CN103811020B - A kind of intelligent sound processing method - Google Patents

Abstract

The invention relates to an intelligent speech processing method, which belongs to the technical field of information processing. The present invention realizes the intelligent identification of the identities of multiple interlocutors in a multi-person voice environment by establishing a voice model library of interlocutors, and at the same time separates the mixed voice to obtain the identity of each interlocutor. Independent voice, according to the user's needs, amplifies the voice of the interlocutor to be heard by the user and eliminates the voice of the interlocutor not required by the user; different from traditional hearing aids, this method can automatically provide the user with the desired voice according to the user's personal needs, The interference of non-target human voices except noise is reduced, reflecting the individuation, interaction and intelligence of the method.

Description

A kind of intelligent voice processing method

technical field

The invention belongs to the technical field of information processing, and in particular relates to an intelligent voice processing method.

Background technique

According to the latest assessment data released by the World Health Organization (WHO) in 2013, there are currently 360 million people in the world with hearing impairments of varying degrees, accounting for 5% of the total global population. The use of hearing aid products can effectively compensate for the hearing loss of hearing-impaired patients and improve their quality of life and work. However, the current research on hearing aid system technology still focuses on noise suppression and source sound amplitude amplification, and rarely involves sound feature-based modeling and multi-sound source automatic separation technology. When the actual application scenario is very complex, for example: at a party, multiple speakers speak at the same time, even accompanied by background sounds such as music, since the hearing aid system cannot separate the sound object of interest from the mixed sound input, it is simple The sound intensity expansion function of the mobile phone can only increase the hearing burden or even damage the user, and will not bring effective sound input and understanding. Therefore, aiming at the technical defects of the current hearing aid system, it is of great significance to design a new hearing aid system with specific sound object recognition function, which is more intelligent and personalized.

Contents of the invention

Aiming at the deficiencies of the existing technology, the present invention proposes an intelligent voice processing method to ensure that users can obtain pure voice reception and amplification according to their own needs, and realize the purpose of intelligence, interaction and personalization of the hearing aid system.

A kind of intelligent speech processing method, comprises the following steps:

Step 1. Collect sample speech segments to build a sample speech library, perform feature extraction on the sample speech, obtain feature parameters, and train the feature parameters;

The specific process is as follows:

Step 1-1, collecting sample speech segments, discretizing the collected speech segments, extracting the Mel frequency cepstral coefficients of the speech signal as the speech signal characteristic parameters, and establishing a Gaussian mixture model;

The model formula is as follows:

$\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; XIG</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Wherein, p (XIG) represents the probability that the sample speech characteristic parameter is in the model of G in model parameter;

G represents the Gaussian mixture model parameter set, G={p _i , μ _i , ∑ _i }, i=1, 2,..., I;

I represents the number of single Gaussian models in the Gaussian mixture model;

p _i represents the weight coefficient of the i-th single Gaussian model,

μ _i represents the mean vector of the i-th single Gaussian model;

∑ _i represents the covariance matrix of the i-th single Gaussian model;

X represents the sample speech feature parameter, X={x ₁ , x ₂ ,..., x _T }, and T represents the number of feature vectors;

b _i (X) represents the density function of the i-th single Gaussian model, b _i (X)=N(μ _i , ∑ _i ), N(.) represents the density function of the standard Gaussian distribution;

Step 1-2, utilize speech signal feature parameter training Gaussian mixture model;

That is, the k-means clustering algorithm is used to cluster the characteristic parameters of the speech signal, and the initial value of the Gaussian mixture model parameter set G ₀ = {p _i ⁰ , μ _i ⁰ , ∑ _i ⁰ }, i=1, 2,... , I; and according to the initial value of the obtained Gaussian mixture model parameter set, adopt the maximum expectation algorithm to estimate the model, and then obtain the Gaussian mixture model parameters, that is, complete the training of the characteristic parameters;

Step 2, adopting a microphone array composed of M microphones to collect the measured environmental audio signal, determining the number of environmental sound sources and the direction of arrival of each sound source beam, that is, the incident angle of the sound source to the microphone array;

The specific process is as follows:

Step 2-1, using a microphone array composed of M microphones to collect the mixed audio signal of the environment under test, and discretize the collected mixed audio signal to obtain the amplitude of each sampling point;

Step 2-2, matrix the amplitude of each sampling point to obtain the mixed audio matrix collected by each microphone; the number of columns of the above mixed audio matrix is one, the number of rows is the number of sampling points, and the elements in the matrix are The amplitude of each sampling point;

Step 2-3, according to the mixed audio matrix collected by each microphone and the number of microphones, obtain the estimated value of the vector covariance matrix of the mixed audio signal of the tested environment;

The formula for estimating the vector covariance matrix is as follows:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; xx</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Wherein, R _xx represents the estimated value of the vector covariance matrix of the mixed audio signal of tested environment;

X(m) represents the mixed audio matrix collected by the mth microphone;

X ^H (m) represents the transpose matrix of the mixed audio matrix collected by the m microphone;

Step 2-4, performing eigenvalue decomposition on the estimated value of the vector covariance matrix, obtaining eigenvalues, and sorting the eigenvalues from large to small, and determining the number of eigenvalues greater than the threshold value, which is the number of sound sources;

Step 2-5, subtracting the number of sound sources from the number of microphones to obtain the number of noise sources, and then correspondingly obtain a noise matrix;

Step 2-6, according to the distance between each microphone and the array center, the wavelength of the mixed audio signal, the direction angle of the microphone to the array center and the beam arrival direction of the sound source to obtain the steering vector of the microphone array, and then according to the noise matrix and the microphone array The steering vector obtains the angular spectrum function of the mixed audio signal;

The angular spectrum function formula of the mixed audio signal is as follows:

$\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}^{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}^{\mathrm{<span\; class="notranslate">\; h</span>}}}_{\mathrm{<span\; class="notranslate">\; u</span>}}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Wherein, P(θ) represents the angular spectrum function of the mixed audio signal;

α(θ) represents the steering vector of the microphone array, α(θ)=(α ₁ (θ), ..., α _m (θ), ..., α _M (θ)), where, j represents the imaginary number unit, k=2π/λ, λ represents the wavelength of the mixed audio signal, d _m represents the distance between the mth microphone and the center of the array, Indicates the direction angle of the mth microphone to the center of the array;

θ represents the beam arrival direction of the sound source;

α ^H (θ) represents the transpose matrix of the steering vector of the microphone array;

V _u represents the noise matrix;

V ^H _u represents the transpose matrix of the noise matrix;

Step 2-7, according to the waveform of the angle spectrum function of the mixed audio signal, select a plurality of peaks of the waveform from large to small, and the number of selected peaks is the number of sound sources;

Steps 2-8, determine the angle value corresponding to the selected peak value, that is, obtain the beam arrival direction of each sound source;

Step 3, according to the audio signal of each sound source, the conversion relationship between the sound source and the microphone, obtain the sound pressure of the microphone array received by the microphone, the sound pressure gradient in the horizontal direction of the microphone array, and the sound pressure gradient in the vertical direction of the microphone array;

The formula of the sound pressure signal of the microphone array is as follows:

${\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; N</span>}}\mathrm{<span\; class="notranslate">\; 0.5</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; mn</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Among them, p _w (t) represents the sound pressure of the microphone array at time t;

N represents the number of sound sources;

t means time;

s _n (t) represents the audio signal of the nth sound source;

h _mn (t) represents the transformation matrix between the nth sound source and the mth microphone, h _mn (t) = p ₀ (t)α _m (θ _n (t)), p ₀ (t) represents t The sound pressure at the center of the microphone array caused by sound waves at time; α _m (θ _n (t)) represents the steering vector of the mth microphone with respect to the _n The beam arrival direction of a sound source;

The formula for the sound pressure gradient in the horizontal direction of the microphone array is as follows:

Wherein, p _x (t) represents the sound pressure gradient in the horizontal direction of the microphone array;

The sound pressure gradient formula in the vertical direction of the microphone array is as follows:

Wherein, p _y (t) represents the sound pressure gradient in the vertical direction of the microphone array;

Step 4, using Fourier transform to convert the sound pressure at the center of the microphone array, the sound pressure gradient in the horizontal direction of the microphone array, and the sound pressure gradient in the vertical direction of the microphone array from the time domain to the frequency domain;

Step 5, according to the microphone array sound pressure in the frequency domain, the microphone array horizontal direction gradient and the microphone array vertical direction sound pressure gradient, obtain the intensity vector formula of the sound pressure signal in the frequency domain, and then deduce the intensity vector direction;

The intensity vector formula of the sound pressure signal in the frequency domain is:

$\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; Re</span>}<span\; class="notranslate">\; \{</span>{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; Re</span>}<span\; class="notranslate">\; \{</span>{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

Wherein, I (ω, t) represents the intensity vector of the sound pressure signal in the frequency domain;

p ₀ represents the measured ambient air density;

c represents the speed of sound;

Re[.] means to take the real part of a complex number;

p _w ^* (ω, t) represents the conjugate matrix of the sound pressure of the microphone array in the frequency domain;

p _x (ω, t) represents the sound pressure gradient in the horizontal direction of the microphone array in the frequency domain;

p _y (ω, t) represents the sound pressure gradient in the vertical direction of the microphone array in the frequency domain;

u _x represents the unit vector in the direction of the abscissa axis;

u _y represents the unit vector in the direction of the ordinate axis;

The intensity vector direction formula is as follows:

$\mathrm{<span\; class="notranslate">\; \&gamma;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; the\; tan</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; [</span>\frac{\mathrm{<span\; class="notranslate">\; Re</span>}<span\; class="notranslate">\; \{</span>{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>}{{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

Wherein, γ (ω, t) represents the intensity vector direction of the sound pressure signal of the mixed sound received by the microphone array;

Step 6. Perform statistics on the intensity vector direction to obtain its probability density distribution, and use the mixed von Misius distribution for fitting to obtain the model parameters that the voice intensity vector direction obeys the mixed von Misius distribution, and then obtain the intensity vector direction function of each sound pressure signal ;

The specific process is as follows:

Step 6-1. Perform statistics on the direction of the intensity vector to obtain its probability density distribution, and use a mixed von Misius distribution for fitting to obtain a model parameter set of the mixed von Misius distribution that the intensity vector direction of the speech obeys;

The formula of the mixed von Misius distribution model is as follows:

$\mathrm{<span\; class="notranslate">\; g</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

in, Represents the mixed von Misius distribution probability density;

Indicates the mixed sound direction angle;

α _n represents the weight of the intensity vector direction function of the sound pressure signal of the nth sound source;

Among them, I ₀ (k _n ) represents the first-order modified Bessel function corresponding to the nth sound source, and k _n represents the concentration parameter corresponding to the single von Misius distribution that the intensity vector direction of the sound pressure signal of the nth sound source obeys, That is, the reciprocal of the variance of the von Misius distribution;

The parameter set of the mixed von Misius distribution function is as follows:

Γ={α _n , k _n }, i=1, . . . , N(11)

Step 6-2, initialize the model parameters, and obtain the initial function parameter set;

Step 6-3, according to the obtained initial model parameters, use the maximum expectation algorithm to estimate the parameters of the mixed von Misius distribution model;

Step 6-4, obtain the intensity vector direction function of each sound pressure signal according to the estimated mixed von Misius distribution model parameters;

The formula of the intensity vector direction function of the sound pressure signal is as follows:

${\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>$

in, Represents the intensity vector direction function of the nth sound source;

Step 7, according to the obtained intensity vector direction function of each sound pressure signal and the sound pressure of the microphone array, obtain the signal of each sound source in the frequency domain, and use the inverse Fourier transform to convert each sound source signal in the frequency domain Convert to a sound source signal in the time domain;

The signal formula for each sound source in the frequency domain is as follows:

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 13</span>}<span\; class="notranslate">\; )</span>$

in, Represents the frequency domain signal of the nth sound source signal obtained after the mixed speech separation;

Will The time domain signal is obtained by inverse Fourier transform

Step 8, calculate the matching probability of each sound source signal and the specified sound source in the sample speech library, select the sound source with the largest probability value as the target sound source, keep the sound source signal, and delete other non-target sound sources;

The matching probability formula of each sound source signal and the specified sound source in the sample speech library is as follows:

$\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; log</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 14</span>}<span\; class="notranslate">\; )</span>$

In the formula: voice after separation The extracted speech feature parameters, that is, the extracted speech Mel frequency cepstral coefficients as speech The characteristic parameters;

Indicates the matching probability of the nth sound source signal and the specified sound source in the sample speech library;

G _c represents the voice model parameter of the user designated person;

Indicates the probability that the voice after separation belongs to the voice of the user-designated person;

Step 9, amplifying the retained sound source signal, that is, completing the amplification of the specified sound source in the tested environment.

The threshold described in step 2-4 ranges from 10 ^-2 to 10 ^-16 .

The α _n described in step 6-1 is a random number between 0 and 1, and satisfies k _n is a random number from 1 to 700.

Advantage of the present invention:

An intelligent voice processing method of the present invention, by establishing a voice model library of interlocutors, realizes the intelligent recognition of the identities of multiple interlocutors in a multi-person voice environment, and simultaneously separates and mixes the voices to obtain the independent voice of each interlocutor, and provides the user with a voice according to user needs. Amplify the voice of the interlocutor to be heard while eliminating the voice of the interlocutor who is not required by the user; unlike traditional hearing aids, this method can automatically provide the user with the sound they need according to the user's personal needs, reducing non-targets except noise The interference of human voice reflects the personalization, interaction and intelligence of the method.

Description of drawings

Fig. 1 is the flow chart of the intelligent voice processing method of an embodiment of the present invention;

Fig. 2 is a schematic diagram of modeling sound source data of an embodiment of the present invention, wherein, figure (a) represents the sound of the first person

Data schematic diagram, figure (b) represents the voice data schematic diagram of the second person, and figure (c) represents the voice data schematic diagram of the third person;

Fig. 3 is a schematic diagram of sound source data used for sound mixing in an embodiment of the present invention, wherein, figure (a) represents the data schematic diagram of the first sound source, figure (b) represents the data schematic diagram of the second sound source, and figure (c) ) represents the data schematic diagram of the third sound source;

Fig. 4 is a schematic diagram of a microphone array according to an embodiment of the present invention;

Fig. 5 is a schematic diagram of data received by four microphones in an embodiment of the present invention, wherein, figure (a) represents a schematic diagram of the mixed sound signal received by the first microphone, and figure (b) represents the signal received by the second microphone Mixed sound signal schematic diagram, figure (c) represents the mixed sound signal schematic diagram that the 3rd microphone receives, and figure (d) represents the mixed sound signal schematic diagram that the 4th microphone receives;

Fig. 6 is the schematic diagram of the data sampling received by four microphones of an embodiment of the present invention, wherein, figure (a) represents the schematic diagram of the mixed sound signal received by the first microphone after sampling, and figure (b) represents the second Schematic diagram after the mixed sound signal sampling that a microphone receives, figure (c) represents the schematic diagram after the mixed sound signal sampling that the 3rd microphone receives, and figure (d) represents the schematic diagram after the mixed sound signal sampling that the 4th microphone receives;

FIG. 7 is a schematic diagram of spatial spectrum estimation of a mixed signal according to an embodiment of the present invention;

Fig. 8 is a probability density diagram of the direction distribution of mixed sound vectors according to an embodiment of the present invention;

Fig. 9 is a schematic diagram of a maximum likelihood estimation mixed von Misius model according to an embodiment of the present invention;

Fig. 10 is the ideal speech of an embodiment of the present invention and obtains the speech contrast chart after separation, and wherein, figure (a) is the original sound signal of the first sound source, and figure (b) is the original sound of the first sound source after separation Signal, figure (c) is the original sound signal of the second sound source, figure (d) is the original sound signal of the second sound source after separation, figure (e) is the original sound signal of the third sound source, figure (f) is the original sound signal of the third sound source after separation.

detailed description

An embodiment of the present invention will be further described below in conjunction with the accompanying drawings.

In the embodiment of the present invention, the model system is mainly divided into two modules: a speech modeling module and a speech dynamic real-time processing module, wherein the speech modeling module realizes speaker speech modeling, and the speech dynamic real-time processing module realizes mixed human Direction localization and separation of sound, hybrid speech recognition and extraction (that is, extraction and amplification of the target sound and shielding of other sounds).

A kind of intelligent speech processing method, method flowchart as shown in Figure 1, comprises the following steps:

Step 1. Collect sample speech segments to build a sample speech library, perform feature extraction on sample speech, obtain feature parameters, and train feature parameters; the specific process is as follows:

Step 1-1, recording the sample speech segment in a quiet indoor environment, discretizing the collected speech segment, extracting the Mel frequency cepstral coefficient (MFCC) of the speech signal as the speech signal characteristic parameter, and establishing a Gaussian mixture model;

In the embodiment of the present invention, adopt Windows self-contained recorder to record the voices of 3 people respectively, each person records 2 sections, wherein 1 section is used for sound separation and recognition, and the other 1 section is used for speaker's voice modeling, and the target sound source is set as No. 1 sound source; as shown in Figure 2 (a) to (c), take a piece of speech from three people respectively, establish a Gaussian mixture model for it, and store the obtained model parameters in the model library.

The model formula is as follows:

$\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; XIG</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Wherein, p (XIG) represents the probability that the sample speech characteristic parameter is in the model of G in model parameter;

G represents the Gaussian mixture model parameter set, G={p _i , μ _i , ∑ _i }, i=1, 2,..., I;

I represents the number of single Gaussian models in the Gaussian mixture model;

p _i represents the weight coefficient of the i-th single Gaussian model,

μ _i represents the mean vector of the i-th single Gaussian model;

∑ _i represents the covariance matrix of the i-th single Gaussian model;

X represents the sample speech feature parameter, X={x ₁ , x ₂ ,..., x _T }, and T represents the number of feature vectors;

b _i (X) represents the density function of the i-th single Gaussian model, b _i (X)=N(μ _i , ∑ _i ), N(.) represents the density function of the standard Gaussian distribution;

Step 1-2, utilize speech signal feature parameter training Gaussian mixture model;

That is, the k-means clustering algorithm is used to cluster the characteristic parameters of the speech signal, and the initial value of the Gaussian mixture model parameter set G ₀ = {p _i ⁰ , μ _i ⁰ , ∑ _i ⁰ }, i=1, 2,... , I;

In this example, 16 single Gaussian models are used to form a Gaussian mixture model. Randomly generate 16 vectors as cluster centers, the length of each vector is the number of speech frames, assign the feature parameters of each frame to one of the 16 cluster centers according to the minimum distance criterion, and then recalculate each cluster center vector Take it as the new clustering center until the algorithm converges and calculates. At this time, the obtained clustering center is the initial Gaussian mixture model mean parameter μ _i ⁰ . Calculate the characteristic parameter covariance to obtain the initial ∑ _i ⁰ , p _i ⁰ , the initial value is

The maximum expectation algorithm is used to estimate the model. The principle is that the probability of occurrence of the observed value is the largest. By deriving the model function with respect to the parameters p _i ⁰ , μ _i ⁰ , ∑ _i ⁰ respectively, the parameters p _i , μ _i , ∑ _i are equal to zero. The revaluation of the algorithm until the end of the algorithm convergence calculation, at this time, the training of the characteristic parameters is completed.

Step 2, adopting a microphone array composed of 4 microphones to collect the measured environmental audio signal, determining the number of the environmental sound sources and the arrival direction of each sound source beam, that is, the incident angle of the sound source to the microphone array;

The specific process is as follows:

Step 2-1, using a microphone array composed of 4 microphones to collect the measured environmental audio signal, and discretizing the collected mixed audio signal to obtain the amplitude of each sampling point;

In the embodiment of the present invention, as shown in Figures (a) to (c) in Figure 3, another section of voice of three people is respectively taken as the sound data source of the mixed audio, and four microphones are used, and the array formed by the four microphones is As shown in Figure 4, the No. 1 microphone and the No. 2 microphone are symmetrically distributed on both sides of the horizontal direction around the center of the array, and the No. 3 and No. 4 microphones are symmetrically distributed on both sides of the vertical direction around the center of the array; The mixed data is shown in Figures (a) to (d) in Figure 5, discretize the speech received by the four microphones, the discretization frequency is 12500Hz, and determine the amplitude of each sampling point, as shown in Figure 6 Figures (a) to (d) shown in the middle.

Step 2-2, matrix the amplitude of each sampling point to obtain the mixed audio matrix collected by each microphone; the number of columns of the above mixed audio matrix is one, the number of rows is the number of sampling points, and the elements in the matrix are The amplitude of each sampling point;

Step 2-3, according to the mixed audio matrix collected by each microphone and the number of microphones, obtain the estimated value of the vector covariance matrix of the mixed audio signal of the tested environment;

The formula for estimating the vector covariance matrix is as follows:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; xx</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 4</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Wherein, R _xx represents the estimated value of the vector covariance matrix of the mixed audio signal of tested environment;

X(m) represents the mixed audio matrix collected by the mth microphone;

XH(m) represents the transpose matrix of the mixed audio matrix collected by the mth microphone;

Step 2-4. In this example, perform eigenvalue decomposition on the estimated value of the vector covariance matrix to obtain the eigenvalue [0.00000.01900.03630.1128], and sort the eigenvalues from large to small, and compare with the threshold 10 ^-7 , that is, get 3 eigenvalues, so the number of sound sources is 3;

Step 2-5, subtracting the number of sound sources from the number of microphones to obtain the number of noise sources, and then correspondingly obtain a noise matrix;

In the embodiment of the present invention, the eigenvalues and corresponding eigenvectors equal to the number of sound sources 3 are regarded as the signal part space, and the remaining 4-3, that is, 1 eigenvalue and eigenvector are regarded as the noise part space, namely The number of noise sources is 1, and the noise matrix can be obtained according to the elements corresponding to the noise eigenvalues

V _u =[-0.1218-0.4761i-0.1564+0.4659i-0.5070-0.0374i-0.5084];

Step 2-6, according to the distance between each microphone and the array center, the wavelength of the mixed audio signal, the direction angle of the microphone to the array center and the beam arrival direction of the sound source to obtain the steering vector of the microphone array, and then according to the noise matrix and the microphone array The steering vector obtains the angular spectrum function of the mixed audio signal;

As shown in Figure 4, the distance between each microphone and the array center is 0.02m; in the embodiment of the present invention, the wavelength of the mixed audio signal is 30000; The directional angle of the center is 180°, the directional angle of the No. 3 microphone to the array center is 90°, and the directional angle of the No. 1 microphone to the array center is 270°;

The angular spectrum function formula of the mixed audio signal is as follows:

$\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}^{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}^{\mathrm{<span\; class="notranslate">\; h</span>}}}_{\mathrm{<span\; class="notranslate">\; u</span>}}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Wherein, P(θ) represents the angular spectrum function of the mixed audio signal;

α(θ) represents the steering vector of the microphone array, α(θ)=(α ₁ (θ), α ₂ (θ), α ₃ (θ), α ₄ (θ)), where α ₁ (θ)= e ^{jk0.02cos(0°-θ)} ，α ₂ (θ)＝e ^{jk002cos(180°-θ)} ，α ₃ (θ)＝e ^{jk002cos(90°-θ)} ，α ₄ (θ)＝e ^{jk002cos( 270°-θ)} , j represents the imaginary number unit, k=2π/λ, and λ represents the wavelength of the mixed audio signal;

θ represents the beam arrival direction of the sound source;

α ^H (θ) represents the transpose matrix of the steering vector of the microphone array;

V _u represents the noise matrix;

V ^H _u represents the transpose matrix of the noise matrix;

Step 2-7, according to the waveform of the angle spectrum function of the mixed audio signal, select a plurality of peaks of the waveform from large to small, and the number of selected peaks is the number of sound sources;

Steps 2-8, determine the angle value corresponding to the selected peak value, that is, obtain the beam arrival direction of each sound source;

As shown in Figure 7, the waveform of the angular spectrum function P(θ) of the mixed audio signal, the beam arrival directions of the three sound sources in the mixed sound are [50°, 200°, 300°] respectively.

Step 3, according to the audio signal of each sound source, the conversion relationship between the sound source and the microphone, obtain the sound pressure of the microphone array received by the microphone, the sound pressure gradient in the horizontal direction of the microphone array, and the sound pressure gradient in the vertical direction of the microphone array;

The sound pressure formula of the microphone array is as follows:

${\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 3</span>}}\mathrm{<span\; class="notranslate">\; 0.5</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 4</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; mn</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Among them, p _w (t) represents the sound pressure of the microphone array at time t;

N represents the number of sound sources;

t means time;

s _n (t) represents the audio signal of the nth sound source;

h _mn (t) represents the transformation matrix between the nth sound source and the mth microphone, h _mn (t) = p ₀ (t)α _m (θ _n (t)), p ₀ (t) represents t The sound pressure at the center of the microphone array caused by sound waves at time; α _m (θ _n (t)) represents the steering vector of the mth microphone with respect to the _n The beam arrival direction of a sound source;

The formula for the sound pressure gradient in the horizontal direction of the microphone array is as follows:

Wherein, p _x (t) represents the sound pressure gradient in the horizontal direction of the microphone array;

The sound pressure gradient formula in the vertical direction of the microphone array is as follows:

Wherein, p _y (t) represents the sound pressure gradient in the vertical direction of the microphone array;

Step 4, using Fourier transform to convert the sound pressure at the center of the microphone array, the sound pressure gradient in the horizontal direction of the microphone array, and the sound pressure gradient in the vertical direction of the microphone array from the time domain to the frequency domain;

Step 5, according to the microphone array sound pressure in the frequency domain, the microphone array horizontal direction gradient and the microphone array vertical direction sound pressure gradient, obtain the intensity vector formula of the sound pressure signal in the frequency domain, and then obtain the intensity vector direction;

The intensity vector formula of the sound pressure signal in the frequency domain is:

The intensity vector formula of the sound pressure signal in the frequency domain is:

$\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; Re</span>}<span\; class="notranslate">\; \{</span>{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; Re</span>}<span\; class="notranslate">\; \{</span>{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

Wherein, I (ω, t) represents the intensity vector of the sound pressure signal in the frequency domain;

_ρ0 represents the measured ambient air density;

c represents the speed of sound;

Re[.] means to take the real part of a complex number;

p _w ^* (ω, t) represents the conjugate matrix of the sound pressure of the microphone array in the frequency domain;

p _x (ω, t) represents the sound pressure gradient in the horizontal direction of the microphone array in the frequency domain;

p _y (ω, t) represents the sound pressure gradient in the vertical direction of the microphone array in the frequency domain;

u _x represents the unit vector in the direction of the abscissa axis;

u _y represents the unit vector in the direction of the ordinate axis;

The intensity vector direction formula is as follows:

$\mathrm{<span\; class="notranslate">\; \&gamma;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; the\; tan</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; [</span>\frac{\mathrm{<span\; class="notranslate">\; Re</span>}<span\; class="notranslate">\; \{</span>{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>}{\mathrm{<span\; class="notranslate">\; Re</span>}<span\; class="notranslate">\; \{</span>{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

Wherein, γ (ω, t) represents the intensity vector direction of the sound pressure signal of the mixed sound received by the microphone array;

Step 6. Perform statistics on the intensity vector direction to obtain its probability density distribution, and use the mixed von Misius distribution for fitting to obtain the model parameters that the voice intensity vector direction obeys the mixed von Misius distribution, and then obtain the intensity vector direction function of each sound pressure signal ;

The specific process is as follows:

Step 6-1. Perform statistics on the direction of the intensity vector to obtain its probability density distribution, and use a mixed von Misius distribution for fitting to obtain a model parameter set of the mixed von Misius distribution that the intensity vector direction of the speech obeys;

In the embodiment of the present invention, as shown in Figure 8, the distribution probability density map of γ (ω, t); According to the number and angle of the sound source sought above, the mixed Von Misius distribution conforming to the probability density distribution can be obtained by three Composed of a single von Misius distribution, and the central angles of these three distributions are [50°, 200°, 300°], respectively.

The formula of the mixed von Misius distribution model is as follows:

$\mathrm{<span\; class="notranslate">\; g</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

in, Represents the mixed von Misius distribution probability density;

Indicates the mixed sound direction angle;

α _n represents the weight of the intensity vector direction function of the sound pressure signal of the nth sound source;

Among them, I ₀ (k _n ) represents the first-order modified Bessel function corresponding to the nth sound source, and k _n represents the concentration parameter corresponding to the single von Misius distribution that the intensity vector direction of the sound pressure signal of the nth sound source obeys, That is, the reciprocal of the variance of the von Misius distribution;

The parameter set of the mixed von Misius distribution function is as follows:

Γ={α _n , k _n }, i=1, 2, 3 (11)

Step 6-2, initialize the model parameters, and obtain the initial function parameter set;

In the embodiment of the present invention, the value of α is [1/3, 1/3, 1/3], and the value of k is [8, 6, 3];

Step 6-3. According to the obtained initial model parameters, an initial mixed von Misius distribution function is established, and the function formula is obtained as:

The maximum expectation algorithm is used to estimate the parameters of the mixed von Misius distribution model. The principle is that the probability of the observed value is the largest. By deriving the model function with respect to the parameters α and k equal to zero, the revaluation of the parameters α and k is calculated.

Take γ(ω,t) as substitute Take the logarithm to get the initial log-likelihood value -3.0249e+004. By calculating the proportion of each current single von Misius distribution to the mixed von Misius distribution, the revalued α parameter [0.2267, 0.2817, 0.4516] can be obtained, and at the same time according to The method of calculating the parameter k obtained by deriving the value of the revalued k is [5.1498, 4.0061, 3.1277]. At this time, the new logarithmic likelihood value can be obtained -2.9887e+004, and the difference between the old and new likelihood values is 362.3362. The threshold value is 0.1, so the new likelihood value is assigned to the old likelihood value, and then the two newly obtained re-evaluation parameters are used to repeat the previous steps until the old and new likelihood values are less than the threshold value, and the algorithm is considered to converge. In this example Finally, the α parameter [0.2689, 0.2811, 0.4500] is obtained, and the value of k is [4.3508, 3.3601, 2.8332]. At this time, the mixed von Misius distribution function satisfying the direction distribution of the intensity vector is obtained, as shown in Figure 9. Mixed von Misius distribution.

Step 6-4, obtain the intensity vector direction function of each sound pressure signal according to the estimated mixed von Misius distribution model parameters;

The formula of the intensity vector direction function of the sound pressure signal is as follows:

${\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>$

in, Represents the intensity vector direction function of the nth sound source;

Step 7, according to the obtained intensity vector direction function of each sound pressure signal and the sound pressure of the microphone array, obtain the signal of each sound source in the frequency domain, and use the inverse Fourier transform to convert each sound source signal in the frequency domain Convert to a sound source signal in the time domain;

The signal formula for each sound source in the frequency domain is as follows:

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 13</span>}<span\; class="notranslate">\; )</span>$

in, Represents the frequency domain signal of the nth sound source signal obtained after the mixed speech separation;

Will The time domain signal is obtained by inverse Fourier transform

Step 8, calculate the matching probability of each sound source signal and the specified sound source in the sample speech library, think that the sound source with the largest probability value is the target sound source, keep the sound source signal, and delete other non-target sound sources;

In the embodiment of the present invention, assuming that the first person is the target sound source, the logarithm values of the matching probabilities of the final separated three voices and the target sound model are [-2.0850-2.8807-3.5084]×10 ⁴ , where the largest matching sound After separating the sound for No. 1, the target sound source is found.

The matching probability formula of each sound source signal and the specified sound source in the sample speech library is as follows:

$\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; log</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 14</span>}<span\; class="notranslate">\; )</span>$

In the formula: voice after separation The extracted speech feature parameters, that is, the extracted speech Mel frequency cepstral coefficients as speech The characteristic parameters;

Indicates the matching probability of the nth sound source signal and the specified sound source in the sample speech library;

G _c represents the voice model parameter of the user designated person;

Indicates the probability that the voice after separation belongs to the voice of the user-designated person;

Step 9, amplifying the retained sound source signal, that is, completing the amplification of the specified sound source in the tested environment.

In the embodiment of the present invention, finally, according to the obtained mixed von Misius distribution model parameters, the direction function of each sound source is obtained, and the original sound is further separated, as shown in Fig. 10 (a) to (f), which is ideal Compared with the data obtained after separation, it can be seen that the similarity is extremely high.

Claims (3)

1. An intelligent speech processing method, comprising the steps of:

step 1, collecting sample voice sections to construct a sample voice library, performing feature extraction on the sample voice to obtain feature parameters, and training the feature parameters;

the specific process is as follows:

step 1-1, collecting sample voice sections, carrying out discretization processing on the collected voice sections, extracting Mel frequency cepstrum coefficients of voice signals as voice signal characteristic parameters, and establishing a Gaussian mixture model;

the model formula is as follows:

$p\left(X\right|G)={\mathrm{\&Sigma;}}_{i=1}^M{p}_i{b}_i\left(X\right)---\left(1\right)$

wherein p (X | G) represents the probability of the sample speech feature parameter in the model with the model parameter G;

g denotes a gaussian mixture model parameter set, G ═ p_i，μ_i，∑_i}，i＝1，2，...，I；

I represents the number of single Gaussian models in the Gaussian mixture model;

p_irepresenting the weighting coefficients of the ith single gaussian model,

μ_ia mean vector representing the ith single Gaussian model;

∑_ia covariance matrix representing the ith single gaussian model;

x represents a sample speech feature parameter, X ═ X₁，x₂，...，x_TT represents the number of feature vectors;

b_i(X) a density function representing the ith single Gaussian model, b_i(X)＝N(μ_i，∑_i) N (·) represents a density function of a standard gaussian distribution;

step 1-2, training a Gaussian mixture model by using the characteristic parameters of the voice signals;

i.e. adopt k are allClustering the characteristic parameters of the voice signals by using a value clustering algorithm to obtain an initial value G of a Gaussian mixture model parameter set₀＝{p_i ⁰，μ_i ⁰，∑_i ⁰1, 2, ·, I; estimating the model by adopting a maximum expectation algorithm according to the obtained initial value of the parameter set of the Gaussian mixture model, and further obtaining parameters of the Gaussian mixture model, namely finishing the training of the characteristic parameters;

step 2, collecting the audio signal of the detected environment by adopting a microphone array consisting of M microphones, and determining the number of the environmental sound sources and the arrival direction of each sound source beam, namely the incident angle from the sound source to the microphone array;

the specific process is as follows:

step 2-1, collecting mixed audio signals of the tested environment by using a microphone array consisting of M microphones, and carrying out discretization processing on the collected mixed audio signals to obtain the amplitude of each sampling point;

step 2-2, performing matrixing on the amplitude of each sampling point to obtain a mixed audio matrix collected by each microphone; the number of columns of the mixed audio matrix is one, the number of rows is the number of sampling points, and elements in the matrix are the amplitude of each sampling point;

step 2-3, obtaining an estimated value of a vector covariance matrix of the mixed audio signal of the tested environment according to the mixed audio matrix and the number of the microphones acquired by each microphone;

the estimated value of the vector covariance matrix is expressed as follows:

$R_{xx}=\frac{1}{M}{\mathrm{\&Sigma;}}_{m=1}^{M}X\left(m\right)X^H\left(m\right)---\left(2\right)$

wherein R is_xxAn estimate of a vector covariance matrix of the mixed audio signal representing the measured environment;

x (m) represents the mixed audio matrix collected by the mth microphone;

X^H(m) a transpose matrix representing the mixed audio matrix collected by the mth microphone;

step 2-4, performing eigenvalue decomposition on the estimated value of the vector covariance matrix to obtain eigenvalues, sequencing the eigenvalues from large to small, and determining the number of the eigenvalues larger than a threshold, namely the number of the sound sources;

2-5, subtracting the number of the sound sources from the number of the microphones to obtain the number of the noise sources, and further correspondingly obtaining a noise matrix;

step 2-6, obtaining a steering vector of the microphone array according to the distance between each microphone and the array center, the wavelength of the mixed audio signal, the direction angle of the microphone to the array center and the arrival direction of the sound source beam, and obtaining an angle spectrum function of the mixed audio signal according to the noise matrix and the steering vector of the microphone array;

the angle spectral function formula of the mixed audio signal is as follows:

$P\left(\mathrm{\&theta;}\right)=\frac{1}{{\mathrm{\&alpha;}}^H\left(\mathrm{\&theta;}\right)V_u{V^H}_u\mathrm{\&alpha;}\left(\mathrm{\&theta;}\right)}---\left(3\right)$

wherein P (θ) represents an angular spectral function of the mixed audio signal;

α (theta) represents the steering vector of the microphone array, α (theta) ═ α₁(θ)，...，α_m(θ)，...，α_M(theta)), wherein, among others,j denotes an imaginary unit, k is 2 pi/λ, λ denotes a wavelength of the mixed audio signal, d_mIndicating the distance of the mth microphone from the center of the array,represents the direction angle of the mth microphone to the center of the array;

θ represents a beam arrival direction of the sound source;

α^H(θ) a transposed matrix representing steering vectors of the microphone array;

V_urepresenting a noise matrix;

V^H _ua transposed matrix representing the noise matrix;

2-7, selecting a plurality of peak values of the waveform from big to small according to the waveform of the angle spectrum function of the mixed audio signal, wherein the number of the selected peak values is the number of the sound sources;

step 2-8, determining an angle value corresponding to the selected peak value, namely obtaining the arrival direction of the wave beam of each sound source;

step 3, obtaining microphone array sound pressure received by the microphone, sound pressure gradient in the horizontal direction of the microphone array and sound pressure gradient in the vertical direction of the microphone array according to the audio signal of each sound source and the conversion relation between the sound sources and the microphone;

the microphone array sound pressure signal formula is as follows:

$p_w\left(t\right)={\mathrm{\&Sigma;}}_{n=1}^{N}0.5{\mathrm{\&Sigma;}}_{m=1}^M{h}_{mn}\left(t\right)s_n\left(t\right)---\left(4\right)$

wherein p is_w(t) represents microphone array sound pressure at time t;

n represents the number of sound sources;

t represents time;

s_n(t) represents an audio signal of an nth sound source;

h_mn(t) denotes a conversion matrix between the nth sound source and the mth microphone, h_mn(t)＝p₀(t)α_m(θ_n(t))，p₀(t) represents the sound pressure at the center of the microphone array caused by the sound wave at time t, α_m(θ_n(t)) represents the steering vector of the mth microphone with respect to the nth sound source at time t, where θ_n(t) represents the beam arrival direction of the nth sound source at time t;

the sound pressure gradient formula in the horizontal direction of the microphone array is as follows:

wherein p is_x(t) representing a sound pressure gradient in a horizontal direction of the microphone array;

the sound pressure gradient in the vertical direction of the microphone array is formulated as follows:

wherein p is_y(t) representing a sound pressure gradient in a vertical direction of the microphone array;

step 4, converting the central sound pressure of the microphone array, the sound pressure gradient in the horizontal direction of the microphone array and the sound pressure gradient in the vertical direction of the microphone array from a time domain to a frequency domain by adopting Fourier transform;

step 5, obtaining an intensity vector formula of a sound pressure signal in a frequency domain according to sound pressure of the microphone array in the frequency domain, horizontal direction gradient of the microphone array and vertical direction sound pressure gradient of the microphone array, and further deducing an intensity vector direction;

the intensity vector formula of the sound pressure signal in the frequency domain is:

$I(\mathrm{\&omega;},t)=\frac{1}{{\mathrm{\&rho;}}_{0}c}\&lsqb;\mathrm{Re}\left\{{p_w}^{*}(\mathrm{\&omega;},t)p_x(\mathrm{\&omega;},t)\right\}{u}_x+\mathrm{Re}\left\{{p_w}^{*}(\mathrm{\&omega;},t)p_y(\mathrm{\&omega;},t)\right\}{u}_y\&rsqb;---\left(7\right)$

wherein I (ω, t) represents an intensity vector of the sound pressure signal in the frequency domain;

ρ₀representing the air density of the tested environment;

c represents the speed of sound;

re [ ] represents taking the real part of the complex number;

p_w ^*(ω, t) represents a conjugate matrix of microphone array sound pressures in the frequency domain;

p_x(ω, t) represents the microphone array horizontal direction sound pressure gradient in the frequency domain;

p_y(ω, t) represents the microphone array vertical direction sound pressure gradient in the frequency domain;

u_xrepresenting unit vectors in the direction of the abscissa axis;

u_ya unit vector representing the direction of the ordinate axis;

the intensity vector direction formula is as follows:

$\mathrm{\&gamma;}(\mathrm{\&omega;},t)={\tan }^{-1}\&lsqb;\frac{\mathrm{Re}\left\{{p_w}^{*}(\mathrm{\&omega;},t)p_y(\mathrm{\&omega;},t)\right\}}{\mathrm{Re}\left\{{p_w}^{*}(\mathrm{\&omega;},t)p_x(\mathrm{\&omega;},t)\right\}}\&rsqb;---\left(8\right)$

wherein γ (ω, t) represents the intensity vector direction of the sound pressure signal of the mixed sound received by the microphone array;

step 6, counting the intensity vector direction to obtain probability density distribution, fitting by adopting mixed Von Milius distribution to obtain model parameters of the voice intensity vector direction obeying the mixed Von Milius distribution, and further obtaining an intensity vector direction function of each sound pressure signal;

the specific process is as follows:

step 6-1, counting the intensity vector direction to obtain probability density distribution of the intensity vector direction, and fitting by adopting mixed Von Milius distribution to obtain a model parameter set of the mixed Von Milius distribution to which the intensity vector direction of the voice obeys;

the formula of the mixed von mises distribution model is as follows:

wherein,representing a mixed von mises distribution probability density;

representing a mixed sound direction angle;

α_na weight representing an intensity vector direction function of a sound pressure signal of an nth sound source;

wherein, I₀(k_n) First order modified Bessel function, k, representing the corresponding nth sound source_nThe concentration parameter corresponding to the single von mises distribution to which the intensity vector direction of the nth sound source sound pressure signal obeys is represented, namely the reciprocal of the variance of the von mises distribution;

the hybrid von mises distribution function parameter set is as follows:

＝{α_n，k_n}，i＝1，...，N(11)

6-2, initializing model parameters to obtain an initial function parameter set;

6-3, estimating parameters of the mixed von Michels distribution model by adopting a maximum expectation algorithm according to the obtained initial model parameters;

6-4, solving an intensity vector direction function of each sound pressure signal according to the estimated mixed Von Milius distribution model parameters;

the intensity vector direction function formula of the sound pressure signal is as follows:

wherein,representing the intensity vector direction function of the nth sound source;

step 7, obtaining a signal of each sound source in a frequency domain according to the obtained intensity vector direction function of each sound pressure signal and the sound pressure of the microphone array, and converting each sound source signal in the frequency domain into a sound source signal in a time domain by adopting Fourier inverse transformation;

the signal formula of each sound source in the frequency domain is as follows:

wherein,a frequency domain signal representing an nth sound source signal obtained after separation of the mixed speech;

will be provided withObtaining a time domain signal through Fourier inverse transformation

Step 8, calculating the matching probability of each sound source signal and the designated sound source in the sample sound library, selecting the sound source with the maximum probability value as a target sound source, reserving the sound source signal, and deleting other non-target sound sources;

the matching probability formula of each sound source signal and the designated sound source in the sample voice library is as follows:

$C\left({\tilde{X}}_n\right)=log\&lsqb;P\left({\tilde{X}}_n\right|G_c)\&rsqb;---\left(14\right)$

in the formula:representing speech by separationExtracted speech feature parameters, i.e. extracting speechAs speech, Mel frequency cepstrum coefficientsThe characteristic parameters of (1);

representing the matching probability of the nth sound source signal and the specified sound source in the sample voice library;

G_cacoustic model parameters representing a user-specified person;

representing the probability that the separated speech belongs to the voice of the user-specified person;

and 9, amplifying the reserved sound source signals, namely completing the amplification of the specified sound source in the tested environment.

2. The intelligent speech processing method of claim 1, wherein the speech processing method comprises a speech processing method, and a speech processing methodThe threshold value range in the steps 2-4 is 10^-2～10^-16。

3. The intelligent speech processing method according to claim 1, wherein α of step 6-1_nTaking a random number within 0-1 and satisfyingk_nTaking random numbers within 1-700.