KR20190129805A - Hearing Aid Having Noise Environment Classification and Reduction Function and Method thereof - Google Patents

Abstract

The present invention relates to a hearing aid having a noise environment classification and cancellation function and a method thereof. According to the present invention, the hearing aid having a noise environment classification and cancellation function comprises: a noise classification estimation unit for determining the type of a noise environment corresponding to a characteristic value of a signal for a divided noise section of an input audio signal and updating an estimator parameter matching the type of the noise environment; an input signal process unit for generating fast Fourier transform (FFT) data of the input audio signal; and a noise power estimator for estimating power of noise from the FFT data and estimating power of noise in the noise environment based on the estimator parameter.

Description

Hearing aid with noise environment classification and removal function and method thereof

The present invention relates to a hearing aid, and more particularly, to a hearing aid and a method for maximizing the performance of a noise canceling algorithm for hearing aids in various noise environments.

을 사용하여 잡음을 제거한다.Hearing aid technology aims to remove the noise signal by noise environment by estimating the noise signal for the changing environment. Specifically, the digital signal is divided into a voice signal and a noise in the signal processor to update the noise signal, and based on the updated noise signal.

Use to remove the noise.

는 잡음신호,

는 사용자에게 제공할 최종 디지털신호이며, δ는 입력된 디지털 신호의 주파수 대역과 유사도를 보이는 잡음환경을 매칭해주는 매칭부의 결과에 따라 결정되는 값이다. Where X (f, t) is the total input signal, δ is the weight value for each noise environment, α is the weight value according to signal-to-noise ratio (SNR) for each frequency,

Is the noise signal,

Is the final digital signal to be provided to the user, and δ is a value determined according to the result of the matching unit that matches the noise environment showing similarity with the frequency band of the input digital signal.

However, since the conventional hearing aid technology does not remove noise quickly and effectively by adapting to various noise environments, it is necessary to improve noise removal performance for various noise environments.

As related prior documents, International Patent Publication No. WO2010-015371, Korean Patent Registration No. 10-11517460000, Korean Patent Publication No. 10-2013-0118513 and the like can be referred to.

In order to maximize the performance of the noise canceling algorithm for hearing aids in various noise environments, accurate noise power estimation is very important. The accuracy of this noise power estimation algorithm is determined by internal parameters, which can have different optimization values depending on the type of noise environment. Therefore, the present invention provides a hearing aid and a method for quickly and effectively removing noise by adapting to various noise environments in real time by accurately determining the noise power estimator parameters by classifying the types of noise to be removed to improve the performance of the noise cancellation algorithm. To provide.

First, to summarize the features of the present invention, a hearing aid according to an aspect of the present invention for achieving the above object, determines the type of noise environment corresponding to the characteristic value of the signal for the divided noise interval of the input audio signal A noise classification estimator for updating an estimator parameter matching the type of the noise environment; An input signal processor configured to generate FFT data of the input audio signal; And a noise power estimator estimating power of noise from the FFT data, and estimating power of noise in a corresponding noise environment based on the estimator parameter.

The noise classification estimator may extract the characteristic value of only the signal for the noise section extracted using the voice detector, not the overall characteristic value of the input audio signal.

The noise classification estimating unit determines a noise environment for a signal characteristic value for the corresponding noise section by comparing a characteristic value for each noise environment obtained in advance using a machine learning method with a characteristic value of the signal for the noise section. Can be.

The noise classification estimator may output, to the noise power estimator, the estimator parameter corresponding to the determined noise environment among the noise environment estimator parameters previously obtained by preliminary evaluation.

The noise classification estimator may include: a voice detector configured to generate a separator of each section for distinguishing between a voice signal section and a noise section of the input audio signal; A feature extractor for extracting a feature value of a signal for a noise section of the input audio signal based on the delimiter; A noise signal characteristic buffer for sequentially storing and outputting the output of the characteristic extractor; And a noise environment classifier for determining a type of the noise environment corresponding to the characteristic value for a predetermined time stored in the noise signal characteristic buffer, and updating and outputting a corresponding estimator parameter.

The input signal processor may include: an input buffer configured to sequentially store and output the input audio signal; A window application unit for selecting valid data from an output of the input buffer; And an FFT unit generating data obtained by FFT processing the output of the window application unit.

The hearing aid includes: a gain calculator which calculates a gain for the estimated power of the noise; A multiplier for multiplying the gain by the FFT processed data; An IFFT unit for generating data obtained by IFFT processing the output of the multiplier; And an output buffer for sequentially storing and outputting output data of the IFFT unit.

According to another aspect of the present invention, a hearing aid signal processing method includes determining a type of noise environment corresponding to a characteristic value of a signal for a divided noise section of an input audio signal, and an estimator parameter matching the type of the noise environment. Updating; Generating FFT data of the input audio signal; And estimating power of noise from the FFT data, and estimating power of noise in a corresponding noise environment based on the estimator parameter.

In the updating step, the characteristic value is extracted with respect to only the signal for the noise section extracted using the voice detector, not the overall characteristic value for the input audio signal.

In the updating step, by comparing the characteristic value of the signal for the noise section with the characteristic value obtained in advance using the machine learning method in advance to determine the noise environment for the characteristic value of the signal for the noise section do.

In the updating step, the estimator parameter corresponding to the determined noise environment is output to the noise power estimator, among the noise environment estimator parameters previously obtained by preliminary evaluation.

The updating may include generating a separator of each section for distinguishing between a voice signal section and a noise section of the input audio signal; Extracting a characteristic value of a signal for a noise section of the input audio signal based on the delimiter; Sequentially storing and outputting the characteristic values in a buffer; And determining the type of the noise environment corresponding to the characteristic value for a predetermined time stored in the buffer, and updating and outputting the corresponding estimator parameter.

The hearing aid signal processing method includes calculating a gain for an estimated power of the noise; Multiplying the gain by the FFT processed data; Generating data obtained by IFFT processing the result of the multiplication; And sequentially storing and outputting data by the IFFT process in a buffer.

The hearing aid and the method according to the present invention determine the parameters of the noise power estimator optimized according to the noise environment and calculate the correct noise power to derive the gain optimized for the noise environment to improve the voice recognition ability of the hearing aid user in the noise situation. Can be improved.

In addition, instead of extracting the characteristics of the entire signal, a voice detector is used to derive the characteristics of the noise signal section only, which requires less computation than the structure of deriving the characteristics of the entire signal.

In addition, unlike the structure that trains various voice and noise situations and derives gain (noise power, SNR) based on this, the training process is simple in that only the environmental noise is trained and only the parameters of the noise power estimator are derived. And efficient.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are included as part of the detailed description in order to provide a thorough understanding of the present invention, provide examples of the present invention and together with the description, describe the technical idea of the present invention.
1 is a view for explaining the operation structure of a general hearing aid.
2 is a view for explaining the structure of the hearing aid having a noise environment classification and removal function according to an embodiment of the present invention.
FIG. 3 is a view for explaining characteristic values of noise environments by using a machine learning method in advance used in the noise environment classifier of the present invention.
FIG. 4 is a diagram for explaining a noise environment estimator parameter previously obtained by prior evaluation used in the noise environment classifier of the present invention.
5 is a flowchart illustrating a signal processing operation of a hearing aid having a noise environment classification and cancellation function according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail with respect to the present invention. In this case, the same components in each drawing are represented by the same reference numerals as much as possible. In addition, detailed descriptions of already known functions and / or configurations are omitted. The following description focuses on parts necessary for understanding the operation according to various embodiments, and descriptions of elements that may obscure the gist of the description are omitted. In addition, some components of the drawings may be exaggerated, omitted, or schematically illustrated. The size of each component does not entirely reflect the actual size, and thus the contents described herein are not limited by the relative size or spacing of the components drawn in the respective drawings.

In describing the embodiments of the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. In addition, terms to be described below are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of a user or an operator. Therefore, the definition should be made based on the contents throughout the specification. The terminology used in the description is for the purpose of describing particular embodiments only and should not be limiting. Unless expressly used otherwise, the singular forms “a,” “an,” and “the” include plural forms of meaning. In this description, expressions such as "comprises" or "equipment" are intended to indicate certain features, numbers, steps, actions, elements, portions or combinations thereof, and one or more than those described. It should not be construed to exclude the presence or possibility of other features, numbers, steps, actions, elements, portions or combinations thereof.

In addition, terms such as first and second may be used to describe various components, but the components are not limited by the terms, and the terms are used to distinguish one component from another component. Used only as

1 is a view for explaining the operation structure of a general hearing aid.

Referring to FIG. 1, a hearing aid generally includes a digital signal processor, a hardware accelerator for feedback cancellation, and a central processor that is responsible for overall control of the components.

Hearing aids are made in the form of chips to assist hearing loss so that users who can't hear well can hear them when they are plugged into their ears. To do this, the hearing aid removes and amplifies (eg, Wide Dynamic Range Compression (WDRC) compression) from the audio input signal generated by another person's voice and outputs the audio output signal through the micro speaker.

The hearing aid includes a digital signal processor for performing a hearing aid algorithm for improving the signal-to-noise ratio in a noisy environment. However, the speaker output is fed back at a close distance and becomes an input signal, thereby generating a ringing sound such as howling and echo. Therefore, in order to remove the influence of the feedback, the hearing aid includes a hardware accelerator using an adaptive filter, and the feedback cancellation hardware accelerator and the digital signal processing unit are interworked through an audio interface, so that the feedback removal and the hearing aid algorithm are simultaneously performed. It operates to output clean, enhanced audio signals.

Hereinafter, a hearing aid having a noise environment classification and cancellation function according to an embodiment of the present invention will be described with reference to FIGS. 2 to 5. The noise environment classification and cancellation function according to an embodiment of the present invention may be performed by the digital signal processor of FIG. 1.

2 is a view for explaining the structure of the hearing aid 100 having a noise environment classification and removal function according to an embodiment of the present invention.

Referring to FIG. 2, the hearing aid 100 according to an exemplary embodiment of the present invention includes a noise classification estimator 110, an input signal processor 120, a noise power estimator 180, a gain calculator 190, and a multiplier. 200, the IFFT unit 210 and the output buffer 220.

The noise classification estimator 110 determines the type of noise environment corresponding to the characteristic value of the signal for the divided noise section of the digital input audio signal (see S510 and S520 of FIG. 5), and determines the type of the corresponding noise environment. The matching estimator parameter is updated and provided to the noise power estimator 180 (see S530 of FIG. 5). To this end, the noise classification estimator 110 includes a voice detector 111, a feature extractor 112, a noise signal feature buffer 113, and a noise environment classifier 114. In the present invention, the noise classification estimator 110 extracts the characteristic value of only the signal for the noise section extracted using the voice detector 111, not the overall characteristic value of the input audio signal. The noise environment can be determined.

The audio signal from the hearing aid through a microphone (not shown) is converted into a digital signal through an analog to digital (A / D) converter. The modified digital signal may directly perform voice detection on the digital input audio signal according to the type of voice detector (e.g., VAD of standard codec G.729 in the time domain), or voice detection after frequency conversion (e.g., local speech). Absent Probability, Global Speech Absent Probability, etc.) or voice detection may be performed after feature extraction (eg, machine learning based algorithm). The following description will focus on the case where voice detection is performed directly on the digital input audio signal.

The voice detector 111 generates a delimiter for each section to distinguish between a voice signal section and a noise section of the digital input audio signal. For example, the voice detector 111 analyzes the input audio signal at a predetermined period such as a frame unit or a plurality of frame units to determine whether the corresponding unit section is a voice signal section or a noise section, and thus the corresponding delimiter for each unit section. Can be determined and corresponding identifier information / data (eg 0 or 1) can be generated. In addition, the voice detector 111 determines whether information on each frequency corresponds to a voice signal or noise through frequency-specific analysis for each unit section, thereby determining a separator for each frequency, and identifying the corresponding separator information / data. You can also create more.

The feature extractor 112 extracts the feature value of the signal for the noise section of the input audio signal based on the delimiter generated by the voice detector 111 (see S510 of FIG. 5). The feature extractor 112 is a feature value of an input signal by an algorithm for speech recognition in a noise environment. For example, the feature extractor 112 may extract one or more feature values such as a mean, a covariance matrix, and a weight. Can be. That is, in case of Gaussian Mixture Model (GMM) learning, the characteristic value of the speech signal may include an average of each component (Gaussian probability density function), a covariance, and a mixing coefficient (a ratio of the components in the overall mixing probability density function). In the case of neural network learning, weights and biases between the layers may be included.

In addition, the characteristic value of the signal for the noise section may include a characteristic value in the frequency domain or a characteristic value in the time domain. The characteristic value of the frequency domain may include Mel-frequency Cepstral Coefficients, Relative Spectra Perceptual Linear Prediction Coefficients, Discrete Fourier transform results, and the like. The characteristic value of the time domain may include Pitch, Linear Prediction Coefficients.

The noise signal characteristic buffer 113 sequentially stores and outputs the output of the characteristic extractor 112. The noise environment classifier 114 determines the type of the current noise environment corresponding to the characteristic value for a predetermined time (e.g., all or part of the buffer) stored in the noise signal characteristic buffer 113, and updates the corresponding estimator parameter to determine the noise. To the power estimator 180.

The noise environment classifier 114 is a reference characteristic value to be compared with the characteristic value output from the noise signal characteristic buffer 113. The noise environment classifier 114 stores the characteristic value of each noise environment previously obtained by using a machine learning method in a storage means such as a memory. In addition, the estimator parameter for each noise environment obtained by prior evaluation is stored in a storage means such as a memory.

FIG. 3 is a diagram for explaining characteristic values of noise environments for each noise environment previously obtained by using a machine learning method used in the noise environment classifier 114 of the present invention.

For example, by computer (e.g. Personal Computer, etc.), noise data (e.g., noisy environment, noise environment 1, noise environment 2, noise environment 3, ..) noise data (e.g., actual recording, Noisex-92) Etc.), and records and manages the data in a database S110, and extracts voice signal characteristic values such as mean, covariance matrix, and weight with respect to data of each noise environment using the feature extractor S120. do. For the extracted speech signal characteristic values, a teacher learning method may be applied or a non-private learning method may be used for clustering. For example, by using a machine learning algorithm such as SVM (Support Vector Machine), GMM (Gausian Mixture Model), Neural Network (Neural Network) of the machine learner (S130), the learning results of the voice signal characteristic values are stored in the memory (S140). By updating to, the noise signal characteristic value optimized for each noise environment can be provided to the noise environment classifier 114. In order to provide more accurate voice signal characteristic values of each noise environment, noise data for as many kinds of noise environments as possible can be collected, and the corresponding voice signal characteristic values can be learned and updated to be provided to the noise environment classifier 114. have.

The noise environment classifier 114 of the noise classification estimator 110 is configured to generate a signal for a noise section output from the noise signal characteristic buffer 113 and a noise environment characteristic value previously obtained by using a machine learning method. By comparing the characteristic values, it is possible to determine the noise environment of the characteristic values of the signal for the corresponding noise section (see S520 of FIG. 5).

FIG. 4 is a diagram for explaining the noise parameter for each noise environment, which is obtained in advance by prior evaluation used in the noise environment classifier 114 of the present invention.

For example, first, by the computer (eg, Personal Computer, etc.), by collecting the noise data for each noise environment (eg, noisy environment, noise environment 1, noise environment 2, noise environment 3, ..) database (S210) The recording management is performed in the recording process, and data for a predetermined voice signal without noise is collected and recorded in the database S220. Next, a noise contaminated with various signal-to-noise ratios is generated by synthesizing the noise of each noise environment from the databases S210 and S220 with a voice signal (S230), and then changing the estimator parameter for each noise environment to change the power of the noise. By estimating the noise power (see, for example, estimating the noise power of the noise power estimator 180), and calculating the gain for the power of the noise (for example, referring to the gain calculation of the gain calculator 190) to remove the noise (S240), The performance evaluation is performed on the result (S250) and updated in the memory (S260). When the noise canceling performance by the estimator parameter selected to optimize the noise canceling result for each noise environment satisfies a predetermined criterion, the corresponding noise environment estimator parameter may be provided to the noise environment classifier 114. Here, the performance evaluation of the noise cancellation algorithm may be based on a subjective method (MOS, mean opinion score) or an objective method (Perceptual Evaluation Speech Quality, Log-Likelihood Ratio, Segmental Signal to Noise Ratio, Weighted-Slop spectral, etc.). It can be made by measuring with a linear combined composite measure.

The noise environment classifier 114 of the noise classification estimator 110 determines the noise environment classifier 114 previously determined by using the noise environment characteristic values previously acquired by using the machine learning method, among the noise environment estimator parameters previously obtained by prior evaluation. The estimator parameter corresponding to the noise environment is output to the noise power estimator 180 (see S530 of FIG. 5).

In FIG. 1, the input signal processor 120 generates fast fourier transform (FFT) data of an input audio signal. To this end, the input signal processor 120 includes an input buffer 121, a window application unit 122, and an FFT unit 123.

The input buffer 121 sequentially stores and outputs an input audio signal. The window application unit 122 selects valid data from the output of the input buffer. The window application unit 122 may select and output valid data of a corresponding shape by using a predetermined FFT window function. The FFT unit 123 generates data (data in the frequency domain) obtained by FFT processing the output of the window application unit 122.

Accordingly, the noise power estimator 180 estimates the power of the noise from the FFT data of the FFT unit 123, and estimates the power of the noise in the corresponding noise environment based on the estimator parameter provided by the noise environment classifier 114. Can be.

The gain calculator 190 calculates a gain for the power of the noise estimated by the noise power estimator 180. The multiplier 200 multiplies the gain from the gain calculator 190 by the FFT-processed data and outputs a result of removing the noise. The IFFT unit 210 generates data (data in a time domain) obtained by IFFT (Inverse FFT) processing of the output of the multiplier 200. The output buffer 220 sequentially stores the output data of the IFFT unit 210 and outputs an improved audio signal to a speaker or the like (see S540 of FIG. 5).

The noise power estimator 180 uses the algorithm of noise N (k, l) using algorithms such as minimum tracking, minimum recursive averaging, improved MCRA, and histogram based algorithms. The power can be estimated. The estimator parameter is the power of the noise E {| N (k, l) | ² } may correspond to a coefficient such as α.

After the noise power estimation, the gain calculator 190 calculates the gain G (k, l) of the noise for noise cancellation based on this. The gain calculation algorithm may use Spectral Subtraction, Winner Filter, maximum-likelihood (ML), minimum mean square error (MMSE), log-MMSE, etc. The calculated gain is applied to Equation 1 to remove noise do.

[Equation 1]

로 나타낼 수 있고, 주파수 영역에서는

로 나타낼 수 있다. 여기에서 k와 l은 각각 주파수 영역에서 인덱스와 시간 영역에서 프레임 인덱스를 나타내며

는 출려되는 향상된 오디오 신호를 나타낸다. 여기서, 이득 계산부(190)의 잡음 제거 알고리즘과 관련된 처리는 다음과 같은 함수들 중 어느 하나 이상을 이용하여 수행될 수 있다. 하기에서

는 기대값 연산자(expectation operator)를 나타낸다. In the time domain, s (t) is the clear voice and n (t) is the noise.

In the frequency domain

It can be represented as. Where k and l represent the index in the frequency domain and the frame index in the time domain, respectively

Denotes the enhanced audio signal to be emitted. Here, the processing related to the noise canceling algorithm of the gain calculator 190 may be performed using any one or more of the following functions. In the following

Represents an expectation operator.

(1) a priori SNR:

(2) a posteriori SNR:

(3) Winer filter:

(4) spectral subtraction:

(5) Maximum-Likelihood:

(6) MMSE:

, I₀와 I₁은 각각 0차와 1차의 Modified Bessel 함수.here,

, I ₀ and I ₁ are Modified Bessel functions of 0th and 1st order, respectively.

(7) log-MMSE:

(8) decision-directed: a priori SNR estimation

은 이전 프레임의 향상된 오디오 신호이고,

은 추정된 잡음 전력이다.here,

Is the enhanced audio signal of the previous frame,

Is the estimated noise power.

을 이용하여 잡음 제거를 수행한다는 차이가 있다.On the other hand, the biggest difference between the conventional Korean Patent Publication No. 10-2013-0118513 and the present invention, in the case of the present invention by updating the parameters of the noise power estimator 180 optimized according to the noise environment through the noise environment classification While noise cancellation is performed by obtaining a noise power and calculating a gain based on this, in case of Korean Patent Publication No. 10-2013-0118513

There is a difference in that noise cancellation is performed using.

In addition, WO2010-015371 obtains information required for speech enhancement, which can be expressed as noise power, SNR, or gain, using machine learning, but the present invention learns the characteristics of noise to distinguish the noise environment and accordingly The difference is that it provides a parameter to calculate the optimal noise power. In addition, the present invention is different from the international patent publication WO2010-015371 in that it extracts only the noise signal using the voice detector 111.

As described above, the hearing aid 100 according to the present invention determines a parameter of the noise power estimator 180 optimized according to the noise environment and calculates accurate noise power to derive a gain optimized for the noise environment to obtain a noise situation. Improve the voice recognition ability of hearing aid users. In addition, instead of extracting the characteristic of the entire signal, the voice detector 111 is used to derive the characteristic of only the noise signal section, which requires less computation than the structure that requires derivation of the characteristic of the entire signal. In addition, unlike the structure that derives gains (noise power, SNR) based on the situation where various voices and noises are mixed and trains the situation, only the environmental noise is trained to obtain only the parameters of the noise power estimator 180 accordingly. The process is simple and efficient.

In the present invention as described above has been described by the specific embodiments, such as specific components and limited embodiments and drawings, but this is provided to help a more general understanding of the present invention, the present invention is not limited to the above embodiments. For those skilled in the art, various modifications and variations may be made without departing from the essential features of the present invention. Accordingly, the spirit of the present invention should not be limited to the described embodiments, and all technical ideas having equivalent or equivalent modifications to the claims as well as the following claims are included in the scope of the present invention. It should be interpreted as.

Noise classification estimator 110
Voice Detector (111)
Feature Extractor (112)
Noise Signal Characteristic Buffer (113)
Noise Environment Classifier (114)
Input signal processing unit 120
Input buffer (121)
Window application part (122)
FFT section (123)
Noise Power Estimator (180)
Gain calculator 190
Multiplier (200)
IFFT Part (210)
Output buffer (220)

Claims (13)

A noise classification estimator for determining a type of noise environment corresponding to the characteristic value of the signal for the divided noise section of the input audio signal and updating an estimator parameter matching the type of the noise environment;
An input signal processor configured to generate FFT data of the input audio signal; And
A noise power estimator estimating power of noise from the FFT data, and estimating power of noise in a corresponding noise environment based on the estimator parameter
Hearing aid comprising a. The method of claim 1,
And the noise classification estimator extracts the characteristic value from only the signal for the noise section extracted using a voice detector, not from the overall characteristic value of the input audio signal. The method of claim 1,
The noise classification estimator determines a noise environment of a signal characteristic value corresponding to the noise section by comparing a characteristic value of a noise environment obtained in advance using a machine learning method with a characteristic value of a signal for the noise section. Hearing aid, characterized in that. The method of claim 3,
And the noise classification estimator outputs, to the noise power estimator, the estimator parameter corresponding to the determined noise environment among the noise environment estimator parameters previously obtained by preliminary evaluation. The method of claim 1,
The noise classification estimator,
A voice detector for generating a separator of each section for distinguishing between a voice signal section and a noise section of the input audio signal;
A feature extractor for extracting a feature value of a signal for a noise section of the input audio signal based on the delimiter;
A noise signal characteristic buffer for sequentially storing and outputting the output of the characteristic extractor; And
A noise environment classifier which determines the type of the noise environment corresponding to the characteristic value for a predetermined time stored in the noise signal characteristic buffer and updates and outputs the corresponding estimator parameter.
Hearing aid comprising a. The method of claim 1,
The input signal processor,
An input buffer for sequentially storing and outputting the input audio signal;
A window application unit for selecting valid data from an output of the input buffer; And
FFT unit for generating data obtained by FFT processing the output of the window application unit
Hearing aid comprising a. The method of claim 6,
A gain calculator for calculating an estimated gain for the power of the noise;
A multiplier for multiplying the gain by the FFT processed data;
An IFFT unit for generating data obtained by IFFT processing the output of the multiplier; And
An output buffer for sequentially storing and outputting the output data of the IFFT unit
Hearing aid further comprising a. Determining a type of noise environment corresponding to the characteristic value of the signal for the divided noise section of the input audio signal and updating an estimator parameter matching the type of the noise environment;
Generating FFT data of the input audio signal; And
Estimating power of noise from the FFT data, and estimating power of noise in a corresponding noise environment based on the estimator parameter
Hearing aid signal processing method comprising a. The method of claim 8,
And in the updating step, extracting the characteristic value of only the signal for the noise section extracted using a voice detector, not the overall characteristic value of the input audio signal. The method of claim 8,
In the updating step, by comparing the characteristic values of the signal for the noise section with the noise values obtained in advance by using the machine learning method to determine the noise environment for the characteristic value of the signal for the noise section Hearing aid signal processing method characterized in that. The method of claim 10,
And in the updating step, outputting the estimator parameter corresponding to the determined noise environment from among the noise environment estimator parameters previously obtained by a prior evaluation, to the noise power estimator. The method of claim 8,
The updating step,
Generating a separator for each section for distinguishing between a voice signal section and a noise section of the input audio signal;
Extracting a characteristic value of a signal for a noise section of the input audio signal based on the delimiter;
Sequentially storing and outputting the characteristic values in a buffer; And
Determining a type of the noise environment corresponding to the characteristic value for a predetermined time stored in the buffer, and updating and outputting a corresponding estimator parameter
Hearing aid signal processing method comprising a. The method of claim 8,
Calculating a gain for the estimated power of the noise;
Multiplying the gain by the FFT processed data;
Generating data obtained by IFFT processing the result of the multiplication; And
Sequentially storing and outputting data by the IFFT process in a buffer
Hearing aid signal processing method further comprising.

Images