公开(公告)号：WO2011087376A1

公开(公告)日：2011-07-21

申请号：PCT/NO2011/000017

申请日：2011-01-18

Applicant: TANDBERG TELECOM AS ,  BIRKENES, Øystein

Inventor： BIRKENES, Øystein

IPC: H04M1/20 ,  H04M9/08 ,  H04B3/23

CPC classification number: H04M9/085

Abstract: A method for processing an audio signal executed by an audio echo suppression apparatus, the method including: receiving, at the audio echo suppression apparatus, the audio signal; generating a subband signal from the audio signal; delaying, at the audio echo suppression apparatus, the subband signal with a plurality of different delay values to form a plurality of time lag signals; multiplying, at the audio echo suppression apparatus, the plurality of time lag signals with first respective filter coefficients to generate a first signal; calculating, at the audio echo suppression apparatus, a complex product between pairs of the plurality of time lag signals to generate complex product signals; multiplying, at the audio echo suppression apparatus, each of a real part and imaginary part of the complex product signals with second respective filter coefficients, and taking a sum thereof, to generate a second signal; and estimating an echo subband signal from the first signal and the second signal.

Abstract translation: 一种用于处理由音频回声抑制装置执行的音频信号的方法，所述方法包括：在音频回声抑制装置处接收音频信号; 从所述音频信号产生子带信号; 在音频回声抑制装置处延迟具有多个不同延迟值的子带信号以形成多个时滞信号; 在音频回声抑制装置处将具有第一相应滤波器系数的多个时滞信号相乘以产生第一信号; 在所述音频回声抑制装置处计算所述多个时滞信号对之间的复数乘积以产生复乘产品信号; 在音频回声抑制装置处，将具有第二相应滤波器系数的复乘积信号的实部和虚部中的每一者相乘并产生第二信号; 以及从所述第一信号和所述第二信号估计回波子带信号。

公开(公告)号：WO2011081527A1

公开(公告)日：2011-07-07

申请号：PCT/NO2010/000470

申请日：2010-12-17

Applicant: TANDBERG TELECOM AS ,  SOLVANG, Audun

Inventor： SOLVANG, Audun

IPC: G01S3/80 ,  H04R3/00 ,  H04N7/15

CPC classification number: H04R3/005 ,  G01S3/8083 ,  H04N7/142 ,  H04N7/15 ,  H04R1/406 ,  H04R2201/403

Abstract: A method and system for determining a direction between a detection point, e.g. at a camera in a video conference equipment, and an acoustic source, e.g. an active speaker participating in a video conference. The method comprises receiving acoustic signals originating from the acoustic source at a first and second pair of microphone elements, arranged symmetrically about the detection point; calculating a first cross correlation signal from the first pair of microphone elements; and calculating a second cross correlation of signals from the second pair of microphone elements. The direction is then calculated based on both the first and second cross correlation signals, e.g. by convolution. Further symmetrically arranged pairs of microphone elements may also be used.

Abstract translation: 一种用于确定检测点之间的方向的方法和系统。 在视频会议设备中的相机处，以及声源，例如， 参与视频会议的主动演讲者。 该方法包括在第一和第二对麦克风元件处接收源自声源的声信号，围绕检测点对称地布置; 计算来自所述第一对麦克风元件的第一互相关信号; 以及计算来自所述第二对麦克风元件的信号的第二互相关。 然后基于第一和第二互相关信号，例如， 通过卷积 还可以使用对称布置的麦克风元件对。

公开(公告)号：WO2011078682A1

公开(公告)日：2011-06-30

申请号：PCT/NO2010/000448

申请日：2010-12-07

Applicant: TANDBERG TELECOM AS ,  SCHATVET, Petter

Inventor： SCHATVET, Petter

IPC: H04N5/235

CPC classification number: H04N5/235 ,  H04N5/2357

Abstract: The present invention discloses a method providing removal of the distractive illuminative flickering due to discrepancy between line frequency, screen updating rate and camera exposure time in video conferencing and video recording by primarily adjusting the screen updating rate and secondary the camera exposure time to achieve a flickering free experience of the video captured by the camera.

Abstract translation: 本发明公开了一种方法，通过主要调整屏幕更新速率和辅助照相机曝光时间来实现闪烁，提供了由于视频会议和视频记录中的线路频率，屏幕更新率和照相机曝光时间之间的差异而导致的分散照明闪烁的去除 摄像机拍摄的视频的免费体验。

公开(公告)号：WO2011002304A1

公开(公告)日：2011-01-06

申请号：PCT/NO2010/000256

申请日：2010-07-01

Applicant: TANDBERG TELECOM AS ,  HANSEN, Robert ,  WITTY, Paul

Inventor： HANSEN, Robert ,  WITTY, Paul

IPC: H04L12/06 ,  H04L12/56

CPC classification number: H04Q11/0457 ,  H04Q3/62 ,  H04Q2213/1307 ,  H04Q2213/13097 ,  H04Q2213/13102 ,  H04Q2213/13103 ,  H04Q2213/13196 ,  H04Q2213/13204 ,  H04Q2213/1322 ,  H04Q2213/13251 ,  H04Q2213/13348 ,  H04Q2213/13389

Abstract: The present invention introduces a novel system and novel methods for allowing an individual using an IP communication device to contact an individual in an external organization using a single identifier, such as a telephone number, without the caller having any knowledge of the IP capabilities of the called individual. This is utilised by communicating with a registry device, e.g. a Global Address Database (GAD). The GAD contains contact information, such as IP capabilities, for registered numbers. When trying to make a call to a given number, a call server or similar at the site at the caller side, requests contact information relating to that number from the GAD, then sets up a call based on the information received from the GAD.

Abstract translation: 本发明引入了一种新颖的系统和新颖的方法，用于允许使用IP通信设备的个人使用诸如电话号码的单个标识符来联系外部组织中的个体，而呼叫者不知道IP地址的IP能力 叫个人 这通过与注册设备通信来利用，例如。 全球地址数据库（GAD）。 GAD包含注册号码的联系信息，如IP功能。 当呼叫方呼叫一个呼叫服务器或类似呼叫者时，请求从GAD发送与该号码有关的联系信息，然后根据从GAD接收到的信息建立呼叫。

公开(公告)号：WO2010077148A1

公开(公告)日：2010-07-08

申请号：PCT/NO2009/000447

申请日：2009-12-28

Applicant: TANDBERG TELECOM AS ,  ENDRESEN, Lars, Petter

Inventor： ENDRESEN, Lars, Petter

IPC: H04N7/26 ,  H04N7/50

CPC classification number: H04N19/93 ,  H04N19/129 ,  H04N19/132 ,  H04N19/176 ,  H04N19/18 ,  H04N19/60

Abstract: A process for calculating run-and-level representations of quantized transform coefficients includes packing each quantized transform coefficients in a value interval (Max, Min) by setting all quantized transform coefficients greater than Max equal to Max, and all quantized transform coefficients less than Min equal to Min; reordering the quantized transform coefficients resulting in an array C of reordered quantized transform coefficients; masking C by generating an array M containing ones in positions corresponding to positions of C having non-zero values, and zeros in positions corresponding to positions of C having zero values; and for each position containing a one in M, generating a run and a level representation by setting the level value equal to an occurring value in a corresponding position C, and setting the run value equal to the number of proceeding positions relative to a current position in M since a previous occurrence of one in M.

Abstract translation: 用于计算量化变换系数的运行和级别表示的过程包括通过将大于Max的所有量化变换系数设置为等于Max，并将所有量化变换系数小于Min来将每个经量化的变换系数包括在值间隔（Max，Min）中 等于闵; 重新排序量化的变换系数，得到重新排序的量化变换系数的数组C. 通过在对应于具有非零值的C的位置的位置中生成包含一个数组的阵列M来掩蔽C，并且在与具有零值的C的位置对应的位置中为零; 并且对于包含M中的一个的每个位置，通过将等级值设置为等于相应位置C中的出现值来生成游程和水平表示，并且将游程值设置为等于当前位置的进行位置数 在M中，因为先前在M.

公开(公告)号：WO2010071448A1

公开(公告)日：2010-06-24

申请号：PCT/NO2009/000440

申请日：2009-12-17

Applicant: TANDBERG TELECOM AS ,  ENDRESEN, Lars, Petter

Inventor： ENDRESEN, Lars, Petter

IPC: G06T1/60 ,  G09G5/393 ,  H04N5/14

CPC classification number: H04N19/117 ,  H04N19/182 ,  H04N19/59 ,  H04N19/61 ,  H04N19/80

Abstract: The present invention is related to an implementation of a filter process in compression/decompression of digital video systems in multi-purpose processors. It provides a method significantly reducing the number of memory loads/stores and address computations. To achieve known memory behavior within a loop, one filter is implemented for each resolution. Prior to filtering it is determined which format the video content is adapted to, and the width and height of each frame is known before executing the loop, and consequently also the memory address of the pixel values to be filtered. For a given resolution the exact memory location of each pixel in the frame is known in compile-time and need not be calculated on the fly i.e. in each loop of the filter function. Thus loading and storing becomes extremely efficient. This will in particular boost the in-order execution engines that cannot reorder the instructions in the executable on the fly, but it may also lead to dramatic speedup for out-of-order execution processors that thus may find many more independent instructions that can be executed simultaneously.

Abstract translation: 本发明涉及多用途处理器中的数字视频系统的压缩/解压缩中的滤波处理的实现。 它提供了显着减少内存加载/存储和地址计算数量的方法。 为了在循环内实现已知的记忆行为，为每个分辨率实现一个过滤器。 在进行滤波之前，确定视频内容适合的格式，并且在执行循环之前已知每个帧的宽度和高度，并且因此也确定要滤波的像素值的存储器地址。 对于给定的分辨率，帧中每个像素的精确存储器位置在编译时是已知的，并且不需要在飞行中即在滤波器函数的每个循环中计算。 因此，加载和存储变得非常有效。 这将特别提升无法重新排序可执行文件中的指令的顺序执行引擎，但它也可能导致无序执行处理器的显着加速，从而可能会发现更多独立的指令，可以是 同时执行。

公开(公告)号：WO2010068115A1

公开(公告)日：2010-06-17

申请号：PCT/NO2009/000427

申请日：2009-12-11

Applicant: TANDBERG TELECOM AS ,  HAGEN, Einar, Andreas, Hustadnes ,  FUREY, Sean

Inventor： HAGEN, Einar, Andreas, Hustadnes ,  FUREY, Sean

IPC: H04N7/15 ,  H04M3/56 ,  H04L29/06

CPC classification number: H04N7/152 ,  H04L65/1009 ,  H04L65/4015 ,  H04L69/14 ,  H04N7/144

Abstract: The present invention relates to video conferencing systems and telepresence. More specifically, the invention introduces a novel method of setting up communication sessions in a telepresence call comprising of a multiple point-to-point connections between at least two telepresence systems, wherein the information required for setting up the communication sessions is embedded in a control protocol message flow establishing a first communication session between the two telepresence systems.

Abstract translation: 本发明涉及视频会议系统和远程呈现。 更具体地，本发明引入了一种在远程呈现呼叫中建立通信会话的新颖方法，其包括在至少两个远程呈现系统之间的多个点对点连接，其中将建立通信会话所需的信息嵌入到控制 协议消息流建立两个远程呈现系统之间的第一通信会话。

公开(公告)号：WO2010041954A1

公开(公告)日：2010-04-15

申请号：PCT/NO2009/000344

申请日：2009-10-05

Applicant: TANDBERG TELECOM AS ,  HAGEN, Einar

Inventor： HAGEN, Einar

IPC: H04N7/15 ,  H04N5/45

CPC classification number: H04N7/15 ,  H04N7/157

Abstract: A method is disclosed for processing images during a conference between a plurality of video conferencing terminals. The method comprises providing a first image from a first camera at a first video conferencing terminal; providing a second image from a second camera at a second video conferencing terminal; and providing a third image from a third camera at a third video conferencing terminal. Further, the method comprises generating a first composite image, including inserting said third image at a first position into said second image; and generating a second composite image, including inserting said third image at a second position into said first image. The first and second positions are located in horizontally opposite portions of the first and second composite images, respectively. Further, the first and second composite images are supplied to be displayed on a display of the first and second video conferencing terminals, respectively. The method improves telepresence experience when a regular video conferencing terminal is connected to the conference.

Abstract translation: 公开了一种在多个视频会议终端之间的会议期间处理图像的方法。 该方法包括在第一视频会议终端处从第一相机提供第一图像; 在第二视频会议终端处从第二相机提供第二图像; 以及在第三视频会议终端从第三相机提供第三图像。 此外，该方法包括生成第一合成图像，包括将第一图像在第一位置插入到所述第二图像中; 以及生成第二合成图像，包括将所述第三图像在第二位置插入到所述第一图像中。 第一和第二位置分别位于第一和第二合成图像的水平相对的部分。 此外，第一和第二合成图像被提供以分别显示在第一和第二视频会议终端的显示器上。 当常规视频会议终端连接到会议时，该方法提高了远程呈现体验。

公开(公告)号：WO2009082239A1

公开(公告)日：2009-07-02

申请号：PCT/NO2008/000464

申请日：2008-12-19

Applicant: TANDBERG TELECOM AS ,  BJØNTEGAARD, Gisle ,  FULDSETH, Arild

Inventor： BJØNTEGAARD, Gisle ,  FULDSETH, Arild

IPC: H04N7/26 ,  H03M7/40

CPC classification number: H04N19/93 ,  G11B27/10 ,  H03M7/42 ,  H04N7/00 ,  H04N19/13 ,  H04N19/61 ,  H04N19/91

Abstract: The present invention provides a method and a device for lossless coding of event tables with dynamically matching VLC tables. The most probable event should be assigned the shortest code and the respective events should have increasing code length as the associated probability of occurrence decrease. The present invention takes into account that the probability distribution of the event table may not be stable throughout the different parts of a video sequence. Each time an event has occurred, this event is moved one position up in the event table. The present invention results in more efficient coding of digital compressed video by dynamically reordering event tables to obtain a better match between event probabilities and VLC code words. This is particularly useful when coding video with light and color conditions temporarily or constantly differing from the expected conditions from which static VLCs are derived.

Abstract translation: 本发明提供一种用于具有动态匹配VLC表的事件表的无损编码的方法和装置。 最可能的事件应该被分配最短的代码，相应的事件应该具有增加的代码长度，因为相关的发生概率降低。 本发明考虑到视频序列的不同部分的事件表的概率分布可能不稳定。 每次发生事件时，事件表都会向上移动一个位置。 本发明通过动态重新排序事件表来实现数字压缩视频的更有效的编码，以获得事件概率与VLC码字之间更好的匹配。 当使用光和颜色条件暂时或不断地与导出静态VLC的预期条件不同的视频进行编码时，这是特别有用的。

公开(公告)号：WO2009078736A1

公开(公告)日：2009-06-25

申请号：PCT/NO2008/000463

申请日：2008-12-19

Applicant: TANDBERG TELECOM AS ,  HVIDSTEN, Knut, Inge

Inventor： HVIDSTEN, Knut, Inge

IPC: H04N7/26 ,  G06T5/00

CPC classification number: G06T5/003 ,  G06T7/90 ,  G06T2207/10024 ,  H04N19/117 ,  H04N19/132 ,  H04N19/154 ,  H04N19/186 ,  H04N19/44 ,  H04N19/59 ,  H04N19/60 ,  H04N19/61 ,  H04N19/86

Abstract: The present invention is related to video and still-image compression systems. It addresses the problem of the occur-rence blurry edges in images exposed to conventional decod-ing and coding process. The present invention provides a method allowing some 'leakage' of edges and high-frequency content from the full-resolution luma-channel into the low-resolution chroma channels. It is adjusted to operate on parts of a image to be decoded (blocks, lines etc), to in-vestigate the conformity between the available decimated chroma-information with decimated luma-information. If a good fit can be found, i.e. appropriate parameters can be determined to express decimated chroma values with deci-mated luma values, those parameters are applied on the full-resolution luma values for obtaining estimated full-resolution chroma values instead of interpolation. If a good fit cannot be found, the full-resolution chroma values will gradually deteriorate to a standard fallback method interpolation method. This process is repeated for the en-tire picture and for each chroma-channel.

Abstract translation: 本发明涉及视频和静止图像压缩系统。 它解决了暴露于常规解码和编码过程的图像中出现的模糊边缘的问题。 本发明提供了允许边缘和高频内容的一些“泄漏”从全分辨率亮度通道进入低分辨率色度通道的方法。 它被调整为对待解码的图像的部分（块，线等）进行操作，以便调节可用的抽取色度信息与抽取的亮度信息之间的一致性。 如果可以找到良好的拟合，则可以确定适当的参数来表示具有决定亮度值的抽取色度值，将这些参数应用于全分辨率亮度值，以获得估计的全分辨率色度值而不是内插。 如果找不到很好的匹配，则全分辨率色度值将逐渐劣化为标准的回退方法插值方法。 对于轮胎图像和每个色度通道重复该过程。