WO2007140724A1 - procédé et appareil pour transmettre et recevoir un bruit de fond et système de compression de silence - Google Patents

procédé et appareil pour transmettre et recevoir un bruit de fond et système de compression de silence Download PDF

Info

Publication number
WO2007140724A1
WO2007140724A1 PCT/CN2007/070063 CN2007070063W WO2007140724A1 WO 2007140724 A1 WO2007140724 A1 WO 2007140724A1 CN 2007070063 W CN2007070063 W CN 2007070063W WO 2007140724 A1 WO2007140724 A1 WO 2007140724A1
Authority
WO
WIPO (PCT)
Prior art keywords
noise
background noise
different
sid frame
current
Prior art date
Application number
PCT/CN2007/070063
Other languages
English (en)
French (fr)
Inventor
Libin Zhang
Jun Zhang
Tong Jin
Original Assignee
Huawei Technologies Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Publication of WO2007140724A1 publication Critical patent/WO2007140724A1/zh

Links

Classifications

    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/012Comfort noise or silence coding

Definitions

  • the present invention relates to communication technologies, and more particularly to a method and apparatus for transmitting and receiving background noise and a silent compression system. Background of the invention
  • the transmission of voice signals is not continuous.
  • human voices are not continuous, so the transmitted voice signals can be divided into voiced signals and unvoiced signals.
  • the audible signal refers to various audio signals such as speaking and vocalization.
  • the silent signal refers to the signal generated in the vocalization gap, including various background noises such as white noise, background noise and mute.
  • the audible signal is a useful signal and is the carrier of the communication content
  • the unvoiced signal that is, the background noise is an unwanted signal, and does not contain the main information of the communication.
  • the audible signal is a useful signal and is the carrier of the communication content
  • the unvoiced signal that is, the background noise is an unwanted signal, and does not contain the main information of the communication.
  • people only pay attention to useful audible signals. Considering the pressure of bandwidth, it is not desirable to transmit unwanted background noise, which can effectively reduce the transmission bandwidth. However, if only the transmitted signal does not transmit background noise, it will cause the background noise to be discontinuous, which will make the listener feel very uncomfortable. In the case of strong background noise, the feeling will be more obvious, sometimes making the speech difficult. understanding. Therefore, at present, the usual processing method is: using a silent compression technique in the codec to compress and transmit the background noise, thereby ensuring the transmission quality while ensuring the coding quality.
  • FIG. 1 is a schematic structural diagram of a system for implementing silent compression in the prior art.
  • 2 is a flow chart of a prior art transmission and reception of voice signals.
  • the process of transmitting background noise by using a silent compression technique includes the following steps: Step 201: At the transmitting end, for any frame of the signal, the voice detector (VAD) ) analyzing and detecting the signal, detecting whether the signal is an audible signal or a background noise, if If there is an audible signal, go to step 202. If it is background noise, go to step 203.
  • Step 202 The current signal is used as a voice frame, and the voice frame is encoded by the voice encoder, and then sent to the receiving end, and step 204 is performed.
  • VAD voice detector
  • Step 203 The current signal is used as a non-speech frame, and the non-speech frame is encoded by a non-speech encoder (DTX) to generate a silence description (SID) frame describing the background noise, and the generated SID frame is sent to the receiving end.
  • DTX non-speech encoder
  • SID silence description
  • Step 204 At the receiving end, the received speech frame is decoded by the voice decoder to obtain a voice signal, that is, a voice signal, and is decoded by the non-speech decoder (CNG) according to the feature parameters carried in the received SID frame. Synthetic background noise.
  • CNG non-speech decoder
  • the speech signal can be reconstructed using the decoded speech signal and background noise.
  • the principle of synthesizing the comfortable background noise is the same as that of the speech synthesis.
  • the principle of speech synthesis is:
  • the excitation signal of the background noise can be regarded as a simple random noise sequence, and these sequences can be generated by the random noise generating module at the codec end, and then these sequences are controlled by the energy parameters.
  • the magnitude of the excitation signal can be generated, so that when synthesizing comfortable background noise, the excitation parameters can be simply represented by energy parameters without further description of other characteristic parameters.
  • the characteristic parameters carried in the SID frame include the synthesis filter.
  • Parameters and excitation parameters, wherein the excitation parameters are The energy parameter of the current background noise frame, the synthesis filter parameter in the SID frame is the same as the speech signal, and is the line spectrum frequency LSF quantization parameter.
  • the bit allocation in the SID frame is as shown in Table 1.
  • the speech encoder G.729B adopts a single SID frame description manner when transmitting background noise. That is to say, in the voice communication system, regardless of the characteristics of the current background noise, each encoder uses only one description and quantization method to represent the characteristics of the current background noise, so that regardless of whether the characteristics of the background noise are smooth, The structure of the SID frame of the background noise is fixed, and the feature parameters in the SID frame and the number of bits quantized are identical.
  • the encoder faces a variety of occasions, and the background noise characteristics are different for each occasion.
  • the background noise is relatively simple, generally simple white noise and quiet, and the noise characteristics are relatively stable; in some noisy occasions, such as on the street, the background noise is more complicated.
  • the noise characteristics such as car sound and human voice are very unstable. Therefore, the prior art does not distinguish the characteristics of the background noise at the time of encoding, and the characteristics of all background noises are described by using a SID frame, which lacks pertinence and cannot accurately express the characteristics of various background noises.
  • the synthesized comfortable background noise cannot be well matched with the background noise characteristic of the encoding end, so that the synthesized speech quality is better. Poor, resulting in a very uncomfortable feeling of the human ear, thereby greatly reducing the quality of service of voice communication.
  • the characteristic parameters carried in the SID frame are only various quantization of the line spectrum frequency LSF. Parameters and energy parameters, but no other characteristic parameters.
  • the noise characteristics are more complex, and there are many specific noise characteristics. Only the various quantization parameters and energy parameters of the line spectrum frequency LSF are used in the SID frame, and the background noise is often not fully expressed.
  • the synthesized comfortable background noise can not completely match the background noise characteristic of the encoding end, so the quality of the synthesized speech is further deteriorated, resulting in the feeling of the human ear. Comfortable, thereby further reducing the quality of service for audio communications.
  • a first object of the embodiments of the present invention is to provide a method for transmitting background noise.
  • a second object of the embodiments of the present invention is to provide a method for receiving background noise.
  • a device for transmitting background noise, a fourth object of the present invention is to provide a device for receiving background noise, and a fifth object of the present invention is to provide a system for transmitting and receiving background noise, which is an embodiment of the present invention.
  • a sixth object is to provide a silent compression system for accurately expressing the characteristics of different background noises through different SID frames.
  • a method for transmitting background noise including:
  • the receiving end receives the current background noise, and determines the background noise category to which the current background noise belongs according to the noise characteristic of the current background noise and the correspondence between the established noise characteristics and the background noise category;
  • the transmitting end constructs a SID frame according to the SID frame structure corresponding to the determined background noise category, and sends the constructed SID frame.
  • a method for receiving background noise comprising: dividing each background noise into different categories according to noise characteristics of each background noise, and setting different noise generation rules corresponding to different background noise categories; receiving SID frames at the receiving end Then, determining a background noise category corresponding to the SID frame, and generating a current background noise according to the determined noise generation rule corresponding to the background noise category.
  • An apparatus for transmitting background noise comprising: a noise classification unit and a plurality of noise coding units, wherein different noise coding units hold different SID frame structures corresponding to different types of background noise, and
  • a noise classification unit configured to receive current background noise, and determine a background noise category to which the current background noise belongs according to a noise characteristic of the current background noise and a correspondence between the stored noise characteristics and a background noise category, and determine the current background noise according to the current The background noise category to which the background noise belongs and the correspondence between the saved background noise category and the noise coding unit, and the current background noise is sent to the corresponding noise coding unit;
  • Any one of the noise coding units is configured to construct a SID frame describing the current background noise according to the saved SID frame structure when the current background noise is received, and send the constructed SID frame.
  • An apparatus for receiving background noise comprising: a noise recognition unit and a plurality of noise decoding units, wherein different noise decoding units hold different noise generation rules corresponding to different background noise categories, and
  • a noise identifying unit configured to receive a current SID frame, determine a background noise category corresponding to the current SID frame, and according to the determined background noise category corresponding to the current SID frame and the correspondence between the saved background noise category and the noise decoding unit, The current SID frame is sent to the corresponding noise decoding unit; Any one of the noise decoding units is configured to generate a current background noise according to the saved noise generation rule and the received SID frame when the SID frame is received.
  • a system for transmitting and receiving background noise comprising: a noise classification unit, a plurality of noise coding units, a noise recognition unit, and a plurality of noise decoding units, wherein different noise coding units hold different types of background noise corresponding to each other Different SID frame structures, different noise decoding units store different noise generation rules corresponding to different background noise categories, and a noise classification unit is configured to receive current background noise, and according to the noise characteristics of the current background noise and the saved noise characteristics and Corresponding relationship of the background noise category, determining the background noise category to which the current background noise belongs, and transmitting the current background noise to the corresponding noise coding unit according to the determined background noise category and the correspondence between the stored background noise category and the noise coding unit ;
  • a noise identifying unit configured to receive a current SID frame, determine a background noise category corresponding to the current SID frame, and determine a background noise category corresponding to the current SID frame and a correspondence between the saved background noise category and the noise decoding unit according to the determined Sending the current SID frame to the corresponding noise decoding unit;
  • Any one of the noise coding units configured to, when receiving the current background noise, construct a SID frame describing the current background noise according to the saved SID frame structure, and send the constructed SID frame;
  • Any one of the noise decoding units is configured to generate a current background noise according to the saved noise generation rule and the received SID frame when the SID frame is received.
  • a silent compression system comprising: a transmitting end and a receiving end, wherein the transmitting end comprises: a voice detector, a voice encoder and a non-speech encoder, wherein the non-speech encoder comprises a noise classifying unit and a plurality of noise encoding units And,
  • the voice detector detects whether the signal frame is a voice frame or a background noise when the signal frame is received. After detecting the voice frame, the voice frame is sent to the voice encoder for encoding, The speech encoder transmits the encoded speech frame, and when the background noise is detected, sends the current background noise to the noise classification unit in the non-speech encoder; the noise classification unit according to the noise characteristics of the current background noise and the saved noise Corresponding relationship between the characteristic and the background noise category, determining the background noise category to which the current background noise belongs, and transmitting the current background noise to the corresponding noise according to the determined background noise category and the correspondence between the stored background noise category and the noise coding unit a coding unit; when receiving the current background noise, constructing a SID frame describing the current background noise according to the saved SID frame structure, and transmitting the constructed SID frame;
  • the receiving end includes a voice decoder and a non-speech decoder, and the non-speech decoder includes a noise identifying unit and a plurality of noise decoding units, and
  • a voice decoder configured to receive a voice frame, and decode the voice frame
  • a noise identifying unit configured to receive a current SID frame, determine a background noise category corresponding to the current SID frame, and according to the determined background noise category corresponding to the current SID frame and the correspondence between the saved background noise category and the noise decoding unit,
  • the current SID frame is sent to the corresponding noise decoding unit; any one of the noise decoding units is configured to generate a current background noise according to the saved noise generation rule and the received SID frame when the SID frame is received.
  • the characteristics of the background noise are distinguished, that is, according to different noise characteristics of different background noises at the transmitting end, SID frames of different structures are respectively used to describe, so that each can be accurately expressed.
  • the characteristics of background noise when the comfortable background noise is synthesized at the decoding end, the synthesized comfortable background noise can be well matched with the background noise characteristic of the encoding end, so that the synthesized speech quality is high, and the human ear feels very comfortable, thereby greatly Improve the quality of business services for voice communications.
  • FIG. 1 is a schematic structural diagram of a system for implementing silent compression in the prior art.
  • FIG. 2 is a flow chart of a prior art transmitting and receiving voice signals.
  • FIG. 3 is a schematic structural diagram of a silent compression system in an embodiment of the present invention.
  • the SID frame In order to make the SID frame accurately describe the characteristics of the background noise, so that the synthesized audio quality is better, it is necessary to break the encoder of the transmitting end in the prior art, and adopt a SID for all non-speech frames, that is, various background noises. Frame limit. That is to say, in an encoder, when describing the background noise characteristics, various structures can be adopted according to different characteristics of background noise, such as simple white noise or mute, or more complicated street noises. The SID frame describes these background noises with different characteristics separately.
  • the embodiment of the present invention provides a method for transmitting background noise, and the core idea is to divide each background noise into different categories according to the noise characteristics of each background noise, and set different background noises corresponding to different categories.
  • SID frame structure when detecting that the current signal frame is background noise, the transmitting end determines the category to which the current background noise belongs; the transmitting end constructs the SID frame according to the SID frame structure corresponding to the category of the current background noise, and transmits the constructed SID frame.
  • the embodiment of the present invention further provides a method for receiving background noise, and the core idea is: dividing each background noise into different categories according to the noise characteristics of each background noise. And setting different background noise categories corresponding to different noise generation rules; when the receiving end receives
  • the background noise category corresponding to the SID frame is determined, and the current background noise is generated according to the determined noise generation rule corresponding to the background noise category.
  • an embodiment of the present invention also provides a silent compression system.
  • Figure 3 is a block diagram showing the structure of a silent compression system in the present invention.
  • the silent compression system proposed by the present invention includes a transmitting end and a receiving end.
  • the transmitting end includes a voice detector (VAD), a voice coder and a non-speech encoder (DTX); and the receiving end includes a voice decoder and a non-speech decoder (CNG).
  • VAD voice detector
  • DTX non-speech encoder
  • CNG non-speech decoder
  • the non-speech encoder (DTX) at the transmitting end is a device for transmitting background noise in the present invention, which internally includes a noise classifying unit and a plurality of noise encoding units; and, at the receiving end, a non-speech decoder (CNG), That is, the apparatus for receiving background noise in the present invention includes a noise identifying unit and a plurality of noise decoding units.
  • CNG non-speech decoder
  • the voice detectors are respectively connected to the noise classifying units in the speech encoder and the non-speech encoder, and the noise classifying units in the non-speech encoder are respectively connected to the respective noise encoding units;
  • the speech encoder is connected to the speech decoder through the communication channel, and each of the noise encoding units is connected to the noise recognition unit through the communication channel, and at the same time, the noise recognition unit is connected to each of the noise decoding units.
  • Step 401 According to the noise characteristics of each background noise, each background noise is divided into different categories, and each noise is established. Correspondence between the characteristics and the background noise category, and the established correspondence is stored in the noise classification unit.
  • the noise characteristic of each background noise may be the magnitude of the correlation between the background noise signals, and the correlation between the background noise signals. Can be obtained using the correlation function. For example, all background noises with correlations between signals greater than 0.5 are classified into the first category, and all background noises with correlations between signals less than 0.5 are classified into the second category. In this way, the established correspondence may be corresponding to the first type of background noise whose correlation is greater than 0.5 in the noise characteristic, and the corresponding second type of background noise having a correlation less than 0.5.
  • the noise characteristic of each background noise may be based on whether or not there is periodicity in each background noise. For example, all background noise with periodic characteristics in the signal is divided into the first category, and all background noises with no periodic characteristics in the signal are divided into the second category.
  • the established correspondence may be that the first type of background noise of the signal having a periodic characteristic in the noise characteristic, and the second type of background noise of the signal having no periodic characteristic in the noise characteristic.
  • the classified categories of background noise may include two or more classes.
  • Step 402 According to the noise characteristics of various background noises, different background noise types are set corresponding to different SID frame structures, and the set different SID frame structures are respectively stored in different noise coding units.
  • the different SID frame structures may be in the following multiple ways:
  • the SID frame structure is different.
  • Mode B Between different SID frames, there are completely different feature parameters. Therefore, the SID frame structure is different.
  • the present invention in order to fully express the characteristics of the background noise, it is also possible to set new characteristic parameters for some specific noise characteristics of the background noise. Therefore, in the above manners B and C, there are different SID frames.
  • the different characteristic parameters may be not only existing different feature parameters, but also a newly set feature parameter included in other SID frames in one SID frame.
  • Step 403 Presetting different noise generation rules corresponding to different background noise categories, and storing the set different noise generation rules in different noise decoding units.
  • Step 404 When a voice signal needs to be sent, at the transmitting end, for a signal of any frame, a voice detector (VAD) analyzes and detects the signal, and detects whether the signal is an audible signal or a background noise, and if it is an audible signal, Then, step 405 is performed. If it is background noise, step 408 is performed.
  • VAD voice detector
  • Step 405 The voice detector transmits the current signal as a voice frame to the voice encoder.
  • Step 406 The speech encoder encodes the received speech frame, and sends the encoded speech frame to the speech decoder of the receiving end through the communication channel.
  • Step 407 At the receiving end, the voice decoder decodes the received voice frame, obtains the original voice signal, and ends the current process.
  • Step 408 The voice detector transmits the current signal as a non-speech frame, that is, background noise, to the non-speech encoder.
  • Step 409 The noise classification unit in the non-speech encoder receives the current background noise, and the root According to the noise characteristics of the current background noise and the corresponding relationship between the stored noise characteristics and the background noise category, the background noise category to which the current background noise belongs is determined.
  • the noise characteristic of the data is the correlation between the signals.
  • the noise characteristic according to the noise classification unit is the correlation between the signals of the current background noise.
  • the noise classification unit can use the correlation function to obtain the correlation between the signals of the current background noise.
  • Step 410 The noise classification unit sends the current background noise to the corresponding noise coding unit according to the determined background noise category and the correspondence between the pre-stored background noise category and the noise coding unit.
  • the determined current background noise belongs to the class 1 and the noise coding unit 1 has a correspondence relationship with the type 1 background noise, that is, the noise coding unit 1 is responsible for coding the type 1 background noise,
  • the current background noise is sent to the noise encoding unit 1.
  • Step 411 The corresponding noise coding unit receives the current background noise, and constructs a SID frame describing the current background noise according to the pre-stored SID frame structure.
  • the process of constructing a SID frame describing the current background noise includes: the corresponding noise coding unit according to the SID frame structure corresponding to the category of the current background noise, that is, the SID frame structure saved in itself, from the current background noise Extracting corresponding feature parameters, and quantifying the extracted feature parameters according to the quantization precision of each feature parameter in the SID frame structure corresponding to the category of the current background noise, that is, the SID frame structure saved by itself, and then using the quantized The number of bits constructs a SID frame.
  • Step 412 The corresponding noise coding unit sends the constructed SID frame describing the current background noise to the noise recognition unit at the receiving end through the communication channel.
  • the transmitting end may send a plurality of SID frames having different frame structures in an equal interval manner or an unequal interval manner or a continuous manner.
  • Step 413 The noise recognition unit in the non-speech decoder receives the SID frame of the current background noise, and determines the background noise category corresponding to the SID frame.
  • Step 414 The noise recognition unit sends the received SID frame to the corresponding noise decoding unit according to the correspondence between the preset background noise category and the noise decoding unit.
  • the corresponding noise decoding unit is a noise decoding unit responsible for decoding the SID frame corresponding to the category to which the current background noise belongs.
  • Step 415 At the receiving end, the corresponding noise decoding unit decodes the received SID frame according to the saved noise generating rule to generate current background noise.
  • the speech signal can be reconstructed using the decoded speech signal and background noise.
  • the voice signal includes not only voice signals when a person is talking, but also other various forms of audio signals. That is to say, the technical solution for transmitting background noise of the present invention can be applied not only when a voice signal is transmitted while a person is talking, but also when transmitting other audio signals.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Description

一种发送和接收背景噪声的方法和装置及静音压缩系统 技术领域
本发明涉及通信技术, 特别是涉及一种发送和接收背景噪声的方法 和装置及静音压缩系统。 发明背景
在各种语音通信中, 语音信号的传输并不是连续的, 比如, 人发声 并不是连续的, 因此所传输的语音信号可以分为有声信号和无声信号。 有声信号指说话、 发声等各种音频信号, 无声信号指在发声间隙产生的 信号, 包括各种背景噪声, 比如白噪声、 背景嘈杂声和静音等。
对通信双方来说, 有声信号为有用信号, 是通信内容的载体, 而无 声信号即背景噪声为无用信号, 不包含通信的主要信息。 在通信时, 人 们关注的只是有用的有声信号, 考虑到带宽的压力, 因此不希望传送无 用的背景噪声, 这样就可以有效的降低传输带宽。 但如果只传有声信号 不传背景噪声, 就会导致背景噪声的不连续, 会使收听的人感觉非常不 舒服, 在背景噪声较强的情况下这种感觉会更明显, 有时会令语音难以 理解。 因此, 目前, 通常的处理方式为: 在编解码器中采用静音压缩技 术, 对背景噪声进行压缩后传输, 从而在节省传输带宽的同时, 较好地 保证编码质量。
图 1是现有技术实现静音压缩的系统结构示意图。 图 2是现有技术 发送和接收语音信号的流程图。 参见图 1和图 2, 现有技术在发送和接 收语音信号时, 采用静音压缩技术传输背景噪声的过程包括以下步骤: 步骤 201 : 在发送端, 对于任意一帧的信号, 语音检测器(VAD ) 对该信号进行分析和检测, 检测该信号为有声信号还是背景噪声, 如果 是有声信号, 则执行步骤 202, 如果是背景噪声, 则执行步骤 203。 步骤 202: 将当前的信号作为语音帧, 由语音编码器对该语音帧进 行编码后, 发送给接收端, 执行步骤 204。
步骤 203: 将当前的信号作为非语音帧, 由非语音编码器 (DTX ) 对该非语音帧进行编码, 产生描述背景噪声的静音描述(SID ) 帧, 将 所产生的 SID帧发送给接收端。
步骤 204: 在接收端, 由语音解码器对接收到的语音帧进行解码, 得到有声信号即语音信号,并由非语音解码器(CNG )根据接收到的 SID 帧中携带的特征参数进行解码, 合成背景噪声。
此后, 在接收端, 则可利用解码后的语音信号以及背景噪声重构语 音信号。
目前, 在接收端对非语音帧进行解码时, 舒适背景噪声的合成原理 与语音合成原理相同。 语音的合成原理是: 语音^ 可以看成是一个激 励信号 激励一个合成滤波器 v(")所产生的输出, 即 = * ν(") , 这样, 合成背景噪声时也可以用这个模型。 所以 SID帧中所传输的描述 背景噪声的特征参数与语音编码帧基本相同, 包括信号合成时的合成滤 波器参数和激励参数。 由于 SID帧描述的是背景噪声特性, 背景噪声的 激励信号可以认为是简单的噪声随机序列, 而这些序列在编解码端均可 以筒单的用随机噪声产生模块产生, 然后用能量参数控制这些序列的幅 度, 就可产生最终的激励信号, 因此在合成舒适背景噪声时, 其中的激 励参数可以简单的用能量参数来表示, 而不需要用其它的一些特征参数 来进一步描述。
根据上述接收端对非语音帧进行解码时,舒适背景噪声的合成原理, 在上述的步骤 203中, 在对非语音帧进行编码产生 SID帧时, 该 SID帧 中携带的特征参数包括合成滤波器参数和激励参数, 其中, 激励参数为 当前背景噪声帧的能量参数, SID帧中的合成滤波器参数与语音信号相 同, 为线谱频率 LSF量化参数。 比如, 以 G.729B语音编码器为例, 其 SID帧中的比特分配如表 1所示。
Figure imgf000005_0001
G.729B中的 SID帧比特分配
但是, 参见图 1和图 2, 在现有技术中, 比如语音编码器 G.729B, 在传输背景噪声时, 采用的是单一 SID帧描述方式。 也就是说, 在语音 通信系统中, 无论当前背景噪声是什么特性, 每个编码器只采用一种描 述和量化方法来表示当前背景噪声的特性, 这样, 无论背景噪声的特性 是否平稳, 描述该背景噪声的 SID帧的结构都是固定不变的, 在 SID帧 中的特征参数及对其量化的比特数完全相同。
然而, 在实际的语音通信中, 编码器面对的是各种各样的场合, 每 种场合的背景噪声特性均不一样。 在一些较为安静的场合, 比如在办公 室, 背景噪声就较为筒单, 一般为简单的白噪声和静音, 其噪声特性较 为平稳; 而在一些比较嘈杂的场合, 比如在街头, 背景噪声就较为复杂, 如车声和人声等嘈杂声, 其噪声特性很不平稳。 因此, 现有技术在编码 时不区分背景噪声的特性, 对所有背景噪声的特性均釆用一种 SID帧来 描述的方式, 则缺乏针对性, 无法准确地表达出各种背景噪声的特性。 相应的, 在解码端合成舒适背景噪声时, 所合成的舒适背景噪声就不能 与编码端的背景噪声特性很好的匹配, 因此会造成所合成的语音质量较 差, 导致人耳的感觉很不舒服, 从而大大降低了语音通信的业务服务质 量。
另外, 在现有技术中, 在发送端描述背景噪声的特性时, 只考虑了 合成滤波器参数和激励参数, 也就是说, SID帧中所携带的特征参数只 有线谱频率 LSF的各种量化参数和能量参数, 而没有其他的特征参数。 然而, 对于一些较为复杂的背景噪声, 其噪声特性较为复杂, 存在很多 特定的噪声特性,仅仅在 SID帧中使用线谱频率 LSF的各种量化参数和 能量参数, 往往无法完全表达背景噪声的所有特性, 相应的, 在解码端 合成舒适背景噪声时, 所合成的舒适背景噪声也就不能与编码端的背景 噪声特性完全匹配, 因此会进一步造成所合成的语音质量差, 导致人耳 的感觉艮不舒服, 从而进一步降低了音频通信的业务服务质量。 发明内容
有鉴于此, 本发明实施例的第一目的在于提供一种发送背景噪声的 方法, 本发明实施例的第二目的在于提供一种接收背景噪声的方法, 本 发明实施例的第三目的在于提供一种发送背景噪声的装置, 本发明实施 例的第四目的在于提供一种接收背景噪声的装置, 本发明实施例的第五 目的在于提供一种发送和接收背景噪声的系统, 本发明实施例的第六目 的在于提供一种静音压缩系统, 以便通过不同 SID帧准确地表达出不同 背景噪声的特性。
为了达到上述目的, 本发明实施例的技术方案是这样实现的: 一种发送背景噪声的方法, 包括:
A、 发送端接收到当前背景噪声, 根据当前背景噪声的噪声特性以 及所建立的噪声特性与背景噪声类别的对应关系, 确定当前背景噪声所 属的背景噪声类别; B、发送端根据所确定背景噪声类别对应的 SID帧结构,构造 SID帧, 并发送所构造出的 SID帧。
一种接收背景噪声的方法, 该方法包括: 根据各背景噪声的噪声特 性, 将各背景噪声划分为不同的类别, 并设置不同背景噪声类别对应不 同的噪声产生规则; 当接收端接收到 SID帧后, 确定该 SID帧对应的背景 噪声类别, 并根据所确定的背景噪声类别对应的噪声产生规则, 产生当 前背景噪声。
一种发送背景噪声的装置, 该装置包括: 噪声分类单元和多个噪声 编码单元, 其中, 不同噪声编码单元中保存不同类别的背景噪声所对应 的不同 SID帧结构, 并且,
噪声分类单元, 用于接收当前背景噪声, 并根据当前背景噪声的噪 声特性以及所保存的噪声特性与背景噪声类别的对应关系, 确定当前背 景噪声所属的背景噪声类别, 才艮据所确定的当前背景噪声所属的背景噪 声类别以及所保存的背景噪声类别与噪声编码单元的对应关系, 将当前 背景噪声发送至对应的噪声编码单元;
任意一个噪声编码单元, 用于在接收到当前背景噪声时, 根据所保 存的 SID帧结构, 构造描述当前背景噪声的 SID帧, 并发送所构造的 SID 帧。
一种接收背景噪声的装置, 该装置包括: 噪声识别单元和多个噪声 解码单元, 其中, 不同噪声解码单元中保存不同背景噪声类别对应的不 同噪声产生规则, 并且,
噪声识别单元, 用于接收当前的 SID帧, 确定当前 SID帧对应的背景 噪声类别, 并根据确定的当前 SID帧对应的背景噪声类别以及所保存的 背景噪声类别与噪声解码单元的对应关系, 将当前 SID帧发送至对应的 噪声解码单元; 任意一个噪声解码单元, 用于在接收到 SID帧时, 根据所保存的噪 声产生规则以及所接收到的 SID帧, 产生当前背景噪声。
一种发送和接收背景噪声的系统, 该系统包括: 噪声分类单元、 多 个噪声编码单元、 噪声识别单元和多个噪声解码单元, 其中, 不同噪声 编码单元中保存不同类别的背景噪声所对应的不同 SID帧结构, 不同噪 声解码单元中保存不同背景噪声类别对应的不同噪声产生规则, 并且, 噪声分类单元, 用于接收当前背景噪声, 并根据当前背景噪声的噪 声特性以及所保存的噪声特性与背景噪声类别的对应关系, 确定当前背 景噪声所属的背景噪声类别, 根据所确定的背景噪声类别及所保存的背 景噪声类别与噪声编码单元的对应关系, 将当前背景噪声发送至对应的 噪声编码单元;
噪声识别单元, 用于接收当前的 SID帧, 确定当前 SID帧对应的背 景噪声类别, 并才艮据确定的当前 SID帧对应的背景噪声类别以及所保存 的背景噪声类别与噪声解码单元的对应关系, 将当前 SID帧发送至对应 的噪声解码单元;
任意一个噪声编码单元, 用于在接收到当前背景噪声时, 根据所保 存的 SID帧结构, 构造描述当前背景噪声的 SID帧, 并发送所构造的 SID 帧;
任意一个噪声解码单元, 用于在接收到 SID帧时, 根据所保存的噪 声产生规则以及所接收到的 SID帧, 产生当前背景噪声。
一种静音压缩系统, 该系统包括: 发送端和接收端, 其中, 发送端包括, 语音检测器、 语音编码器和非语音编码器, 非语音编 码器中包括噪声分类单元和多个噪声编码单元, 并且,
语音检测器, 在接收到信号帧时检测该信号帧为语音帧还是背景噪 声, 在检测出为语音帧后, 将该语音帧发送至语音编码器进行编码, 由 语音编码器发送编码后的语音帧, 在检测出为背景噪声时, 将当前的背 景噪声发送至非语音编码器中的噪声分类单元; 噪声分类单元根据当前 背景噪声的噪声特性以及所保存的噪声特性与背景噪声类别的对应关 系, 确定当前背景噪声所属的背景噪声类别, 根据所确定的背景噪声类 别及所保存的背景噪声类别与噪声编码单元的对应关系, 将当前背景噪 声发送至对应的噪声编码单元; 任意一个噪声编码单元在接收到当前背 景噪声时, 根据所保存的 SID帧结构, 构造描述当前背景噪声的 SID帧, 并发送所构造的 SID帧;
接收端包括, 语音解码器以及非语音解码器, 非语音解码器中包括 噪声识别单元和多个噪声解码单元, 并且,
语音解码器, 用于接收语音帧, 对该语音帧进行解码;
噪声识别单元, 用于接收当前的 SID帧, 确定当前 SID帧对应的背景 噪声类别, 并根据确定的当前 SID帧对应的背景噪声类别以及所保存的 背景噪声类别与噪声解码单元的对应关系 , 将当前 SID帧发送至对应的 噪声解码单元; 任意一个噪声解码单元, 用于在接收到 SID帧时, 根据 所保存的噪声产生规则以及所接收到的 SID帧, 产生当前背景噪声。
由此可见, 在本发明实施例中, 区分了背景噪声的特性, 即在发送 端才 据不同背景噪声的不同噪声特性, 采用不同结构的 SID帧来分别描 述, 因此, 能够准确地表达出各种背景噪声的特性。 相应的, 在解码端 合成舒适背景噪声时, 所合成的舒适背景噪声就能够与编码端的背景噪 声特性很好的匹配,使得所合成的语音质量较高,人耳的感觉也很舒服, 从而大大提高了语音通信的业务服务质量。
另外, 在本发明实施例中, 不仅在 SID帧中使用线谱频率 LSF的各种 量化参数和能量参数, 而且还可以针对一些背景噪声特定的噪声特性, 设置了新的特征参数来表示这些特定的噪声特性, 因此, 能够完全表达 背景噪声的所有特性, 相应的, 在解码端合成舒适背景噪声时, 所合成 的舒适背景噪声则能与编码端的背景噪声特性完全匹配, 因此进一步提 高了所合成的语音质量, 从而进一步提高了音频通信的业务服务质量。 附图简要说明
图 1是现有技术实现静音压缩的系统结构示意图。
图 2是现有技术发送和接收语音信号的流程图。
图 3是在本发明实施例中静音压缩系统的结构示意图。
图 4是在本发明实施例中发送和接收背景噪声的流程图。 实施本发明的方式
为了使 SID帧准确的描述出背景噪声的特性, 从而使合成的音频质 量较好, 则必须突破现有技术中发送端的编码器对于所有的非语音帧即 各种背景噪声, 均采用一种 SID帧的限制。也就是说,在一个编码器内, 在对背景噪声特性进行描述时, 可以根据背景噪声的不同特性, 比如是 简单的白噪声或静音, 还是较为复杂的街头嘈杂声等, 而采用多种结构 的 SID帧对这些具有不同特性的背景噪声分别进行描述。
因此, 本发明实施例提出了一种发送背景噪声的方法, 其核心思想 是:预先根据各背景噪声的噪声特性,将各背景噪声划分为不同的类别, 并设置不同类别的背景噪声对应不同的 SID帧结构; 发送端在检测出当 前的信号帧为背景噪声时, 确定当前背景噪声所属的类别; 发送端根据 当前背景噪声所属类别对应的 SID帧结构, 构造 SID帧, 并发送所构造 出的 SID帧。
对应的, 本发明实施例还提出了一种接收背景噪声的方法, 其核心 思想是:根据各背景噪声的噪声特性,将各背景噪声划分为不同的类别, 并设置不同背景噪声类别对应不同的噪声产生规则; 当接收端接收到
SID帧后, 确定该 SID帧对应的背景噪声类别, 并根据所确定的背景噪 声类别对应的噪声产生规则, 产生当前背景噪声。
对应的, 本发明实施例还提出了一种静音压缩系统。 图 3是在本发 明中静音压缩系统的结构示意图。 参见图 3, 本发明所提出的静音压缩 系统包括发送端和接收端。 其中, 发送端包括, 语音检测器 (VAD )、 语音编码器和非语音编码器 (DTX ); 接收端包括, 语音解码器以及非 语音解码器(CNG )。 其中, 发送端的非语音编码器(DTX ), 即为本发 明中实现发送背景噪声的装置, 其内部包括噪声分类单元和多个噪声编 码单元; 并且, 在接收端的非语音解码器 (CNG ), 即为本发明中实现 接收背景噪声的装置, 其内部包括噪声识别单元和多个噪声解码单元。
参见图 3, 在本发明的静音压缩系统中, 语音检测器分别与语音编 码器和非语音编码器中的噪声分类单元相连, 非语音编码器中的噪声分 类单元分别与各个噪声编码单元相连; 并且, 语音编码器通过通信信道 与语音解码器相连, 各个噪声编码单元通过通信信道与噪声识别单元相 连, 同时, 噪声识别单元与各个噪声解码单元相连。
为使本发明的目的、 技术方案和优点更加清楚, 下面结合附图及具 体实施例对本发明作进一步地详细描述。
图 4是在本发明实施例中发送和接收背景噪声的流程图。 参见图 3 和图 4, 在本发明中, 实现发送和接收背景噪声的过程包括以下步骤: 步骤 401 : 预先根据各背景噪声的噪声特性, 将各背景噪声划分为 不同的类别, 并建立各噪声特性与背景噪声类别的对应关系, 将所建立 的对应关系保存在噪声分类单元中。
这里, 在划分背景噪声类别时, 所根据的各背景噪声的噪声特性可 以是各背景噪声信号间相关性的大小, 各背景噪声信号间相关性的大小 可以利用相关函数来得到。 比如, 将信号间相关性大于 0.5的所有背景 噪声划分为第一类, 将信号间相关性小于 0.5的所有背景噪声划分为第 二类。 这样, 所建立的对应关系可以为, 噪声特性中相关性大于 0.5的 对应第一类背景噪声, 相关性小于 0.5的对应第二类背景噪声。
或者, 在划分背景噪声类别时, 所根据的各背景噪声的噪声特性还 可以是各背景噪声中有无周期性特性。 比如, 将信号中有周期性特性的 所有背景噪声划分为第一类, 将信号中没有周期性特性的所有背景噪声 间划分为第二类。 这样, 所建立的对应关系可以为, 噪声特性中信号有 周期性特性的对应第一类背景噪声, 噪声特性中信号没有周期性特性的 对应第二类背景噪声。
在本步骤中, 所划分的背景噪声的类别可以包括两类或更多类。 步骤 402: 预先根据各类背景噪声的噪声特性, 设置不同的背景噪 声类别对应不同的 SID帧结构,将所设置的不同 SID帧结构分别保存在 不同的噪声编码单元中。
这里, 所述不同的 SID帧结构可以为以下的多种方式:
方式 A、 不同 SID帧之间, 具有完全相同的特征参数, 但一个或多 个特征参数的量化精度不同。
这里, 在每一类别背景噪声所对应的 SID帧中, 虽然采用相同的背 景噪声特征参数, 但这些相同背景噪声特征参数中一个或多个的量化精 度不同, 因此, SID帧结构不同。
方式 B、 不同 SID帧之间, 具有完全不同的特征参数, 因此, SID 帧结构不同。
方式 C、不同 SID帧之间,具有相同的特征参数和不同的特征参数, 因此, SID帧结构不同。
在方式 C中, 在不同 SID帧之间, 一个或多个相同特征参数的量化 精度可以相同或不同。
在本发明中, 为了完整地表达出背景噪声的特性, 也可以为背景噪 声的一些特定的噪声特性设置新的特征参数, 因此, 在上述方式 B和方 式 C中, 不同 SID帧之间所具有的不同特征参数, 不仅可以是已有的不 同特征参数, 也可以是一个 SID帧中包括其他 SID帧中所没有的、 新设 置的特征参数。
步骤 403: 预先设置不同背景噪声类别对应不同的噪声产生规则, 将所设置的不同噪声产生规则分别保存在不同的噪声解码单元中。
这里, 因为在上述步骤 402中, 针对不同的背景噪声类别设置了不 同的 SID帧结构, 因此, 在本步骤中, 设置不同背景噪声类别对应不同 的噪声产生规则,从而使得在后续过程中,能够针对不同背景噪声的 SID 帧结构, 准确地解码出背景噪声。
步骤 404: 当需要发送语音信号时, 在发送端, 对于任意一帧的信 号, 语音检测器(VAD )对该信号进行分析和检测, 检测该信号为有声 信号还是背景噪声, 如果是有声信号, 则执行步骤 405, 如果是背景噪 声, 则执行步骤 408。
步骤 405: 语音检测器将当前的信号作为语音帧, 传输给语音编码 器。
步骤 406: 语音编码器对接收到的语音帧进行编码, 并通过通信信 道将编码后的语音帧发送给接收端的语音解码器。
步骤 407: 在接收端, 语音解码器对接收到的语音帧进行解码, 获 取原始的有声信号, 结束当前流程。
步骤 408: 语音检测器将当前的信号作为非语音帧, 即背景噪声, 传输给非语音编码器。
步骤 409: 非语音编码器中的噪声分类单元接收当前背景噪声, 根 据当前背景噪声的噪声特性以及所保存的各噪声特性与背景噪声类别 的对应关系, 确定当前背景噪声所属的背景噪声类别。
这里, 比如, 在步骤 401中, 所才艮据的噪声特性为信号间的相关性 大小, 那么, 在本步骤中, 噪声分类单元所根据的噪声特性是当前背景 噪声的信号间相关性大小。 其中, 噪声分类单元可以利用相关函数来得 到当前背景噪声的信号间相关性大小。
步骤 410: 噪声分类单元根据所确定的背景噪声类别以及预先保存 的背景噪声类别与噪声编码单元之间的对应关系, 将当前背景噪声发送 至对应的噪声编码单元。
这里, 比如, 所确定的当前背景噪声所属的类别为 1类, 而噪声编 码单元 1与 1类背景噪声存在对应关系, 即应该由噪声编码单元 1负责 对 1类背景噪声进行编码, 则在本步骤中, 将当前背景噪声发送至噪声 编码单元 1。
步骤 411 : 该对应的噪声编码单元接收到当前背景噪声, 根据预先 保存的 SID帧结构, 构造描述当前背景噪声的 SID帧。
这里, 所述构造描述当前背景噪声的 SID帧的过程包括: 该对应的 噪声编码单元才艮据当前背景噪声所属类别对应的 SID帧结构, 即自身中 保存的 SID帧结构, 从当前背景噪声中提取对应的特征参数, 并根据当 前背景噪声所属类别对应的 SID帧结构中,即自身保存的 SID帧结构中, 各特征参数的量化精度, 对所提取的特征参数进行量化, 然后利用量化 后的比特数构造 SID帧。
步骤 412: 该对应的噪声编码单元通过通信信道, 将所构造出的、 描述当前背景噪声的 SID帧发送至接收端的噪声识别单元。
这里, 发送端可以采用等间隔方式或不等间隔方式或连续方式发送 多个具有不同帧结构的 SID帧。 步骤 413:非语音解码器中的噪声识别单元接收当前背景噪声的 SID 帧, 并确定该 SID帧对应的背景噪声类别。
步骤 414: 噪声识别单元根据预先保存的背景噪声类别与噪声解码 单元的对应关系, 将接收到的 SID帧发送至对应的噪声解码单元。
这里, 所述的对应噪声解码单元为负责对当前背景噪声所属类别对 应的 SID帧进行解码的噪声解码单元。
步骤 415: 在接收端, 该对应噪声解码单元根据所保存的噪声产生 规则对所接收到的 SID帧进行解码, 产生当前背景噪声。
此后, 在接收端, 则可利用解码后的语音信号以及背景噪声重构语 音信号。
需要说明的是, 在本发明中, 所述的语音信号不仅包括人通话时的 话音信号, 而且还包括其他各种形式的音频信号。 也就是说, 本发明传 输背景噪声的技术方案, 不仅可以在人通话时传输话音信号时应用, 而 且还可以在传输其他音频信号时应用。
总之, 以上所述仅为本发明的较佳实施例而已, 并非用于限定本发 明的保护范围。 凡在本发明的精神和原则之内, 所作的任何修改、 等同 替换、 改进等, 均应包含在本发明的保护范围之内。

Claims

权利要求书
1、 一种发送背景噪声的方法, 其特征在于, 该方法还包括:
A、 发送端接收到当前背景噪声, 根据当前背景噪声的噪声特性以 及所建立的噪声特性与背景噪声类别的对应关系, 确定当前背景噪声所 属的背景噪声类别;
B、 发送端根据所确定背景噪声类别对应的 SID帧结构, 构造 SID 帧, 并发送所构造出的 SID帧。
2、 根据权利要求 1所述的方法, 其特征在于, 在步骤 A前, 该方 法进一步包括: 根据各背景噪声的噪声特性, 将各背景噪声划分为不同 的类别, 建立噪声特性与背景噪声类别的对应关系, 并设置不同的背景 噪声类别对应不同的静音描述 SID帧结构。
3、 根据权利要求 1所述的方法, 其特征在于, 所述噪声特性为: 背 景噪声信号间的相关性大小; 或, 背景噪声有无周期性特性。
4、 根据权利要求 1所述的方法, 其特征在于, 所述不同的 SID帧 结构包括: 不同 SID帧之间, 具有完全相同的特征参数, 且一个或多个 特征参数的量化精度不同;
或, 不同 SID帧之间, 具有完全不同的特征参数;
或, 不同 SID帧之间, 具有相同的特征参数和不同的特征参数。
5、 根据权利要求 4所述的方法, 其特征在于, 当所述不同的 SID 帧结构为不同 SID帧之间具有相同的特征参数和不同的特征参数时, 所 述不同的 SID帧结构进一步包括: 一个或多个相同特征参数的量化精度 相同; 或, 一个或多个相同特征参数的量化精度不同。
6、 根据权利要求 4或 5所述的方法, 其特征在于, 当所述不同的 SID帧结构中包括不同 SID帧之间具有不同的特征参数时, 所述不同的 SID帧结构进一步包括: 不同的特征参数中包括新设置的特征参数。
7、 根据权利要求 1所述的方法, 其特征在于, 在步骤 B中, 所述 构造 SID帧的步骤包括:发送端根据所确定背景噪声类别对应的 SID帧 结构, 从当前背景噪声中提取对应的特征参数, 并根据所确定背景噪声 类别对应的 SID帧结构中各特征参数的量化精度, 对所提取的特征参数 进行量化, 然后利用量化后的比特数构造 SID帧。
8、 根据权利要求 1所述的方法, 其特征在于, 该方法进一步包括: 发送端采用等间隔方式或不等间隔方式或连续方式发送多个具有不同 帧结构的 SID帧。
9、 根据权利要求 1所述的方法, 其特征在于, 该方法进一步包括: 设置不同背景噪声类别对应不同的噪声产生规则; 当接收端接收到 SID 帧后, 确定该 SID帧对应的背景噪声类别, 并才艮据所确定背景噪声类别 对应的噪声产生规则, 产生当前背景噪声。
10、 一种接收背景噪声的方法, 其特征在于, 该方法包括: 根据各 背景噪声的噪声特性, 将各背景噪声划分为不同的类别, 并设置不同背 景噪声类别对应不同的噪声产生规则; 当接收端接收到 SID帧后, 确定 该 SID帧对应的背景噪声类别, 并才艮据所确定的背景噪声类别对应的噪 声产生规则, 产生当前背景噪声。
11、 一种发送背景噪声的装置, 其特征在于, 该装置包括: 噪声分 类单元和多个噪声编码单元, 其中, 不同噪声编码单元中保存不同类别 的背景噪声所对应的不同 SID帧结构, 并且,
噪声分类单元, 用于接收当前背景噪声, 并根据当前背景噪声的噪 声特性以及所保存的噪声特性与背景噪声类别的对应关系, 确定当前背 景噪声所属的背景噪声类别, 根据所确定的当前背景噪声所属的背景噪 声类别以及所保存的背景噪声类别与噪声编码单元的对应关系, 将当前 背景噪声发送至对应的噪声编码单元; 任意一个噪声编码单元, 用于在接收到当前背景噪声时, 根据所保 存的 SID帧结构, 构造描述当前背景噪声的 SID帧, 并发送所构造的 SID帧。
12、 一种接收背景噪声的装置, 其特征在于, 该装置包括: 噪声识 别单元和多个噪声解码单元, 其中, 不同噪声解码单元中保存不同背景 噪声类别对应的不同噪声产生规则, 并且,
噪声识别单元, 用于接收当前的 SID帧, 确定当前 SID帧对应的背 景噪声类别, 并才艮据确定的当前 SID帧对应的背景噪声类别以及所保存 的背景噪声类别与噪声解码单元的对应关系, 将当前 SID帧发送至对应 的噪声解码单元;
任意一个噪声解码单元, 用于在接收到 SID帧时, 根据所保存的噪 声产生规则以及所接收到的 SID帧, 产生当前背景噪声。
13、 一种发送和接收背景噪声的系统, 其特征在于, 该系统包括: 噪声分类单元、多个噪声编码单元、噪声识别单元和多个噪声解码单元, 其中, 不同噪声编码单元中保存不同类别的背景噪声所对应的不同 SID 帧结构, 不同噪声解码单元中保存不同背景噪声类别对应的不同噪声产 生规则, 并且,
噪声分类单元, 用于接收当前背景噪声, 并根据当前背景噪声的噪 声特性以及所保存的噪声特性与背景噪声类别的对应关系, 确定当前背 景噪声所属的背景噪声类别, 根据所确定的背景噪声类别及所保存的背 景噪声类别与噪声编码单元的对应关系, 将当前背景噪声发送至对应的 噪声编码单元;
任意一个噪声编码单元, 用于在接收到当前背景噪声时, 根据所保 存的 SID帧结构, 构造描述当前背景噪声的 SID帧, 并发送所构造的 SID帧; 噪声识别单元, 用于接收当前的 SID帧, 确定当前 SID帧对应的背 景噪声类别, 并根据确定的当前 SID帧对应的背景噪声类别以及所保存 的背景噪声类别与噪声解码单元的对应关系, 将当前 SID帧发送至对应 的噪声解码单元;
任意一个噪声解码单元, 用于在接收到 SID帧时, 根据所保存的噪 声产生规则以及所接收到的 SID帧, 产生当前背景噪声。
14、 一种静音压缩系统, 其特征在于, 该系统包括: 发送端和接收 端, 其中,
发送端包括, 语音检测器、 语音编码器和非语音编码器, 非语音编 码器中包括噪声分类单元和多个噪声编码单元, 并且,
语音检测器, 在接收到信号帧时检测该信号帧为语音帧还是背景噪 声, 在检测出为语音帧后, 将该语音帧发送至语音编码器进行编码, 由 语音编码器发送编码后的语音帧, 在检测出为背景噪声时, 将当前的背 景噪声发送至非语音编码器中的噪声分类单元; 噪声分类单元根据当前 背景噪声的噪声特性以及所保存的噪声特性与背景噪声类别的对应关 系, 确定当前背景噪声所属的背景噪声类别, 根据所确定的背景噪声类 别及所保存的背景噪声类别与噪声编码单元的对应关系, 将当前背景噪 声发送至对应的噪声编码单元; 任意一个噪声编码单元在接收到当前背 景噪声时,根据所保存的 SID帧结构,构造描述当前背景噪声的 SID帧, 并发送所构造的 SID帧;
接收端包括, 语音解码器以及非语音解码器, 非语音解码器中包括 噪声识别单元和多个噪声解码单元, 并且,
语音解码器, 用于接收语音帧, 对该语音帧进行解码;
噪声识别单元, 用于接收当前的 SID帧, 确定当前 SID帧对应的背 景噪声类别, 并才艮据确定的当前 SID帧对应的背景噪声类别以及所保存 的背景噪声类别与噪声解码单元的对应关系, 将当前 SID帧发送至对应 的噪声解码单元; 任意一个噪声解码单元, 用于在接收到 SID帧时, 根 据所保存的噪声产生规则以及所接收到的 SID帧, 产生当前背景噪声。
PCT/CN2007/070063 2006-06-05 2007-05-29 procédé et appareil pour transmettre et recevoir un bruit de fond et système de compression de silence WO2007140724A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN 200610083533 CN101087319B (zh) 2006-06-05 2006-06-05 一种发送和接收背景噪声的方法和装置及静音压缩系统
CN200610083533.6 2006-06-05

Publications (1)

Publication Number Publication Date
WO2007140724A1 true WO2007140724A1 (fr) 2007-12-13

Family

ID=38801072

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2007/070063 WO2007140724A1 (fr) 2006-06-05 2007-05-29 procédé et appareil pour transmettre et recevoir un bruit de fond et système de compression de silence

Country Status (2)

Country Link
CN (1) CN101087319B (zh)
WO (1) WO2007140724A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102395030A (zh) * 2011-11-18 2012-03-28 杭州海康威视数字技术股份有限公司 基于视频压缩码流的运动分析方法、码流转换方法及其装置
KR101248535B1 (ko) * 2008-03-20 2013-04-03 후아웨이 테크놀러지 컴퍼니 리미티드 배경 노이즈 생성 방법 및 노이즈 처리 장치

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101483042B (zh) * 2008-03-20 2011-03-30 华为技术有限公司 一种噪声生成方法以及噪声生成装置
CN101990082B (zh) * 2010-11-03 2014-07-16 中兴通讯股份有限公司 一种实现可视电话的方法及装置
CN103137133B (zh) * 2011-11-29 2017-06-06 南京中兴软件有限责任公司 非激活音信号参数估计方法及舒适噪声产生方法及系统
CN103187065B (zh) * 2011-12-30 2015-12-16 华为技术有限公司 音频数据的处理方法、装置和系统
CN103873714B (zh) * 2012-12-14 2017-12-26 联想(北京)有限公司 通信方法、以及通话发起端设备和通话接收端设备
CN105101109B (zh) * 2014-05-15 2019-12-03 哈尔滨海能达科技有限公司 警用数字集群系统的不连续发送的实现方法、终端和系统
EP2980790A1 (en) * 2014-07-28 2016-02-03 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Apparatus and method for comfort noise generation mode selection
CN104184485B (zh) * 2014-08-08 2016-05-25 北京海格云熙技术有限公司 民航收信机的静音方法和系统
CN105721656B (zh) * 2016-03-17 2018-10-12 北京小米移动软件有限公司 背景噪声生成方法及装置
CN107391498B (zh) * 2017-07-28 2020-10-27 深圳市沃特沃德股份有限公司 语音翻译方法和装置
WO2019019135A1 (zh) * 2017-07-28 2019-01-31 深圳市沃特沃德股份有限公司 语音翻译方法和装置

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1155801A (zh) * 1995-10-13 1997-07-30 法国电信公司 在语音数字传输系统中产生悦耳噪声的方法与装置
CN1398126A (zh) * 2001-07-18 2003-02-19 华为技术有限公司 在通用移动电信系统中实现多种语音编解码的方法
US6577862B1 (en) * 1999-12-23 2003-06-10 Ericsson Inc. System and method for providing comfort noise in a mobile communication network
US6807525B1 (en) * 2000-10-31 2004-10-19 Telogy Networks, Inc. SID frame detection with human auditory perception compensation

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6782361B1 (en) * 1999-06-18 2004-08-24 Mcgill University Method and apparatus for providing background acoustic noise during a discontinued/reduced rate transmission mode of a voice transmission system
DE69931783T2 (de) * 1999-10-18 2007-06-14 Lucent Technologies Inc. Verbesserung bei digitaler Kommunikationseinrichtung

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1155801A (zh) * 1995-10-13 1997-07-30 法国电信公司 在语音数字传输系统中产生悦耳噪声的方法与装置
US6577862B1 (en) * 1999-12-23 2003-06-10 Ericsson Inc. System and method for providing comfort noise in a mobile communication network
US6807525B1 (en) * 2000-10-31 2004-10-19 Telogy Networks, Inc. SID frame detection with human auditory perception compensation
CN1398126A (zh) * 2001-07-18 2003-02-19 华为技术有限公司 在通用移动电信系统中实现多种语音编解码的方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101248535B1 (ko) * 2008-03-20 2013-04-03 후아웨이 테크놀러지 컴퍼니 리미티드 배경 노이즈 생성 방법 및 노이즈 처리 장치
US8494846B2 (en) 2008-03-20 2013-07-23 Huawei Technologies Co., Ltd. Method for generating background noise and noise processing apparatus
CN102395030A (zh) * 2011-11-18 2012-03-28 杭州海康威视数字技术股份有限公司 基于视频压缩码流的运动分析方法、码流转换方法及其装置
US9380309B2 (en) 2011-11-18 2016-06-28 Hangzhou Hikvision Digital Technology Co. Ltd Motion analysis method and code stream conversion method based on video compression code stream and apparatus thereof

Also Published As

Publication number Publication date
CN101087319B (zh) 2012-01-04
CN101087319A (zh) 2007-12-12

Similar Documents

Publication Publication Date Title
WO2007140724A1 (fr) procédé et appareil pour transmettre et recevoir un bruit de fond et système de compression de silence
US6295009B1 (en) Audio signal encoding apparatus and method and decoding apparatus and method which eliminate bit allocation information from the encoded data stream to thereby enable reduction of encoding/decoding delay times without increasing the bit rate
TW563094B (en) Method and apparatus for high performance low bit-rate coding of unvoiced speech
KR101364983B1 (ko) Sid 프레임을 인코딩하기 위한 방법
WO2005081232A1 (ja) 通信装置及び信号符号化/復号化方法
KR20030076646A (ko) 음성 비활동 동안에 보이스 송신 시스템들 사이에상호운용성을 제공하는 방법 및 장치
JP2002530705A (ja) 音声の無声セグメントの低ビットレート符号化
EP2359365B1 (en) Apparatus and method for encoding at least one parameter associated with a signal source
CN103187065A (zh) 音频数据的处理方法、装置和系统
US6424942B1 (en) Methods and arrangements in a telecommunications system
EP1726006A2 (en) Method of comfort noise generation for speech communication
US8949121B2 (en) Method and means for encoding background noise information
WO2008138263A1 (fr) Procédé et dispositif de génération de paramètres de bruit de confort
CN103680509A (zh) 一种语音信号非连续传输及背景噪声生成方法
US7233893B2 (en) Method and apparatus for transmitting wideband speech signals
CN112767955A (zh) 音频编码方法及装置、存储介质、电子设备
CN101170590B (zh) 一种背景噪声的编码码流传输的方法、系统及装置
KR101166650B1 (ko) 배경 잡음 정보를 디코딩하기 위한 방법 및 수단
US7584096B2 (en) Method and apparatus for encoding speech
US20030065512A1 (en) Communication device and a method for transmitting and receiving of natural speech
JP3496618B2 (ja) 複数レートで動作する無音声符号化を含む音声符号化・復号装置及び方法
Nishimura Steganographic band width extension for the AMR codec of low-bit-rate modes
Lin A Synchronization Scheme for Hiding Information in Encoded Bitstream of Inactive Speech Signal.
CN116137151A (zh) 低码率网络连接中提供高质量音频通信的系统和方法
Amro Higher Compression Rates for GSM 6.10 Standard Using Lossless Compression

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 07721686

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 07721686

Country of ref document: EP

Kind code of ref document: A1