CN100407292C - Method for converting speech code between differential speech agreement - Google Patents
Method for converting speech code between differential speech agreement Download PDFInfo
- Publication number
- CN100407292C CN100407292C CN031539335A CN03153933A CN100407292C CN 100407292 C CN100407292 C CN 100407292C CN 031539335 A CN031539335 A CN 031539335A CN 03153933 A CN03153933 A CN 03153933A CN 100407292 C CN100407292 C CN 100407292C
- Authority
- CN
- China
- Prior art keywords
- agreement
- parameter
- stream
- code
- target code
- Prior art date
- Legal status (The legal status 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 status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The present invention discloses a method for converting speech codes among different speech agreements, which comprises the steps: A, data bit streams of a source code agreement are decoded according to the agreement, index streams of the source code agreement are obtained after decoding, a vector codebook is searched for the index streams of the corresponding agreement, and parameter streams of the source code agreement are obtained; B, the parameter streams are used as parameter streams of a target code agreement; C, the vector codebook is searched according to the parameter streams of the target code agreement, index streams of the target code agreement are obtained, the index streams of the target code agreement are encoded, and finally, data bit streams of the target code agreement are obtained. The method avoids phonetic damages and reduces the time delay of a communication system, and the high density integration of the communication system is realized.
Description
Technical field
The present invention relates to the encoding and decoding speech technology of voice Internet protocol (VOIP), the conversion method of voice coding between particularly a kind of different voice protocol.
Background technology
Along with the technology based on Internet protocol (IP) constantly develops, what the VOIP technology was used must get more and more.When using the VOIP technology, different protocol networks usually uses different speech coding technologies to carry out the compression transmission of voice, for the interconnection and interflow that realizes networks of different protocol just needs the conversion between the different code/decode format of realization.For example: the network that has can use g.723.1 agreement, and the network that has can use g.729 agreement, and these two networks with different agreement will carry out the conversion between the code/decode format when carrying out intercommunication.
As shown in Figure 1, Fig. 1 is for two kinds of networks of different protocol of prior art carry out the synoptic diagram of encoding and decoding conversion, and what suppose that network 1 uses is agreement g.723.1, and network 2 uses is agreement g.729, when these two networks carry out encoding and decoding conversion, carry out following operation:
At first decode to having g.723.1 the coded data of agreement in the network 1, generate linear data, the coded data decoding back that is about to g.723.1 generates synthetic speech;
To linear data, promptly synthetic speech is encoded then, weaves into to have the g.729 coded data of agreement, and g.729 this have that the coded data of agreement is applicable to network 2.
As shown in Figure 2, Fig. 2 carries out the concrete implementing procedure figure of encoding and decoding conversion for two kinds of networks of different protocol of prior art:
The method clear thinking of present this encoding and decoding conversion, simple, but, on realizing, this method also exists shortcoming: shortcoming one, the time-delay of whole communication system is bigger, since have when linear data carried out encoding and decoding algorithm intrinsic encoding and decoding time-delay, this can strengthen the time delay of communication system; Shortcoming two, whole communication system sound effect sustains damage, g.723.1 protocol code and g.729 protocol code all be the lossy compression method that voice signal is carried out, the operation that this method is decoded and encoded is equivalent to have carried out the squeeze operation that diminishes for twice, these have produced influence to voice quality; Shortcoming three: adopt the method for present this encoding and decoding conversion to carry out a decode operation and an encoding operation to data, this needs bigger calculated amount, take more communication system resource, is unfavorable for that like this high density of communication system is integrated.
Summary of the invention
In view of this, fundamental purpose of the present invention is to provide the conversion method of voice coding between a kind of different voice protocol, to avoid the damage of voice, reduces the time delay of communication system, realizes that the high density of communication system is integrated.
According to above-mentioned purpose, technical scheme of the present invention is achieved in that
The conversion method of voice coding between a kind of different voice protocol, this method comprises:
A, the data bit flow of source code agreement is decoded by its agreement, decoding obtains the index stream of source code agreement, and the index stream of this respective protocol is searched the vector code book, obtains the parameter stream of source code agreement;
B, the parameter that parameter is flowed as the target code agreement flow;
C, with each parameter in target code protocol parameter described in the step B stream as input, search the vector code book of target code agreement, if the parameter in the parameter that finds and the target code protocol parameter stream is identical, then directly this parameter is converted to the index stream parameter of target code agreement according to the vector code book; Otherwise, the vector code book find with target code protocol parameter stream in the parameter of parameter equivalence, the parameter that this is of equal value is converted to the index stream parameter of target code agreement according to the vector code book, index stream to the target code agreement after the conversion is encoded, and finally obtains the data bit flow of target code agreement.
Described equivalent parameter is the square error minimum of the parameter in parameter that the vector code book finds and target code protocol parameter stream.
Include excitation information and filter information in the parameter stream of described source code agreement.
Described source code protocol type is during for agreement g.723.1, and the target code protocol type is g.729 agreement or ARM agreement;
Described source code protocol type is during for agreement g.729, and the target code protocol type is g.723.1 agreement or ARM agreement;
When described source code protocol type was the ARM agreement, the target code protocol type was g.723.1 agreement or g.729 agreement.
Because several coding methods of using all have same or analogous mathematical model at present, it specifically all is the mathematical model of the Code Excited Linear Prediction (celp) of employing, the parameter basically identical that it extracts when encoding, so, method provided by the invention does not need this step of synthetic speech, and directly utilize corresponding to parameter that the raw parameter circulation is changed to purpose parameter stream, and carry out the coding of purpose parameter stream again, realize the conversion between the parameter coding.Method provided by the invention has been omitted this step of synthetic speech, this step of synthetic speech has caused the time delay of communication system, voice are caused damage, the process of handling synthetic speech has taken the resource of communication system, therefore, omitted this step of synthetic speech and avoided the damage of voice, reduced the time delay of communication system, realized that the high density of communication system is integrated.
Description of drawings
Fig. 1 carries out the synoptic diagram of encoding and decoding conversion for two kinds of networks of different protocol of prior art.
Fig. 2 carries out the concrete implementing procedure figure of encoding and decoding conversion for two kinds of networks of different protocol of prior art.
Fig. 3 carries out the process flow diagram of encoding and decoding conversion for two kinds of networks of different protocol of the present invention.
Embodiment
In order to make the purpose, technical solutions and advantages of the present invention clearer, by the following examples and with reference to accompanying drawing, the present invention is described in more detail.
Middle low rate coding at present commonly used, for example g.723.1, what g.729 use with the AMR agreement all is identical mathematical model, can list the coding techniques of Code Excited Linear Prediction (CELP) in.Specifically, these several coded systems all have following characteristics:
Characteristics one all are based on the steady characteristics in short-term of voice signal, voice signal is carried out the branch frame handle, and think that in a frame voice signal can adopt the signal system of lumped parameter to represent;
Characteristics two extract the excitation sample of voice signal according to the present frame voice sample information in cataloged procedure, calculate filter coefficient;
Characteristics three for the ease of the stability that transmits, all will be carried out certain variation to the coefficient of wave filter, are changed to more stable representation;
Characteristics four in order to reduce the information number of transmission, all need the mode by vector quantization, and the parameter that obtains is carried out vector coding.
Because the model of these several codings is basic identical, therefore, can be not when changing mode by synthetic speech, directly carry out the conversion of parameter aspect, as shown in Figure 3, Fig. 3 is the process flow diagram that two kinds of networks of different protocol of the present invention carry out encoding and decoding conversion, and its concrete steps are:
The vector code book is a parameter table of comparisons, promptly this g.723.1 the parameter in the agreement index stream corresponding real speech sample parameter is all arranged in the vector code book, 40HZ as the 20 corresponding real speech sample parametric excitation information of the sequence number in the agreement index stream g.723.1, therefore, find corresponding excitation information and filter information according to the indexing parameter in the index stream of agreement g.723.1;
Because the model of these several coded system structures is similar substantially, then g.723.1 the excitation information of protocol parameter stream and filter information are similar substantially with filter information with the excitation information that protocol parameter g.729 flows, promptly can utilize the excitation information of the stream of protocol parameter g.723.1 and filter information as input, search the g.729 vector code book of agreement, according to the vector code book table of comparisons of agreement g.729, excitation information and filter information in the parameter stream are converted to corresponding indexing parameter, thereby obtain the g.729 index stream of agreement;
G.723.1 the index stream of agreement is converted to the g.729 index stream of agreement, and the two representation is different, but these two kinds of parameter streams have identical or close physical significance;
If excitation information in the parameter stream and filter information do not have in the code book of agreement g.729, then in the vector code book of agreement g.729, seek the value of the most close excitation information and filter information, be the square error minimum of this hithermost value and this excitation information and filter information, in the vector code book of agreement g.729, find corresponding indexing parameter with this hithermost value again.
When g.723.1 the Parameters Transformation of agreement is the parameter of g.729 agreement, at first find relevant parameter in the parameter stream of agreement g.723.1 according to the vector code book of agreement g.723.1, as linear predictor coefficient (1pc), excitation information and filter information, with of the parameter input of this relevant parameter as the stream of protocol parameter g.729, according to agreement g.729 this relevant parameters is operated, promptly find the indexing parameter value of this relevant parameter correspondence, obtain needed g.729 agreement index stream according to the code book of agreement g.729.
For example: suppose to be the parameter of agreement g.729 from the Parameters Transformation of agreement g.723.1, then g.723.1 the data of agreement are decoded and are obtained excitation information and filter information, and the excitation information of acquisition is: a; The filter information that obtains is: b0, and b1, b2 ... b9; Directly with a and b0, b1, b2 ... b9 is as input, according to the vector code book of agreement g.729 with a and b0, b1, b2 ... b9 finds to the indexing parameter value of correspondence, finally obtains the index stream of the g.729 agreement that needs.
Method provided by the present invention, without this step of synthetic speech, directly the conversion through parameter is changed to purpose parameter stream from the raw parameter circulation, has saved this process of synthetic speech and extracting parameter.
Method provided by the invention does not have synthetic speech and this process of extracting parameter, has avoided the damage to voice: because synthetic speech and extracting parameter itself are the processes of damage voice; Method provided by the invention has reduced time delay; Method provided by the invention is convenient to carry out highdensity integrated, the computational resource that synthetic speech and Parameter Extraction account in the process of Code And Decode is bigger, adopt the mode of Parameters Transformation, can avoid the work of this part, comparatively speaking, the system resource that the conversion of parameter needs is much smaller, and therefore can carry out highdensity integrated.
The above only is preferred embodiment of the present invention, not in order to restriction the present invention, all any modifications of being made within the spirit and principles in the present invention, is equal to and replaces and improvement etc., all should be included within protection scope of the present invention.
Claims (4)
1. the conversion method of voice coding between a different voice protocol is characterized in that this method comprises:
A, the data bit flow of source code agreement is decoded by its agreement, decoding obtains the index stream of source code agreement, and the index stream of this respective protocol is searched the vector code book, obtains the parameter stream of source code agreement;
B, the parameter that parameter is flowed as the target code agreement flow;
C, with each parameter in target code protocol parameter described in the step B stream as input, search the vector code book of target code agreement, if the parameter in the parameter that finds and the target code protocol parameter stream is identical, then directly this parameter is converted to the index stream parameter of target code agreement according to the vector code book; Otherwise, the vector code book find with target code protocol parameter stream in the parameter of parameter equivalence, the parameter that this is of equal value is converted to the index stream parameter of target code agreement according to the vector code book, index stream to the target code agreement after the conversion is encoded, and finally obtains the data bit flow of target code agreement.
2. the method for claim 1 is characterized in that, described equivalent parameter is the square error minimum of the parameter in parameter that the vector code book finds and target code protocol parameter stream.
3. the method for claim 1 is characterized in that, includes excitation information and filter information in the parameter stream of described source code agreement.
4. the method for claim 1 is characterized in that, described source code protocol type is during for agreement g.723.1, and the target code protocol type is g.729 agreement or ARM agreement;
Described source code protocol type is during for agreement g.729, and the target code protocol type is g.723.1 agreement or ARM agreement;
When described source code protocol type was the ARM agreement, the target code protocol type was g.723.1 agreement or g.729 agreement.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN031539335A CN100407292C (en) | 2003-08-20 | 2003-08-20 | Method for converting speech code between differential speech agreement |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN031539335A CN100407292C (en) | 2003-08-20 | 2003-08-20 | Method for converting speech code between differential speech agreement |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1584983A CN1584983A (en) | 2005-02-23 |
CN100407292C true CN100407292C (en) | 2008-07-30 |
Family
ID=34597925
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN031539335A Expired - Fee Related CN100407292C (en) | 2003-08-20 | 2003-08-20 | Method for converting speech code between differential speech agreement |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN100407292C (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0944194A (en) * | 1995-08-03 | 1997-02-14 | Kokusai Electric Co Ltd | Voice encoding communication system |
US5903866A (en) * | 1997-03-10 | 1999-05-11 | Lucent Technologies Inc. | Waveform interpolation speech coding using splines |
CN1435817A (en) * | 2002-01-29 | 2003-08-13 | 富士通株式会社 | Voice coding converting method and device |
-
2003
- 2003-08-20 CN CN031539335A patent/CN100407292C/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0944194A (en) * | 1995-08-03 | 1997-02-14 | Kokusai Electric Co Ltd | Voice encoding communication system |
US5903866A (en) * | 1997-03-10 | 1999-05-11 | Lucent Technologies Inc. | Waveform interpolation speech coding using splines |
CN1435817A (en) * | 2002-01-29 | 2003-08-13 | 富士通株式会社 | Voice coding converting method and device |
Also Published As
Publication number | Publication date |
---|---|
CN1584983A (en) | 2005-02-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
TWI453734B (en) | Method for encoding a symbol, method for decoding a symbol, method for transmitting a symbol from a transmitter to a receiver, encoder, decoder and system for transmitting a symbol from a transmitter to a receiver | |
CN102119414B (en) | Device and method for quantizing and inverse quantizing LPC filters in a super-frame | |
CN105993178B (en) | Pyramid vector quantization for audio/video sample vector, which is indexed, conciliates the method and apparatus of index | |
CN101176148B (en) | Encoder, decoder, and their methods | |
KR20070038041A (en) | Method and apparatus for voice trans-rating in multi-rate voice coders for telecommunications | |
CN102089813A (en) | Audio encoder and audio decoder | |
KR101170137B1 (en) | Reduced-complexity vector indexing and de-indexing | |
WO2010007211A1 (en) | Method and apparatus for fast nearestneighbor search for vector quantizers | |
MXPA06006621A (en) | Transcoding between the indices of multipulse dictionaries used for coding in digital signal compression. | |
CN101572093B (en) | Method and device for transcoding | |
CN102761543B (en) | A kind of method and apparatus realizing the general encoding and decoding of Session Initiation Protocol | |
KR20120096541A (en) | Method, system, and apparatus for compression or decompression of digital signals | |
US8571112B2 (en) | Specification method and apparatus for coding and decoding | |
US6223157B1 (en) | Method for direct recognition of encoded speech data | |
CN100407292C (en) | Method for converting speech code between differential speech agreement | |
US9318115B2 (en) | Efficient coding of binary strings for low bit rate entropy audio coding | |
WO2011150810A1 (en) | Data encoding method, decoding method, encoder and decoder | |
KR100451622B1 (en) | Voice coder and communication method using the same | |
CN114429200A (en) | Standardized Huffman coding and decoding method and neural network computing chip | |
CN102422541B (en) | Method, device and system of coding, and method, device and system of decoding | |
Wernik et al. | Analysis of inter-channel dependencies in audio lossless block coding. | |
CN103474079A (en) | Voice encoding method | |
MA et al. | Description and measurement of semantic information for the intelligent machine communication | |
JP2002073097A (en) | Celp type voice coding device and celp type voice decoding device as well as voice encoding method and voice decoding method | |
Aksu et al. | Design, performance, and complexity analysis of residual trellis-coded vector quantizers |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080730 Termination date: 20190820 |
|
CF01 | Termination of patent right due to non-payment of annual fee |