A method, device and computer equipment for processing speech signals

Abstract

The embodiment of the present invention discloses a speech signal processing method, device and computer equipment, including: obtaining the original speech signal, reference signal and near-end speech signal of the current frame; calculating the noise power of the near-end speech signal; calculating the prior probability that the target speech signal of the current frame does not exist and the posterior signal-to-noise ratio of the target speech signal of the current frame according to the first cross-correlation coefficient, the second cross-correlation coefficient, the noise power and the posterior probability that the target speech signal of the previous frame exists; calculating the prior signal-to-noise ratio according to the posterior signal-to-noise ratio, and then calculating the posterior probability that the target speech signal of the current frame exists according to the prior probability that the target speech signal of the current frame does not exist; calculating the hybrid suppression factor according to the posterior probability; and finally performing residual echo suppression and noise suppression processing on the near-end speech signal of the current frame according to the hybrid suppression factor. The technical solution of the embodiment of the present invention can reduce the computational complexity of speech signal processing while ensuring the speech signal processing performance.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of speech processing technology, and in particular to a speech signal processing method, apparatus, computer equipment and storage medium. Background Art

[0002] In the speech signal processing process, speech enhancement processing is the front-end signal processing method to achieve smooth speech interaction. Linear echo cancellation, residual echo suppression and noise suppression are the three main parts of the current front-end signal processing, and they also occupy most of the speech signal processing computing resources. Linear echo cancellation uses adaptive filtering technology to suppress most of the echoes in the speech signal; residual echo suppression mainly eliminates the residual nonlinear echo components in the speech signal through specific nonlinear means; noise suppression uses nonlinear algorithms to eliminate environmental noise in the speech signal.

[0003] At present, the above three parts of speech processing operations are completed independently and serially in the speech signal processing process, which requires a large amount of calculation. Especially in some cases where computing resources are limited, the above speech signal processing algorithm cannot achieve the optimal speech processing effect. Summary of the invention

[0004] The embodiments of the present invention provide a speech signal processing method, apparatus, computer equipment and storage medium to reduce the computational complexity of speech signal processing while ensuring speech signal processing performance.

[0005] In a first aspect, an embodiment of the present invention provides a method for processing a speech signal, comprising:

[0006] Obtaining a current frame original speech signal, a current frame reference signal, and a current frame near-end speech signal;

[0007] Calculating the noise power of the near-end speech signal of the current frame;

[0008] Calculating the prior probability that the target speech signal of the current frame does not exist according to the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame;

[0009] Calculate the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame;

[0010] Calculating the a priori signal-to-noise ratio of the target speech signal of the current frame according to the a priori signal-to-noise ratio of the target speech signal of the current frame, and calculating the a priori probability that the target speech signal of the current frame exists according to the a priori signal-to-noise ratio of the target speech signal of the current frame and the a priori probability that the target speech signal of the current frame does not exist;

[0011] Calculating a mixed suppression factor according to the posterior probability of the existence of the target speech signal of the current frame;

[0012] A speech signal is calculated according to the mixed suppression factor and the near-end speech signal of the current frame after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression.

[0013] In a second aspect, an embodiment of the present invention further provides a speech signal processing device, including:

[0014] A signal acquisition module, used to acquire the original speech signal of the current frame, the reference signal of the current frame and the near-end speech signal of the current frame;

[0015] A noise power calculation module, used to calculate the noise power of the near-end speech signal of the current frame;

[0016] A priori probability calculation module, used to calculate the priori probability that the target speech signal of the current frame does not exist based on the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame;

[0017] A posterior signal-to-noise ratio calculation module is used to calculate the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame;

[0018] a posterior probability calculation module, used to calculate the priori signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculate the posterior probability of the existence of the target speech signal of the current frame according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist;

[0019] A mixed suppression factor calculation module, used to calculate the mixed suppression factor according to the posterior probability of the existence of the target speech signal in the current frame;

[0020] The speech signal processing module is used to calculate the speech signal obtained by performing residual echo suppression and noise suppression on the near-end speech signal of the current frame according to the mixed suppression factor and the near-end speech signal of the current frame.

[0021] In a third aspect, an embodiment of the present invention further provides a computer device, the computer device comprising: one or more processors;

[0022] A storage device for storing one or more programs;

[0023] When the one or more programs are executed by the one or more processors, the one or more processors implement the speech signal processing method provided by any embodiment of the present invention.

[0024] In a fourth aspect, an embodiment of the present invention further provides a computer storage medium on which a computer program is stored, and when the program is executed by a processor, the speech signal processing method provided by any embodiment of the present invention is implemented.

[0025] The embodiment of the present invention calculates the noise power of the near-end speech signal of the current frame, calculates the priori probability that the target speech signal of the current frame does not exist according to the first cross-correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second cross-correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power, and calculates the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability that the target speech signal of the previous frame exists, and then calculates the priori signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculates the posterior probability that the target speech signal of the current frame exists according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist, thereby calculating the hybrid suppression factor according to the posterior probability that the target speech signal of the current frame exists, and calculating the speech signal obtained by the near-end speech signal of the current frame after the residual echo suppression and noise suppression processing according to the hybrid suppression factor and the near-end speech signal of the current frame, thereby solving the problems of large calculation amount and unsatisfactory processing effect in the existing independent residual echo suppression and noise suppression processing of the speech signal, and achieving the reduction of the computational complexity of speech signal processing under the premise of ensuring speech signal processing performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of a speech signal processing method provided by Embodiment 1 of the present invention;

[0027] Figure 2 is a flow chart of a speech signal processing method provided by Embodiment 2 of the present invention;

[0028] Figure 3 is a flow chart of a speech signal processing method provided by Embodiment 2 of the present invention;

[0029] Figure 4 is a schematic diagram of a speech signal processing device provided by Embodiment 3 of the present invention;

[0030] Figure 5 A schematic diagram of the structure of a computer device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0031] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention.

[0032] It should also be noted that, for ease of description, only the part relevant to the present invention but not all content is shown in the accompanying drawings. It should be mentioned before discussing exemplary embodiments in more detail that some exemplary embodiments are described as processing or methods depicted as flow charts. Although the flow chart describes each operation (or step) as sequential processing, many operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of each operation can be rearranged. When its operation is completed, the processing can be terminated, but it can also have additional steps not included in the accompanying drawings. The processing can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0033] The terms "first" and "second" and the like in the description, claims and drawings of the embodiments of the present invention are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may include steps or units that are not listed.

[0034] Embodiment 1

[0035] Figure 1 is a flow chart of a speech signal processing method provided by the first embodiment of the present invention. This embodiment is applicable to the case where residual echo suppression and noise suppression are performed on speech signals at the same time. The method can be executed by a speech signal processing device, which can be implemented by software and / or hardware and can generally be integrated in a computer device. Figure 1 As shown, the method includes the following operations:

[0036] S110, obtaining a current frame original speech signal, a current frame reference signal, and a current frame near-end speech signal.

[0037] Among them, the original voice signal of the current frame can be a voice signal that needs to be processed by residual echo suppression and noise suppression. Exemplarily, the current frame voice command signal (that is, the current frame microphone signal) input by the user obtained by the vehicle-mounted terminal through the microphone device or the current frame voice command signal collected by other intelligent terminals can be used as the original voice signal of the current frame. The original voice signal of the current frame may include but is not limited to the target voice signal, the noise signal, the echo signal or the residual echo signal. Among them, the residual echo signal is the echo signal remaining after the echo cancellation of the original voice signal of the current frame. The target voice signal is the voice command signal issued by the user. The current frame reference signal can be the system audio signal of the current frame, such as the audio signal in wav format played by the terminal. Correspondingly, the echo signal included in the original voice signal of the current frame can be the audio signal played by the terminal collected by the voice collection device (such as a microphone). The near-end voice signal of the current frame can be the current frame voice signal obtained after the original voice signal of the current frame is processed by AEC (Adaptive Echo Cancellation). In an embodiment of the present invention, the so-called current frame is also the voice frame that needs to be processed at present.

[0038] Considering that both residual echo suppression and noise suppression use nonlinear algorithms to process speech signals, and the residual echo signal can be approximated as a special noise signal, therefore, in the embodiment of the present invention, a hybrid suppression factor can be calculated to perform residual echo suppression and noise suppression on the speech signal at the same time. When calculating the hybrid suppression factor, three types of speech signals are needed: the original speech signal of the current frame, the reference signal of the current frame, and the near-end speech signal of the current frame.

[0039] S120: Calculate the noise power of the near-end speech signal of the current frame.

[0040] After obtaining the near-end speech signal of the current frame, the noise power of the near-end speech signal of the current frame can be further calculated. Optionally, the noise power of the near-end speech signal of the current frame can be calculated according to a minimum value tracking method. The minimum value tracking method tracks the minimum power value at each frequency point within a time interval in real time. Optionally, the time interval can be 2 seconds or 3 seconds, etc. The embodiment of the present invention does not limit the specific time range of the time interval.

[0041] S130. Calculate the prior probability that the target speech signal of the current frame does not exist based on the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame.

[0042] The first cross-correlation coefficient may be a cross-correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, and the second cross-correlation coefficient may be a cross-correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame.

[0043] Correspondingly, after the noise power of the near-end speech signal of the current frame is calculated, the prior probability that the target speech signal of the current frame does not exist can be calculated based on the first cross-correlation coefficient, the second cross-correlation coefficient and the noise power of the near-end speech signal of the current frame.

[0044] S140, calculating the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame.

[0045] In the embodiment of the present invention, the posterior signal-to-noise ratio of the target speech signal of the current frame may also be calculated based on the posterior probability of the existence of the target speech signal of the previous frame.

[0046] S150, calculating the priori signal-to-noise ratio of the target speech signal of the current frame according to the posteriori signal-to-noise ratio of the target speech signal of the current frame, and calculating the posteriori probability that the target speech signal of the current frame exists according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist.

[0047] Correspondingly, after calculating the posterior signal-to-noise ratio of the target speech signal of the current frame, the prior signal-to-noise ratio of the target speech signal of the current frame can be calculated based on the posterior signal-to-noise ratio of the target speech signal of the current frame, and the posterior probability of the existence of the target speech signal of the current frame can be calculated based on the prior signal-to-noise ratio of the target speech signal of the current frame and the prior probability that the target speech signal of the current frame does not exist.

[0048] S160: Calculate a mixed suppression factor according to the posterior probability that the target speech signal of the current frame exists.

[0049] Correspondingly, after calculating the posterior probability that the target speech signal of the current frame exists, the hybrid suppression factor can be calculated according to the posterior probability that the target speech signal of the current frame exists.

[0050] S170. Calculate, according to the mixed suppression factor and the near-end speech signal of the current frame, a speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression.

[0051] In an embodiment of the present invention, after the hybrid suppression factor is calculated, the speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression can be calculated based on the hybrid suppression factor and the near-end speech signal of the current frame, so as to realize simultaneous residual echo suppression and noise suppression processing on the near-end speech signal of the current frame.

[0052] The embodiment of the present invention calculates the noise power of the near-end speech signal of the current frame, calculates the priori probability that the target speech signal of the current frame does not exist according to the first cross-correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second cross-correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power, and calculates the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability that the target speech signal of the previous frame exists, and then calculates the priori signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculates the posterior probability that the target speech signal of the current frame exists according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist, thereby calculating the hybrid suppression factor according to the posterior probability that the target speech signal of the current frame exists, and calculating the speech signal obtained by the near-end speech signal of the current frame after the residual echo suppression and noise suppression processing according to the hybrid suppression factor and the near-end speech signal of the current frame, thereby solving the problems of large calculation amount and unsatisfactory processing effect in the existing independent residual echo suppression and noise suppression processing of the speech signal, and achieving the reduction of the computational complexity of speech signal processing under the premise of ensuring speech signal processing performance.

[0053] Embodiment 2

[0054] Figure 2 is a flow chart of a speech signal processing method provided by Embodiment 2 of the present invention, Figure 3 is a flow chart of a speech signal processing method provided by the second embodiment of the present invention. This embodiment is specific based on the above embodiment. In this embodiment, the noise power, the cross-correlation coefficient, the prior probability that the target speech signal of the current frame does not exist, the posterior signal-to-noise ratio and the prior signal-to-noise ratio of the target speech signal of the current frame, the posterior probability that the target speech signal of the current frame exists, and the specific calculation method of the mixed suppression factor are given. Accordingly, as Figure 2 and Figure 3 As shown, the method of this embodiment may include:

[0055] S210, obtaining a current frame original speech signal, a current frame reference signal, and a current frame near-end speech signal.

[0056] S220: Calculate the noise power of the near-end speech signal of the current frame.

[0057] S230, calculating a first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, and a second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame.

[0058] Accordingly, S230 may specifically include:

[0059] S231. Calculate a first cross-correlation spectrum between the original speech signal of the current frame and the reference signal of the current frame.

[0060] In an optional embodiment of the present invention, calculating the first cross-correlation spectrum between the current frame original speech signal and the current frame reference signal may include:

[0061] The power spectrum of the original speech signal of the current frame and the reference signal of the current frame is calculated based on the following formula:

[0062]

[0063]

[0064] Among them, S d (i, j) represents the power spectrum of the jth frequency point of the original speech signal of the current frame, S d (i-1, j) represents the power spectrum of the jth frequency point of the original speech signal of the previous frame, β represents the smoothing coefficient, and β can be optionally set to 0.85. The embodiment of the present invention does not limit the specific value of β. i,j represents the spectrum of the jth frequency point of the original speech signal of the current frame, represents the conjugate complex number of the spectrum of the jth frequency point of the original speech signal of the current frame, S x (i, j) represents the power spectrum of the jth frequency point of the current frame reference signal; S x (i-1,j) represents the power spectrum of the jth frequency point of the previous frame reference signal, x i,j represents the spectrum of the jth frequency point of the current frame reference signal, Represents the conjugate complex number of the spectrum of the j-th frequency point of the current frame reference signal.

[0065] It should be noted that, in the embodiment of the present invention, the “current frame” refers to the i-th frame. d (i,j) can also represent the power spectrum of the jth frequency point of the original speech signal of the i-th frame, that is, the power spectrum of the jth frequency point of the original speech signal of the current frame is the power spectrum of the jth frequency point of the original speech signal of the i-th frame. Correspondingly, "previous frame" also means the i-1th frame. For example, S d (i-1,j) can also represent the power spectrum of the jth frequency point of the original speech signal of the i-1th frame, that is, the power spectrum of the jth frequency point of the original speech signal of the previous frame is the power spectrum of the jth frequency point of the original speech signal of the i-1th frame.

[0066] The first cross-correlation spectrum is calculated based on the following formula:

[0067]

[0068] Among them, S xd (i, j) the first cross-correlation spectrum between the jth frequency point of the original speech signal of the current frame and the jth frequency point of the reference signal of the current frame, S xd (i-1, j) represents the first cross-correlation spectrum between the j-th frequency point of the original speech signal of the previous frame and the j-th frequency point of the reference signal of the previous frame.

[0069] S232: Calculate a first cross-correlation coefficient according to the first cross-correlation spectrum, the spectrum of the original speech signal of the current frame, and the spectrum of the reference signal of the current frame.

[0070] In an optional embodiment of the present invention, calculating the first cross-correlation coefficient according to the first cross-correlation spectrum, the spectrum of the original speech signal of the current frame, and the spectrum of the reference signal of the current frame may include: calculating the first cross-correlation coefficient based on the following formula:

[0071]

[0072] Among them, C xd (i, j) represents the first correlation coefficient, represents the conjugate complex number of the first cross-correlation spectrum.

[0073] S233: Calculate a second cross-correlation spectrum between the original speech signal of the current frame and the near-end speech signal of the current frame.

[0074] In an optional embodiment of the present invention, calculating the second cross-correlation spectrum between the original speech signal of the current frame and the near-end speech signal of the current frame may include:

[0075] The power spectrum of the near-end speech signal of the current frame is calculated based on the following formula:

[0076]

[0077] Among them, S e (i, j) represents the power spectrum of the jth frequency point of the near-end speech signal in the current frame, S e (i-1, j) represents the power spectrum of the jth frequency point of the near-end speech signal of the previous frame, e i,j represents the spectrum of the jth frequency point of the near-end speech signal of the current frame, A conjugate complex number representing the frequency spectrum of the jth frequency point of the near-end speech signal of the current frame;

[0078] The second cross-correlation spectrum is calculated based on the following formula:

[0079]

[0080] Among them, Sde (i, j) represents the second cross-correlation spectrum of the jth frequency point of the original speech signal of the current frame and the jth frequency point of the near-end speech signal of the current frame, S de (i-1, j) represents the second cross-correlation spectrum between the j-th frequency point of the original speech signal of the previous frame and the j-th frequency point of the near-end speech signal of the previous frame.

[0081] S234: Calculate a second cross-correlation coefficient according to the second cross-correlation spectrum, the spectrum of the original speech signal of the current frame, and the spectrum of the near-end speech signal of the current frame.

[0082] In an optional embodiment of the present invention, calculating the second cross-correlation coefficient according to the second cross-correlation spectrum, the spectrum of the original speech signal of the current frame, and the spectrum of the near-end speech signal of the current frame may include: calculating the second cross-correlation coefficient based on the following formula:

[0083]

[0084] Among them, C de (i, j) represents the second correlation coefficient, represents the conjugate complex number of the second cross-correlation spectrum.

[0085] In the embodiment of the present invention, C de (i, j) can be used as the proportion of the target speech signal at the jth frequency point in the current frame, C xd (i, j) can be used as the proportion of the residual echo signal at the jth frequency point in the current frame.

[0086] S240. Calculate the prior probability that the target speech signal of the current frame does not exist based on the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame.

[0087] In an optional embodiment of the present invention, calculating the prior probability that the target speech signal of the current frame does not exist based on the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame may include: calculating the power ratio between the target speech signal of the current frame and the noise signal of the current frame based on the noise power of the near-end speech signal of the current frame, the first mutual correlation coefficient, and the second mutual correlation coefficient; calculating the prior probability that the target speech signal of the current frame does not exist based on the power ratio between the target speech signal of the current frame and the noise signal of the current frame.

[0088] In an optional embodiment of the present invention, calculating the power ratio between the current frame target speech signal and the current frame noise signal according to the noise power of the current frame near-end speech signal, the first cross-correlation coefficient, and the second cross-correlation coefficient may include: calculating the power ratio between the current frame target speech signal and the current frame noise signal based on the following formula:

[0089]

[0090] Among them, η1(i,j) represents the preliminary power ratio between the current frame target speech signal and the current frame noise signal. The so-called preliminary power ratio can be understood as a staged power ratio, that is, a more accurate and reasonable target power ratio can be calculated based on the preliminary power ratio. noise (i, j) represents the noise power of the near-end speech signal of the current frame, C de (i, j) represents the second correlation coefficient, C xd (i, j) represents the first correlation coefficient, e i,j Represents the spectrum of the jth frequency point of the near-end speech signal of the current frame.

[0091] In an optional embodiment of the present invention, calculating the power ratio between the current frame target speech signal and the current frame noise signal according to the noise power of the current frame near-end speech signal, the first cross-correlation coefficient, and the second cross-correlation coefficient may include: calculating the power ratio between the current frame target speech signal and the current frame noise signal based on the following formula:

[0092] η(i,j)=min((C de (i,j) / C xd (i,j)) 2 ,η1(i,j))

[0093] Among them, η(i,j) is the target power ratio between the target speech signal of the current frame and the noise signal of the current frame.

[0094] In an embodiment of the present application, considering that the residual echo signal does not always exist, the preliminary power ratio between the current frame target speech signal and the current frame noise signal can be further refined to obtain the target power ratio between the current frame target speech signal and the current frame noise signal.

[0095] In an optional embodiment of the present invention, calculating the prior probability that the target speech signal of the current frame does not exist according to the power ratio between the target speech signal of the current frame and the noise signal of the current frame may include: calculating the prior probability that the target speech signal of the current frame does not exist based on the following formula:

[0096]

[0097] Wherein, q(i,j) represents the prior probability that the near-end speech signal of the current frame does not exist, ν0 represents the threshold value, and optionally, ν0 can be 5. The embodiment of the present invention does not limit the specific value of ν0. η* represents the power ratio between the target speech signal of the current frame and the noise signal of the current frame, and η* is η1(i,j) or η(i,j).

[0098] It can be seen from the calculation formula of the power ratio above that the larger the power ratio is, the greater the probability that the target speech signal exists, and the smaller the power ratio is, the smaller the probability that the target speech signal exists. When the power ratio is less than 1, it can be considered that the target speech signal does not exist at this time. Therefore, the prior probability that the target speech signal does not exist in the current frame can be calculated according to the power ratio between the target speech signal of the current frame and the noise signal of the current frame according to this criterion.

[0099] S250, calculating the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame.

[0100] In an optional embodiment of the present invention, calculating the posterior signal-to-noise ratio of the target speech signal of the current frame based on the posterior probability of the existence of the target speech signal of the previous frame may include: calculating the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame based on the posterior probability of the existence of the near-end speech signal of the previous frame; calculating the posterior signal-to-noise ratio of the target speech signal of the current frame based on the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame.

[0101] The joint power spectrum may include the power spectra of the current frame residual echo signal and the current frame noise signal at the same time.

[0102] In an optional embodiment of the present invention, calculating the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame according to the posterior probability of the existence of the near-end speech signal of the previous frame may include: calculating the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame based on the following formula:

[0103]

[0104]

[0105] Wherein, λ(i,j) represents the joint power spectrum of the current frame residual echo signal and the current frame noise signal, represents the variable smoothing factor of the previous frame near-end speech signal, λ(i-1,j) represents the joint power spectrum of the previous frame residual echo signal and the previous frame noise signal, and p(i-1,j) represents the posterior probability of the existence of the previous frame near-end speech signal; αn represents a fixed smoothing coefficient; α n The value of can be set according to actual needs, and the embodiment of the present invention does not limit this.

[0106] Since the existence probability of the target speech signal affects the residual echo signal and the noise signal, it can be seen from the above formula that the variable smoothing factor used to calculate the joint power spectrum of the current frame residual echo signal and the current frame noise signal is directly related to the posterior probability of the existence of the near-end speech signal.

[0107] In an optional embodiment of the present invention, calculating the a posteriori signal-to-noise ratio of the target speech signal of the current frame according to the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame may include: calculating the a posteriori signal-to-noise ratio of the target speech signal of the current frame based on the following formula:

[0108]

[0109] Wherein, γ(i,j) represents the posterior signal-to-noise ratio of the target speech signal of the current frame.

[0110] S260, calculating the priori signal-to-noise ratio of the target speech signal of the current frame according to the posteriori signal-to-noise ratio of the target speech signal of the current frame, and calculating the posteriori probability that the target speech signal of the current frame exists according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist.

[0111] The a priori signal-to-noise ratio of the target speech signal of the current frame can be understood as the ratio between the power spectrum of the target speech signal and the joint power spectrum of the residual echo signal and the noise signal.

[0112] In an optional embodiment of the present invention, calculating the a priori signal-to-noise ratio of the target speech signal of the current frame according to the a posteriori signal-to-noise ratio of the target speech signal of the current frame may include: calculating the a priori signal-to-noise ratio of the target speech signal of the current frame based on the following formula:

[0113] ξ(i,j)=αG1 2 (i-1,j)γ(i-1,j)+(1-α)max{γ(i,j)-1,0}

[0114]

[0115] In an optional embodiment of the present invention, calculating the a posteriori probability of the existence of the target speech signal of the current frame according to the a priori signal-to-noise ratio of the target speech signal of the current frame and the a priori probability that the target speech signal of the current frame does not exist may include: calculating the a posteriori probability of the existence of the target speech signal of the current frame based on the following formula:

[0116]

[0117] Wherein, ξ(i,j) represents the a priori signal-to-noise ratio of the target speech signal of the current frame, α represents the smoothing coefficient, and optionally, α can be taken as 0.9. The embodiment of the present invention does not limit the specific value of α. G1(i-1,j) represents the intermediate value of the mixed suppression factor of the near-end speech signal of the previous frame, γ(i-1,j) represents the a posteriori signal-to-noise ratio of the target speech signal of the previous frame, ξ(i-1,j) represents the a priori signal-to-noise ratio of the target speech signal of the previous frame, and p(i,j) represents the a posteriori probability of the existence of the target speech signal of the current frame.

[0118] S270. Calculate a mixed suppression factor according to the posterior probability that the target speech signal of the current frame exists.

[0119] In an optional embodiment of the present invention, calculating the mixed suppression factor according to the posterior probability of the target speech signal in the current frame may include: calculating the mixed suppression factor based on the following formula:

[0120] G(i,j)=(G1(i,j)) p(i,j) G min (i,j) (1-p(i,j))

[0121] Wherein, G(i,j) represents the mixed suppression factor, G1(i,j) represents the middle value of the mixed suppression factor of the near-end speech signal of the current frame, G min (i, j) represents the threshold control value of the mixed suppression factor.

[0122] S280. Calculate, according to the mixed suppression factor and the near-end speech signal of the current frame, a speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression.

[0123] In an optional embodiment of the present invention, calculating the speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression according to the mixed suppression factor and the near-end speech signal of the current frame may include: calculating the speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression based on the following formula:

[0124] M(i,j)=e i,j G(i,j)

[0125] Wherein, M(i, j) represents the speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression.

[0126] It can be seen that the speech signal processing method provided in the embodiment of the present invention obtains the existence of the residual echo signal preliminarily according to the mutual correlation coefficient between the microphone signal, the reference signal and the speech signal after linear echo elimination, and then estimates the existence probability of the target speech signal, as well as the priori signal-to-noise ratio and the posterior signal-to-noise ratio in combination with the noise power, and finally obtains the mixed suppression parameters of the residual echo signal and the noise signal in combination with the above data, so as to suppress the residual echo signal and the noise signal in the speech signal at the same time through the mixed suppression parameters, which can significantly reduce the computational complexity while ensuring the speech enhancement performance, and is more conducive to engineering applications.

[0127] The embodiment of the present invention calculates the noise power of the near-end speech signal of the current frame, calculates the priori probability that the target speech signal of the current frame does not exist according to the first cross-correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second cross-correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power, and calculates the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability that the target speech signal of the previous frame exists, and then calculates the priori signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculates the posterior probability that the target speech signal of the current frame exists according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist, thereby calculating the hybrid suppression factor according to the posterior probability that the target speech signal of the current frame exists, and calculating the speech signal obtained by the near-end speech signal of the current frame after the residual echo suppression and noise suppression processing according to the hybrid suppression factor and the near-end speech signal of the current frame, thereby solving the problems of large calculation amount and unsatisfactory processing effect in the existing independent residual echo suppression and noise suppression processing of the speech signal, and achieving the reduction of the computational complexity of speech signal processing under the premise of ensuring speech signal processing performance.

[0128] It should be noted that any arrangement and combination of the technical features in the above embodiments also falls within the protection scope of the present invention.

[0129] Embodiment 3

[0130] Figure 4 is a schematic diagram of a speech signal processing device provided by Embodiment 3 of the present invention, such as Figure 4 As shown, the device includes: a signal acquisition module 310, a noise power calculation module 320, a priori probability calculation module 330, a posteriori signal-to-noise ratio calculation module 340, a posteriori probability calculation module 350, a mixed suppression factor calculation module 360 ​​and a speech signal processing module 370, wherein:

[0131] The signal acquisition module 310 is used to acquire the original speech signal of the current frame, the reference signal of the current frame and the near-end speech signal of the current frame;

[0132] A noise power calculation module 320, configured to calculate the noise power of the near-end speech signal of the current frame;

[0133] A priori probability calculation module 330 is used to calculate the priori probability that the target speech signal of the current frame does not exist based on the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame;

[0134] A posterior signal-to-noise ratio calculation module 340 is used to calculate the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame;

[0135] The posterior probability calculation module 350 is used to calculate the priori signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculate the posterior probability that the target speech signal of the current frame exists according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist;

[0136] A mixed suppression factor calculation module 360, used to calculate the mixed suppression factor according to the posterior probability of the existence of the target speech signal in the current frame;

[0137] The speech signal processing module 370 is used to calculate the speech signal obtained by performing residual echo suppression and noise suppression processing on the near-end speech signal of the current frame according to the mixed suppression factor and the near-end speech signal of the current frame.

[0138] The embodiment of the present invention calculates the noise power of the near-end speech signal of the current frame, calculates the priori probability that the target speech signal of the current frame does not exist according to the first cross-correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second cross-correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power, and calculates the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability that the target speech signal of the previous frame exists, and then calculates the priori signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculates the posterior probability that the target speech signal of the current frame exists according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist, thereby calculating the hybrid suppression factor according to the posterior probability that the target speech signal of the current frame exists, and calculating the speech signal obtained by the near-end speech signal of the current frame after the residual echo suppression and noise suppression processing according to the hybrid suppression factor and the near-end speech signal of the current frame, thereby solving the problems of large calculation amount and unsatisfactory processing effect in the existing independent residual echo suppression and noise suppression processing of the speech signal, and achieving the reduction of the computational complexity of speech signal processing under the premise of ensuring speech signal processing performance.

[0139] Optionally, the prior probability calculation module 330 is specifically used to: calculate the power ratio between the current frame target speech signal and the current frame noise signal based on the noise power of the current frame near-end speech signal, the first mutual correlation coefficient, and the second mutual correlation coefficient; calculate the prior probability that the current frame target speech signal does not exist based on the power ratio between the current frame target speech signal and the current frame noise signal.

[0140] Optionally, the prior probability calculation module 330 is specifically used to calculate the power ratio between the current frame target speech signal and the current frame noise signal based on the following formula:

[0141]

[0142] Wherein, η1(i,j) represents the preliminary power ratio between the target speech signal of the current frame and the noise signal of the current frame, λ noise (i, j) represents the noise power of the near-end speech signal of the current frame, C de (i, j) represents the second correlation coefficient, C xd (i, j) represents the first correlation coefficient, e i,j Represents the spectrum of the jth frequency point of the near-end speech signal of the current frame.

[0143] Optionally, the prior probability calculation module 330 is specifically used to calculate the power ratio between the current frame target speech signal and the current frame noise signal based on the following formula:

[0144] η(i,j)=min((C de (i,j) / C xd (i,j)) 2 ,η1(i,j))

[0145] Among them, η(i,j) is the target power ratio between the target speech signal of the current frame and the noise signal of the current frame.

[0146] Optionally, the prior probability calculation module 330 is specifically used to calculate the prior probability that the target speech signal of the current frame does not exist based on the following formula:

[0147]

[0148] Among them, q(i,j) represents the prior probability that the near-end speech signal of the current frame does not exist, ν0 represents the threshold value, η* represents the power ratio between the target speech signal of the current frame and the noise signal of the current frame, and η* takes η1(i,j) or η(i,j).

[0149] Optionally, the a posteriori signal-to-noise ratio calculation module 340 is specifically used to: calculate the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame based on the a posteriori probability of the existence of the near-end speech signal of the previous frame; calculate the a posteriori signal-to-noise ratio of the target speech signal of the current frame based on the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame.

[0150] Optionally, the a posteriori signal-to-noise ratio calculation module 340 is specifically used to calculate the joint power spectrum of the current frame residual echo signal and the current frame noise signal based on the following formula:

[0151]

[0152]

[0153] Wherein, λ(i,j) represents the joint power spectrum of the current frame residual echo signal and the current frame noise signal, represents the variable smoothing factor of the previous frame near-end speech signal, λ(i-1,j) represents the joint power spectrum of the previous frame residual echo signal and the previous frame noise signal, and p(i-1,j) represents the posterior probability of the existence of the previous frame near-end speech signal; α n represents a fixed smoothing coefficient;

[0154] Calculating a posterior signal-to-noise ratio of the target speech signal of the current frame according to a joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame, including:

[0155] The posterior signal-to-noise ratio of the target speech signal of the current frame is calculated based on the following formula:

[0156]

[0157] Wherein, γ(i,j) represents the posterior signal-to-noise ratio of the target speech signal of the current frame.

[0158] Optionally, the posterior probability calculation module 350 is specifically used to calculate the prior signal-to-noise ratio of the target speech signal of the current frame based on the following formula:

[0159] ξ(i,j)=αG1 2 (i-1,j)γ(i-1,j)+(1-α)max{γ(i,j)-1,0}

[0160]

[0161] Calculating the posterior probability of the existence of the target speech signal of the current frame according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist includes:

[0162] The posterior probability of the target speech signal in the current frame existing is calculated based on the following formula:

[0163]

[0164] Among them, ξ(i,j) represents the prior signal-to-noise ratio of the target speech signal of the current frame, α represents the smoothing coefficient, G1(i-1,j) represents the intermediate value of the mixed suppression factor of the near-end speech signal of the previous frame, γ(i-1,j) represents the posterior signal-to-noise ratio of the target speech signal of the previous frame, ξ(i-1,j) represents the prior signal-to-noise ratio of the target speech signal of the previous frame, and p(i,j) represents the posterior probability of the existence of the target speech signal of the current frame.

[0165] Optionally, the hybrid suppression factor calculation module 360 ​​is specifically configured to calculate the hybrid suppression factor based on the following formula:

[0166] G(i,j)=(G1(i,j)) p(i,j) G min (i,j) (1-p(i,j))

[0167] Wherein, G(i,j) represents the mixed suppression factor, G1(i,j) represents the middle value of the mixed suppression factor of the near-end speech signal of the current frame, G min (i, j) represents the threshold control value of the mixed suppression factor.

[0168] Optionally, the speech signal processing module 370 is specifically used to calculate the speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression based on the following formula:

[0169] M(i,j)=e i,j G(i,j)

[0170] Wherein, M(i, j) represents the speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression.

[0171] The above-mentioned speech signal processing device can execute the speech signal processing method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not described in detail in this embodiment, please refer to the speech signal processing method provided by any embodiment of the present invention.

[0172] Since the above-mentioned speech signal processing device is a device that can execute the speech signal processing method in the embodiment of the present invention, based on the speech signal processing method introduced in the embodiment of the present invention, the technicians in the field can understand the specific implementation of the speech signal processing device in the present embodiment and its various variations, so how the speech signal processing device implements the speech signal processing method in the embodiment of the present invention is not described in detail here. As long as the technicians in the field implement the device used by the speech signal processing method in the embodiment of the present invention, it belongs to the scope of protection of this application.

[0173] Embodiment 4

[0174] Figure 5 A schematic diagram of the structure of a computer device provided in Embodiment 4 of the present invention. Figure 5 A block diagram of a computer device 412 suitable for use in implementing embodiments of the present invention is shown. Figure 5 The computer device 412 shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention. The computer device 412 can typically be a terminal device that performs voice processing functions.

[0175] like Figure 5 As shown, the computer device 412 is in the form of a general-purpose computing device. The components of the computer device 412 may include, but are not limited to: one or more processors 416, a storage device 428, and a bus 418 connecting various system components (including the storage device 428 and the processor 416).

[0176] Bus 418 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

[0177] The computer device 412 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 412, including volatile and non-volatile media, removable and non-removable media.

[0178] The storage device 428 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 430 and / or cache memory 432. The computer device 412 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 434 may be used to read and write non-removable, non-volatile magnetic media ( Figure 5 not shown, usually called a "hard drive"). Although Figure 5 Not shown in the figure, a disk drive for reading and writing a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing a removable non-volatile optical disk (e.g., a read-only optical disk (Compact Disc-Read Only Memory, CD-ROM), a digital video disk (Digital Video Disc-Read Only Memory, DVD-ROM) or other optical media) may be provided. In these cases, each drive may be connected to the bus 418 via one or more data medium interfaces. The storage device 428 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0179] A program 436 having a set (at least one) of program modules 426 may be stored, for example, in a storage device 428, such program modules 426 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination thereof may include an implementation of a network environment. Program modules 426 generally perform the functions and / or methods of the embodiments described herein.

[0180] The computer device 412 may also communicate with one or more external devices 414 (e.g., keyboard, pointing device, camera, display 424, etc.), may also communicate with one or more devices that enable a user to interact with the computer device 412, and / or communicate with any device that enables the computer device 412 to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication may be performed through an input / output (I / O) interface 422. In addition, the computer device 412 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN) and / or a public network, such as the Internet) through a network adapter 420. As shown, the network adapter 420 communicates with other modules of the computer device 412 through a bus 418. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in conjunction with the computer device 412, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, disk arrays (Redundant Arrays of Independent Disks, RAID) systems, tape drives, and data backup storage systems.

[0181] The processor 416 executes various functional applications and data processing by running the programs stored in the storage device 428, such as implementing the speech signal processing method provided in the above embodiment of the present invention.

[0182] That is, when the processing unit executes the program, it realizes: obtaining the original speech signal of the current frame, the reference signal of the current frame and the near-end speech signal of the current frame; calculating the noise power of the near-end speech signal of the current frame; calculating the prior probability that the target speech signal of the current frame does not exist according to the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame and the noise power of the near-end speech signal of the current frame; calculating the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability that the target speech signal of the previous frame exists; calculating the prior signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculating the posterior probability that the target speech signal of the current frame exists according to the prior signal-to-noise ratio of the target speech signal of the current frame and the prior probability that the target speech signal of the current frame does not exist; calculating the hybrid suppression factor according to the posterior probability that the target speech signal of the current frame exists; and calculating the speech signal obtained by the near-end speech signal of the current frame after the residual echo suppression and noise suppression processing according to the hybrid suppression factor and the near-end speech signal of the current frame.

[0183] Embodiment 5

[0184] Embodiment 6 of the present invention further provides a computer storage medium storing a computer program, and when the computer program is executed by a computer processor, it is used to execute the speech signal processing method described in any of the above embodiments of the present invention: obtaining the original speech signal of the current frame, the reference signal of the current frame, and the near-end speech signal of the current frame; calculating the noise power of the near-end speech signal of the current frame; calculating the priori that the target speech signal of the current frame does not exist based on the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame. probability; calculating the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame; calculating the prior signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculating the posterior probability of the existence of the target speech signal of the current frame according to the prior signal-to-noise ratio of the target speech signal of the current frame and the prior probability that the target speech signal of the current frame does not exist; calculating the hybrid suppression factor according to the posterior probability of the existence of the target speech signal of the current frame; calculating the speech signal obtained by the near-end speech signal of the current frame after the residual echo suppression and noise suppression processing according to the hybrid suppression factor and the near-end speech signal of the current frame.

[0185] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0186] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0187] The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, radio frequency (RF), etc., or any suitable combination of the foregoing.

[0188] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0189] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A speech signal processing method, characterized in that: include: Obtaining a current frame original speech signal, a current frame reference signal, and a current frame near-end speech signal; Calculating the noise power of the near-end speech signal of the current frame; Calculating the prior probability that the target speech signal of the current frame does not exist according to the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame; Calculate the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame; Calculating the a priori signal-to-noise ratio of the target speech signal of the current frame according to the a priori signal-to-noise ratio of the target speech signal of the current frame, and calculating the a priori probability that the target speech signal of the current frame exists according to the a priori signal-to-noise ratio of the target speech signal of the current frame and the a priori probability that the target speech signal of the current frame does not exist; Calculating a mixed suppression factor according to the posterior probability of the existence of the target speech signal of the current frame; A speech signal is calculated according to the mixed suppression factor and the near-end speech signal of the current frame after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression.

2. The method according to claim 1, characterized in that Calculating the prior probability that the target speech signal of the current frame does not exist according to the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame, comprising: Calculate the power ratio between the current frame target speech signal and the current frame noise signal according to the noise power of the current frame near-end speech signal, the first cross-correlation coefficient, and the second cross-correlation coefficient; The prior probability that the target speech signal of the current frame does not exist is calculated according to the power ratio between the target speech signal of the current frame and the noise signal of the current frame.

3. The method according to claim 2, characterized in that Calculating a power ratio between a target speech signal of a current frame and a noise signal of a current frame according to the noise power of the near-end speech signal of the current frame, the first cross-correlation coefficient, and the second cross-correlation coefficient, comprising: The power ratio between the current frame target speech signal and the current frame noise signal is calculated based on the following formula: Wherein, η1(i,j) represents the preliminary power ratio between the target speech signal of the current frame and the noise signal of the current frame, λ noise (i, j) represents the noise power of the near-end speech signal of the current frame, C de (i, j) represents the second correlation coefficient, C xd (i, j) represents the first correlation coefficient, e i,j Represents the spectrum of the jth frequency point of the near-end speech signal of the current frame.

4. The method according to claim 3, characterized in that Calculating a power ratio between a target speech signal of a current frame and a noise signal of a current frame according to the noise power of the near-end speech signal of the current frame, the first cross-correlation coefficient, and the second cross-correlation coefficient, comprising: The power ratio between the current frame target speech signal and the current frame noise signal is calculated based on the following formula: η(i,j)=min((C de (i,j) / C xd (i,j)) 2 ,η1(i,j)) Among them, η(i,j) is the target power ratio between the target speech signal of the current frame and the noise signal of the current frame.

5. The method according to claim 3 or 4, characterized in that: Calculating the prior probability that the target speech signal of the current frame does not exist according to the power ratio between the target speech signal of the current frame and the noise signal of the current frame includes: The prior probability that the target speech signal of the current frame does not exist is calculated based on the following formula: Among them, q(i,j) represents the prior probability that the near-end speech signal of the current frame does not exist, ν0 represents the threshold value, η* represents the power ratio between the target speech signal of the current frame and the noise signal of the current frame, and η* takes η1(i,j) or η(i,j).

6. The method according to claim 1, characterized in that The posterior signal-to-noise ratio of the target speech signal of the current frame is calculated according to the posterior probability of the existence of the target speech signal of the previous frame, including: Calculate the joint power spectrum of the current frame residual echo signal and the current frame noise signal according to the posterior probability of the existence of the previous frame near-end speech signal; The posterior signal-to-noise ratio of the target speech signal of the current frame is calculated according to the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame.

7. The method according to claim 6, characterized in that Calculating the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame according to the posterior probability of the existence of the near-end speech signal of the previous frame, including: The joint power spectrum of the current frame residual echo signal and the current frame noise signal is calculated based on the following formula: Wherein, λ(i,j) represents the joint power spectrum of the current frame residual echo signal and the current frame noise signal, represents the variable smoothing factor of the previous frame near-end speech signal, λ(i-1,j) represents the joint power spectrum of the previous frame residual echo signal and the previous frame noise signal, and p(i-1,j) represents the posterior probability of the existence of the previous frame near-end speech signal; α n represents a fixed smoothing coefficient; Calculating a posterior signal-to-noise ratio of the target speech signal of the current frame according to a joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame, including: The posterior signal-to-noise ratio of the target speech signal of the current frame is calculated based on the following formula: Wherein, γ(i,j) represents the posterior signal-to-noise ratio of the target speech signal of the current frame.

8. The method according to claim 1, characterized in that Calculating the a priori signal-to-noise ratio of the target speech signal of the current frame according to the a posteriori signal-to-noise ratio of the target speech signal of the current frame includes: The priori signal-to-noise ratio of the target speech signal of the current frame is calculated based on the following formula: ξ(i,j)=αG1 2 (i-1,j)γ(i-1,j)+(1-α)max{γ(i,j)-1,0} Calculating the posterior probability of the existence of the target speech signal of the current frame according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist includes: The posterior probability of the target speech signal in the current frame existing is calculated based on the following formula: Among them, ξ(i,j) represents the prior signal-to-noise ratio of the target speech signal of the current frame, α represents the smoothing coefficient, G1(i-1,j) represents the intermediate value of the mixed suppression factor of the near-end speech signal of the previous frame, γ(i-1,j) represents the posterior signal-to-noise ratio of the target speech signal of the previous frame, ξ(i-1,j) represents the prior signal-to-noise ratio of the target speech signal of the previous frame, and p(i,j) represents the posterior probability of the existence of the target speech signal of the current frame.

9. The method according to claim 1, characterized in that: Calculating a mixed suppression factor according to the posterior probability of the existence of the target speech signal of the current frame includes: The mixed inhibition factor is calculated based on the following formula: Wherein, G(i,j) represents the mixed suppression factor, G1(i,j) represents the middle value of the mixed suppression factor of the near-end speech signal of the current frame, G min (i, j) represents the threshold control value of the mixed suppression factor.

10. The method according to claim 1, characterized in that Calculating, according to the mixed suppression factor and the current frame near-end speech signal, a speech signal obtained after the current frame near-end speech signal is subjected to residual echo suppression and noise suppression processing, comprising: The voice signal obtained after the near-end voice signal of the current frame is processed by residual echo suppression and noise suppression is calculated based on the following formula: M(i,j)=e i,j G(i,j) Wherein, M(i, j) represents the speech signal obtained after the near-end speech signal of the current frame is processed by residual echo suppression and noise suppression.

11. A speech signal processing device, characterized in that: include: A signal acquisition module, used to acquire the original speech signal of the current frame, the reference signal of the current frame and the near-end speech signal of the current frame; A noise power calculation module, used to calculate the noise power of the near-end speech signal of the current frame; A priori probability calculation module, used to calculate the priori probability that the target speech signal of the current frame does not exist based on the first mutual correlation coefficient between the original speech signal of the current frame and the reference signal of the current frame, the second mutual correlation coefficient between the original speech signal of the current frame and the near-end speech signal of the current frame, and the noise power of the near-end speech signal of the current frame; A posterior signal-to-noise ratio calculation module is used to calculate the posterior signal-to-noise ratio of the target speech signal of the current frame according to the posterior probability of the existence of the target speech signal of the previous frame; a posterior probability calculation module, used to calculate the priori signal-to-noise ratio of the target speech signal of the current frame according to the posterior signal-to-noise ratio of the target speech signal of the current frame, and calculate the posterior probability of the existence of the target speech signal of the current frame according to the priori signal-to-noise ratio of the target speech signal of the current frame and the priori probability that the target speech signal of the current frame does not exist; A mixed suppression factor calculation module, used to calculate the mixed suppression factor according to the posterior probability of the existence of the target speech signal in the current frame; The speech signal processing module is used to calculate the speech signal obtained by performing residual echo suppression and noise suppression on the near-end speech signal of the current frame according to the mixed suppression factor and the near-end speech signal of the current frame.

12. The device according to claim 11, characterized in that The prior probability calculation module is specifically used for: Calculate the power ratio between the current frame target speech signal and the current frame noise signal according to the noise power of the current frame near-end speech signal, the first cross-correlation coefficient, and the second cross-correlation coefficient; The prior probability that the target speech signal of the current frame does not exist is calculated according to the power ratio between the target speech signal of the current frame and the noise signal of the current frame.

13. The device according to claim 12, characterized in that The prior probability calculation module is specifically used for: The power ratio between the current frame target speech signal and the current frame noise signal is calculated based on the following formula: Wherein, η1(i,j) represents the preliminary power ratio between the target speech signal of the current frame and the noise signal of the current frame, λ noise (i, j) represents the noise power of the near-end speech signal of the current frame, C de (i, j) represents the second correlation coefficient, C xd (i, j) represents the first correlation coefficient, e i,j Represents the spectrum of the jth frequency point of the near-end speech signal of the current frame.

14. The device according to claim 13, characterized in that The prior probability calculation module is specifically used for: The power ratio between the current frame target speech signal and the current frame noise signal is calculated based on the following formula: η(i,j)=min((C de (i,j) / C xd (i,j)) 2 ,η1(i,j)) Among them, η(i,j) is the target power ratio between the target speech signal of the current frame and the noise signal of the current frame.

15. The device according to claim 11, characterized in that The a posteriori signal-to-noise ratio calculation module is specifically used for: Calculate the joint power spectrum of the current frame residual echo signal and the current frame noise signal according to the posterior probability of the existence of the previous frame near-end speech signal; The posterior signal-to-noise ratio of the target speech signal of the current frame is calculated according to the joint power spectrum of the residual echo signal of the current frame and the noise signal of the current frame.

16. A computer device, characterized in that: The computer device comprises: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the speech signal processing method as described in any one of claims 1-10.

Images