-
公开(公告)号：KR1020130032980A
公开(公告)日：2013-04-03
申请号：KR1020110096750
申请日：2011-09-26
Applicant: 한국전자통신연구원
IPC: G10L19/00
CPC classification number: G10L19/0208 , G10L19/0017 , G10L19/002 , G10L19/038 , H03M7/3082 , H04N19/94
Abstract: PURPOSE: A coding device using residual bits and a method thereof are provided to remove noise caused by excessively estimated gain by quantizing all gains of sub bands to bits which are not allocated in an AVQ(Algebraic Vector Quantization) process. CONSTITUTION: An AVQ performing unit(101) receives frequency coefficients converted from a voice/audio signal to perform an AVQ process. First performance improving units(102-106) quantize all gains of sub bands to bits which are not allocated in the AVQ process according to residual bits in order to improve performance. [Reference numerals] (101) AVQ performing unit; (102) Gain parameter calculating unit; (103) Residual bit(n_bits) calculating unit; (104) Whole band gain calculating unit; (105) Whole band gain code book selecting unit; (106) Whole band gain quantizing unit; (107) Sub band gain code book selecting unit; (108) Sub band gain calculating unit; (109) Sub band gain quantizing unit; (110) Shape parameter calculating unit; (111) Residual frequency coefficient calculating unit; (112) Residual frequency coefficient quantizing unit; (113) Multiplexing unit; (AA) Frequency coefficients; (BB) Bit stream;
Abstract translation: 目的：提供使用残余比特的编码装置及其方法，以通过将子带的所有增益量化为在AVQ（代数向量量化）处理中未分配的比特来消除由过度估计的增益引起的噪声。 构成：AVQ执行单元（101）接收从语音/音频信号转换的频率系数，以执行AVQ处理。 第一性能改进单元（102-106）将子频带的所有增益量化到根据残余比特未被分配在AVQ进程中的比特，以便提高性能。 （101）AVQ执行单元; （102）增益参数计算单元; （103）剩余位（n_bits）计算单元; （104）全带增益计算单位; （105）全带增益码本选择单元; （106）全带增益量化单位; （107）子带增益码本选择单元; （108）子带增益计算单元; （109）子带增益量化单元; （110）形状参数计算单位; （111）残余频率系数计算单位; （112）残余频率系数量化单位; （113）复用单元; （AA）频率系数; （BB）比特流;
-
公开(公告)号：KR1020120063421A
公开(公告)日：2012-06-15
申请号：KR1020110064854
申请日：2011-06-30
Applicant: 한국전자통신연구원
CPC classification number: G10K11/16 , G10K2210/108 , G10K2210/3028 , G10L21/0208
Abstract: PURPOSE: An acoustic echo removing device using an adaptive filter and method thereof are provided to improve an echo signal removing speed by omitting a process required for renewing a filter coefficient because values of the filter coefficients of predetermined numbers are maintained previous values when being decided the filter coefficients are collected in a process of removing acoustic echo using an adaptive filter algorithm. CONSTITUTION: A method for removing acoustic echo is same like next. Error signals are obtained by subtracting similar echo signals generated by long-distance signals from short distance signals. Whether a low power mode is set or not is decided based on the error signals(S130). If the low power mode is set, one or more filter coefficients used for an adaptive filter algorithm to remove acoustic echo is set as coefficient values previously used.
Abstract translation: 目的：提供一种使用自适应滤波器的声学回波消除装置及其方法，用于通过省略更新滤波器系数所需的处理来改善回波信号去除速度，因为当确定预定数量的滤波器系数的值被保持为先前值时 在使用自适应滤波器算法去除声学回波的过程中收集滤波器系数。 构成：去除声学回声的方法与下一样相同。 通过从短距离信号中减去长距离信号产生的相似回波信号来获得误差信号。 基于误差信号决定低功率模式是否被设定（S130）。 如果设置了低功率模式，则将用于去除声学回声的自适应滤波器算法所使用的一个或多个滤波器系数设置为先前使用的系数值。
-
公开(公告)号：KR1020120055945A
公开(公告)日：2012-06-01
申请号：KR1020100117407
申请日：2010-11-24
Applicant: 한국전자통신연구원
CPC classification number: H04L12/66
Abstract: PURPOSE: A packet route management apparatus, voice communication system based on the Internet, and voice quality guarantee method are provided to prevent packet delay and packet loss by processing VoIP(Voice over Internet Protocol) voice packets in real time. CONSTITUTION: A communication unit(2000) collects bandwidth information and the address information of a packet route apparatus. A control unit(2010) determines the bandwidth availability of the packet route apparatus by using the collected address information and bandwidth information. The control unit establishes a packet route which transmits voice packets according to determination results. A database(2020) stores the bandwidth information and the address information of the packet apparatus.
Abstract translation: 目的：提供一种分组路由管理装置，基于互联网的语音通信系统和语音质量保证方法，通过实时处理VoIP（Voice over Internet Protocol，语音电话）语音包来防止分组延迟和分组丢失。 构成：通信单元（2000）收集分组路由装置的带宽信息和地址信息。 控制单元（2010）通过使用收集的地址信息和带宽信息来确定分组路由装置的带宽可用性。 控制单元建立根据确定结果发送语音分组的分组路由。 数据库（2020）存储分组装置的带宽信息和地址信息。
-
公开(公告)号：KR1020120033847A
公开(公告)日：2012-04-09
申请号：KR1020100095577
申请日：2010-09-30
Applicant: 한국전자통신연구원
CPC classification number: H04L65/608 , H04L49/90 , H04L65/80
Abstract: PURPOSE: A jitter buffer controller, an electronic device, and a method thereof are provided to predict the future jitter buffer size by applying past jitter buffer size based a determined jitter buffer size which is determined by an adaptive jitter buffer size algorithm. CONSTITUTION: A jitter bugger controller(100) estimates the future jitter buffer size by using the calculated jitter buffer size. The jitter buffer size is calculated by a jitter buffer unit(24). A jitter buffer size predictor(102) estimates the jitter buffer size by using an ARMA(Auto-Regressive Moving Average) models. A jitter buffer size selector(104) selects the smallest jitter buffer size as the jitter buffer size to buffer among the estimated jitter buffer sizes and the current jitter buffer size.
Abstract translation: 目的：提供抖动缓冲控制器，电子设备及其方法，以通过基于由自适应抖动缓冲器大小算法确定的确定的抖动缓冲器大小应用过去的抖动缓冲器大小来预测未来的抖动缓冲器大小。 构成：抖动bugger控制器（100）通过使用计算的抖动缓冲区大小来估计未来的抖动缓冲区大小。 抖动缓冲器大小由抖动缓冲器单元（24）计算。 抖动缓冲器大小预测器（102）通过使用ARMA（自回归移动平均）模型来估计抖动缓冲器大小。 抖动缓冲器大小选择器（104）选择最小的抖动缓冲器大小作为在估计的抖动缓冲器大小和当前抖动缓冲器大小之间缓冲的抖动缓冲器大小。
-
公开(公告)号：KR1020110068961A
公开(公告)日：2011-06-22
申请号：KR1020110045248
申请日：2011-05-13
Applicant: 한국전자통신연구원
IPC: G10L19/005 , G10L19/00
CPC classification number: G10L21/0208 , G10L19/24 , G10L21/0316 , G10L21/038 , G11B20/10527 , G11B2020/10555 , H03G3/341
Abstract: PURPOSE: A quality improving apparatus of audio codec and method thereof are provided to reduce noises due to the quantization error of mute section and to increase the quality of audio codec. CONSTITUTION: A first energy calculating unit(300) adds up energy about each sample. The first energy calculator obtains the energy of one frame. In case a low band improvement mode operates, a second energy calculating unit(310) obtains the energy of a signal that is decoded through the low band improvement mode. A scaling unit(320) scales the size of the signal that is decoded through a kernel codec.
Abstract translation: 目的：提供音频编解码器的质量改进装置及其方法，以减少由于静音部分的量化误差引起的噪声并提高音频编解码器的质量。 构成：第一能量计算单元（300）将关于每个样品的能量相加。 第一个能量计算器获得一帧的能量。 在低频带改善模式工作的情况下，第二能量计算单元（310）获得通过低频带改善模式解码的信号的能量。 缩放单元（320）缩放通过内核编解码器解码的信号的大小。
-
Patent Agency Ranking146.发明公开광대역 음성 인터넷 전화 서비스를 제공하는 셋탑박스 및 이를 이용한 광대역 음성 인터넷 전화 서비스 제공방법 有权包括宽带语音互联网电话的INERNET SETTOP-BOX的系统和方法
公开(公告)号：KR1020110067972A
公开(公告)日：2011-06-22
申请号：KR1020090124774
申请日：2009-12-15
Applicant: 한국전자통신연구원
CPC classification number: H04L12/66 , H04N21/478 , H04N21/6437
Abstract: PURPOSE: A set-top box and wideband voice internet phone service supplying method using the same are provided to implement a wideband voice internet phone service through wideband voice codec. CONSTITUTION: An interface unit(210) transmits and receives a broadcasting signal or an internet phone signal. A connection processing unit(220) outputs the broadcasting signal or the internet phone signal. A wideband voice processing unit(260) converts the internet phone signal into a voice signal through a wideband voice codec. The wideband voice processing unit converts the inputted voice signal into the internet phone signal.
Abstract translation: 目的：提供一种使用该机顶盒和宽带语音互联网电话服务提供方法，通过宽带语音编解码器实现宽带语音互联网电话业务。 构成：接口单元（210）发送和接收广播信号或互联网电话信号。 连接处理单元（220）输出广播信号或互联网电话信号。 宽带语音处理单元（260）通过宽带语音编解码器将互联网电话信号转换成语音信号。 宽带语音处理单元将输入的语音信号转换为互联网电话信号。
-
公开(公告)号：KR100948840B1
公开(公告)日：2010-03-22
申请号：KR1020070132730
申请日：2007-12-17
Applicant: 한국전자통신연구원
CPC classification number: H04W28/22 , H04L1/0014 , H04W28/02 , H04W88/181
Abstract: 무선랜 환경에서의 음성 품질 제어 방법 및 그 장치가 개시된다. 액세스 포인트에 연결되어 있는 단말들의 채널 점유 시간을 파악하기 위한 채널상태 정보를 수집하여 파악한, 채널 총 용량에 대한 단말들의 채널 점유율을 기초로 단말들의 코덱 비트율을 조정한 후, 그 조정값을 각 단말들에게 전송함으로써 단말의 코덱 비트율을 제어한다.
-
公开(公告)号：KR100922897B1
公开(公告)日：2009-10-20
申请号：KR1020070128525
申请日：2007-12-11
Applicant: 한국전자통신연구원
CPC classification number: G10L19/26 , G10L19/0212
Abstract: 본 발명은 MDCT 영역에서 동작하는 후처리 필터장치 및 필터방법에 관한 것으로, 더욱 상세하게는 과거와 현재의 MDCT 계수를 이용하여 실제 음성 스펙트럼과 유사한 스펙트럼 계수를 얻고 그 계수 크기가 작은 곳에서는 미분값이 크도록, 계수 크기가 큰 곳에서는 미분값이 작도록 볼록 함수로 변환하여 후처리 필터 계수를 구하고 이를 MDCT 계수에 적용하여 음성 신호 왜곡 없이 코딩 잡음을 줄이는 후처리 필터장치 및 필터방법에 관한 것이다. 이에 의하면, 현재와 과거의 MDCT 값을 모두 사용하기 때문에 실제 음성 스펙트럼과 유사한 계수를 획득하는 것이 가능하고 더 정확한 필터 계수를 얻을 수 있다. 또한, 볼록 함수에 의해 계수를 적절히 변환하였기 때문에 음질을 향상시킬 수 있다.
후처리 필터(post-filter), 코덱(codec), MDCT(modified discrete cosine transform), 볼록 함수, 음성 스펙트럼-
公开(公告)号：KR1020090061499A
公开(公告)日：2009-06-16
申请号：KR1020070128525
申请日：2007-12-11
Applicant: 한국전자통신연구원
CPC classification number: G10L19/26 , G10L19/0212
Abstract: A post-processing filter apparatus for improving sound quality in an MDCT(Modified Discrete Cosine Transform) area and a filter method are provided to use MDCT functions of both previous frame and current frame, thereby obtaining more similar coefficients to substantial voice spectrum. A spectrum coefficient generator(101) generates spectrum coefficients by using MDCT coefficients of the current voice frame and a previous voice frame. A normalizing unit(102) normalizes the generated spectrum coefficients. A transforming unit(103) maps the normalized spectrum coefficients with convex functions to generate converted spectrum coefficients. A filter coefficient generator(104) generates filter coefficients by controlling reflecting degrees of the converted spectrum coefficients. An MDCT coefficient generator(105) generates new MDCT coefficients by multiplying the filter coefficients by the MDCT coefficients of the current voice frame.
Abstract translation: 提供了一种用于改善MDCT（改进离散余弦变换）区域中的声音质量的后处理滤波器装置和滤波器方法，以使用先前帧和当前帧的MDCT功能，从而获得与实质语音频谱更相似的系数。 频谱系数生成器（101）通过使用当前语音帧和先前语音帧的MDCT系数来生成频谱系数。 归一化单元（102）对所生成的频谱系数进行归一化。 变换单元（103）将具有凸函数的归一化频谱系数映射以产生转换的频谱系数。 滤波器系数发生器（104）通过控制转换的频谱系数的反射程度来产生滤波器系数。 MDCT系数生成器（105）通过将滤波器系数乘以当前语音帧的MDCT系数来生成新的MDCT系数。
-
公开(公告)号：KR1020090060100A
公开(公告)日：2009-06-11
申请号：KR1020080008590
申请日：2008-01-28
Applicant: 한국전자통신연구원
IPC: G10L19/02 , G10L19/032
CPC classification number: G10L21/0208 , G10L19/24 , G10L21/0316 , G10L21/038 , G11B20/10527 , G11B2020/10555 , H03G3/341
Abstract: An apparatus for enhancing quality of a voice codec and a method therefor are provided to reduce noise due to quantization error of a combination interval when coding voice, thereby entirely increasing hearing quality. An apparatus for enhancing quality of a voice codec comprises the first energy calculator(300) and a scaling unit(320). The first energy calculator obtains the first energy of a signal decoded through a core codec. The scaling unit scales a size of the decoded signal when the first energy is smaller than the first threshold value.
Abstract translation: 提供一种用于提高语音编解码器的质量的装置及其方法，用于在语音编码时减少由于组合间隔的量化误差引起的噪声，由此完全提高了听觉质量。 一种用于提高语音编解码器质量的装置包括第一能量计算器（300）和缩放单元（320）。 第一能量计算器获得通过核心编解码器解码的信号的第一能量。 当第一能量小于第一阈值时，缩放单元缩放解码信号的大小。
-
-
-
-
-
-
-
-
-
Search Results
219
records
(took 0.034 seconds)
