WO2022217978A1 - 一种抑制啸叫的方法、装置、芯片及模组设备 - Google Patents

一种抑制啸叫的方法、装置、芯片及模组设备 Download PDF

Info

Publication number
WO2022217978A1
WO2022217978A1 PCT/CN2021/143642 CN2021143642W WO2022217978A1 WO 2022217978 A1 WO2022217978 A1 WO 2022217978A1 CN 2021143642 W CN2021143642 W CN 2021143642W WO 2022217978 A1 WO2022217978 A1 WO 2022217978A1
Authority
WO
WIPO (PCT)
Prior art keywords
howling
spectral
signal
spectrum
detection threshold
Prior art date
Application number
PCT/CN2021/143642
Other languages
English (en)
French (fr)
Inventor
潘思伟
罗本彪
雍雅琴
董斐
林福辉
Original Assignee
展讯通信（上海）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 展讯通信（上海）有限公司 filed Critical 展讯通信（上海）有限公司
Publication of WO2022217978A1 publication Critical patent/WO2022217978A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G10L21/0216—Noise filtering characterised by the method used for estimating noise
    • G10L21/0232—Processing in the frequency domain
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R27/00—Public address systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R3/00—Circuits for transducers, loudspeakers or microphones
    • H04R3/02—Circuits for transducers, loudspeakers or microphones for preventing acoustic reaction, i.e. acoustic oscillatory feedback

Definitions

  • the present invention relates to the field of communications, and in particular, to a method, device, chip and module equipment for suppressing howling.
  • the sound reinforcement system can process the voice signal with lower volume, so as to output the voice signal with higher volume.
  • the microphone receives the external voice signal, it converts the voice signal into an input signal, so that the input signal is amplified by the power amplifier and then output by the loudspeaker, and the output signal passes through the boundary
  • the reflection and refraction are picked up by the microphone again, forming a closed loop. This process is called acoustic feedback.
  • the main methods are: notch method, frequency shift method and adaptive feedback cancellation method.
  • the notch method is the most widely used in practice.
  • the notch method is mainly composed of two parts, the first part is howling detection, and the second part is notch filter design.
  • the basic principle of howling suppression by the notch method is to accurately detect howling frequencies appearing in the acoustic feedback and perform notch processing to reduce the gain at the howling frequencies, destroy the gain conditions for howling, and achieve the purpose of suppressing howling. . Since the howling suppression by the notch method is highly dependent on howling detection, the existing howling detection methods are difficult to quickly and accurately detect howling or broadband howling with more than two howling frequency points at the same time. The actual application effect is not ideal.
  • the present application provides a method, device, chip and module device for suppressing howling, which is beneficial to suppressing multi-frequency howling and broadband howling.
  • the present application provides a method for suppressing howling, the method comprising: receiving a first signal input by a microphone; acquiring a spectrum signal corresponding to the first signal; determining a howling detection threshold based on the first spectrum amplitude, the a spectrum amplitude is the maximum spectrum amplitude of the spectrum signal in the preset reference frequency band; determine the howling spectrum component in the spectrum signal based on the howling detection threshold; based on the howling detection threshold and the howling spectrum component corresponding
  • the spectral amplitude determines a suppression parameter; the spectral signal is filtered based on the suppression parameter.
  • the method can simultaneously detect the howling of multiple frequency points in the spectrum signal Or broadband howling, which is conducive to suppressing multi-frequency howling and broadband howling.
  • the first spectrum component is the whistling spectrum component
  • the first frequency point is any part of the spectrum signal. one or more frequency points.
  • the howling detection threshold corresponding to each frequency point is determined based on the first spectrum amplitude and the first adjustment function.
  • the first frequency point is determined to be the frequency spectrum component of the whistling, and the first frequency point is determined as the howling spectrum component.
  • a point is any one or more frequency points of the spectrum signal.
  • the suppression parameter corresponding to the howling spectral component is determined based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component; the corresponding frequency points other than the howling spectral component are determined.
  • the suppression parameter of 1 is 1.
  • the suppression parameter corresponding to the howling spectral component is determined based on the howling detection threshold, the spectral amplitude corresponding to the howling spectral component, and a second adjustment function.
  • the first signal is subjected to frame-by-frame processing to obtain multiple time-domain signals; the multiple time-domain signals are converted into multiple spectrum signals.
  • the present application provides a howling suppression device, which includes a howling detection module and a howling filtering module.
  • the howling detection module is used for: receiving a first signal input by a microphone; acquiring a spectrum signal corresponding to the first signal; the maximum spectral amplitude in the reference frequency band; determine the howling spectral component in the spectrum signal based on the howling detection threshold; determine the suppression parameter based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component; the howling filter a module for filtering the spectral signal based on the suppression parameter.
  • the present application provides an electronic device comprising a loudspeaker, a microphone, a memory and a processor: the loudspeaker is used to output a processed audio signal; the microphone is used to acquire the audio signal; The memory is used to store a computer program; the processor is specifically used to call the computer program from the memory to execute the method described in the first aspect.
  • the present application provides a chip, which is used for: receiving a first signal input by a microphone; acquiring a spectrum signal corresponding to the first signal; determining a howling detection threshold based on the amplitude of the first spectrum.
  • the amplitude is the maximum spectral amplitude of the spectral signal in the preset reference frequency band; the howling spectral component in the spectral signal is determined based on the howling detection threshold; the spectral amplitude corresponding to the howling detection threshold and the howling spectral component is determined Determine a suppression parameter; filter the spectral signal based on the suppression parameter.
  • the present application provides a module device including a power module, a storage module and a chip module.
  • the power module is used to provide power for the module equipment;
  • the storage module is used to store data and instructions;
  • the chip module is used to: receive the first signal input by the microphone; obtain the spectrum signal corresponding to the first signal; Determine a howling detection threshold based on a first spectral amplitude, which is the maximum spectral amplitude of the spectral signal within the preset reference frequency band; determine a howling spectral component in the spectral signal based on the howling detection threshold;
  • the howling detection threshold and the spectral amplitude corresponding to the howling spectral component determine a suppression parameter; the spectral signal is filtered based on the suppression parameter.
  • the present application provides a storage medium on which computer instructions are stored, and when the computer instructions are executed, the method described in the first aspect above is executed.
  • the present application provides a computer program or computer program product, comprising codes or instructions, which, when run on a computer, cause the computer to perform the method described in the first aspect.
  • Fig. 1 is the structural representation of a kind of existing sound reinforcement system
  • FIG. 2 is a schematic structural diagram of a sound reinforcement system provided by an embodiment of the present application.
  • FIG. 3 is a schematic flowchart of a method for suppressing howling provided by an embodiment of the present application
  • FIG. 4 is a schematic diagram of a spectrum signal provided by an embodiment of the present application.
  • FIG. 5 is a schematic diagram of another spectrum signal provided by an embodiment of the present application.
  • FIG. 6 is a schematic diagram of another spectrum signal provided by an embodiment of the present application.
  • FIG. 7 is a comparison diagram before and after howling processing of a voice signal provided by an embodiment of the present application.
  • FIG. 8 is a schematic flowchart of another method for suppressing howling provided by an embodiment of the present application.
  • FIG. 9 is a schematic diagram of an acoustic feedback scenario provided by an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.
  • FIG. 11 is a schematic structural diagram of a device provided by an embodiment of the present application.
  • FIG. 12 is a schematic structural diagram of a module device provided by an embodiment of the present application.
  • FIG. 2 is a schematic structural diagram of a sound reinforcement system according to an embodiment of the present application.
  • the near-end speech signal 201 is picked up by the microphone 202 to obtain the input signal 203 ( 205 ), and the amplified signal 207 is obtained after being acted by the forward gain unit 206 of the power amplifier, and then played by the loudspeaker 208 to obtain
  • the amplified output signal 209 is more easily received by the human ear than the speech signal 201 .
  • the amplified output signal 209 forms a feedback voice signal 211 through the acoustic feedback loop 210, and the feedback voice signal 211 is picked up by the microphone 202 to form a closed loop.
  • the sound reinforcement system 20 in FIG. 2 adds a howling detection module 212 and a filtering system module 204 to the original sound reinforcement system 10, and can filter the input signal 203 containing howling, thereby outputting an output signal without howling 209.
  • the howling detection module 212 detects the input signal 203 of the microphone 202.
  • the howling component If the howling component is detected, it outputs the suppression parameter 213 to the filtering system 204. After the input signal 203 is filtered by the filtering system 204, the signal 205 after the howling suppression is output. The amplified signal 207 is obtained after passing through the forward gain unit 206 , and then played through the loudspeaker 208 to obtain the output voice 209 . If the howling component is not detected, the command or the corresponding suppression parameter 213 is output to the filtering system 204 without filtering the input signal 203 .
  • FIG. 3 is a schematic flowchart of a howling suppression method provided by an embodiment of the present application.
  • the method is applied to an electronic device or a chip in the electronic device.
  • FIG. 3 takes the electronic device as the execution body of the method as an example for description.
  • the execution subject of the howling suppression method shown in other drawings in the embodiment of the present invention is the same, and will not be repeated hereafter. Steps 301 to 306 of the howling suppression method in this embodiment of the present application:
  • An electronic device receives a first signal input by a microphone.
  • the electronic device may be a device including a microphone and a loudspeaker, such as a mobile phone, a stereo or a walkie-talkie.
  • the electronic device can amplify the voice signal input by the microphone so as to make the loudspeaker output, which is beneficial for the user to hear the voice signal more clearly.
  • the first signal may be a voice signal received by the microphone from the outside world, or may be an audio signal obtained by the microphone through a communication interface or a memory, etc., which is not limited in this embodiment of the present application.
  • the electronic device acquires a spectrum signal corresponding to the first signal.
  • step 302 may specifically include: the electronic device performs frame-by-frame processing on the first signal to obtain multiple time-domain signals; and the electronic device converts the multiple time-domain signals into multiple spectrum signals.
  • the frame division processing requires not less than 40 frames per second.
  • the frame-by-frame processing facilitates the subsequent detection of howling spectral components more accurately.
  • the converted spectrum signals corresponding to the time-domain signals of different frames are different, and the different spectrum signals can be distinguished by using different frame identifiers.
  • the electronic device may also directly convert the first signal into a spectrum signal, without first acquiring the time domain signal, and the embodiment of the present application does not limit the manner of acquiring the spectrum signal corresponding to the first signal.
  • the electronic device determines a howling detection threshold based on a first spectral amplitude, where the first spectral amplitude is the maximum spectral amplitude of the spectral signal within a preset reference frequency band.
  • the preset reference frequency band is usually a low frequency band with a low occurrence probability of howling frequency points. Because the low frequency band with a low probability of howling frequency points is selected, it can avoid the situation that the spectrum amplitude corresponding to the howling frequency points in the signal is smaller than the howling detection threshold, resulting in the inability to filter out the howling in the signal.
  • the preset reference frequency band can be determined in two ways. One way is to determine only one cutoff frequency f c , and the preset reference frequency band is expressed as (0, f c ); the other way is to determine an initial frequency f s and cutoff frequency f c , the preset reference frequency band can be expressed as (f s , f c ). This embodiment of the present application does not limit the manner of how to determine the preset reference frequency band.
  • the first spectrum amplitude is the maximum spectrum amplitude of the spectrum signal within the preset reference frequency band.
  • An actual example is used to introduce the first spectrum amplitude.
  • FIG. 4 is a schematic diagram of the spectrum amplitude of a spectrum signal according to an embodiment of the present application.
  • the vertical axis A represents the spectral amplitude of the spectral signal
  • the horizontal axis f i represents the frequency point of the spectral signal
  • the solid line represents the spectral amplitude A(k, f i ) of the spectral signal
  • k is the frame identification.
  • the howling detection threshold is a fixed value.
  • the spectral amplitude of the spectral signal can be expressed as A(k, f i ) by a formula, and the reference frequency band is (0, f c ), where k is a frame identification, and f i is a frequency point.
  • the first spectral magnitude Thr_base can be represented by the following formula:
  • Thr_base max[A(k, f i )] f i ⁇ f c (1)
  • the size of the howling detection threshold is equal to the first spectrum amplitude.
  • the howling detection threshold may be a frequency-dependent segment threshold.
  • the corresponding howling detection threshold of each frequency point may be determined based on the first frequency spectrum amplitude and the first adjustment function.
  • the spectral amplitude of the spectral signal is A(k, f i )
  • the reference frequency band is (0, f c )
  • the first adjustment function is a(k, f i )
  • the first spectral amplitude is Thr_base.
  • the first adjustment function a(k, f i ) needs to satisfy the following conditions:
  • the electronic device can set different howling detection thresholds based on different frequencies, which is beneficial to more accurately detect howling spectral components.
  • the electronic device determines a howling spectral component in the spectral signal based on the howling detection threshold.
  • the howling frequency spectrum component may be one howling frequency point in the spectrum signal, or may be multiple howling frequency points, or may be a segment of howling frequency band, which is not limited in the embodiment of the present application. .
  • the specific implementation step of step 304 is: if the spectrum amplitude corresponding to the first frequency point is greater than the howling detection threshold, determine the first frequency
  • the spectrum component is the howling spectrum component
  • the first frequency point is any one or more frequency points of the spectrum signal.
  • the vertical axis A represents the spectral amplitude of the spectral signal
  • the horizontal axis f i represents the frequency point of the spectral signal
  • the solid line represents the spectral amplitude A(k, f i ) of the spectral signal
  • k is the frame identification
  • (f s , f c ) are preset reference frequency bands
  • the howling detection threshold is Thr_base.
  • the frequency band (f 1 , f 2 ) can be determined as the howling spectral component.
  • the specific implementation steps of step 304 are: if the spectrum amplitude corresponding to the first frequency point is greater than the frequency corresponding to the first frequency point
  • the first frequency point is the frequency spectrum component of the howling
  • the first frequency point is any one or more frequency points of the spectrum signal.
  • the vertical axis A represents the spectral amplitude
  • the horizontal axis f i represents the frequency point
  • the solid line A(k, f i ) represents the spectral amplitude of the spectral signal
  • the dotted line Thr(k, f i ) represents each frequency
  • the howling detection threshold corresponding to the point Comparing the howling detection threshold Thr(k, f i ) with the spectral amplitude of the spectral signal, it can be seen that the spectral amplitude corresponding to the frequency point in (f 1 , f 2 ) is greater than the frequency in (f 1 , f 2 ).
  • the howling detection threshold corresponding to the point for example, A(k, f 3 ) corresponding to f 3 is greater than Thr(k, f 3 ). Therefore, the frequency band (f 1 , f 2 ) can be determined as the howling spectral component.
  • the electronic device determines a suppression parameter based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component.
  • the specific implementation manner of step 305 is: based on the howling detection threshold and the spectrum amplitude corresponding to the howling spectral component, determine the suppression parameter corresponding to the howling spectral component; determine and divide the howling spectral component
  • the suppression parameter corresponding to other frequency points is 1. If the suppression parameter corresponding to a frequency point is 1, it means that the frequency point is not a howling frequency point, and in the filtering process, the frequency point in the spectrum signal will not be processed. If the suppression parameters corresponding to all frequency points are 1, it means that there is no howling spectrum component in the spectrum signal, and the signal does not need to be filtered.
  • the suppression parameter G(k, f i ) can be expressed by formula (4) or formula (5):
  • formula (4) is the expression of the suppression parameter when the howling detection threshold is a fixed value
  • formula (5) is the expression of the suppression parameter when the howling detection threshold is a frequency-related segmental threshold. way of representation.
  • the suppression parameter may also be determined based on the howling detection threshold, the spectral amplitude corresponding to the howling spectral component, and the second adjustment function. If the second adjustment function is expressed as ⁇ (k, f i ), the second adjustment function ⁇ (k, f i ) satisfies 0 ⁇ (k, f i ) ⁇ 1, then the above formula (4) and formula ( 5) can be replaced by formula (6) and formula (7) respectively to express:
  • formula (6) is the expression of the suppression parameter when the howling detection threshold is a fixed value
  • formula (7) is the expression of the suppression parameter when the howling detection threshold is a frequency-related segmental threshold. way of representation.
  • the suppression parameter determined by this implementation is beneficial to filtering out howling more accurately.
  • the electronic device performs filtering processing on the spectrum signal based on the suppression parameter.
  • the electronic device performs filtering processing on the spectrum signal based on the suppression parameter, and obtains a spectrum signal after howling suppression is obtained after filtering processing. Then the howling-suppressed spectral signal A s (k, f i ) can be expressed as:
  • this method can simultaneously detect the spectrum signal Howling or broadband whistling at multiple frequencies in the middle is beneficial to suppressing multi-frequency whistling and broadband whistling.
  • the upper picture in Fig. 7 is the speech signal before howling processing, the sampling rate is 8000Hz, the duration is about 6.7s, and the main howling components have been marked by dotted boxes. It can be seen that the howling components are distributed in multiple frequencies or broadband, the maximum bandwidth is not less than 1000Hz, and the maximum amplitude is close to the average amplitude of the near-end voice part below 2000Hz.
  • the lower picture in FIG. 7 is the speech signal processed by the howling suppression method proposed in the present application. It can be seen that all howling components are effectively suppressed with almost no delay. Because howling is effectively suppressed, there is no continuous howling component in the input speech signal before howling processing.
  • the near-end speech parts other than the howling components are not significantly suppressed or distorted, and the RMS average of all spectral energy is only reduced by about 1dB, indicating that this method can suppress howling quickly and effectively at the same time. , it is not easy to cause sound quality deterioration, and at the same time, it has less restriction on system gain.
  • FIG. 8 is a schematic flowchart of another howling suppression method provided by an embodiment of the present application.
  • the howling suppression method includes steps 801 to 807 . in:
  • the electronic device determines that the maximum spectral amplitude of the signal to be output in the loudspeaker exceeds a preset threshold.
  • the maximum spectral amplitude of the signal to be output in the loudspeaker exceeds the preset threshold, indicating that the signal to be amplified is about to output a signal with a larger spectral amplitude, and the output signal may be picked up by the microphone of the electronic device, Acoustic feedback is formed, resulting in howling.
  • howling suppression processing it is beneficial to further protect the input signal when there is no sound feedback while reducing the calculation. For example, when the amplifier of the electronic device has no output signal, at this time No acoustic feedback is formed, and no howling occurs, so howling suppression processing is not required.
  • the to-be-output signal in the loudspeaker not only satisfies the requirement that the maximum spectral amplitude exceeds the preset threshold, but also meets other preset requirements before performing howling suppression processing.
  • the average spectral amplitude of the to-be-output signal exceeds the preset threshold.
  • the threshold is set, which is not limited in this embodiment of the present application.
  • the first electronic device and the second electronic device may be terminal devices, walkie-talkies and other devices. If the user of the first electronic device and the user of the second electronic device are in a voice call, the voice signal received by the microphone of the first electronic device is processed by the processor and then transmitted to the second electronic device. The output from the loudspeaker and the voice signal received by the microphone of the second electronic device are processed in the same way.
  • the loudspeaker of the first electronic device does not need to output a signal, that is, for the first electronic device, in this case, there is no acoustic feedback and no whistling occurs. Call.
  • the processor of the first electronic device will perform howling suppression processing on the voice signal input by the microphone, this will affect the audio quality of the voice signal to a certain extent. Therefore, if the spectral amplitude of the signal to be output does not exceed the preset threshold, it means that the amplitude of the signal output by the loudspeaker is small or almost no. Howling suppression processing.
  • By limiting howling suppression processing to be performed when the maximum spectral amplitude of the signal to be output in the loudspeaker exceeds a preset threshold it is beneficial to further protect the input signal during silent feedback while reducing computation.
  • the electronic device receives the first signal input by the microphone.
  • the electronic device acquires a spectrum signal corresponding to the first signal.
  • the electronic device determines a howling detection threshold based on a first spectral amplitude, where the first spectral amplitude is the maximum spectral amplitude of the spectral signal within a preset reference frequency band.
  • the electronic device determines a howling spectral component in the spectral signal based on the howling detection threshold.
  • the electronic device determines a suppression parameter based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component.
  • the electronic device performs filtering processing on the spectrum signal based on the suppression parameter.
  • steps 802 to 807 are the same as the specific implementation manners of the above-mentioned steps 301 to 306, and are not repeated here.
  • FIG. 10 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.
  • the electronic device may be a terminal device, an audio system, or the like.
  • the electronic device 100 includes a processor 1010 , a loudspeaker 1020 , a microphone 1030 and a memory 1040 .
  • the processor 1010 can be a central processing unit (Central Processing Unit, CPU), and the processor 1010 can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC) ), off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • a general-purpose processor can be a microprocessor, alternatively, the processor 1010 can also be any conventional processor or the like.
  • Memory 1040 may include read-only memory and random access memory, and provides instructions and data to processor 1010 .
  • a portion of memory 1040 may also include non-volatile random access memory.
  • the electronic device 100 may further include a device other than the device described above, such as a communication interface, which is not limited in this embodiment of the present application.
  • the processor 1010 is configured to invoke the program instructions stored in the memory 1040 .
  • a loudspeaker 1020 for outputting the processed audio signal is provided.
  • the microphone 1030 is used for acquiring audio signals.
  • Memory 1040 for storing program instructions.
  • the processor 1010 invokes the program instructions stored in the memory 1040 to make the electronic device 100 perform the following operations: receive the first signal input by the microphone 1030; acquire a spectrum signal corresponding to the first signal; determine a howling detection threshold based on the first spectrum amplitude , the first spectral amplitude is the maximum spectral amplitude of the spectral signal in the preset reference frequency band; the howling spectral component in the spectral signal is determined based on the howling detection threshold; the howling spectral component is determined based on the howling detection threshold and the howling spectral component
  • the corresponding spectral amplitude determines a suppression parameter; the spectral signal is filtered based on the suppression parameter.
  • FIG. 11 shows an apparatus 110 provided by an embodiment of the present application, which is used to implement the functions of the electronic device in the foregoing FIG. 3 and FIG. 8 .
  • the apparatus may be an electronic device or an apparatus for an electronic device.
  • the apparatus for an electronic device may be a system-on-a-chip or a chip within the electronic device. Wherein, the chip system may be composed of chips, and may also include chips and other discrete devices.
  • the 11 may include a howling detection module 1101 and a howling filtering module 1102, wherein: the howling detection module 1101 is used to receive the first signal input by the microphone; the howling detection module 1101 is also used to obtain the the frequency spectrum signal corresponding to the first signal; the howling detection module 1101 is further configured to determine the howling detection threshold based on the first frequency spectrum amplitude, the first frequency spectrum amplitude being the maximum frequency spectrum amplitude of the frequency spectrum signal in the preset reference frequency band; howling The detection module 1101 is further configured to determine the howling spectral component in the spectrum signal based on the howling detection threshold; the howling detection module 1101 is also used to determine the howling spectral component based on the howling detection threshold and the corresponding spectral amplitude of the howling spectral component Determine the suppression parameter; the howling filtering module 1102 is used for filtering the spectrum signal based on the suppression parameter.
  • the howling detection module 1101 is used to receive the first signal input by the microphone; the howling detection
  • the howling detection module 1101 is further configured to determine that the first spectral component is the howling spectral component if the spectral amplitude corresponding to the first frequency point is greater than the howling detection threshold, and the first spectral component is the howling spectral component.
  • a frequency point is any one or more frequency points of the spectrum signal.
  • the howling detection module 1101 is further configured to determine the howling detection threshold corresponding to each frequency point based on the first frequency spectrum amplitude and the first adjustment function.
  • the howling detection module 1101 is further configured to determine that the first frequency Howling spectrum components, the first frequency point is any one or more frequency points of the spectrum signal.
  • the howling detection module 1101 is further configured to determine the suppression parameter corresponding to the howling spectral component based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component;
  • the howling detection module 1101 is further configured to determine the suppression corresponding to the howling spectral component based on the howling detection threshold, the spectral amplitude corresponding to the howling spectral component, and the second adjustment function parameter.
  • the apparatus further includes a framing module and a frequency domain analysis module.
  • the framing module is used to perform frame processing on the first signal to obtain multiple time domain signals; the frequency domain analysis and conversion module is used to convert the multiple time domain signals into multiple spectrum signals.
  • the howling detection module 1101 is further configured to determine that the maximum spectral amplitude of the signal to be output in the loudspeaker exceeds a preset threshold.
  • each module/unit included in each device and product described in the above-mentioned embodiments it may be a software module/unit, a hardware module/unit, or a part of a software module/unit and a part of a hardware module/unit .
  • each module/unit included in the product may be implemented by hardware such as a circuit, or at least some modules/units may be implemented by a software program, and the software program runs Since the processor is integrated inside the chip, the remaining (if any) modules/units can be implemented by hardware such as circuits; for each device and product applied to or integrated in the chip module, each module/unit included can be implemented using It is realized by hardware such as circuits, and different modules/units can be located in the same piece of the chip module (such as chips, circuit modules, etc.) or in different components, or at least some modules/units can be realized by software programs.
  • the remaining (if any) part of the modules/units can be implemented by hardware such as circuits; for each device and product applied to or integrated in the terminal, the modules/units contained therein can be all It is implemented by means of hardware such as circuits, and different modules/units may be located in the same component (eg, chip, circuit module, etc.) or in different components in the terminal, or at least some modules/units may be implemented by means of software programs.
  • the program runs on the processor integrated inside the terminal, and the remaining (if any) part of the modules/units can be implemented by hardware such as circuits.
  • Embodiments of the present application further provide a chip, and the chip can execute the relevant steps of the electronic device in the foregoing method embodiments.
  • This chip is used for:
  • a microphone receiving a first signal input by a microphone; acquiring a spectrum signal corresponding to the first signal; determining a howling detection threshold based on a first spectrum amplitude, where the first spectrum amplitude is the maximum spectrum amplitude of the spectrum signal in a preset reference frequency band;
  • the howling detection threshold determines the howling spectral component in the spectrum signal; the suppression parameter is determined based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component; and the spectral signal is filtered based on the suppression parameter.
  • the chip is further configured to determine that the first frequency spectrum component is the howling spectrum component if the spectrum amplitude corresponding to the first frequency point is greater than the howling detection threshold, and the first frequency point is any one or more frequency points of the spectral signal.
  • the chip is further configured to determine the howling detection threshold corresponding to each frequency point based on the first spectrum amplitude and the first adjustment function.
  • the chip is further configured to determine that the first frequency point is the howling spectrum component if the spectrum amplitude corresponding to the first frequency point is greater than the howling detection threshold corresponding to the first frequency point , the first frequency point is any one or more frequency points of the spectrum signal.
  • the chip is further configured to determine the suppression parameter corresponding to the howling spectral component based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component;
  • the suppression parameter corresponding to other frequency points is 1.
  • the chip is further configured to determine the suppression parameter corresponding to the howling spectral component based on the howling detection threshold, the spectral amplitude corresponding to the howling spectral component, and the second adjustment function.
  • the first signal is subjected to frame-by-frame processing to obtain multiple time-domain signals; the multiple time-domain signals are converted into multiple spectrum signals.
  • the chip is further configured to determine that the maximum spectral amplitude of the signal to be output in the loudspeaker exceeds a preset threshold.
  • FIG. 12 is a schematic structural diagram of a module device provided by an embodiment of the present application.
  • the module device 120 can perform the relevant steps of the terminal device in the foregoing method embodiments, and the module device 120 includes: a communication module 1201 , a power module 1202 , a storage module 1203 , and a chip module 1204 .
  • the power module 1202 is used to provide power for the module device;
  • the storage module 1203 is used to store data and instructions;
  • the communication module 1201 is used to perform internal communication of the module device, or to The module device communicates with external devices;
  • the chip module 1204 is used for:
  • a microphone receiving a first signal input by a microphone; acquiring a spectrum signal corresponding to the first signal; determining a howling detection threshold based on a first spectrum amplitude, where the first spectrum amplitude is the maximum spectrum amplitude of the spectrum signal in a preset reference frequency band;
  • the howling detection threshold determines the howling spectral component in the spectrum signal; the suppression parameter is determined based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component; and the spectral signal is filtered based on the suppression parameter.
  • the chip module 1204 is further configured to determine that the first spectral component is the howling spectral component if the spectral amplitude corresponding to the first frequency point is greater than the howling detection threshold, and the first spectral component is the howling spectral component.
  • the frequency point is any one or more frequency points of the spectrum signal.
  • the chip module 1204 is further configured to determine the howling detection threshold corresponding to each frequency point based on the first frequency spectrum amplitude and the first adjustment function.
  • the chip module 1204 is further configured to determine that the first frequency point is the howling if the spectrum amplitude corresponding to the first frequency point is greater than the howling detection threshold corresponding to the first frequency point It is called spectrum component, and the first frequency point is any one or more frequency points of the spectrum signal.
  • the chip module 1204 is further configured to determine the suppression parameter corresponding to the howling spectral component based on the howling detection threshold and the spectral amplitude corresponding to the howling spectral component;
  • the suppression parameter corresponding to the frequency points other than the spectral components is 1.
  • the chip module 1204 is further configured to determine the suppression parameter corresponding to the howling spectral component based on the howling detection threshold, the spectral amplitude corresponding to the howling spectral component, and the second adjustment function .
  • the chip module 1204 is further configured to perform frame division processing on the first signal to obtain multiple time domain signals; and convert the multiple time domain signals into multiple spectrum signals.
  • the chip module 1204 is further configured to determine that the maximum spectral amplitude of the signal to be output in the loudspeaker exceeds a preset threshold.
  • Embodiments of the present application further provide a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when the computer-readable storage medium is executed on a processor, the method flow of the foregoing method embodiment is implemented.
  • the embodiments of the present application further provide a computer program product, when the computer program product runs on a processor, the method flow of the above method embodiments is realized.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Computational Linguistics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Multimedia (AREA)
  • General Health & Medical Sciences (AREA)
  • Otolaryngology (AREA)
  • Circuit For Audible Band Transducer (AREA)
  • Telephone Function (AREA)

Abstract

一种抑制啸叫的方法、装置、芯片及模组设备，该方法包括：接收传声器输入的第一信号（301）；获取该第一信号对应的频谱信号（302）；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度（303）；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分（304）；基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数（305）；基于该抑制参数对该频谱信号进行滤波处理（306）。采用该方法，有利于抑制多频啸叫和宽频啸叫。

Description

一种抑制啸叫的方法、装置、芯片及模组设备 技术领域
本发明涉及通信领域，尤其涉及一种抑制啸叫的方法、装置、芯片及模组设备。
背景技术
在通常情况下，手机、音响等电子设备都会应用到扩声系统。扩声系统能够对音量较小的语音信号进行处理，从而输出音量更大的语音信号。如图1所示，在扩声系统10中，传声器接收到外界的语音信号后，将该语音信号转化为输入信号，使该输入信号经功率放大器放大后由扩音器输出，输出信号经边界反射、折射，重新被传声器拾取，形成了一个闭环回路，这个过程称为声反馈。根据奈奎斯特稳定性判据，当某频点的开环传递函数的幅度大于等于1，且该扩声系统的开环传递函数的相位为2π的整数倍时，系统是不稳定的，典型的不稳定现象是产生啸叫。啸叫的产生会对扩声系统造成一定的危害，例如，影响扩音器的扩音效果，限制扩声系统的最大稳定增益(Maximum Stable Gain,MSG)，在最严重的情况下，甚至也可能损坏整个扩音系统。针对啸叫抑制的研究已有50多年的历史，先后经历了手动控制，模拟自动控制和数字自动控制等发展阶段。目前主要的方法有：陷波法、移频法和自适应反馈抵消法等。
其中，陷波法在实际中应用最为广泛。陷波法主要由两部分组成，第一部分为啸叫检测，第二部分为陷波器设计。陷波法抑制啸叫的基本原理是通过准确检测声反馈中出现的啸叫频点并进行陷波处理，降低啸叫频点处增益，破坏啸叫产生的增益条件，达到抑制啸叫的目的。由于陷波法抑制啸叫对啸叫检测高度依赖，而现有的啸叫检测方法难以快速且准确地检出同时出现两个以上啸叫频点的啸叫或宽频啸叫，因此陷波法的实际应用效果并不理想。
发明内容
本申请提供一种抑制啸叫的方法、装置、芯片及模组设备，有利于抑制多频啸叫和宽频啸叫。
第一方面，本申请提供一种抑制啸叫的方法，该方法包括：接收传声器输入的第一信 号；获取该第一信号对应的频谱信号；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；基于该抑制参数对该频谱信号进行滤波处理。
基于第一方面所描述的方法，由于通过啸叫检测阈值对频谱信号中除预设频带外所有频点均进行检测，因此，该方法能够同时检测出该频谱信号中多个频点的啸叫或宽频啸叫，有利于抑制多频啸叫和宽频啸叫。
在一种可能的实现方式中，若第一频点对应的频谱幅度大于该啸叫检测阈值，则确定该第一频谱成分为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，基于该第一频谱幅度和第一调节函数确定各个频点所对应的该啸叫检测阈值。
在一种可能的实现方式中，若第一频点对应的频谱幅度大于该第一频点对应的该啸叫检测阈值，则确定该第一频点为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定该啸叫频谱成分对应的该抑制参数；确定除该啸叫频谱成分以外的频点对应的该抑制参数为1。
在一种可能的实现方式中，基于该啸叫检测阈值、该啸叫频谱成分所对应的频谱幅度和第二调节函数确定该啸叫频谱成分对应的该抑制参数。
在一种可能的实现方式中，对该第一信号进行分帧处理，得到多个时域信号；将该多个时域信号转化为多个频谱信号。
在一种可能的实现方式中，确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
第二方面，本申请提供一种抑制啸叫装置，该装置包括啸叫检测模块和啸叫滤波模块。该啸叫检测模块，用于：接收传声器输入的第一信号；获取该第一信号对应的频谱信号；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；基于该啸叫检测 阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；该啸叫滤波模块，用于基于该抑制参数对该频谱信号进行滤波处理。
第三方面，本申请提供一种电子设备，该电子设备包括扩声器、传声器、存储器和处理器：该扩声器，用于输出处理后的音频信号；该传声器，用于获取音频信号；该存储器，用于存储计算机程序；该处理器，具体用于从该存储器中调用该计算机程序执行上述第一方面所描述的方法。
第四方面，本申请提供一种芯片，该芯片，用于：接收传声器输入的第一信号；获取该第一信号对应的频谱信号；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；基于该抑制参数对该频谱信号进行滤波处理。
第五方面，本申请提供一种模组设备，该模组设备包括电源模组、存储模组以及芯片模组。该电源模组用于为该模组设备提供电能；该存储模组用于存储数据和指令；该芯片模组用于：接收传声器输入的第一信号；获取该第一信号对应的频谱信号；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；基于该抑制参数对该频谱信号进行滤波处理。
第六方面，本申请提供一种存储介质，其上存储有计算机指令，该计算机指令运行时执行上述第一方面所描述的方法。
第七方面，本申请提供一种计算机程序或计算机程序产品，包括代码或指令，当代码或指令在计算机上运行时，使得计算机执行如第一方面所描述的方法。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是现有的一种扩声系统的结构示意图；
图2是本申请实施例提供的一种扩声系统的结构示意图；
图3是本申请实施例提供的一种抑制啸叫方法的流程示意图；
图4是本申请实施例提供的一种频谱信号的示意图；
图5是本申请实施例提供的又一种频谱信号的示意图；
图6是本申请实施例提供的又一种频谱信号的示意图；
图7是本申请实施例提供的一种语音信号的啸叫处理前后对比图；
图8是本申请实施例提供的又一种抑制啸叫方法的流程示意图；
图9是本申请实施例提供的一种的声反馈场景示意图；
图10是本申请实施例提供的一种电子设备的结构示意图；
图11是本申请实施例提供的一种装置的结构示意图；
图12是本申请实施例提供的一种模组设备的结构示意图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
请参阅图2，图2为本申请实施例提供的一种扩声系统的结构示意图。在不考虑声反馈时，近端语音信号201被传声器202拾取得到输入信号203(205)，经过功放的前向增益单元206作用后得到放大后的信号207，再经过扩声器208播放后得到放大后的输出信号209，输出信号209相比语音信号201更容易被人耳接收到。当考虑声反馈时，放大后的输出信号209经声反馈回路210形成反馈语音信号211，反馈语音信号211被传声器202拾取形成了一个闭环回路。当功放的前向增益206和声反馈回路210满足奈奎斯特准则时，系统会趋向不稳定，产生啸叫。图2中的扩声系统20，在原有的扩声系统10中加入啸叫检测模块212和滤波系统模块204，可以对含有啸叫的输入信号203进行过滤，从而输出不具有啸叫的输出信号209。啸叫检测模块212对传声器202的输入信号203进行检测，若检测到啸叫成分，则输出抑制参数213给滤波系统204，输入信号203经滤波系统204 滤波后输出啸叫抑制后的信号205，经前向增益单元206后得到放大信号207，再经扩声器208播放得到输出语音209。若未检测到啸叫成分，则输出指令或相应的抑制参数213给滤波系统204，不对输入信号203做滤波，此时前向增益206的输入信号就是传声器202的输入信号203。
请参阅图3，图3是本申请实施例提供的一种啸叫抑制方法的流程示意图。该方法应用于电子设备或电子设备中的芯片，具体的，如图3所示，图3以电子设备为方法的执行主体为例进行说明。本发明实施例的其他附图所示的啸叫抑制方法的执行主语同理，后文不再赘述。本申请实施例的啸叫抑制方法步骤301～步骤306：
301、电子设备接收传声器输入的第一信号。
本申请实施例中，电子设备可以为手机、音响或对讲机等包含有传声器和扩声器的设备。该电子设备能够放大传声器输入的语音信号从而使扩声器输出，有利于用户能够更清晰地听到该语音信号。
本申请实施例中，该第一信号可以为传声器从外界接收到的语音信号，也可以是传声器通过通信接口或存储器中获取到的音频信号等，本申请实施例对此不作限定。
302、电子设备获取该第一信号对应的频谱信号。
本申请实施例中，步骤302具体可以包括：电子设备对该第一信号进行分帧处理，得到多个时域信号；电子设备将该多个时域信号转化为多个频谱信号。其中，该分帧处理要求每秒不少于40帧。通过分帧处理有利于使后续更精确地检测啸叫频谱成分。不同分帧的时域信号所对应转换的频谱信号不同，不同频谱信号可以通过采用不同的分帧标识进行区分。
可选的，电子设备也可以直接将第一信号转换为频谱信号，无需先获取时域信号，本申请实施例对获取第一信号对应的频谱信号的方式不作限定。
303、电子设备基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度。
本申请实施例中，预设基准频带通常为啸叫频点出现概率较低的低频带。因为选择啸叫频点出现概率较低的低频带，可以避免出现该信号中啸叫频点所对应频谱幅度小于啸叫检测阈值，导致无法过滤掉信号中的啸叫的情况。其中，该预设基准频带可以通过两种方 式确定，一种方式为，仅确定一个截止频率f c，预设基准频带表示为(0,f c)；另一种方式为，确定一个初始频率f s和截止频率f c，则该预设基准频带可以表示为(f s,f c)。本申请实施例对如何确定预设基准频带的方式不作限定。
本申请实施例中，第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度。采用一个实际的例子对第一频谱幅度进行介绍，如图4所示，图4为本申请实施例提供的一个频谱信号的频谱幅度的示意图。纵轴A表示频谱信号的频谱幅度，横轴f i表示频谱信号的频点，实线表示频谱信号的频谱幅度A(k,f i)，k为分帧标识。设定预设基准频带为(f s,f c)，则根据(f s,f c)所对应的频谱幅度的大小，可见，f 0所对应的A(k,f 0)是(f s,f c)所对应的频谱幅度中的最大值，由此可以确定为A(k,f 0)为第一频谱幅度。
在一种可能实现方式中，该啸叫检测阈值为一个固定的值。频谱信号的频谱幅度可以通过公式表示为A(k,f i)，该基准频带为(0,f c)，其中k为分帧标识，f i为频点。该第一频谱幅度Thr_base可以通过下列公式表示：
Thr_base＝max[A(k，f i)]  f i≤f c  (1)
其中，啸叫检测阈值的大小等于该第一频谱幅度。
在另一种可能的实现方式中，该啸叫检测阈值可以为与频率有关的分段阈值。此时各个频点所述对应的该啸叫检测阈值可以基于该第一频谱幅度和第一调节函数确定。频谱信号的频谱幅度为A(k,f i)，该基准频带为(0,f c)，第一调节函数为a(k,f i),第一频谱幅度为Thr_base。则该啸叫检测阈值Thr(k,f i)可以通过下列公式表示：
Thr(k，f i)＝a(k，f i)×Thr_base  (2)
其中，第一调节函数a(k,f i)需满足下列条件：
Figure PCTCN2021143642-appb-000001
通过该实现方式，电子设备可以在基于不同频率的情况下，设定不同的啸叫检测阈值的大小，有利于能够更精确地检测出啸叫频谱成分。
304、电子设备基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分。
本申请实施例中，该啸叫频谱成分可以为该频谱信号中的一个啸叫频点，也可以为多个啸叫频点，也可以为一段啸叫频带，本申请实施例对此不作限定。
在一种可能的实现方式中，啸叫检测阈值为一个固定的值时，该步骤304的具体实现步骤为：若第一频点对应的频谱幅度大于该啸叫检测阈值，则确定该第一频谱成分为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。通过将该啸叫检测阈值与所有频点对应的频谱幅度进行比较，确定出啸叫频谱成分，这样的方式能够有利于同时检测出多个频点的啸叫或宽频啸叫。
例如，如图5所示，纵轴A表示频谱信号的频谱幅度，横轴f i表示频谱信号的频点，实线表示频谱信号的频谱幅度A(k,f i)，k为分帧标识，(f s,f c)为预设基准频带，啸叫检测阈值为Thr_base。将Thr_base与各个频点所对应的频谱幅度进行比较，可见(f 1,f 2)内的频点所对应的频谱幅度大于Thr_base，例如，f 3所对应的A(k,f 3)即大于Thr_base。因此，即可确定频带(f 1,f 2)为啸叫频谱成分。
在另一种可能的实现方式中，啸叫检测阈值为与频率有关的分段阈值时，该步骤304的具体实现步骤为：若第一频点对应的频谱幅度大于该第一频点对应的该啸叫检测阈值，则确定该第一频点为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。通过将所有频点对应的该啸叫检测阈值和频谱幅度进行比较，确定出啸叫频谱成分，这样的方式能够有利于同时检测出多个频点的啸叫或宽频啸叫。除此以外，由于啸叫检测阈值为与频率有关的分段阈值，能够更加精确地检测出频谱信号中的啸叫频谱成分。
例如，如图6所示，纵轴A表示频谱幅度，横轴f i表示频点，实线A(k,f i)表示频谱信号的频谱幅度，虚线Thr(k,f i)表示各个频点所对应的啸叫检测阈值。将啸叫检测阈值Thr(k,f i)与该频谱信号的频谱幅度进行比较，可见(f 1,f 2)内的频点所对应的频谱幅度大于(f 1,f 2)内的频点所对应啸叫检测阈值，例如，f 3所对应的A(k,f 3)即大于Thr(k,f 3)。因此，即可确定频带(f 1,f 2)为啸叫频谱成分。
305、电子设备基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数。
本申请实施例中，步骤305的具体实现方式为：基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定该啸叫频谱成分对应的该抑制参数；确定除该啸叫频谱成分以外的频点对应的该抑制参数为1。一个频点所对应的抑制参数为1即表示该频点不是啸叫频点，且在滤波处理中，不会对频谱信号中的该频点进行处理。如果所有的频点所对应的抑 制参数均为1，则说明该频谱信号中不存在啸叫频谱成分，该信号无需执行滤波处理。
其中，该抑制参数G(k,f i)可以通过公式(4)或公式(5)来表示：
Figure PCTCN2021143642-appb-000002
Figure PCTCN2021143642-appb-000003
其中，公式(4)为当啸叫检测阈值为一个固定的值时，该抑制参数的表示方式；公式(5)为当啸叫检测阈值为与频率有关的分段阈值时，该抑制参数的表示方式。
可选的，该抑制参数还可以基于该啸叫检测阈值、该啸叫频谱成分所对应的频谱幅度和第二调节函数确定。若将第二调节函数表示为β(k,f i)，该第二调节函数β(k,f i)满足0≤β(k,f i)≤1，则上述公式(4)和公式(5)分别可以替换表示为公式(6)和公式(7)来表示：
Figure PCTCN2021143642-appb-000004
Figure PCTCN2021143642-appb-000005
其中，公式(6)为当啸叫检测阈值为一个固定的值时，该抑制参数的表示方式；公式(7)为当啸叫检测阈值为与频率有关的分段阈值时，该抑制参数的表示方式。通过该实现方式所确定出的抑制参数，有利于更精确过滤掉啸叫。
306、电子设备基于该抑制参数对该频谱信号进行滤波处理。
本申请实施例中，电子设备基于该抑制参数对该频谱信号进行滤波处理，滤波处理后得到啸叫抑制后的频谱信号。则该啸叫抑制后的频谱信号A s(k,f i)可以表示为：
A s(k，f i)＝G(k，f i)×A(k，f i)  (8)
通过上述所描述的步骤301～步骤306所描述的方法，可见，由于通过啸叫检测阈值对频谱信号中除预设频带外所有频点均进行检测，因此，该方法能够同时检测出该频谱信号中多个频点的啸叫或宽频啸叫，有利于抑制多频啸叫和宽频啸叫。
例如，如图7所示，图7中上图为啸叫处理前的语音信号，采样率为8000Hz，时长约6.7s，其中主要啸叫成分已由虚线框标示。可以看出啸叫成分呈多频或宽频分布，最大带宽不低于1000Hz，其最大幅度与2000Hz以下近端语音部分的平均幅度接近。图7中下图为经过本申请所提出的啸叫抑制方法所处理后的语音信号，可以看出，所有啸叫成分都得到了几乎无延迟地有效抑制。正因为啸叫被有效抑制了，啸叫处理前的输入语音信号里才 没有出现持续性的啸叫成分。经过啸叫抑制处理后，啸叫成分以外的近端语音部分未出现明显压制或失真，全部频谱能量的均方根平均值仅降低约1dB，说明本方法在快速、有效地抑制啸叫的同时，不易引起音质恶化，同时对系统增益的限制较小。
请参见图8，图8是本申请实施例提供的另一种啸叫抑制方法的流程示意图。该啸叫抑制方法包括步骤801～步骤807。其中：
801、电子设备确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
本申请实施例中，扩音器中的待输出信号的最大频谱幅度超过预设阈值，说明待扩音信号即将输出频谱幅度较大的信号，该输出信号可能会被该电子设备的传声器拾取，形成声反馈，导致出现啸叫。通过限定在该情况下执行啸叫抑制处理，有利于在减少计算的同时，进一步实现对无声反馈时的输入信号的保护，例如，当该电子设备的扩音器没有输出信号的时候，此时没有形成声反馈，也不会出现啸叫，因此无需进行啸叫抑制处理。
其中，该扩音器中的待输出信号除了满足最大频谱幅度超过预设阈值这个要求，还可以满足其他预设要求再执行啸叫抑制处理，例如，该待输出的信号的平均频谱幅度超过预设阈值，本申请实施例对此不作限定。
其中，该实现方式非常适用于移动通信终端应用场景。如图9所示，第一电子设备和第二电子设备可以为终端设备、对讲机等设备。若该第一电子设备和第二电子设备的用户正在进行语音通话，此时第一电子设备的传声器接收到的语音信号经过处理器处理后，传输到第二电子设备，通过第二电子设备的扩声器输出，第二电子设备的传声器接收到的语音信号的处理方式同理。在该场景下，若第二电子设备的用户未发出声音，则第一电子设备的扩声器无需输出信号，即对于第一电子设备而言，在该情况下没有声反馈，不会发生啸叫。但若第一电子设备的用户正在讲话，且第一电子设备的处理器将对传声器输入的语音信号进行啸叫抑制处理，这会对该语音信号的音频质量造成一定影响。因此，若该待输出的信号的频谱幅度未超过预设阈值，则表示此时该扩音器输出的信号幅度较小或几乎没有，在该情况下没有声反馈，无需对输入的语音信号进行啸叫抑制处理。通过限定在扩音器中的待输出信号的最大频谱幅度超过预设阈值情况下执行啸叫抑制处理，有利于在减少计算的同时，进一步实现对无声反馈时的输入信号的保护。
802、电子设备接收传声器输入的第一信号。
803、电子设备获取该第一信号对应的频谱信号。
804、电子设备基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度。
805、电子设备基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分。
806、电子设备基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数。
807、电子设备基于该抑制参数对该频谱信号进行滤波处理。
其中，步骤802～步骤807的具体实现方式与上述步骤301～步骤306的具体实现方式相同，在此不赘述。
请参阅图10，图10是本申请实施例提供的一种电子设备的结构示意图，该电子设备可以为终端设备、音响等。该电子设备100中包括处理器1010、扩音器1020、传声器1030和存储器1040。
处理器1010可以是中央处理单元(Central Processing Unit，CPU)，该处理器1010还可以是其他通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现成可编程门阵列(Field-Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器，可选的，该处理器1010也可以是任何常规的处理器等。
存储器1040可以包括只读存储器和随机存取存储器，并向处理器1010提供指令和数据。存储器1040的一部分还可以包括非易失性随机存取存储器。
可选的，该电子设备100还可以包括除上述所描述的器件以外的器件，例如通信接口，本申请实施例对此不作限定。
其中：
处理器1010，用于调用存储器1040中存储的程序指令。
扩音器1020，用于输出处理后的音频信号。
传声器1030，用于获取音频信号。
存储器1040，用于存储程序指令。
处理器1010调用存储器1040中存储的程序指令，使该电子设备100执行以下操作：接收传声器1030输入的第一信号；获取该第一信号对应的频谱信号；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；基于该抑制参数对该频谱信号进行滤波处理。
如图11所示为本申请实施例提供的一种装置110，用于实现上述图3和图8中电子设备的功能。该装置可以是电子设备或用于电子设备的装置。用于电子设备的装置可以为电子设备内的芯片系统或芯片。其中，芯片系统可以由芯片构成，也可以包含芯片和其他分立器件。图11所示的装置110可以包括啸叫检测模块1101和啸叫滤波模块1102，其中：啸叫检测模块1101，用于接收传声器输入的第一信号；啸叫检测模块1101，还用于获取该第一信号对应的频谱信号；啸叫检测模块1101，还用于基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；啸叫检测模块1101，还用于基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；啸叫检测模块1101，还用于基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；啸叫滤波模块1102，用于基于该抑制参数对该频谱信号进行滤波处理。
在一种可能的实现方式中，啸叫检测模块1101，还用于若第一频点对应的频谱幅度大于该啸叫检测阈值，则确定该第一频谱成分为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，啸叫检测模块1101，还用于基于该第一频谱幅度和第一调节函数确定各个频点所对应的该啸叫检测阈值。
在一种可能的实现方式中，啸叫检测模块1101，还用于若第一频点对应的频谱幅度大于该第一频点对应的该啸叫检测阈值，则确定该第一频点为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，啸叫检测模块1101，还用于基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定该啸叫频谱成分对应的该抑制参数；
确定除该啸叫频谱成分以外的频点对应的该抑制参数为1。
在一种可能的实现方式中，啸叫检测模块1101，还用于基于该啸叫检测阈值、该啸叫 频谱成分所对应的频谱幅度和第二调节函数确定该啸叫频谱成分对应的该抑制参数。
在一种可能的实现方式中，该装置还包括分帧模块和频域分析模块。其中分帧模块用于对该第一信号进行分帧处理，得到多个时域信号；频域分析转换模块用于将该多个时域信号转化为多个频谱信号。
在一种可能的实现方式中，啸叫检测模块1101，还用于确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
关于上述实施例中描述的各个装置、产品包含的各个模块/单元，其可以是软件模块/单元，也可以是硬件模块/单元，或者也可以部分是软件模块/单元，部分是硬件模块/单元。例如，对于应用于或集成于芯片的各个装置、产品其包含的各个模块/单元可以都采用电路等硬件的方式实现，或者，至少部分模块/单元可以采用软件程序的方式实现，该软件程序运行于芯片内部集成处理器，剩余的(如果有)部分模块/单元可以采用电路等硬件方式实现；对于应用于或集成于芯片模组的各个装置、产品，其包含的各个模块/单元可以都采用电路等硬件的方式实现，不同模块/单元可以位于芯片模组的同一件(例如芯片、电路模块等)或者不同组件中，或者，至少部分模块/单元可以采用软件程序的方式实现，该软件程序运行于芯片模组内部集成的处理器，剩余的(如果有)部分模块/单元可以采用电路等硬件方式实现；对于应用于或集成于终端的各个装置、产品，其包含的模块/单元可以都采用电路等硬件的方式实现，不同的模块/单元可以位于终端内同一组件(例如，芯片、电路模块等)或者不同组件中，或者，至少部分模块/单元可以采用软件程序的方式实现，该软件程序运行于终端内部集成的处理器，剩余的(如果有)部分模块/单元可以采用电路等硬件方式实现。
本申请实施例还提供一种芯片，该芯片可以执行前述方法实施例中电子设备的相关步骤。该芯片用于：
接收传声器输入的第一信号；获取该第一信号对应的频谱信号；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；基于该抑制参数对该频谱信号进行滤波处理。
在一种可能的实现方式中，该芯片还用于若第一频点对应的频谱幅度大于该啸叫检测 阈值，则确定该第一频谱成分为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，该芯片还用于基于该第一频谱幅度和第一调节函数确定各个频点所对应的该啸叫检测阈值。
在一种可能的实现方式中，该芯片还用于若第一频点对应的频谱幅度大于该第一频点对应的该啸叫检测阈值，则确定该第一频点为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，该芯片还用于基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定该啸叫频谱成分对应的该抑制参数；确定除该啸叫频谱成分以外的频点对应的该抑制参数为1。
在一种可能的实现方式中，该芯片还用于基于该啸叫检测阈值、该啸叫频谱成分所对应的频谱幅度和第二调节函数确定该啸叫频谱成分对应的该抑制参数。
在一种可能的实现方式中，对该第一信号进行分帧处理，得到多个时域信号；将该多个时域信号转化为多个频谱信号。
在一种可能的实现方式中，该芯片还用于确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
如图12所示，图12是本申请实施例提供的一种模组设备的结构示意图。该模组设备120可以执行前述方法实施例中终端设备的相关步骤，该模组设备120包括：通信模组1201、电源模组1202、存储模组1203以及芯片模组1204。
其中，所述电源模组1202用于为所述模组设备提供电能；所述存储模组1203用于存储数据和指令；所述通信模组1201用于进行模组设备内部通信，或者用于所述模组设备与外部设备进行通信；所述芯片模组1204用于：
接收传声器输入的第一信号；获取该第一信号对应的频谱信号；基于第一频谱幅度确定啸叫检测阈值，该第一频谱幅度为该频谱信号在预设基准频带内的最大频谱幅度；基于该啸叫检测阈值确定该频谱信号中的啸叫频谱成分；基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定抑制参数；基于该抑制参数对该频谱信号进行滤波处理。
在一种可能的实现方式中，该芯片模组1204还用于若第一频点对应的频谱幅度大于该 啸叫检测阈值，则确定该第一频谱成分为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，该芯片模组1204还用于基于该第一频谱幅度和第一调节函数确定各个频点所对应的该啸叫检测阈值。
在一种可能的实现方式中，该芯片模组1204还用于若第一频点对应的频谱幅度大于该第一频点对应的该啸叫检测阈值，则确定该第一频点为该啸叫频谱成分，该第一频点为该频谱信号的任意一个或多个频点。
在一种可能的实现方式中，该芯片模组1204还用于基于该啸叫检测阈值和该啸叫频谱成分所对应的频谱幅度确定该啸叫频谱成分对应的该抑制参数；确定除该啸叫频谱成分以外的频点对应的该抑制参数为1。
在一种可能的实现方式中，该芯片模组1204还用于基于该啸叫检测阈值、该啸叫频谱成分所对应的频谱幅度和第二调节函数确定该啸叫频谱成分对应的该抑制参数。
在一种可能的实现方式中，该芯片模组1204还用于对该第一信号进行分帧处理，得到多个时域信号；将该多个时域信号转化为多个频谱信号。
在一种可能的实现方式中，该芯片模组1204还用于确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
本申请实施例还提供一种计算机可读存储介质，该计算机可读存储介质中存储有指令，当其在处理器上运行时，上述方法实施例的方法流程得以实现。
本申请实施例还提供一种计算机程序产品，当所述计算机程序产品在处理器上运行时，上述方法实施例的方法流程得以实现。
需要说明的是，对于前述的各方法实施例，为了简单描述，故将其都表述为一系列的动作组合，但是本领域技术人员应该知悉，本申请并不受所描述的动作顺序的限制，因为依据本申请，某些操作可以采用其他顺序或者同时进行。其次，本领域技术人员也应该知悉，说明书中所描述的实施例均属于优选实施例，所涉及的动作和模块并不一定是本申请所必须的。
本申请提供的各实施例的描述可以相互参照，对各个实施例的描述都各有侧重，某个实施例中没有详述的部分，可以参见其他实施例的相关描述。为描述的方便和简洁，例如 关于本申请实施例提供的各装置、设备的功能以及执行的操作可以参照本申请方法实施例的相关描述，各方法实施例之间、各装置实施例之间也可以互相参考、结合或引用。
最后应说明的是：以上各实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述各实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分或者全部技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。

Claims (24)

  1. 一种抑制啸叫的方法，其特征在于，所述方法包括：
    接收传声器输入的第一信号；
    获取所述第一信号对应的频谱信号；
    基于第一频谱幅度确定啸叫检测阈值，所述第一频谱幅度为所述频谱信号在预设基准频带内的最大频谱幅度；
    基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分；
    基于所述啸叫检测阈值和所述啸叫频谱成分所对应的频谱幅度确定所述啸叫频谱成分对应的抑制参数；
    确定除所述啸叫频谱成分以外的频点对应的所述抑制参数为1；
    基于所述抑制参数对所述频谱信号进行滤波处理。
  2. 根据权利要求1所述的方法，其特征在于，所述基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分，包括：
    若第一频点对应的频谱幅度大于所述啸叫检测阈值，则确定所述第一频点为所述啸叫频谱成分，所述第一频点为所述频谱信号的任意一个或多个频点。
  3. 根据权利要求1所述的方法，其特征在于，所述基于第一频谱幅度确定啸叫检测阈值，包括：
    基于所述第一频谱幅度和第一调节函数确定各个频点所对应的所述啸叫检测阈值。
  4. 根据权利要求3所述的方法，其特征在于，所述基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分，包括：
    若第一频点对应的频谱幅度大于所述第一频点对应的所述啸叫检测阈值，则确定所述第一频点为所述啸叫频谱成分，所述第一频点为所述频谱信号的任意一个或多个频点。
  5. 根据权利要求1所述的方法，其特征在于，所述基于所述啸叫检测阈值和所述啸叫频谱成分所对应的频谱幅度确定抑制参数，包括：
    基于所述啸叫检测阈值、所述啸叫频谱成分所对应的频谱幅度和第二调节函数确定所述啸叫频谱成分对应的所述抑制参数。
  6. 根据权利要求1～4中任意一项所述的方法，其特征在于，所述获取所述第一信号对应 的频谱信号，包括：
    对所述第一信号进行分帧处理，得到多个时域信号；
    将所述多个时域信号转化为多个频谱信号。
  7. 根据权利要求1～4中任意一项所述的方法，其特征在于，所述接收传声器输入的第一信号之前，包括：
    确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
  8. 一种抑制啸叫的装置，其特征在于，所述装置包括啸叫检测模块和啸叫滤波模块：
    所述啸叫检测模块，用于：
    接收传声器输入的第一信号；
    获取所述第一信号对应的频谱信号；
    基于第一频谱幅度确定啸叫检测阈值，所述第一频谱幅度为所述频谱信号在预设基准频带内的最大频谱幅度；
    基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分；
    基于所述啸叫检测阈值和所述啸叫频谱成分所对应的频谱幅度确定所述啸叫频谱成分对应的抑制参数；
    确定除所述啸叫频谱成分以外的频点对应的所述抑制参数为1；
    所述啸叫滤波模块，用于基于所述抑制参数对所述频谱信号进行滤波处理。
  9. 根据权利要求8所述的装置，其特征在于，所述基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分，包括：
    若第一频点对应的频谱幅度大于所述啸叫检测阈值，则确定所述第一频点为所述啸叫频谱成分，所述第一频点为所述频谱信号的任意一个或多个频点。
  10. 根据权利要求8所述的装置，其特征在于，所述基于第一频谱幅度确定啸叫检测阈值，包括：
    基于所述第一频谱幅度和第一调节函数确定各个频点所对应的所述啸叫检测阈值。
  11. 根据权利要求10所述的装置，其特征在于，所述基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分，包括：
    若第一频点对应的频谱幅度大于所述第一频点对应的所述啸叫检测阈值，则确定所述 第一频点为所述啸叫频谱成分，所述第一频点为所述频谱信号的任意一个或多个频点。
  12. 根据权利要求8所述的装置，其特征在于，所述基于所述啸叫检测阈值和所述啸叫频谱成分所对应的频谱幅度确定抑制参数，包括：
    基于所述啸叫检测阈值、所述啸叫频谱成分所对应的频谱幅度和第二调节函数确定所述啸叫频谱成分对应的所述抑制参数。
  13. 根据权利要求8～12中任意一项所述的装置，其特征在于，所述获取所述第一信号对应的频谱信号，包括：
    对所述第一信号进行分帧处理，得到多个时域信号；
    将所述多个时域信号转化为多个频谱信号。
  14. 根据权利要求8～13中任意一项所述的装置，其特征在于，在所述啸叫检测模块，用于接收传声器输入的第一信号之前，
    所述啸叫检测模块，还用于确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
  15. 一种电子设备，其特征在于，该电子设备包括扩声器、传声器、存储器和处理器：
    所述扩声器，用于输出处理后的音频信号；
    所述传声器，用于获取音频信号；
    所述存储器，用于存储计算机程序；
    所述处理器，具体用于从所述存储器中调用所述计算机程序执行如权利要求1～8中任一项所述的方法。
  16. 一种芯片，其特征在于，所述芯片，用于：
    接收传声器输入的第一信号；
    获取所述第一信号对应的频谱信号；
    基于第一频谱幅度确定啸叫检测阈值，所述第一频谱幅度为所述频谱信号在预设基准频带内的最大频谱幅度；
    基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分；
    基于所述啸叫检测阈值和所述啸叫频谱成分所对应的频谱幅度确定所述啸叫频谱成分对应的抑制参数；
    确定除所述啸叫频谱成分以外的频点对应的所述抑制参数为1；
    基于所述抑制参数对所述频谱信号进行滤波处理。
  17. 一种模组设备，其特征在于，所述模组设备包括电源模组、存储模组以及芯片模组，其中：
    所述电源模组用于为所述模组设备提供电能；
    所述存储模组用于存储数据和指令；
    所述芯片模组用于：
    接收传声器输入的第一信号；
    获取所述第一信号对应的频谱信号；
    基于第一频谱幅度确定啸叫检测阈值，所述第一频谱幅度为所述频谱信号在预设基准频带内的最大频谱幅度；
    基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分；
    基于所述啸叫检测阈值和所述啸叫频谱成分所对应的频谱幅度确定所述啸叫频谱成分对应的抑制参数；
    确定除所述啸叫频谱成分以外的频点对应的所述抑制参数为1；
    基于所述抑制参数对所述频谱信号进行滤波处理。
  18. 根据权利要求17所述的模组设备，其特征在于，所述基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分，包括：
    若第一频点对应的频谱幅度大于所述啸叫检测阈值，则确定所述第一频点为所述啸叫频谱成分，所述第一频点为所述频谱信号的任意一个或多个频点。
  19. 根据权利要求17所述的模组设备，其特征在于，所述基于第一频谱幅度确定啸叫检测阈值，包括：
    基于所述第一频谱幅度和第一调节函数确定各个频点所对应的所述啸叫检测阈值。
  20. 根据权利要求19所述的模组设备，其特征在于，所述基于所述啸叫检测阈值确定所述频谱信号中的啸叫频谱成分，包括：
    若第一频点对应的频谱幅度大于所述第一频点对应的所述啸叫检测阈值，则确定所述第一频点为所述啸叫频谱成分，所述第一频点为所述频谱信号的任意一个或多个频点。
  21. 根据权利要求17所述的模组设备，其特征在于，所述基于所述啸叫检测阈值和所述啸叫频谱成分所对应的频谱幅度确定抑制参数，包括：
    基于所述啸叫检测阈值、所述啸叫频谱成分所对应的频谱幅度和第二调节函数确定所述啸叫频谱成分对应的所述抑制参数。
  22. 根据权利要求17～21中任意一项所述的模组设备，其特征在于，所述获取所述第一信号对应的频谱信号，包括：
    对所述第一信号进行分帧处理，得到多个时域信号；
    将所述多个时域信号转化为多个频谱信号。
  23. 根据权利要求17～22中任意一项所述的模组设备，其特征在于，在所述芯片模组用于接收传声器输入的第一信号之前，
    所述芯片模组，还用于确定扩音器中的待输出信号的最大频谱幅度超过预设阈值。
  24. 一种存储介质，存储有计算机指令，其特征在于，所述计算机指令运行时执行权利要求1～7任一项所述方法的步骤。
PCT/CN2021/143642 2021-04-14 2021-12-31 一种抑制啸叫的方法、装置、芯片及模组设备 WO2022217978A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110397343.6A CN112802492B (zh) 2021-04-14 2021-04-14 一种抑制啸叫的方法、装置、芯片及模组设备
CN202110397343.6 2021-04-14

Publications (1)

Publication Number Publication Date
WO2022217978A1 true WO2022217978A1 (zh) 2022-10-20

Family

ID=75817084

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/143642 WO2022217978A1 (zh) 2021-04-14 2021-12-31 一种抑制啸叫的方法、装置、芯片及模组设备

Country Status (2)

Country Link
CN (1) CN112802492B (zh)
WO (1) WO2022217978A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112802492B (zh) * 2021-04-14 2021-07-27 展讯通信（上海）有限公司 一种抑制啸叫的方法、装置、芯片及模组设备

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190043467A1 (en) * 2017-08-04 2019-02-07 Cirrus Logic International Semiconductor Ltd. Tone and howl suppression in an anc system
CN109637552A (zh) * 2018-11-29 2019-04-16 河北远东通信系统工程有限公司 一种抑制音频设备啸叫的语音处理方法
CN109658945A (zh) * 2018-12-19 2019-04-19 Oppo广东移动通信有限公司 防止啸叫的方法、装置、电子设备及存储介质
US10284728B1 (en) * 2018-02-21 2019-05-07 Motorola Solutions, Inc. Adaptive proximity thresholds for dynamic howling suppression
CN110536215A (zh) * 2019-09-09 2019-12-03 普联技术有限公司 音频信号处理的方法、装置、计算设置及存储介质
CN111294711A (zh) * 2019-07-18 2020-06-16 展讯通信（上海）有限公司 信号处理方法及装置
CN111800725A (zh) * 2020-05-29 2020-10-20 展讯通信（上海）有限公司 啸叫检测方法及装置、存储介质、计算机设备
CN112802492A (zh) * 2021-04-14 2021-05-14 展讯通信（上海）有限公司 一种抑制啸叫的方法、装置、芯片及模组设备

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5442712A (en) * 1992-11-25 1995-08-15 Matsushita Electric Industrial Co., Ltd. Sound amplifying apparatus with automatic howl-suppressing function
JP3235925B2 (ja) * 1993-11-19 2001-12-04 松下電器産業株式会社 ハウリング抑制装置
JP2005109533A (ja) * 2003-09-26 2005-04-21 Yamaha Corp ハウリング抑制装置
US8996365B2 (en) * 2009-03-19 2015-03-31 Yugengaisya Cepstrum Howling canceller
JP2012175453A (ja) * 2011-02-22 2012-09-10 Sony Corp 音声処理装置、音声処理方法、及びプログラム
CN105228056B (zh) * 2015-10-21 2018-06-19 西安航空学院 一种消除麦克风啸叫的方法及系统
CN109671445A (zh) * 2018-12-28 2019-04-23 广东美电贝尔科技集团股份有限公司 一种音频系统声音啸叫的抑制方法
CN109788400B (zh) * 2019-03-06 2020-12-18 哈尔滨工业大学（深圳） 一种用于数字助听器的神经网络啸叫抑制方法、系统及存储介质
CN110248300B (zh) * 2019-05-16 2021-03-19 苏州茹声电子有限公司 一种基于自主学习的啸叫抑制方法及扩声系统
CN110782910B (zh) * 2019-11-06 2022-05-20 大连理工大学 一种高检出率的啸叫音频检测系统
CN111402911B (zh) * 2019-12-23 2023-01-31 佛山慧明电子科技有限公司 一种啸叫检测与抑制方法
CN111182431A (zh) * 2019-12-27 2020-05-19 中山大学花都产业科技研究院 一种会议扩声系统啸叫抑制方法
CN111583949A (zh) * 2020-04-10 2020-08-25 南京拓灵智能科技有限公司 啸叫抑制的方法、装置和设备
CN112037816B (zh) * 2020-05-06 2023-11-28 珠海市杰理科技股份有限公司 语音信号频域频率的校正、啸叫检测、抑制方法及装置

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190043467A1 (en) * 2017-08-04 2019-02-07 Cirrus Logic International Semiconductor Ltd. Tone and howl suppression in an anc system
US10284728B1 (en) * 2018-02-21 2019-05-07 Motorola Solutions, Inc. Adaptive proximity thresholds for dynamic howling suppression
CN109637552A (zh) * 2018-11-29 2019-04-16 河北远东通信系统工程有限公司 一种抑制音频设备啸叫的语音处理方法
CN109658945A (zh) * 2018-12-19 2019-04-19 Oppo广东移动通信有限公司 防止啸叫的方法、装置、电子设备及存储介质
CN111294711A (zh) * 2019-07-18 2020-06-16 展讯通信（上海）有限公司 信号处理方法及装置
CN110536215A (zh) * 2019-09-09 2019-12-03 普联技术有限公司 音频信号处理的方法、装置、计算设置及存储介质
CN111800725A (zh) * 2020-05-29 2020-10-20 展讯通信（上海）有限公司 啸叫检测方法及装置、存储介质、计算机设备
CN112802492A (zh) * 2021-04-14 2021-05-14 展讯通信（上海）有限公司 一种抑制啸叫的方法、装置、芯片及模组设备

Also Published As

Publication number Publication date
CN112802492B (zh) 2021-07-27
CN112802492A (zh) 2021-05-14

Similar Documents

Publication Publication Date Title
JP4681163B2 (ja) ハウリング検出抑圧装置、これを備えた音響装置、及び、ハウリング検出抑圧方法
CN105612576B (zh) 限制有源噪声消除输出
JP5704470B2 (ja) オーディオ明瞭度増大方法および装置とコンピュータ装置
CA2766196C (en) Apparatus, method and computer program for controlling an acoustic signal
US8085930B2 (en) Communication system
US9672843B2 (en) Apparatus and method for improving an audio signal in the spectral domain
US9491545B2 (en) Methods and devices for reverberation suppression
WO2015080927A1 (en) Detecting nonlinear amplitude processing
US10516941B2 (en) Reducing instantaneous wind noise
WO2022217978A1 (zh) 一种抑制啸叫的方法、装置、芯片及模组设备
US20110137644A1 (en) Decoding speech signals
US9779753B2 (en) Method and apparatus for attenuating undesired content in an audio signal
TW201317983A (zh) 增進語音即時輸出之方法及助聽器
CN107750038B (zh) 音量调节方法、装置、设备及存储介质
US9392365B1 (en) Psychoacoustic hearing and masking thresholds-based noise compensator system
CN113990283A (zh) 降噪处理方法、装置、耳机及计算机可读存储介质
US9313582B2 (en) Hearing aid and method of enhancing speech output in real time
US9961441B2 (en) Near-end listening intelligibility enhancement
CN107170461B (zh) 语音信号处理方法及装置
WO2023040322A1 (zh) 回声消除方法、终端设备及存储介质
US11955939B2 (en) Control device, control method, and recording medium
US11616873B2 (en) Communication device and output sidetone adjustment method thereof
US11804221B2 (en) Audio device and method of audio processing with improved talker discrimination
CN116170721A (zh) 一种音量调节方法、装置、耳机及计算机可读存储介质
CN117116281A (zh) 声学反馈抑制方法、装置、设备及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21936844

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21936844

Country of ref document: EP

Kind code of ref document: A1