sidetone microphone 121 and the multiple functional modules in a working state.

[0076] On the contrary, if the sidetone microphone 121 and any one of the functional modules are in a turned-off state, the sidetone channel 120 will not work. Therefore, in an embodiment of the present application, turning off the sidetone channel 120 includes: turning off the sidetone microphone 121 and at least one of the multiple functional modules, that is, turning off at least one of the sidetone microphone 121, the sidetone gain module, the ADC 163, the low-pass and down-sampling filter 164, and the limiter 165.

[0077] In this way, the control logic can be simplified, and when the sidetone channel 120 needs to be turned on, it is only necessary to control the sidetone microphone 121 or some functional modules to be turned on, without having to reopen all modules, thereby effectively improving the efficiency of turning on the sidetone channel 120 and reducing delay.

[0078] In an embodiment of the present application, the sidetone channel 120 includes a sidetone gain module 1221, which is used to amplify the sidetone signal. The sidetone gain module 1221 can be as follows: Figure 2 The side tone gain module shown in Figure 1 is Figure 2 As shown, the sidetone channel 120 includes a low-pass and down-sampling filter and a sidetone filter 122. Figure 3 As shown, the sidetone filter 122 includes an IIR configurable filter 1222 and a FIR adaptive filter 1223, that is, the sidetone channel 120 includes a plurality of filter circuits, and a gain module is usually configured in the filter circuit. In the embodiment of the present application, the sidetone gain module 1221 may also be a low-pass and down-sampling filter module, or a gain module before, after, or in the sidetone filter 122, that is, the sidetone gain module 1221 is arranged between the sidetone microphone 121 and the filter circuit, or the sidetone gain module 1221 is arranged between any two filter circuits.

[0079] In one embodiment, the sidetone microphone 121, the sidetone gain module 1221 and each filter circuit are in working state. In this embodiment, opening the sidetone channel 120 includes: adjusting the gain of the sidetone gain module 1221 to be greater than or equal to a first preset value. Conversely, closing the sidetone channel 120 includes: adjusting the gain of the sidetone gain module 1221 to be less than or equal to a second preset value, or making the gain of the sidetone gain module 0. The second preset value is less than the first preset value.

[0080] In this embodiment, the gain of the sidetone gain module 1221 is made greater than or equal to the first preset value, so that the sidetone channel 120 can normally output the sidetone signal, and the gain of the sidetone gain module 1221 is made less than or equal to the second preset value or 0, so that the sidetone gain module 1221 cannot output the sidetone signal that can be used to offset the sidetone component, thereby also realizing the opening and closing of the sidetone channel 120.

[0081] In this embodiment, the sidetone channel 120 is turned on and off by adjusting the gain. During the off period, the components of the sidetone channel 120 are still in working state. Therefore, there is no need to wait for the components to be turned on, thereby reducing the delay of turning on the sidetone channel 120.

[0082] In some embodiments of the present application, the sidetone microphone 121 is an external ear microphone, which may collect sounds other than the wearer's voice. Therefore, in an embodiment of the present application, multiple sidetone microphones 121 can be set on the active noise reduction headset 100, and before S110, the sidetone sub-signals collected by each sidetone microphone 121 are obtained respectively; and beamforming processing is performed on each sidetone sub-signal in a preset direction to obtain a sidetone signal.

[0083] In this embodiment, the preset direction is used to represent the direction in which the active noise reduction headset 100 points to the wearer's mouth. Beamforming processing based on the preset direction can make the sidetone signal more representative of the wearer's voice, thereby improving the accuracy of the sidetone signal, and then improving the accuracy of canceling the sidetone component based on the sidetone signal. In addition, the sidetone channel 120 can provide the wearer with more accurate sidetone, reducing the possibility of the wearer hearing noise.

[0084] In some embodiments of the present application, the preset direction may be an angle range representing the wearer's mouth obtained through experiments on a large number of volunteers, or may be a direction / angle configured by the wearer according to his or her own circumstances, and is not limited here.

[0085] As mentioned above, the sound collected by the external microphone may contain other noises. Therefore, in other embodiments of the present application, a bandpass filter (implemented by a circuit) may be provided in the sidetone channel 120 or the sidetone filter 122. The bandwidth range of the bandpass filter is a preset bandwidth range, and the preset bandwidth range represents the bandwidth range in which the human voice is located. The sidetone signal is filtered based on the bandpass filter, thereby retaining the sound in the sidetone signal within the preset bandwidth range. In this way, the human voice in the sidetone signal can be retained, and other noises can be filtered out, thereby improving the accuracy of the sidetone signal.

[0086] The preset bandwidth range can use the well-known human voice range, such as 300Hz to 3.5KHz, or it can be obtained by volunteer testing, or it can be configured by the wearer according to his or her own speaking habits (such as the difference between male and female voices, the different voice ranges of people with different habits, etc.), and there is no limitation here.

[0087] S120: When the feedback noise reduction channel receives the sidetone signal, control the feedback noise reduction channel to cancel the sidetone component in the feedback audio signal based on the sidetone signal.

[0088] In this embodiment, the sidetone signal can be output to the feedback noise reduction channel 130 through the audio echo module 161 to offset the sidetone component in the feedback audio signal collected by the feedback noise reduction channel 130. As a result, the feedback audio signal output by the feedback noise reduction channel 130 will not include the sidetone component, and the sidetone cannot be offset and reduced, so that the wearer can hear the sidetone.

[0089] The above process can refer to the working principles of the existing audio echo module 161 and the feedback noise reduction channel 130, which will not be elaborated here.

[0090] S130: Control the feedback noise reduction filter to filter the feedback audio signal that cancels the sidetone component.

[0091] In this embodiment, the feedback noise reduction filter 132 can output the feedback audio signal that cancels the sidetone component to the DAC 170 after being processed by the feedback filter, and the DAC 170 converts the signal into an analog signal, which is finally output and played by the speaker 150, thereby canceling the noise and achieving feedback noise reduction. For details, please refer to the principle of feedback noise reduction, which will not be elaborated here.

[0092] As mentioned above, the playback of the sidetone requires a low delay so that the wearer can hear the voice of the wearer and the sidetone synchronously, otherwise an echo may occur. Therefore, in the embodiment of the present application, the sidetone channel 120 (including the sidetone filter 122) is implemented by hardware, and the sidetone filter 122 includes a FIR adaptive filter 1223, which needs to adjust its own filter coefficient according to the environment, so as to provide a noise reduction effect adapted to the environment.

[0093] In an embodiment of the present application, the processor 110 can also calculate the filter coefficients that need to be iteratively updated for the FIR adaptive filter 1223 and update the FIR adaptive filter 1223. The process may include: obtaining a sidetone signal; converting the sidetone signal into a frequency domain signal; performing speech noise reduction on the sidetone signal in the frequency domain to obtain a frequency domain gain vector; converting the frequency domain gain vector in the time domain to obtain a time domain gain filter; the time domain gain filter includes filter coefficients for updating the filter of the FIR adaptive filter 1223 circuit; mapping the filter coefficients of the time domain gain filter to the filter coefficients of the FIR adaptive filter 1223 to update the filter coefficients of the FIR adaptive filter 1223.

[0094] like Figure 5 As shown, in this embodiment, on the one hand, the sidetone channel 120 is implemented by a hardware circuit, which effectively reduces the time delay of sidetone signal acquisition, processing and emission, and improves the synchronization with the user's voice. On the other hand, the processor 110 calculates the updated filter coefficients of the FIR adaptive filter 1223, and the FIR adaptive filter 1223 itself does not need to calculate the filter coefficients, which makes the noise reduction function and the filter coefficient update function of the FIR adaptive filter 1223 independent of each other, and avoids the power consumption caused by the mutual influence when the two are executed by the same device.

[0095] In this embodiment, the acquired sidetone signal may be the sidetone signal output by the low-pass and down-sampling filter 164 of the sidetone channel 120 (or the sidetone signal input to the sidetone filter 122). The sidetone signal is a time domain signal and may be converted into a frequency domain signal using Fourier transform or short-time Fourier transform.

[0096] The method of performing speech noise reduction on the sidetone signal in the frequency domain can refer to the existing technology, such as spectrum subtraction, Wiener filtering, minimum mean square error short-time spectrum amplitude estimation, adaptive noise elimination, deep learning methods, etc., which will not be expanded here. During speech noise reduction, the frequency domain signal (the sidetone signal in the frequency domain) is multiplied by the frequency domain gain vector to obtain a noise reduction result. The frequency domain gain vector is calculated by the frequency domain signal and the various methods provided above, that is, the frequency domain gain vector is an intermediate result of the speech noise reduction process, not the noise reduction result. In an embodiment of the present application, the frequency domain gain vector can be extracted by speech noise reduction to convert it to the time domain to obtain a time domain gain filter, and the time domain gain vector includes the filter coefficients that the FIR adaptive filter 1223 needs to update.

[0097] Next, the frequency domain gain vector is transformed in the time domain to convert it into time domain data, such as using an inverse Fourier transform. In the present application, the time domain gain vector can be called a time domain gain filter, which also includes filter coefficients for updating the filter of the FIR adaptive filter 1223 circuit.

[0098] Finally, the filter coefficients of the time domain gain filter are mapped to the FIR adaptive filter 1223 , and the corresponding filter coefficients in the FIR adaptive filter 1223 are adjusted to update the FIR adaptive filter 1223 .

[0099] In some embodiments of the present application, after obtaining the sidetone signal, the sidetone signal may be low-passed and down-sampled to reduce the sampling rate, thereby improving the efficiency of calculating the updated filter coefficients. Accordingly, after obtaining the time domain gain filter, the time domain gain filter is up-sampled so that the sampling rate of the time domain gain filter matches the FIR adaptive filter 1223.

[0100] Based on the same inventive concept, the embodiment of the present application further provides a sidetone-based noise reduction device, which can be applied to the processor 110 of the aforementioned active noise reduction headset 100.

[0101] See also Figure 6 , Figure 6 Schematic diagram of a sidetone-based noise reduction device according to an embodiment of the present application. The sidetone-based noise reduction device includes: a sidetone channel control module 510, a feedback noise reduction channel control module 520 and a noise reduction module.

[0102] The sidetone channel control module 510 is configured to control the sidetone channel 120 to output the collected sidetone signal to the feedback noise reduction channel 130 in response to the sidetone channel 120 being turned on.

[0103] The feedback noise reduction channel control module 520 is configured to control the feedback noise reduction channel 130 to cancel the sidetone component in the feedback audio signal based on the sidetone signal when the feedback noise reduction channel 130 receives the sidetone signal.

[0104] The noise reduction control module 530 is used to control the feedback noise reduction filter 132 to filter the feedback audio signal that cancels the sidetone component.

[0105] In one embodiment, the sidetone channel control module 510 is used to detect whether the wearer of the active noise reduction headset 100 is in a target state, where the target state indicates that the wearer is in a call and / or speaking; if the wearer is in the target state, the sidetone channel 120 is turned on.

[0106] In one embodiment, the sidetone channel control module 510 is configured to close the sidetone channel 120 if the wearer is not in the target state.

[0107] In one embodiment, the sidetone channel 120 further includes multiple functional modules, a sidetone gain module 1221 and multiple filter circuits; the sidetone microphone 121 and the multiple functional modules are connected in series; the sidetone channel control module 510 is specifically used to put the sidetone microphone 121 and each of the multiple functional modules in an operating state to turn on the sidetone channel 120, and to turn off the sidetone microphone 121 and at least one of the multiple functional modules to turn off the sidetone channel 120.

[0108] In one embodiment, the sidetone channel 120 further includes a sidetone gain module 1221 and a plurality of filter circuits; the sidetone microphone 121, the sidetone gain module 1221 and the plurality of filter circuits are connected in series, and the sidetone gain module 1221 is disposed between the sidetone microphone 121 and the filter circuit, or the sidetone gain module 1221 is disposed between any two of the filter circuits; the sidetone microphone 121, the sidetone gain module 1221 and the filter circuits are all in working state. The sidetone channel control module 510 is specifically used to adjust the gain of the sidetone gain module 1221 to be greater than or equal to a first preset value to open the sidetone channel 120, and to adjust the gain of the sidetone gain module 1221 to be less than or equal to a second preset value, or to make the gain of the sidetone gain module 1221 0 to close the sidetone channel 120; the second preset value is less than the first preset value.

[0109] In one embodiment, the active noise reduction headset 100 further includes an in-ear microphone, and the sidetone channel control module 510 is used to obtain an in-ear audio signal collected by the in-ear microphone; perform wearer voice activity detection based on the difference between the in-ear audio signal and the sidetone signal; and determine whether the wearer is in the target state of speaking based on the detection result of the wearer's voice activity detection.

[0110] In one embodiment, the active noise reduction headset 100 is communicatively connected to a smart device; the sidetone channel control module 510 is used to receive data and / or instructions transmitted by the smart device; and determine whether it is in the target state of the call process based on the data and / or instructions transmitted by the smart device.

[0111] In one embodiment, the sidetone microphones 121 include multiple ones, and the sidetone channel control module 510 is used to respectively obtain the sidetone sub-signals collected by each of the sidetone microphones 121; perform beamforming processing in a preset direction on each of the sidetone sub-signals to obtain the sidetone signal, and the preset direction is used to represent the direction in which the active noise reduction headset 100 points to the wearer's mouth.

[0112] In one embodiment, the sidetone channel 120 includes a sidetone filter 122, and the sidetone filter 122 includes a FIR adaptive filter 1223. The sidetone channel control module 510 is also used to obtain the sidetone signal; convert the sidetone signal into a frequency domain signal; perform speech noise reduction on the sidetone signal in the frequency domain to obtain a frequency domain gain vector; convert the frequency domain gain vector in the time domain to obtain a time domain gain filter; the time domain gain filter includes a filter coefficient for updating the filter of the FIR adaptive filter 1223 circuit; and map the filter coefficient of the time domain gain filter to the filter coefficient of the FIR adaptive filter 1223 to update the filter coefficient of the FIR adaptive filter 1223.

[0113] Based on the same inventive concept, an embodiment of the present application further provides a readable storage medium, in which instructions are stored. The instructions can be executed by one or more processors 110 to implement the sidetone-based noise reduction method provided in the above embodiment.

[0114] The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD (digital video disc)), or a semiconductor medium (e.g., an SSD (Solid State Disk)).

[0115] If the sidetone-based noise reduction method is implemented in the form of a software function module 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 can essentially or partly contribute to the prior art or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0116] In the embodiments provided in the present application, it should be understood that the disclosed method can also be implemented in other ways. The functional modules in the various embodiments of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

[0117] The above embodiments can be freely combined without conflict, and the embodiments obtained by the combination are included in the protection scope of this application.

[0118] The above detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0119] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0120] In the description of this application, it is also necessary to explain that, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0121] The above description is only an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of protection of the present application.

Claims

1. A noise reduction method based on sidetone, characterized in that: A processor applied to an active noise reduction headset, wherein the active noise reduction headset includes a sidetone channel and a feedback noise reduction channel, wherein the sidetone channel includes a sidetone microphone for collecting a sidetone signal; the feedback noise reduction channel includes a feedback microphone and a feedback noise reduction filter, wherein the feedback microphone is used to collect a feedback audio signal, and the feedback noise reduction filter is used to filter the feedback audio signal; The sidetone-based noise reduction method comprises: In response to the sidetone channel being turned on, controlling the sidetone channel to output the collected sidetone signal to the feedback noise reduction channel; When the feedback noise reduction channel receives the sidetone signal, controlling the feedback noise reduction channel to cancel the sidetone component in the feedback audio signal based on the sidetone signal; The feedback noise reduction filter is controlled to filter the feedback audio signal that cancels the sidetone component.

2. The sidetone-based noise reduction method according to claim 1, characterized in that: In response to the sidetone channel being turned on, controlling the sidetone channel to output the collected sidetone signal to the feedback noise reduction channel, the method further includes: Detecting whether the wearer of the active noise reduction headset is in a target state, wherein the target state indicates that the wearer is in a call and / or is speaking; If the wearer is in the target state, the sidetone channel is turned on.

3. The sidetone-based noise reduction method according to claim 2, characterized in that: After detecting whether the wearer of the active noise reduction headset is in the target state, the method further includes: If the wearer is not in the target state, the sidetone channel is turned off.

4. The sidetone-based noise reduction method according to claim 3, characterized in that: The sidetone channel also includes a plurality of functional modules, a sidetone gain module and a plurality of filter circuits; The sidetone microphone and the plurality of functional modules are connected in series; The opening of the sidetone channel comprises: making the sidetone microphone and each of the plurality of functional modules in a working state; The closing of the sidetone channel includes: closing the sidetone microphone and at least one of the plurality of functional modules.

5. The sidetone-based noise reduction method according to claim 3, characterized in that: The sidetone channel further includes a sidetone gain module and a plurality of filter circuits; the sidetone microphone, the sidetone gain module and the plurality of filter circuits are connected in series, and the sidetone gain module is arranged between the sidetone microphone and the filter circuit, or the sidetone gain module is arranged between any two of the filter circuits; The sidetone microphone, the sidetone gain module and each filter circuit are all in working state; The opening of the sidetone channel includes: adjusting the gain of the sidetone gain module to be greater than or equal to a first preset value; Closing the sidetone channel includes: adjusting the gain of the sidetone gain module to be less than or equal to a second preset value, or making the gain of the sidetone gain module 0; the second preset value is less than the first preset value.

6. The sidetone-based noise reduction method according to claim 2, characterized in that: The active noise reduction headset further includes an in-ear microphone, and the detecting whether the wearer of the active noise reduction headset is in a target state includes: Acquire an in-ear audio signal collected by the in-ear microphone; Performing wearer voice activity detection based on the difference between the in-ear audio signal and the sidetone signal; It is determined whether the wearer is in a target state of speaking according to a detection result of the wearer's voice activity detection.

7. The sidetone-based noise reduction method according to claim 2, characterized in that: The active noise reduction headset is communicatively connected with a smart device; The detecting whether the wearer of the active noise reduction headset is in a target state includes: Receiving data and / or instructions transmitted by the smart device; Based on the data and / or instructions transmitted by the smart device, it is determined whether the call process is in the target state.

8. The sidetone-based noise reduction method according to claim 1, characterized in that: The sidetone microphones include a plurality of sidetone microphones, and before the sidetone channel is controlled to output the collected sidetone signal to the feedback noise reduction channel, the method further includes: Respectively acquiring the sidetone sub-signals collected by the sidetone microphones; A beamforming process is performed on each of the sidetone sub-signals in a preset direction to obtain the sidetone signal, wherein the preset direction is used to represent the direction in which the active noise reduction headset points to the wearer's mouth.

9. The sidetone-based noise reduction method according to claim 1, characterized in that: The sidetone channel further includes a sidetone filter, the sidetone filter includes a bandpass filter, the bandwidth range of the bandpass filter is a preset bandwidth range, and the preset bandwidth range represents the bandwidth range of human voice; The sidetone signal is a signal filtered by the bandpass filter.

10. The sidetone-based noise reduction method according to any one of claims 1 to 9, characterized in that: The sidetone channel includes a sidetone filter, the sidetone filter includes a FIR adaptive filter, and the method further includes: Acquiring the sidetone signal; Converting the sidetone signal into a frequency domain signal; Perform speech noise reduction on the sidetone signal in the frequency domain to obtain a frequency domain gain vector; Converting the frequency domain gain vector in the time domain to obtain a time domain gain filter; the time domain gain filter includes filter coefficients for updating the FIR adaptive filter; The filter coefficients of the time domain gain filter are mapped to the filter coefficients of the FIR adaptive filter to update the filter coefficients of the FIR adaptive filter.

11. An active noise reduction headset, characterized in that: include: A sidetone channel, including a sidetone microphone for collecting a sidetone signal; A feedback noise reduction channel, comprising a feedback microphone and a feedback noise reduction filter, wherein the feedback microphone is used to collect a feedback audio signal, and the feedback noise reduction filter is used to filter the feedback audio signal; A processor, configured to execute the sidetone-based noise reduction method according to any one of claims 1 to 10.

12. A readable storage medium, characterized in that: The readable storage medium stores instructions, which can be executed by a processor to implement the sidetone-based noise reduction method according to any one of claims 1 to 10.

Images