公开(公告)号：US20050203998A1

公开(公告)日：2005-09-15

申请号：US10515286

申请日：2003-05-26

申请人： Kimmo Kinnunen ,  Mika Rantaniva ,  Matti Lehtimaki

发明人： Kimmo Kinnunen ,  Mika Rantaniva ,  Matti Lehtimaki

IPC分类号： G01L11/02 ,  G01L15/00 ,  G06F15/16 ,  G10L15/26 ,  G10L25/78 ,  H04B1/46 ,  H04M9/10 ,  H04W4/10 ,  H04W84/08

CPC分类号： G10L25/78 ,  G10L15/26 ,  H04W4/10 ,  H04W76/45

摘要： The invention concerns a method in a digital network system for controlling the transmission of terminal equipments. Terminal equipment includes a PTT (Push-to-Talk) function in order to at least activate the transmission to be carried out to the said network system, and wherein the terminal equipment for voice control of the said PTT function also includes a VOX (Voice Operated transmission) feature, which can be activated/passivated and which is implemented by a VRE (Voice Recognition Engine) function. In the method stops are performed—the VRE function is used to search for an established keyword from an audio signal,—the established keyword is recognised from the audio signal,—a turn to transmit is requested from the network system,—a turn to transmit is received from the network system,—the transmission is connected and the granted turn to transmit is indicated,—the transmission is carried out, and—the transmission is passivated. In the said VOX feature before the said VRE function the audio signal is monitored by a VAD. (Voice Activity Detection) function arranged in connection with terminal equipments, and whereby when activating the said VOX feature in the terminal equipment steps are performed before the said partial steps the terminal equipment's incoming audio signal is processed with the VAD function searching it for a signal form in accordance with an established criterion, and—when a signal form according to the established criterion is detected in the audio signal, the said VRE function is activated to search for an established keyword.

摘要翻译： 本发明涉及用于控制终端设备传输的数字网络系统中的方法。 终端设备包括PTT（一键通）功能，以便至少激活要执​​行的传输到所述网络系统，并且其中用于所述PTT功能的语音控制的终端设备还包括VOX（语音 操作传输）功能，可以激活/钝化，哪些功能由VRE（语音识别引擎）功能实现。 在方法中，执行停止 - VRE功能用于从音频信号中搜索建立的关键字， - 从音频信号识别已建立的关键字， - 从网络系统请求发送，转向 从网络系统接收到发送， - 传输被连接，并且指示授权转向发送 - 传输被执行，并且传输被钝化。 在所述VRE功能之前的所述VOX功能中，音频信号由VAD监视。 （语音活动检测）功能，并且由此当在终端设备中激活所述VOX特征时，在所述部分步骤之前执行终端设备的输入音频信号被处理的VAD功能，以搜索信号 根据已建立的标准形式，并且当在音频信号中检测到根据建立的标准的信号形式时，激活所述VRE功能以搜索建立的关键字。

公开(公告)号：US20050094795A1

公开(公告)日：2005-05-05

申请号：US10697670

申请日：2003-10-29

申请人： Darwin Rambo

发明人： Darwin Rambo

IPC分类号： H04M9/08 ,  H04M9/10 ,  H04M3/42

CPC分类号： H04M9/10 ,  H04M9/08

摘要： Systems and methods that enable high quality audio teleconferencing are disclosed. In one embodiment of the present invention, a signal processor receives signals from a spatially dispersed set of directional microphones, processing the microphone signals and the far-end received audio into a signal for transmission to a far-end party. The processing may comprise the use of one or more algorithms that reduce conference room noise and may selectively increase participant audio levels by processing the microphone signals using beamforming techniques. An embodiment of the present invention may also comprise one or more omni-directional microphones that may be used in cooperation with the directional microphones to adjust for background noise, acoustic echo, and the existence of side conversations.

摘要翻译： 公开了实现高质量音频电话会议的系统和方法。 在本发明的一个实施例中，信号处理器从空间分散的定向麦克风集合接收信号，将麦克风信号和远端接收到的音频处理成用于传输到远端方的信号。 该处理可以包括使用一种或多种减少会议室噪声的算法，并且可以通过使用波束成形技术处理麦克风信号来选择性地增加参与者音频电平。 本发明的实施例还可以包括一个或多个全向麦克风，其可以与定向麦克风配合使用以调整背景噪声，声学回声和侧面对话的存在。

公开(公告)号：US5857019A

公开(公告)日：1999-01-05

申请号：US662569

申请日：1996-06-11

申请人： David Iglehart ,  Leland Lester ,  Joel Q. Adams ,  Geof Findley

发明人： David Iglehart ,  Leland Lester ,  Joel Q. Adams ,  Geof Findley

IPC分类号： H04R3/02 ,  H04B3/20 ,  H04M1/60 ,  H04M9/08 ,  H04M9/10 ,  H04M9/00

CPC分类号： H04M9/10 ,  H04M9/08

摘要： An improved speakerphone that may be viewed as a modification of a conventional half-duplex speakerphone that includes transmit and receive attenuators for alternatively isolating either a speaker or a microphone from a phone line. The attenuators are controlled by a controller that compares a signal generated by the microphone with a signal received on the phone line to determine which of the transmit or receive attenuators should be turned on. In the present invention, a first variable gain amplifier is connected so as to amplify the signal generated by the microphone prior to the signal being connected to the controller, and a second variable gain amplifier is connected so as to amplify the signal received on the phone line prior to the signal being connected to the controller. The first and second variable gain amplifiers have adjustable gain controls that are connected such that the gain of the first variable amplifier is decreased when the gain of the second variable amplifier is increased and vice versa. The adjustable gain control of one of the adjustable gain amplifiers is accessible to the user of the speakerphone and allows the user to adjust the threshold at which his voice will take over the conversation.

摘要翻译： 一种改进的免提电话，其可被视为传统的半双工免提电话的修改，其包括用于将扬声器或麦克风与电话线隔离的发送和接收衰减器。 衰减器由控制器控制，该控制器将由麦克风产生的信号与在电话线上接收到的信号进行比较，以确定哪个发射或接收衰减器应该被接通。 在本发明中，连接第一可变增益放大器，以便在信号连接到控制器之前放大由麦克风产生的信号，并且连接第二可变增益放大器以放大在电话上接收到的信号 信号连接到控制器之前的线路。 第一和第二可变增益放大器具有连接的可调节增益控制，使得当第二可变放大器的增益增加时第一可变放大器的增益减小，反之亦然。 可调节增益放大器之一的可调节增益控制可由扬声器电话的用户访问，并允许用户调整其声音将接管对话的阈值。

公开(公告)号：US5737407A

公开(公告)日：1998-04-07

申请号：US702937

申请日：1996-08-26

申请人： David L. Graumann

发明人： David L. Graumann

IPC分类号： G10L11/02 ,  H04M9/08 ,  H04M9/10 ,  G10L3/00 ,  G10L9/18

CPC分类号： H04M9/08 ,  H04M9/10 ,  G10L25/78

摘要： A method of detecting voice in an audio signal comprises the steps of determining an average peak value representing an envelope of the audio signal, determining a running instance of audio signal standard deviation, which corresponds to one of a number of overlapping time intervals, and updating a power density function (PDF) by adding instances of noise to the PDF if the average peak of the audio signal exceeds the current level of the audio signal by a certain amount and if the current standard deviation value fails below a threshold for a predetermined time interval. A noise floor is located based on the mean value of the PDF, and, if the audio signal sustains a power level exceeding the noise floor, voice activity is determined to be present in the audio signal. The PDF is updated by a low confidence factor if all of the standard deviation values calculated during a certain period of time are below the threshold value and by a high confidence factor if all standard deviation values within a certain longer period of time period are below the threshold value.

摘要翻译： 一种检测音频信号中的语音的方法包括以下步骤：确定表示音频信号的包络的平均峰值，确定音频信号标准偏差的运行实例，其对应于多个重叠时间间隔中的一个，以及更新 功率密度函数（PDF），如果音频信号的平均峰值超过音频信号的当前电平一定量，并且如果当前标准偏差值在预定时间内故障低于阈值，则向PDF添加噪声实例 间隔。 基于PDF的平均值来设置噪声底层，并且如果音频信号保持超过噪声本底的功率电平，则确定语音活动存在于音频信号中。 如果在一定时间段内计算出的所有标准偏差值都低于阈值，则该PDF将被更新，如果在一定时间段内的所有标准偏差值低于 阈值。

公开(公告)号：US5598466A

公开(公告)日：1997-01-28

申请号：US520305

申请日：1995-08-28

申请人： David L. Graumann

发明人： David L. Graumann

IPC分类号： G10L11/02 ,  H04M9/08 ,  H04M9/10 ,  G10L3/00 ,  G10L9/18

CPC分类号： H04M9/08 ,  H04M9/10 ,  G10L25/78

摘要： A method of detecting voice in an audio signal comprises the steps of determining an average peak value representing an envelope of the audio signal, determining a running instance of audio signal standard deviation, which corresponds to one of a number of overlapping time intervals, and updating a power density function (PDF) by adding instances of noise to the PDF if the average peak of the audio signal exceeds the current level of the audio signal by a certain amount and if the current standard deviation value falls below a threshold for a predetermined time interval. A noise floor is located based on the mean value of the PDF, and, if the audio signal sustains a power level exceeding the noise floor, voice activity is determined to be present in the audio signal. The PDF is updated by a low confidence factor if all of the standard deviation values calculated during a certain period of time are below the threshold value and by a high confidence factor if all standard deviation values within a certain longer period of time period are below the threshold value.

摘要翻译： 一种检测音频信号中的语音的方法包括以下步骤：确定表示音频信号的包络的平均峰值，确定音频信号标准偏差的运行实例，其对应于多个重叠时间间隔中的一个，以及更新 功率密度函数（PDF），如果音频信号的平均峰值超过音频信号的当前水平一定量，并且如果当前标准偏差值在预定时间内低于阈值，则向PDF添加噪声实例 间隔。 基于PDF的平均值来设置噪声底层，并且如果音频信号保持超过噪声本底的功率电平，则确定语音活动存在于音频信号中。 如果在一定时间段内计算出的所有标准偏差值都低于阈值，则该PDF将被更新，如果在一定时间段内的所有标准偏差值低于 阈值。

公开(公告)号：US5459786A

公开(公告)日：1995-10-17

申请号：US347106

申请日：1994-11-22

申请人： Yutaka Tomiyori ,  Tadashi Eguchi

发明人： Yutaka Tomiyori ,  Tadashi Eguchi

IPC分类号： H04M1/60 ,  H04M9/08 ,  H04M9/10

CPC分类号： H04M9/10 ,  H04M9/08

摘要： Controlled by a signal level produced by a microphone (12), a noise suppressing circuit (25) supplies a first attenuator (16) with a local level which is equal to the signal level and is suppressed to a predetermined output level when the signal level is lower and higher than a predetermined input level. Cooperating with a drive level produced by a second attenuator (17) to drive a loudspeaker (13), the local level makes a comparator unit (21-23) produce a control signal for making the first attenuator give smaller and greater amounts of attenuation to the signal level with the drive level rendered by the second attenuator low and high when positive and negative values are had, respectively, by an input difference equal to the local level less the drive level. Preferably, the local level is suppressed by a level difference had by the signal level above the predetermined input level.

摘要翻译： 由麦克风（12）产生的信号电平控制，噪声抑制电路（25）为等于信号电平的局部电平提供第一衰减器（16），当信号电平 低于预定输入电平。 与由第二衰减器（17）产生的驱动扬声器（13）的驱动电平相协调，局部电平使得比较器单元（21-23）产生控制信号，以使第一衰减器产生较小和更大的衰减量 当正和负值分别具有等于减小驱动电平的本地电平的输入差时，由第二衰减器呈现的驱动电平的信号电平为低和高。 优选地，通过高于预定输入电平的信号电平来抑制电平差。

公开(公告)号：US5199065A

公开(公告)日：1993-03-30

申请号：US599657

申请日：1990-10-18

申请人： Andreas von Zitzewitz ,  Siegmund Redl

发明人： Andreas von Zitzewitz ,  Siegmund Redl

IPC分类号： H04M9/10

CPC分类号： H04M9/10

摘要： A method for automatic speech direction reversal includes supplying a reception signal with variable damping to a loudspeaker, outputting a microphone signal with variable damping from a microphone as a transmission signal; continuously classifying each of the reception signal and the microphone signal as a speech signal or noise; setting the damping of one of the signals classified as the speech signal to a first damping value, setting the damping of the other of the signals to a second damping value being higher than the first damping value, and maintaining the set damping until the one signal is classified as noise. The preceding damping values are maintained if both of the signals are classified as a speech signal. Both damping vlaues are set to a third damping value being located between the first and second damping values if both of the signals are classified as noise. A transition is performed from one of the first and second damping values to the third damping value more slowly than a transition from the third damping value to one of the first and second damping values, more slowly than a transition from the first to the second damping value, and more slowly than a transition from the second to the first damping value. A configuration for performing the method includes a loudspeaker, a controllable reception attenuator, a microphone, a controllable transmission attenuator, two signal/noise detectors, and a control logic.

摘要翻译： 一种用于自动语音方向反转的方法包括向扬声器提供具有可变阻尼的接收信号，从麦克风输出具有可变阻尼的麦克风信号作为发送信号; 将每个接收信号和麦克风信号连续分类为语音信号或噪声; 将分类为所述语音信号的信号中的一个的阻尼设置为第一阻尼值，将所述另一个信号的阻尼设置为高于所述第一阻尼值的第二阻尼值，并将所述设定阻尼保持直到所述一个信号 被归类为噪音。 如果两个信号被分类为语音信号，则保持前述阻尼值。 如果两个信号被分类为噪声，则两个阻尼阀都被设置为位于第一和第二阻尼值之间的第三阻尼值。 比从第三阻尼值到第一和第二阻尼值之一的转变更缓慢地从第一和第二阻尼值之一到第三阻尼值的转变比从第一阻尼值到第二阻尼值的转变慢得多 值，并且比从第二阻尼值到第一阻尼值的过渡更慢。 用于执行该方法的配置包括扬声器，可控接收衰减器，麦克风，可控传输衰减器，两个信号/噪声检测器和控制逻辑。

公开(公告)号：US4982425A

公开(公告)日：1991-01-01

申请号：US313209

申请日：1989-02-21

申请人： Ikio Yoshida

发明人： Ikio Yoshida

IPC分类号： H04M1/60 ,  H04M9/08 ,  H04M9/10 ,  H04Q7/20 ,  H04Q7/38

CPC分类号： H04M9/08 ,  H04M9/10

摘要： In order to perform voice-switched telephoning, a hands-free telephone includes a speaker, a microphone, a receive variable attenuator (R-ATT), a transmit variable attenuator (T-ATT), a receive signal detector, a transmit signal detector and an attenuation control circuit. The telephone also includes an auxiliary control circuit which prevents the output of the transmit signal detector from reaching the attenuation control circuit during a transient period between a call origination and the beginning of a conversation. During the transient period, only the speaker is enabled to output a ringback tone therethrough. When the output level of the transmit signal detector exceeds a predetermined level, the auxiliary control circuit passes the output of the transmit signal detector to the attenuation control circuit to start the voice-switched telephoning. Once the auxiliary control circuit passes the output of the transmit signal detector, it holds this state until the conversation finishes.

公开(公告)号：US4726062A

公开(公告)日：1988-02-16

申请号：US807596

申请日：1985-12-11

申请人： Donald Martz ,  Robert D. Talbot

发明人： Donald Martz ,  Robert D. Talbot

IPC分类号： H04B3/20 ,  H04M9/10

CPC分类号： H04B3/20 ,  H04M9/10

摘要： A circuit for reducing the effect of acoustic and electrical feedback in duplex "hands free" telephone sets. The circuit uses relatively simple and inexpensive gain control circuits and expanders to reduce the effects of feedback by controlling the gain of a transmitted signal in inverse relationship to the strength of a the received signal.

摘要翻译： 一种用于减少双重“免提”电话机中声和电反馈影响的电路。 该电路使用相对简单和便宜的增益控制电路和扩展器来通过控制与接收信号的强度成反比关系的发送信号的增益来减少反馈的影响。

公开(公告)号：US4492824A

公开(公告)日：1985-01-08

申请号：US522106

申请日：1983-08-11

申请人： Steve W. Bell ,  Milton L. Embree

发明人： Steve W. Bell ,  Milton L. Embree

IPC分类号： H04M1/60 ,  H04M9/10 ,  H04M9/08

CPC分类号： H04M9/10

摘要： There is disclosed a precision linear comparator for a speakerphone circuit having two rectifying constant voltage nodes for separately summing two groups of signals derived from a speakerphone transmit channel and receive channel and supplying at an output terminal a signal representative of the difference between two unblocked summed signals. Each node blocks its summed signal if the signal is not representative of a true voice signal. A null signal is supplied at the output terminal during the absence of unblocked summed signals.

摘要翻译： 公开了一种用于具有两个整流恒压节点的扬声器电话的精密线性比较器，用于分别对从扬声器电话发射信道和接收信道导出的两组信号进行求和，并在输出端提供表示两个未阻塞求和信号之间的差异的信号 。 如果信号不代表真实的语音信号，则每个节点阻塞其求和信号。 在没有阻塞的求和信号的情况下，在输出端提供空信号。