CN101276587A - Audio encoding apparatus and method thereof, audio decoding device and method thereof - Google Patents

Audio encoding apparatus and method thereof, audio decoding device and method thereof Download PDF

Info

Publication number
CN101276587A
CN101276587A CNA2007100888785A CN200710088878A CN101276587A CN 101276587 A CN101276587 A CN 101276587A CN A2007100888785 A CNA2007100888785 A CN A2007100888785A CN 200710088878 A CN200710088878 A CN 200710088878A CN 101276587 A CN101276587 A CN 101276587A
Authority
CN
China
Prior art keywords
frequency sub
low frequency
band
band territory
signal
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.)
Granted
Application number
CNA2007100888785A
Other languages
Chinese (zh)
Other versions
CN101276587B (en
Inventor
潘兴德
徐恒宇
李靓
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
BEIJING TIANLAI CHUANYIN DIGITAL TECHNOLOGY Co Ltd
Original Assignee
BEIJING TIANLAI CHUANYIN DIGITAL TECHNOLOGY 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 BEIJING TIANLAI CHUANYIN DIGITAL TECHNOLOGY Co Ltd filed Critical BEIJING TIANLAI CHUANYIN DIGITAL TECHNOLOGY Co Ltd
Priority to CN2007100888785A priority Critical patent/CN101276587B/en
Publication of CN101276587A publication Critical patent/CN101276587A/en
Application granted granted Critical
Publication of CN101276587B publication Critical patent/CN101276587B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Transmission Systems Not Characterized By The Medium Used For Transmission (AREA)

Abstract

The invention discloses a single acoustic path sound coding device using an analysis sub-band filter set module to decompound a single acoustic path sound signal into a low-frequency sub-band domain signal comprising at least two sub-bands and a high-frequency sub-band domain signal comprising at least two sub-bands; the low-frequency sub-band domain signal is performed with prediction analysis by a low-frequency sub-band domain time-varying prediction analysis module, is performed with time domain transformation by a low-frequency sub-band domain time-frequency transformation module, and is performed with quantification coding by a low-frequency sub-band domain waveform coding module so as to obtain low-frequency sub-band domain waveform coding data; high-frequency sub-band domain parameter coding data is extracted from the high-frequency sub-band domain signal by parameter coding of a high-frequency sub-band domain parameter coding module; the low-frequency sub-band domain waveform coding data and the high-frequency sub-band domain parameter coding data are multiplexed by a bit-stream multiplexing module to output a sound code stream. The invention also discloses a single acoustic path sound decoding device, a single acoustic path sound coding/decoding method, a dimensional sound coding/decoding device and method, all of which can realize high-quality coding of wide band sound in a low code rate.

Description

Sound coder and method thereof and sound decoding device and method thereof
Technical field
The present invention relates to the sound encoding and decoding technique, be specifically related to the code device and the method thereof of the decoding device of the code device of monophonic sounds and method thereof, monophonic sounds and method thereof, stereo sound, and the decoding device and the method thereof of stereo sound.
Background technology
At present, high-quality compression coding technology in broadband all adopts the waveform coding technology.The waveform coding technology roughly is divided into 3 classes: predictive coding, sub-band coding and transition coding.Wherein, the actual value that predictive coding will be transmitted voice signal changes the transmission error value into, can express voice signal with less coding figure place.Sub-band coding is divided into several continuous subbands with voice signal, just can adopt independent encoding scheme to encode to the signal on each subband, for example, adopt different quantization steps, to greatest extent the lower quantization error at the subband signal of varying strength.Transition coding with the time domain space signal transformation to such as the transform domain of frequency domain, it is more concentrated to make that signal distributes in transform domain, can improve code efficiency.Above-mentioned encoding and decoding technique can also mutually combine, and forms present subband predictive coding, sub-band transforms coding and predictive transformation coding.
Wherein, the subband predictive coding APT-X100 of system that adopted of digital home's cinema system of Digital Theater System company (DTS, Digital TheaterSystems) is the representative of subband predictive coding.The APT-X100 system at first adopts two-stage tandem quadrature mirror filter bank (QMF, QuadratureMirror Filter) sound signal is divided into 4 subband signals, and subband signal is carried out time domain linear predict, to obtain to remove the subband pumping signal of subband redundancy, adopt the adaptive quantizing coding then, obtain sound signal through the subband predictive coding.
The Advanced Audio Coding (AAC, Advanced Audio Coding) of the AC-3 of Dolby Labs and mobile motion picture expert group version (MPEG, Moving Picture ExpertsGroup) tissue is the representative of adopting the transition coding scheme.The transition coding scheme of AC-3 and AAC is at first revised discrete cosine transform/correction discrete sine transform (MDCT to sound signal, Modified Discrete Cosine Transform/MDST, Modified Discrete Sine Transform), and conversion coefficient is carried out Bit Allocation in Discrete according to psychoacoustic model, adopt the scalar Huffman encoding to carry out quantization encoding then, obtain sound signal through transition coding.
The MP3 of MPEG tissue is the representative of sub-band transforms coding.The MPEG layers 1 and 2 adopts pseudo-mirror filter group (PQMF, Pseudo Quadrature Mirror Filter) that sound signal is divided into 32 subband signals, and subband signal is carried out scalar quantization and entropy coding; The 3rd layer of MPEG, promptly MP3 then on the basis of 32PQMF sub-band filter, further adopts MDCT to reject signal redundancy, carries out scalar quantization and entropy coding then, obtains the sound signal through the sub-band transforms coding.
MPGE TwinVQ is the representative of predictive transformation coding, can obtain very high code efficiency equally.
Practice shows, subband predictive coding, transition coding, sub-band transforms coding and predictive transformation coding can be with the audio signal compression of monophony 48kHz and the 16 bit sample code checks to about the 64kb/s, thereby effectively improve code efficiency, be well suited for high-quality acoustic coding and use.But, if further compression bit rate then must adopt the parameter coding technology that is more suitable in low code check acoustic coding application.
At present, in Low Bit-rate Coding was used, the typical practice was based on waveform coding " waveform-parameter " encoding scheme.The combination of waveform and parameter coding technology can effectively improve code efficiency.Wherein, by third generation partner program (3GPP, 3rd Generation Partnership Project) the enhanced advanced audio coding (EAAC+ of the expansion of adopting, Enhanced Advanced AudioCoding Plus) and the expansion the many speed (AMR-WB+ of wideband adaptive, Extended AdaptiveMulti-Rate-Wideband) coding and decoding scheme is to be proved to be the acoustic coding technology that low bandwidth such as the most suitable mobile communication is used so far.The decoder architecture of the coder structure of EAAC+ scheme and EAAC+ scheme is respectively referring to Fig. 1 and Fig. 2.The decoder architecture of the coder structure of AMR-WB+ scheme and AMR-WB+ scheme is respectively referring to Fig. 3 and Fig. 4.
Wherein, in the EAAC+ scheme, on MPEG ACC technical foundation based on the waveform coding technology, increase frequency spectrum and duplicated (SBR, Spectral Band Replication) coding techniques, this technology can be carried out effective spectrum structure adjustment to the high frequency copy pumping signal that low-frequency excitation signal copy obtains.The basic model of EAAC+ scheme has adopted the higher waveform coding technology of efficient, and than the more suitable music signal coding of AMR-WB+ scheme, but decoding complex degree is high more a lot of than AMR-WB+.Simultaneously, owing to do not have to adopt the actual parameter module that meets speech production mechanism, i.e. " prediction+excitation ", the voice coding mass ratio AMR-WB+ under the low code check is low.
Parameter stereo coding technology in the EAAC+ encoding scheme is the coding method with higher compression ratio and encode sound quality, but this encoding scheme is based on the implementation method of subband signal frequency spectrum.And the parameter stereo coding technology among the present EAAC+ does not relate to the stereo coding solution based on the subband excitation spectrum.
In the AMR-WB+ scheme, by simple filtering method input signal is decomposed into two subbands, first subband is low frequency (LF, Low Frequency) signal, through threshold sampling, its sample frequency is Fs/2, and wherein Fs is the sample frequency through the resampling module institute picked up signal of AMR-WB+ scrambler, in stereo coding, this low frequency signal is decomposed into the intermediate-freuqncy signal of the low frequency signal and a subband of a subband again; Second subband is high frequency (HF, High Frequency) signal, and equally by threshold sampling, its sample frequency is Fs/2.Wherein, the HF signal adopts simple high frequency extended coding (BWE, Band Width Extension) technology, this BWE technology is the implementation method of a time domain, decoding complex degree is lower, but can not carry out effective spectrum structure adjustment to the high frequency copy pumping signal that low-frequency excitation signal copy obtains, the sound quality that the back of therefore encoding generates is lower, and particularly big distortion can appear in the encode sound to some music signals.
Stereo coding technology in the AMR-WB+ scheme adopts time-domain filtering method to realize, its resolution is not as the EAAC+ height, and the sound quality that the decoding back generates is low than EAAC+.
As seen, in the Low Bit-rate Coding of digital audio was used, all there were certain defective in existing EAAC+ technology and AMR-WB+ technology aspect coding quality and complexity.At present, need a kind of more rational sound encoding and decoding technique, make under lower code check and lower implementation complexity constraint, realize high-quality acoustic coding, overcome the problem that above technology exists.
Summary of the invention
In view of this, fundamental purpose of the present invention is to provide a kind of monophonic sounds code device, can realize that the high-quality to wideband voice is encoded under the low code check.
Second fundamental purpose of the present invention is to provide a kind of monophonic sounds coding method, can realize that the high-quality to wideband voice is encoded under the low code check.
The 3rd fundamental purpose of the present invention is to provide a kind of monophonic sound sound decoding device, can realize that the high-quality to wideband voice is decoded under the low code check.
The 4th fundamental purpose of the present invention is to provide a kind of monophonic sounds coding/decoding method, can realize that the high-quality to wideband voice is decoded under the low code check.
The 5th fundamental purpose of the present invention is to provide a kind of stereo encoding apparatus, can realize that the high-quality to wideband voice is encoded under the low code check.
The 6th fundamental purpose of the present invention is to provide a kind of stereo encoding method, can realize that the high-quality to wideband voice is encoded under the low code check.
The 7th fundamental purpose of the present invention is to provide a kind of stereo decoding apparatus, can realize that the high-quality to wideband voice is decoded under the low code check.
The 8th fundamental purpose of the present invention is to provide a kind of stereo decoding method, can realize that the high-quality to wideband voice is decoded under the low code check.
According to first aspect of above-mentioned purpose, the invention provides a kind of monophonic sounds code device, comprising:
Analyze the sub-filter pack module, be used for the monophonic sound tone signal is carried out sub-band division, be decomposed into the low frequency sub-band territory signal that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises at least 2 subbands;
Become the forecast analysis module during low frequency sub-band territory, be used for described low frequency sub-band territory signal is carried out forecast analysis, to obtain low frequency sub-band territory pumping signal;
Low frequency sub-band territory time-frequency conversion module is used for described low frequency sub-band territory pumping signal is carried out time-frequency conversion, to obtain low frequency sub-band territory excitation spectrum;
Low frequency sub-band domain waveform coding module is used for described low frequency sub-band territory excitation spectrum is carried out quantization encoding, to obtain low frequency sub-band domain waveform coded data;
High-frequency sub-band field parameter coding module, be used for being used for from the high-frequency sub-band field parameter of low frequency sub-band territory excitation spectrum recovery high-frequency sub-band territory signal according to described low frequency sub-band territory excitation spectrum and described high-frequency sub-band territory calculated signals, after this high-frequency sub-band field parameter carried out quantization encoding, to obtain high-frequency sub-band field parameter coded data; Perhaps, be used for from the high-frequency sub-band field parameter of low frequency sub-band territory pumping signal recovery high-frequency sub-band territory signal according to described low frequency sub-band territory pumping signal and described high-frequency sub-band territory calculated signals, after this high-frequency sub-band field parameter carried out quantization encoding, to obtain high-frequency sub-band field parameter coded data;
The bit stream Multiplexing module, be used for described low frequency sub-band domain waveform coded data and high-frequency sub-band field parameter coded data are carried out multiplexing, with the output sound encoding code stream.
This monophonic sounds code device further comprises low frequency sub-band territory signal type analysis module, and this module is used for the described low frequency sub-band of frame territory signal is carried out the signal type analysis, and output low frequency subband domain signal type analysis result; If described low frequency sub-band territory signal is tempolabile signal then output signal type; If fast changed signal then further obtains fast height position, and output signal type and described fast height position;
Becoming the forecast analysis module during described low frequency sub-band territory and be further used for according to described low frequency sub-band territory signal type analysis result, is subframe one or more, that be used for forecast analysis with described low frequency sub-band territory division of signal;
Described low frequency sub-band territory time-frequency conversion module is further used for according to described low frequency sub-band territory signal type analysis result described low frequency sub-band territory pumping signal being divided into one or more, as to be used for time-frequency conversion subframe;
Described bit stream Multiplexing module is further used for carrying out multiplexing to described low frequency sub-band territory signal type analysis result.
Second aspect according to above-mentioned purpose the invention provides a kind of monophonic sounds coding method, comprising:
A, the monophonic sound tone signal is carried out sub-band division, be decomposed into the low frequency sub-band territory signal that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises at least 2 subbands;
B, described low frequency sub-band territory signal is carried out forecast analysis and time-frequency conversion, to obtain low frequency sub-band territory excitation spectrum; Described low frequency sub-band territory excitation spectrum is carried out quantization encoding, to obtain low frequency sub-band domain waveform coded data;
C, according to described high-frequency sub-band territory signal and low frequency sub-band territory excitation spectrum, calculating is used for recovering from low frequency sub-band territory excitation spectrum the high-frequency sub-band field parameter of high-frequency sub-band territory signal, and after this high-frequency sub-band field parameter carried out quantization encoding, obtain high-frequency sub-band field parameter coded data; Perhaps, according to described high-frequency sub-band territory signal with through low frequency sub-band territory pumping signal that described forecast analysis obtains, calculating is used for recovering from low frequency sub-band territory pumping signal the high-frequency sub-band field parameter of high-frequency sub-band territory signal, and after the high-frequency sub-band field parameter carried out quantization encoding, obtain high-frequency sub-band field parameter coded data;
D, described low frequency sub-band domain waveform coded data and described high-frequency sub-band field parameter coded data are carried out multiplexing, the output sound encoding code stream.
This monophonic sounds coding method further comprises step: a frame low frequency sub-band territory signal is carried out the signal type analysis, and definite low frequency sub-band territory signal type analysis result; If described low frequency sub-band territory signal is a tempolabile signal, then with signal type as low frequency sub-band territory signal type analysis result; If fast changed signal then further obtains fast height position, with signal type and fast height position as the signal type analysis result;
Step B is described to carry out forecast analysis and time-frequency conversion comprises to described low frequency sub-band territory signal: according to low frequency sub-band territory signal type analysis result, be subframe one or more, that be used for forecast analysis with a frame low frequency sub-band territory division of signal; By subframe described low frequency sub-band territory signal is carried out forecast analysis, obtain the low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory signal type analysis result, a described frame low frequency sub-band territory pumping signal is divided into one or more, as to be used for time-frequency conversion subframe; By subframe described low frequency sub-band territory pumping signal is carried out time-frequency conversion, obtain the low frequency sub-band territory excitation spectrum of subframe;
Step D further comprises and carries out multiplexing to described low frequency sub-band territory signal type analysis result.
The 3rd aspect according to above-mentioned purpose the invention provides a kind of monophonic sound sound decoding device, comprising:
The bit stream demultiplexing module is used for the acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band domain waveform coded data that comprises at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises at least 2 subbands;
Low frequency sub-band domain waveform decoder module is used for described low frequency sub-band domain waveform coded data is carried out the re-quantization decoding, to obtain low frequency sub-band territory excitation spectrum;
Low frequency sub-band territory frequency-time domain transformation module is used for described low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation, to obtain low frequency sub-band territory pumping signal;
Become the comprehensive module of prediction during the low frequency sub-band territory, be used for described low frequency sub-band territory pumping signal is predicted comprehensively, to obtain low frequency sub-band territory signal;
High-frequency sub-band field parameter decoder module, be used for described high-frequency sub-band field parameter coded data is carried out the re-quantization decoding with acquisition high-frequency sub-band field parameter, and from the excitation spectrum of described low frequency sub-band territory, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter; Perhaps, from the pumping signal of described low frequency sub-band territory, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter;
Comprehensive sub-filter pack module, it is synthetic to be used for that described low frequency sub-band territory signal and described high-frequency sub-band territory signal are carried out subband, to obtain the monophonic sound tone signal of decoding.
Wherein, described bit stream demultiplexing module is further used for, and obtains the low frequency sub-band territory signal type analysis result that is used to recover monophonic sounds from the described acoustic coding code stream of demultiplexing;
Described low frequency sub-band territory frequency-time domain transformation module is further used for according to described low frequency sub-band territory signal type analysis result, and the low frequency sub-band territory excitation spectrum that receives is divided into one or more, as to be used for frequency-time domain transformation subframe;
Become the comprehensive module of prediction during described low frequency sub-band territory and be further used for according to low frequency sub-band territory signal type analysis result, with the low frequency sub-band territory pumping signal that receives be divided into one or more, be used to predict comprehensive subframe.
The 4th aspect according to above-mentioned purpose the invention provides a kind of monophonic sounds coding/decoding method, comprising:
A, the acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band domain waveform coded data that comprises at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises at least 2 subbands;
B, described low frequency sub-band domain waveform coded data is carried out re-quantization decoding, to obtain low frequency sub-band territory excitation spectrum; Described low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation and predicts comprehensively, to obtain low frequency sub-band territory signal;
C, described high-frequency sub-band field parameter coded data is carried out re-quantization decoding obtaining the high-frequency sub-band field parameter, and from the excitation spectrum of described low frequency sub-band territory, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter; Perhaps, from the low frequency sub-band territory pumping signal that obtains through described frequency-time domain transformation, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter;
D, that described low frequency sub-band territory signal and described high-frequency sub-band territory signal are carried out subband is synthetic, the monophonic sound tone signal of output decoder.
Wherein, described steps A further comprises obtain the low frequency sub-band territory signal type analysis result that is used to recover monophonic sounds from the acoustic coding code stream of demultiplexing;
Step B is described to carry out low frequency sub-band territory excitation spectrum frequency-time domain transformation and predicts comprehensively comprising: according to described low frequency sub-band territory signal type analysis result, the described low frequency sub-band of frame territory excitation spectrum is divided into one or more, as to be used for frequency-time domain transformation subframe; By subframe described low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation, obtain the low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory signal type analysis result, with a described frame low frequency sub-band territory pumping signal be divided into one or more, be used to predict comprehensive subframe; By subframe described low frequency sub-band territory signal is predicted comprehensively, obtained the low frequency sub-band territory signal of subframe.
The 5th aspect according to above-mentioned purpose the invention provides a kind of stereo encoding apparatus, comprising:
Analyze the sub-filter pack module, the left and right sound channels that is used for stereophonic signal is carried out sub-band division respectively, be decomposed into the low frequency sub-band territory signal of the left and right sound channels that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises the left and right sound channels of at least 2 subbands;
Become the forecast analysis module during low frequency sub-band territory, be used for respectively the low frequency sub-band territory signal of described left and right sound channels is carried out forecast analysis, to obtain the low frequency sub-band territory pumping signal of left and right sound channels;
Low frequency sub-band territory time-frequency conversion module is used for respectively described left and right sound channels low frequency sub-band territory pumping signal being carried out time-frequency conversion, to obtain the low frequency sub-band territory excitation spectrum of left and right sound channels;
Low frequency sub-band territory stereo coding module is used for the low frequency sub-band territory excitation spectrum of described left and right sound channels is carried out stereo coding, to obtain low frequency sub-band territory stereo coding data;
High-frequency sub-band field parameter coding module, be used for respectively according to the low frequency sub-band territory excitation spectrum of described left and right sound channels and the high-frequency sub-band territory signal of described left and right sound channels, calculating is used for from the high-frequency sub-band field parameter of the left and right sound channels of the high-frequency sub-band territory signal of the low frequency sub-band territory excitation spectrum recovery left and right sound channels of left and right sound channels, and respectively the high-frequency sub-band field parameter of described left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels; Perhaps, respectively according to the high-frequency sub-band territory signal of the low frequency sub-band territory pumping signal and the described left and right sound channels of described left and right sound channels, calculating is used for from the high-frequency sub-band field parameter of the left and right sound channels of the high-frequency sub-band territory signal of the low frequency sub-band territory pumping signal recovery left and right sound channels of left and right sound channels, and respectively the high-frequency sub-band field parameter of described left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels;
The bit stream Multiplexing module, be used for the high-frequency sub-band field parameter coded data of described low frequency sub-band territory stereo coding data and described left and right sound channels is carried out multiplexing, to export stereosonic acoustic coding code stream.
This stereo encoding apparatus further comprises low frequency sub-band territory and signal type analysis module, this module is used for going out a frame low frequency sub-band territory and a signal according to the low frequency sub-band territory calculated signals of a frame left and right sound channels, the signal type analysis is carried out in this low frequency sub-band territory and signal, and output low frequency subband domain and signal type analysis result; If described low frequency sub-band territory and signal are tempolabile signal then output signal type; If fast changed signal then further obtains fast height position, and output signal type and described fast height position;
Become the forecast analysis module during described low frequency sub-band territory and be further used for according to described low frequency sub-band territory and signal type analysis result, the low frequency sub-band territory division of signal of the described left and right sound channels of a frame that will receive respectively is a subframe one or more, that be used for forecast analysis;
Described low frequency sub-band territory time-frequency conversion module is further used for according to described low frequency sub-band territory and signal type analysis result, and the described left and right sound channels low frequency sub-band territory pumping signal that will receive respectively is divided into one or more, as to be used for time-frequency conversion subframe;
Described bit stream Multiplexing module is further used for carrying out multiplexing to described low frequency sub-band territory and signal type analysis result.
When carrying out stereo coding, described low frequency sub-band territory stereo coding module is further used for selecting from more than one optional stereo coding pattern, adopt the stereo coding pattern of selecting to encode, and export coding mode selection information to described bit stream Multiplexing module
The 6th aspect according to above-mentioned purpose the invention provides a kind of stereo encoding method, comprising:
The left and right sound channels of A, stereophonic signal is carried out sub-band division respectively, is decomposed into the low frequency sub-band territory signal of the left and right sound channels that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises the left and right sound channels of at least 2 subbands;
B, the low frequency sub-band territory signal to described left and right sound channels carries out forecast analysis and time-frequency conversion respectively, to obtain the low frequency sub-band territory excitation spectrum of left and right sound channels; Low frequency sub-band territory excitation spectrum to described left and right sound channels carries out stereo coding, to obtain low frequency sub-band territory stereo coding data;
C, respectively according to the high-frequency sub-band territory signal and the described left and right sound channels low frequency sub-band territory excitation spectrum of described left and right sound channels, calculating is used for recovering from left and right sound channels low frequency sub-band territory excitation spectrum the high-frequency sub-band field parameter of the left and right sound channels of left and right sound channels high-frequency sub-band territory signal, and respectively the high-frequency sub-band field parameter of left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels; Perhaps, the low frequency sub-band territory pumping signal of the left and right sound channels that obtains according to the high-frequency sub-band territory signal of described left and right sound channels with through described forecast analysis respectively, calculating is used for recovering from left and right sound channels low frequency sub-band territory pumping signal the high-frequency sub-band field parameter of the left and right sound channels of left and right sound channels high-frequency sub-band territory signal, and respectively the high-frequency sub-band field parameter of described left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels;
D, the high-frequency sub-band field parameter coded data of described low frequency sub-band territory stereo coding data and described left and right sound channels is carried out multiplexing, to export stereosonic acoustic coding code stream.
This stereo encoding method further comprises step: go out a frame low frequency sub-band territory and a signal according to steps A through the low frequency sub-band territory calculated signals of a frame left and right sound channels of sub-band division acquisition, the signal type analysis is carried out in this low frequency sub-band territory and signal, and definite low frequency sub-band territory and signal type analysis result; If described low frequency sub-band territory and signal are tempolabile signals, then with signal type as described low frequency sub-band territory and signal type analysis result; If fast changed signal then further obtains fast height position, with signal type and described fast height position as described low frequency sub-band territory and signal type analysis result;
The described low frequency sub-band territory signal to a described left side/R channel of step B carries out forecast analysis and time-frequency conversion comprises: according to low frequency sub-band territory and signal type analysis result, be subframe one or more, that be used for forecast analysis with a frame left side/R channel low frequency sub-band territory division of signal; By subframe the low frequency sub-band territory signal of a described left side/R channel is carried out forecast analysis, obtain a left side/R channel low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame left side/R channel low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory and signal type analysis result, a described frame left side/R channel low frequency sub-band territory pumping signal is divided into one or more, as to be used for time-frequency conversion subframe; By subframe a described left side/R channel low frequency sub-band territory pumping signal is carried out time-frequency conversion, obtain a left side/R channel low frequency sub-band territory excitation spectrum of subframe;
Step D further comprises and carries out multiplexing to described low frequency sub-band territory and signal type analysis result.
When carrying out stereo coding, the described low frequency sub-band territory excitation spectrum to left and right sound channels of step B carries out stereo coding and comprises:
B2, the low frequency sub-band territory excitation spectrum with described left and right sound channels is divided into plurality of sub-bands respectively, selects a kind of stereo coding pattern to carry out stereo coding to each sub-band;
Step D further comprises and carries out multiplexing to coding mode selection information.
The 7th aspect according to above-mentioned purpose the invention provides a kind of stereo decoding apparatus, comprising:
The bit stream demultiplexing module is used for stereosonic acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band territory stereo coding data that comprise at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises the left and right sound channels of at least 2 subbands;
Low frequency sub-band territory stereo decoding module is used for described low frequency sub-band territory stereo coding data are carried out stereo decoding, to obtain the low frequency sub-band territory excitation spectrum of left and right sound channels;
Conversion module during described low frequency sub-band territory is used for respectively the low frequency sub-band territory excitation spectrum of described left and right sound channels is carried out frequency-time domain transformation, to obtain the low frequency sub-band territory pumping signal of left and right sound channels;
Become the comprehensive module of prediction during the low frequency sub-band territory, be used for respectively the low frequency sub-band territory pumping signal of described left and right sound channels being predicted comprehensively, to obtain the low frequency sub-band territory signal of left and right sound channels;
High-frequency sub-band field parameter decoder module, be used for the high-frequency sub-band field parameter coded data of described left and right sound channels is carried out the high-frequency sub-band field parameter of re-quantization decoding with the acquisition left and right sound channels, and from the low frequency sub-band territory excitation spectrum of described left and right sound channels, recover the high-frequency sub-band territory signal of left and right sound channels respectively according to the high-frequency sub-band field parameter of described left and right sound channels; Perhaps, from the low frequency sub-band territory pumping signal of described left and right sound channels, recover the high-frequency sub-band territory signal of left and right sound channels respectively according to the high-frequency sub-band field parameter of described left and right sound channels;
Comprehensive sub-filter pack module, being used for high-frequency sub-band territory signal with the low frequency sub-band territory signal of described left and right sound channels and described left and right sound channels, to carry out subband synthetic, with the stereophonic signal of the left and right sound channels of obtaining decoding.
Wherein, described bit stream demultiplexing module is further used for, and obtains to be used to recover stereosonic low frequency sub-band territory and signal type analysis result from the described stereosonic acoustic coding code stream of demultiplexing;
Described low frequency sub-band territory frequency-time domain transformation module is further used for according to described low frequency sub-band territory and signal type analysis result, and the low frequency sub-band territory excitation spectrum with the left and right sound channels that receives is divided into one or more, as to be used for frequency-time domain transformation subframe respectively;
Become the comprehensive module of prediction during described low frequency sub-band territory and be further used for according to described low frequency sub-band territory and signal type analysis result, with the low frequency sub-band territory pumping signal of the left and right sound channels that receives be divided into one or more, be used to predict comprehensive subframe.
Described bit stream Multiplexing module is further used for, and obtains the coding mode that is used for stereo decoding and select information from the described stereosonic acoustic coding code stream of demultiplexing;
When carrying out stereo coding, described low frequency sub-band territory stereo decoding module is further used for, and adopts described coding mode to select the stereo decoding pattern of information correspondence to carry out stereo decoding.
The 8th aspect according to above-mentioned purpose the invention provides a kind of stereo decoding method, comprising:
A, stereosonic acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band territory stereo coding data that comprise at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises the left and right sound channels of at least 2 subbands
B, the low frequency sub-band territory stereo coding data that described demultiplexing is obtained are carried out stereo decoding, obtain the low frequency sub-band territory excitation spectrum of left and right sound channels; Low frequency sub-band territory excitation spectrum to described left and right sound channels carries out frequency-time domain transformation and predicts comprehensively respectively, to obtain the low frequency sub-band territory signal of left and right sound channels;
The high-frequency sub-band field parameter coded data of C, left and right sound channels that described demultiplexing is obtained is carried out the re-quantization decoding obtaining the high-frequency sub-band field parameter of left and right sound channels, and recovers the high-frequency sub-band territory signal of left and right sound channels respectively from the low frequency sub-band territory excitation spectrum of described left and right sound channels according to the high-frequency sub-band field parameter of described left and right sound channels; Perhaps respectively according to the high-frequency sub-band field parameter of described left and right sound channels from the described high-frequency sub-band territory signal that through the low frequency sub-band territory pumping signal of the left and right sound channels that frequency-time domain transformation obtains, recovers left and right sound channels;
D, that the high-frequency sub-band territory signal of the low frequency sub-band territory signal of described left and right sound channels and described left and right sound channels is carried out subband is synthetic, obtains the stereophonic signal of the left and right sound channels of decoding.
Wherein, steps A further comprises obtaining from the acoustic coding code stream of demultiplexing and is used to recover stereosonic low frequency sub-band territory and signal type analysis result;
The described low frequency sub-band territory excitation spectrum with a left side/R channel of step B carries out frequency-time domain transformation and prediction comprehensively comprises: according to described low frequency sub-band territory and signal type analysis result, the described left side of a frame/R channel low frequency sub-band territory excitation spectrum is divided into one or more, as to be used for frequency-time domain transformation subframe; By subframe a described left side/R channel low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation, obtain a left side/R channel low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame left side/R channel low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory and signal type analysis result, with a described frame left side/R channel low frequency sub-band territory pumping signal be divided into one or more, be used to predict comprehensive subframe; By subframe a described left side/R channel low frequency sub-band territory signal is predicted comprehensively, obtained a left side/R channel low frequency sub-band territory signal of subframe
When carrying out stereo decoding, described steps A comprises that further obtaining the coding mode that is used for stereo decoding from the acoustic coding code stream of demultiplexing selects information;
The described stereo decoding that carries out of step B is: adopt described coding mode to select the stereo decoding pattern of information correspondence, the low frequency sub-band territory stereo coding data of each sub-band k are decoded.
Compared with prior art, monophonic sounds coding and decoding scheme provided by the present invention carries out sub-band division to voice signal earlier, the high-frequency sub-band territory signal that is decomposed into the low frequency sub-band territory signal that comprises at least 2 subbands and comprises 2 subbands at least adopts different disposal routes to encode to low frequency sub-band territory signal and high-frequency sub-band territory signal respectively then.When low frequency sub-band territory signal is carried out waveform coding, low frequency sub-band territory signal has been adopted the actual parameter model of " prediction+excitation ", improved the voice coding quality under low constrained code rate; Then low frequency sub-band territory pumping signal is transformed to low frequency sub-band territory excitation spectrum, signal is more concentrated in the distribution of transformation space, can represent identical voice signal with lower code check, has therefore reduced code check, at last with subband domain excitation spectrum quantization encoding and output.Adopted the combining form of high efficiency sub-band coding, predictive coding and transition coding when as seen, the low-frequency sound signal being carried out waveform coding.When high-frequency sub-band territory signal is handled, adopted the high efficiency factors coding form.And in the process of parameter coding, carry out the gain of effective spectrum structure adjustment and time domain and adjust, the distortion of sound after not only improving code efficiency but also reducing decoding.
The present invention also provides stereosonic coding and decoding scheme, this scheme not only has the above-mentioned advantage that is had based on the monophonic sounds coding and decoding scheme of the principle of the invention, multiple parameter stereo coding method based on the subband excitation spectrum also is provided, has been adapted at the stereo coding under the extremely low code check.
Description of drawings
Fig. 1 is the structured flowchart of EAAC+ scrambler in the prior art;
Fig. 2 is the structured flowchart of EAAC+ demoder in the prior art;
Fig. 3 is the structured flowchart of AMR-WB+ scrambler in the prior art;
Fig. 4 is the structured flowchart of AMR-WB+ demoder in the prior art;
Fig. 5 is the structured flowchart of preferred embodiment of the present invention monophonic sounds code device;
When being low frequency sub-band shown in Figure 5 territory, Fig. 6 becomes the structured flowchart of forecast analysis module;
Fig. 7 is the structured flowchart of high-frequency sub-band field parameter coding module shown in Figure 5;
Fig. 8 is the coding method process flow diagram based on monophonic sounds code device of the present invention;
Fig. 9 a is the time-frequency plane figure of the tempolabile signal behind the time-frequency conversion;
Fig. 9 b is the time-frequency plane figure of the fast changed signal behind the time-frequency conversion;
Figure 10 is the structured flowchart of preferred embodiment of the present invention monophonic sound sound decoding device;
When being low frequency sub-band shown in Figure 10 territory, Figure 11 becomes the structured flowchart of the comprehensive module of prediction;
Figure 12 is the structured flowchart of high-frequency sub-band field parameter decoder module shown in Figure 10;
Figure 13 is the coding/decoding method process flow diagram based on monophonic sound sound decoding device of the present invention;
Figure 14 is the structured flowchart of preferred embodiment of the present invention stereo encoding apparatus;
Figure 15 is the coding method process flow diagram based on stereo encoding apparatus of the present invention;
Figure 16 is the illustraton of model of the present invention and difference stereo coding pattern;
Figure 17 is the illustraton of model of parameter stereo coding pattern of the present invention;
Figure 18 is the illustraton of model of parameter error stereo coding pattern of the present invention;
Figure 19 is the structured flowchart of preferred embodiment of the present invention stereo decoding apparatus;
Figure 20 is the coding/decoding method process flow diagram based on stereo sound decoding device of the present invention.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with embodiment and accompanying drawing, the present invention is described in more detail.
Fig. 5 is the structured flowchart as the monophonic sounds code device of the preferred embodiment of the present invention.As shown in Figure 5, the monophonic sounds code device of the preferred embodiment of the present invention comprises change forecast analysis module 503, low frequency sub-band territory time-frequency conversion module 504, low frequency sub-band domain waveform coding module 505, high-frequency sub-band field parameter coding module 506 and bit stream Multiplexing module 507 when analyzing sub-filter pack module 501, low frequency sub-band territory signal type analysis module 502, low frequency sub-band territory.
At first, introduce the annexation and the function of each module among Fig. 5 in detail, wherein:
Analyze sub-filter pack module 501, be used for the monophonic sound tone signal M of input is carried out sub-band division, make the corresponding specific frequency range of each subband signal, generate the subband domain signal M of each frequency range 1～M K1+k2, then with low frequency sub-band territory signal M 1～M K1Be that unit becomes forecast analysis module 503 when being input to low frequency sub-band territory signal type analysis module 502 and low frequency sub-band territory with the frame, with high-frequency sub-band territory signal M K1+1～M K1+k2With the frame is that unit is input to high-frequency sub-band field parameter coding module 506.Wherein k1 is more than or equal to 2, and k2 is more than or equal to 2.
This analysis sub-filter pack module 501 can adopt mirror filter (QMF), pseudo-mirror filter (PQMF), cosine modulation wave filter (CMF, Cosine Modulated Filter) and discrete wavelet (Discrete Wavelet) transducer etc. to carry out sub-band division and handle.
Low frequency sub-band territory signal type analysis module 502, be used for carrying out the signal type analysis to being received from every frame low frequency sub-band territory signal of analyzing sub-filter pack module 501, the type of judging this frame signal is tempolabile signal or fast changed signal, if tempolabile signal, then direct output signal type, for example, the type of this frame signal of output expression is gradual sign; If fast changed signal, then the position that fast height takes place is calculated in continuation, and exports the position of corresponding signal type and fast height generation.Become when the signal type analysis result outputs to the low frequency sub-band territory on the one hand and be used for sub-frame division control in the forecast analysis module 503, output on the other hand in the low frequency sub-band territory time-frequency conversion module 504 and be used for the exponent number of time-frequency conversion is controlled, simultaneously, the signal type analysis result can also output in the bit stream Multiplexing module 507 as the side information (Side Information) of acoustic coding code stream.This acoustic coding code stream is the data of this code device to the decoding device transmission, comprises sound coding data and side information.Wherein, side information is a kind of of code rate data, is generally control information or parameter coding information, is used for recovering voice signal in decoding end.This low frequency sub-band territory signal type analysis module 502 can adopt the signal perceptual entropy to judge signal type, judges signal type etc. by the energy of signal calculated frame.In practice, the monophonic sounds code device can not comprise this module in accordance with the principles of the present invention.
Become forecast analysis module 503 during the low frequency sub-band territory, be used for receiving low frequency sub-band territory signal from analyzing sub-filter pack module 501, and according to the low frequency sub-band territory signal type analysis result that receives from low frequency sub-band territory signal type analysis module 502, the low frequency sub-band territory signal that receives is carried out the branch subframe to be handled, by subframe low frequency sub-band territory signal is carried out forecast analysis then, be linear prediction filtering, obtain low frequency sub-band territory pumping signal, or claim low frequency sub-band territory residual signals, be that unit outputs to low frequency sub-band territory time-frequency conversion module 504 with the frame with the low frequency sub-band territory pumping signal that obtains then.When adopting linear prediction filtering, become 503 pairs of low frequency sub-band territories of forecast analysis module signal during this low frequency sub-band territory and carry out linear prediction analysis, obtain one group of predictive coefficient, be converted into one group of line spectral frequencies (LSF, Linear SpectrumFrequency), should organize the LSF parameter again and carry out vector quantization, the LSF vector quantization index that obtains is used to make up the linear prediction synthesis filter in decoding end, so this LSF vector quantization index outputs to bit stream Multiplexing module 507 as side information.
Low frequency sub-band territory time-frequency conversion module 504, become forecast analysis module 503 when being used for and receive low frequency sub-band territory pumping signal from the low frequency sub-band territory, and according to the low frequency sub-band territory signal type analysis result that receives from low frequency sub-band territory signal type analysis module 502, adopt the conversion of different length exponent number, with low frequency sub-band territory pumping signal by spatial transform to frequency domain, obtain the frequency domain representation of low frequency sub-band territory pumping signal, i.e. low frequency sub-band territory excitation spectrum.The low frequency sub-band territory excitation spectrum that time-frequency conversion is obtained outputs to low frequency sub-band domain waveform coding module 505 and high-frequency sub-band field parameter coding module 506.If the monophonic sounds code device does not comprise low frequency sub-band territory signal type analysis module 502 in accordance with the principles of the present invention, then when time-frequency conversion not match exponents control.
Low frequency sub-band domain waveform coding module 505, be used for receiving low frequency sub-band territory excitation spectrum from low frequency sub-band territory time-frequency conversion module 504, it is carried out quantization encoding, obtain low frequency sub-band domain waveform coded data, output to bit stream Multiplexing module 507 as the sound coding data of acoustic coding code stream.
High-frequency sub-band field parameter coding module 506, be used for receiving low frequency sub-band territory excitation spectrum from low frequency sub-band territory time-frequency conversion module 504, receive high-frequency sub-band territory signal from analyzing sub-filter pack module 501, according to low frequency sub-band territory excitation spectrum and high-frequency sub-band territory signal extraction high-frequency sub-band field parameter, this high-frequency sub-band field parameter is used for recovering high-frequency sub-band territory signal in decoding end from low frequency sub-band territory excitation spectrum, the high-frequency sub-band field parameter of 506 pairs of extractions of this high-frequency sub-band field parameter coding module carries out quantization encoding then, and the high-frequency sub-band field parameter coded data that obtains is outputed to bit stream Multiplexing module 507 as side information.
Bit stream Multiplexing module 507, be used for becoming will be from low frequency sub-band territory signal type analysis module 502, low frequency sub-band territory the time sound coding data that forecast analysis module 503, low frequency sub-band domain waveform coding module 505 and high-frequency sub-band field parameter coding module 506 receive and side information and carry out multiplexingly, form the acoustic coding code stream.
Because low frequency sub-band territory signal and high-frequency sub-band territory signal all comprise the subband more than 2 respectively, therefore each subband of 502 pairs of low frequency sub-band territories of signal type analysis module, above-mentioned low frequency sub-band territory signal carries out the signal type analysis; Each subband that becomes 503 pairs of low frequency sub-band territories of forecast analysis module signal during the low frequency sub-band territory carries out forecast analysis; Each subband of 504 pairs of low frequency sub-band territories of time-frequency conversion module, low frequency sub-band territory pumping signal carries out time-frequency conversion; Each subband of 505 pairs of low frequency sub-band territories of low frequency sub-band domain waveform coding module excitation spectrum carries out quantization encoding, exports to bit stream Multiplexing module 507 with the code stream form; High-frequency sub-band field parameter coding module 506 obtains the high-frequency sub-band field parameter of each subband and exports to bit stream Multiplexing module 507 with the code stream form.Following detailed description in the present embodiment is just no longer distinguished description to each subband.
Below, become forecast analysis module 503, low frequency sub-band territory time-frequency conversion module 504, low frequency sub-band domain waveform coding module 505, high-frequency sub-band field parameter coding module 506 during to the low frequency sub-band territory of above-mentioned monophonic sounds code device and specifically explain.
When being low frequency sub-band territory among Fig. 5, Fig. 6 becomes the structured flowchart of forecast analysis module 503.Becoming forecast analysis module 503 during the low frequency sub-band territory is made of sub-frame division device 600, linear prediction analysis device 601, converter 602, vector quantizer 603, inverse converter 604 and linear prediction filter 605.Particularly, at first by the low frequency sub-band territory signal type analysis result of sub-frame division device 600 according to reception, the branch subframe is carried out in low frequency sub-band territory signal y (n) of input to be handled, one frame y (n) is divided into one or more, as to be used for linear prediction analysis subframe, be that unit carries out linear prediction analysis to low frequency sub-band territory signal y (n) by linear prediction analysis device 601 with the subframe again, obtain one group of predictive coefficient a i, should organize a by converter 602 then iConvert one group of line spectral frequencies (LSF) to, should organize LSF again and send into and carry out vector quantization in the vector quantizer 603 and obtain LSF vector quantization index, and one group of line spectral frequencies after obtaining quantizing according to LSF vector quantization index
Figure A20071008887800361
With this group that obtains
Figure A20071008887800362
Process inverse converter 604 is obtained one group of predictive coefficient after the quantification
Figure A20071008887800363
605 pairs of linear prediction filters that constitutes with this group predictive coefficient after quantizing carry out filtering through the subframe signal that sub-frame division device 600 carries out the sub-frame division processing at last, obtain low frequency sub-band territory pumping signal x e(n), with the frame be unit output.If be divided into a plurality of subframes, then subframe is merged into frame output by stripe sequence.In the forecast analysis process, adopted vector quantization, above-mentioned LSF vector quantization index carries out linear prediction when comprehensive in decoding end, will be as the information that constitutes the linear prediction synthesis filter, so, at coding side, this LSF vector quantization index outputs in the bit stream Multiplexing module 507 as side information.Each carries out respectively corresponding its LSF vector quantization index separately of subframe of linear prediction analysis.
Wherein, when division was used for the subframe of linear prediction analysis, if signal type is a tempolabile signal, a frame was only divided a subframe, does not promptly divide subframe; If signal type is a fast changed signal, then this frame low frequency sub-band territory signal is divided into different subframes according to fast height position, specifically can be: several the 256 sampling point frames before the fast height be divided into a subframe, the 256 sampling point frames that comprise fast height are divided into a subframe, and several the 256 sampling point frames after fast height takes place are divided into a subframe.If the monophonic sounds code device does not comprise low frequency sub-band territory signal type analysis module 502 in accordance with the principles of the present invention, then become prediction module 503 during the low frequency sub-band territory and do not comprise sub-frame division device 600.
The method that low frequency sub-band territory time-frequency conversion module 504 is carried out time-frequency conversion can adopt discrete Fourier transform (DFT) (DFT, Discrete Fourier Transform), discrete cosine transform (DCT, DiscreteCosine Transform), correction discrete cosine transform transform methods such as (MDCT).When concrete time-frequency conversion, according to low frequency sub-band territory signal type analysis result low frequency sub-band territory pumping signal is repartitioned subframe, one frame low frequency sub-band territory pumping signal is divided into one or more, as to be used for time-frequency conversion subframe, low frequency sub-band territory pumping signal is carried out conversion by subframe.
Wherein, when division is used for the subframe of conversion,, then be divided into a subframe, and select the conversion of longer exponent number if signal type is a tempolabile signal; If signal type is a fast changed signal, then is divided into different subframes or is divided into the subframe of equal length, and select the conversion of shorter exponent number according to fast height position.When dividing subframe according to fast height position, division methods becomes 503 divisions of forecast analysis module during with the low frequency sub-band territory, and to be used for the subframe method of linear prediction analysis identical.
The method that 505 pairs of low frequency sub-band territories of low frequency sub-band domain waveform coding module excitation spectrum carries out quantization encoding can adopt the scalar among the similar MPEG AAC to add the quantization scheme that Huffman (Huffman) is encoded, and also can adopt the vector quantization scheme.When adopting the vector quantization scheme, the codebook vector word indexing that produces in the vector quantization process sends to bit stream Multiplexing module 507 (not shown among Fig. 5) as side information, and this codebook vector word indexing will be as the information of decoding end decoded low frequency subband domain waveform coding data.
Fig. 7 is the structured flowchart of high-frequency sub-band field parameter coding module 506 shown in Figure 5, high-frequency sub-band field parameter coding module 506 comprise signal type analyzer 701, the time become predictive analyzer 702, spectrum parametric encoder 703, the time become prediction synthesizer 704 and parameter extractor 705 is adjusted in the time-domain adaptive gain.
Wherein, signal type analyzer 701, be used for carrying out the signal type analysis to being received from every vertical frame dimension frequency subband domain signal of analyzing sub-filter pack module 501, become when high-frequency sub-band territory signal type analysis result is outputed to predictive analyzer 702, the time become prediction synthesizer 704 and parameter extractor 705 is adjusted in the time-domain adaptive gain, and output to bit stream Multiplexing module 507 as side information.The method of high-frequency sub-band territory signal being carried out the signal type analysis is identical with the method principle that low frequency sub-band territory signal type analysis module 502 carries out the signal type analysis.In practice, based on also can not comprising signal type analyzer 701 in the high-frequency sub-band field parameter coding module of the present invention.
The time become predictive analyzer 702, the high-frequency sub-band territory signal that is used for 501 outputs of receiving and analyzing sub-filter pack module, and according to high-frequency sub-band territory signal type analysis result from signal type analyzer 701, with the high-frequency sub-band territory division of signal that receives is one or more subframe, each subframe is one group of high-frequency sub-band territory sample of signal, is the unit of forecast analysis.Here,, then be divided into a subframe, can think that also not making to divide subframe handles, if fast changed signal then according to fast height occurrence positions, is divided into more than one subframe if high-frequency sub-band territory signal type analysis result is a tempolabile signal.Then each subframe is carried out linear prediction filtering, obtain high-frequency sub-band territory pumping signal, or claim high-frequency sub-band territory residual signals, then the high-frequency sub-band territory pumping signal that obtains is outputed to spectrum parametric encoder 703.This time become predictive analyzer 702 concrete composition structure become the composition structure of forecast analysis module 503 can adopt low frequency sub-band territory shown in above-mentioned Fig. 6 the time, according to the high-frequency sub-band territory signal of input through dividing subframe, linear prediction analysis, predictive coefficient inverse conversion to the conversion of line spectral frequencies, vector quantization, line spectral frequencies to predictive coefficient, generate linear prediction filter at last, adopt this linear prediction filter by subframe high-frequency sub-band territory signal to be carried out filtering, obtaining high-frequency sub-band territory pumping signal the most at last is unit output with the frame.If carried out sub-frame division, the subframe after then will handling is merged into frame output by stripe sequence.In the vector quantization process therein, become prediction synthesizer 704 when the high frequency LSF vector quantization index that generates is outputed to, and output to (not shown among Fig. 7) in the bit stream Multiplexing module 507 as side information.For with low frequency sub-band territory signal processing in the LSF vector quantization index that generates distinguish, will be called high frequency LSF vector quantization index to the LSF vector quantization index that high-frequency sub-band territory signal obtains through linear forecast analysis.
Spectrum parametric encoder 703, be used for from the time become predictive analyzer 702 and receive vertical frame dimension subband domain pumping signal frequently, receive low frequency sub-band territory excitation spectrums from low frequency sub-band territory time-frequency conversion module 504; Part corresponding with above-mentioned low frequency sub-band territory excitation spectrum in the pumping signal of high-frequency sub-band territory is made discrete Fourier transform (DFT) (DFT) obtain high-frequency sub-band territory excitation spectrum, according to gained high-frequency sub-band territory excitation spectrum described low frequency sub-band territory excitation spectrum is composed adjustment then, the adjustment of this spectrum comprises tonality adjustment and gain adjustment; To compose the adjustment parameter at last and behind quantization encoding, output to bit stream Multiplexing module 507, and will compose adjusted low frequency sub-band territory excitation spectrum and make contrary discrete Fourier transform (DFT) (IDFT), obtain composing becoming when adjusted low frequency sub-band territory pumping signal outputs to and predict synthesizer 704.In this module, except DFT can also adopt other time-frequency conversion method to obtain high-frequency sub-band territory excitation spectrum to high-frequency sub-band territory pumping signal.In like manner, the frequency-time domain transformation corresponding with above-mentioned time-frequency conversion also adopted in inverse transformation.
This spectrum parametric encoder 703 comprises DFT transducer 703a, spectrum tonality adjustment and parameter extractor 703b, spectrum gain adjustment and parameter extractor 703c and IDFT transducer 703d.Wherein, the vertical frame dimension frequency subband domain pumping signal that will become predictive analyzer 702 when DFT transducer 703a will be received from is divided into the subframe of the one or more DFT of being used for conversion, division methods is identical with the method for dividing low frequency sub-band territory pumping signal, and the basis of division can be the low frequency sub-band territory signal type analysis result (not shown among Fig. 5) from low frequency sub-band territory signal type analysis module 502; By subframe high-frequency sub-band territory pumping signal is obtained the frequency spectrum of this subframe as DFT, i.e. high-frequency sub-band territory excitation spectrum, and export the high-frequency sub-band territory excitation spectrum of gained to spectrum tonality adjustment and parameter extractor 703b and spectrum gain and adjust and parameter extractor 703c;
Spectrum tonality adjustment and parameter extractor 703b, be used for high-frequency sub-band territory excitation spectrum according to the subframe of DFT transducer 703a output, the low frequency sub-band territory excitation spectrum that receives from low frequency sub-band territory time-frequency conversion module 504 is composed the tonality adjustment, and the parameter of the spectrum tonality adjustment that will obtain outputs to bit stream Multiplexing module 507 as side information behind quantization encoding, will compose the adjusted low frequency sub-band of tonality territory excitation spectrum and export the spectrum gain to and adjust and parameter extractor 703c;
The spectrum gain is adjusted and parameter extractor 703c, be used for the high-frequency sub-band territory excitation spectrum of basis from the subframe of DFT transducer 703a, to composing gain adjustment from the adjusted low frequency sub-band of the spectrum tonality territory excitation spectrum of the correspondence of composing tonality adjustment and parameter extractor 703b, and will compose the parameter that gain adjusts and behind quantization encoding, output to bit stream Multiplexing module 507 as side information, will compose the adjusted low frequency sub-band territory excitation spectrum that gains and export IDFT transducer 703d to;
IDFT transducer 703d, be used for the adjusted low frequency sub-band territory excitation spectrum that gains from the spectrum of composing gain adjustment and parameter extractor 703c is carried out the IDFT conversion, obtain composing adjusted low frequency sub-band territory pumping signal, described spectrum adjustment comprises spectrum tonality adjustment and spectrum gain adjustment; Then the adjusted low frequency sub-band of the spectrum of each subframe territory pumping signal is reassembled into the adjusted low frequency sub-band of the spectrum territory pumping signal of a whole frame, becomes prediction synthesizer 704 when this signal is outputed to.
The time become prediction synthesizer 704, according to high-frequency sub-band territory signal type analysis result from signal type analyzer 701, be reclassified as one or more subframe to composing adjusted low frequency sub-band territory pumping signal from a frame of spectrum parameter coding module 703, each subframe is one group and composes adjusted low frequency sub-band territory pumping signal sample, is the unit of integrated filter; Again according to from the time become predictive analyzer 701 high frequency LSF vector quantization index obtain the linear prediction synthesis filter, and the adjusted low frequency sub-band of the spectrum of each subframe territory pumping signal carried out integrated filter, obtain rebuilding high-frequency sub-band territory signal, this signal is that unit exports time-domain adaptive gain adjustment parameter extractor 705 to the frame;
Parameter extractor 705 is adjusted in the time-domain adaptive gain, according to high-frequency sub-band territory signal type analysis result from signal type analyzer 701, will from the time become prediction synthesizer 704 the reconstruction high-frequency sub-band territory division of signal of every frame be one or more subframe, each subframe is one group and rebuilds high-frequency sub-band territory sample of signal, is the unit that the time domain gain is adjusted; And receive from the high-frequency sub-band territory signal of analyzing sub-filter pack module 501, the time domain gain of the high-frequency sub-band territory signal of the time domain gain of the reconstruction high-frequency sub-band territory signal of each subframe and this subframe correspondence is compared, obtain the time domain gain and adjust parameter, behind quantization encoding, output to bit stream Multiplexing module 507.Described time domain gain refers to the mean value of signal energy in the subframe.
Therefore, the high-frequency sub-band field parameter of high-frequency sub-band field parameter coding module 506 extractions comprises that spectrum is adjusted parameter, parameter, high frequency LSF vector quantization index and high-frequency sub-band territory signal type analysis result are adjusted in the time domain gain.These high-frequency sub-band field parameters generate high-frequency sub-band field parameter coded data behind quantization encoding, output to bit stream Multiplexing module 507 as side information.
Bit stream Multiplexing module 507 will be received from the sound coding data of above-mentioned each module and side information and carry out multiplexingly, form the acoustic coding code stream.Wherein, sound coding data is the low frequency sub-band domain waveform coded data of low frequency sub-band domain waveform coding module 505 outputs.Side information comprises the low frequency sub-band territory signal type analysis result of low frequency sub-band territory signal type analysis module 502 outputs, become the LSF vector quantization index of forecast analysis module 503 outputs during the low frequency sub-band territory, and the high-frequency sub-band field parameter coded data of high-frequency sub-band field parameter coding module 506 outputs.This high-frequency sub-band field parameter coded data comprises that the spectrum of quantization encoding is adjusted parameter, parameter, high frequency LSF vector quantization index and high-frequency sub-band territory signal type analysis result are adjusted in the time domain gain.When low frequency sub-band domain waveform coding module adopted vector quantization coding, side information also comprised the codebook vector word indexing.
The high-frequency sub-band field parameter coding module 506 of this preferred embodiment extracts the high-frequency sub-band field parameter that is used for recovering from low frequency sub-band territory excitation spectrum high-frequency sub-band territory signal according to high-frequency sub-band territory signal and low frequency sub-band territory excitation spectrum.In practice, high-frequency sub-band field parameter coding module 506 becomes the low frequency sub-band territory pumping signal that forecast analysis module 503 generates when receiving the low frequency sub-band territory, and does not receive low frequency sub-band territory excitation spectrum.In this case, high-frequency sub-band field parameter coding module 506 can not composed adjustment, be that high-frequency sub-band field parameter coding module 506 does not comprise spectrum parametric encoder 703, but directly be used for from the high-frequency sub-band field parameter of high-frequency sub-band territory pumping signal recovery high-frequency sub-band territory signal according to low frequency sub-band territory pumping signal and the signal extraction of high-frequency sub-band territory, this high-frequency sub-band field parameter does not comprise spectrum adjustment parameter.
Fig. 8 is the coding method process flow diagram based on monophonic sounds code device of the present invention.As shown in Figure 8, this method may further comprise the steps:
Step 11: the monophonic signal to input carries out sub-band division, makes the corresponding specific frequency range of each subband signal; Monophonic sound tone signal after the sub-band division can simply be divided into low frequency sub-band territory signal and high-frequency sub-band territory signal by frequency range.Wherein, low frequency sub-band territory signal comprises the subband domain signal of at least 2 subbands; High-frequency sub-band territory signal also comprises the subband domain signal of at least 2 subbands.
Step 12: low frequency sub-band territory signal is carried out the signal type analysis, if gradual type signal then directly is defined as signal type low frequency sub-band territory signal type analysis result; If become type signal soon, then continue to calculate the position that fast height takes place, at last signal type and fast height position are defined as low frequency sub-band territory signal type analysis result.
Step 13: according to low frequency sub-band territory signal type analysis result, low frequency sub-band territory signal is carried out the branch subframe handle, carry out forecast analysis by subframe, i.e. linear prediction filtering obtains low frequency sub-band territory pumping signal.
Step 14: according to low frequency sub-band territory signal type analysis result, adopt the different length exponent number that low frequency sub-band territory pumping signal is carried out time-frequency conversion, obtain low frequency sub-band territory excitation spectrum.
Step 15: low frequency sub-band territory excitation spectrum is carried out quantization encoding, obtain low frequency sub-band domain waveform coded data.
Step 16: be used for from the high-frequency sub-band field parameter of low frequency sub-band territory excitation spectrum recovery high-frequency sub-band territory signal according to low frequency sub-band territory excitation spectrum and the signal extraction of high-frequency sub-band territory, the high-frequency sub-band field parameter is carried out quantization encoding, obtain high-frequency sub-band field parameter coded data.
Step 17: above-mentioned low frequency sub-band domain waveform coded data and side information are carried out multiplexing, obtain the acoustic coding code stream.Wherein, side information comprises low frequency sub-band territory signal type analysis result and high-frequency sub-band field parameter coded data, also comprises the LSF vector quantization index that generates in the linear prediction filtering.
Wherein, because low frequency sub-band territory signal and high-frequency sub-band territory signal all comprise the subband more than 2 respectively, so each subband of 12 pairs of low frequency sub-band territories of above-mentioned steps signal carries out the signal type analysis; Each subband of step 13 pair low frequency sub-band territory signal carries out forecast analysis; Each subband of step 14 pair low frequency sub-band territory pumping signal carries out time-frequency conversion; Each subband of step 15 pair low frequency sub-band territory excitation spectrum carries out quantization encoding; Step 16 calculates the high-frequency sub-band field parameter of each subband.Following detailed description in the present embodiment is just no longer distinguished description to each subband.
Below the step among Fig. 8 is elaborated.
The Methods of Subband Filter Banks of step 11 can adopt technology such as QMF, PQMF, CMF and wavelet transform.Preferably, with PQMF be the process of example explanation sub-band division.
The impulse response of analysis filter is among the PQMF:
h k ( n ) = p 0 ( n ) cos ( π 2 M ( 2 k + 1 ) ( t - N 2 - M 2 ) )
Wherein, p 0(n) be prototype filter, M is a sub band number, and N is the prototype filter exponent number, and M 〉=4,0≤k＜M-1,0≤n＜2KM-1, N=2KM.
The step of carrying out sub-band division with PQMF is specific as follows:
Step 11a: will import the data translation of PQMF buffer zone, then data shift out buffer zone in the buffer zone, and shifting out the data number is M, identical with sub band number, that is:
x[n]＝x[n-M]，n＝N-1～M
Wherein, x[n] be pulse code modulation (PCM) (PCM, the Pulse Code Modulate) sample of input audio signal;
Step 11b:M newly-increased sample input block obtains to treat filter vector
Step 11c: the data in the buffer zone are carried out windowing process, obtain new vector
Figure A20071008887800423
That is:
u[n]＝x[n]×c[n]，n＝0～N-1
Wherein, c[n] be window function;
Step 11d: compute vector
Figure A20071008887800431
In the weighted sum z (t) of each element:
z ( t ) = Σ s = 0 k - 1 u [ 2 M × s + t ] , t = 0 ~ 2 M - 1
Step 11e: calculate subband data y k(n):
y k ( n ) = Σ t = 0 2 M - 1 z ( t ) × cos ( π 2 M ( 2 k + 1 ) ( t - N 2 - M 2 ) ) , k = 0 ≈ K - 1
Wherein, 0≤k＜M-1,0≤n＜2KM-1.
Low frequency sub-band territory signal type in the step 12 judges that many kinds of methods are arranged, and for example, judges signal type by the signal perceptual entropy, judges signal type etc. by the energy of signal calculated subframe.Preferably, adopt the method for judging signal type by the signal calculated subframe energy, detailed process is as follows:
Step 12a: a frame low frequency sub-band territory signal y (n) is carried out high-pass filtering, and with low frequency part, for example the following frequency of 500Hz filters out;
Step 12b: the signal after the high-pass filtering is divided into several subframes yi (n), is convenience of calculation, usually a frame signal is divided into an integer subframe, when being at 2048 as a frame, can 256 be a subframe;
Step 12c: calculate the ENERGY E i of each subframe yi (n) respectively, wherein i is the sequence number of subframe.Obtain the energy ratio of current subframe and last subframe again, when energy than greater than certain threshold value Te the time, judge that then this frame signal type is a fast changed signal, if the energy of all subframes and former frame than all less than Te the time, judges that then this frame signal type is a tempolabile signal.If fast changed signal then continues execution in step 12d, otherwise execution in step 12d not is defined as low frequency sub-band territory signal type analysis result with gradual signal type.Threshold value Te in the method can adopt the well-known process in some signal Processing to obtain, and as the mean ratio of statistics coded signal energy, and multiply by certain constant and obtains Te;
Step 12d:, the subframe of energy maximum is judged as the position of fast height generation for fast changed signal.Signal type that becomes soon and the position that fast height takes place are defined as low frequency sub-band territory signal type analysis result.
If need not divide subframe then need not execution in step 12.
In the step 13, for the frame signal on each low frequency sub-band territory, if this frame signal type is a fast changed signal, then the low frequency sub-band territory signal basis with this frame is divided into subframe in the position of height soon, specifically can be that several 256 sampling point frames before the fast height are divided into a subframe, comprise that 256 sampling point frames of fast height are divided into a subframe, several the 256 sampling point frames after fast height takes place be divided into a subframe, otherwise, be divided into a subframe and do not do the sub-frame division processing in other words conj.or perhaps.Then this frame signal in low frequency sub-band territory is carried out linear prediction filtering by subframe, obtain the low frequency sub-band territory pumping signal on each low frequency sub-band territory of this frame and the LSF vector quantization index of each subframe.
Wherein, carrying out the linear prediction analysis of p rank for the N point low frequency sub-band territory signal y (n) of a subframe handles and comprises following steps:
Step 13a: calculate the coefficient of autocorrelation r (k) of current subframe low frequency sub-band territory signal y (n), r ( k ) = Σ n = k N y ( n ) × y ( n - k ) , k ∈ [ 0 , p ] ;
Step 13b: adopt the coefficient of autocorrelation r (k) that obtains, carry out the Levinson-Durbin algorithm by recursion and obtain one group of predictive coefficient a i, and by predictive coefficient a iConstitute linear prediction filter A ( z ) = 1 - Σ i - 1 p a i z - i ;
Step 13c: by to two polynomial expressions f 1 ( z ) = A ( z ) + z - p A ( z - 1 ) f 2 ( z ) = A ( z ) - z - p A ( z - 1 ) Rooting is with a iConvert one group of line spectrum pair LSP to i, and by line spectrum pair LSP iObtain line spectral frequencies LSF i
Step 13d: line spectral frequencies is carried out vector quantization, the line spectral frequencies after obtaining quantizing
Figure A20071008887800444
And be converted to line spectrum pair after the quantification
Figure A20071008887800445
LSF vector quantization index is outputed in the bit stream Multiplexing module 507 as side information, be used for generating the linear prediction synthesis filter at the decoding device end;
Step 13e: by the line spectrum pair after quantizing
Figure A20071008887800446
By calculating f 1(z) and f 2(z) obtain coefficient of linear prediction wave filter after the quantification
Figure A20071008887800447
And the wave filter after the formation quantification A ^ ( z ) = 1 - Σ i = 1 P a ^ i z - i ;
Step 13f: low frequency sub-band territory signal y (n) is calculated low frequency sub-band territory pumping signal after the prediction by the wave filter after quantizing e ( n ) = y ( n ) - Σ i = 1 P a ^ i y ( n - i ) .
In the step 14, the method for low frequency sub-band territory pumping signal being carried out time-frequency conversion has a lot, as discrete Fourier transform (DFT) (DFT), discrete cosine transform (DCT), correction discrete cosine transform (MDCT) etc.Preferably, with discrete Fourier transform (DFT) (DFT) and correction discrete cosine transform (MDCT) be the process of example explanation time-frequency conversion.
Carry out the situation of time-frequency conversion for adopting discrete Fourier transform (DFT) (DFT), at first according to the signal type analysis result present frame being carried out the branch subframe handles, choose the low frequency sub-band territory pumping signal of M+N sample from the reference position of the current subframe of pumping signal, use x here e(n) expression, wherein M is current sub-frame data length, N is the overlapping length of next subframe.The length of M and N is determined by the signal type of present frame: when signal type is tempolabile signal, be divided into a subframe, M and N select long exponent number, suppose that in the present embodiment a frame length is 2048, then this moment M=2048, N=256; N can be decided to be 1/8 of M; When signal type is fast changed signal, can divide subframe according to the position of fast height, with the length of subframe as M, N=M/8 perhaps is divided into a frame isometric a plurality of subframes, and M and N select short exponent number, in the present embodiment a frame is divided into 8 isometric subframes, M=256 then, N=32.Again the low frequency sub-band territory pumping signal of M+N sample is carried out windowing operation, obtain the signal x after the windowing w(n)=w (n) x e(n).Wherein w (n) is a window function, can use various window functions in realization, for example can adopt Cosine Window, promptly
w ( n ) = cos ( n&pi; 2 N 0 ) , 0 &le; n &le; N 0 1 , N 0 &le; n &le; M cos ( ( n - M ) &pi; 2 N ) , M < n &le; M + N
Wherein, N 0Be the overlapping length of present frame, determine by the signal type of previous frame.Then the signal after the process windowing is carried out the DFT conversion, thereby obtains M+N low frequency sub-band territory excitation spectrum coefficient, X ( k ) = &Sigma; n = 0 M + N - 1 x w ( n ) e - j ( 2 &pi; / ( M + N ) ) kn , k∈[0，M+N-1]
For adopting the situation that time-frequency conversion is carried out in discrete cosine transform (MDCT) of revising, at first choose the low frequency sub-band territory pumping signal of a last subframe M sample and a current subframe M sample, again the low frequency sub-band territory pumping signal of common 2M the sample of this two subframe is carried out the windowing operation, then the signal after the process windowing is carried out the MDCT conversion, thereby obtain M low frequency sub-band territory excitation spectrum coefficient.
The impulse response of MDCT analysis filter is:
h k ( n ) = w ( n ) 2 M cos [ ( 2 n + M + 1 ) ( 2 k + 1 ) &pi; 4 M ] ,
Then MDCT is transformed to: X ( k ) = &Sigma; n = 0 2 M - 1 x e ( n ) h k ( n ) , 0≤k≤M-1, wherein: w (n) is a window function; x e(n) be the input low frequency sub-band territory pumping signal of MDCT conversion; X (k) is the low frequency sub-band territory excitation spectrum of the output of MDCT conversion.
For satisfying the condition of the complete reconstruct of signal, the window function w (n) of MDCT conversion must satisfy following two conditions:
W (2M-1-n)=w (n) and w 2(n)+w 2(n+M)=1.
In practice, can select for use the Sine window as window function.Certainly, also can revise above-mentioned restriction with specific analysis filter and synthesis filter by using biorthogonal conversion to window function.
Like this, these frame data that adopt MDCT to carry out time-frequency conversion just obtain different time-frequency plane figure according to signal type.For example, suppose that the time-frequency conversion exponent number when present frame is tempolabile signal is 2048, the time-frequency conversion exponent number during for the fast changed signal type is 256, then time-frequency plane figure as shown in Figure 9, wherein Fig. 9 a is the time-frequency plane figure of tempolabile signal; Fig. 9 b is the time-frequency plane figure of fast changed signal.
In the step 15, the waveform quantization coding that low frequency sub-band territory excitation spectrum is carried out can adopt the scalar among the similar MPEGAAC to add the quantization scheme of Huffman encoding, also can adopt the vector quantization scheme.In the constant bit rate coding, vector quantizer is a rational selection scheme.In a specific implementation method of the present invention, this preferred embodiment has adopted 8 n dimensional vector n quantization schemes.The vector quantization module is carried out the quantification of 8 n dimensional vector ns to the pumping signal frequency spectrum in low frequency sub-band territory.The result of vector quantization outputs to bit stream Multiplexing module 507.
The method of vector quantization specifically may further comprise the steps: at first the frequency coefficient on the low frequency sub-band territory is constituted a plurality of 8 n dimensional vector n signals; Then estimate criterion and in code book, search code word with vector distance minimum to be quantified, obtain its codebook vector word indexing by all direction search method according to perceived distance.This codebook vector word indexing adds the acoustic coding code stream, is used in decoding end decoded low frequency subband domain waveform coding data.Wherein perceived distance can adopt Euclidean distance to estimate.
In the step 16, high-frequency sub-band field parameter coding is a kind of method that is used for recovering the high-frequency sub-band field parameter of high-frequency sub-band territory signal according to low frequency sub-band territory excitation spectrum and the signal extraction of high-frequency sub-band territory.Coding method comprises following steps to the high-frequency sub-band field parameter among the present invention:
Step 16-1: the high-frequency sub-band territory signal that step 11 is obtained carries out the signal type analysis, determines high-frequency sub-band territory signal type analysis result;
Wherein, the signal type analysis of high-frequency sub-band territory comprises: the type of judging described high-frequency sub-band territory signal is tempolabile signal or fast changed signal, if fast changed signal is then further analyzed the position that fast height takes place, determines high-frequency sub-band territory signal type analysis result.Here, judgement, analytical approach are identical with the signal type analytical approach of low frequency sub-band territory signal, see the specific implementation process of step 12.
Step 16-2: according to high-frequency sub-band territory signal type analysis result, with high-frequency sub-band territory division of signal is subframe one or more, that be used for forecast analysis, carry out linear prediction filtering by subframe, obtain the high frequency LSF vector quantization index and the high-frequency sub-band territory pumping signal of high-frequency sub-band territory signal.Its acquisition methods is identical with the method for LSF vector quantization index that calculates low frequency sub-band territory signal and low frequency sub-band territory pumping signal, sees the specific implementation process of step 13.
Step 16-3: high-frequency sub-band territory pumping signal is made the time-frequency conversion of different rank, obtain high-frequency sub-band territory excitation spectrum.For example can adopt the DFT conversion.Exponent number when the exponent number of time-frequency conversion carries out time-frequency conversion with step 14 pair low frequency sub-band territory pumping signal is identical, and the high-frequency sub-band territory excitation spectrum that makes this step obtain is corresponding with the low frequency sub-band territory excitation spectrum that step 14 obtains.See the specific implementation process of step 14.
Step 16-4: according to high-frequency sub-band territory excitation spectrum low frequency sub-band territory excitation spectrum is composed spectrum tonality adjustment in the adjustment, and obtain composing tonality and adjust low frequency sub-band territory excitation spectrum after parameter and the adjustment of spectrum tonality.
Wherein, the adjustment of spectrum tonality comprises:
Step 16-4a: high-frequency sub-band territory excitation spectrum and low frequency sub-band territory excitation spectrum are divided into several frequency ranges, and calculate the tonality T of the high-frequency sub-band territory excitation spectrum of each frequency sub-band RefAnd the tonality T of the low frequency sub-band territory excitation spectrum of corresponding band Est, and judge T EstWith T RefSize, if T Est＜T Ref-T 0, then adjust type for adding the string adjustment, and execution in step 16-4b; If T Ref-T Ref-T 0≤ T Est≤ T 1, then to adjust type and be set to not carry out the tonality adjustment, the spectrum gain that directly goes to step 16-5 is adjusted; If T Est＞T Ref+ T 1, then adjust type for adding the adjustment of making an uproar, and execution in step 16-4c; T wherein 0, T 1Be the constant that sets in advance;
Step 16-4b: adjust type and be set to add the string adjustment, will adjust type and add the string factor p ^ = T ref - T est 1 + T est Adjust parameter as the spectrum tonality, with adding the string energy &Delta; E T = E est &CenterDot; ( T ref - T est ) 1 + T est The spectrum gain that described low frequency sub-band territory excitation spectrum adds the string adjustment and goes to step 16-5 is adjusted;
Step 16-4c: adjust type and be set to add the adjustment of making an uproar, will adjust type and add the factor of making an uproar p ^ = T ref - T est T ref &CenterDot; ( 1 + T est ) Adjust parameter as the spectrum tonality, with adding the energy of making an uproar &Delta; E N = E est &CenterDot; ( T est - T ref ) T ref &CenterDot; ( 1 + T est ) Described low frequency sub-band territory excitation spectrum is added the spectrum gain of adjusting and go to step 16-5 of making an uproar to be adjusted.
Step 16-5: adjust according to the spectrum gain that high-frequency sub-band territory excitation spectrum is composed in the adjustment spectrum tonality adjusted low frequency sub-band territory excitation spectrum, and obtain composing gain and adjust parameter and the spectrum gain low frequency sub-band territory excitation spectrum after adjusting.Its step is as follows:
Step 16-5a: high-frequency sub-band territory excitation spectrum and the adjusted low frequency sub-band of spectrum tonality territory excitation spectrum are divided into plurality of sub-bands, high-frequency sub-band territory excitation spectrum and the adjusted low frequency sub-band of spectrum tonality territory excitation spectrum for each sub-band calculate its energy respectively;
Step 16-5b: for any one frequency band, calculate the high-frequency sub-band territory excitation spectrum of this frequency band and the energy ratio of the adjusted low frequency sub-band of spectrum tonality territory excitation spectrum, the square root of described energy ratio is adjusted parameter as the frequency domain gain of this frequency band;
Step 16-5c: to any one sub-band of the adjusted low frequency sub-band of spectrum tonality territory excitation spectrum, every spectral line in this sub-band be multiply by the spectrum gain corresponding with this sub-band respectively adjust parameter, obtain the adjusted low frequency sub-band of spectrum gain territory excitation spectrum; And the set of the spectrum gain of each sub-frequency bands being adjusted parameter is adjusted parameter as composing gain.
Step 16-6: carry out frequency-time domain transformation to composing adjusted low frequency sub-band territory excitation spectrum, obtain adjusted low frequency sub-band territory pumping signal.Wherein spectrum adjustment comprises spectrum tonality adjustment and composes gain adjustment.Corresponding to the time-frequency conversion of step 16-3, this step adopts the IDFT conversion.
Step 16-7: according to high-frequency sub-band territory signal type analysis result, adjusted low frequency sub-band territory pumping signal is carried out integrated filter, obtain rebuilding high-frequency sub-band territory signal.This step specifically comprises:
Step 16-7a: according to high-frequency sub-band territory signal type analysis result, with the adjusted low frequency sub-band territory pumping signal of a frame be divided into one or more, be used to predict comprehensive subframe, its division methods is identical with sub-frame division method principle among the step 16-2;
Step 16-7b: to each subframe, the interior one group of high frequency LSF vector quantization index of this subframe that adopts step 16-2 to obtain, make the line spectral frequencies that inverse vector quantizes to obtain vector quantization, and this line spectral frequencies is converted into coefficient of linear prediction wave filter behind the vector quantization, and constitute the linear prediction synthesis filter by these coefficients of gained;
Step 16-7c: by subframe, the linear prediction synthesis filter of adjusted low frequency sub-band territory pumping signal by this subframe correspondence carried out integrated filter, obtain the high-frequency sub-band territory signal that a frame is rebuild.
Step 16-8: according to high-frequency sub-band territory signal type analysis result, and the gain of the time domain of the high-frequency sub-band territory signal of described reconstruction is adjusted parameter with the time domain gain extraction time domain gain of the high-frequency sub-band territory signal of step 11 acquisition.Specific practice is: according to described high-frequency sub-band territory signal type analysis result, be one or more subframe with the reconstruction high-frequency sub-band territory division of signal of each frame of gained; Calculate the ratio of average energy of the high-frequency sub-band territory signal of the average energy of reconstruction high-frequency sub-band territory signal of each subframe and this subframe correspondence, the square root of described ratio is adjusted parameter as the time domain gain of adjusting this subframe.
Step 16-9: after will comprising that high-frequency sub-band territory signal type analysis result that step 16-1 obtains, high frequency LSF vector quantization index that step 16-2 obtains, the time domain gain that step 16-6 gained spectrum is adjusted parameter and step 16-8 gained are adjusted the high-frequency sub-band field parameter of parameter and carried out quantization encoding, output.
In this preferred embodiment, if low frequency sub-band territory pumping signal and high-frequency sub-band territory signal extraction high-frequency sub-band field parameter that step 16 directly obtains according to step 13, then in leaching process, do not comprise the step that spectrum is adjusted, so do not comprise spectrum adjustment parameter in the high-frequency sub-band field parameter.
Below introduce the monophonic sound sound decoding device and the method for the preferred embodiment of the present invention, because decode procedure is the inverse process of cataloged procedure, so only simply introduce decode procedure.Figure 10 is the structured flowchart as the monophonic sound sound decoding device of the preferred embodiment of the present invention.The sound decoding device of the preferred embodiment of the present invention comprises: become the comprehensive module 1004 of prediction, high-frequency sub-band field parameter decoder module 1005 and comprehensive sub-filter pack module 1006 when bit stream demultiplexing module 1001, low frequency sub-band domain waveform decoder module 1002, low frequency sub-band territory frequency-time domain transformation module 1003, low frequency sub-band territory.
Below, the annexation of summarized introduction each module shown in Figure 10 and function.
Bit stream demultiplexing module 1001 is used for the acoustic coding code stream that receives is carried out demultiplexing, obtains the sound coding data and the side information of corresponding data frame, to the corresponding sound coding data of low frequency sub-band domain waveform decoder module 1002 outputs; When low frequency sub-band territory frequency-time domain transformation module 1003, low frequency sub-band territory, become comprehensive module 1004 of prediction and the corresponding side information of high-frequency sub-band field parameter decoder module 1005 outputs.
Wherein, the data that output to low frequency sub-band domain waveform decoder module 1002 are low frequency sub-band domain waveform coded data; The side information that outputs to low frequency sub-band territory frequency-time domain transformation module 1003 is a low frequency sub-band territory signal type analysis result; The side information that becomes the comprehensive module 1004 of prediction when outputing to the low frequency sub-band territory comprises low frequency sub-band territory signal type analysis result and LSF vector quantization index; The side information that outputs to high-frequency sub-band field parameter decoder module 1005 is a high-frequency sub-band field parameter coded data, comprising: the high-frequency sub-band territory signal type analysis result of quantization encoding, high frequency LSF vector quantization index, spectrum adjust parameter and parameter is adjusted in the time domain gain.
Because what this bit stream demultiplexing module 1001 received is the acoustic coding code stream that the monophonic sounds code device shown in Fig. 5 sends, and therefore this acoustic coding code stream demultiplexing is obtained the coded data that low frequency sub-band domain waveform coded data comprises at least 2 subbands; High-frequency sub-band territory coded data comprises the coded data of at least 2 subbands; Low frequency sub-band territory signal type analysis result and LSF vector quantization index also comprise the corresponding data of at least 2 subbands respectively.
Low frequency sub-band domain waveform decoder module 1002, be used for the low frequency sub-band domain waveform coded data that is received from bit stream demultiplexing module 1001 is carried out the re-quantization decoding, and the low frequency sub-band territory excitation spectrum that will obtain behind the re-quantization outputs to low frequency sub-band territory frequency-time domain transformation module 1003 and high-frequency sub-band field parameter decoder module 1005.Here, the method for re-quantization decoding is to adopt the inverse process of quantization encoding in the coding side medium and low frequency subband domain waveform coding module 505.
Low frequency sub-band territory frequency-time domain transformation module 1003, be used for the low frequency sub-band territory excitation spectrum that is received from low frequency sub-band domain waveform decoder module 1002 is carried out frequency-time domain transformation, frequency-time domain transformation is according to the low frequency sub-band territory signal type analysis result in the bit stream demultiplexing module 1001 output side informations, adopt the conversion of different length exponent number, obtaining low frequency sub-band territory pumping signal, is that unit becomes the comprehensive module 1004 of prediction when exporting the low frequency sub-band territory to the frame.The method of frequency-time domain transformation is the inverse process of time-frequency conversion in the time-frequency conversion module 504 of coding side low frequency sub-band territory, comprise contrary discrete Fourier transform (DFT) (IDFT, Inverse Discrete FourierTransform), inverse discrete cosine transform (IDCT, Inverse Discrete Cosine Transform), against revising discrete cosine transform (IMDCT, Inverse Modified Discrete Cosine Transform) etc.
Become the comprehensive module 1004 of prediction during the low frequency sub-band territory, according to the low frequency sub-band territory signal type analysis result in the bit stream demultiplexing module 1001 output side informations, a frame low frequency sub-band territory pumping signal that is received from low frequency sub-band territory frequency-time domain transformation module 1003 is carried out the branch subframe to be handled, predict comprehensively by subframe then, obtain this subframe low frequency sub-band territory signal.If this frame signal is a fast changed signal, then the position with the fast height of this frame low frequency sub-band territory pumping signal basis is divided into more than one subframe, and the low frequency sub-band territory pumping signal of corresponding subframe is predicted comprehensively according to each the subframe LSF vector quantization index in the side information, obtain the low frequency sub-band territory signal of each subframe, the low frequency sub-band territory signal combination with each subframe becomes this frame low frequency sub-band territory signal at last; If tempolabile signal is divided into a subframe, be equivalent to not divide subframe and handle, therefore this frame low frequency sub-band territory pumping signal is predicted comprehensively according to the LSF vector quantization index of this frame, obtain the low frequency sub-band territory signal of this frame.
High-frequency sub-band field parameter decoder module 1005 is used for the high-frequency sub-band field parameter coded data that is received from bit stream demultiplexing module 1001 is carried out re-quantization, obtains the high-frequency sub-band field parameter; Low frequency sub-band territory excitation spectrum and high-frequency sub-band field parameter according to 1002 outputs of low frequency sub-band domain waveform decoder module recover high-frequency sub-band territory signal.When recovering high-frequency sub-band territory signal, recover the high-frequency sub-band territory signal of a subband according to the low frequency sub-band territory excitation spectrum of a subband according to the high-frequency sub-band field parameter; The high-frequency sub-band territory signal of a plurality of subbands can adopt the low frequency sub-band territory excitation spectrum of same subband to recover, be high-frequency sub-band territory signal, be used for the high frequency LSF vector quantization index difference of restoring signal for the different sub-band of the low frequency sub-band territory excitation spectrum recovery of adopting same subband.
Comprehensive sub-filter pack module 1006, it is synthetic that the high-frequency sub-band territory signal of the low frequency sub-band territory signal of comprehensive module 1004 outputs of change prediction and 1005 outputs of high-frequency sub-band field parameter decoder module carries out subband when being used for the low frequency sub-band territory, with the monophonic sound tone signal of acquisition decoding.
Corresponding with coding side, each subband of 1002 pairs of low frequency sub-band domain waveforms of low frequency sub-band domain waveform decoder module coded data carries out the re-quantization decoding; Each subband of 1003 pairs of low frequency sub-band territories of frequency-time domain transformation module, low frequency sub-band territory excitation spectrum carries out frequency-time domain transformation; Each subband that becomes 1004 pairs of low frequency sub-band territories of the comprehensive module of prediction pumping signal during the low frequency sub-band territory is predicted comprehensively; High-frequency sub-band field parameter decoder module 1005 obtains the high-frequency sub-band territory signal of corresponding subband according to the high-frequency sub-band field parameter coded data of each subband.The following detailed content of present embodiment is just no longer distinguished description to each subband.
Below, become the comprehensive module 1004 of prediction during to the low frequency sub-band territory of above-mentioned monophonic sound sound decoding device and high-frequency sub-band field parameter decoder module 1005 specifically explains.
Become the structured flowchart of the comprehensive module 1004 of prediction when Figure 11 is low frequency sub-band shown in Figure 10 territory, this module is formed by separating vector quantizer 1101, converter 1102 and linear prediction synthesizer 1103.Wherein, separate the LSF vector quantization index that vector quantizer 1101 receives the subframe in the side informations, solve line spectral frequencies after the quantification by LSF vector quantization index
Figure A20071008887800521
Export converter 1102 to.Line spectral frequencies after the quantification that converter 1102 will receive
Figure A20071008887800522
Be converted to one group of coefficient of linear prediction wave filter after the quantification
Figure A20071008887800523
With these coefficient input linear prediction synthesizers 1103.Linear prediction synthesizer 1103 is according to the coefficient that receives
Figure A20071008887800524
-* constitutes the linear prediction synthesis filter
Figure A20071008887800525
According to the low frequency sub-band territory signal type analysis result in the side information, the low frequency sub-band territory pumping signal that is received from low frequency sub-band territory frequency-time domain transformation module 1003 is divided into subframe again, adopts the linear prediction synthesis filter
Figure A20071008887800526
By subframe low frequency sub-band territory pumping signal is carried out integrated filter, the low frequency sub-band territory signal that is restored.When comprehensively being coding side low frequency sub-band territory, prediction becomes the inverse process of forecast analysis in the forecast analysis module 503.
Figure 12 is the structured flowchart of high-frequency sub-band field parameter decoder module 1005 shown in Figure 10, it comprise spectrum parameter decoder 1201, adaptive time domain gain demoder 1202 and the time become prediction synthesizer 1203.
Wherein, spectrum parameter decoder 1201, at first obtain low frequency sub-band territory excitation spectrum from low frequency sub-band domain waveform decoder module 1002, adjusting parameter (comprising that tonality is adjusted parameter and parameter is adjusted in gain) according to the spectrum that obtains from bit stream demultiplexing module 1001 adjusts accordingly to low frequency sub-band territory excitation spectrum, then adjusted low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation such as IDFT, obtain adjusted low frequency sub-band territory pumping signal and output to adaptive time domain gain demoder 1202.
This spectrum parameter decoder 1201 comprises spectrum tonality adjuster 1201a, spectrum fader 1201b and IDFT transducer 1201c.Wherein, spectrum tonality adjuster 1201a, be used for receiving the spectrum tonality adjustment parameter of side information, adjust parameter according to the spectrum tonality and compose the tonality adjustment obtaining low frequency sub-band territory excitation spectrums from low frequency sub-band domain waveform decoder module 1002 from bit stream demultiplexing module 1001; Spectrum fader 1201b is used for adjusting parameter from the spectrum gain of bit stream demultiplexing module 1001 reception side informations, according to spectrum gain adjustment parameter the adjusted low frequency sub-band of spectrum tonality territory excitation spectrum is composed gain adjustment; IDFT transducer 1201c will compose gain adjusted low frequency sub-band territory excitation spectrum and carry out the IDFT conversion and obtain low frequency sub-band territory pumping signal, export adaptive time domain gain demoder 1202 to.
Adaptive time domain gain demoder 1202, high-frequency sub-band territory signal type analysis result according in the bit stream demultiplexing module 1001 output side informations will be divided into one or more subframes from low frequency sub-band territory pumping signal after the adjustment of spectrum parameter decoder 1201; The position that fast height when described high-frequency sub-band territory signal type analysis result comprises signal type and fast changed signal takes place, the division methods of subframe is identical with the sub-frame division method that the gain of the time-domain adaptive of coding side is adjusted in the parameter extractor 705; Adaptive time domain gain demoder 1202 is adjusted parameter according to the time domain gain of the corresponding subframe that obtains from bit stream demultiplexing module 1001 and the low frequency sub-band territory pumping signal of each subframe is carried out the time domain gain is adjusted then, and resulting time domain gained becomes prediction synthesizer 1203 when adjusted low frequency sub-band territory pumping signal exports to.
Described time domain gain adjusting method is: parameter is adjusted in the time domain gain after the quantification of the subframe that will carry out time domain gain adjustment that at first obtains from bit stream demultiplexing module 1001, de-quantization should be adjusted parameter, and the pumping signal in this subframe all be multiply by this adjustment parameter behind the de-quantization.The gained result is the time domain adjusted low frequency sub-band territory pumping signal that gains.
The time become prediction synthesizer 1203, to become a frame low frequency sub-band territory pumping signal according to its time sequential combination from the gain low frequency sub-band territory pumping signal of one or several subframes of demoder 1202 of adaptive time domain, and this frame low frequency sub-band territory pumping signal is divided subframe again according to the high-frequency sub-band territory signal type analysis result from bit stream demultiplexing module 1001 output side informations, time the method in the change predictive analyzer 702 of the division methods of subframe and coding side is identical; Constitute the linear prediction synthesis filter of each subframe again according to the high frequency LSF vector quantization index that from bit stream demultiplexing module 1001, obtains, and the low frequency sub-band territory pumping signal of corresponding subframe carried out integrated filter, obtain the reconstruction high-frequency sub-band territory signal of corresponding subframe.The method of integrated filter predicts that with the time change of coding side synthesizer 704 is identical; And the reconstruction high-frequency sub-band territory signal of the corresponding subframe of gained outputed to comprehensive sub-filter pack module 1006 as decoded high-frequency sub-band territory signal.
Corresponding with coding side, when not comprising spectrum, the high-frequency sub-band field parameter do not adjust parameter, then high-frequency sub-band field parameter decoder module 1005 also can obtain low frequency sub-band territory pumping signal from low frequency sub-band territory frequency-time domain transformation module 1003, therefore high-frequency sub-band field parameter decoder module 1005 is not composed adjustment, be that high-frequency sub-band field parameter decoder module 1005 does not comprise spectrum parameter decoder 1201 yet, the linear prediction synthesis filter that constitutes according to high frequency LSF vector quantization index only, the time domain adjusted low frequency sub-band territory pumping signal that gains is carried out integrated filter, thereby obtain the high-frequency sub-band territory signal of reconstruction, realized the recovery of high-frequency sub-band territory signal.
Figure 13 is the coding/decoding method process flow diagram based on monophonic sound sound decoding device of the present invention.As shown in figure 13, this method may further comprise the steps:
Step 21: the acoustic coding code stream is carried out demultiplexing, all side informations that obtain low frequency sub-band domain waveform coded data and decode used.
Step 22:, obtain the low frequency sub-band territory excitation spectrum behind the re-quantization to the low frequency sub-band domain waveform coded data re-quantization of decoding.
Step 23: the low frequency sub-band territory excitation spectrum behind the re-quantization is carried out frequency-time domain transformation, obtain low frequency sub-band territory pumping signal.
Step 24: obtain the linear prediction synthesis filter according to the LSF vector quantization index that from side information, reads, low frequency sub-band territory pumping signal is carried out integrated filter, recover low frequency sub-band territory signal.
Step 25: recover high-frequency sub-band territory signal according to low frequency sub-band territory excitation spectrum that obtains in the step 22 and the high-frequency sub-band field parameter coded data from side information.
Step 26: it is synthetic that the high-frequency sub-band territory signal combination that the low frequency sub-band territory signal that step 24 is obtained and step 25 obtain uses comprehensive Methods of Subband Filter Banks to carry out subband together, recovers the monophonic sound tone signal.
Corresponding with coding side, the low frequency sub-band domain waveform coded data that step 21 is obtained comprises the waveform coding data of at least two subbands, comprises the high-frequency parameter coded data of at least two subbands in the side information.Each subband of step 22 pair low frequency sub-band domain waveform coded data carries out the re-quantization decoding; Each subband of step 23 pair low frequency sub-band territory excitation spectrum carries out frequency-time domain transformation; Each subband of step 24 pair low frequency sub-band territory pumping signal is predicted comprehensively; Step 25 is obtained the high-frequency sub-band territory signal of corresponding subband according to the high-frequency sub-band field parameter coded data of each subband.The following detailed content of present embodiment is just no longer distinguished description to each subband.
Below each step among Figure 13 is elaborated.
In the step 22, low frequency sub-band domain waveform coding/decoding method is corresponding to the coding method of low frequency sub-band domain waveform.If the specific embodiments in low frequency sub-band territory signal encoding part is the method for the vector quantization of employing, then Dui Ying low frequency sub-band territory re-quantization need obtain the codebook vector word indexing from code stream, finds corresponding vector according to codewords indexes in fixed code book.Low frequency sub-band domain waveform vector is combined into low frequency sub-band territory excitation spectrum behind the re-quantization in order.
In the step 23,, be the process of example explanation frequency-time domain transformation with contrary discrete Fourier transform (DFT) (IDFT) and contrary correction discrete cosine transform (IMDCT) corresponding to the specific implementation method of monophonic sounds coded portion low frequency sub-band territory time-frequency conversion among the present invention.
For contrary discrete Fourier transform (DFT) (IDFT), the frequency-time domain transformation process comprises three steps: IDFT conversion, time-domain windowed are handled and time domain stack computing.
In the present embodiment, at first low frequency sub-band territory excitation spectrum is divided into the one or more subframe that is used for frequency-time domain transformation, by subframe low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation again according to the low frequency sub-band territory signal type analysis result in the side information.In division methods and the coding side low frequency sub-band territory time-frequency conversion module 504, adopt the division methods of DFT conversion identical.
Secondly, the low frequency sub-band territory excitation spectrum behind the re-quantization is carried out the IDFT conversion, obtain the time-domain signal x after the conversion I, nThe expression formula of IDFT conversion is: x i , n = 1 N + M &Sigma; k = 0 N + M - 1 X ( k ) e j ( 2 &pi; / ( M + N ) ) kn , Wherein, M represents current subframe sample number, and N represents next subframe interpenterating sample number, and the length of M and N is determined by the signal type of present frame, is consistent with the value in the scrambler low frequency sub-band territory time-frequency conversion module 504; N represents the sample sequence number, and 0≤n＜N+M, and i represents frame number; K represents to compose sequence number.
Then, the time-domain signal after the IDFT conversion is carried out windowing process, the data after the windowing, back N point keeps, and is next subframe superposition of data.Windowed function is corresponding with coding side.Cosine Window for example:
w ( n ) = cos ( n&pi; 2 N 0 ) , 0 &le; n < N 0 1 , N 0 &le; n &le; M cos ( ( n - M ) &pi; 2 N ) , M < n &le; M + N
Wherein, N 0Length for current subframe stack.Signal type by previous frame is determined.At last for the calculating that superposes of the preceding M point of the time-domain signal of top windowing.Computing method are for to preserve last N with last subframe 0N before some time domain data and the current subframe 0The time domain data stack of point, all the other M-N 0Point data is constant, and the gained data are the time-domain signal of current subframe behind the frequency-time domain transformation, i.e. low frequency sub-band territory pumping signal.
For contrary correction discrete cosine transform (IMDCT), the frequency-time domain transformation process comprises three steps: IMDCT conversion, time-domain windowed are handled and time domain stack computing.
In the present embodiment, at first low frequency sub-band territory excitation spectrum is divided into the one or more subframe that is used for frequency-time domain transformation according to the low frequency sub-band territory signal type analysis result in the side information.In division methods and the coding side low frequency sub-band territory time-frequency conversion module 504, the division methods that employing MDCT carries out time-frequency conversion is identical.
Secondly, low frequency sub-band territory excitation spectrum is carried out the IMDCT conversion, obtain the time-domain signal x after the conversion by subframe I, nThe expression formula of IMDCT conversion is: x i , n = 2 N &Sigma; k = 0 N 2 - 1 spec [ i ] [ k ] cos ( 2 &pi; N ( n + n 0 ) ( k + 1 2 ) ) , Wherein, n represents the sample sequence number, and 0≤n＜N, and N represents the time domain samples number, and value is 2048, n 0=(N/2+1)/2; I represents frame number; K represents to compose sequence number.
Then, the time-domain signal that the IMDCT conversion is obtained carries out windowing process in time domain.For satisfying complete reconstruction condition, window function w (n) must satisfy following two condition: w (2M-1-n)=w (n) and w 2(n)+w 2(n+M)=1
Typical window function has sine (Sine) window, Kai Saer-Bezier to derive (KBD, Kaiser-BesselDerived) window etc.Can utilize biorthogonal conversion in addition, adopt specific analysis filter and composite filter to revise above-mentioned restriction window function.
At last, above-mentioned windowing time-domain signal is carried out overlap-add procedure, obtain time-domain audio signal.Specifically:, obtain the time-domain audio sample of N/2 output, i.e. timeSam with preceding N/2 sample of the signal of acquisition and back N/2 sample overlap-add of last subframe signal after the windowing operation I, n=preSam I, n+ preSam I-1, n+N/2, wherein i represents frame number, n represents the sample sequence number, has 0 &le; n &le; N 2 . Thereby obtain low frequency sub-band territory pumping signal.
In the step 24, the prediction of low frequency sub-band territory pumping signal comprehensively is the inverse process of in the coded portion low frequency sub-band territory signal estimation being analyzed, and its effect is the low frequency sub-band territory signal after low frequency sub-band territory pumping signal is comprehensively obtained synthesizing by linear prediction.
Predicting when comprehensive, at first according to the low frequency sub-band territory signal type analysis result in the side information, with a frame low frequency sub-band territory pumping signal be reclassified as one or more, be used to predict comprehensive subframe, the division subframe method that division methods becomes forecast analysis module 503 during with coding side low frequency sub-band territory is identical.Be the low frequency sub-band territory pumping signal x of the present invention's realization below to a subframe e(n) predict comprehensive detailed process, may further comprise the steps:
Step 24a:, LSF vector quantization index is decoded as line spectral frequencies after the quantification, and is converted to line spectrum pair by the LSF vector quantization index that reads in the side information;
Step 24b: pass through to calculate f by the line spectrum pair after quantizing 1(z) and f 2(z) obtain predictive coefficient after the quantification
Figure A20071008887800572
And the linear prediction synthesis filter after the formation quantification 1 A ^ ( z ) = 1 1 - &Sigma; i = 1 p a ^ i z - i , wherein, p is prediction order, and is identical with coding side;
Step 24c, with low frequency sub-band territory pumping signal x e(n) the low frequency sub-band territory signal y (n) that obtains synthesizing by the linear prediction synthesis filter: y ( n ) = x e ( n ) + &Sigma; i = 1 p a ^ i y ( n - i ) .
In the step 25, high-frequency sub-band field parameter coding/decoding method comprises following steps:
Step 25-1: with the high-frequency sub-band field parameter coded data re-quantization decoding of side information, obtain the high-frequency sub-band field parameter, from the high-frequency sub-band field parameter, read the spectrum tonality and adjust parameter, and adjust parameter according to this spectrum tonality the low frequency sub-band territory excitation spectrum that step 22 obtains is composed the tonality adjustment.Wherein, spectrum tonality adjustment parameter comprises the adjustment type and adjusts parameter
Figure A20071008887800582
The tonality adjustment may further comprise the steps:
Step 25-1a: low frequency sub-band territory excitation spectrum is divided into one or more sub-band, and calculates the ENERGY E of each sub-band respectively Est
Step 25-1b: judge the adjustment type for each sub-band,, then this frequency band is not dealt with if adjust type for not adjusting; Handle for adding string if adjust type, then the center position at this frequency band adds string, adds the string energy to be &Delta; E T = E est &CenterDot; P ^ , To add the phase place and the former frame phase place of string continuous; If the adjustment type then adds random noise at this frequency band for adding the adjustment of making an uproar, add the energy of making an uproar and be &Delta; E N = E est &CenterDot; P ^ ; Dispose and then finish the adjustment of spectrum tonality behind all frequency bands.
Step 25-2 reads the spectrum gain and adjusts parameter from the high-frequency sub-band field parameter, and the adjusted low frequency sub-band of spectrum tonality territory excitation spectrum is composed gain adjustment, comprising:
Step 25-2a: low frequency sub-band territory excitation spectrum is divided into one or more sub-band;
Step 25-2b: for any sub-band, adjust parameter with the gain of the spectrum of this sub-band correspondence and multiply by every spectral line in this sub-band, the spectrum that obtains this sub-band adjusted low frequency sub-band territory excitation spectrum that gains is the adjusted low frequency sub-band of spectrum gain territory excitation spectrum altogether with all group of subbands.
Step 25-3: adjusted low frequency sub-band territory excitation spectrum is carried out the IDFT conversion, obtain low frequency sub-band territory pumping signal, transform method is identical with low frequency sub-band territory frequency-time domain transformation method, and embodiment is seen step 23.
Step 25-4: from the high-frequency sub-band field parameter, read the time domain gain and adjust parameter, according to the time domain gain adjustment parameter of correspondence gained low frequency sub-band territory pumping signal is carried out the time domain gain and adjust, and the adjusted low frequency sub-band of output gained territory pumping signal.Specific practice is: earlier low frequency sub-band territory pumping signal is divided into one or more subframe; Read time domain gain again and adjust parameter, the low frequency sub-band territory pumping signal of each subframe be multiply by corresponding time domain gain respectively adjust parameter, obtain the time domain adjusted low frequency sub-band territory pumping signal that gains.
Step 25-5: the gained time domain adjusted low frequency sub-band territory pumping signal that gains is carried out integrated filter, the high-frequency sub-band territory signal that obtains rebuilding.Specifically comprise:
Step 25-5a: read high-frequency sub-band territory signal type analysis result, and according to the high-frequency sub-band territory signal type analysis result that is read with the low frequency sub-band territory pumping signal of a frame be divided into one or more, be used to predict comprehensive subframe;
Step 25-5b:, read the linear prediction synthesis filter that high frequency LSF vector quantization index constitutes each subframe to described each subframe;
Step 25-5c: by described linear prediction synthesis filter low frequency sub-band territory pumping signal is carried out integrated filter, the high-frequency sub-band territory signal that obtains rebuilding has promptly recovered high-frequency sub-band territory signal, realizes the band spread decoding, and continues to carry out subsequent processing steps.
In the step 26, analyzing Methods of Subband Filter Banks corresponding to monophonic sounds coded portion of the present invention and adopt PQMF to carry out the specific embodiment of sub-band division, is the synthetic process of example explanation subband with PQMF.
The impulse response of synthesis filter is among the PQMF:
f k ( n ) = p 0 ( n ) cos ( &pi; 2 M ( 2 k + 1 ) ( t - N 2 + M 2 ) )
Wherein, p 0(n) be prototype filter, M is a sub band number, and N is the prototype filter exponent number, and 0≤k＜M-1,0≤n＜2KM-1, N=2KM.
Step with the synthetic time domain waveform signal of PQMF is specific as follows:
Step 26a: with the data translation in the buffer zone of PQMF, that is:
z[t]＝z[t-M]，t＝N-1～M
Step 26b: calculate M new coefficient, that is:
z [ t ] = 1 M &Sigma; k = 0 2 M - 1 y k [ t ] &times; cos ( &pi; 2 M ( 2 k + 1 ) ( t - N 2 + M 2 ) ) ,
Wherein, t=M-1～0.
Step 26c: generate weighing vector, that is:
u[Mn+k]＝z[2Mn+k]
u[Mn+64+k]＝z[2Mn+192+k]
Wherein, k=0～M-1, n=0～K.
Step 26d: the windowing of weighing vector data, that is:
W[n]=u[n] * c[n], n=0～N-1 wherein
Step 26e: the PCM sample that calculates monophonic sounds
x ^ ( n ) = &Sigma; k = 0 K - 1 w [ Mk + n ] , n = 0 ~ M
Step 26f: order is exported PCM sample, the monophonic sound tone signal that is restored.
In this preferred embodiment, step 25 is when recovering high-frequency sub-band territory signal, if do not compose in the high-frequency sub-band field parameter and adjust parameter, then this step 25 can be recovered high-frequency sub-band territory signal according to the high-frequency sub-band field parameter from the low frequency sub-band territory pumping signal that step 23 obtains, and does not also comprise spectrum adjustment in rejuvenation.
Below introduce the stereo encoding apparatus and the method for the preferred embodiment of the present invention.
Figure 14 is the structured flowchart as the stereo encoding apparatus of the preferred embodiment of the present invention, the stereo encoding apparatus of the preferred embodiment of the present invention comprises: become forecast analysis module 1403, low frequency sub-band territory time-frequency conversion module 1404, low frequency sub-band territory stereo coding module 1405, high-frequency sub-band field parameter coding module 1406 and bit stream Multiplexing module 1407 when analyzing sub-filter pack module 1401, low frequency sub-band territory and signal type analysis module 1402, low frequency sub-band territory.
Below, specifically introduce the annexation and the function of each module among Figure 14, wherein:
Analyze sub-filter pack module 1401, be used for left and right sound channels (L with the stereophonic signal of input, R) carry out sub-band division respectively, make all corresponding specific frequency range of each subband signal in two sound channels, generate the subband domain signal (L of each frequency range 1, R 1)～(L K1+k2, R K1+k2), wherein k1 is more than or equal to 2, and k2 is more than or equal to 2.Then with the low frequency sub-band signal (L in two sound channels 1, R 1)～(L K1, R K1) be that unit becomes forecast analysis module 1403 when being input to low frequency sub-band territory and signal type analysis module 1402 and low frequency sub-band territory with the frame, with the high-frequency sub-band signal (L in two sound channels K1+1, R K1+1)～(L K1, R K1) be that unit is input to high-frequency sub-band field parameter coding module 1406 with the frame.
Low frequency sub-band territory and signal type analysis module 1402, be used for from analyzing the low frequency sub-band territory signal that sub-filter pack module 1401 receives two sound channels, and go out low frequency sub-band territory and signal by the low frequency sub-band calculated signals of these two sound channels, the signal type analysis is carried out in this low frequency sub-band territory and signal, the type of judging this frame and signal is tempolabile signal or fast changed signal, if tempolabile signal, then direct output signal type, for example, the type of output this frame of expression and signal is gradual sign; If fast changed signal, then the position that fast height takes place is calculated in continuation, and exports the position of corresponding signal type and fast height generation.Become when the result of low frequency sub-band territory and signal type analysis outputs to the low frequency sub-band territory on the one hand and be used for sub-frame division control in the forecast analysis module 1403, output on the other hand in the low frequency sub-band territory time-frequency conversion module 1404 and be used for the exponent number of time-frequency conversion is controlled, simultaneously and the signal type analysis result also output in the bit stream Multiplexing module 1407 as the side information of acoustic coding code stream.In practice, stereo encoding apparatus can not comprise this module in accordance with the principles of the present invention.
Become forecast analysis module 1403 during the low frequency sub-band territory, be used for from analyzing the low frequency sub-band territory signal that sub-filter pack module 1401 receives two sound channels, and according to the low frequency sub-band territory and the signal type analysis result of low frequency sub-band territory and signal type analysis module 1402 outputs, low frequency sub-band signal in two sound channels that receive is carried out the branch subframe to be handled, by subframe the low frequency sub-band territory signal in two sound channels is carried out forecast analysis then, be linear prediction filtering, obtain the low frequency sub-band territory pumping signal in described two sound channels, the low frequency sub-band territory pumping signal in two sound channels that will obtain again outputs to low frequency sub-band territory time-frequency conversion module 1404.When output, subframe is merged framing output according to stripe sequence.
Wherein, when dividing subframe, and if signal type be tempolabile signal, a frame is only divided a subframe, does not promptly divide subframe; And if signal type is fast changed signal, then this frame left side/R channel low frequency sub-band territory signal is divided into different subframes according to fast height position.Become 503 pairs of monophonic sound tone signal of forecast analysis module when specifically the method for the different subframes of division is referring to the low frequency sub-band territory and carry out the concrete grammar that sub-frame division is handled.
Becoming forecast analysis module 1403 during the low frequency sub-band territory can adopt the composition structure in the monophonic sounds code device of the present invention by subframe low frequency sub-band territory signal to be carried out linear prediction analysis.The LSF vector quantization index of two sound channels obtaining in the linear prediction analysis process is sent to bit stream Multiplexing module 1407 as side information.
Low frequency sub-band territory time-frequency conversion module 1404, become the low frequency sub-band territory pumping signal that forecast analysis module 1403 receives described two sound channels when being used for from the low frequency sub-band territory, and according to the low frequency sub-band territory and the signal type analysis result of low frequency sub-band territory and signal type analysis module 1402 outputs, adopt the conversion of different length exponent number, respectively with the low frequency sub-band territory pumping signal in two sound channels by spatial transform to frequency domain, obtain the frequency domain representation of the low frequency sub-band territory pumping signal in described two sound channels, i.e. the low frequency sub-band territory excitation spectrum of two sound channels.The low frequency sub-band territory excitation spectrum of two sound channels that again time-frequency conversion obtained outputs to low frequency sub-band domain waveform coding module 1405 and high-frequency sub-band field parameter coding module 1406.If stereo encoding apparatus does not comprise low frequency sub-band territory and signal type analysis module 1402 in accordance with the principles of the present invention, then when time-frequency conversion not match exponents control.
Wherein, concrete time-frequency conversion method such as discrete Fourier transform (DFT) (DFT), discrete cosine transform (DCT), revise discrete cosine transform (MDCT) etc.And, when concrete time-frequency conversion, with the exponent number control of low frequency sub-band territory and signal type analysis result as time-frequency conversion.Fast changed signal is that time-frequency conversion is done by unit with the subframe, selects the conversion of shorter exponent number; Tempolabile signal is that time-frequency conversion is done by unit with the frame, selects the conversion of longer exponent number.The low frequency sub-band territory excitation spectrum of described two sound channels that time-frequency conversion is obtained outputs to low frequency sub-band territory stereo coding module 1405.
Low frequency sub-band territory stereo coding module 1405, be used for receiving the low frequency sub-band territory excitation spectrum of described two sound channels from low frequency sub-band territory time-frequency conversion module 1404, and excitation spectrum is divided into plurality of sub-bands, adopt the stereo coding pattern to carry out stereo coding to each sub-band respectively, obtain low frequency sub-band territory stereo coding data, and output to bit stream Multiplexing module 1407 as the sound coding data in the acoustic coding code stream.Wherein, the stereo coding pattern comprises and differs from stereo coding pattern, parameter stereo coding pattern and parameter error stereo coding pattern.When carrying out stereo coding, each sub-band is selected a kind of stereo coding that carries out in above-mentioned three kinds of coding modes.Output in the bit stream Multiplexing module 1407 as side information when wherein, coding mode is selected information of same.
High-frequency sub-band field parameter coding module 1406, be used for receiving the low frequency sub-band territory excitation spectrum of two sound channels from low frequency sub-band territory time-frequency conversion module 1404, from analyzing the high-frequency sub-band territory signal that Methods of Subband Filter Banks 1401 receives two sound channels, high-frequency sub-band field parameter according to two sound channels of high-frequency sub-band territory signal extraction of the low frequency sub-band territory excitation spectrum of two sound channels and two sound channels, this high-frequency sub-band field parameter is used for from the high-frequency sub-band territory signal of two sound channels of low frequency sub-band territory excitation spectrum recovery of two sound channels, after the high-frequency sub-band field parameter of 1406 pairs of extractions of this high-frequency sub-band field parameter coding module carries out quantization encoding then, obtain high-frequency sub-band field parameter coded data, this high-frequency sub-band field parameter coded data is outputed to bit stream Multiplexing module 1407 as side information.
Bit stream Multiplexing module 1407, be used for becoming will be from low frequency sub-band territory and signal type analysis module 1402, low frequency sub-band territory the time sound coding data that forecast analysis module 1403, low frequency sub-band territory stereo coding module 1405 and high-frequency sub-band field parameter coding module 1406 receive and side information and carry out multiplexingly, form stereosonic acoustic coding code stream.
In the present embodiment, becoming forecast analysis module 1403, low frequency sub-band territory time-frequency conversion module 1404, high-frequency sub-band field parameter coding module 1406 during the low frequency sub-band territory need handle respectively stereosonic left and right sound channels, and its disposal route is identical with resume module method of the same name in the monophonic sounds code device.Therefore, each module in above-mentioned three modules is all passed through the module combinations of the same name in two monophonic sounds code devices, thereby realizes stereosonic processing.
As seen, be with the monophonic sounds code device difference of the preferred embodiment of the present invention, when the monophonic sounds code device generates the sound coding data of acoustic coding code stream, employing be low frequency sub-band domain waveform coding module 505; And stereo encoding apparatus is when generating the sound coding data of acoustic coding code stream, employing be low frequency sub-band territory stereo coding module 1405.This module also is that each subband of low frequency sub-band territory stereo coding data is divided sub-band and stereo coding.
Certainly, identical with the high-frequency sub-band field parameter coding module 506 in the monophonic sounds code device, become the low frequency sub-band territory pumping signal of two sound channels of forecast analysis module 1403 generations when the high-frequency sub-band field parameter coding module 1406 of this preferred embodiment also can receive the low frequency sub-band territory, and do not receive the low frequency sub-band territory excitation spectrum of two sound channels.In this case, directly according to the high-frequency sub-band field parameter of two sound channels of high-frequency sub-band territory signal extraction of the low frequency sub-band territory pumping signal of two sound channels and two sound channels.This high-frequency sub-band field parameter is used for recovering from the high-frequency sub-band territory pumping signal of two sound channels the high-frequency sub-band territory signal of two sound channels, and this high-frequency sub-band field parameter does not comprise spectrum adjustment parameter.
Figure 15 is the coding method process flow diagram based on stereo encoding apparatus of the present invention.As shown in figure 15, this method may further comprise the steps:
Step 31: the left and right sound channels of stereophonic signal to input is carried out sub-band division respectively, makes the corresponding specific frequency range of each subband domain signal in two sound channels.Decomposition obtains the low frequency sub-band territory signal of left and right sound channels and the high-frequency sub-band territory signal of left and right sound channels, and wherein the low frequency sub-band territory signal of a left side/R channel comprises the subband domain signal of at least 2 subbands; The high-frequency sub-band territory signal of a left side/R channel also comprises the subband domain signal of at least 2 subbands.
Step 32: go out low frequency sub-band territory and signal by the low frequency sub-band territory calculated signals in two sound channels, the signal type analysis is carried out in this low frequency sub-band territory and signal, if gradual type signal then directly is defined as signal type low frequency sub-band territory and signal type analysis result; If become type signal soon, then continue to calculate the position that fast height takes place, at last signal type and fast height position are defined as low frequency sub-band territory and signal type analysis result.
Step 33: according to low frequency sub-band territory and signal type analysis result, low frequency sub-band territory signal in described two sound channels is carried out the branch subframe to be handled, carry out forecast analysis by subframe, i.e. linear prediction filtering, and obtain low frequency sub-band territory pumping signal in described two sound channels.
Step 34: according to low frequency sub-band territory and signal type analysis result, adopt the different length exponent number that the low frequency sub-band territory pumping signal in described two sound channels is carried out time-frequency conversion, obtain the low frequency sub-band territory excitation spectrum in described two sound channels.
Step 35: the low frequency sub-band territory excitation spectrum in described two sound channels is divided into plurality of sub-bands, each sub-band is carried out stereo coding, obtain low frequency sub-band territory stereo coding data.
Step 36: according to the high-frequency sub-band territory signal in the low frequency sub-band territory excitation spectrum in described two sound channels and described two sound channels, extraction is used for low frequency sub-band territory excitation spectrum from described two sound channels and recovers the high-frequency sub-band field parameter of the high-frequency sub-band territory signal in described two sound channels, the high-frequency sub-band field parameter is carried out quantization encoding, obtain high-frequency sub-band field parameter coded data.
Step 37: above-mentioned low frequency sub-band territory stereo coding data and side information are carried out multiplexing, obtain the acoustic coding code stream.Wherein, side information comprises the high-frequency sub-band field parameter coded data of low frequency sub-band territory and signal type analysis result and left and right sound channels, also comprises the LSF vector quantization index of two sound channels that low frequency sub-band territory signal is carried out generating in the linear prediction filtering.
Wherein, sub-band division method in the step 31, the signal type determination methods in the step 32, the forecast analysis processing in the step 33, the time-frequency conversion method in the step 34 and the high-frequency sub-band field parameter coding method in the step 36, all in the embodiment of the coding method of monophony code device of the present invention, introduced, in the embodiment of the coding method of stereo encoding apparatus of the present invention, adopt identical method, therefore do not do introduction.
Wherein, the process of the low frequency sub-band territory stereo coding of step 35 is, at first the low frequency sub-band territory excitation spectrum in described two sound channels is divided into plurality of sub-bands, to each sub-band promptly and select a kind ofly difference stereo coding pattern, parameter stereo coding pattern and the parameter error stereo coding pattern, the spectrum of the low-frequency excitation in two sound channels in this sub-band is encoded then from three kinds of coding modes.When dividing, respectively each subband of the low frequency sub-band territory excitation spectrum of two sound channels is divided.At first provide the implementation method that two kinds of coding modes are selected below:
Coding mode is selected implementation method 1: with identical bit number the low frequency sub-band territory excitation spectrum in described two sound channels is carried out Code And Decode with three kinds of coding modes respectively, the error of low frequency sub-band territory excitation spectrum in two sound channels that the calculating decoding recovers and the preceding low frequency sub-band territory of coding excitation spectrum, and the coding mode of Select Error minimum is as the coding mode of stereo coding.Select information to output in the bit stream Multiplexing module 1407 coding mode as side information;
Coding mode is selected implementation method 2: the lower frequency sub-band that is lower than a determined value for frequency in the low frequency sub-band territory, the following sub-band of 1kHz for example, adopt respectively and differ from the stereo coding pattern and the parameter stereo coding pattern is carried out Code And Decode, calculate the error of the low frequency sub-band territory excitation spectrum in two sound channels recovering and the preceding low frequency sub-band territory excitation spectrum of encoding, and the less coding mode of Select Error, select information to output in the bit stream Multiplexing module 1407 coding mode as side information, the upper frequency sub-band that is higher than above-mentioned determined value for frequency, as the sub-band more than the 1kHz, adopt the parameter stereo coding pattern.At this moment, bit stream Multiplexing module 1407 can be exported or do not exported to the selection information of parameter stereo coding pattern.
Certainly, also can adopt fixing stereo coding pattern in actual applications, in this case, not need to select information to output in the bit stream Multiplexing module 1407 coding mode as side information.
Respectively the implementation method of three kinds of stereo coding patterns is elaborated below.
Figure 16 is and differs from the illustraton of model of stereo coding pattern.With differing from the stereo coding pattern is according to the low frequency sub-band territory excitation spectrum in the sub-band in described two sound channels, calculates one and excitation spectrum and a poor excitation spectrum in this sub-band.Specific implementation method is as follows:
Excitation spectrum by left and right sound channels With
Figure A20071008887800662
Calculate corresponding and excitation spectrum
Figure A20071008887800663
With the difference excitation spectrum
Figure A20071008887800664
And will
Figure A20071008887800665
With
Figure A20071008887800666
After carrying out the waveform quantization coding, with what obtain With
Figure A20071008887800668
Output to bit stream Multiplexing module 1407 as low frequency sub-band territory stereo coding data.
Figure A20071008887800669
With
Figure A200710088878006610
Calculating formula be:
M &RightArrow; = ( L &RightArrow; + R &RightArrow; ) / 2
S &RightArrow; = ( L &RightArrow; - R &RightArrow; ) / 2
Wherein, right
Figure A200710088878006613
With
Figure A200710088878006614
Carrying out the waveform quantization coding can adopt 505 pairs of low frequency sub-band territories of low frequency sub-band domain waveform coding module excitation spectrum of monophonic sounds code device to carry out the method for quantization encoding.
Figure 17 is the illustraton of model of parameter stereo coding pattern.The parameter stereo coding pattern is according to the low frequency sub-band territory excitation spectrum in the sub-band k in described two sound channels, calculate a monaural excitation spectrum in this sub-band k, calculate the parameter that is used for recovering the low frequency sub-band territory excitation spectrum in this sub-band k of described two sound channels simultaneously by this sub-band monophony excitation spectrum.Enumerate the specific implementation method of two kinds of parameter stereo codings below.
Parameter stereo coding specific implementation method 1 comprises following steps:
Step 35-1a: in sub-band k, for certain sound channel, as R channel
Figure A200710088878006615
Calculate the weighting parameters g of this sound channel rAnd obtain the excitation spectrum of this sound channel behind the convergent-divergent (k), R &RightArrow; &prime; = 1 g r ( k ) R &RightArrow; , Make behind the convergent-divergent
Figure A200710088878006617
With
Figure A200710088878006618
Energy equate; g r(k) computing method can adopt following formula:
g r ( k ) = E R ( k ) E L ( k )
Wherein, E R(k) and E L(k) be respectively the energy of R channel in the sub-band k, L channel.
Step 35-1b:, calculate the weighted sum excitation spectrum of this Frequency point for each the Frequency point i in the sub-band k
Figure A200710088878006620
With the weighted difference excitation spectrum
Figure A200710088878006621
Because behind convergent-divergent, the energy of the left and right acoustic channels of each Frequency point is upward approximate more identical than statistics in the sub-band k, so by
Figure A200710088878006622
With Energy approximation equates, so the weighted sum excitation spectrum
Figure A20071008887800671
With the weighted difference excitation spectrum
Figure A20071008887800672
Near normal.Computing formula is as follows:
M &RightArrow; &prime; = ( L &RightArrow; + R &RightArrow; &prime; ) / 2 = [ L &RightArrow; + 1 g r ( k ) R &RightArrow; ] / 2
S &RightArrow; &prime; = ( L &RightArrow; - R &RightArrow; &prime; ) / 2
Step 35-1c: produce and the weighted sum excitation spectrum
Figure A20071008887800675
The quadrature excitation spectrum that constant amplitude is vertical Compose according to quadrature excitation
Figure A20071008887800677
With the weighted difference excitation spectrum
Figure A20071008887800678
Calculating quadrature excitation spectrum
Figure A20071008887800679
Weighting parameters g d(k), the feasible g that adopts d(k) spectrum of the quadrature excitation behind the convergent-divergent
Figure A200710088878006710
With Energy equate.g d(k) computing method can adopt following formula:
g d ( k ) = E S ( k ) E D ( k )
Wherein, E S(k) and E D(k) be respectively weighted difference excitation spectrum in the sub-band k Compose with quadrature excitation
Figure A200710088878006714
Energy.
Step 35-1d: above-mentioned weighted sum excitation spectrum
Figure A200710088878006715
And g r(k) and g d(k) behind quantization encoding, output to bit stream Multiplexing module 1407 respectively.Wherein, behind the quantization encoding
Figure A200710088878006716
Be low frequency sub-band territory stereo coding data, the g behind the quantization encoding r(k) and g d(k) be side information.
With respect to specific implementation method 1, the parameter g in the parameter stereo coding specific implementation method 2 r(k), g d(k) and the weighted sum excitation spectrum
Figure A200710088878006717
Obtain according to the minimum principle of error, comprise following steps:
Step 35-2a:,, calculate first parameter g according to following formula for sub-band k d(k):
g d ( k ) = - b ( k ) + b 2 ( k ) + a 2 ( k ) a ( k )
Wherein,
a ( k ) = &Sigma; i &Element; band ( k ) ( x r [ i , k ] y l [ i , k ] - x l [ i , k ] y r [ i , k ] ) ,
b ( k ) = &Sigma; i &Element; band ( k ) ( x l [ i , k ] x r [ i , k ] + y l [ i , k ] y r [ i , k ] )
Wherein, x lAnd y lBe respectively the real part and the imaginary part of L channel low frequency sub-band territory excitation spectrum, x rAnd y rBe respectively the real part and the imaginary part of R channel low frequency sub-band territory excitation spectrum;
Step 35-2b:,, calculate second parameter g according to following formula for sub-band k r(k):
g r ( k ) = - ( c ( k ) - d ( k ) ) + ( c ( k ) - d ( k ) ) 2 + g ( k ) m 2 ( k ) g ( k ) m 2 ( k )
Wherein,
c ( k ) = &Sigma; i &Element; band ( k ) ( x l [ i , k ] x l [ i , k ] + y l [ i , k ] y l [ i , k ] ) ;
d ( k ) = &Sigma; i &Element; band ( k ) ( x r [ i , k ] x r [ i , k ] + y r [ i , k ] y r [ i , k ] ) ;
m ( k ) = 2 b ( k ) ( 1 - g d 2 ( k ) ) + 2 a ( k ) g d ( k ) 1 + g d 2 ( k )
Step 35-2c: each the Frequency point i in the sub-band k calculates the weighted sum excitation spectrum according to following formula
Figure A20071008887800685
x m [ i , k ] = x l [ i , k ] + g d ( k ) y l [ i , k ] + g ( k ) g r ( k ) ( x r [ i , k ] - g d ( k ) y r [ i , k ] ) ( 1 + g d 2 ( k ) ) ( 1 + g ( k ) g r 2 ( k )
y m [ i , k ] = - g d ( k ) x l [ i , k ] + y l [ i , k ] + g ( k ) g r ( k ) ( g d ( k ) x r [ i , k ] + y r [ i , k ] ) ( 1 + g d 2 ( k ) ) ( 1 + g ( k ) g r 2 ( k )
Wherein, x mAnd y mRepresent the weighted sum excitation spectrum respectively
Figure A20071008887800688
Real part and imaginary part, M &RightArrow; &prime; [ i , k ] = x m [ i , k ] + j y m [ i , k ] ; G (k) is the importance factors of sub-band k intrinsic parameter stereo coding, has reflected the distribution of parameter stereo coding error at left and right acoustic channels, can select according to characteristics of signals, and for example g (k) can equal that the ratio of L channel and the energy of R channel is E in the sub-band k L(k)/E R(k).
Step 35-2d: above-mentioned weighted sum excitation spectrum g r(k) and g d(k) behind quantization encoding, output to bit stream Multiplexing module 1407 respectively.Wherein, behind the quantization encoding
Figure A200710088878006811
Be low frequency sub-band territory stereo coding data, the g behind the quantization encoding r(k) and g d(k) be side information.
Figure 18 is the illustraton of model of parameter error stereo coding pattern.Parameter error stereo coding pattern is according to the low frequency sub-band territory excitation spectrum in the sub-band in described two sound channels, the parameter that calculates a monaural excitation spectrum in this sub-band, an error spectrum and compose the low frequency sub-band territory excitation spectrum in the sub-band that recovers in described two sound channels by this monophony excitation spectrum, error.
Compared to the computation model of parameter stereo coding pattern, improve encoding precision if desired and then adopt parameter error stereo coding pattern, further calculate the error of excitation spectrum, i.e. the error spectrum
Figure A20071008887800691
And error composed
Figure A20071008887800692
Also carry out the waveform quantization coding.The specific implementation method of parameter error stereo coding pattern may further comprise the steps:
Step 35-3a: for certain sound channel in the sub-band k, as R channel
Figure A20071008887800693
Calculate the weighting parameters g of this sound channel rAnd obtain the excitation spectrum of this sound channel behind the convergent-divergent (k),
Figure A20071008887800694
Because the energy of the left and right acoustic channels of each Frequency point i is upward approximate more identical than statistics in the parameter extraction frequency band, so
Figure A20071008887800695
With
Figure A20071008887800696
Energy approximation equates, so the weighted sum excitation spectrum
Figure A20071008887800697
With the weighted difference excitation spectrum
Figure A20071008887800698
Near normal; Wherein, g r(k) g among computing method and the step 35-1a r(k) computing method are identical.
Step 35-3b:, calculate the weighted sum excitation spectrum of this Frequency point for each the Frequency point i in this sub-band
Figure A20071008887800699
With the weighted difference excitation spectrum
Step 35-3c: produce and the weighted sum excitation spectrum
Figure A200710088878006911
The quadrature excitation spectrum that constant amplitude is vertical
Figure A200710088878006912
Step 35-3d: compose according to quadrature excitation
Figure A200710088878006913
With the weighted difference excitation spectrum
Figure A200710088878006914
Calculate weighting parameters g dAnd obtain (k), according to g d(k) spectrum of the quadrature excitation behind the convergent-divergent
Figure A200710088878006915
Wherein, g d(k) g among computing method and the step 35-1c d(k) computing method are identical.
Step 35-3e: by calculating the weighted difference excitation spectrum
Figure A200710088878006916
With the quadrature excitation spectrum behind the convergent-divergent
Figure A200710088878006917
Difference can obtain the error excitation spectrum
Figure A200710088878006918
Promptly E &RightArrow; = S &RightArrow; &prime; - D &RightArrow; &prime; .
Step 35-3f: above-mentioned weighted sum excitation spectrum
Figure A200710088878006920
The error excitation spectrum
Figure A200710088878006921
Parameter g r(k) and g d(k) behind quantization encoding, output to bit stream Multiplexing module 1407 respectively.Wherein, behind the quantization encoding
Figure A200710088878006922
With
Figure A200710088878006923
Be low frequency sub-band territory stereo coding data, the g behind the quantization encoding r(k) and g d(k) be side information.
In this preferred embodiment, if the low frequency sub-band territory pumping signal of two sound channels that step 36 directly obtains according to step 33 and the high-frequency sub-band territory signal extraction high-frequency sub-band field parameter of two sound channels, then in leaching process, do not comprise the step that spectrum is adjusted, therefore do not comprise spectrum adjustment parameter in the high-frequency sub-band field parameter of two sound channels.
Below introduce the stereo decoding apparatus and the method for the preferred embodiment of the present invention.
Figure 19 is the structured flowchart as the stereo decoding apparatus of the preferred embodiment of the present invention.As shown in figure 19, the stereo decoding apparatus of the preferred embodiment of the present invention comprises: become the comprehensive module 1904 of prediction, high-frequency sub-band field parameter decoder module 1905 and comprehensive sub-filter pack module 1906 when bit stream demultiplexing module 1901, low frequency sub-band territory stereo decoding module 1902, low frequency sub-band territory frequency-time domain transformation module 1903, low frequency sub-band territory.
Below, specifically introduce the annexation and the function of each module shown in Figure 19, wherein,
Bit stream demultiplexing module 1901 is used for the acoustic coding code stream that receives is carried out demultiplexing, obtains the sound coding data and the side information of corresponding data frame, to the corresponding sound coding data of low frequency sub-band territory stereo decoding module 1902 outputs; When low frequency sub-band territory frequency-time domain transformation module 1903, low frequency sub-band territory, become comprehensive module 1904 of prediction and the corresponding side information of high-frequency sub-band field parameter decoder module 1905 outputs.
Wherein, the sound coding data that outputs to low frequency sub-band territory stereo decoding module 1902 is low frequency sub-band territory stereo coding data; The side information that outputs to low frequency sub-band territory frequency-time domain transformation module 1903 is low frequency sub-band territory and signal type analysis result; The side information that becomes the comprehensive module 1904 of prediction when outputing to the low frequency sub-band territory comprises low frequency sub-band territory and signal type analysis result and low frequency sub-band territory LSF vector quantization index; The side information that outputs to high-frequency sub-band field parameter decoder module 1905 is a high-frequency sub-band field parameter coded data, comprising: high-frequency sub-band territory signal type analysis result, high frequency LSF vector quantization index, spectrum adjust parameter and parameter is adjusted in the time domain gain.Exported coding mode when the low frequency sub-band territory of coding side stereo coding module 1405 and selected information, coding mode selects information also will export low frequency sub-band territory stereo decoding module 1902 (not shown among Figure 19) to as side information.
Low frequency sub-band territory stereo decoding module 1902, be used for selecting information that low frequency sub-band territory stereo coding data are carried out stereo decoding according to the coding mode of bit stream demultiplexing module 1901 output side informations, obtain the low frequency sub-band territory excitation spectrum in described two sound channels, send to low frequency sub-band territory frequency-time domain transformation module 1903 and high-frequency sub-band field parameter decoder module 1905.
Low frequency sub-band territory frequency-time domain transformation module 1903, be used for the low frequency sub-band territory excitation spectrum of two sound channels being received from low frequency sub-band territory stereo decoding module 1902 is carried out frequency-time domain transformation, frequency-time domain transformation is according to low frequency sub-band territory and signal type analysis result in the bit stream demultiplexing module 1901 output side informations, adopt the conversion of different length exponent number, obtaining the low frequency sub-band territory pumping signal in described two sound channels, is that unit becomes the comprehensive module 1904 of prediction when sending to the low frequency sub-band territory with the frame.The method of frequency-time domain transformation is the inverse process of time-frequency conversion in the time-frequency conversion module 1404 of coding side low frequency sub-band territory, comprises contrary discrete Fourier transform (DFT) (IDFT), inverse discrete cosine transform (IDCT), contrary correction discrete cosine transform (IMDCT) etc.
Become the comprehensive module 1904 of prediction during the low frequency sub-band territory, be used for low frequency sub-band territory and signal type analysis result according to bit stream demultiplexing module 1901 output side informations, with a frame low frequency sub-band territory excitation spectrum that is received from low frequency sub-band territory frequency-time domain transformation module 1903 be divided into one or more, be used to predict comprehensive subframe, each subframe is carried out integrated filter, obtain the low frequency sub-band territory signal of each subframe, form a frame low frequency sub-band territory signal according to stripe sequence.The sub-frame division method that becomes forecast analysis module 1403 during the low frequency sub-band territory of division methods and stereo encoding apparatus is identical.
High-frequency sub-band field parameter decoder module 1905 is used for recovering high-frequency sub-band territory signal in described two sound channels according to the high-frequency sub-band field parameter coded data of two sound channels in low frequency sub-band territory excitation spectrum of described two sound channels that is received from low frequency sub-band territory stereo decoding module 1902 and the bit stream demultiplexing module 1901 output side informations.
Comprehensive sub-filter pack module 1906, it is synthetic that high-frequency sub-band territory signal when being used for the low frequency sub-band territory in described two sound channels of the low frequency sub-band territory signal of described two sound channels of comprehensive module 1904 outputs of change prediction and 1905 outputs of high-frequency sub-band field parameter decoder module carries out subband, with the stereophonic signal of acquisition decoding.
In the present embodiment, become comprehensive module 1904 of prediction and high-frequency sub-band field parameter decoder module 1905 when low frequency sub-band territory frequency-time domain transformation module 1903, low frequency sub-band territory and adopt the module of the same name of two cover monophonic sound sound decoding devices respectively the left and right sound channels signal to be handled respectively.
Corresponding with stereo decoding apparatus, when not comprising spectrum, the high-frequency sub-band field parameter do not adjust parameter, then high-frequency sub-band field parameter decoder module 1905 also can obtain the low frequency sub-band territory pumping signal of two sound channels from low frequency sub-band territory frequency-time domain transformation module 1903, therefore high-frequency sub-band field parameter decoder module 1905 is not composed adjustment, the linear prediction synthesis filter separately that constitutes according to the high frequency LSF vector quantization index of two sound channels only, carry out integrated filter to only having passed through the gain low frequency sub-band territory pumping signal of adjusted two sound channels of time domain, thereby obtain the reconstruction high-frequency sub-band territory signal of two sound channels, realized the recovery of the high-frequency sub-band territory signal of two sound channels.
Figure 20 is the coding/decoding method process flow diagram based on stereo sound decoding device of the present invention.As shown in figure 20, this method may further comprise the steps:
Step 41: the acoustic coding code stream is carried out demultiplexing, all side informations that obtain low frequency sub-band territory stereo coding data and decode used.
Step 42: according to the stereo coding model selection information in the side information low frequency sub-band territory stereo coding data are carried out stereo decoding, obtain the low frequency sub-band territory excitation spectrum in described two sound channels.
Step 43: according to low frequency sub-band territory in the side information and signal type analysis result, the low frequency sub-band territory excitation spectrum in described two sound channels according to different rank, is carried out frequency-time domain transformation, obtain the low frequency sub-band territory pumping signal in described two sound channels.
Step 44: the LSF vector quantization index according to two sound channels in the side information obtains low frequency sub-band territory linear prediction synthesis filter, adopt low frequency sub-band territory and signal type analysis result in the side information that the low frequency sub-band territory pumping signal in two sound channels is divided the subframe processing, adopt the linear prediction synthesis filter that obtains the low frequency sub-band territory pumping signal in described two sound channels to be carried out integrated filter then, obtain the decoded low frequency sub-band territory signal in described two sound channels by subframe.
Step 45: according to low frequency sub-band territory excitation spectrum in described two sound channels and the high-frequency sub-band field parameter coded data in the side information, recover the high-frequency sub-band territory signal in described two sound channels, obtain the decoded high-frequency sub-band territory signal in described two sound channels.
Step 46: low frequency sub-band territory signal in described two sound channels and the high-frequency sub-band territory signal combination in described two sound channels are in the same place, carry out subband with comprehensive Methods of Subband Filter Banks and synthesize.
Wherein, frequency-time domain transformation method in the step 43, the prediction overall treatment in the step 44, the high-frequency sub-band field parameter coding/decoding method in the step 45, with the subband synthetic method in the step 46, all in the embodiment of the coding/decoding method of monophony decoding device of the present invention, introduced, in the embodiment of the coding/decoding method of stereo decoding apparatus of the present invention, adopt identical method, therefore do not do introduction.
Wherein, step 42 is carried out stereo decoding according to coding mode selection information, selects implementation method 1 corresponding to coding mode, and coding/decoding method is for selecting information that the low frequency sub-band territory stereo coding data of each sub-band are decoded according to coding mode; Select implementation method 2 corresponding to coding mode, coding/decoding method is for selecting information that the low frequency sub-band territory stereo coding data of each sub-band in the lower frequency sub-band are decoded according to coding mode, for the sub-band of upper frequency, adopt the parameter stereo decoding schema.Wherein, low frequency sub-band territory stereo decoding comprises three kinds of stereo decoding patterns.
With difference stereo decoding pattern is to recover low frequency sub-band territory excitation spectrum in described two sound channels in this sub-band by the low frequency sub-band territory in the sub-band with excitation spectrum and difference excitation spectrum.Specific implementation method is as follows:
Low frequency sub-band territory stereo decoding module 1902 will receive after low frequency sub-band territory stereo coding data carry out re-quantization decoding from bit stream demultiplexer 1901, obtain the low frequency sub-band territory and excitation spectrum
Figure A20071008887800731
With the difference excitation spectrum
Figure A20071008887800732
Adopt following formula to recover the low frequency sub-band territory excitation spectrum of left and right sound channels.
L &RightArrow; ^ = M &RightArrow; ^ + S &RightArrow; ^
R &RightArrow; ^ = M &RightArrow; ^ - S &RightArrow; ^
The parameter stereo decoding schema is weighted sum excitation spectrum in the sub-band that receives according to low frequency sub-band territory stereo decoding module 1902 and the relevant parameter in the side information
Figure A20071008887800735
With
Figure A20071008887800736
Recover the left and right sound channels low frequency sub-band territory excitation spectrum in this sub-band.Corresponding to embodiment 1 in the parameter stereo coding method of coded portion and embodiment 2, but the decode procedure of two kinds of embodiments is identical, comprises following steps:
Step 42-1a: low frequency sub-band territory stereo decoding module 1902 will receive after low frequency sub-band territory stereo coding data and relevant parameter carry out the re-quantization decoding from bit stream demultiplexer 1901, obtain the weighted sum excitation spectrum Parameter
Figure A20071008887800738
With
Figure A20071008887800739
Step 42-1b: produce and the weighted sum excitation spectrum The quadrature excitation spectrum that constant amplitude is vertical
Figure A200710088878007311
Wherein, D &RightArrow; [ i , k ] = - y m [ i , k ] + jx m [ i , k ] ;
Step 42-1c: according to the parameter that obtains
Figure A200710088878007313
Quadrature excitation is composed
Figure A200710088878007314
Convergent-divergent obtains the quadrature excitation spectrum behind the convergent-divergent
Step 42-1d: by the weighted sum excitation spectrum
Figure A200710088878007316
With the quadrature excitation spectrum behind the convergent-divergent
Figure A200710088878007317
Obtain the excitation spectrum of left and right sound channels, the excitation spectrum of one of them sound channel (R channel) through convergent-divergent after; Computing formula is as follows:
R &RightArrow; ^ &prime; = M &RightArrow; ^ &prime; + D &RightArrow; ^ &prime;
L &RightArrow; ^ &prime; = M &RightArrow; ^ &prime; - D &RightArrow; ^ &prime;
Step 42-1e: by the parameter that from side information, obtains A sound channel of convergent-divergent is carried out again convergent-divergent return original size, obtain
Figure A20071008887800744
Parameter error stereo decoding pattern is the sub-band weighted sum excitation spectrum that obtains according to low frequency sub-band territory stereo decoding module 1902
Figure A20071008887800745
The error excitation spectrum
Figure A20071008887800746
With relevant parameters in the side information With
Figure A20071008887800748
Recover this sub-band left and right acoustic channels excitation spectrum.Specific implementation method comprises following steps:
Step 42-2a: low frequency sub-band territory stereo decoding module 1902 will receive after low frequency sub-band territory stereo coding data and relevant parameter carry out the re-quantization decoding from bit stream demultiplexer 1901, obtain the weighted sum excitation spectrum
Figure A20071008887800749
The error excitation spectrum
Figure A200710088878007410
And parameter
Figure A200710088878007411
With
Step 42-2b: produce and the weighted sum excitation spectrum
Figure A200710088878007413
The quadrature excitation spectrum that constant amplitude is vertical
Figure A200710088878007414
Step 42-2c: according to the parameter that obtains
Figure A200710088878007415
Quadrature excitation is composed
Figure A200710088878007416
Convergent-divergent obtains the quadrature excitation spectrum behind the convergent-divergent
Figure A200710088878007417
Step 42-2d: the quadrature excitation spectrum behind the convergent-divergent
Figure A200710088878007418
With the error excitation spectrum
Figure A200710088878007419
Addition, the weighted difference excitation spectrum that is restored
Figure A200710088878007420
Step 42-2e: by the weighted sum excitation spectrum
Figure A200710088878007421
With the weighted difference excitation spectrum
Figure A200710088878007422
Obtain the excitation spectrum of left and right acoustic channels, the excitation spectrum of one of them sound channel (R channel) through convergent-divergent after;
Step 42-2f: pass through parameter
Figure A200710088878007423
The sound channel of convergent-divergent is carried out again convergent-divergent return original size.
In this preferred embodiment, step 45 is when the signal of the high-frequency sub-band territory of two sound channels of recovery, if do not compose in the high-frequency sub-band field parameter of two sound channels and adjust parameter, then this step 45 can be recovered the high-frequency sub-band territory signal of two sound channels according to the high-frequency sub-band field parameter of two sound channels from the low frequency sub-band territory pumping signal of two sound channels of step 43 acquisition, does not comprise the step that spectrum is adjusted in rejuvenation.
By the above as can be seen, monophonic sounds coding and decoding scheme provided by the present invention has adopted the combining form of high efficiency sub-band coding, predictive coding and transition coding when low frequency sub-band territory signal is carried out waveform coding, improved code efficiency; When being handled, adopted high-frequency sub-band territory signal the high efficiency factors coding form, also carry out in the process of parameter coding in high-frequency sub-band territory pumping signal, carry out the gain of effective spectrum structure adjustment and time domain and adjust, the distortion of sound after not only having improved code efficiency but also having reduced decoding.
In addition, stereo coding/decoding scheme of the present invention not only has the above-mentioned advantage that is had based on the monophonic sounds coding and decoding scheme of the principle of the invention, multiple parameter stereo coding method based on the subband excitation spectrum also is provided, can reduce code check, be adapted at the stereo coding under the extremely low code check.
In sum, more than be preferred embodiment of the present invention only, be not to be used to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (31)

1. a monophonic sounds code device is characterized in that, this code device comprises:
Analyze the sub-filter pack module, be used for the monophonic sound tone signal is carried out sub-band division, be decomposed into the low frequency sub-band territory signal that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises at least 2 subbands;
Become the forecast analysis module during low frequency sub-band territory, be used for described low frequency sub-band territory signal is carried out forecast analysis, to obtain low frequency sub-band territory pumping signal;
Low frequency sub-band territory time-frequency conversion module is used for described low frequency sub-band territory pumping signal is carried out time-frequency conversion, to obtain low frequency sub-band territory excitation spectrum;
Low frequency sub-band domain waveform coding module is used for described low frequency sub-band territory excitation spectrum is carried out quantization encoding, to obtain low frequency sub-band domain waveform coded data;
High-frequency sub-band field parameter coding module, be used for according to described low frequency sub-band territory excitation spectrum and described high-frequency sub-band territory signal, calculating is used for recovering from low frequency sub-band territory excitation spectrum the high-frequency sub-band field parameter of high-frequency sub-band territory signal, after this high-frequency sub-band field parameter carried out quantization encoding, to obtain high-frequency sub-band field parameter coded data; Perhaps, according to described low frequency sub-band territory pumping signal and described high-frequency sub-band territory signal, calculating is used for recovering from low frequency sub-band territory pumping signal the high-frequency sub-band field parameter of high-frequency sub-band territory signal, after this high-frequency sub-band field parameter carried out quantization encoding, to obtain high-frequency sub-band field parameter coded data;
The bit stream Multiplexing module, be used for described low frequency sub-band domain waveform coded data and high-frequency sub-band field parameter coded data are carried out multiplexing, with the output sound encoding code stream.
2. code device as claimed in claim 1, it is characterized in that, this code device further comprises low frequency sub-band territory signal type analysis module, and this module is used for the described low frequency sub-band of frame territory signal is carried out the signal type analysis, and output low frequency subband domain signal type analysis result; If described low frequency sub-band territory signal is tempolabile signal then output signal type; If fast changed signal then further obtains fast height position, and output signal type and described fast height position;
Becoming the forecast analysis module during described low frequency sub-band territory and be further used for according to described low frequency sub-band territory signal type analysis result, is subframe one or more, that be used for forecast analysis with described low frequency sub-band territory division of signal;
Described low frequency sub-band territory time-frequency conversion module is further used for according to described low frequency sub-band territory signal type analysis result described low frequency sub-band territory pumping signal being divided into one or more, as to be used for time-frequency conversion subframe;
Described bit stream Multiplexing module is further used for carrying out multiplexing to described low frequency sub-band territory signal type analysis result.
3. code device as claimed in claim 1 is characterized in that, when described high-frequency sub-band field parameter coding module during according to low frequency sub-band territory excitation spectrum and high-frequency sub-band territory calculated signals high-frequency sub-band field parameter, described high-frequency sub-band field parameter coding module comprises:
The time become predictive analyzer, be used for the described high-frequency sub-band territory signal that receives is carried out linear prediction analysis, obtain the high frequency LSF vector quantization index and the high-frequency sub-band territory pumping signal of linear prediction filter; Export described high frequency LSF vector quantization index to described bit stream Multiplexing module after encoded;
The spectrum parametric encoder, be used for described high-frequency sub-band territory pumping signal is carried out time-frequency conversion, obtain high-frequency sub-band territory excitation spectrum, according to described high-frequency sub-band territory excitation spectrum the described low frequency sub-band territory excitation spectrum that receives is composed adjustment, the spectrum of extracting is adjusted parameter behind quantization encoding, export described bit stream Multiplexing module to, carry out frequency-time domain transformation to composing adjusted low frequency sub-band territory excitation spectrum, obtain composing adjusted low frequency sub-band territory pumping signal;
The time become the prediction synthesizer, be used to adopt the linear prediction synthesis filter that obtains according to described high frequency LSF vector quantization index, the adjusted low frequency sub-band of described spectrum territory pumping signal is carried out integrated filter, obtain rebuilding high-frequency sub-band territory signal;
Parameter extraction module is adjusted in the time-domain adaptive gain, is used for described high-frequency sub-band territory signal and described reconstruction high-frequency sub-band territory signal are compared, and obtains the time domain gain and adjusts parameter, exports described bit stream Multiplexing module behind quantization encoding to.
4. a monophonic sounds coding method is characterized in that, this method comprises:
A, the monophonic sound tone signal is carried out sub-band division, be decomposed into the low frequency sub-band territory signal that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises at least 2 subbands;
B, described low frequency sub-band territory signal is carried out forecast analysis and time-frequency conversion, to obtain low frequency sub-band territory excitation spectrum; Described low frequency sub-band territory excitation spectrum is carried out quantization encoding, to obtain low frequency sub-band domain waveform coded data;
C, according to described high-frequency sub-band territory signal and low frequency sub-band territory excitation spectrum, calculating is used for recovering from low frequency sub-band territory excitation spectrum the high-frequency sub-band field parameter of high-frequency sub-band territory signal, and after this high-frequency sub-band field parameter carried out quantization encoding, obtain high-frequency sub-band field parameter coded data; Perhaps, according to described high-frequency sub-band territory signal with through low frequency sub-band territory pumping signal that described forecast analysis obtains, calculating is used for recovering from low frequency sub-band territory pumping signal the high-frequency sub-band field parameter of high-frequency sub-band territory signal, and after the high-frequency sub-band field parameter carried out quantization encoding, obtain high-frequency sub-band field parameter coded data;
D, described low frequency sub-band domain waveform coded data and described high-frequency sub-band field parameter coded data are carried out multiplexing, the output sound encoding code stream.
5. coding method as claimed in claim 4 is characterized in that, this method further comprises step: a frame low frequency sub-band territory signal is carried out the signal type analysis, and definite low frequency sub-band territory signal type analysis result; If described low frequency sub-band territory signal is a tempolabile signal, then with signal type as low frequency sub-band territory signal type analysis result; If fast changed signal then further obtains fast height position, with signal type and fast height position as the signal type analysis result;
Step B is described to carry out forecast analysis and time-frequency conversion comprises to described low frequency sub-band territory signal: according to low frequency sub-band territory signal type analysis result, be subframe one or more, that be used for forecast analysis with a frame low frequency sub-band territory division of signal; By subframe described low frequency sub-band territory signal is carried out forecast analysis, obtain the low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory signal type analysis result, a described frame low frequency sub-band territory pumping signal is divided into one or more, as to be used for time-frequency conversion subframe; By subframe described low frequency sub-band territory pumping signal is carried out time-frequency conversion, obtain the low frequency sub-band territory excitation spectrum of subframe;
Step D further comprises and carries out multiplexing to described low frequency sub-band territory signal type analysis result.
6. coding method as claimed in claim 4 is characterized in that, when described step C calculated the high-frequency sub-band field parameter according to high-frequency sub-band territory signal and low frequency sub-band territory excitation spectrum, described step C comprised:
C1, obtain high-frequency sub-band territory pumping signal, and high-frequency sub-band territory pumping signal is obtained high-frequency sub-band territory excitation spectrum as time-frequency conversion according to high-frequency sub-band territory signal; According to high-frequency sub-band territory excitation spectrum described low frequency sub-band territory excitation spectrum is composed adjustment, obtain spectrum and adjust parameter and compose adjusted low frequency sub-band territory excitation spectrum; Carry out frequency-time domain transformation and obtain composing adjusted low frequency sub-band territory pumping signal composing adjusted low frequency sub-band territory excitation spectrum;
C2, the adjusted low frequency sub-band of described spectrum territory pumping signal is carried out integrated filter, obtain rebuilding high-frequency sub-band territory signal, and extract the time domain gain according to the time domain gain of described reconstruction high-frequency sub-band territory signal with the time domain gain of high-frequency sub-band territory signal and adjust parameter; Obtain comprising the high-frequency sub-band field parameter of step C1 gained spectrum adjustment parameter and described time domain gain adjustment parameter.
7. a monophonic sound sound decoding device is characterized in that, this decoding device comprises:
The bit stream demultiplexing module is used for the acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band domain waveform coded data that comprises at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises at least 2 subbands;
Low frequency sub-band domain waveform decoder module is used for described low frequency sub-band domain waveform coded data is carried out the re-quantization decoding, to obtain low frequency sub-band territory excitation spectrum;
Low frequency sub-band territory frequency-time domain transformation module is used for described low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation, to obtain low frequency sub-band territory pumping signal;
Become the comprehensive module of prediction during the low frequency sub-band territory, be used for described low frequency sub-band territory pumping signal is predicted comprehensively, to obtain low frequency sub-band territory signal;
High-frequency sub-band field parameter decoder module, be used for described high-frequency sub-band field parameter coded data is carried out the re-quantization decoding with acquisition high-frequency sub-band field parameter, and from the excitation spectrum of described low frequency sub-band territory, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter; Perhaps, from the pumping signal of described low frequency sub-band territory, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter;
Comprehensive sub-filter pack module, it is synthetic to be used for that described low frequency sub-band territory signal and described high-frequency sub-band territory signal are carried out subband, to obtain the monophonic sound tone signal of decoding.
8. decoding device as claimed in claim 7 is characterized in that, described bit stream demultiplexing module is further used for, and obtains the low frequency sub-band territory signal type analysis result that is used to recover monophonic sounds from the described acoustic coding code stream of demultiplexing;
Described low frequency sub-band territory frequency-time domain transformation module is further used for according to described low frequency sub-band territory signal type analysis result, and the low frequency sub-band territory excitation spectrum that receives is divided into one or more, as to be used for frequency-time domain transformation subframe;
Become the comprehensive module of prediction during described low frequency sub-band territory and be further used for according to low frequency sub-band territory signal type analysis result, with the low frequency sub-band territory pumping signal that receives be divided into one or more, be used to predict comprehensive subframe.
9. decoding device as claimed in claim 7, it is characterized in that, when described high-frequency sub-band field parameter decoder module recovered high-frequency sub-band territory signal according to the high-frequency sub-band field parameter from the excitation spectrum of low frequency sub-band territory, described high-frequency sub-band field parameter comprised that high frequency LSF vector quantization index, spectrum are adjusted parameter and parameter is adjusted in the time domain gain; This high-frequency sub-band field parameter decoder module comprises:
The spectrum parameter decoder is used for adjusting parameter according to described spectrum described low frequency sub-band territory excitation spectrum is composed adjustment, will compose adjusted low frequency sub-band territory excitation spectrum and carry out frequency-time domain transformation, obtains composing adjusted low frequency sub-band territory pumping signal;
Adaptive time domain gain demoder is used for adjusting parameter according to described time domain gain and described low frequency sub-band territory pumping signal is carried out the time domain gain adjusts, and obtains the time domain adjusted low frequency sub-band territory pumping signal that gains;
The time become the prediction synthesizer, be used for obtaining the linear prediction synthesis filter according to described high frequency LSF vector quantization index, the described time domain adjusted low frequency sub-band territory pumping signal that gains is carried out integrated filter, obtain rebuilding high-frequency sub-band territory signal, and export described comprehensive sub-filter pack module to.
10. a monophonic sounds coding/decoding method is characterized in that, this method comprises:
A, the acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band domain waveform coded data that comprises at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises at least 2 subbands;
B, described low frequency sub-band domain waveform coded data is carried out re-quantization decoding, to obtain low frequency sub-band territory excitation spectrum; Described low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation and predicts comprehensively, to obtain low frequency sub-band territory signal;
C, described high-frequency sub-band field parameter coded data is carried out re-quantization decoding obtaining the high-frequency sub-band field parameter, and from the excitation spectrum of described low frequency sub-band territory, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter; Perhaps, from the low frequency sub-band territory pumping signal that obtains through described frequency-time domain transformation, recover high-frequency sub-band territory signal according to described high-frequency sub-band field parameter;
D, that described low frequency sub-band territory signal and described high-frequency sub-band territory signal are carried out subband is synthetic, the monophonic sound tone signal of output decoder.
11. coding/decoding method as claimed in claim 10 is characterized in that, described steps A further comprises obtains the low frequency sub-band territory signal type analysis result that is used to recover monophonic sounds from the acoustic coding code stream of demultiplexing;
Step B is described to carry out low frequency sub-band territory excitation spectrum frequency-time domain transformation and predicts comprehensively comprising: according to described low frequency sub-band territory signal type analysis result, the described low frequency sub-band of frame territory excitation spectrum is divided into one or more, as to be used for frequency-time domain transformation subframe; By subframe described low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation, obtain the low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory signal type analysis result, with a described frame low frequency sub-band territory pumping signal be divided into one or more, be used to predict comprehensive subframe; By subframe described low frequency sub-band territory signal is predicted comprehensively, obtained the low frequency sub-band territory signal of subframe.
12. coding/decoding method as claimed in claim 10, it is characterized in that, when described step C recovered high-frequency sub-band territory signal according to the high-frequency sub-band field parameter from the excitation spectrum of low frequency sub-band territory, described high-frequency sub-band field parameter comprised that high frequency LSF vector quantization index, spectrum are adjusted parameter and parameter is adjusted in the time domain gain; Described step C comprises:
C1, adjust parameter according to spectrum described low frequency sub-band territory excitation spectrum is composed adjustment; Adjusted low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation, obtain low frequency sub-band territory pumping signal;
C2, described low frequency sub-band territory pumping signal is carried out time domain gain adjustment, obtain adjusted low frequency sub-band territory pumping signal;
C3, employing are carried out integrated filter to adjusted low frequency sub-band territory pumping signal, the high-frequency sub-band territory signal that obtains rebuilding according to the linear prediction synthesis filter that described high frequency LSF vector quantization index obtains.
13. a stereo encoding apparatus is characterized in that, this code device comprises:
Analyze the sub-filter pack module, the left and right sound channels that is used for stereophonic signal is carried out sub-band division respectively, be decomposed into the low frequency sub-band territory signal of the left and right sound channels that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises the left and right sound channels of at least 2 subbands;
Become the forecast analysis module during low frequency sub-band territory, be used for respectively the low frequency sub-band territory signal of described left and right sound channels is carried out forecast analysis, to obtain the low frequency sub-band territory pumping signal of left and right sound channels;
Low frequency sub-band territory time-frequency conversion module is used for respectively described left and right sound channels low frequency sub-band territory pumping signal being carried out time-frequency conversion, to obtain the low frequency sub-band territory excitation spectrum of left and right sound channels;
Low frequency sub-band territory stereo coding module is used for the low frequency sub-band territory excitation spectrum of described left and right sound channels is carried out stereo coding, to obtain low frequency sub-band territory stereo coding data;
High-frequency sub-band field parameter coding module, be used for respectively according to the low frequency sub-band territory excitation spectrum of described left and right sound channels and the high-frequency sub-band territory signal of described left and right sound channels, calculating is used for from the high-frequency sub-band field parameter of the left and right sound channels of the high-frequency sub-band territory signal of the low frequency sub-band territory excitation spectrum recovery left and right sound channels of left and right sound channels, and respectively the high-frequency sub-band field parameter of described left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels; Perhaps, respectively according to the high-frequency sub-band territory signal of the low frequency sub-band territory pumping signal and the described left and right sound channels of described left and right sound channels, calculating is used for from the high-frequency sub-band field parameter of the left and right sound channels of the high-frequency sub-band territory signal of the low frequency sub-band territory pumping signal recovery left and right sound channels of left and right sound channels, and respectively the high-frequency sub-band field parameter of described left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels;
The bit stream Multiplexing module, be used for the high-frequency sub-band field parameter coded data of described low frequency sub-band territory stereo coding data and described left and right sound channels is carried out multiplexing, to export stereosonic acoustic coding code stream.
14. code device as claimed in claim 13, it is characterized in that, this code device further comprises low frequency sub-band territory and signal type analysis module, this module is used for going out a frame low frequency sub-band territory and a signal according to the low frequency sub-band territory calculated signals of a frame left and right sound channels, the signal type analysis is carried out in this low frequency sub-band territory and signal, and output low frequency subband domain and signal type analysis result; If described low frequency sub-band territory and signal are tempolabile signal then output signal type; If fast changed signal then further obtains fast height position, and output signal type and described fast height position;
Become the forecast analysis module during described low frequency sub-band territory and be further used for according to described low frequency sub-band territory and signal type analysis result, the low frequency sub-band territory division of signal of the described left and right sound channels of a frame that will receive respectively is a subframe one or more, that be used for forecast analysis;
Described low frequency sub-band territory time-frequency conversion module is further used for according to described low frequency sub-band territory and signal type analysis result, and the described left and right sound channels low frequency sub-band territory pumping signal that will receive respectively is divided into one or more, as to be used for time-frequency conversion subframe;
Described bit stream Multiplexing module is further used for carrying out multiplexing to described low frequency sub-band territory and signal type analysis result.
15. code device as claimed in claim 13 is characterized in that, described high-frequency sub-band field parameter coding module comprises two identical high-frequency sub-band field parameter coding submodules, is respectively applied for the high-frequency sub-band field parameter that extracts left and right sound channels; When described high-frequency sub-band field parameter coding submodule during according to the high-frequency sub-band field parameter of a high-frequency sub-band territory calculated signals left side/R channel of the low frequency sub-band territory excitation spectrum of a left side/R channel and a left side/R channel, described high-frequency sub-band field parameter coding submodule comprises:
The time become predictive analyzer, be used for the high-frequency sub-band territory signal of a described left side/R channel of receiving is carried out linear prediction analysis, obtain the high frequency LSF vector quantization index and the high-frequency sub-band territory pumping signal of the linear prediction filter of a left side/R channel; Export the high frequency LSF vector quantization index of a described left side/R channel to described bit stream Multiplexing module after encoded;
The spectrum parametric encoder, be used for the high-frequency sub-band territory pumping signal of a described left side/R channel is carried out time-frequency conversion, obtain the high-frequency sub-band territory excitation spectrum of a left side/R channel, high-frequency sub-band territory excitation spectrum according to a described left side/R channel is composed adjustment to the low frequency sub-band territory excitation spectrum of a described left side/R channel of reception, the spectrum of a left side/R channel of extracting is adjusted parameter behind quantization encoding, export described bit stream Multiplexing module to, the adjusted low frequency sub-band of the spectrum of a left side/R channel territory excitation spectrum is carried out frequency-time domain transformation, obtain composing the low frequency sub-band territory pumping signal of an adjusted left side/R channel;
The time become the prediction synthesizer, be used to adopt the linear prediction synthesis filter of a left side/R channel that obtains according to described high frequency LSF vector quantization index, low frequency sub-band territory pumping signal to an adjusted left side/R channel of described spectrum is carried out integrated filter, obtains the reconstruction high-frequency sub-band territory signal of a left side/R channel;
Parameter extractor is adjusted in the time-domain adaptive gain, be used for the high-frequency sub-band territory signal of a described left side/R channel and the reconstruction high-frequency sub-band territory signal of a described left side/R channel are compared, parameter is adjusted in the time domain gain that obtains a left side/R channel, exports described bit stream Multiplexing module behind quantization encoding to.
16. code device as claimed in claim 13, it is characterized in that, described low frequency sub-band territory stereo coding module is further used for selecting from more than one optional stereo coding pattern, adopt the stereo coding pattern of selecting to encode, and export coding mode selection information to described bit stream Multiplexing module.
17. a stereo encoding method is characterized in that, this method comprises:
The left and right sound channels of A, stereophonic signal is carried out sub-band division respectively, is decomposed into the low frequency sub-band territory signal of the left and right sound channels that comprises at least 2 subbands, and the high-frequency sub-band territory signal that comprises the left and right sound channels of at least 2 subbands;
B, the low frequency sub-band territory signal to described left and right sound channels carries out forecast analysis and time-frequency conversion respectively, to obtain the low frequency sub-band territory excitation spectrum of left and right sound channels; Low frequency sub-band territory excitation spectrum to described left and right sound channels carries out stereo coding, to obtain low frequency sub-band territory stereo coding data;
C, respectively according to the high-frequency sub-band territory signal and the described left and right sound channels low frequency sub-band territory excitation spectrum of described left and right sound channels, calculating is used for recovering from left and right sound channels low frequency sub-band territory excitation spectrum the high-frequency sub-band field parameter of the left and right sound channels of left and right sound channels high-frequency sub-band territory signal, and respectively the high-frequency sub-band field parameter of left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels; Perhaps, the low frequency sub-band territory pumping signal of the left and right sound channels that obtains according to the high-frequency sub-band territory signal of described left and right sound channels with through described forecast analysis respectively, calculating is used for recovering from left and right sound channels low frequency sub-band territory pumping signal the high-frequency sub-band field parameter of the left and right sound channels of left and right sound channels high-frequency sub-band territory signal, and respectively the high-frequency sub-band field parameter of described left and right sound channels is carried out quantization encoding, to obtain the high-frequency sub-band field parameter coded data of left and right sound channels;
D, the high-frequency sub-band field parameter coded data of described low frequency sub-band territory stereo coding data and described left and right sound channels is carried out multiplexing, to export stereosonic acoustic coding code stream.
18. coding method as claimed in claim 17, it is characterized in that, this method further comprises step: the low frequency sub-band territory calculated signals of a frame left and right sound channels that obtains according to steps A goes out a frame low frequency sub-band territory and a signal, the signal type analysis is carried out in this low frequency sub-band territory and signal, and definite low frequency sub-band territory and signal type analysis result; If described low frequency sub-band territory and signal are tempolabile signals, then with signal type as described low frequency sub-band territory and signal type analysis result; If fast changed signal then further obtains fast height position, with signal type and described fast height position as described low frequency sub-band territory and signal type analysis result;
The described low frequency sub-band territory signal to a described left side/R channel of step B carries out forecast analysis and time-frequency conversion comprises: according to low frequency sub-band territory and signal type analysis result, be subframe one or more, that be used for forecast analysis with a frame left side/R channel low frequency sub-band territory division of signal; By subframe the low frequency sub-band territory signal of a described left side/R channel is carried out forecast analysis, obtain a left side/R channel low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame left side/R channel low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory and signal type analysis result, a described frame left side/R channel low frequency sub-band territory pumping signal is divided into one or more, as to be used for time-frequency conversion subframe; By subframe a described left side/R channel low frequency sub-band territory pumping signal is carried out time-frequency conversion, obtain a left side/R channel low frequency sub-band territory excitation spectrum of subframe;
Step D further comprises and carries out multiplexing to described low frequency sub-band territory and signal type analysis result.
19. coding method as claimed in claim 17 is characterized in that, when described step C calculated the high-frequency sub-band field parameter of a left side/R channel according to the low frequency sub-band territory excitation spectrum of the high-frequency sub-band territory signal of a left side/R channel and a left side/R channel, described step C comprised:
C1, obtain the high-frequency sub-band territory pumping signal of a left side/R channel, and the high-frequency sub-band territory pumping signal of a described left side/R channel is obtained the high-frequency sub-band territory excitation spectrum of a left side/R channel as time-frequency conversion according to the high-frequency sub-band territory signal of a left side/R channel; High-frequency sub-band territory excitation spectrum according to a described left side/R channel is composed adjustment to the low frequency sub-band territory excitation spectrum of a described left side/R channel, obtains the spectrum adjustment parameter of a left side/R channel and the low frequency sub-band territory excitation spectrum of composing an adjusted left side/R channel; The low frequency sub-band territory excitation spectrum of composing an adjusted left side/R channel is carried out the low frequency sub-band territory pumping signal that frequency-time domain transformation obtains composing an adjusted left side/R channel;
C2, the low frequency sub-band territory pumping signal of an adjusted left side/R channel of described spectrum is carried out integrated filter, obtain the reconstruction high-frequency sub-band territory signal of a left side/R channel, and adjust parameter with the time domain gain that a left side/R channel is extracted in the time domain gain of the high-frequency sub-band territory signal of a left side/R channel according to the time domain gain of the reconstruction high-frequency sub-band territory signal of a described left side/R channel; The high-frequency sub-band field parameter of a left side/R channel of parameter is adjusted in the time domain gain that the spectrum that obtains comprising a step C1 gained left side/R channel is adjusted parameter and a described left side/R channel.
20. coding method as claimed in claim 17 is characterized in that, the described low frequency sub-band territory excitation spectrum to left and right sound channels of step B carries out stereo coding and comprises:
B2, the low frequency sub-band territory excitation spectrum with described left and right sound channels is divided into plurality of sub-bands respectively, selects a kind of stereo coding pattern to carry out stereo coding to each sub-band;
Step D further comprises and carries out multiplexing to coding mode selection information.
21. coding method as claimed in claim 20 is characterized in that, step B2 is described to select a kind of stereo coding pattern to comprise to sub-band k:
B21, the optional stereo coding pattern that adopts more than one are respectively carried out stereo coding and decoding to the low frequency sub-band territory excitation spectrum of the left and right sound channels of described sub-band k, the error of the low frequency sub-band territory excitation spectrum of the left and right sound channels of sub-band k before the low frequency sub-band territory excitation spectrum of the left and right sound channels of the sub-band k that calculating decodes recovers and the coding, by comparing the coding mode of Select Error minimum, as the stereo coding pattern of sub-band k; Perhaps
B22, in the low frequency sub-band territory, the sub-band that frequency is higher than a determined value adopts fixing stereo coding pattern; Frequency is lower than the sub-band of described determined value and carries out described step B21 selection stereo coding pattern.
22. coding method as claimed in claim 21 is characterized in that, when the stereo coding pattern was the parameter stereo coding pattern, the described stereo coding that carries out comprised:
B31, in sub-band k, the first sound channel low frequency sub-band territory excitation spectrum in the left and right sound channels is calculated weighting parameters g r(k), make the second sound channel low frequency sub-band territory excitation spectrum and adopt g r(k) the first sound channel low frequency sub-band territory excitation spectrum energy equates behind the convergent-divergent;
B32, according to the first sound channel low frequency sub-band territory excitation spectrum behind second sound road subband domain excitation spectrum and the convergent-divergent, calculate weighted sum excitation spectrum and weighted difference excitation spectrum;
B33, the calculating quadrature excitation spectrum vertical with described weighted sum excitation spectrum constant amplitude;
B34, according to described quadrature excitation spectrum and described weighted difference excitation spectrum, calculate weighting parameters g d(k), make described weighted difference excitation spectrum and adopt g d(k) the quadrature excitation spectrum energy behind the convergent-divergent equates;
B35, with described weighted sum excitation spectrum, g r(k) and g d(k) carry out quantization encoding respectively, with the weighted sum excitation spectrum of quantization encoding as low frequency sub-band territory stereo coding data, with the g of quantization encoding r(k) and g d(k) as the parameter that is used for stereo decoding;
Step D further comprises described weighting parameters g r(k) and g d(k) carry out multiplexing;
Perhaps, when described stereo coding pattern was described parameter stereo coding pattern, the described stereo coding that carries out comprised:
B41, in sub-band k, calculate weighting parameters g according to following formula d(k):
g d ( k ) = - b ( k ) + b 2 ( k ) + a 2 ( k ) a ( k )
Wherein, a ( k ) = &Sigma; i &Element; band ( k ) ( x r [ i , k ] y l [ i , k ] - x l [ i , k ] y r [ i , k ] ) ;
b ( k ) = &Sigma; i &Element; band ( k ) ( x l [ i , k ] x r [ i , k ] + y l [ i , k ] y r [ i , k ] ) ;
x lAnd y lBe respectively the real part and the imaginary part of L channel low frequency sub-band territory excitation spectrum, x rAnd y rBe respectively the real part and the imaginary part of R channel low frequency sub-band territory excitation spectrum;
B42, calculate weighting parameters g according to following formula r(k):
g r ( k ) = - ( c ( k ) - d ( k ) ) + ( c ( k ) - d ( k ) ) 2 + g ( k ) m 2 ( k ) g ( k ) m 2 ( k ) ;
Wherein, c ( k ) = &Sigma; i &Element; band ( k ) ( x l [ i , k ] x l [ i , k ] + y l [ i , k ] y l [ i , k ] ) ;
d ( k ) = &Sigma; i &Element; band ( k ) ( x r [ i , k ] x r [ i , k ] + y r [ i , k ] y r [ i , k ] ) ;
m ( k ) = 2 b ( k ) ( 1 - g d 2 ( k ) ) + 2 a ( k ) g d ( k ) 1 + g d 2 ( k ) ;
G (k) is the importance factors of parameter stereo coding error in the distribution of left and right sound channels;
B43, calculate the weighted sum excitation spectrum according to following formula,
x m [ i , k ] = x l [ i , k ] + g d ( k ) y l [ i , k ] + g ( k ) g r ( k ) ( x r [ i , k ] - g d ( k ) y r [ i , k ] ) ( 1 + g d 2 ( k ) ) ( 1 + g ( k ) g r 2 ( k )
y m [ i , k ] = - g d ( k ) x l [ i , k ] + y l [ i , k ] + g ( k ) g r ( k ) ( g d ( k ) x r [ i , k ] + y r [ i , k ] ) ( 1 + g d 2 ( k ) ) ( 1 + g ( k ) g r 2 ( k ) ;
Wherein, x mAnd y mReal part and the imaginary part of representing the weighted sum excitation spectrum respectively;
Step D further comprises described weighting parameters g r(k) and g d(k) carry out multiplexing;
When described stereo coding pattern was parameter error stereo coding pattern, the described stereo coding that carries out comprised:
B51, in sub-band k, the first sound channel low frequency sub-band territory excitation spectrum in the left and right sound channels is calculated weighting parameters g r(k), make the second sound channel low frequency sub-band territory excitation spectrum and adopt g r(k) the first sound channel low frequency sub-band territory excitation spectrum energy equates behind the convergent-divergent;
B52, according to the first sound channel low frequency sub-band territory excitation spectrum behind second sound road subband domain excitation spectrum and the convergent-divergent, calculate weighted sum excitation spectrum and weighted difference excitation spectrum;
B53, the calculating quadrature excitation spectrum vertical with described weighted sum excitation spectrum constant amplitude;
B54, according to described quadrature excitation spectrum and described weighted difference excitation spectrum, calculate weighting parameters g d(k), make described weighted difference excitation spectrum and adopt g d(k) energy of the spectrum of the quadrature excitation behind the convergent-divergent equates;
The error excitation spectrum of quadrature excitation spectrum behind B55, the described weighted difference excitation spectrum of calculating and the convergent-divergent;
B56, with described weighted sum excitation spectrum, error excitation spectrum, g r(k) and g d(k) carry out quantization encoding respectively, with the weighted sum excitation spectrum of quantization encoding and error excitation spectrum as low frequency sub-band territory stereo coding data, with the g of quantization encoding r(k) and g d(k) as the parameter that is used for stereo decoding;
Step D further comprises described weighting parameters g r(k) and g d(k) carry out multiplexing.
23. a stereo decoding apparatus is characterized in that, this decoding device comprises:
The bit stream demultiplexing module is used for stereosonic acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band territory stereo coding data that comprise at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises the left and right sound channels of at least 2 subbands;
Low frequency sub-band territory stereo decoding module is used for described low frequency sub-band territory stereo coding data are carried out stereo decoding, to obtain the low frequency sub-band territory excitation spectrum of left and right sound channels;
Conversion module during described low frequency sub-band territory is used for respectively the low frequency sub-band territory excitation spectrum of described left and right sound channels is carried out frequency-time domain transformation, to obtain the low frequency sub-band territory pumping signal of left and right sound channels;
Become the comprehensive module of prediction during the low frequency sub-band territory, be used for respectively the low frequency sub-band territory pumping signal of described left and right sound channels being predicted comprehensively, to obtain the low frequency sub-band territory signal of left and right sound channels;
High-frequency sub-band field parameter decoder module, be used for the high-frequency sub-band field parameter coded data of described left and right sound channels is carried out the high-frequency sub-band field parameter of re-quantization decoding with the acquisition left and right sound channels, and from the low frequency sub-band territory excitation spectrum of described left and right sound channels, recover the high-frequency sub-band territory signal of left and right sound channels respectively according to the high-frequency sub-band field parameter of described left and right sound channels; Perhaps, from the low frequency sub-band territory pumping signal of described left and right sound channels, recover the high-frequency sub-band territory signal of left and right sound channels respectively according to the high-frequency sub-band field parameter of described left and right sound channels;
Comprehensive sub-filter pack module, being used for high-frequency sub-band territory signal with the low frequency sub-band territory signal of described left and right sound channels and described left and right sound channels, to carry out subband synthetic, with the stereophonic signal of the left and right sound channels of obtaining decoding.
24. decoding device as claimed in claim 23 is characterized in that, described bit stream demultiplexing module is further used for, and obtains to be used to recover stereosonic low frequency sub-band territory and signal type analysis result from the described stereosonic acoustic coding code stream of demultiplexing;
Described low frequency sub-band territory frequency-time domain transformation module is further used for according to described low frequency sub-band territory and signal type analysis result, and the low frequency sub-band territory excitation spectrum with the left and right sound channels that receives is divided into one or more, as to be used for frequency-time domain transformation subframe respectively;
Become the comprehensive module of prediction during described low frequency sub-band territory and be further used for according to described low frequency sub-band territory and signal type analysis result, with the low frequency sub-band territory pumping signal of the left and right sound channels that receives be divided into one or more, be used to predict comprehensive subframe.
25. decoding device as claimed in claim 23 is characterized in that, described high-frequency sub-band field parameter decoder module comprises two identical high-frequency sub-band field parameter decoding submodules, is respectively applied for the high-frequency sub-band territory signal that recovers left and right sound channels; When described high-frequency sub-band field parameter coding module according to the high-frequency sub-band field parameter of left and right sound channels when the low frequency sub-band territory excitation spectrum of left and right sound channels recovers high-frequency sub-band territory signal, the high-frequency sub-band field parameter of a described left side/R channel comprises that high frequency LSF vector quantization index, spectrum are adjusted parameter and parameter is adjusted in the time domain gain; Described high-frequency sub-band field parameter decoding submodule comprises:
The spectrum parameter decoder, be used for adjusting parameter the low frequency sub-band territory excitation spectrum of a described left side/R channel is composed adjustment according to the spectrum of a described left side/R channel, the low frequency sub-band territory excitation spectrum of an adjusted left side/R channel of spectrum is carried out frequency-time domain transformation, obtain composing the low frequency sub-band territory pumping signal of an adjusted left side/R channel;
Adaptive time domain gain demoder, be used for according to the time domain of a described left side/R channel gain and adjust parameter and the low frequency sub-band territory pumping signal of a described left side/R channel is carried out the time domain gain adjust, obtain the gain low frequency sub-band territory pumping signal of an adjusted left side/R channel of time domain;
The time become the prediction synthesizer, be used to adopt the linear prediction synthesis filter that obtains a left side/R channel according to the high frequency LSF vector quantization index of a described left side/R channel, the gain low frequency sub-band territory pumping signal of an adjusted left side/R channel of described time domain is carried out integrated filter, obtain the reconstruction high-frequency sub-band territory signal of a left side/R channel, and output to described comprehensive sub-filter pack module.
26. decoding device as claimed in claim 23 is characterized in that, described bit stream Multiplexing module is further used for, and obtains the coding mode that is used for stereo decoding and select information from the described stereosonic acoustic coding code stream of demultiplexing;
Described low frequency sub-band territory stereo decoding module is further used for, and adopts described coding mode to select the stereo decoding pattern of information correspondence to carry out stereo decoding.
27. a stereo decoding method is characterized in that, this method comprises:
A, stereosonic acoustic coding code stream is carried out demultiplexing, obtaining the low frequency sub-band territory stereo coding data that comprise at least 2 subbands, and the high-frequency sub-band field parameter coded data that comprises the left and right sound channels of at least 2 subbands;
B, described low frequency sub-band territory stereo coding data are carried out stereo decoding, obtain the low frequency sub-band territory excitation spectrum of left and right sound channels; Low frequency sub-band territory excitation spectrum to described left and right sound channels carries out frequency-time domain transformation and predicts comprehensively respectively, to obtain the low frequency sub-band territory signal of left and right sound channels;
C, the high-frequency sub-band field parameter coded data of described left and right sound channels is carried out re-quantization decoding obtaining the high-frequency sub-band field parameter of left and right sound channels, and from the low frequency sub-band territory excitation spectrum of described left and right sound channels, recover the high-frequency sub-band territory signal of left and right sound channels respectively according to the high-frequency sub-band field parameter of described left and right sound channels; Perhaps respectively according to the high-frequency sub-band field parameter of described left and right sound channels from the described high-frequency sub-band territory signal that through the low frequency sub-band territory pumping signal of the left and right sound channels that frequency-time domain transformation obtains, recovers left and right sound channels;
D, that the high-frequency sub-band territory signal of the low frequency sub-band territory signal of described left and right sound channels and described left and right sound channels is carried out subband is synthetic, obtains the stereophonic signal of the left and right sound channels of decoding.
28. coding/decoding method as claimed in claim 27 is characterized in that, steps A further comprises obtaining from the acoustic coding code stream of demultiplexing and is used to recover stereosonic low frequency sub-band territory and signal type analysis result;
The described low frequency sub-band territory excitation spectrum with a left side/R channel of step B carries out frequency-time domain transformation and prediction comprehensively comprises: according to described low frequency sub-band territory and signal type analysis result, the described left side of a frame/R channel low frequency sub-band territory excitation spectrum is divided into one or more, as to be used for frequency-time domain transformation subframe; By subframe a described left side/R channel low frequency sub-band territory excitation spectrum is carried out frequency-time domain transformation, obtain a left side/R channel low frequency sub-band territory pumping signal of subframe, according to stripe sequence each subframe combination is generated a frame left side/R channel low frequency sub-band territory pumping signal then;
According to described low frequency sub-band territory and signal type analysis result, with a described frame left side/R channel low frequency sub-band territory pumping signal be divided into one or more, be used to predict comprehensive subframe; By subframe a described left side/R channel low frequency sub-band territory signal is predicted comprehensively, obtained a left side/R channel low frequency sub-band territory signal of subframe.
29. coding/decoding method as claimed in claim 27, it is characterized in that, when described step C recovered the high-frequency sub-band territory signal of a left side/R channel according to the high-frequency sub-band field parameter of a left side/R channel from the low frequency sub-band territory excitation spectrum of a left side/R channel, the high-frequency sub-band field parameter of a described left side/R channel comprised that high frequency LSF vector quantization index, spectrum are adjusted parameter and parameter is adjusted in the time domain gain; Described step C comprises:
C1, adjust parameter according to the spectrum of a described left side/R channel the low frequency sub-band territory excitation spectrum of a described left side/R channel is composed adjustment; Low frequency sub-band territory excitation spectrum to an adjusted left side/R channel carries out frequency-time domain transformation, obtains the low frequency sub-band territory pumping signal of a left side/R channel;
C2, gain is adjusted parameter and the low frequency sub-band territory pumping signal of a described left side/R channel is carried out the time domain gain is adjusted according to the time domain of a described left side/R channel, obtains the low frequency sub-band territory pumping signal of an adjusted left side/R channel;
C3, the linear prediction synthesis filter that adopts the high frequency LSF vector quantization index according to a described left side/R channel to obtain, low frequency sub-band territory pumping signal to an adjusted left side/R channel is carried out integrated filter, the high-frequency sub-band territory signal of a left side/R channel that obtains rebuilding.
30. coding/decoding method as claimed in claim 27 is characterized in that, described steps A comprises that further obtaining the coding mode that is used for stereo decoding from the acoustic coding code stream of demultiplexing selects information;
The described stereo decoding that carries out of step B is: adopt described coding mode to select the stereo decoding pattern of information correspondence, the low frequency sub-band territory stereo coding data of each sub-band k are decoded.
31. coding/decoding method as claimed in claim 30, it is characterized in that, when described stereo decoding pattern was the parameter stereo decoding schema, described steps A further comprised the weighting parameters g that obtains the quantization encoding that is used for stereo decoding from the acoustic coding code stream of demultiplexing r(k) and g d(k), described low frequency sub-band territory stereo coding data are the weighted sum excitation spectrum of quantization encoding; The described stereo decoding that carries out comprises:
B11, with the weighted sum excitation spectrum of described quantization encoding, the weighting parameters g of quantization encoding r(k) and g d(k) carry out the re-quantization decoding; Obtain weighted sum excitation spectrum and weighting parameters g r(k) and g d(k);
B12, the calculating quadrature excitation spectrum vertical with described weighted sum excitation spectrum constant amplitude;
B13, according to g d(k) obtain quadrature excitation spectrum behind the convergent-divergent;
B14, according to the quadrature excitation behind described weighted sum excitation spectrum and described convergent-divergent spectrum, calculate in the left and right acoustic channels the first sound channel low frequency sub-band territory excitation spectrum behind the second sound channel low frequency sub-band territory excitation spectrum and convergent-divergent;
B15, according to g r(k) obtain the first sound channel low frequency sub-band territory excitation spectrum that reduces;
When described stereo decoding pattern was parameter error stereo decoding pattern, steps A further comprised the weighting parameters g that obtains the quantization encoding that is used for stereo decoding from the acoustic coding code stream of demultiplexing r(k) and g d(k), described low frequency sub-band territory stereo coding data are the weighted sum excitation spectrum and the error excitation spectrum of quantization encoding; The described stereo decoding that carries out comprises:
B21, weighted sum excitation spectrum, the error excitation spectrum of quantization encoding, the weighting parameters g of quantization encoding with described quantization encoding r(k) and g d(k) carry out the re-quantization decoding; Obtain weighted sum excitation spectrum, error excitation spectrum and weighting parameters g r(k) and g d(k);
B22, the calculating quadrature excitation spectrum vertical with described weighted sum excitation spectrum constant amplitude;
B23, according to g d(k) obtain quadrature excitation spectrum behind the convergent-divergent;
B24, with the quadrature excitation behind described convergent-divergent spectrum and the addition of described weighted sum excitation spectrum, the weighted difference excitation spectrum that is restored;
B25, according to described weighted sum excitation spectrum and described weighted difference excitation spectrum, calculate in the left and right sound channels the first sound channel low frequency sub-band territory excitation spectrum behind the second sound channel low frequency sub-band territory excitation spectrum and convergent-divergent;
B26, according to g r(k) the described first sound channel low frequency sub-band territory excitation spectrum of reduction.
CN2007100888785A 2007-03-27 2007-03-27 Audio encoding apparatus and method thereof, audio decoding device and method thereof Active CN101276587B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2007100888785A CN101276587B (en) 2007-03-27 2007-03-27 Audio encoding apparatus and method thereof, audio decoding device and method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2007100888785A CN101276587B (en) 2007-03-27 2007-03-27 Audio encoding apparatus and method thereof, audio decoding device and method thereof

Publications (2)

Publication Number Publication Date
CN101276587A true CN101276587A (en) 2008-10-01
CN101276587B CN101276587B (en) 2012-02-01

Family

ID=39995948

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2007100888785A Active CN101276587B (en) 2007-03-27 2007-03-27 Audio encoding apparatus and method thereof, audio decoding device and method thereof

Country Status (1)

Country Link
CN (1) CN101276587B (en)

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101436407B (en) * 2008-12-22 2011-08-24 西安电子科技大学 Method for encoding and decoding audio
CN102177545A (en) * 2009-04-09 2011-09-07 弗兰霍菲尔运输应用研究公司 Apparatus and method for generating a synthesis audio signal and for encoding an audio signal
CN102737636A (en) * 2011-04-13 2012-10-17 华为技术有限公司 Audio coding method and device thereof
CN102006483B (en) * 2009-09-03 2013-05-01 中兴通讯股份有限公司 Video coding and decoding method and device
CN103119648A (en) * 2010-09-22 2013-05-22 杜比实验室特许公司 Efficient implementation of phase shift filtering for decorrelation and other applications in an audio coding system
CN103714822A (en) * 2013-12-27 2014-04-09 广州华多网络科技有限公司 Sub-band coding and decoding method and device based on SILK coder decoder
CN103854650A (en) * 2012-11-30 2014-06-11 中兴通讯股份有限公司 Stereo audio coding method and device
CN104078048A (en) * 2013-03-29 2014-10-01 北京天籁传音数字技术有限公司 Acoustic decoding device and method thereof
CN104103276A (en) * 2013-04-12 2014-10-15 北京天籁传音数字技术有限公司 Sound coding device, sound decoding device, sound coding method and sound decoding method
CN105164749A (en) * 2013-04-30 2015-12-16 杜比实验室特许公司 Hybrid encoding of multichannel audio
CN107134280A (en) * 2013-09-12 2017-09-05 杜比国际公司 The coding of multichannel audio content
CN107221334A (en) * 2016-11-01 2017-09-29 武汉大学深圳研究院 The method and expanding unit of a kind of audio bandwidth expansion
CN109036457A (en) * 2018-09-10 2018-12-18 广州酷狗计算机科技有限公司 Restore the method and apparatus of audio signal
WO2020051786A1 (en) * 2018-09-12 2020-03-19 Shenzhen Voxtech Co., Ltd. Signal processing device having multiple acoustic-electric transducers
CN110941415A (en) * 2019-11-08 2020-03-31 北京达佳互联信息技术有限公司 Audio file processing method and device, electronic equipment and storage medium
CN112365897A (en) * 2020-11-26 2021-02-12 北京百瑞互联技术有限公司 Method, device and medium for self-adaptively adjusting interframe transmission code rate of LC3 encoder
WO2021179788A1 (en) * 2020-03-11 2021-09-16 腾讯科技（深圳）有限公司 Speech signal encoding and decoding methods, apparatuses and electronic device, and storage medium
CN113646836A (en) * 2019-03-27 2021-11-12 诺基亚技术有限公司 Sound field dependent rendering
WO2022012554A1 (en) * 2020-07-17 2022-01-20 华为技术有限公司 Multi-channel audio signal encoding method and apparatus
WO2022012628A1 (en) * 2020-07-17 2022-01-20 华为技术有限公司 Multi-channel audio signal encoding/decoding method and device
US11665482B2 (en) 2011-12-23 2023-05-30 Shenzhen Shokz Co., Ltd. Bone conduction speaker and compound vibration device thereof

Cited By (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101436407B (en) * 2008-12-22 2011-08-24 西安电子科技大学 Method for encoding and decoding audio
CN102177545B (en) * 2009-04-09 2013-03-27 弗兰霍菲尔运输应用研究公司 Apparatus and method for generating a synthesis audio signal and for encoding an audio signal
CN102177545A (en) * 2009-04-09 2011-09-07 弗兰霍菲尔运输应用研究公司 Apparatus and method for generating a synthesis audio signal and for encoding an audio signal
CN102006483B (en) * 2009-09-03 2013-05-01 中兴通讯股份有限公司 Video coding and decoding method and device
CN103119648B (en) * 2010-09-22 2015-06-17 杜比实验室特许公司 Efficient implementation of phase shift filtering for decorrelation and other applications in an audio coding system
CN103119648A (en) * 2010-09-22 2013-05-22 杜比实验室特许公司 Efficient implementation of phase shift filtering for decorrelation and other applications in an audio coding system
WO2012139401A1 (en) * 2011-04-13 2012-10-18 华为技术有限公司 Audio coding method and device
CN102737636B (en) * 2011-04-13 2014-06-04 华为技术有限公司 Audio coding method and device thereof
CN102737636A (en) * 2011-04-13 2012-10-17 华为技术有限公司 Audio coding method and device thereof
US11665482B2 (en) 2011-12-23 2023-05-30 Shenzhen Shokz Co., Ltd. Bone conduction speaker and compound vibration device thereof
CN103854650A (en) * 2012-11-30 2014-06-11 中兴通讯股份有限公司 Stereo audio coding method and device
CN104078048A (en) * 2013-03-29 2014-10-01 北京天籁传音数字技术有限公司 Acoustic decoding device and method thereof
CN104078048B (en) * 2013-03-29 2017-05-03 北京天籁传音数字技术有限公司 Acoustic decoding device and method thereof
CN104103276A (en) * 2013-04-12 2014-10-15 北京天籁传音数字技术有限公司 Sound coding device, sound decoding device, sound coding method and sound decoding method
CN104103276B (en) * 2013-04-12 2017-04-12 北京天籁传音数字技术有限公司 Sound coding device, sound decoding device, sound coding method and sound decoding method
CN105164749B (en) * 2013-04-30 2019-02-12 杜比实验室特许公司 The hybrid coding of multichannel audio
CN105164749A (en) * 2013-04-30 2015-12-16 杜比实验室特许公司 Hybrid encoding of multichannel audio
CN107134280A (en) * 2013-09-12 2017-09-05 杜比国际公司 The coding of multichannel audio content
US11410665B2 (en) 2013-09-12 2022-08-09 Dolby International Ab Methods and apparatus for decoding encoded audio signal(s)
US11776552B2 (en) 2013-09-12 2023-10-03 Dolby International Ab Methods and apparatus for decoding encoded audio signal(s)
CN107134280B (en) * 2013-09-12 2020-10-23 杜比国际公司 Encoding of multi-channel audio content
CN103714822A (en) * 2013-12-27 2014-04-09 广州华多网络科技有限公司 Sub-band coding and decoding method and device based on SILK coder decoder
CN107221334A (en) * 2016-11-01 2017-09-29 武汉大学深圳研究院 The method and expanding unit of a kind of audio bandwidth expansion
CN109036457B (en) * 2018-09-10 2021-10-08 广州酷狗计算机科技有限公司 Method and apparatus for restoring audio signal
CN109036457A (en) * 2018-09-10 2018-12-18 广州酷狗计算机科技有限公司 Restore the method and apparatus of audio signal
WO2020051786A1 (en) * 2018-09-12 2020-03-19 Shenzhen Voxtech Co., Ltd. Signal processing device having multiple acoustic-electric transducers
US11875815B2 (en) 2018-09-12 2024-01-16 Shenzhen Shokz Co., Ltd. Signal processing device having multiple acoustic-electric transducers
US11373671B2 (en) 2018-09-12 2022-06-28 Shenzhen Shokz Co., Ltd. Signal processing device having multiple acoustic-electric transducers
CN113646836A (en) * 2019-03-27 2021-11-12 诺基亚技术有限公司 Sound field dependent rendering
CN110941415A (en) * 2019-11-08 2020-03-31 北京达佳互联信息技术有限公司 Audio file processing method and device, electronic equipment and storage medium
CN110941415B (en) * 2019-11-08 2023-11-28 北京达佳互联信息技术有限公司 Audio file processing method and device, electronic equipment and storage medium
WO2021179788A1 (en) * 2020-03-11 2021-09-16 腾讯科技（深圳）有限公司 Speech signal encoding and decoding methods, apparatuses and electronic device, and storage medium
WO2022012628A1 (en) * 2020-07-17 2022-01-20 华为技术有限公司 Multi-channel audio signal encoding/decoding method and device
WO2022012554A1 (en) * 2020-07-17 2022-01-20 华为技术有限公司 Multi-channel audio signal encoding method and apparatus
CN112365897A (en) * 2020-11-26 2021-02-12 北京百瑞互联技术有限公司 Method, device and medium for self-adaptively adjusting interframe transmission code rate of LC3 encoder

Also Published As

Publication number Publication date
CN101276587B (en) 2012-02-01

Similar Documents

Publication Publication Date Title
CN101276587B (en) Audio encoding apparatus and method thereof, audio decoding device and method thereof
JP4950210B2 (en) Audio compression
CN101067931B (en) Efficient configurable frequency domain parameter stereo-sound and multi-sound channel coding and decoding method and system
CN101086845B (en) Sound coding device and method and sound decoding device and method
KR101589942B1 (en) Cross product enhanced harmonic transposition
CN102177426B (en) Multi-resolution switched audio encoding/decoding scheme
CN101089951B (en) Band spreading coding method and device and decode method and device
CN103366749B (en) A kind of sound codec devices and methods therefor
EP1852851A1 (en) An enhanced audio encoding/decoding device and method
KR101411901B1 (en) Method of Encoding/Decoding Audio Signal and Apparatus using the same
CN109448741B (en) 3D audio coding and decoding method and device
CN103366750B (en) A kind of sound codec devices and methods therefor
CN101432802A (en) Method and apparatus for lossless encoding of a source signal, using a lossy encoded data stream and a lossless extension data stream
CN103366755A (en) Method and apparatus for encoding and decoding audio signal
CN101615393B (en) Method and apparatus for encoding or decoding a speech and/or non-speech audio input signal
CN103366751B (en) A kind of sound codec devices and methods therefor
CN102460574A (en) Method and apparatus for encoding and decoding audio signal using hierarchical sinusoidal pulse coding
RU2414009C2 (en) Signal encoding and decoding device and method
RU2409874C2 (en) Audio signal compression
US20100280830A1 (en) Decoder
Sinha et al. A fractal self-similarity model for the spectral representation of audio signals
CN105336334B (en) Multi-channel sound signal coding method, decoding method and device
CN111210832B (en) Bandwidth expansion audio coding and decoding method and device based on spectrum envelope template
CN111210831B (en) Bandwidth extension audio encoding and decoding method and device based on spectrum stretching

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
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20081001

Assignee: Pan Xingde

Assignor: Beijing Tianlai Chuanyin Digital Technology Co., Ltd.

Contract record no.: 2013990000772

Denomination of invention: Audio encoding apparatus and method thereof, audio decoding device and method thereof

Granted publication date: 20120201

License type: Common License

Record date: 20131119

LICC Enforcement, change and cancellation of record of contracts on the licence for exploitation of a patent or utility model