KR930010851A - Adaptive Extraction Method of Start Point and End Point of Speech Signal - Google Patents

Adaptive Extraction Method of Start Point and End Point of Speech Signal Download PDF

Info

Publication number
KR930010851A
KR930010851A KR1019910020912A KR910020912A KR930010851A KR 930010851 A KR930010851 A KR 930010851A KR 1019910020912 A KR1019910020912 A KR 1019910020912A KR 910020912 A KR910020912 A KR 910020912A KR 930010851 A KR930010851 A KR 930010851A
Authority
KR
South Korea
Prior art keywords
frame
point
speech signal
signal
end point
Prior art date
Application number
KR1019910020912A
Other languages
Korean (ko)
Other versions
KR940002853B1 (en
Inventor
이순건
Original Assignee
강진구
삼성전자 주식회사
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 강진구, 삼성전자 주식회사 filed Critical 강진구
Priority to KR1019910020912A priority Critical patent/KR940002853B1/en
Publication of KR930010851A publication Critical patent/KR930010851A/en
Application granted granted Critical
Publication of KR940002853B1 publication Critical patent/KR940002853B1/en

Links

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/78—Detection of presence or absence of voice signals
    • G10L25/84—Detection of presence or absence of voice signals for discriminating voice from noise

Landscapes

  • Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Telephonic Communication Services (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Abstract

내용 없음No content

Description

음성신호의 시작점 및 끝점의 적응적 추출방법Adaptive Extraction Method of Start Point and End Point of Speech Signal

본 내용은 요부공개 건이므로 전문내용을 수록하지 않았음Since this is an open matter, no full text was included.

제1도는 본 발명에 적용되는 음성신호처리 시스템도.1 is a voice signal processing system applied to the present invention.

제2도는 본 발명의 음성신호처리 제어 흐름도.2 is a flowchart of a voice signal processing control of the present invention.

제3도는 본 발명을 설명하기 위해 적용된 음성신호의 에너지 분포도.3 is an energy distribution diagram of a speech signal applied to explain the present invention.

제4도는 제2도중 통계 파라미터 추출과정의 세부 제어 흐름도.4 is a detailed control flowchart of the statistical parameter extraction process of FIG.

제5도는 제2도에 따른 블록 이동을 위한 제1도중의 메모리(20) 구성도.5 is a configuration diagram of the memory 20 in FIG. 1 for block movement according to FIG.

제6도는 제2도중 시작점 추출과정의 세부 제어 흐름도.6 is a detailed control flowchart of the starting point extraction process of FIG.

제7도는 제2도중 음성전처리 및 음성분석 과정의 세부 제어 흐름도.7 is a detailed control flowchart of the voice preprocessing and voice analysis process in FIG.

제8도는 제2도중 음성의 끝점 추출 과정의 세부 제어 흐름도.8 is a detailed control flowchart of a process of extracting the endpoint of speech during FIG. 2.

Claims (4)

디지탈 음성신호의 처리방법에 있어서, 음성신호에 대하여 주변잡음이 배제된 음성만을 추출하기 위하여 상기 디지탈 음성신호를 소정 프레임 단위로 입력하여 상기 주변 잡음의 에너지 레벨을 계산하는 통계파라미터 추출과정과, 상기 디지탈 음성신호에 대한 프레임의 블록 이동후 상기 에너지 레벨을 이용하여 상기 음성신호에 대한 프레임의 시작점을 추출하기 위한 시작점 추출과정과, 상기 에너지 레벨을 이용하여 상기 음성신호에 대한 프레임의 끝점을 추출하기 위한 끝점 추출과정으로 이루어져 주변잡음에 따라 적응적으로 음성구간을 추출하는 것을 특징으로 하는 음성신호의 시작점 및 끝점의 적응적 추출방법.In the method of processing a digital voice signal, Statistical parameter extraction step of calculating the energy level of the ambient noise by inputting the digital voice signal in a predetermined frame unit to extract only the voice of the ambient noise is excluded from the voice signal, Starting point extraction process for extracting the starting point of the frame for the speech signal using the energy level after block movement of the frame for the digital speech signal, and extracting the end point of the frame for the speech signal using the energy level. Adaptive extraction method of the starting point and the end point of the speech signal, characterized in that the extraction process is performed by the end point extraction process adaptively extracting the speech section. 제1항에 있어서, 상기 통계 파라미터 추출과정이 상기 에너지 레벨을 계산하기 위해 상기 디지탈 음성신호의샘플링 데이타가 입력되면 상기 디지탈 음성신호에 대한 단위 프레임당 평균에너지를 구한뒤 1.6배함에 의해 LTL을 산출하고 상기 LTL을 상수(2,3,4)와 곱함에 의해 UTL을 산출하는 것을 특징으로 하는 음성신호의 시작점 및 끝점의 적응적 추출방법.The method of claim 1, wherein the statistical parameter extraction process calculates the LTL by obtaining 1.6 times the average energy per unit frame for the digital voice signal when sampling data of the digital voice signal is input to calculate the energy level. And calculating the UTL by multiplying the LTL by a constant (2, 3, 4). 제2항에 있어서, 상기 시작점 추출과정이 상기 LTL보다 현재 입력되는 상기 디지탈 음성 데이타의 프레임의 에너지가 크고 적어도 상기 UTL보다 1번 이상 클 경우에 최초의 음성신호의 시작점으로 정하여 추출하는 것을 특징으로 하는 음성신호의 시작점 및 끝점의 적응적 추출방법.3. The method of claim 2, wherein the start point extraction process extracts the start point of the first audio signal when the energy of the frame of the digital voice data currently input is greater than the LTL and at least one time larger than the UTL. Adaptive extraction method of the start and end points of the speech signal. 제2항에 있어서, 상기 끝점 추출과정이 상기 시작점 추출과정 수행후 상기 LTL보다 현재 입력되는 상기 디지탈 음성 데이타의 프레임의 에너지가 작을 경우에 음성신호의 끝점으로 정하여 추출하는 것을 특징으로 하는 음성신호의 시작점 및 끝점의 적응적 추출방법.3. The method of claim 2, wherein the end point extracting step extracts the end point of the voice signal when the energy of the frame of the digital voice data currently input is smaller than the LTL after performing the start point extracting step. Adaptive extraction of start and end points. ※ 참고사항 : 최초출원 내용에 의하여 공개하는 것임.※ Note: The disclosure is based on the initial application.
KR1019910020912A 1991-11-22 1991-11-22 Adaptationally sampling method for starting and finishing points of a sound signal KR940002853B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1019910020912A KR940002853B1 (en) 1991-11-22 1991-11-22 Adaptationally sampling method for starting and finishing points of a sound signal

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1019910020912A KR940002853B1 (en) 1991-11-22 1991-11-22 Adaptationally sampling method for starting and finishing points of a sound signal

Publications (2)

Publication Number Publication Date
KR930010851A true KR930010851A (en) 1993-06-23
KR940002853B1 KR940002853B1 (en) 1994-04-04

Family

ID=19323291

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1019910020912A KR940002853B1 (en) 1991-11-22 1991-11-22 Adaptationally sampling method for starting and finishing points of a sound signal

Country Status (1)

Country Link
KR (1) KR940002853B1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101718073B1 (en) * 2016-01-26 2017-03-20 주식회사 시그널웍스 Method for detecting unusual sound and apparatus for executing the method

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101718073B1 (en) * 2016-01-26 2017-03-20 주식회사 시그널웍스 Method for detecting unusual sound and apparatus for executing the method

Also Published As

Publication number Publication date
KR940002853B1 (en) 1994-04-04

Similar Documents

Publication Publication Date Title
CN106653056B (en) Fundamental frequency extraction model and training method based on LSTM recurrent neural network
US4811399A (en) Apparatus and method for automatic speech recognition
KR910020641A (en) Noise Prediction Device and Signal Processing Device Using It
US5671330A (en) Speech synthesis using glottal closure instants determined from adaptively-thresholded wavelet transforms
US3509280A (en) Adaptive speech pattern recognition system
EP0780828A3 (en) Method and system for performing speech recognition
KR970049740A (en) Phoneme Segmentation Using Multi-Layer Neural Networks
KR910015962A (en) Voice signal processing device
JP2000047696A (en) Information processing method, information processor and storage medium therefor
KR930010851A (en) Adaptive Extraction Method of Start Point and End Point of Speech Signal
CN115331703A (en) Song voice detection method and device
KR20220012473A (en) Apparatus and method unsupervised pretraining speaker embedding extraction system using mutual information neural estimator, computer-readable storage medium and computer program
JPH0114599B2 (en)
Chelloug et al. Real Time Implementation of Voice Activity Detection based on False Acceptance Regulation.
KR100273395B1 (en) Voice duration detection method for voice recognizing system
JP3346200B2 (en) Voice recognition device
Chelloug et al. Robust Voice Activity Detection Against Non Homogeneous Noisy Environments
JP3190231B2 (en) Apparatus and method for extracting pitch period of voiced sound signal
KR20060095237A (en) Noise exclusion method of voice recognition system on vehicle
JPH07334189A (en) Sound information analysis device
KR100523905B1 (en) Dual Speech Detection Method of The Startpoint and The Endpoint in Speech Recognition
JP2772598B2 (en) Audio coding device
CN118197347A (en) Audio noise reduction method and device, electronic equipment and storage medium
CN115526180A (en) Method and system for processing context-dependent word components of local information
CN117746831A (en) Emotion controllable speech synthesis method and system based on few samples of specific characters

Legal Events

Date Code Title Description
A201 Request for examination
G160 Decision to publish patent application
E701 Decision to grant or registration of patent right
GRNT Written decision to grant
FPAY Annual fee payment

Payment date: 20030328

Year of fee payment: 10

LAPS Lapse due to unpaid annual fee