In order to detect a degree of voicing of a speech signal, an input speech signal is converted to a speech signal in the frequency domain, a pitch value is calculated from the speech signal, a plurality of harmonic peaks existing in the speech signal are detected, and a difference obtained by comparing the pitch value to an interval between adjacent harmonic peaks among the detected harmonic peaks is detected as the degree of voicing included in the speech signal.

Apparatus and method for detecting degree of voicing of speech signal

A quantization encoding decoding method and device, mainly include: in the quantization encoding end, if the closest point value is not included in a base codebook, the encoding operation is performed by a processing mode of splitting the closest point value; in the quantization decoding end, the decoding operation is performed by a corresponding processing mode.

A quantization encoding decoding method and device

본 발명은 음성 신호 부호화 장치의 다중 밴드 여기 보코더에 관한 것으로, 특히 다중 밴드 여기 보코더의 매개변수 추정시 피치, 밴드 크기, 밴드별 유성음/무성음의 최적값을 동시에 추정할 수 있도록, 합성 스펙트럼을 이용하여 피치를 정의하는 피치 후보값 선택부(10)와 ; 주어진 피치(τ) 에 대하여 주기적인 펄스 열과 윈도우 스펙트럼을 이용해서 고조파 여기를 생성하는 고조파 여기 발생부(20) ; 전체 밴드가 유성음(V)인 경우에 대하여 고조파 여기로부터 m-번째 밴드의 스펙트럼 크기를 구하는 유성음 밴드 크기 추정부(30) ; 유성음 밴드 크기와 고조파 여기 스펙트럼으로부터 유성음 스펙트럼을 합성하는 유성음 스펙트럼 합성부(40) ; 랜덤 스펙트럼과 윈도우 스펙트럼으로부터 무성음 스펙트럼을 생성하는 랜덤 여기 발생부(50) ; 전체 밴드가 무성음(UV)인 경우에 대하여 랜덤 여기로부터 m-번째 밴드의 스펙트럼 크기를 구하는 무성음 밴드 크기 추정부(60) ; 무성음 밴드 크기와 랜덤 여기 스펙트럼으로부터 무성음 스펙트럼을 합성하는 무성음 스펙트럼 합성부(70) ; 각 밴드에서 구한 유성음 스펙트럼과 무성음 스펙트럼 중에서 원 스펙트럼과의 스펙트럼 차가 작은 쪽을 취하는 유성음/무성음 결정부(80) 및 ; 최적의 매개변수 추정을 하기 위하여 후보값 매개변수를 평가하는 최소화부(90)로 구성한 것을 특징으로 하는, 다중 밴드 여기 음성 부호화기에서 매개변수 추정 장치 및 방법에 관한 것이다.

다중 밴드 여기 음성 부호화기에서 매개변수 추정 장치 및 방법

PURPOSE: A method for predictive-linked quantization for split LSF vectors is provided to realize a low mean SD(Spectral Distortion) and improve an outlier percentage. CONSTITUTION: A first quantizer(30) quantizes middle LSF(Line Spectrum Frequencies) vectors for generating the quantized middle LSF vectors and an index 1. A first classifier(32) inputs the first middle LSF vector for deciding an appropriate codebook out of three COLs. A second classifier(34) inputs the first middle LSF vector for deciding an appropriate codebook out of three COUs. A second quantizer(36) encodes lower LSF vectors by using the codebook decided by the first classifier for generating an index 2. A third quantizer(38) encodes upper LSFs by using the codebook decided by the second classifier for generating an index 3.

Method for predictive-linked quantization for split lsf vectors

PURPOSE: A method for quantizing a linear spectrum pair coefficient in coding voice is provided to select a code word minimizing a whole spectrum error by combining an error generated in an interpolation subframe with quantization. CONSTITUTION: A pth linear spectrum pair coefficient is extracted from a voice signal of an ith frame(201,202). A linear spectrum pair coefficient is extracted from the voice signal of the ith frame by subframes(207,208). Another linear spectrum pair coefficient used for synthesizing a voice of a previous frame is obtained. A third linear spectrum pair coefficient interpolated by subframes by linear interpolation is obtained. An index c minimizing a quantization error is obtained by using a codebook of linear spectrum pair coefficients.

Method for quantizing linear spectrum pair coefficient in coding voice

PURPOSE: A coding and decoding device robust against a random error is provided to convert a coded bit string to be of a robust type against errors before transmitting or recovering the bit string, thereby minimizing fading of tone quality. CONSTITUTION: A coding unit(200) performs vector quantization for sampled voice signals by referring to a vector quantization code book(202). A vector quantization code ordering unit(208) rearranges the sequence of vector quantization codewords, such that the codewords having high similarity outputted from the coding unit(200) are adjacent to each other. A mapping unit maps a created sequential binary code to a gray code. An inverse mapping unit inversely maps the gray code to the original sequential binary code. A vector quantization code inverse ordering unit(210) recovers the sequence of the vector quantization codewords rearranged such that the codewords having high similarity outputted from the inverse mapping unit are adjacent to each other. A decoding unit(204) performs inverse vector quantization for the resulting sequential binary code by referring to the vector quantization code book(202), to recover the sampled voice signals.

Encoding and decoding apparatus being robust to random error

PURPOSE: An adaptive codebook searching method based on a correlation function in code-excited linear prediction coding is provided to design an adaptive codebook with a scalar gain while having multi-tap structure CONSTITUTION: A past excite signal is modeled through a second-order smoothing filter A factor that minimizes the sum of squares of error signals is calculated An excite signal is obtained using the calculated factor A pitch delay and a gain value are obtained from the excite signal A gain vector is estimated from the past excite signal and a three-dimensional gain vector is represented in scalar from the estimated value Multi-gain is modeled into single-gain using correlation of the excite signal

Kim Hong Guk

Papers

다중 밴드 여기 음성 부호화기에서 매개변수 추정 장치 및 방법

Method for predictive-linked quantization for split lsf vectors

Method for quantizing linear spectrum pair coefficient in coding voice

Encoding and decoding apparatus being robust to random error

Adaptive codebook searching method based on correlation function in code-excited linear prediction coding