CN117116281A - Acoustic feedback suppression method, device, equipment and storage medium - Google Patents
Acoustic feedback suppression method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN117116281A CN117116281A CN202311197516.5A CN202311197516A CN117116281A CN 117116281 A CN117116281 A CN 117116281A CN 202311197516 A CN202311197516 A CN 202311197516A CN 117116281 A CN117116281 A CN 117116281A
- Authority
- CN
- China
- Prior art keywords
- value
- signal
- power
- spectral line
- power spectrum
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 62
- 230000001629 suppression Effects 0.000 title claims abstract description 47
- 230000003595 spectral effect Effects 0.000 claims description 213
- 238000001228 spectrum Methods 0.000 claims description 207
- 238000009499 grossing Methods 0.000 claims description 34
- 230000003044 adaptive effect Effects 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 16
- 238000013507 mapping Methods 0.000 claims description 12
- 230000009466 transformation Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 description 16
- 230000003321 amplification Effects 0.000 description 11
- 238000003199 nucleic acid amplification method Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 5
- 238000009432 framing Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 241000197727 Euscorpius alpha Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000010363 phase shift Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
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
-
- 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
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/26—Pre-filtering or post-filtering
-
- 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
-
- 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
- G10L2021/02082—Noise filtering the noise being echo, reverberation of the speech
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
Abstract
The application discloses an acoustic feedback suppression method, an acoustic feedback suppression device, acoustic feedback suppression equipment and a storage medium, and relates to the technical field of voice processing. The method comprises the following steps: obtaining a residual signal according to an input voice signal and a reference signal of a current frame through a linear filter; the linear filter is used for suppressing an acoustic feedback signal in an input voice signal, and the reference signal is obtained by performing acoustic feedback suppression on the input voice signal of the previous frame; estimating an acoustic feedback signal in the residual signal to obtain a residual feedback signal; and carrying out nonlinear processing on the residual signal according to the residual feedback signal to obtain an output voice signal. The method suppresses the linear part and the nonlinear part of the acoustic feedback signal in the input voice signal, so that the loudspeaker can obtain larger maximum stable gain when playing the output voice signal.
Description
Technical Field
The embodiment of the application relates to the technical field of voice processing, in particular to an acoustic feedback suppression method, an acoustic feedback suppression device, acoustic feedback suppression equipment and a storage medium.
Background
Acoustic feedback refers to positive feedback in the same acoustic environment, where sound emitted by a speaker returns to a microphone and is played back again by the speaker. This feedback may cause howling to occur, which may lead to poor amplification.
In the related art, a linear filtering manner may be used to suppress an acoustic feedback signal from a speaker in a voice signal collected by a speaker, thereby improving MSG (Maximum Stable Gain ) of an acoustic amplification system.
However, when the gain of the sound amplifying system is large, since the nonlinear portion of the acoustic feedback signal in the speech signal is also correspondingly strong, it is difficult for the above method to suppress the acoustic feedback signal effectively.
Disclosure of Invention
The embodiment of the application provides an acoustic feedback suppression method, an acoustic feedback suppression device, acoustic feedback suppression equipment and a storage medium. The technical scheme provided by the embodiment of the application is as follows:
according to an aspect of an embodiment of the present application, there is provided an acoustic feedback suppression method, the method including:
obtaining a residual signal according to an input voice signal and a reference signal of a current frame through a linear filter; the linear filter is used for suppressing an acoustic feedback signal in the input voice signal, and the reference signal is obtained by performing acoustic feedback suppression on the input voice signal of the previous frame;
Estimating an acoustic feedback signal in the residual signal to obtain a residual feedback signal;
and carrying out nonlinear processing on the residual signal according to the residual feedback signal to obtain an output voice signal.
According to an aspect of an embodiment of the present application, there is provided an acoustic feedback suppression apparatus, the apparatus including:
the linear module is used for obtaining a residual signal according to the input voice signal and the reference signal of the current frame through a linear filter; the linear filter is used for suppressing an acoustic feedback signal in the input voice signal, and the reference signal is obtained by performing acoustic feedback suppression on the input voice signal of the previous frame;
the estimation module is used for estimating the acoustic feedback signal in the residual signal to obtain a residual feedback signal;
and the nonlinear module is used for carrying out nonlinear processing on the residual signal according to the residual feedback signal to obtain an output voice signal.
According to an aspect of an embodiment of the present application, there is provided an electronic device including a processor and a memory, the memory storing a computer program loaded and executed by the processor to implement the above-described acoustic feedback suppression method.
According to an aspect of an embodiment of the present application, there is provided a computer-readable storage medium having stored therein a computer program loaded and executed by a processor to implement the above-described acoustic feedback suppression method.
According to an aspect of an embodiment of the present application, there is provided a computer program product comprising a computer program stored in a computer readable storage medium, from which a processor reads and executes the computer program to implement the above-mentioned acoustic feedback suppression method.
The technical scheme provided by the embodiment of the application at least comprises the following beneficial effects:
and suppressing the acoustic feedback signal in the input voice signal of the current frame according to the reference signal obtained by performing acoustic feedback suppression on the input voice signal of the previous frame through a linear filter to obtain a residual signal. And estimating the acoustic feedback signal which is not filtered by the linear filter in the residual signal, so as to obtain the residual feedback signal. Finally, according to the residual feedback signal, the acoustic feedback signal in the residual signal is further suppressed through nonlinear processing, and an output voice signal is obtained. The method suppresses the linear part and the nonlinear part of the acoustic feedback signal in the input voice signal, so that the loudspeaker can obtain larger maximum stable gain when playing the output voice signal.
Drawings
FIG. 1 is a schematic illustration of an implementation environment for an embodiment of the present application;
FIG. 2 is a schematic diagram of an acoustic enhancement system according to an embodiment of the present application;
FIG. 3 is a flow chart of an acoustic feedback suppression method provided by one embodiment of the present application;
FIG. 4 is a flow chart of an acoustic feedback suppression method provided by another embodiment of the present application;
FIG. 5 is a schematic diagram of an acoustic feedback suppression scheme provided by one embodiment of the present application;
FIG. 6 is a block diagram of an acoustic feedback suppression apparatus provided in one embodiment of the present application;
fig. 7 is a block diagram of an electronic device according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the embodiments of the present application will be described in further detail with reference to the accompanying drawings.
Referring to fig. 1, a schematic diagram of an implementation environment of an embodiment of the present application is shown, where the implementation environment of the embodiment may include: an electronic device 10.
The electronic device 10 may be an electronic device such as a mobile phone, a tablet computer, a PC (Personal Computer ), a vehicle-mounted terminal, a smart home appliance, a multimedia playing device, or the like. The electronic device 10 may be in a closed, semi-closed, or open acoustic environment. The electronic device 10 may be in an in-vehicle environment, an indoor environment, or an outdoor environment, for example. For example, in an in-vehicle scenario, the electronic device 10 may collect a voice signal of a user, or obtain a voice signal of a user recorded by another device. And, acoustic feedback suppression processing is performed on the user's voice signal. Further, the electronic device 10 may amplify and play the processed voice signal, or transmit the processed voice signal to other devices for playing.
Referring to fig. 2, a schematic diagram of an acoustic amplification system according to an embodiment of the present application is shown, where the acoustic amplification system may include: a microphone 21 and a speaker 22.
In this scenario, after the microphone 21 collects the user's voice signal, the voice signal is amplified and amplified, and then is sent to the speaker 22 for playing, and a part of the played voice signal is absorbed by the physical structure in the acoustic environment, and a part of the played voice signal is fed back (including directly transmitted to and reflected from) the microphone 21 to form an acoustic feedback signal. This forms a closed loop signal circuit in the sound amplifying system. The characteristics of the loop make the voice signal collected by the microphone 21 and the voice signal played by the loudspeaker 22 have a strong correlation. With the increase of the gain of the sound amplifying system, the energy of the sound feedback signal is continuously accumulated, and when the energy of the sound feedback signal exceeds the system bearing capacity at a certain frequency or a plurality of frequencies, the howling phenomenon is generated at the frequency. Therefore, acoustic feedback suppression is required in an acoustic amplification system to avoid the occurrence of the howling phenomenon described above.
Note that, the microphone 21 and the speaker 22 may be provided in different electronic devices, or may be provided in the same electronic device, and the electronic device may be the electronic device 10 of the above embodiment, which is not limited to this aspect of the application.
Referring to fig. 3, a flowchart of an acoustic feedback suppression method according to an embodiment of the present application is shown. The main body of execution of each step of the method may be an electronic device, and for example, the main body of execution of each step may be the electronic device 10 described above. The method may include at least one of the following steps 310-330.
Step 310, obtaining a residual signal according to the input voice signal and the reference signal of the current frame through a linear filter; the linear filter is used for suppressing an acoustic feedback signal in the input voice signal, and the reference signal is obtained by performing acoustic feedback suppression on the input voice signal of the previous frame.
The linear filter is used for performing linear operation on the input voice signal to restrain the acoustic feedback signal. The input voice signal is a voice signal collected by the microphone.
In processing speech signals, framing techniques are required, and audio is framed because it is a long, unsteady sequence. In order to obtain a relatively stable characteristic parameter for unstable audio, we often do some framing operation. The goal of framing is to have unsteady audio in a steady state for short periods of time, i.e., short-time analysis techniques for speech signals. In short-time analysis, the voice information is analyzed by dividing the voice information into sections, each section is called a frame, and a frame can take 10-30 ms generally, and the application is not limited to this. In the embodiment of the present application, the input speech signal of the user is processed according to frames as well, that is, the reference signal may be understood as a signal obtained after the input speech signal of the previous frame is processed according to the acoustic feedback suppression method provided by the present application.
In some embodiments, the linear filter is a linear adaptive filter, which refers to a filter that uses an adaptive algorithm to change the parameters and structure of the filter according to changes in the environment.
In some embodiments, the weight parameters of the linear adaptive filter are updated based on the reference signal and the residual signal, the weight parameters of the linear adaptive filter being used to estimate the acoustic feedback signal in the input speech signal. The updated weight parameters are used to estimate the acoustic feedback signal in the input speech signal for the next frame. For example, in the case of using a normalized least mean square (NLMS, normalized Least Mean Square) adaptive filter, the weight parameter of the (k+1) th frameWherein w (k) is a weight parameter of the kth frame, a and mu are preset parameters, x (k) is a reference signal of the kth frame, e * (k) Is the complex conjugate of the residual signal of the kth frame.
Illustratively, the linear adaptive filter may be a least mean square (LMS, least Mean Square) adaptive filter, a normalized least mean square adaptive filter, a recursive least square (Recursive Least Square, RLS) adaptive filter, or a Kalman (Kalman) adaptive filter. The application is not limited in this regard.
In addition, it should be noted that, because the acoustic feedback suppression scheme provided by the embodiment of the application adopts a single filter architecture, in the case that the filter is a linear adaptive filter, the mutual influence of a plurality of filters in parameter update and the hysteresis of parameter update do not need to be considered, so that the acoustic feedback suppression scheme has stronger tracking capability, can obtain faster convergence speed, and can better suppress the linear part of the acoustic feedback signal. In addition, the single-filter architecture has lower operand and is favorable for equipment (such as embedded equipment) requiring calculation power.
Referring to FIG. 4, in some embodiments, step 310 may include at least one of the following sub-steps 312-314.
In a sub-step 312, a first estimated signal is obtained from the reference signal by means of a linear filter, the first estimated signal being an estimate of the acoustic feedback signal in the input speech signal of the current frame by means of the linear filter.
In some embodiments, substep 312 includes the steps of:
1. and pre-emphasis processing is carried out on the reference signal to obtain an emphasis signal.
2. And obtaining an emphasis estimation signal according to the emphasis signal through a linear filter.
3. And performing de-emphasis processing on the emphasis estimated signal to obtain a first estimated signal.
For voice signals, the low-frequency band energy of voice is large, the energy is mainly distributed in the low-frequency band, and the power spectrum density of voice is reduced along with the increase of frequency. Thus, the pre-emphasis process is used to compensate for the high frequency component of the reference signal before the acoustic feedback signal is estimated, in order to facilitate the subsequent operation of the linear filter. Accordingly, after the emphasis estimated signal is obtained, it needs to be subjected to de-emphasis processing to obtain a first estimated signal without distortion.
In a substep 314, the first estimated signal is subtracted from the input speech signal of the current frame to obtain a residual signal.
I.e. residual signal e=d-echo 1 Where d is the input speech signal of the current frame, echo 1 Is the first estimated signal.
In step 320, the acoustic feedback signal in the residual signal is estimated, resulting in a residual feedback signal.
The residual feedback signal is an estimate of the acoustic feedback signal in the residual signal, that is, the residual feedback signal is an estimate of the acoustic feedback signal of the residual portion of the input speech signal of the current frame that the linear filter fails to filter.
In some embodiments, step 320 includes sub-step 322.
In a substep 322, a residual feedback signal is obtained from the residual signal and the first estimated signal.
In some embodiments, deriving the residual feedback signal refers to deriving frequency domain characteristics of the residual feedback signal, including deriving a power spectrum of the residual feedback signal.
In some embodiments, there is a mapping between the first estimation signal and the residual feedback signal, and the residual feedback signal may be estimated by estimating the mapping. In the above process, the idea of noise reduction may be introduced, where the output speech signal (i.e., the residual signal after nonlinear processing) is estimated first, and then the ratio of the estimated output speech signal to the residual signal is estimated, so as to finally obtain the mapping relationship.
In some embodiments, substep 322 comprises the steps of:
1. calculating a first proportional value corresponding to each spectral line according to the spectral value of each spectral line in the first power spectrum and the spectral value of each spectral line in the second power spectrum; the first power spectrum is the power spectrum of the residual signal, the second power spectrum is the power spectrum of the first estimation signal, and the first proportion value corresponding to the spectral line is obtained by dividing the spectral value of the spectral line in the first power spectrum by the spectral value of the spectral line in the second power spectrum.
Illustratively, the residual signal is subjected to K-point fourier transform, so that a spectrum including K-half plus 1 spectral line having effective frequency domain information can be obtained, and the power spectrum of the residual signal, that is, the first power spectrum, can be obtained by multiplying the spectrum value of each spectral line in the spectrum by the conjugate thereof. Likewise, the second power spectrum may be obtained in a similar manner. In addition, in the embodiment of the present application, each spectral line corresponds to a different frequency, so the spectrum value of each spectral line may also be expressed as the spectrum value of each frequency, that is, the frequencies corresponding to the same spectral line in different power spectrums are the same. The spectral value of a spectral line in the power spectrum is the power value of the signal at the frequency corresponding to the spectral line.
In some embodiments, the spectrum value of the first spectral line in the first power spectrum may be divided by the spectrum value of the first spectral line in the second power spectrum, so as to calculate a first proportion value corresponding to the first spectral line.
That is, the first scale value=e corresponding to the first spectral line 1 ÷echo 1 Wherein e is 1 For the spectral value of the first spectral line in the first power spectrum, echo 1 Is the spectral value of the first spectral line in the second power spectrum. The first spectral line may be any one of the respective spectral lines.
2. And determining the first proportional value corresponding to each spectral line and the smaller value in the set first parameter as the second proportional value corresponding to each spectral line.
For example, a first scale value corresponding to the first spectral line and a smaller value of the set first parameter may be determined as a second scale value corresponding to the first spectral line.
I.e. alpha 1 ＝(e 1 ÷echo 1 Nk), where alpha 1 And (3) for the second proportion value corresponding to the first spectral line, nk is a set first parameter, and the value range of the nk is between 1 and 64. The specific value of nk may be determined by the actual acoustic environment and the specific processing mode of the speech signal, and in general, the higher the nk value, the more the acoustic feedback signal is suppressed.
3. And obtaining the first power and the second power according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum, the second proportion value corresponding to each spectral line and a preset frequency interval.
The first power refers to the total power of the residual signal in a preset frequency interval, and the second power refers to the total power of the estimated output voice signal in the preset frequency interval.
In some embodiments, the preset frequency interval may be a frequency range that can be heard by the human ear. The preset frequency interval may be, for example, 150Hz to 1000Hz.
In some embodiments, the first power may be calculated according to a spectrum value of each spectral line in the first power spectrum and a preset frequency interval.
For example, the spectral values of all spectral lines in the first power spectrum within a preset frequency interval may be added to obtain the first power. That is, the first power ep= Σe k Wherein e is k The spectrum value of the spectral line in the preset frequency interval in the first power spectrum.
In some embodiments, the second power may be calculated according to a spectrum value of each spectral line in the first power spectrum, a spectrum value of each spectral line in the second power spectrum, a second proportion value corresponding to each spectral line, and a preset frequency interval.
In some embodiments, the second power may be calculated by the following steps.
(1) And obtaining the spectrum value of each spectral line in a third power spectrum according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum and the second proportion value corresponding to each spectral line, wherein the third power spectrum refers to the estimated power spectrum of the output voice signal.
For a first spectral line of the spectral lines, the product of the spectral value of the first spectral line in the second power spectrum and a second proportional value corresponding to the first spectral line is subtracted from the spectral value of the first spectral line in the first power spectrum to obtain the spectral value of the first spectral line in the third power spectrum.
I.e. the spectral value of the first spectral line in the third power spectrum=e 1 -alpha 1 ×echo 1 。
(2) And carrying out smoothing treatment on the third power spectrum to obtain a smoothed third power spectrum.
In the embodiment of the application, the smoothing processing is performed on the data so as to reduce the gap between the data frames and further improve the readability and the interpretability of the data when the input voice signal is subjected to framing processing.
In some embodiments, the first spectral line in the third power spectrum may be smoothed by:
and multiplying the spectrum value of the first spectral line in the third power spectrum by a set second parameter for the first spectral line in each spectral line to obtain a first intermediate value.
And multiplying the spectrum value of the first spectral line in the smoothed third power spectrum obtained in the previous frame by the difference between 1 and the second parameter to obtain a second intermediate value.
And adding the first intermediate value and the second intermediate value to obtain a spectrum value of the first spectral line in the smoothed third power spectrum.
The above steps can be expressed by the following formula: wherein, ps 1 For smoothing the spectral value, ps, of the first spectral line in the third power spectrum 1 ' is the spectral value of the first spectral line in the smoothed third power spectrum obtained from the previous frame, ">For the second parameter to be set, a value between 0 and 1 may be taken.
(3) And carrying out smoothing treatment on the first power spectrum to obtain a smoothed first power spectrum.
In some embodiments, the first spectral line in the first power spectrum may be smoothed by the following formula:wherein Pe is 1 To smooth the spectrum value of the first spectral line in the processed first power spectrum, pe 1 ' is the spectral value of the first spectral line in the smoothed first power spectrum obtained from the previous frame.
(4) And calculating to obtain the second power according to the spectrum value of each spectral line in the third power spectrum after the smoothing, the spectrum value of each spectral line in the first power spectrum and a preset frequency interval.
In some embodiments, this step comprises the steps of:
and obtaining a third proportion value corresponding to each spectral line according to the spectral value of each spectral line in the third power spectrum after the smoothing processing and the spectral value of each spectral line in the first power spectrum after the smoothing processing.
And adding products of the spectral values of the spectral lines in the preset frequency interval and the third proportion value in the first power spectrum to obtain the second power.
The above steps can be expressed by the following formula: sP= Σ (Ps k ÷Pe k ×e k ) Wherein sP is the second power, ps k For smoothing spectral values of spectral lines in a preset frequency interval in the third power spectrum, pe k For smoothing spectral values, ps, of spectral lines in a predetermined frequency interval in the processed first power spectrum k ÷Pe k And the third ratio value corresponding to the spectral line is obtained.
4. And obtaining a power spectrum of the residual feedback signal according to the first power, the second power and the first power spectrum.
In some embodiments, this step includes the sub-steps of:
(1) And carrying out smoothing treatment on the first power to obtain the smoothed first power.
In some embodiments, the first power may be smoothed by:
and multiplying the first power by the set third parameter to obtain a third intermediate value.
And multiplying the smoothed first power obtained in the previous frame by the difference between 1 and the third parameter to obtain a fourth intermediate value.
And adding the third intermediate value and the fourth intermediate value to obtain the first power after the smoothing.
The above steps can be expressed by the following formula: ePs =γχ ePs '+ (1- γ) ×ep, where ePs is the first power after smoothing, ePs' is the first power after smoothing obtained in the previous frame, and γ is a third parameter set, and may take a value between 0 and 1.
(2) And carrying out smoothing treatment on the second power to obtain smoothed second power.
In some embodiments, the second power may be smoothed by the following formula: sPs =γ× sPs '+ (1- γ) ×sp, where sPs is the smoothed second power and sPs' is the smoothed second power obtained in the previous frame.
(3) Dividing the smoothed second power by the smoothed first power to obtain a first gain value. I.e. the first gain value mask= sPs +. ePs.
(4) And carrying out nonlinear mapping according to the first gain value to obtain a second gain value.
Nonlinear mapping is understood in the present application as mapping the first gain value to a higher dimensional space by nonlinear operation, making it close to the gain value of the actual dimension where the residual signal is nonlinear processed.
In some embodiments, the step comprises:
and converting the first gain value into a decibel form to obtain a first decibel value. That is, the first decibel value mask db=10×log 10 mask。
And carrying out linear transformation on the first decibel value to obtain a first linear value. That is, limt= - α×mask db- β, where limt is a first linear value, and α and β are preset parameters.
And taking 10 as a bottom, taking the result obtained by dividing the first linear value by 10 as a power, and performing power operation to obtain the second gain value. I.e. the second gain value mask up=10 limt-10 。
(5) And limiting the second gain value within a preset range to obtain a third gain value.
In some embodiments, the second gain value may be limited to within 1 to 120, i.e. the third gain value mask up' =min (max (1), 120).
(6) And determining the smaller value of the first proportional value and the third gain value corresponding to each spectral line as a fourth proportional value corresponding to each spectral line.
Illustratively, a fourth scale value alpha corresponding to the first spectral line 1 ′＝(e 1 ÷echo 1 ，mask up′)。
(7) And multiplying the spectrum value of each spectral line in the first power spectrum by a fourth proportional value corresponding to each spectral line to obtain the power spectrum of the residual feedback signal.
Illustratively, in the power spectrum of the residual feedback signal, the spectral value of the first spectral line remains echo 1 ＝alpha 1 ′×echo 1 。
By the method, the (power spectrum of the) residual acoustic feedback signal can be accurately estimated, so that the residual acoustic feedback signal can be better restrained.
And 330, performing nonlinear processing on the residual signal according to the residual feedback signal to obtain an output voice signal.
Nonlinear processing of the residual signal means suppressing the acoustic feedback signal in the residual signal, which is difficult to filter out once by a linear filter, since this part of the acoustic feedback signal comprises a nonlinear part of the acoustic feedback signal of the input speech signal. The output voice signal is used for gain amplification and then provided for a loudspeaker to play.
In some embodiments, the gain value of the residual feedback signal by the nonlinear process may be estimated by wiener (wiener) filtering, based on the residual feedback signal and the power spectrum of the residual signal, to further suppress the acoustic feedback signal in the residual signal. The above is merely exemplary, and the present application is not limited to the specific steps of the nonlinear processing described above.
In some embodiments, referring to fig. 5, after performing nonlinear processing on the residual signal to obtain an output speech signal, the output speech signal needs to be returned to the linear filter as a reference signal of the next frame.
In some embodiments, the acoustic feedback suppression method further includes the steps of:
1. And adjusting a divergence value according to the energy value of the first estimation signal, the energy value of the reference signal and the energy value of the residual signal, wherein the divergence value is used for measuring the divergence degree of the linear filter.
For example, the initial divergence value may be 0, and the quality of the filter's estimation of the acoustic feedback signal in the input speech signal of the current frame may be judged based on a magnitude relation, a proportional relation, or a difference between the energy value of the first estimation signal and the energy value of the reference signal, the energy value of the residual signal. Further, when the three energy values indicate that the filter has poor estimation of the acoustic feedback signal in the input speech signal of the current frame, the divergence value is increased by 1; when the three energy values indicate that the filter estimates the acoustic feedback signal very poorly in the input speech signal of the current frame, the divergence value is increased by 50; the divergence value is set to 0 when the three energy values indicate that the filter estimates the acoustic feedback signal in the input speech signal of the current frame better.
2. And under the condition that the divergence value is larger than the threshold value, resetting the weight parameter of the linear filter, wherein the weight parameter of the linear filter is used for estimating the acoustic feedback signal in the input voice signal.
For example, the threshold value may be 2000, i.e. when the divergence value is greater than 2000, the weight parameter of the linear filter is adjusted back to the initial state.
In some embodiments, the value of the first estimated signal is limited in the event that the divergence value is greater than a threshold value.
The divergence value being greater than the threshold value indicates that the first estimated signal of the current frame is not trusted. For example, in case the divergence value is greater than the threshold value, the amplitude of the first estimated signal is limited within a preset range of values (e.g. -1 to 1).
According to the technical scheme provided by the embodiment of the application, the linear filter is used for suppressing the acoustic feedback signal in the input voice signal of the current frame according to the reference signal obtained by performing acoustic feedback suppression on the input voice signal of the previous frame, so as to obtain the residual signal. And estimating the acoustic feedback signal which is not filtered by the linear filter in the residual signal, so as to obtain the residual feedback signal. Finally, according to the residual feedback signal, the acoustic feedback signal in the residual signal is further suppressed through nonlinear processing, and an output voice signal is obtained. The method suppresses the linear part and the nonlinear part of the acoustic feedback signal in the input voice signal, so that the loudspeaker can obtain larger maximum stable gain when playing the output voice signal.
In an acoustic amplification system, the maximum stable gain is understood as the maximum gain value of the system without howling, and can be used to measure the overall performance of the system. In the related art, if the acoustic feedback suppression is performed by the acoustic amplification system by adopting a frequency/phase shift scheme, the maximum stable gain is usually 1-3dB; if the acoustic amplification system adopts an automatic notch scheme to perform acoustic feedback suppression, the maximum stable gain is usually 3-6dB; if the acoustic amplification system only adopts an adaptive filtering scheme for acoustic feedback suppression, the maximum stable gain is usually 6-10dB. When the acoustic feedback suppression scheme provided by the embodiment of the application is adopted by the sound amplification system, the maximum stable gain is usually 8-12dB, and the sound quality is better.
The following are examples of the apparatus of the present application that may be used to perform the method embodiments of the present application. For details not disclosed in the embodiments of the apparatus of the present application, please refer to the embodiments of the method of the present application.
Referring to fig. 6, a block diagram of an acoustic feedback suppression apparatus according to an embodiment of the present application is shown. The device has the function of realizing the acoustic feedback suppression method, and the function can be realized by hardware or by executing corresponding software by the hardware. The device can be an electronic device or can be arranged in the electronic device. The apparatus 600 may include: a linear module 610, an estimation module 620, and a nonlinear module 630.
A linear module 610, configured to obtain a residual signal according to the input speech signal and the reference signal of the current frame by using a linear filter; the linear filter is used for suppressing an acoustic feedback signal in the input voice signal, and the reference signal is obtained by performing acoustic feedback suppression on the input voice signal of the previous frame.
And an estimation module 620, configured to estimate the acoustic feedback signal in the residual signal, so as to obtain a residual feedback signal.
And the nonlinear module 630 is configured to perform nonlinear processing on the residual signal according to the residual feedback signal, so as to obtain an output speech signal.
In some embodiments, the linear module 610 includes: an estimation sub-module and a subtraction operation sub-module.
And the estimation sub-module is used for obtaining a first estimation signal according to the reference signal through the linear filter, wherein the first estimation signal is an estimation value of the linear filter on an acoustic feedback signal in the input voice signal of the current frame.
And the subtraction operation submodule is used for subtracting the first estimation signal from the input voice signal of the current frame to obtain the residual signal.
In some embodiments, the estimation sub-module is configured to perform pre-emphasis processing on the reference signal to obtain an emphasized signal; obtaining an emphasis estimation signal according to the emphasis signal through the linear filter; and de-emphasis processing is carried out on the emphasis estimation signal to obtain the first estimation signal.
In some embodiments, the estimation module 620 is configured to obtain the residual feedback signal according to the residual signal and the first estimation signal.
In some embodiments, the estimation module 620 includes: the system comprises a proportion calculation sub-module, a power calculation sub-module and a spectrum calculation sub-module.
The proportion calculation sub-module is used for calculating a first proportion value corresponding to each spectral line according to the spectral value of each spectral line in the first power spectrum and the spectral value of each spectral line in the second power spectrum; the first power spectrum is a power spectrum of the residual signal, the second power spectrum is a power spectrum of the first estimation signal, and the first proportion value corresponding to the spectral line is obtained by dividing a spectral value of the spectral line in the first power spectrum by a spectral value of the spectral line in the second power spectrum; and determining the first proportional value corresponding to each spectral line and the smaller value in the set first parameter as the second proportional value corresponding to each spectral line.
The power calculation sub-module is used for obtaining first power and second power according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum, a second proportion value corresponding to each spectral line and a preset frequency interval; the first power refers to the total power of the residual signal in the preset frequency interval, and the second power refers to the estimated total power of the output voice signal in the preset frequency interval.
And the spectrum calculation sub-module is used for obtaining the power spectrum of the residual feedback signal according to the first power, the second power and the first power spectrum.
In some embodiments, the power calculation submodule includes: a first unit and a second unit.
The first unit is used for calculating the first power according to the spectrum value of each spectral line in the first power spectrum and the preset frequency interval.
The second unit is configured to calculate the second power according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum, the second proportion value corresponding to each spectral line, and the preset frequency interval.
In some embodiments, the second unit comprises: an estimation subunit, a smoothing subunit and a power calculation subunit.
And the estimation subunit is used for obtaining the spectrum value of each spectral line in a third power spectrum according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum and the second proportion value corresponding to each spectral line, wherein the third power spectrum refers to the estimated power spectrum of the output voice signal.
A smoothing processing subunit, configured to perform smoothing processing on the third power spectrum, to obtain a smoothed third power spectrum; and carrying out smoothing treatment on the first power spectrum to obtain a smoothed first power spectrum.
The power calculation subunit is configured to calculate the second power according to the spectrum value of each spectral line in the third power spectrum after the smoothing process, the spectrum value of each spectral line in the first power spectrum, and the preset frequency interval.
In some embodiments, the estimation subunit is configured to subtract, for a first spectral line of the spectral lines, a product of a spectral value of the first spectral line in the first power spectrum and a second ratio value corresponding to the first spectral line in the second power spectrum, to obtain a spectral value of the first spectral line in the third power spectrum.
In some embodiments, the smoothing subunit is configured to multiply, for a first spectral line of the spectral lines, a spectral value of the first spectral line in the third power spectrum by a set second parameter to obtain a first intermediate value; multiplying the spectrum value of the first spectral line in the smoothed third power spectrum obtained in the previous frame by the difference between 1 and the second parameter to obtain a second intermediate value; and adding the first intermediate value and the second intermediate value to obtain a spectrum value of the first spectral line in the smoothed third power spectrum.
In some embodiments, the power calculating subunit is configured to obtain a third proportion value corresponding to each spectral line according to a spectral value of each spectral line in the third smoothed power spectrum and a spectral value of each spectral line in the first smoothed power spectrum; and adding products of the spectral values of the spectral lines and the third proportion value in the first power spectrum in the preset frequency interval to obtain the second power.
In some embodiments, the spectrum calculation submodule includes: a smoothing processing unit, a gain calculation unit, a nonlinear mapping unit, a limiting unit, a proportion calculation unit and a spectrum calculation unit.
The smoothing processing unit is used for carrying out smoothing processing on the first power to obtain smoothed first power; and carrying out smoothing treatment on the second power to obtain smoothed second power.
And the gain calculation unit is used for dividing the smoothed second power by the smoothed first power to obtain a first gain value.
And the nonlinear mapping unit is used for carrying out nonlinear mapping according to the first gain value to obtain a second gain value.
And the limiting unit is used for determining the smaller value of the first proportion value and the third gain value corresponding to each spectral line as a fourth proportion value corresponding to each spectral line.
And the spectrum calculation unit is used for multiplying the spectrum value of each spectral line in the first power spectrum by a fourth proportion value corresponding to each spectral line to obtain the power spectrum of the residual feedback signal.
In some embodiments, the smoothing unit is configured to multiply the first power by a set third parameter to obtain a third intermediate value; multiplying the smoothed first power obtained in the previous frame by the difference between 1 and the third parameter to obtain a fourth intermediate value; and adding the third intermediate value and the fourth intermediate value to obtain the smoothed first power.
In some embodiments, the nonlinear mapping unit is configured to convert the first gain value into a decibel form, to obtain a first decibel value; performing linear transformation on the first decibel value to obtain a first linear value; and taking 10 as a bottom, taking the result obtained by dividing the first linear value by 10 as a power, and performing power operation to obtain the second gain value.
In some embodiments, the apparatus 600 further comprises a divergence protection module (not shown in fig. 6).
And the divergence protection module is used for adjusting a divergence value according to the energy value of the first estimation signal, the energy value of the reference signal and the energy value of the residual signal, wherein the divergence value is used for measuring the divergence degree of the linear filter. And under the condition that the divergence value is larger than a threshold value, resetting the weight parameter of the linear filter, wherein the weight parameter of the linear filter is used for estimating an acoustic feedback signal in the input voice signal.
In some embodiments, the divergence protection module is further configured to limit the value of the first estimated signal if the divergence value is greater than a threshold value.
In some embodiments, the linear filter is a linear adaptive filter, and the apparatus 600 further includes a parameter adjustment module (not shown in fig. 6).
And the parameter adjustment module is used for updating the weight parameters of the linear adaptive filter according to the reference signal and the residual signal, and the weight parameters of the linear adaptive filter are used for estimating the acoustic feedback signal in the input voice signal.
It should be noted that, in the apparatus provided in the foregoing embodiment, when implementing the functions thereof, only the division of the foregoing functional modules is used as an example, in practical application, the foregoing functional allocation may be implemented by different functional modules, that is, the internal structure of the device is divided into different functional modules, so as to implement all or part of the functions described above. In addition, the apparatus and the method embodiments provided in the foregoing embodiments belong to the same concept, and specific implementation processes of the apparatus and the method embodiments are detailed in the method embodiments and are not repeated herein.
Referring to fig. 7, a block diagram of an electronic device according to an embodiment of the present application is shown.
In general, the electronic device 700 includes: a processor 701 and a memory 702.
Processor 701 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and the like. The processor 701 may be implemented in at least one hardware form of DSP (Digital Signal Processing ), FPGA (Field Programmable Gate Array, field programmable gate array), PLA (Programmable Logic Array ). The processor 701 may also include a main processor, which is a processor for processing data in an awake state, also referred to as a CPU (Central Processing Unit ); a coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 701 may integrate a GPU (Graphics Processing Unit, image processor) for rendering and drawing of content required to be displayed by the display screen. In some embodiments, the processor 701 may also include an AI processor for processing computing operations related to machine learning.
Memory 702 may include one or more computer-readable storage media, which may be tangible and non-transitory. The memory 702 may also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In some embodiments, a non-transitory computer readable storage medium in memory 702 stores a computer program that is loaded and executed by processor 701 to implement the acoustic feedback suppression method described above.
Those skilled in the art will appreciate that the architecture shown in fig. 7 is not limiting of the computer device 700, and may include more or fewer components than shown, or may combine certain components, or employ a different arrangement of components.
In some embodiments, a computer readable storage medium having a computer program stored therein, the computer program being loaded and executed by a processor to implement the acoustic feedback suppression method described above is also provided.
Alternatively, the computer-readable storage medium may include: ROM (Read-Only Memory), RAM (Random-Access Memory), SSD (Solid State Drives, solid State disk), optical disk, or the like. The random access memory may include ReRAM (Resistance Random Access Memory, resistive random access memory) and DRAM (Dynamic Random Access Memory ), among others.
In some embodiments, there is also provided a computer program product comprising a computer program stored in a computer readable storage medium, from which a processor reads and executes the computer program to implement the acoustic feedback suppression method described above.
It should be understood that references herein to "a plurality" are to two or more. "and/or", describes an association relationship of an association object, and indicates that there may be three relationships, for example, a and/or B, and may indicate: a exists alone, A and B exist together, and B exists alone. The character "/" generally indicates that the context-dependent object is an "or" relationship. In addition, the step numbers described herein are merely exemplary of one possible execution sequence among steps, and in some other embodiments, the steps may be executed out of the order of numbers, such as two differently numbered steps being executed simultaneously, or two differently numbered steps being executed in an order opposite to that shown, which is not limiting.
The foregoing is illustrative of the present application and is not to be construed as limiting thereof, but rather, any modification, equivalent replacement, improvement or the like which comes within the spirit and principles of the present application are intended to be included within the scope of the present application.
Claims (20)
1. A method of acoustic feedback suppression, the method comprising:
obtaining a residual signal according to an input voice signal and a reference signal of a current frame through a linear filter; the linear filter is used for suppressing an acoustic feedback signal in the input voice signal, and the reference signal is obtained by performing acoustic feedback suppression on the input voice signal of the previous frame;
estimating an acoustic feedback signal in the residual signal to obtain a residual feedback signal;
and carrying out nonlinear processing on the residual signal according to the residual feedback signal to obtain an output voice signal.
2. The method of claim 1, wherein the obtaining, by the linear filter, a residual signal from the input speech signal of the current frame and the reference signal, comprises:
obtaining a first estimated signal according to the reference signal through the linear filter, wherein the first estimated signal is an estimated value of the linear filter on an acoustic feedback signal in an input voice signal of the current frame;
subtracting the first estimated signal from the input voice signal of the current frame to obtain the residual signal.
3. The method of claim 2, wherein said obtaining, by said linear filter, a first estimated signal from said reference signal, comprises:
pre-emphasis processing is carried out on the reference signal to obtain an emphasis signal;
obtaining an emphasis estimation signal according to the emphasis signal through the linear filter;
and de-emphasis processing is carried out on the emphasis estimation signal to obtain the first estimation signal.
4. The method of claim 2, wherein estimating noise in the residual signal to obtain a residual feedback signal comprises:
and obtaining the residual feedback signal according to the residual signal and the first estimation signal.
5. The method of claim 4, wherein said deriving the residual feedback signal from the residual signal and the first estimated signal comprises:
calculating a first proportional value corresponding to each spectral line according to the spectral value of each spectral line in the first power spectrum and the spectral value of each spectral line in the second power spectrum; the first power spectrum is a power spectrum of the residual signal, the second power spectrum is a power spectrum of the first estimation signal, and the first proportion value corresponding to the spectral line is obtained by dividing a spectral value of the spectral line in the first power spectrum by a spectral value of the spectral line in the second power spectrum;
Determining a first proportion value corresponding to each spectral line and a smaller value in the set first parameter as a second proportion value corresponding to each spectral line;
obtaining first power and second power according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum, a second proportion value corresponding to each spectral line and a preset frequency interval; wherein the first power refers to the total power of the residual signal in the preset frequency interval, and the second power refers to the estimated total power of the output voice signal in the preset frequency interval;
and obtaining the power spectrum of the residual feedback signal according to the first power, the second power and the first power spectrum.
6. The method of claim 5, wherein the obtaining the first power and the second power according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum, the second proportion value corresponding to each spectral line, and the preset frequency interval includes:
calculating to obtain the first power according to the spectrum value of each spectral line in the first power spectrum and the preset frequency interval;
And calculating to obtain the second power according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum, the second proportion value corresponding to each spectral line and the preset frequency interval.
7. The method of claim 6, wherein the calculating the second power according to the spectral value of each spectral line in the first power spectrum, the spectral value of each spectral line in the second power spectrum, the second ratio value corresponding to each spectral line, and the preset frequency interval includes:
obtaining a spectrum value of each spectral line in a third power spectrum according to the spectrum value of each spectral line in the first power spectrum, the spectrum value of each spectral line in the second power spectrum and a second proportion value corresponding to each spectral line, wherein the third power spectrum refers to the estimated power spectrum of the output voice signal;
performing smoothing treatment on the third power spectrum to obtain a smoothed third power spectrum;
smoothing the first power spectrum to obtain a smoothed first power spectrum;
and calculating to obtain the second power according to the spectrum value of each spectral line in the third power spectrum after the smoothing, the spectrum value of each spectral line in the first power spectrum and the preset frequency interval.
8. The method of claim 7, wherein the obtaining the spectral value of each spectral line in the third power spectrum according to the spectral value of each spectral line in the first power spectrum, the spectral value of each spectral line in the second power spectrum, and the second ratio value corresponding to each spectral line, comprises:
and for a first spectral line in the spectral lines, subtracting a product of the spectral value of the first spectral line in the second power spectrum and a second proportion value corresponding to the first spectral line from the spectral value of the first spectral line in the first power spectrum to obtain the spectral value of the first spectral line in the third power spectrum.
9. The method of claim 7, wherein smoothing the third power spectrum to obtain a smoothed third power spectrum comprises:
multiplying the spectrum value of the first spectral line in the third power spectrum by a set second parameter for the first spectral line in each spectral line to obtain a first intermediate value;
multiplying the spectrum value of the first spectral line in the smoothed third power spectrum obtained in the previous frame by the difference between 1 and the second parameter to obtain a second intermediate value;
and adding the first intermediate value and the second intermediate value to obtain a spectrum value of the first spectral line in the smoothed third power spectrum.
10. The method of claim 7, wherein the calculating the second power according to the spectral value of each spectral line in the smoothed third power spectrum, the spectral value of each spectral line in the smoothed first power spectrum, the spectral value of each spectral line in the first power spectrum, and the preset frequency interval includes:
obtaining a third proportion value corresponding to each spectral line according to the spectral value of each spectral line in the third power spectrum after smoothing and the spectral value of each spectral line in the first power spectrum after smoothing;
and adding products of the spectral values of the spectral lines and the third proportion value in the first power spectrum in the preset frequency interval to obtain the second power.
11. The method of claim 5, wherein said deriving a power spectrum of said residual feedback signal from said first power, said second power, and said first power spectrum comprises:
smoothing the first power to obtain smoothed first power;
smoothing the second power to obtain smoothed second power;
Dividing the smoothed second power by the smoothed first power to obtain a first gain value;
according to the first gain value, nonlinear mapping is carried out to obtain a second gain value;
limiting the second gain value within a preset range to obtain a third gain value;
determining smaller values of the first proportional value and the third gain value corresponding to each spectral line as fourth proportional values corresponding to each spectral line;
multiplying the spectrum value of each spectral line in the first power spectrum by a fourth proportion value corresponding to each spectral line to obtain the power spectrum of the residual feedback signal.
12. The method of claim 11, wherein smoothing the first power to obtain a smoothed first power comprises:
multiplying the first power by a set third parameter to obtain a third intermediate value;
multiplying the smoothed first power obtained in the previous frame by the difference between 1 and the third parameter to obtain a fourth intermediate value;
and adding the third intermediate value and the fourth intermediate value to obtain the smoothed first power.
13. The method of claim 11, wherein said non-linearly mapping according to said first gain value to obtain a second gain value comprises:
Converting the first gain value into a decibel form to obtain a first decibel value;
performing linear transformation on the first decibel value to obtain a first linear value;
and taking 10 as a bottom, taking the result obtained by dividing the first linear value by 10 as a power, and performing power operation to obtain the second gain value.
14. The method according to claim 2, wherein the method further comprises:
according to the energy value of the first estimation signal, the energy value of the reference signal and the energy value of the residual signal, adjusting a divergence value, wherein the divergence value is used for measuring the divergence degree of the linear filter;
and under the condition that the divergence value is larger than a threshold value, resetting the weight parameter of the linear filter, wherein the weight parameter of the linear filter is used for estimating an acoustic feedback signal in the input voice signal.
15. The method of claim 14, wherein the method further comprises:
and limiting the value of the first estimation signal under the condition that the divergence value is larger than a threshold value.
16. The method of claim 1, wherein the linear filter is a linear adaptive filter, the method further comprising:
And updating weight parameters of the linear adaptive filter according to the reference signal and the residual signal, wherein the weight parameters of the linear adaptive filter are used for estimating an acoustic feedback signal in the input voice signal.
17. An acoustic feedback suppression apparatus, the apparatus comprising:
the linear module is used for obtaining a residual signal according to the input voice signal and the reference signal of the current frame through a linear filter; the linear filter is used for suppressing an acoustic feedback signal in the input voice signal, and the reference signal is obtained by performing acoustic feedback suppression on the input voice signal of the previous frame;
the estimation module is used for estimating the acoustic feedback signal in the residual signal to obtain a residual feedback signal;
and the nonlinear module is used for carrying out nonlinear processing on the residual signal according to the residual feedback signal to obtain an output voice signal.
18. An electronic device comprising a processor and a memory, the memory having stored therein a computer program that is loaded and executed by the processor to implement the method of any of the preceding claims 1 to 16.
19. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein a computer program, which is loaded and executed by a processor to implement the method of any of the preceding claims 1 to 16.
20. A computer program product, characterized in that it comprises a computer program stored in a computer readable storage medium, from which a processor reads and executes the computer program to implement the method according to any of the preceding claims 1 to 16.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311197516.5A CN117116281A (en) | 2023-09-15 | 2023-09-15 | Acoustic feedback suppression method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311197516.5A CN117116281A (en) | 2023-09-15 | 2023-09-15 | Acoustic feedback suppression method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117116281A true CN117116281A (en) | 2023-11-24 |
Family
ID=88796404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311197516.5A Pending CN117116281A (en) | 2023-09-15 | 2023-09-15 | Acoustic feedback suppression method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117116281A (en) |
-
2023
- 2023-09-15 CN CN202311197516.5A patent/CN117116281A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP2845189B1 (en) | A universal reconfigurable echo cancellation system | |
| US9437180B2 (en) | Adaptive noise reduction using level cues | |
| CN110996203B (en) | Earphone noise reduction method, device and system and wireless earphone | |
| US10553236B1 (en) | Multichannel noise cancellation using frequency domain spectrum masking | |
| CN110176244B (en) | Echo cancellation method, device, storage medium and computer equipment | |
| US10115411B1 (en) | Methods for suppressing residual echo | |
| CN111883156A (en) | Audio processing method and device, electronic equipment and storage medium | |
| WO2020097828A1 (en) | Echo cancellation method, delay estimation method, echo cancellation apparatus, delay estimation apparatus, storage medium, and device | |
| CN110782914B (en) | Signal processing method and device, terminal equipment and storage medium | |
| KR102040986B1 (en) | Method and apparatus for noise reduction in a portable terminal having two microphones | |
| KR102190833B1 (en) | Echo suppression | |
| CN108540895A (en) | Intelligent equalization device design method and noise cancelling headphone with intelligent equalization device | |
| WO2019239977A1 (en) | Echo suppression device, echo suppression method, and echo suppression program | |
| CN111756906B (en) | Echo suppression method and device for voice signal and computer readable medium | |
| CN111524532B (en) | Echo suppression method, device, equipment and storage medium | |
| CN110503973B (en) | Audio signal transient noise suppression method, system and storage medium | |
| CN109215672B (en) | Method, device and equipment for processing sound information | |
| CN112997249B (en) | Voice processing method, device, storage medium and electronic equipment | |
| CN110021289B (en) | Sound signal processing method, device and storage medium | |
| CN115884030A (en) | Noise reduction method and device, earphone equipment, storage medium and chip | |
| CN117116281A (en) | Acoustic feedback suppression method, device, equipment and storage medium | |
| CA2840730C (en) | Maintaining spatial stability utilizing common gain coefficient | |
| CN115240700B (en) | Acoustic device and sound processing method thereof | |
| EP2816817B1 (en) | Sound field spatial stabilizer with spectral coherence compensation | |
| CN115440236A (en) | Echo suppression method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |