公开(公告)号：KR101565633B1

公开(公告)日：2015-11-13

申请号：KR1020120076634

申请日：2012-07-13

Applicant: 한국전자통신연구원 ,  광운대학교 산학협력단

Inventor： 이태진 ,  백승권 ,  김민제 ,  장대영 ,  서정일 ,  강경옥 ,  홍진우 ,  박호종 ,  박영철

IPC: G10L19/00

CPC classification number: G10L19/12 ,  G10L19/008 ,  G10L19/02 ,  G10L19/20 ,  G10L19/22 ,  G11C2207/16

Abstract: 음성/음악통합신호의부호화/복호화방법및 장치가개시된다. 음성/음악통합신호의부호화장치는, 입력신호의특성을분석하는입력신호분석부, 상기입력신호가음악특성신호인경우, 상기입력신호를주파수영역으로변환하여부호화하는제1 변환부호화부, 상기입력신호가음성특성신호인경우, 상기입력신호를 LPCLinear Predictive Coding)기반부호화하는 LPC 부호화부및 상기제1 변환부호화부의출력신호및 상기 LPC 부호화부의출력신호를이용하여비트스트림을생성하는비트스트림생성부를포함한다.

Abstract translation: 提供了一种用于对语音信号和音频信号进行整体编码和解码的装置和方法。 编码装置可以包括：输入信号分析器，用于分析输入信号的特性; 第一转换编码器，用于将输入信号转换为频域信号，并且在输入信号是音频特征信号时对输入信号进行编码; 线性预测编码（LPC）编码器，用于当输入信号是语音特征信号时执行输入信号的LPC编码; 以及比特流发生器，用于使用第一转换编码器的输出信号和LPC编码器的输出信号来产生比特流。

公开(公告)号：KR101527441B1

公开(公告)日：2015-06-11

申请号：KR1020110017283

申请日：2011-02-25

Applicant: 한국전자통신연구원

Inventor： 김민제 ,  백승권 ,  장인선 ,  이태진 ,  강경옥

IPC: H04R3/00 ,  G10L21/02

Abstract: 복수의채널을통해공간감을제공하는오디오신호에서특정한음원이채널사이의관계파라미터를기준으로특정한분포를가진다는가정을기반으로해당음원의분포를학습하여혼합신호로부터해당음원의에너지기여도만큼의분량을분리할수 있는음원분리장치및 방법이개시된다. 본발명에따르면, 특정음원의일반적인채널분포정보를대략적으로모델링한상황에서입력혼합신호에포함되어있는특정음원의채널분포를보다정밀하게예측할수 있으며, 기존의채널기반음원분리방식보다정교한분리가가능하다.

公开(公告)号：KR1020110029055A

公开(公告)日：2011-03-22

申请号：KR1020090122218

申请日：2009-12-10

Applicant: 한국전자통신연구원

Inventor： 김민제 ,  백승권 ,  강경옥 ,  장대영 ,  이태진 ,  장인선 ,  홍진우

IPC: G10L21/0272 ,  G10L25/81

Abstract: PURPOSE: A method and system for separating a music sound source without using a sound source database are provided to separate a sound source generated using a rhythm musical instrument based on characteristics of the rhythm musical instrument repeated in an aspect of time, and thereby may separate a sound source included in a mixed signal even when a learning database generated using a specific sound source is absent. CONSTITUTION: An NMPCF(Nonnegative Matrix Partial Co-Factorization) analysis unit(130) obtains info nation shared by the plurality of segments in accordance with an NMPCF algorithm. A rhythm instrument signal has frequency characteristics such as a pitch, that may not be easily changed, and may be repeatedly generated, whereby the shared information may correspond to information of a rhythm musical instrument. A target instrument signal separating unit(140) separates a target instrument signal corresponding to a specific sound source from the mixed signal by calculating an inner product between the entity matrices obtained by the NMPCF analysis unit. The target instrument signal is a signal including sounds generated using the rhythm musical instrument. A signal association unit(150) associates the target instrument signals for each of the plurality of segments separated in the target instrument signal separating unit.

Abstract translation: 目的：提供一种用于在不使用声源数据库的情况下分离音乐声源的方法和系统，用于基于在一段时间内重复的节奏乐器的特征来分离使用节奏乐器产生的声源，从而可以分离 即使当使用特定声源生成的学习数据库不存在时，也包含在混合信号中的声源。 构成：NMPCF（非负矩阵部分协同因子分解）分析单元（130）根据NMPCF算法获得由多个段共享的信息国。 节奏乐器信号具有诸如间距的频率特性，其可能不容易改变，并且可以重复地产生，由此共享信息可以对应于节奏乐器的信息。 目标仪器信号分离单元（140）通过计算由NMPCF分析单元获得的实体矩阵之间的内积，将与特定声源相对应的目标仪器信号与混合信号分离。 目标仪器信号是包括使用节奏乐器产生的声音的信号。 信号关联单元（150）将目标仪器信号分离单元中分离出的多个分段中的每个分段的目标仪器信号相关联。

公开(公告)号：KR1020110023693A

公开(公告)日：2011-03-08

申请号：KR1020090127946

申请日：2009-12-21

Applicant: 한국전자통신연구원

Inventor： 이태진 ,  김민제 ,  강경옥 ,  장대영 ,  장인선 ,  백승권 ,  홍진우

IPC: G10L19/20 ,  H04N5/60 ,  H04N21/233 ,  G11B20/10

Abstract: PURPOSE: An apparatus and a method for structuring bitstream for an object-based audio service are provided to reduce a degradation in the sound quality occurring due to an excessive volume control by designating an upper bound value and a lower bound value of a reproduction volume in an object-based audio service. CONSTITUTION: A bitstream splitting unit divides a bitstream into a file header, and an audio object frame through sound source splitting(110). A reproduction level information storage unit stores the reconstruction level information for the regeneration of the audio object within the file header(120). A preset storage unit stores the preset information for the regeneration of the audio object within the file header(130).

Abstract translation: 目的：提供一种用于构建用于基于对象的音频服务的比特流的装置和方法，以通过指定再现音量的上限值和下限值来减少由于过度音量控制而产生的声音质量下降 一种基于对象的音频服务。 构成：比特流分割单元通过声源分割将比特流分成文件头和音频对象帧（110）。 再现级别信息存储单元将用于音频对象再生的重建级别信息存储在文件头部（120）内。 预设存储单元将用于音频对象的再生的预设信息存储在文件头部（130）内。

公开(公告)号：KR1020110023688A

公开(公告)日：2011-03-08

申请号：KR1020090122217

申请日：2009-12-10

Applicant: 한국전자통신연구원 ,  포항공과대학교 산학협력단

Inventor： 김민제 ,  최승진 ,  유지호 ,  강경옥 ,  장인선 ,  홍진우

IPC: G10L21/0272 ,  G10L25/81

Abstract: PURPOSE: A system for separating musical sound source is provided to efficiently divide a specific sound source in a mixed sound source by reorganizing the mixed sound source into the specific sound source and other sound source. CONSTITUTION: A system for separating musical sound source includes a database(110), a time-frequency domain conversion unit(120), an NMPCF(Nonnegative Matrix Partial Co-Factorization) analysis unit(130), a target instrument signal separating unit(140), and a time domain signal conversion unit(150). The database stores information about a solo performance using a predetermined musical instrument, and transmits the information about the solo performance as a type of a predetermined sound source signal(x1). The predetermined sound source has a significantly great amount of data to include various characteristics of the predetermined sound source. In this case, a great amount of database signals may need to be processed for each sound source separation operation.

Abstract translation: 目的：提供一种用于分离音乐声源的系统，通过将混合声源重组为特定声源和其他声源来有效地分割混合声源中的特定声源。 用于分离音乐声源的系统包括数据库（110），时间 - 频域转换单元（120），NMPCF（非负矩阵部分因子分解）分析单元（130），目标仪器信号分离单元 140）和时域信号转换单元（150）。 数据库使用预定的乐器存储关于个人演奏的信息，并且发送关于个性表演的信息作为预定声源信号的类型（x1）。 预定声源具有显着大量的数据以包括预定声源的各种特性。 在这种情况下，可能需要为每个声源分离操作处理大量的数据库信号。

公开(公告)号：KR1020100059726A

公开(公告)日：2010-06-04

申请号：KR1020090114783

申请日：2009-11-25

Applicant: 한국전자통신연구원 ,  광운대학교 산학협력단

Inventor： 백승권 ,  이태진 ,  김민제 ,  강경옥 ,  장대영 ,  서정일 ,  홍진우 ,  안치득 ,  박호종 ,  박영철

IPC: G10L19/04 ,  G10L19/18 ,  H03M7/30 ,  H04N21/233

CPC classification number: G10L19/22 ,  G10L19/022 ,  G10L19/20

Abstract: PURPOSE: A unified speech/audio coder(USAC) processing windows sequence based mode switching is provided to perform encoding or decoding by processing different window sequences for different situation, thereby improving coding efficiency. CONSTITUTION: A mode switching unit performs inter-linear prediction domain mode switching for the sub-frames which constitute a frame of an input signal. An encoding unit applies a window sequence to current encoding subframes, wherein the window sequence is based on a switched LPD(Linear Prediction Domain) mode. In addition, the encoding unit applies a changed window sequence to the current encoding subframe according to the LPD modes of previous and following subframes.

Abstract translation: 目的：提供统一的语音/音频编码器（USAC）处理基于Windows序列的模式切换，通过对不同情况处理不同的窗口序列进行编码或解码，从而提高编码效率。 构成：模式切换单元对构成输入信号的帧的子帧进行线间预测域模式切换。 编码单元将窗口序列应用于当前编码子帧，其中窗口序列基于切换的LPD（线性预测域）模式。 此外，编码单元根据前一个和后一个子帧的LPD模式将改变的窗口序列应用于当前编码子帧。

公开(公告)号：KR1020100007749A

公开(公告)日：2010-01-22

申请号：KR1020090062070

申请日：2009-07-08

Applicant: 한국전자통신연구원 ,  광운대학교 산학협력단

Inventor： 이태진 ,  백승권 ,  김민제 ,  장대영 ,  서정일 ,  강경옥 ,  홍진우 ,  박호종 ,  박영철

IPC: G10L19/04 ,  G10L19/18 ,  G10L19/008

CPC classification number: G10L19/12 ,  G10L19/008 ,  G10L19/02 ,  G10L19/20 ,  G10L19/22 ,  G11C2207/16

Abstract: PURPOSE: An apparatus and a method for encoding and decoding an integrated voice and music signal are provided to effectively select internal modules according to a characteristic of an input signal, thereby providing excellent sound quality to both a voice signal and a music signal at various bit rates. CONSTITUTION: An input signal analysis unit(110) analyzes a characteristic of an input signal. When the input signal is a music characteristic signal, the first conversion encoding unit(120) converts the input signal into a frequency domain to encode. When the input signal is a voice characteristic signal, an LPC(Linear Predictive Coding) encoder encodes the input signal based on LPC. A bit stream generator(140) uses output signals of the first conversion encoding unit and the LPC encoder to generate a bit stream.

Abstract translation: 目的：提供一种用于编码和解码集成语音和音乐信号的装置和方法，以根据输入信号的特性有效地选择内部模块，从而为语音信号和音乐信号在各种位提供优良的音质 率。 构成：输入信号分析单元（110）分析输入信号的特性。 当输入信号是音乐特征信号时，第一转换编码单元（120）将输入信号转换成频域进行编码。 当输入信号是语音特征信号时，LPC（线性预测编码）编码器基于LPC对输入信号进行编码。 比特流生成器（140）使用第一转换编码单元和LPC编码器的输出信号来产生比特流。

公开(公告)号：KR1020100007741A

公开(公告)日：2010-01-22

申请号：KR1020090061645

申请日：2009-07-07

Applicant: 한국전자통신연구원 ,  광운대학교 산학협력단

Inventor： 백승권 ,  이태진 ,  김민제 ,  장대영 ,  강경옥 ,  서정일 ,  홍진우 ,  박호종 ,  박영철

IPC: G10L19/18 ,  G10L19/04 ,  G10L19/12

CPC classification number: G10L19/22 ,  G10L19/0212 ,  G10L19/04 ,  G10L19/20

Abstract: PURPOSE: An apparatus for determining a signal state of an audio signal is provided to suitably select a voice encoder based on LPC(Linear Predictive Coding) and an audio encoder based on conversion according to a characteristic of an input signal. CONSTITUTION: A signal state observer(101) classifies characteristics of an input signal to output each state observation probability. A state chain unit(102) outputs a state identifier of a frame of the input signal based on the state observation probability. An encoder encoding a frame of the input signal is determined according to the state identifier. The signal state observer comprises a characteristic extracting unit, an entropy base determining tree and a silence state determining unit.

Abstract translation: 目的：提供一种用于确定音频信号的信号状态的装置，以根据输入信号的特性，基于LPC（线性预测编码）和基于转换的音频编码器适当地选择语音编码器。 构成：信号状态观察器（101）对输入信号的特性进行分类以输出每个状态观测概率。 状态链单元（102）基于状态观察概率输出输入信号的帧的状态标识符。 根据状态标识符确定编码输入信号的帧的编码器。 信号状态观察器包括特征提取单元，熵基准确定树和静默状态确定单元。

公开(公告)号：KR100653173B1

公开(公告)日：2006-12-05

申请号：KR1020050114409

申请日：2005-11-28

Applicant: 한국전자통신연구원 ,  학교법인 포항공과대학교

Inventor： 장인선 ,  김민제 ,  백승권 ,  강경옥 ,  최승진 ,  홍진우

IPC: G10L21/0272 ,  G10L21/02

Abstract: A method and an apparatus for solving the permutation ambiguity of a separation coefficient of a multi-channel mixed signal are provided to prevent the mixing of different sound source signals, thereby accurately separating an original signal from the mixed signal. An ICA(Independent Component Analysis) input matrix is created in a row vector type of a back-mixing matrix corresponding to a separation coefficient(401). A base vector matrix and a deflection coefficient matrix are extracted using the created ICA method(402). A base vector index value regarding a column vector of a mixing vector is selected by comparing the extracted base vector matrix and deflection coefficient matrix with each other(403). A corresponding clustering including the column vector of the mixing vector is determined based on the selected base vector index value(404).

Abstract translation: 提供了一种用于解决多声道混合信号的分离系数的置换模糊的方法和装置，以防止不同声源信号的混合，从而将混合信号中的原始信号准确地分离。 在与分离系数相对应的反向混合矩阵的行向量类型中创建ICA（独立分量分析）输入矩阵（401）。 使用所创建的ICA方法提取基矢量矩阵和偏转系数矩阵（402）。 通过将提取的基矢量矩阵和偏转系数矩阵相互比较来选择关于混合矢量的列矢量的基矢量索引值（403）。 基于所选择的基矢量索引值（404）确定包括混合矢量的列矢量的相应聚类。

公开(公告)号：KR1020210012031A

公开(公告)日：2021-02-02

申请号：KR1020210010462

申请日：2021-01-25

Applicant: 한국전자통신연구원 ,  광운대학교 산학협력단

Inventor： 백승권 ,  이태진 ,  김민제 ,  강경옥 ,  장대영 ,  홍진우 ,  박영철 ,  박호종

IPC: G10L19/16 ,  G10L19/02 ,  G10L19/18

Abstract: MDCT 기반의코더와이종의코더간 변환에서의인코딩장치및 디코딩장치가개시된다. 인코딩장치는 MDCT 기반의코더와이종의코더간 변환이발생할때, MDCT 기반의코딩방식에따라인코딩된입력신호를복원하기위해부가정보를코딩할수 있다. 인코딩장치에의하면, 불필요한비트스트림의생성을방지하는한편, 최소한의부가정보를인코딩할수 있다.