CN112201273A - Noise power spectral density calculation method, system, equipment and medium - Google Patents
Noise power spectral density calculation method, system, equipment and medium Download PDFInfo
- Publication number
- CN112201273A CN112201273A CN201910612851.4A CN201910612851A CN112201273A CN 112201273 A CN112201273 A CN 112201273A CN 201910612851 A CN201910612851 A CN 201910612851A CN 112201273 A CN112201273 A CN 112201273A
- Authority
- CN
- China
- Prior art keywords
- noise signal
- signal
- domain noise
- spectral density
- power spectral
- 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
- 230000003595 spectral effect Effects 0.000 title claims abstract description 79
- 238000004364 calculation method Methods 0.000 title claims abstract description 33
- 238000012545 processing Methods 0.000 claims abstract description 60
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 48
- 238000000034 method Methods 0.000 claims abstract description 44
- 238000001914 filtration Methods 0.000 claims abstract description 39
- 230000003044 adaptive effect Effects 0.000 claims description 55
- 230000008569 process Effects 0.000 claims description 16
- 230000004044 response Effects 0.000 claims description 16
- 238000005311 autocorrelation function Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 6
- 238000009432 framing Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 abstract description 8
- 230000000694 effects Effects 0.000 abstract description 4
- 238000001228 spectrum Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000003775 Density Functional Theory Methods 0.000 description 3
- 230000000903 blocking effect Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 102100035971 Molybdopterin molybdenumtransferase Human genes 0.000 description 2
- 101710119577 Molybdopterin molybdenumtransferase Proteins 0.000 description 2
- 238000005314 correlation function Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011410 subtraction method Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G10L21/0224—Processing in the time domain
-
- 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
-
- 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
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/21—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being power information
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)
- Soundproofing, Sound Blocking, And Sound Damping (AREA)
Abstract
A method of noise power spectral density calculation, comprising: collecting a time domain noise signal, and processing the time domain noise signal to obtain a frequency domain noise signal; processing the frequency domain noise signal by using a self-adaptive filter to obtain a time domain error signal, solving the error signal by using a normalized minimum mean square error algorithm in an algorithm part of the self-adaptive filter, and further solving the error signal by using a block frequency domain self-adaptive filtering algorithm; and calculating the power spectral density of the time domain error signal, and calculating the noise power spectral density according to the power spectral density of the error signal. The invention also discloses a noise power spectral density calculation system, electronic equipment and a storage medium. The invention enables the noise power spectral density estimation to be more accurate, reduces the calculated amount and has better subsequent noise reduction effect.
Description
Technical Field
The present invention relates to the field of noise processing, and in particular, to a method, system, device, and medium for calculating a noise power spectral density.
Background
With the development of communication technology, the voice interaction technology is mature day by day and is widely applied to intelligent devices such as mobile phones, intelligent sound boxes and intelligent homes. However, in practical use scenarios, there are still some problems, such as in far-field, noisy scenarios, where noise and interference techniques are required to obtain a clean speech signal for wake-up and speech recognition.
Under far-field conditions, background noise in a signal received by a microphone occupies a large amount of components, and noise suppression algorithms (such as spectral subtraction, wiener filtering, and energy-based filtering algorithms) need to be used for estimating the power spectral density of the background noise, but these algorithms have some defects, for example, the wiener filtering algorithm has a large calculation amount and a low calculation accuracy. Therefore, there is a need to provide an improved noise power spectral density calculation method to reduce the amount of calculation and improve the calculation accuracy.
Disclosure of Invention
In view of the technical problems at present, it is a primary object of the present invention to provide a method, system, device and medium for calculating noise power spectral density, which at least partially solve the above technical problems.
A first aspect of an embodiment of the present invention provides a method for calculating a noise power spectral density, including: collecting a time domain noise signal, and processing the time domain noise signal to obtain a frequency domain noise signal; processing the frequency domain noise signal by adopting a self-adaptive filter to obtain a time domain error signal; and calculating the power spectral density of the time domain error signal, and calculating the noise power spectral density according to the power spectral density of the error signal.
Optionally, the frequency-domain noise signal includes a first frequency-domain noise signal and a second frequency-domain noise signal, and the processing the frequency-domain noise signal with the adaptive filter to obtain the error signal includes: carrying out normalization minimum mean square error processing on the first frequency domain noise signal to obtain a filtering impact response signal corresponding to the first frequency domain noise signal; performing convolution processing on the first frequency domain noise signal and the corresponding filtering impact response signal to obtain an estimation signal; and performing subtraction operation on the second frequency domain noise signal and the estimation signal, and performing inverse Fourier transform to obtain a time domain error signal.
Optionally, the frequency-domain noise signal includes a first frequency-domain noise signal and a second frequency-domain noise signal, and the processing the frequency-domain noise signal with the adaptive filter to obtain the error signal includes: and performing block adaptive filtering processing on the first frequency domain noise signal by adopting a block frequency domain adaptive filtering algorithm so as to reduce the calculated amount.
Optionally, the method further comprises: and updating the filter coefficient of the adaptive filter according to the error signal.
Optionally, the acquiring a time-domain noise signal, and the processing the time-domain noise signal includes: collecting a first time domain noise signal and a second time domain noise signal; and framing the first time domain noise signal and the second time domain noise signal, windowing, and performing fast Fourier transform to obtain a first frequency domain noise signal corresponding to the first time domain noise signal and a second frequency domain noise signal corresponding to the second time domain noise signal.
Optionally, in the process of performing the block adaptive filtering on the first frequency domain noise signal, partial removal and zero padding are performed on the point location data corresponding to the first frequency domain noise signal.
Optionally, the power spectral density of the time domain error signal is calculated as follows:
ΓEE(k)＝FFT(γee(τ))＝FFT(e(n)·e(n+τ))
wherein e (n) represents a time-domain error signal, γee(τ) is the autocorrelation function of e (n), τ is the time delay, FFT () represents a fast Fourier transform operation, ΓEE(k) Denotes the power spectral density of e (n).
A first aspect of an embodiment of the present invention provides a noise power spectral density calculation system, including: the first processing module is used for collecting a time domain noise signal and processing the time domain noise signal to obtain a frequency domain noise signal; the second processing module is used for processing the frequency domain noise signal by adopting the self-adaptive filter to obtain a time domain error signal; and the calculating module is used for calculating the power spectral density of the time-domain error signal and calculating the noise power spectral density according to the power spectral density of the error signal.
Optionally, the frequency-domain noise signal includes a first frequency-domain noise signal and a second frequency-domain noise signal, and the second processing module processes the frequency-domain noise signal by using an adaptive filter, and obtaining the error signal includes: carrying out normalization minimum mean square error processing on the first frequency domain noise signal to obtain a filtering impact response signal corresponding to the first frequency domain noise signal; performing convolution processing on the first frequency domain noise signal and the corresponding filtering impact response signal to obtain an estimation signal; and performing subtraction operation on the second frequency domain noise signal and the estimation signal, and performing inverse Fourier transform to obtain a time domain error signal.
Optionally, the frequency-domain noise signal includes a first frequency-domain noise signal and a second frequency-domain noise signal, and the second processing module processes the frequency-domain noise signal by using an adaptive filter, and obtaining the error signal includes: and performing block adaptive filtering processing on the first frequency domain noise signal by adopting a block frequency domain adaptive filtering algorithm so as to reduce the calculated amount.
A third aspect of an embodiment of the present invention provides an electronic device, including: a memory, a processor and a computer program stored on the memory and operable on the processor, wherein the processor executes the program to implement the method for calculating a noise power spectral density according to the first aspect of the embodiment of the present invention.
A fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method for calculating the noise power spectral density according to the first aspect of the embodiments of the present invention.
It can be known from the foregoing embodiments of the present invention that, in the noise power spectral density calculation method, system, device, and medium provided by the present invention, when a noise signal is processed, an adaptive filter is used to process the signal, and an NLMS algorithm in the adaptive filter is used to replace wiener solution, so that noise estimation is more accurate, the calculated amount is reduced, and the subsequent noise reduction effect is better. Furthermore, in the adaptive filter algorithm part, the error signal is solved by the block frequency domain adaptive filtering algorithm, so that the calculated amount can be further reduced.
Drawings
For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:
fig. 1 schematically shows a flow chart of a noise power spectral density calculation method according to an embodiment of the present invention.
Fig. 2 schematically shows a flow chart of a noise power spectral density calculation method according to an embodiment of the present invention.
Fig. 3 schematically shows a flow chart of NLMS algorithm according to an embodiment of the present invention.
Fig. 4 schematically shows a flow chart of a noise power spectral density calculation method according to an embodiment of the present invention.
Fig. 5 schematically shows a flow chart of a block frequency domain adaptive filtering algorithm according to an embodiment of the present invention.
Fig. 6 schematically shows a structural diagram of a noise power spectral density calculation system according to an embodiment of the present invention.
Fig. 7 schematically shows a block diagram of a hardware structure of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Noise signals are mostly from random and unspecifically generated noise sources, such as vibration sources, vehicles being rushed through, and loud sounds of people speaking. Public places such as parks, schools, and sports have noises generated by human activities, and sounds generated by various household appliances and household appliances in a room, for example, sounds generated by television, video, air conditioner, range hood, air conditioner, and tableware collision at the time of eating are all noises.
Noise can affect the interactive transmission of normal speech signals. Such as: when the smart speaker is used to listen to a song, when a voice command of 'playing simple song' is given, the voice command generally contains noise generated by household appliances or household appliances, so that the smart speaker may not accurately recognize the voice command. The noise power spectral density calculating method, the system, the equipment and the medium provided by the invention can quickly and accurately calculate the power spectral density of the noise so as to perform noise reduction treatment, can greatly improve the accuracy of intelligent sound box identification, and provide better experience for the public. For another example: when the noise reduction processing is carried out by using the double-wheat spectrum subtraction method, a power spectrum X of a signal with noise is requiredPSDSum noise signal power spectrum NPSDAnd then, carrying out subtraction processing, and then recovering by using the phase of the original signal with noise so as to obtain the signal after noise reduction. In this case, N can be obtained by using the method, system, apparatus and medium for calculating noise power spectral densityPSDAnd substituting into spectral subtraction. The algorithm requires NPSDThe estimate is as close to the true value as possible. Also for example: when the wiener filtering algorithm is used for noise reduction, the wiener solution W (also called gain or weight) of each frequency band also needs XPSDAnd NPSDAnd updating by using the calculated values each time. And substituting the weight into a formula to obtain weight, then carrying out conjugate weighting on the received signal X to obtain Y, finally carrying out inverse transformation back to a time domain, and recovering the time domain noise reduction signal Y by using an overlap-add or overlap-save method. Where N isPSDThe noise power spectral density calculation method, the system, the equipment and the medium provided by the invention can be used for more accurately estimating and reducing the calculation amount.
The noise power spectrum density calculation method, the system, the equipment and the medium provided by the invention have wide application scenes and relate to a noise signal power spectrum NPSDThe present invention can be applied to the estimation of (d). The present invention will be described in detail below.
Referring to fig. 1, fig. 1 is a schematic flow chart of a method for calculating a noise power spectral density according to an embodiment of the present invention, the method mainly includes the following steps:
s101, collecting a time domain noise signal, and processing the time domain noise signal to obtain a frequency domain noise signal.
The device for collecting the time-domain noise signal may be a sound sensor, a microphone, etc., and the present invention is not limited thereto. In this embodiment, two time domain noise signals are collected. When the time-domain noise signal is solved at a later stage to obtain an error signal, the specific solving process is realized in a frequency domain, so that the time-domain noise signal needs to be processed and converted into a frequency-domain noise signal.
And S102, processing the frequency domain noise signal by adopting a self-adaptive filter to obtain a time domain error signal.
An adaptive filter is a filter that automatically adjusts the filter parameters at the current time in an iterative process according to some predetermined criteria using the results of the filter parameters obtained at the previous time to adapt to the unknown or time-varying statistical characteristics of the signal and noise, thereby achieving optimal filtering. The method does not need prior knowledge about input signals, the calculation amount is small, the solution is a process of gradual convergence, and the error is smaller and smaller along with the convergence of the filter. The convergence range is the state where the result can approach the steady state only when the algorithm sends out how many sampling points.
In this implementation, the adaptive filter is used to replace the wiener solution portion to solve the time-domain noise signal to obtain an error signal. In the solving process, the statistical characteristic of the time domain noise signal adaptively input by the filtering parameter of the adaptive filter along with the time change is continuously iterated, the filtering coefficient is updated, and then the error signal can be obtained.
The algorithm of the adaptive filter can be an NLMS algorithm or a block frequency domain adaptive filtering algorithm, the specific algorithm is not limited, and other algorithms (such as a Leaky NLMS algorithm) are adopted, and the error signal of the time domain noise signal can also be obtained by combining the gradual convergence solving characteristic of the adaptive filter.
And S103, calculating the power spectral density of the time domain error signal, and calculating the noise power spectral density according to the power spectral density of the error signal.
In the embodiment, the adaptive filter replaces a wiener solving part, so that the calculation accuracy of the noise power spectral density can be improved, and the calculation amount of a noise signal can be reduced.
Referring to fig. 2, fig. 2 is a schematic flow chart of a method for calculating a noise power spectral density according to an embodiment of the present invention, the method mainly includes the following steps:
s201, collecting a time domain noise signal, and processing the time domain noise signal to obtain a frequency domain noise signal.
In the above operation S201, a noise signal is collected by using two microphones, where the collected noise signal is a time-domain noise signal, where one microphone collects a first time-domain noise signal, and the other microphone collects a second time-domain noise signal;
respectively framing and windowing the collected first time domain noise signal and the collected second time domain noise signal; and performing fast Fourier transform on the signals subjected to framing and windowing to obtain a first frequency domain noise signal corresponding to the first time domain noise signal and a second frequency domain noise signal corresponding to the second time domain noise signal.
S202, processing the frequency domain noise signal by using a self-adaptive filter to obtain a time domain error signal, wherein the algorithm part of the self-adaptive filter adopts a normalized least mean square error algorithm.
In operation S202, the adaptive filter algorithm used by the adaptive filter is a Normalized Least Mean Square (NLMS) algorithm, where the NLMS redefines a correction speed μ used to adjust the weighting parameter in the Least Mean Square (LMS) algorithm, so that the μ changes with the input filter signal in a regularization manner, and the stability of convergence is effectively improved.
Referring to fig. 3, fig. 3 is a flowchart of the NLMS algorithm, which includes the following steps:
the coefficients of the adaptive filter are initialized to 0, or may be initialized according to other criteria, and the specific initialization process is not limited in the present invention.
Taking the first frequency domain noise signal obtained in step S201 as an adaptive filter input x (n), and performing NLMS algorithm processing on the signal to obtain a filter impulse response signal h (n) of the first frequency domain noise signal, where the formula of the filter impulse response is:
where α is a fixed step size, ξ is a regularization factor, x (n) is an input signal vector, and x (n) ═ x (n-M +1), x (n-M +2).. x (n).
And performing convolution processing on the first frequency domain noise signal and the corresponding filtering impact response signal to obtain an estimation signal, namely an estimation signal y (n) ═ x (n) × h (n), wherein the middle asterisk represents convolution.
The second frequency-domain noise signal obtained in operation S201 is subtracted from the estimated signal, and is inverse fast fourier transformed to obtain a time-domain error signal, i.e., a time-domain error signal e (n) ═ d (n) — y (n) ═ d (n) — x (n) × h (n).
S203, calculating the power spectral density of the time domain error signal, and calculating the noise power spectral density according to the power spectral density of the error signal.
Calculating the power spectral density of the time domain error signal as follows:
ΓEE(k)＝FFT(γee(τ))＝FFT(e(n)·e(n+τ))
wherein e (n) represents a time-domain error signal, γee(τ) is the autocorrelation function of e (n), τ is the time delay, ΓEE(k) Power spectral density, FFT (gamma) representing e (n)ee(τ)) represents a pair γee(τ) performing a fast Fourier transform operation. A fast fourier transform operation refers to the transformation of a signal from the original domain (usually time or space) to a representation in the frequency domain or vice versa. The FFT can rapidly calculate such a transform by decomposing the DFT matrix into products of sparse (mostly zero) factors, and from the power spectral density function theory (PSD function), the fast fourier transform operation is performed on the autocorrelation function of a time domain signal to obtain the power spectral density function of the time domain signal, and further obtain the power of the time domain signalSpectral density.
The noise power spectral density is calculated as follows:
ΓRR(k)＝FFT(γrr(τ))＝FFT(r(n)·r(n+τ))
ΓLL(k)＝FFT(γll(τ))＝FFT(l(n)·l(n+τ))
wherein, gamma isroot(k) Is an intermediate variable, and Re represents a real part; gamma rayrr(τ) is an autocorrelation function of the first time domain noise signal, τ is a time delay, γll(τ) is an autocorrelation function of the second time-domain noise signal, τ being a time delay; gamma-shapedEE(k) Denotes the power spectral density, Γ, of e (n)RR(k) Representing the power spectral density, Γ, of the first time domain noise signalLL(k) Representing the power spectral density, Γ, of the second time-domain noise signalNN(k) Represents the power spectral density of the resulting noise;
model representing the spatial correlation function in the scattered noise field environment, dLRRepresenting the distance between the two microphones in meters. c represents the speed of sound propagation in the medium, and the unit is meter/second; h (k) is the adaptive filter frequency domain coefficient, k represents the frequency domain index.
Compared with wiener solution, the noise power spectral density calculation method based on the embodiment has higher precision, lower calculation amount and better subsequent noise reduction effect.
Referring to fig. 4, fig. 4 is a schematic flow chart of a noise power spectral density calculation method according to an embodiment of the present invention, which is different from the previous embodiment in that: in the adaptive filter algorithm portion, a block frequency domain adaptive filtering algorithm is adopted to further reduce the calculation amount, the following detailed description is given to the difference portion, the other portions are not described again, and details are not given, please refer to the first embodiment, and the method mainly includes the following steps:
s401, collecting a time domain noise signal, and processing the time domain noise signal to obtain a frequency domain noise signal.
S402, processing the frequency domain noise signal by using a self-adaptive filter to obtain a time domain error signal, wherein an algorithm part of the self-adaptive filter adopts a block frequency domain self-adaptive filtering algorithm.
According to the digital signal processing theory, the overlap-save and overlap-add algorithms provide two efficient algorithms for fast convolution operation, namely calculating linear convolution by using DFT. And when the filters are overlapped by fifty percent (when the block size is equal to the number of the weights), the operation efficiency is highest, so that when the adaptive filter is adopted to solve the power spectral density of the noise signal, a block frequency domain adaptive filtering algorithm is adopted to further reduce the calculation amount.
Referring to fig. 5, fig. 5 is a specific flowchart of the block frequency domain adaptive filtering algorithm, and as shown in fig. 5, the following specific example illustrates the algorithm process:
the first frequency domain noise signal obtained in operation S401 is used as an adaptive filter input x (n), and if the block length of the block is 64 sampling points, two pieces of data are connected in series to obtain 128-bit time domain point data, and the 128-bit time domain point data x (k) is obtained after fast fourier transform.
The length w (k) of the adaptive filter is 64 points, and if zero is padded to 128 points later, the length w (k) is obtained after fast fourier transform. Multiplying the corresponding points of X (k) and W (k) to obtain a signal of Y (k), and performing inverse fast Fourier transform to obtain an estimated signal y (n), wherein the obtained data is 128 points, and since the first 64 points are damaged data as a result of circular convolution, the last 64 points are taken as an output y (n), and then subtracting y (n) from an expected response d (n) (second frequency domain noise signal) to obtain an error signal e (n), and wherein the error signal y (n) is discarded by y (n). Therefore, the e (n) signal needs to be transformed into the frequency domain E (k) by fast Fourier transform after the zero padding to 128-point data. When x (k) is multiplied by W (k), the first 64 points are discarded, then x (k) is transposed, e (k) is multiplied by points, then 64 points are discarded after inverse fast fourier transform, similarly, zero is filled to 128 points after transform to frequency domain, and then the filter coefficient W (k +1) of the filter in the frequency domain is updated. This is repeated iteratively, and as the filter converges, the error will be smaller and smaller.
In this embodiment, in the block frequency domain adaptive filtering algorithm, the time domain length of each block is 128 points, the effective length of the frequency domain is 65 points, and the number of Blocks used by each filter may be 7 to 13. The specific time domain length, frequency domain effective length and fast number are set according to the actual signal processing requirements without limitation.
And S403, calculating the power spectral density of the time domain error signal, and calculating the noise power spectral density according to the power spectral density of the error signal.
In this embodiment, a blocking frequency domain adaptive filtering algorithm is further adopted to perform blocking processing on the noise signal, so that the calculation amount can be further reduced.
Referring to fig. 6, fig. 6 is a diagram illustrating a noise power spectral density calculation system according to an embodiment of the present invention, which can be embedded in an electronic device, the noise power spectral density calculation system mainly includes: a first processing module 601, a second processing module 602, and a calculating module 603.
The first processing module 601 collects a time domain noise signal, and processes the time domain noise signal to obtain a frequency domain noise signal. Specifically, the first processing module 601 collects noise signals by using dual microphones, where the collected noise signals are time-domain signals, one microphone collects a first time-domain noise signal, and the other microphone collects a second time-domain noise signal; respectively framing and windowing the collected first time domain noise signal and the collected second time domain noise signal; and converting the signals subjected to framing and windowing into a first frequency domain noise signal corresponding to the first time domain noise signal and a second frequency domain noise signal corresponding to the second time domain noise signal through fast Fourier transform.
The second processing module 602 processes the frequency domain noise signal by using an adaptive filter to obtain a time domain error signal, wherein an algorithm portion of the adaptive filter is a normalized least mean square error algorithm.
Specifically, the coefficients of the adaptive filter may be initialized to 0, or may be initialized according to other criteria.
The first frequency domain noise signal obtained by the first processing module 601 is used as an adaptive filter input x (n) of the second processing module 602, and the signal is processed by the NLMS algorithm to obtain a filter impulse response signal h (n) of the first frequency domain noise signal, wherein the filter impulse response formula is as follows:
where α is a fixed step size, ξ is a regularization factor, x (n) is an input signal vector, and x (n) ═ x (n-M +1), x (n-M +2) … x (n) ].
And performing convolution processing on the first frequency domain noise signal and the corresponding filtering impact response signal to obtain an estimation signal, namely an estimation signal y (n) ═ x (n) × h (n), wherein the middle asterisk represents convolution.
The second processing module 602 subtracts the second frequency-domain noise signal obtained by the first processing module 601 from the estimated signal, and performs inverse fourier transform to obtain a time-domain error signal, i.e., a time-domain error signal e (n) ═ d (n) — x (n) × (n) ((n)).
In order to further reduce the calculation amount, when the adaptive filter processes signals, a blocking frequency domain adaptive filtering algorithm is adopted to process the acquired noise signals.
Specifically, the first frequency domain noise signal obtained by the first processing module 601 is used as the adaptive filter input x (n) of the first processing module 602, and if the block length of the block is 64 sampling points, two pieces of data are connected in series to obtain 128-bit time domain point data, and the 128-bit time domain point data x (k) is obtained after fast fourier transform.
The length w (k) of the adaptive filter is 64 points, and if zero is padded to 128 points later, the length w (k) is obtained after fast fourier transform. Multiplying the corresponding points of X (k) and W (k) to obtain a signal of Y (k), and performing inverse fast Fourier transform to obtain an estimated signal y (n), wherein the obtained data is 128 points, and since the first 64 points are damaged data as a result of circular convolution, the last 64 points are taken as an output y (n), and then subtracting y (n) from an expected response d (n) (second frequency domain noise signal) to obtain an error signal e (n), and wherein the error signal y (n) is discarded by y (n). Therefore, the e (n) signal needs to be transformed into the frequency domain E (k) by fast Fourier transform after the zero padding to 128-point data. When x (k) is multiplied by W (k), the first 64 points are discarded, then x (k) is transposed, e (k) is multiplied by points, then 64 points are discarded after inverse fast fourier transform, similarly, zero is filled to 128 points after transform to frequency domain, and then the filter coefficient W (k +1) of frequency domain is updated. This is iterated continuously, and as the filter converges, the error becomes smaller and smaller, with k representing the frequency domain index.
In this embodiment, in the block frequency domain adaptive filtering algorithm, the time domain length of each block is 128 points, the effective length of the frequency domain is 65 points, and the number of Blocks used by each filter may be 7 to 13. The specific time domain length, frequency domain effective length and number of blocks are not limited in the present invention, and are specifically set according to the actual signal processing requirements.
The calculating module 603 calculates a power spectral density of the time-domain error signal, and calculates a noise power spectral density according to the power spectral density of the error signal.
Calculating the power spectral density of the time domain error signal as follows:
ΓEE(k)＝FFT(γee(τ))＝FFT(e(n)·e(n+τ))
e (n) represents a time-domain error signal, γee(τ) is the autocorrelation function of e (n), τ is the time delay, FFT () represents a fast Fourier transform operation, ΓEE(k) Denotes the power spectral density of e (n).
The noise power spectral density is calculated as follows:
ΓRR(k)＝FFT(γrr(τ))＝FFT(r(n)·r(n+τ))
ΓLL(k)＝FFT(γll(τ))＝FFT(l(n)·l(n+τ))
wherein, gamma isroot(k) Is an intermediate variable, and Re represents a real part; gamma rayrr(τ) is an autocorrelation function of the first time domain noise signal, τ is a time delay, γll(f) Is the autocorrelation function of the second time domain noise signal, τ is the time delay; gamma-shapedEE(k) Denotes the power spectral density, Γ, of e (n)RR(k) Representing the power spectral density, Γ, of the first time domain noise signalLL(k) Representing the power spectral density, Γ, of the second time-domain noise signalNN(k) Represents the power spectral density of the resulting noise;
model representing the spatial correlation function in the scattered noise field environment, dLRRepresenting the distance between the two microphones in meters. c represents the speed of sound propagation in the medium in meters per second, and H (k) is the adaptive filter frequency domain coefficient.
Referring to fig. 7, fig. 7 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the invention.
The electronic device described in this embodiment includes:
a memory 71, a processor 72 and a computer program stored on the memory 71 and executable on the processor, the processor when executing the program implementing the noise signal power spectral density calculation method as described in the embodiments of fig. 1 or fig. 2 or fig. 4.
Further, the electronic device further includes:
at least one input device 73; at least one output device 74.
The memory 71, processor 72 input device 73 and output device 74 are connected by a bus 75.
The input device 73 may be a camera, a touch panel, a physical button, a mouse, or the like. The output device 74 may be embodied as a display screen.
The Memory 71 may be a high-speed Random Access Memory (RAM) Memory or a non-volatile Memory (non-volatile Memory), such as a magnetic disk Memory. The memory 71 is used for storing a set of executable program codes, and the processor 72 is coupled to the memory 71.
Further, an embodiment of the present disclosure also provides a computer-readable storage medium, where the computer-readable storage medium may be provided in the terminal in the foregoing embodiments, and the computer-readable storage medium may be the memory in the foregoing embodiment shown in fig. 7. The computer-readable storage medium has stored thereon a computer program which, when executed by a processor, implements the noise signal power spectral density calculation method described in the foregoing embodiments shown in fig. 1 or fig. 2 or fig. 4. Further, the computer-readable storage medium may be various media that can store program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described embodiments are merely illustrative, and for example, the division of the modules is merely a logical division, and an actual implementation may have another division, for example, a plurality of modules or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication link may be through some interfaces, and the indirect coupling or communication link of the modules may be in an electrical, mechanical or other form.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical modules, may be located in one place, or may be distributed on a plurality of network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present disclosure may be integrated into one processing module, or each module may exist alone physically, or two or more modules are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode.
It is noted that while for simplicity of explanation, the foregoing method embodiments have been described as a series of acts or combination of acts, it will be appreciated by those skilled in the art that the present disclosure is not limited by the order of acts, as some steps may, in accordance with the present disclosure, occur in other orders and concurrently. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred and that no acts or modules are necessarily required of the disclosure.
In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
The above description of the method, system, apparatus and medium for calculating the power spectral density of a noise signal provided by the present disclosure is provided for those skilled in the art, and the concepts according to the embodiments of the present disclosure may be changed in the aspects of the specific implementation and the application range.
Claims (12)
1. A method of calculating a noise power spectral density, comprising:
collecting a time domain noise signal, and processing the time domain noise signal to obtain a frequency domain noise signal;
processing the frequency domain noise signal by adopting a self-adaptive filter to obtain a time domain error signal;
and calculating the power spectral density of the time domain error signal, and calculating the noise power spectral density according to the power spectral density of the time domain error signal.
2. The method according to claim 1, wherein the frequency-domain noise signal comprises a first frequency-domain noise signal and a second frequency-domain noise signal, and the processing the frequency-domain noise signal with the adaptive filter to obtain the time-domain error signal comprises:
carrying out normalized minimum mean square error processing on the first frequency domain noise signal to obtain a filtering impact response signal corresponding to the first frequency domain noise signal;
performing convolution processing on the first frequency domain noise signal and the corresponding filtering impact response signal to obtain an estimation signal;
and performing subtraction operation on the second frequency domain noise signal and the estimation signal, and performing inverse Fourier transform to obtain the time domain error signal.
3. The method according to claim 2, wherein the frequency domain noise signal comprises a first frequency domain noise signal and a second frequency domain noise signal, and the processing the frequency domain noise signal with the adaptive filter to obtain the time domain error signal comprises:
and performing block adaptive filtering processing on the first frequency domain noise signal by adopting a block frequency domain adaptive filtering algorithm.
4. The method of calculating the noise power spectral density of claim 1, further comprising:
and updating the filter coefficient of the self-adaptive filter according to the time domain error signal.
5. The method according to claim 1, wherein the time-domain noise signal comprises a first time-domain noise signal and a second time-domain noise signal, and wherein the acquiring the time-domain noise signal and the processing the time-domain noise signal comprises:
and framing and windowing the first time domain noise signal and the second time domain noise signal, and performing fast Fourier transform to obtain a first frequency domain noise signal corresponding to the first time domain noise signal and a second frequency domain noise signal corresponding to the second time domain noise signal.
6. The method according to claim 3, wherein during the adaptive filtering process of the first frequency-domain noise signal, partial removal and zero padding are performed on the point location data corresponding to the first frequency-domain noise signal.
7. The method of claim 1, wherein the power spectral density of the time domain error signal is calculated as follows:
ΓEE(k)＝FFT(γee(τ))＝FFT(e(n)·e(n+τ))
e (n) represents the time-domain error signal, γee(τ) is the autocorrelation function of e (n), τ is the time delay, FFT () represents a fast Fourier transform operation, ΓEE(k) Denotes the power spectral density of e (n).
8. A noise power spectral density computation system, comprising:
the first processing module is used for collecting a time domain noise signal and processing the time domain noise signal to obtain a frequency domain noise signal;
the second processing module is used for processing the frequency domain noise signal by adopting a self-adaptive filter to obtain a time domain error signal;
and the calculating module is used for calculating the power spectral density of the time domain error signal and calculating the noise power spectral density according to the power spectral density of the error signal.
9. The system according to claim 8, wherein the frequency-domain noise signal comprises a first frequency-domain noise signal and a second frequency-domain noise signal, and the second processing module uses an adaptive filter to process the frequency-domain noise signal to obtain a time-domain error signal comprises:
carrying out normalized minimum mean square error processing on the first frequency domain noise signal to obtain a filtering impact response signal corresponding to the first frequency domain noise signal;
performing convolution processing on the first frequency domain noise signal and the corresponding filtering impact response signal to obtain an estimation signal;
and performing subtraction operation on the second frequency domain noise signal and the estimation signal, and performing inverse Fourier transform to obtain the time domain error signal.
10. The method of claim 8, wherein the frequency domain noise signal comprises a first frequency domain noise signal and a second frequency domain noise signal, and the second processing module processes the frequency domain noise signal with an adaptive filter to obtain a time domain error signal comprises:
and performing block adaptive filtering processing on the first frequency domain noise signal by adopting a block frequency domain adaptive filtering algorithm.
11. An electronic device, comprising:
a processor;
a memory storing a computer executable program which, when executed by the processor, causes the processor to perform the noise power spectral density calculation method of any one of claims 1-7.
12. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the noise power spectral density calculation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910612851.4A CN112201273A (en) | 2019-07-08 | 2019-07-08 | Noise power spectral density calculation method, system, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910612851.4A CN112201273A (en) | 2019-07-08 | 2019-07-08 | Noise power spectral density calculation method, system, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112201273A true CN112201273A (en) | 2021-01-08 |
Family
ID=74004516
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910612851.4A Pending CN112201273A (en) | 2019-07-08 | 2019-07-08 | Noise power spectral density calculation method, system, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112201273A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113808608A (en) * | 2021-09-17 | 2021-12-17 | 随锐科技集团股份有限公司 | Single sound channel noise suppression method and device based on time-frequency masking smoothing strategy |
| CN114112006A (en) * | 2021-11-26 | 2022-03-01 | 中科传启（苏州）科技有限公司 | Noise monitoring method and device and electronic equipment |
| WO2023138252A1 (en) * | 2022-01-24 | 2023-07-27 | Oppo广东移动通信有限公司 | Audio signal processing method and apparatus, earphone device, and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1188547A (en) * | 1995-06-21 | 1998-07-22 | 艾利森电话股份有限公司 | Power spectral density estimation method and apparatus |
| CN101369427A (en) * | 2007-08-13 | 2009-02-18 | 哈曼贝克自动系统股份有限公司 | Noise reduction by combined beamforming and post-filtering |
| CN102938254A (en) * | 2012-10-24 | 2013-02-20 | 中国科学技术大学 | Voice signal enhancement system and method |
| KR20150054424A (en) * | 2013-11-12 | 2015-05-20 | 고려대학교 산학협력단 | Residual echo cancellation method and apparatus |
| CN107393550A (en) * | 2017-07-14 | 2017-11-24 | 深圳永顺智信息科技有限公司 | Method of speech processing and device |
| US20190122685A1 (en) * | 2017-10-19 | 2019-04-25 | Nxp B.V. | Signal processor for signal enhancement and associated methods |
| US20190131950A1 (en) * | 2017-10-26 | 2019-05-02 | Bose Corporation | Noise estimation using coherence |
-
2019
- 2019-07-08 CN CN201910612851.4A patent/CN112201273A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1188547A (en) * | 1995-06-21 | 1998-07-22 | 艾利森电话股份有限公司 | Power spectral density estimation method and apparatus |
| CN101369427A (en) * | 2007-08-13 | 2009-02-18 | 哈曼贝克自动系统股份有限公司 | Noise reduction by combined beamforming and post-filtering |
| CN102938254A (en) * | 2012-10-24 | 2013-02-20 | 中国科学技术大学 | Voice signal enhancement system and method |
| KR20150054424A (en) * | 2013-11-12 | 2015-05-20 | 고려대학교 산학협력단 | Residual echo cancellation method and apparatus |
| CN107393550A (en) * | 2017-07-14 | 2017-11-24 | 深圳永顺智信息科技有限公司 | Method of speech processing and device |
| US20190122685A1 (en) * | 2017-10-19 | 2019-04-25 | Nxp B.V. | Signal processor for signal enhancement and associated methods |
| US20190131950A1 (en) * | 2017-10-26 | 2019-05-02 | Bose Corporation | Noise estimation using coherence |
Non-Patent Citations (1)
| Title |
|---|
| 罗侃;王华;张敏;: "抽头动态更新频域LMS算法研究", 计算机应用研究, no. 12 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113808608A (en) * | 2021-09-17 | 2021-12-17 | 随锐科技集团股份有限公司 | Single sound channel noise suppression method and device based on time-frequency masking smoothing strategy |
| CN113808608B (en) * | 2021-09-17 | 2023-07-25 | 随锐科技集团股份有限公司 | Method and device for suppressing mono noise based on time-frequency masking smoothing strategy |
| CN114112006A (en) * | 2021-11-26 | 2022-03-01 | 中科传启（苏州）科技有限公司 | Noise monitoring method and device and electronic equipment |
| WO2023138252A1 (en) * | 2022-01-24 | 2023-07-27 | Oppo广东移动通信有限公司 | Audio signal processing method and apparatus, earphone device, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109727604B (en) | Frequency domain echo cancellation method for speech recognition front end and computer storage medium | |
| EP2360685B1 (en) | Noise suppression | |
| CN112201273A (en) | Noise power spectral density calculation method, system, equipment and medium | |
| CN108010536B (en) | Echo cancellation method, device, system and storage medium | |
| CN111971975B (en) | Active noise reduction method, system, electronic equipment and chip | |
| RU2000120915A (en) | METHODS AND DEVICES FOR PROVIDING COMFORTABLE NOISE IN COMMUNICATION SYSTEMS | |
| CN110211602B (en) | Intelligent voice enhanced communication method and device | |
| Gil-Cacho et al. | Wiener variable step size and gradient spectral variance smoothing for double-talk-robust acoustic echo cancellation and acoustic feedback cancellation | |
| KR102190833B1 (en) | Echo suppression | |
| JP6204312B2 (en) | Sound collector | |
| CN110556125B (en) | Feature extraction method and device based on voice signal and computer storage medium | |
| CN112435683B (en) | Adaptive noise estimation and voice noise reduction method based on T-S fuzzy neural network | |
| CN112602150A (en) | Noise estimation method, noise estimation device, voice processing chip and electronic equipment | |
| WO2022256577A1 (en) | A method of speech enhancement and a mobile computing device implementing the method | |
| CN110021289B (en) | Sound signal processing method, device and storage medium | |
| CN112820312B (en) | Voice separation method and device and electronic equipment | |
| CN109379501B (en) | Filtering method, device, equipment and medium for echo cancellation | |
| CN110913312B (en) | Echo cancellation method and device | |
| JP6517124B2 (en) | Noise suppression device, noise suppression method, and program | |
| CN112997249B (en) | Voice processing method, device, storage medium and electronic equipment | |
| JP5438629B2 (en) | Stereo echo canceling method, stereo echo canceling device, stereo echo canceling program | |
| CN113763984A (en) | Parameterized noise elimination system for distributed multiple speakers | |
| CN109493878B (en) | Filtering method, device, equipment and medium for echo cancellation | |
| CN113205824A (en) | Sound signal processing method, device, storage medium, chip and related equipment | |
| JP3616341B2 (en) | Multi-channel echo cancellation method, apparatus thereof, program thereof, and recording 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 |