阅读说明：本技术 一种啸叫抑制方法、装置、入耳式耳机及存储介质 (Howling suppression method and device, in-ear earphone and storage medium ) 是由 丁智慧 许国军 于 2021-08-25 设计创作，主要内容包括：本申请公开了一种啸叫抑制方法,所述啸叫抑制方法包括：根据麦克风采集的音频信号判断是否存在啸叫事件；若是,则确定所述啸叫事件对应的啸叫类型；若所述啸叫类型为冲击波类型的啸叫,则降低所述啸叫事件所在频段的幅值增益,并对降低幅值增益后的音频信号进行滤波处理得到目标音频信号；控制所述扬声器播放所述目标音频信号。本申请能够提高啸叫事件的抑制效果,降低音频播放对于用户听力的损害。本申请还公开了一种啸叫抑制装置、一种入耳式耳机及一种存储介质,具有以上有益效果。(The application discloses a howling suppression method, which comprises the following steps: judging whether a howling event exists according to an audio signal collected by a microphone; if yes, determining a howling type corresponding to the howling event; if the howling type is shock wave type howling, reducing amplitude gain of a frequency band where the howling event is located, and filtering the audio signal with the amplitude gain reduced to obtain a target audio signal; and controlling the loudspeaker to play the target audio signal. The method and the device can improve the inhibition effect of the howling event and reduce the damage of audio playing to the hearing of the user. The application also discloses a howling suppression device, an in-ear earphone and a storage medium, which have the beneficial effects.)

1. A howling suppression method, comprising:

judging whether a howling event exists according to an audio signal collected by a microphone;

if yes, determining a howling type corresponding to the howling event;

if the howling type is shock wave type howling, reducing amplitude gain of a frequency band where the howling event is located, and filtering the audio signal with the amplitude gain reduced to obtain a target audio signal;

and controlling the loudspeaker to play the target audio signal.

2. The howling suppression method as claimed in claim 1, wherein the step of filtering the audio signal with the amplitude gain reduced to obtain the target audio signal comprises:

determining a speaker reference signal collected by a feedback microphone;

inputting the speaker reference signal and the audio signal with the amplitude gain reduced into an adaptive filter to obtain a first filtering result;

inputting the first filtering result and the audio signal with the amplitude gain reduced into a nonlinear post-processing filter to obtain a second filtering result;

and superposing the first filtering result and the second filtering result to obtain a superposition result, and subtracting the superposition result from the audio signal to obtain the target audio signal.

3. The howling suppression method according to claim 1, wherein the determining whether the howling event exists according to the audio signal collected by the microphone comprises:

and judging whether a howling event exists according to the amplitude of the audio signal collected by the microphone.

4. The howling suppression method according to claim 3, wherein determining the howling type corresponding to the howling event comprises:

extracting a feature vector in the audio signal;

and inputting the characteristic vector into a howling classification model, and determining a howling type corresponding to the howling event according to an output result of the howling classification model.

5. The howling suppression method according to claim 1, after determining the howling type according to the frequency band where the howling event is located, further comprising:

and if the howling type is a non-shock wave type howling, filtering the acquired audio signal by using a self-adaptive filter to obtain a target audio signal.

6. The howling suppression method according to any one of claims 1 to 5, wherein reducing the amplitude gain of the frequency band in which the howling event is located comprises:

determining gain variation according to the span and energy of the frequency band of the howling event;

and reducing the amplitude gain of the frequency band where the howling event is located according to the gain variation.

7. The howling suppression method according to claim 6, after reducing the amplitude gain of the frequency band where the howling event is located according to the gain variation, further comprising;

delaying a preset time, and judging whether a howling event exists at the current moment;

if so, reducing the gain variation by a preset value to obtain a new gain variation, and reducing the amplitude gain of the frequency band where the howling event is located according to the new gain variation.

8. A howling suppression device, comprising:

the squeaking detection module is used for judging whether a squeaking event exists according to the audio signal collected by the microphone;

the howling classification module is used for determining a howling type corresponding to the howling event if the howling event exists;

the signal adjusting module is used for reducing the amplitude gain of the frequency band of the howling event if the howling type is shock wave type howling, and filtering the audio signal with the amplitude gain reduced to obtain a target audio signal;

and the audio playing module is used for controlling the loudspeaker to play the target audio signal.

9. An in-ear headphone comprising a microphone, a loudspeaker, a memory, and a processor, the memory having stored therein a computer program, the processor implementing the steps of the howling suppression method as claimed in any one of claims 1 to 7 when calling the computer program in the memory.

10. A storage medium having stored thereon computer-executable instructions which, when loaded and executed by a processor, carry out the steps of the howling suppression method according to any one of claims 1 to 7.

Technical Field

The present application relates to the field of audio processing technologies, and in particular, to a howling suppression method and apparatus, an in-ear headphone, and a storage medium.

Background

With age, the audible limit of the human ear is reduced. World hearing reports issued by the world health organization indicate that: currently, one fifth of the world is hearing-impaired, and hearing loss affects more than 15 hundred million people worldwide, wherein 4.3 hundred million people have hearing loss of moderate or more degree in their ears with better hearing, and the world health organization suggests providing all people with ear and hearing health services. Under market excitation, the function of assisting hearing under a TWS (True Wireless Stereo) earphone is more and more important, and because hearing of hearing-impaired people to a specific frequency band is lost, the function of assisting hearing of the TWS earphone is realized by amplifying gain under the lost frequency band of the hearing-impaired patients. However, when the auxiliary hearing function is implemented, the distance between the microphone and the speaker is too close, which easily causes howling.

In the related art, adaptive filters are mainly used to suppress howling, but when howling occurs in a higher frequency band, howling cannot be suppressed only by using the adaptive filters, which causes damage to the hearing of the user.

Therefore, how to improve the suppression effect of the howling event and reduce the damage of the audio playing to the hearing of the user is a technical problem to be solved by those skilled in the art at present.

Disclosure of Invention

The application aims to provide a howling suppression method, a device, an in-ear earphone and a storage medium, which can improve the suppression effect of a howling event and reduce the damage of audio playing to the hearing of a user.

In order to solve the above technical problem, the present application provides a howling suppression method, including:

judging whether a howling event exists according to an audio signal collected by a microphone;

if yes, determining a howling type corresponding to the howling event;

if the howling type is shock wave type howling, reducing amplitude gain of a frequency band where the howling event is located, and filtering the audio signal with the amplitude gain reduced to obtain a target audio signal;

and controlling the loudspeaker to play the target audio signal.

Optionally, the filtering the audio signal with the amplitude gain reduced to obtain a target audio signal includes:

determining a speaker reference signal collected by a feedback microphone;

inputting the speaker reference signal and the audio signal with the amplitude gain reduced into an adaptive filter to obtain a first filtering result;

inputting the first filtering result and the audio signal with the amplitude gain reduced into a nonlinear post-processing filter to obtain a second filtering result;

and superposing the first filtering result and the second filtering result to obtain a superposition result, and subtracting the superposition result from the audio signal to obtain the target audio signal.

Optionally, the determining whether a howling event exists according to the audio signal collected by the microphone includes:

and judging whether a howling event exists according to the amplitude of the audio signal collected by the microphone.

Optionally, determining a howling type corresponding to the howling event includes:

extracting a feature vector in the audio signal;

and inputting the characteristic vector into a howling classification model, and determining a howling type corresponding to the howling event according to an output result of the howling classification model.

Optionally, after determining the howling type according to the frequency band where the howling event is located, the method further includes:

and if the howling type is a non-shock wave type howling, filtering the acquired audio signal by using a self-adaptive filter to obtain a target audio signal.

Optionally, reducing the amplitude gain of the frequency band where the howling event is located includes:

determining gain variation according to the span and energy of the frequency band of the howling event;

and reducing the amplitude gain of the frequency band where the howling event is located according to the gain variation.

Optionally, after the amplitude gain of the frequency band where the howling event is located is reduced according to the gain variation, the method further includes;

delaying a preset time, and judging whether a howling event exists at the current moment;

if so, reducing the gain variation by a preset value to obtain a new gain variation, and reducing the amplitude gain of the frequency band where the howling event is located according to the new gain variation.

The present application also provides a howling suppression apparatus, including:

the squeaking detection module is used for judging whether a squeaking event exists according to the audio signal collected by the microphone;

the howling classification module is used for determining a howling type corresponding to the howling event if the howling event exists;

the signal adjusting module is used for reducing the amplitude gain of the frequency band of the howling event if the howling type is shock wave type howling, and filtering the audio signal with the amplitude gain reduced to obtain a target audio signal;

and the audio playing module is used for controlling the loudspeaker to play the target audio signal.

The present application also provides a storage medium having a computer program stored thereon, which when executed implements the steps performed by the above-described howling suppression method.

The application also provides an in-ear earphone, which comprises a microphone, a loudspeaker, a memory and a processor, wherein the memory stores a computer program, and the processor realizes the steps executed by the howling suppression method when calling the computer program in the memory.

The application provides a howling suppression method, which comprises the following steps: judging whether a howling event exists according to an audio signal collected by a microphone; if yes, determining a howling type corresponding to the howling event; if the howling type is shock wave type howling, reducing amplitude gain of a frequency band where the howling event is located, and filtering the audio signal with the amplitude gain reduced to obtain a target audio signal; and controlling the loudspeaker to play the target audio signal.

After the howling event is judged to exist, the howling type corresponding to the howling event is determined, so that the frequency band of the howling event is determined according to the howling type. If the howling type is a shock wave type howling, the energy of the howling event is high, and the howling cannot be completely suppressed only by using the filter. According to the method and the device, the amplitude gain of the frequency band where the howling event is located is reduced, and then the audio signal with the reduced amplitude gain is filtered to obtain the target audio signal. The feedback signals which can cause howling in the target audio signal are less, so that the target audio signal is played by using the loudspeaker, the inhibition effect of the howling event can be improved, and the damage of audio playing to the hearing of a user can be reduced. The application also provides a howling suppression device, a storage medium and an in-ear earphone, which have the beneficial effects and are not repeated herein.

Drawings

In order to more clearly illustrate the embodiments of the present application, the drawings needed for the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings can be obtained by those skilled in the art without inventive effort.

Fig. 1 is a flowchart of a howling suppression method according to an embodiment of the present application;

fig. 2 is a schematic diagram of an audio signal filtering processing method according to an embodiment of the present disclosure;

fig. 3 is a schematic diagram illustrating a principle of howling suppression according to an embodiment of the present application;

fig. 4 is a flowchart of processing a sound signal for suppressing howling according to an embodiment of the present application;

fig. 5 is a schematic structural diagram of a howling suppression apparatus according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

Referring to fig. 1, fig. 1 is a flowchart of a howling suppression method according to an embodiment of the present application.

The specific steps may include:

s101: acquiring an audio signal collected by a microphone;

the present embodiment may be applied to a hearing aid, an earphone, and other devices having a microphone (specifically, a feedback microphone) and a speaker, where when the distance between the microphone and the speaker is too close or the output sound of the speaker is too loud, the microphone collects the output signal of the speaker, so that the output signal of the speaker is amplified again, and a howling event occurs.

S102: judging whether a howling event exists according to an audio signal collected by a microphone; if yes, entering S102; if not, ending the flow;

after the audio signal collected by the microphone is obtained, whether a howling event exists can be judged according to the amplitude of the audio signal; namely: and when the audio signal has a signal with the amplitude larger than the preset amplitude, judging that the howling event is detected.

S103: determining a howling type corresponding to the howling event;

the howling types in this embodiment may include a shock wave type howling and a non-shock wave type howling. The energy of the shock wave type howling is higher than that of the non-shock wave type howling, and the embodiment may divide the shock wave type howling and the non-shock wave type howling according to the adaptive filter's ability to process the howling. For example, the highest howling energy that the adaptive filter can suppress is a, the howling type of the howling event with energy greater than a is a shock wave type howling, and the howling type of the howling event with energy less than or equal to a is a non-shock wave type howling.

S104: if the howling type is shock wave type howling, reducing amplitude gain of a frequency band where the howling event is located, and filtering the audio signal with the amplitude gain reduced to obtain a target audio signal;

the howling event that occurs at present is caused by the shock wave, and cannot be suppressed only by using the adaptive filter, and this embodiment first processes the gain of the audio signal. Specifically, the amplitude gain of the frequency band where the howling event is located in the audio signal may be obtained, and the audio signal with the reduced amplitude gain may be filtered to reduce feedback signals causing the howling, so as to obtain the target audio signal.

S105: and controlling the loudspeaker to play the target audio signal.

After determining that the howling event exists, the embodiment determines the howling type corresponding to the howling event, so as to determine the frequency band where the howling event occurs according to the howling type. If the howling type is a shock wave type howling, the energy of the howling event is high, and the howling cannot be completely suppressed only by using the filter. In this embodiment, the amplitude gain of the frequency band where the howling event is located is first reduced, and then the audio signal with the reduced amplitude gain is filtered to obtain a target audio signal. The feedback signals which can cause howling in the target audio signal are less, so that the target audio signal is played by using the loudspeaker, the inhibition effect of the howling event can be improved, and the damage of audio playing to the hearing of a user can be reduced.

As a possible implementation manner, for the shockwave type howling, the present embodiment may utilize an adaptive filter and a nonlinear post-processing filter to jointly process the audio signal to obtain the target audio signal. Referring to fig. 2, fig. 2 is a schematic diagram of an audio signal filtering method according to an embodiment of the present disclosure, and this embodiment further describes the filtering processing performed on the audio signal with the amplitude gain reduced by S104 in the embodiment corresponding to fig. 1, and specifically includes the following steps:

s201: determining a speaker reference signal collected by a feedback microphone;

in the embodiment, a speaker reference signal collected by a feedback microphone is used as a reference, so as to remove the influence caused by nonlinear distortion howling.

S202: inputting the speaker reference signal and the audio signal with the amplitude gain reduced into an adaptive filter to obtain a first filtering result;

s203: inputting the first filtering result and the audio signal with the amplitude gain reduced into a nonlinear post-processing filter to obtain a second filtering result;

s204: and superposing the first filtering result and the second filtering result to obtain a superposition result, and subtracting the superposition result from the audio signal to obtain the target audio signal.

In this embodiment, the adaptive filter is used to cancel linear echo, the nonlinear post-processing filter is used to cancel nonlinear echo, and the first filtering result and the second filtering result are superimposed to obtain the target audio signal after linear echo and nonlinear echo are cancelled.

As a feasible implementation manner, after determining the howling type according to the frequency band where the howling event is located, if the howling type is a non-shockwave type of howling, filtering the acquired audio signal by using an adaptive filter to obtain a target audio signal.

As a possible implementation, since the howling event refers to a situation where the energy (amplitude) of the output signal of the speaker is too high in the high frequency band, the audio signal collected by the microphone includes the output signal of the speaker and the output signal of the non-speaker (such as the voice of the user, the ambient noise, etc.). The embodiment can judge whether the howling event exists according to the amplitude of the audio signal collected by the microphone.

Correspondingly, the embodiment may also determine the howling type by using the speaker reference signal, and the specific process is as follows: extracting a feature vector in the audio signal; and inputting the characteristic vector into a howling classification model, and determining a howling type corresponding to the howling event according to an output result of the howling classification model.

As a possible implementation manner, the foregoing embodiment may reduce the amplitude gain of the frequency band where the howling event is located by the following process, where the process includes: determining gain variation according to the span and energy of the frequency band of the howling event; and reducing the amplitude gain of the frequency band where the howling event is located according to the gain variation.

Further, after the amplitude gain of the frequency band where the howling event is located is reduced according to the gain variation, a preset time duration can be delayed, and whether the howling event exists at the current moment is judged; if so, reducing the gain variation by a preset value to obtain a new gain variation, and reducing the amplitude gain of the frequency band where the howling event is located according to the new gain variation.

The flow described in the above embodiment is explained below by a howling suppression scheme of an in-ear headphone in practical use. Fig. 3 is a schematic diagram of a howling suppression principle provided in an embodiment of the present application, where a speaker signal collected by a feedback microphone is used as a reference, and a gain of a howling part is reduced according to a howling type to maintain stability of an adaptive filter. As shown in fig. 3, a microphone collects a sound signal, the sound signal passes through a speaker after speech preprocessing, the speech preprocessing includes noise reduction and dynamic range control processing, the signal passes through a speaker feedback microphone to collect sound played by the speaker, at this time, the sound amplified by the speaker is picked up again by the microphone due to the short distance between the microphone and the speaker, the picked-up signal is transmitted to the speaker again by the microphone, the signal is infinitely superposed and amplified under multiple cycles, and howling occurs when the amplitude is infinitely superposed at a certain frequency point.

Assuming that a signal collected by a microphone is X, a pure speech signal is S, and a feedback signal represents N, the relationship of the signals is: x is S + N. A clean speech signal refers to a speech signal without any noise and feedback. The feedback signal refers to the signal output by the loudspeaker and picked up by the microphone, which can be determined from the feedback path and the loudspeaker reference signal, i.e.: n ═ F × Y; where F denotes the feedback path and Y denotes the loudspeaker signal. The feedback path refers to a propagation path of sound; for example, a speaker emits sound, which is picked up by a microphone, and the sound from the speaker is propagated to the microphone as a feedback path. The speaker signal is also called a speaker reference signal because it is used as a reference signal when processed in the filter.

In order to solve the howling problem, accurate estimation of a feedback path and a loudspeaker signal is needed, the estimation of the feedback path is obtained according to the updating of an adaptive filter, and the formula of the feedback path estimation is as follows: f '(t) ═ F' (t-1) + a ((Y × Y)/((Y × E) + d)); where a denotes a preset coefficient, E denotes an error signal at the previous time, t denotes time, F 'denotes estimated feedback path information, and F denotes real feedback path information, the present embodiment may use the feedback path information estimated by the above formula as the real feedback path information, and E (t-1) ═ X-F' (t-1) × Y (t-1); if the coefficient is estimated accurately, the reference signal Y, i.e. the sound emitted by the loudspeaker, must be obtained accurately. The sound of the loudspeaker will pass through the path or be gained again to generate nonlinear distortion, so that the difference between the reference and the actual loudspeaker signal directly used as the reference is large. The embodiment can utilize a Feedback microphone (Feedback Mic) to collect the sound of the loudspeaker as a reference, and can effectively remove the influence caused by nonlinearity. Since the amplified loudspeaker signal has nonlinear distortion through output, the nonlinear distortion cannot be estimated, and the loudspeaker signal using the feedback microphone can completely consider the sound to be the sound played by the loudspeaker.

The updating principle of the adaptive filter is that the updating is stable when the signal is stable, and if the shock wave type howling occurs, the situation that convergence cannot be achieved is brought to the updating of the adaptive filter, so when the shock wave type howling caused by sounds such as a horn occurs, the feedback path is accurately estimated by reducing the gain and then updating the coefficient of the adaptive filter.

The type of howling detection may be classified according to the excitation signal or the normal voice signal, and may include a horn whistle sound, a microwave oven operation sound, and a normal voice signal, for example. In this embodiment, the type detection may be performed by using a howling classification model of a Neural Network such as DNN (Deep Neural Network), RNN (Recurrent Neural Network), CNN (Convolutional Neural Network), and the like. And (3) processing the classified signals, firstly performing frequency point gain control if the signals are microwave ovens or whistling sounds, and then performing filtering processing according to the adaptive filter to remove feedback signals, but because the loudspeaker signals are processed and amplified signals, the signals inevitably have nonlinear change, increasing NLP (non-linear programming) to perform filtering processing on the signals, solving errors caused by nonlinearity and eliminating instability of high-frequency gain.

Referring to fig. 4, fig. 4 is a flowchart of processing a sound signal for suppressing howling according to an embodiment of the present application, where the flowchart may include the following steps:

the microphone collects the audio signal X (n), and voice preprocessing operations such as noise reduction, dynamic range control and the like are executed on the audio signal X (n). In fig. 4, s (n) represents a pure signal without a feedback signal, and vf (n) represents a feedback signal.

And inputting the audio signal X (n) subjected to the voice preprocessing operation into a howling detection classification model, and if the audio signal is of shock wave type howling, adjusting a gain value G of a frequency band where the howling is positioned by performing gain control at a specific frequency point, wherein the gain value G represents a gain range grade value of a howling point. The frequency point of the above characteristic refers to a frequency band in which howling occurs.

The neural network judgment type is mainly judged according to the microphone signals. And (2) extracting the characteristics of the audio signal X (n) collected by the microphone, wherein the length of the extracted characteristic vector is 22, and the characteristic vector is sent to a howling detection classification model to be judged so as to determine whether the howling signal exists in the microphone, and if the howling signal exists in the microphone, namely the characteristic has a larger value, the neural network is biased to judge whether the impact howling exists, and the impact howling is fed back to a dynamic range gain control algorithm module at the moment so as to reduce the gain increase of the microphone signal. The control of the gain G is reduced by gain cycling until the neural network determines that there is no impulse howling. The feature extraction algorithm of the microphone may be Mel-Frequency spectrum feature extraction algorithm (MFCCs), and the feature extraction operation includes: the frequency band is divided into 22 frequency bands, and after the frequency band energy is calculated, discrete cosine transform is calculated, so that 22 numerical values are obtained in total.

Further, if the howling type is normal voice and medium-high frequency howling (i.e. low energy howling), the adaptive filter may be used to process the howling.

The loudspeaker y (n) is collected by a feedback microphone as Fb (n), and the signal is used as a reference ref of the adaptive filter; the adaptive filter has a coefficient W1 and the inputs are the microphone signal x (n) and the reference signal ref. The coefficient of the nonlinear post-processing filter (NLP) is updated to W2, and the input is the output e (n) of the adaptive filter and the reference signal ref; coefficient updates for the non-linear post-processing filter are calculated from the output of the adaptive filter and the reference signal.

The adaptive filter has the function of eliminating linear echo, the nonlinear post-processing filter has the function of eliminating nonlinear echo, and the final output comprises yout1 with linear feedback removed and yout2 with nonlinear feedback removed, so that the finally superposed yout (n) does not comprise a feedback signal, thereby avoiding the generation of howling. When the signal is excessively amplified, due to the fact that vibration molecules around the loudspeaker are increased, nonlinear distortion is serious, the scheme uses the adaptive filter and the nonlinear post-processing filter to carry out combined feedback elimination, feedback signals causing howling can be effectively eliminated, and instability caused by adjustment of a wide dynamic range in the forward processing process can be solved.

Howling occurs in a high frequency portion, but the feedback signal attenuation effect of the adaptive filter on a high frequency signal in combination with signal preprocessing is extremely unstable, so that howling sometimes occurs. In the embodiment, the adaptive filter is updated and controlled in step length by detecting the howling type and controlling the gain of the howling frequency point, and the adaptive filter is stabilized by reducing the gain and then processing the high-frequency howling. The howling generation is basically gain increase, the howling types are classified according to the number of howling frequency points generated by different sound sources and the frequency point characteristics of the howling, and the howling types are judged through a neural network and mainly divided into the howling caused by external excitation sources such as a loudspeaker, a voice source and a noise source. And directly performing gain attenuation control according to the type of the howling, regulating and controlling the step length of the filter, controlling a feedback coefficient according to the volume of data collected by the feedback microphone, and performing combined feedback elimination by matching with a nonlinear post-processing filter to suppress the howling.

Referring to fig. 5, fig. 5 is a schematic structural diagram of a howling suppression device provided in an embodiment of the present application, where the device may include:

a howling detection module 501, configured to determine whether a howling event exists according to an audio signal collected by a microphone;

a howling classification module 502, configured to determine a howling type corresponding to the howling event if the howling event exists;

a signal adjusting module 503, configured to reduce an amplitude gain of a frequency band where the howling event is located if the howling type is a shock wave type howling, and perform filtering processing on the audio signal with the amplitude gain reduced to obtain a target audio signal;

an audio playing module 504, configured to control the speaker to play the target audio signal.

After determining that the howling event exists, the embodiment determines the howling type corresponding to the howling event, so as to determine the frequency band where the howling event occurs according to the howling type. If the howling type is a shock wave type howling, the energy of the howling event is high, and the howling cannot be completely suppressed only by using the filter. In this embodiment, the amplitude gain of the frequency band where the howling event is located is first reduced, and then the audio signal with the reduced amplitude gain is filtered to obtain a target audio signal. The feedback signals which can cause howling in the target audio signal are less, so that the target audio signal is played by using the loudspeaker, the inhibition effect of the howling event can be improved, and the damage of audio playing to the hearing of a user can be reduced.

Further, the signal adjusting module 503 is configured to determine a speaker reference signal collected by a feedback microphone; the adaptive filter is further used for inputting the speaker reference signal and the audio signal with the amplitude gain reduced into the adaptive filter to obtain a first filtering result; the first filtering result and the audio signal with the reduced amplitude gain are input into a nonlinear post-processing filter to obtain a second filtering result; and the second filter module is further configured to superimpose the first filter result and the second filter result to obtain a superimposed result, and subtract the superimposed result from the audio signal to obtain the target audio signal.

Further, the howling detection module 501 is configured to determine whether a howling event exists according to the magnitude of the audio signal collected by the microphone.

Further, the howling classification module 502 is configured to extract a feature vector in the audio signal; and the howling classification model is also used for inputting the characteristic vector into the howling classification model and determining the howling type corresponding to the howling event according to the output result of the howling classification model.

Further, the signal adjusting module 503 is further configured to, after determining the howling type according to the frequency band where the howling event is located, if the howling type is a non-shockwave type of howling, perform filtering processing on the acquired audio signal by using an adaptive filter to obtain a target audio signal.

Further, the signal adjusting module 503 is configured to determine a gain variation according to the span and energy of the frequency band where the howling event is located; and the amplitude gain of the frequency band where the howling event is located is also reduced according to the gain variation.

Further, the method also comprises the following steps;

the gain adjustment module is used for delaying a preset time length after reducing the amplitude gain of the frequency band where the howling event is located according to the gain variation, and judging whether the howling event exists at the current moment; if so, reducing the gain variation by a preset value to obtain a new gain variation, and reducing the amplitude gain of the frequency band where the howling event is located according to the new gain variation.

Since the embodiments of the apparatus portion and the method portion correspond to each other, please refer to the description of the embodiments of the method portion for the embodiments of the apparatus portion, which is not repeated here.

The present application also provides a storage medium having a computer program stored thereon, which when executed, may implement the steps provided by the above-described embodiments. The storage medium may include: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

The present application further provides an in-ear headphone, which includes a feedback microphone, a speaker, a memory, and a processor, where the memory stores a computer program, and the processor calls the computer program in the memory to implement the steps provided in the above embodiments. The in-ear headphones may include wired headphones and wireless headphones (e.g., TWS true wireless headphones).

The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description. It should be noted that, for those skilled in the art, it is possible to make several improvements and modifications to the present application without departing from the principle of the present application, and such improvements and modifications also fall within the scope of the claims of the present application.

It is further noted that, in the present specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.