Howling suppression method, sound pickup terminal, and howling suppression system

Abstract

The application provides a howling suppression method, a sound pickup terminal and a howling suppression system, which are applied to the sound pickup terminal. The method comprises the following steps: monitoring the distance between the sound pickup terminal and a collection terminal in real time; if the distance is less than a preset distance threshold, performing howling suppression according to a first howling suppression strategy; if the distance is greater than or equal to the preset distance threshold, performing noise floor detection and howling detection to obtain a noise floor detection result and a howling detection result, performing howling suppression according to a second howling suppression strategy according to the noise floor detection result, and / or performing howling suppression according to a third howling suppression strategy according to the howling detection result, so as to improve the howling suppression effect and reduce the timbre distortion degree.

Description

Technical Field

[0001] This application relates to the field of howling processing technology, and more specifically, to a howling suppression method, a pickup terminal, and a howling suppression system. Background Technology

[0002] In multi-person speaking scenarios, distributed sound pickup systems are typically used to collect and amplify voice data in order to improve communication efficiency. A distributed sound pickup system is configured with multiple wireless sound pickup terminals and a collection terminal. Each sound pickup terminal transmits the collected voice data to the collection terminal through a short-range wireless communication link, and then the collection terminal connects to a speaker to play the data in real time, achieving full sound coverage.

[0003] However, in distributed sound pickup systems, the sound emitted by the loudspeakers may be recaptured by nearby pickup terminals and transmitted back to the collection terminal for playback, creating a feedback path that can easily cause howling. Currently, frequency domain notch filtering, frequency shifting algorithms, or adaptive filtering methods can be used to suppress howling.

[0004] However, using frequency domain notch filtering or frequency shifting algorithms to suppress howling can easily lead to timbre distortion and reduce speech fidelity. Adaptive filtering methods for howling suppression suffer from difficulties in convergence, unstable performance, and high computational complexity. Summary of the Invention

[0005] The purpose of this application is to address the shortcomings of the prior art by providing a feedback suppression method, a pickup terminal, and a feedback suppression system, so as to solve the practical problems that feedback suppression in the prior art easily leads to timbre distortion, reduced speech fidelity, difficulty in convergence, and high computational complexity.

[0006] To achieve the above objectives, the technical solutions adopted in the embodiments of this application are as follows: In a first aspect, embodiments of this application provide a howling suppression method applied to a microphone terminal, the method comprising: Real-time monitoring of the distance between the pickup terminal and the collection terminal; If the distance is less than a preset distance threshold, then howling suppression is performed according to the first howling suppression strategy; If the distance is greater than or equal to the preset distance threshold, then noise floor detection and howling detection are performed to obtain noise floor detection results and howling detection results. Howling is suppressed according to the noise floor detection results and the second howling suppression strategy, and / or howling is suppressed according to the howling detection results and the third howling suppression strategy.

[0007] As an optional implementation, the squealing suppression according to the first squealing suppression strategy includes: The operating state is switched from the sound pickup state to the mute state to stop sound pickup, and when the distance between the sound pickup terminal and the collection terminal is detected to change to be greater than or equal to the preset distance threshold, the operating state is switched from the mute state to the sound pickup state to resume sound pickup.

[0008] As an optional implementation, the step of performing noise floor detection and howling detection to obtain noise floor detection results and howling detection results includes: Real-time acquisition of audio stream data over a preset time period, wherein the audio stream data includes multiple audio frames; Multiple non-voice audio frames are extracted from the audio stream data, and background noise is detected based on each non-voice audio frame to obtain the background noise detection result. The background noise detection result includes the root mean square value of background noise of the sound pickup environment where the sound pickup terminal is located during the preset time period. Feedback detection is performed based on the audio stream data to obtain the feedback detection result, which includes the number of feedback frames of the pickup terminal in the preset time period.

[0009] As an optional implementation, the step of suppressing howling based on the background noise detection result and according to the second howling suppression strategy includes: Determine whether the root mean square value of the noise floor is greater than a preset noise floor value; If so, howling suppression is performed based on the aforementioned root mean square noise floor value.

[0010] As an optional implementation, the squealing suppression based on the root mean square value of the noise floor includes: When the root mean square value of the noise floor is greater than the preset noise floor value, the pickup gain of the pickup terminal is iteratively reduced according to the preset gain adjustment step size to narrow the pickup range of the pickup terminal.

[0011] As an optional implementation, the step of suppressing howling based on the howling detection result and according to the third howling suppression strategy includes: Determine whether the audio stream data contains howling frames; If so, then howling suppression is performed based on the number of howling frames.

[0012] As an optional implementation, the howling suppression based on the number of howling frames includes: The howling level is determined based on the number of howling frames. Suppress the howling based on the howling level.

[0013] As an optional implementation, the feedback suppression based on the feedback level includes: If the howling level is the first level, the preset frame will be muted, and the sound pickup will be restored after the preset frame is muted; If the feedback level is the second level, then continuous non-voice audio frames are extracted from the audio stream data, and segmented amplitude compression processing is performed on the continuous non-voice audio frames so that there are continuous zero signal segments in the processed continuous non-voice audio frames, wherein the feedback severity of the second level is lower than that of the first level.

[0014] Secondly, embodiments of this application provide a microphone terminal, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the microphone terminal is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the howling suppression method described in the first aspect above.

[0015] Thirdly, embodiments of this application provide a howling suppression system, including the pickup terminal described in the first aspect above and a collection terminal communicatively connected to the pickup terminal, wherein the collection terminal is externally connected to a speaker.

[0016] The beneficial effects of this application are: This application provides a feedback suppression method, a pickup terminal, and a feedback suppression system. The system monitors the distance between the pickup terminal and the aggregation terminal in real time. If the distance is less than a preset distance threshold, the pickup terminal performs feedback suppression according to a first feedback suppression strategy. If the distance is greater than or equal to the preset distance threshold, it performs noise floor detection and feedback detection, obtaining noise floor detection results and feedback detection results. Feedback suppression is then performed based on the noise floor detection results and according to a second feedback suppression strategy, and / or, based on the feedback detection results and according to a third feedback suppression strategy. When the distance between the pickup terminal and the external speaker connected to the aggregation terminal is too close, feedback suppression is performed according to the first feedback suppression strategy to prevent the sound emitted by the speaker from being picked up by the pickup terminal again, quickly cutting off the audio feedback loop. Feedback suppression based on noise floor detection results and / or feedback detection results can end feedback without the user's awareness, while ensuring voice fidelity and avoiding timbre distortion. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1This is a schematic diagram of the architecture of the howling suppression system provided in the embodiments of this application; Figure 2 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 1 ; Figure 3 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 2 ; Figure 4 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 3 ; Figure 5 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 4 ; Figure 6 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 5 ; Figure 7 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 6 ; Figure 8 This is a schematic diagram of the structure of the microphone terminal provided in an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0020] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0021] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0022] In multi-person speaking scenarios, a distributed audio pickup system is configured with multiple wireless pickup terminals and a collection terminal. Each pickup terminal transmits the collected voice data to the collection terminal via a short-range wireless communication link, and the collection terminal then connects to a speaker for real-time playback, achieving full-field sound coverage. However, the sound played by the speaker may be recaptured by a nearby pickup terminal and transmitted back to the collection terminal, creating a feedback path that can easily cause feedback. Currently, frequency domain notch filtering, frequency shifting algorithms, or adaptive filtering methods can be used to suppress feedback. However, using frequency domain notch filtering or frequency shifting algorithms to suppress feedback can easily lead to timbre distortion and reduce voice fidelity. Adaptive filtering methods for suppressing feedback suffer from difficulties in convergence, unstable performance, and high computational complexity.

[0023] Based on the above-mentioned problems, this application provides a feedback suppression method to improve feedback suppression effect, reduce timbre distortion, and ensure voice fidelity.

[0024] Figure 1 This is a schematic diagram of the architecture of the howling suppression system provided in the embodiments of this application, as shown below. Figure 1 As shown, the feedback suppression system includes a pickup terminal and a collection terminal that is communicatively connected to the pickup terminal. The collection terminal is externally connected to a speaker.

[0025] Optionally, refer to Figure 1 In multi-person speaking scenarios, the howling suppression system can include There are one microphone terminal and one aggregation terminal, wherein microphone terminal 1, microphone terminal 2, microphone terminal 3, ..., microphone terminal All devices are connected to the aggregation terminal, which has an external speaker for real-time audio playback. To facilitate easy movement of the pickup terminals, communication between them and the aggregation terminal can be achieved via short-range wireless communication methods such as Bluetooth.

[0026] Reference Figure 1 During the speaker's speech, the microphone (such as...) Figure 1 1) The microphone terminal near the speaker transmits the collected voice to the aggregation terminal via a short-range wireless communication link. The aggregation terminal amplifies and plays the sound through a speaker, allowing other participants to hear the speaker's content. Due to local speaker noise, the sound played by the speaker will be collected again by the microphone terminal and transmitted to the aggregation terminal, forming a feedback path that is prone to howling.

[0027] The microphone terminal monitors the distance between itself and the collection terminal. When the microphone terminal gets too close to the collection terminal, it mutes itself, stops picking up sound, and immediately cuts off the feedback path of the howling, thus suppressing the howling.

[0028] When the distance between the pickup terminal and the receiving terminal is appropriate, the pickup terminal performs background noise detection and feedback detection. The background noise detection result reflects the noise level of the environment, and the feedback detection result reflects the severity of feedback. If the background noise detection result indicates a high level of noise in the environment, and feedback is more likely to occur in noisy environments, the pickup terminal adjusts its pickup gain. By iteratively reducing the gain, the pickup range of the pickup terminal is gradually narrowed, so that the sound played by the speaker cannot be picked up by the pickup terminal and thus cannot be transmitted to the receiving terminal, thereby disrupting the feedback path and preventing feedback from occurring.

[0029] If the feedback detection results indicate that the feedback is severe or frequent, the microphone terminal will mute for a few frames, such as 2-3 frames, to quickly cut off the feedback path of the feedback and end the feedback. After the feedback ends, the microphone terminal 1 will resume microphone pickup to avoid affecting voice acquisition.

[0030] If the feedback detection results show that feedback exists but the feedback is not severe, the pickup terminal will gradually compress the amplitude of the audio signal in the silent segment (i.e., non-voice frame) to zero, thereby eliminating the feedback by breaking the conditions for feedback generation through zeroing.

[0031] Because the signal amplitude is processed in the silent segment to obtain the continuous zero segment, it does not affect the normal acquisition and playback of the voice. The feedback ends without the user (speaker and participants) not noticing, and it does not reduce the fidelity of the voice, avoid timbre distortion, and ensure that other participants can hear the speaker's speech accurately and clearly.

[0032] Figure 2 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 1 The execution subject of this method is Figure 1 Any microphone terminal in the system. For example... Figure 2 As shown, the method includes: S101. Real-time monitoring of the distance between the pickup terminal and the aggregation terminal.

[0033] Optionally, with Figure 1Taking a microphone terminal 1, which is relatively close to the speaker, as an example, the microphone terminal 1 can be used as the main body for implementing the feedback suppression method. The microphone terminal collects the distance between itself and the receiving terminal in real time. Since the receiving terminal is connected to an external speaker, the microphone terminal obtains the distance between itself and the external speaker of the receiving terminal in real time. This distance affects whether the sound emitted by the speaker can be collected by the microphone terminal again. That is, this distance is a precondition for feedback. The closer the microphone terminal is to the external speaker of the receiving terminal, the easier it is for an audio feedback loop to form, and the higher the risk of feedback.

[0034] Specifically, if the pickup terminal and the aggregation terminal communicate via Bluetooth, the pickup terminal can use the Bluetooth Received Signal Strength Indicator (RSSI) positioning function to obtain the distance between the pickup terminal and the external speaker of the aggregation terminal in real time.

[0035] S102. If the distance is less than the preset distance threshold, then the howling suppression is performed according to the first howling suppression strategy.

[0036] Optionally, if the distance is less than a preset distance threshold, it indicates that the distance between the pickup terminal and the external speaker connected to the aggregation terminal is too close, and the sound emitted by the speaker can easily be picked up again by the pickup terminal, which can easily form an audio feedback loop and generate howling. In this case, the pickup terminal will suppress howling according to the first howling suppression strategy. The preset distance threshold is a pre-set distance that ensures that the sound emitted by the speaker cannot be picked up again by the pickup terminal, such as 1 meter.

[0037] Since the probability of feedback is extremely high when the distance between the pickup terminal and the external speaker of the aggregation terminal is too close, in order to suppress feedback, the first feedback suppression strategy is the fastest and most powerful strategy among feedback suppression strategies, so as to quickly cut off the audio feedback loop and prevent feedback from spreading.

[0038] S103. If the distance is greater than or equal to the preset distance threshold, then perform background noise detection and howling detection to obtain background noise detection results and howling detection results. Perform howling suppression according to the background noise detection results and the second howling suppression strategy, and / or perform howling suppression according to the howling detection results and the third howling suppression strategy.

[0039] Optionally, if the distance is greater than or equal to a preset distance threshold, it indicates that feedback will not occur due to the close proximity of the microphone terminal and the external speaker of the aggregation terminal, but feedback may occur due to a noisy environment. In this case, the microphone terminal performs background noise detection to detect the basic noise of the surrounding environment and obtain a background noise detection result. It also performs feedback detection to identify whether feedback exists and its severity, and obtains a feedback detection result. The background noise detection result reflects the noise level of the environment, and the feedback detection result reflects whether feedback occurs and its severity.

[0040] When the microphone terminal determines, based on the background noise detection results, that the ambient noise level is too high, it employs a second feedback suppression strategy to suppress feedback caused by the noisy environment. If the microphone terminal determines that feedback exists based on feedback detection results, it employs a third feedback suppression strategy to suppress feedback.

[0041] In this embodiment, the distance between the pickup terminal and the aggregation terminal is monitored in real time. If the distance is less than a preset distance threshold, the pickup terminal performs feedback suppression according to the first feedback suppression strategy. If the distance is greater than or equal to the preset distance threshold, noise floor detection and feedback detection are performed to obtain noise floor detection results and feedback detection results. Feedback suppression is then performed according to the second feedback suppression strategy based on the noise floor detection results, and / or according to the third feedback suppression strategy based on the feedback detection results. When the distance between the pickup terminal and the external speaker connected to the aggregation terminal is too close, feedback suppression is performed according to the first feedback suppression strategy to prevent the sound emitted by the speaker from being picked up by the pickup terminal again, quickly cutting off the audio feedback loop. Feedback suppression based on noise floor detection results and / or feedback detection results can end feedback without the user's awareness, while ensuring voice fidelity and avoiding timbre distortion.

[0042] As an optional implementation, the above step S102, which involves suppressing howling according to the first howling suppression strategy, includes: Switch the operating state from sound pickup to mute to stop sound pickup, and when the distance between the sound pickup terminal and the collection terminal changes to be greater than or equal to a preset distance threshold, switch the operating state from mute to sound pickup to resume sound pickup.

[0043] Optionally, when the pickup terminal detects that the distance to the external speaker of the aggregation terminal is too close, in order to avoid feedback, it switches its working state from pickup state to mute state, closes the pickup channel, and stops collecting audio signals, that is, stops pickup from the source, which can quickly cut off the feedback path of feedback.

[0044] The microphone terminal can also control its own indicator lights synchronously. For example, when the working state is switched to mute mode, the indicator light can be kept on red to indicate to the speaker that the microphone terminal is in a mute state and cannot pick up sound. This allows the speaker to move the microphone terminal based on the indicator light prompt, or use another microphone terminal that is far away and can pick up sound normally to speak.

[0045] During the muted state, the pickup terminal monitors the distance between itself and the external speaker of the aggregation terminal in real time. If the distance between the pickup terminal and the external speaker of the aggregation terminal changes from less than a preset distance threshold to greater than or equal to the preset distance threshold, that is, the distance between the pickup terminal and the external speaker of the aggregation terminal increases and the risk of feedback is extremely low, then it switches its working state from the muted state to the pickup state, resumes audio acquisition, and at the same time controls the indicator light to switch from a solid red light to a solid green light to indicate to the speaker that the pickup terminal has resumed pickup and can be used normally.

[0046] In this embodiment, when the pickup terminal detects that the distance to the collection terminal is less than a preset distance threshold, it switches its operating state from pickup mode to mute mode to stop pickup. When the distance between the pickup terminal and the collection terminal changes to be greater than or equal to the preset distance threshold, it switches its operating state from mute mode back to pickup mode to resume pickup. By switching to mute mode, the feedback path of howling is quickly cut off, improving howling suppression efficiency.

[0047] Figure 3 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 2 ,like Figure 3 As shown, in step S103 above, noise floor detection and howling detection are performed to obtain noise floor detection results and howling detection results, including: S201. Real-time acquisition of audio stream data for a preset time period, the audio stream data includes multiple audio frames.

[0048] Optionally, the microphone terminal collects continuous audio signals over a preset time period in real time to form audio stream data. For example, the preset time period can be three minutes.

[0049] The audio pickup terminal divides the audio stream data into multiple independent audio frames according to a fixed frame length (e.g., 20ms / frame). The audio frame can be used as the smallest unit for the audio pickup terminal to process the audio stream data.

[0050] S202. Extract multiple non-voice audio frames from the audio stream data, and perform background noise detection based on each non-voice audio frame to obtain background noise detection results. The background noise detection results include the root mean square value of background noise in the sound pickup environment where the sound pickup terminal is located during a preset time period.

[0051] Optionally, since background noise is the basic environmental noise, speech interference must be excluded to reflect the true noise level of the environment. The pickup terminal uses Voice Activity Detection (VAD) to extract multiple non-speech audio frames from the audio stream data, excluding the speech audio frames when the speaker is speaking, thus avoiding interference from speech audio frames with background noise detection.

[0052] The microphone terminal calculates the root mean square value of all non-voice audio frames within a preset time period, which is the root mean square value of the noise floor of the microphone environment in which the microphone terminal is located during the preset time period. The root mean square value of the noise floor of the microphone environment in which the microphone terminal is located during the preset time period is used as the noise floor detection result. The noise floor of the microphone environment in which the microphone terminal is located is quantified through the noise floor detection result, so as to accurately reflect the noise level of the on-site microphone environment.

[0053] S203. Perform feedback detection based on audio stream data to obtain feedback detection results, which include the number of feedback frames of the pickup terminal in a preset time period.

[0054] Optionally, since feedback may occur in either the speech or non-speech stages, full-frame detection is required to avoid missed detection. Therefore, the audio pickup terminal performs feedback detection on all audio frames within a preset time period. By analyzing the frequency flatness and peak-to-average power ratio (PAPR) of each audio frame, it identifies whether each audio frame is a feedback frame. Specifically, the frequency flatness of feedback frames is significantly lower than that of non-feedback frames, and the PAPR of feedback frames is significantly higher than that of non-feedback frames.

[0055] The microphone terminal counts the number of howling frames within a preset time period. The number of howling frames is a key indicator for quantifying the frequency of howling; the more howling frames there are, the more frequent and severe the howling. The number of howling frames from the microphone terminal within the preset time period is used as the howling detection result to quantify the frequency / severity of howling, accurately reflecting whether the microphone terminal generates howling within the preset time period and the severity of howling when it does.

[0056] In this embodiment, the microphone terminal collects audio stream data in real time over a preset time period, extracts multiple non-voice audio frames from the audio stream data, and performs background noise detection on each non-voice audio frame to obtain a background noise detection result including the root mean square value of the background noise of the microphone environment where the microphone terminal is located during the preset time period. Feedback detection is performed based on the audio stream data to obtain a feedback detection result including the number of feedback frames generated by the microphone terminal during the preset time period. The background noise of the microphone environment where the microphone terminal is located is quantified through the background noise detection result, accurately reflecting the noise level of the on-site microphone environment. The feedback detection result accurately reflects whether the microphone terminal generates feedback during the preset time period and the severity of feedback when it occurs.

[0057] Figure 4A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 3 ,like Figure 4 As shown, in step S103 above, howling suppression is performed based on the noise floor detection result and according to the second howling suppression strategy, including: S301. Determine whether the root mean square value of the noise floor is greater than the preset noise floor value.

[0058] Optionally, the pickup terminal compares the root mean square value of the noise floor with a preset noise floor value to determine whether the root mean square value of the noise floor is greater than the preset noise floor value.

[0059] The preset noise floor value is a pre-defined boundary between quiet and noisy environments. For example, the preset noise floor value can be -48dBm.

[0060] S302. If so, then howling suppression is performed based on the root mean square value of the noise floor.

[0061] Optionally, if it is determined that the root mean square value of the noise floor is greater than the preset noise floor value, it indicates that the basic noise of the surrounding environment of the microphone terminal is too high, that is, the noise level of the microphone environment is too high, and it is easy to trigger howling due to the noisy microphone environment. In this case, the microphone terminal suppresses howling according to the root mean square value of the noise floor. At this time, the microphone terminal does not need to stop microphone pickup and suppresses howling without interrupting microphone pickup.

[0062] Since the microphone only suppresses howling when the root mean square value of the noise floor is greater than the preset noise floor value, when the root mean square value of the noise floor is less than or equal to the preset noise floor value, it means that the basic noise of the surrounding environment of the microphone is relatively small, that is, the microphone environment is relatively quiet and it is not easy to trigger howling. Therefore, the microphone does not need to suppress howling.

[0063] In this embodiment, the microphone terminal compares the root mean square (RMS) value of the noise floor with a preset noise floor value to determine whether the RMS value is greater than the preset noise floor value. If so, the microphone terminal performs feedback suppression based on the RMS value. Feedback can be suppressed without interrupting microphone pickup, achieving feedback suppression without the user's awareness.

[0064] As an optional implementation, the howling suppression based on the root mean square value of the noise floor in step S302 above includes: When the root mean square value of the noise floor is greater than the preset noise floor value, the pickup gain of the pickup terminal is iteratively reduced according to the preset gain adjustment step size in order to narrow the pickup range of the pickup terminal.

[0065] Optionally, when the root mean square value of the noise floor is greater than the preset noise floor value, the pickup terminal iteratively reduces the pickup gain of the pickup terminal according to the preset gain adjustment step size. For example, the preset gain adjustment step size is 0.5dB.

[0066] Since a higher pickup gain results in a larger pickup range, iteratively reducing the pickup gain of the microphone terminal can gradually narrow its pickup range, reducing the pickup of distant environmental noise and lowering the probability of feedback, thus preventing feedback in advance. The upper limit for pickup gain compression can be 6dB to avoid excessively reducing the pickup gain and causing the speaker's voice to be unable to be properly captured, thus preventing feedback while ensuring speech clarity.

[0067] After reducing the pickup gain, the pickup terminal continues to detect the background noise. If the root mean square value of the background noise within a preset time period is less than or equal to the preset background noise value, it indicates that the pickup environment has changed from noisy to quiet (e.g., the pickup environment was noisy when the attendees were seated, and became quiet after the attendees were seated). At this time, the probability of the environment triggering feedback decreases, and the pickup terminal can iteratively increase the current pickup gain of the pickup terminal according to the preset gain adjustment step size to gradually restore the pickup range of the pickup terminal.

[0068] In this embodiment, when the root mean square value of the noise floor is greater than a preset noise floor value, the pickup terminal iteratively reduces its pickup gain according to a preset gain adjustment step size to narrow the pickup range. This can reduce the pickup terminal's ability to pick up distant environmental noise, reduce the probability of feedback, prevent feedback from occurring in advance, and ensure pickup quality.

[0069] Figure 5 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 4 ,like Figure 5 As shown, step S103 above, which involves suppressing howling based on the howling detection results and according to the third howling suppression strategy, includes: S401. Determine if there are howling frames in the audio stream data.

[0070] Optionally, the pickup terminal determines whether there are howling frames in the audio stream data within the preset time period based on the number of howling frames in the howling detection result.

[0071] Specifically, if the number of howling frames in the preset time period in the howling detection result is not zero, it is determined that there are howling frames in the audio stream data of the preset time period, that is, the audio pickup terminal generated howling in the preset time period. At this time, howling suppression needs to be performed according to the third howling suppression strategy.

[0072] If the number of howling frames in the preset time period is zero in the howling detection result, it is determined that there are no howling frames in the audio stream data of the preset time period, that is, the pickup terminal did not generate howling in the preset time period. At this time, there is no need to suppress howling, and howling detection can continue for the next time period.

[0073] S402. If so, then suppress howling based on the number of howling frames.

[0074] Optionally, if it is determined that there are howling frames in the audio stream data within a preset time period, then howling suppression is performed in a differentiated manner according to the number of howling frames and the third howling suppression strategy.

[0075] Specifically, the severity / frequency of howling can be measured by the number of howling frames, and different levels of howling can be suppressed in a differentiated manner to dynamically adapt to the howling state.

[0076] In this embodiment, the audio pickup terminal determines whether there are howling frames in the audio stream data. If so, it performs differentiated howling suppression according to the number of howling frames and the third howling suppression strategy. This differentiates howling suppression for different degrees, dynamically adapts to howling states, and improves howling suppression effectiveness.

[0077] Figure 6 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 5 ,like Figure 6 As shown, the feedback suppression in step S402 above, based on the number of feedback frames, includes: S501. Determine the howling level based on the number of howling frames.

[0078] Optionally, the microphone terminal determines the howling level based on the number of howling frames and a preset howling frame number threshold, wherein the howling level includes a first level and a second level.

[0079] Specifically, if the number of howling frames is greater than the threshold for the number of howling frames, it indicates that howling occurs frequently and the howling intensity is high, and the howling level is determined to be the first level.

[0080] Correspondingly, if the number of howling frames is greater than zero and less than or equal to the threshold number of howling frames, it indicates that howling occurs occasionally and the howling intensity is low, and the howling level is determined to be the second level.

[0081] S502. Suppress howling based on howling level.

[0082] Optionally, the microphone terminal can perform differentiated feedback suppression based on the feedback level. Different methods can be used to suppress feedback for different feedback levels, thereby improving the feedback suppression effect.

[0083] In this embodiment, the microphone terminal determines the howling level based on the number of howling frames and performs howling suppression based on the howling level. Different methods are used to achieve differentiated howling suppression for different howling levels, thereby improving the howling suppression effect.

[0084] Figure 7 A flowchart illustrating the howling suppression method provided in this application embodiment. Figure 6 ,like Figure 7As shown, the feedback suppression in step S502 above, based on the feedback level, includes: S601. If the howling level is the first level, the preset frame will be muted, and the sound pickup will be restored after the preset frame is muted.

[0085] Optionally, if the feedback level is Level 1, indicating that the microphone terminal frequently experiences feedback within the current preset time period and the feedback intensity is high, the microphone terminal will mute for a preset frame to briefly disrupt the continuous feedback path required for feedback generation, quickly ending severe / frequent feedback. It will then resume microphone pickup after the preset frame is muted, avoiding prolonged silence from affecting the meeting.

[0086] For example, the preset frame can be 2 frames or 3 frames, that is, the microphone resumes pickup after being muted for 40-60ms.

[0087] S602. If the feedback level is the second level, then extract continuous non-voice audio frames from the audio stream data, and perform segmented amplitude compression processing on the continuous non-voice audio frames so that there are continuous zero signal segments in the processed continuous non-voice audio frames. The feedback severity of the second level is lower than that of the first level.

[0088] Optionally, the severity of feedback at the second level is lower than that at the first level. If the feedback level is the second level, it means that the microphone terminal occasionally experiences feedback within the current preset time period, and the feedback intensity is low. In this case, the microphone terminal performs speech gap detection on the audio stream data and extracts continuous non-speech audio frames from the audio stream data. For example, three continuous non-speech audio frames are extracted from the audio stream data.

[0089] The pickup terminal segments continuous non-voice audio frames into multiple segments to be processed. Based on a pre-set gain factor for each segment, the gain factor is multiplied by the audio signal amplitude of that segment, gradually compressing the amplitude of the continuous non-voice audio frames. This results in consecutive zero-signal segments within the processed continuous non-voice audio frames. These zero-signal segments disrupt the continuous feedback path that generates howling, thus eliminating mild howling. Specifically, at least one frame in the processed continuous non-voice audio frames must have a signal amplitude of zero, meaning the duration of the consecutive zero-signal segments is greater than or equal to 20ms.

[0090] For example, continuous non-voice audio frames can be divided into 10 segments, and the amplitude of the continuous non-voice audio frames can be compressed 10 times. After the 5th compression, the signal amplitude becomes zero. Then, the amplitude of the segments to be processed in segments 6-10 is also zero after compression. That is, the continuous zero signal segments in the continuous non-voice audio frames after processing are the segments to be processed in segments 6-10.

[0091] By eliminating feedback during speech gaps that are imperceptible to the user, the normal acquisition of speech by the microphone terminal and the normal playback of speech by the speaker are not affected, thus ensuring the fidelity of the speech and avoiding timbre distortion.

[0092] In this embodiment, if the feedback level is level one, the pickup terminal will mute a preset frame and resume pickup after the preset frame is muted. If the feedback level is level two, continuous non-voice audio frames are extracted from the audio stream data, and segmented amplitude compression is performed on the continuous non-voice audio frames to ensure that there are continuous zero signal segments in the processed continuous non-voice audio frames. When feedback occurs frequently, the continuous feedback path required for feedback generation is temporarily disrupted by silencing the preset frame and resuming pickup, thus quickly ending severe / frequent feedback. When feedback occurs occasionally, segmented amplitude compression is performed on continuous non-voice audio frames during speech gaps that are imperceptible to the user. The continuous zero signal segments in the processed continuous non-voice audio frames disrupt the continuous feedback path for feedback generation, thus eliminating mild feedback, ensuring speech fidelity, and avoiding timbre distortion.

[0093] This application also provides a microphone terminal, such as... Figure 8 The diagram shown is a schematic representation of the structure of a microphone terminal provided in this embodiment of the application, including a processor 81, a memory 82, and a bus 83. The memory 82 stores machine-readable instructions executable by the processor 81. When the microphone terminal is running, the processor 81 communicates with the memory 82 via the bus 83, and the processor 81 executes the machine-readable instructions to perform the steps of the howling suppression method in the aforementioned embodiment.

[0094] This application also provides a feedback suppression system, including the pickup terminal described in the foregoing embodiments and a collection terminal that is communicatively connected to the pickup terminal, wherein the collection terminal is externally connected to a speaker.

[0095] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0096] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0097] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for suppressing howling, characterized in that, Applied to a microphone terminal, the method includes: Real-time monitoring of the distance between the pickup terminal and the collection terminal; If the distance is less than a preset distance threshold, then howling suppression is performed according to the first howling suppression strategy; If the distance is greater than or equal to the preset distance threshold, then noise floor detection and howling detection are performed to obtain noise floor detection results and howling detection results. Howling is suppressed according to the noise floor detection results and the second howling suppression strategy, and / or howling is suppressed according to the howling detection results and the third howling suppression strategy.

2. The method according to claim 1, characterized in that, The method of suppressing howling according to the first howling suppression strategy includes: The operating state is switched from the sound pickup state to the mute state to stop sound pickup, and when the distance between the sound pickup terminal and the collection terminal is detected to change to be greater than or equal to the preset distance threshold, the operating state is switched from the mute state to the sound pickup state to resume sound pickup.

3. The method according to claim 1, characterized in that, The process of performing background noise detection and howling detection, and obtaining background noise detection results and howling detection results, includes: Real-time acquisition of audio stream data over a preset time period, wherein the audio stream data includes multiple audio frames; Multiple non-voice audio frames are extracted from the audio stream data, and background noise is detected based on each non-voice audio frame to obtain the background noise detection result. The background noise detection result includes the root mean square value of background noise of the sound pickup environment where the sound pickup terminal is located during the preset time period. Feedback detection is performed based on the audio stream data to obtain the feedback detection result, which includes the number of feedback frames of the pickup terminal in the preset time period.

4. The method according to claim 3, characterized in that, The step of suppressing howling based on the background noise detection result and according to the second howling suppression strategy includes: Determine whether the root mean square value of the noise floor is greater than a preset noise floor value; If so, howling suppression is performed based on the aforementioned root mean square noise floor value.

5. The method according to claim 4, characterized in that, The method of suppressing howling based on the root mean square value of the noise floor includes: When the root mean square value of the noise floor is greater than the preset noise floor value, the pickup gain of the pickup terminal is iteratively reduced according to the preset gain adjustment step size to narrow the pickup range of the pickup terminal.

6. The method according to claim 3, characterized in that, The step of suppressing howling based on the howling detection result and according to the third howling suppression strategy includes: Determine whether the audio stream data contains howling frames; If so, then howling suppression is performed based on the number of howling frames.

7. The method according to claim 6, characterized in that, The method of suppressing howling based on the number of howling frames includes: The howling level is determined based on the number of howling frames. Suppress the howling based on the howling level.

8. The method according to claim 7, characterized in that, The feedback suppression based on the feedback level includes: If the howling level is the first level, the preset frame will be muted, and the sound pickup will be restored after the preset frame is muted; If the feedback level is the second level, then continuous non-voice audio frames are extracted from the audio stream data, and segmented amplitude compression processing is performed on the continuous non-voice audio frames so that there are continuous zero signal segments in the processed continuous non-voice audio frames, wherein the feedback severity of the second level is lower than that of the first level.

9. A microphone terminal, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the microphone terminal is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the feedback suppression method as described in any one of claims 1 to 8.

10. A howling suppression system, characterized in that, It includes the pickup terminal as described in claim 9 and a collection terminal that is communicatively connected to the pickup terminal, wherein the collection terminal is externally connected to a speaker.

Images