CN101162584A - Method and apparatus to encode and decode audio signal by using bandwidth extension technique - Google Patents

Method and apparatus to encode and decode audio signal by using bandwidth extension technique Download PDF

Info

Publication number
CN101162584A
CN101162584A CNA2007101533639A CN200710153363A CN101162584A CN 101162584 A CN101162584 A CN 101162584A CN A2007101533639 A CNA2007101533639 A CN A2007101533639A CN 200710153363 A CN200710153363 A CN 200710153363A CN 101162584 A CN101162584 A CN 101162584A
Authority
CN
China
Prior art keywords
frequency
signal
coding
low band
band 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.)
Pending
Application number
CNA2007101533639A
Other languages
Chinese (zh)
Inventor
吴殷美
朱基岘
苗磊
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.)
Samsung Electronics Co Ltd
Original Assignee
Samsung Electronics 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 Samsung Electronics Co Ltd filed Critical Samsung Electronics Co Ltd
Publication of CN101162584A publication Critical patent/CN101162584A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G10L19/0204—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders using subband decomposition
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G10L19/032—Quantisation or dequantisation of spectral components
    • G10L19/035—Scalar quantisation
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
    • G10L19/16—Vocoder architecture
    • G10L19/18—Vocoders using multiple modes
    • G10L19/24—Variable rate codecs, e.g. for generating different qualities using a scalable representation such as hierarchical encoding or layered encoding

Abstract

Provided are a method and apparatus to encode and decode an audio signal. According to the present invention, by splitting an input signal into a low band signal and a high band signal, converting each of the low band signal and the high band signal from the time domain to the frequency domain, performing quantization and context-dependant bitplane encoding on the converted low band signal, generating and encoding bandwidth extension information that represents a characteristic of the converted high band signal, and outputting the encoded bitplane and the encoded bandwidth extension information as an encoded result of the input signal, high frequency components may be efficiently encoded at a restricted bit rate, thereby improving the quality of an audio signal.

Description

The utilized bandwidth expansion technique is to the method and apparatus of audio-frequency signal coding and decoding
Technical field
The present invention relates to a kind of method and apparatus that is used for audio-frequency signal coding and decoding, more particularly, relate to the method and apparatus of a kind of utilized bandwidth expansion technique audio-frequency signal coding and decoding.
Background technology
When to audio-frequency signal coding or when decoding, need make the best in qualityization of sound signal by using limited bit rate.The bit quantity that can use when low bit rate is less, thereby frequency bandwidth that must be by reducing sound signal is to audio-frequency signal coding or decoding.Therefore, the quality of sound signal may deterioration.
Usually, discern sound signal for people, low frequency component is more important than high fdrequency component.Therefore, need a kind of distributing to be used for the bit quantity that low frequency component is encoded and to reduce to distribute to the method for quality that is used for the bit quantity of high fdrequency component coding is improved sound signal by increase.
Summary of the invention
The invention provides a kind of method and apparatus to audio-frequency signal coding, wherein, with limited bit rate efficiently to high fdrequency component coding, thereby improved the quality of sound signal.
The present invention also provides a kind of method and apparatus of efficiently high fdrequency component being decoded from bitstream encoded with limited bit rate.
According to an aspect of the present invention, provide a kind of method, said method comprising the steps of: (a) input signal has been separated into low band signal and high-frequency band signals audio-frequency signal coding; (b) in low band signal and the high-frequency band signals each is transformed into frequency domain from time domain; (c) to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding; (d) by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding; (e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
According to a further aspect in the invention, a kind of computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio-frequency signal coding that records on it is provided, has said method comprising the steps of: (a) input signal is separated into low band signal and high-frequency band signals; (b) in low band signal and the high-frequency band signals each is transformed into frequency domain from time domain; (c) to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding; (d) by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding; (e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
According to a further aspect in the invention, provide a kind of method, said method comprising the steps of: (a) input signal has been separated into low band signal and high-frequency band signals audio-frequency signal coding; (b) by low band signal is carried out the discrete cosine transform of revising (MDCT), low band signal is transformed into frequency domain from time domain; (c) low band signal of having carried out MDCT carry out is quantized and based on contextual Bit-Plane Encoding; (d) in low band signal and the high-frequency band signals each is transformed into frequency domain or time/frequency domain from time domain; (e) by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding; (f) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
According to a further aspect in the invention, a kind of computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio-frequency signal coding that records on it is provided, has said method comprising the steps of: (a) input signal is separated into low band signal and high-frequency band signals; (b) by low band signal is carried out the discrete cosine transform of revising (MDCT), low band signal is transformed into frequency domain from time domain; (c) low band signal of having carried out MDCT carry out is quantized and based on contextual Bit-Plane Encoding; (d) in low band signal and the high-frequency band signals each is transformed into frequency domain or time/frequency domain from time domain; (e) by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding; (f) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
According to a further aspect in the invention, provide a kind of method, said method comprising the steps of: (a) input signal is transformed into frequency domain from time domain audio-frequency signal coding; (b) input signal with conversion is separated into low band signal and high-frequency band signals; (c) to low band signal execution quantification with based on contextual Bit-Plane Encoding; (d) by using low band signal to produce the bandwidth extend information of the characteristic of representing high-frequency band signals and described bandwidth extend information being encoded; (e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
According to a further aspect in the invention, provide a kind of computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio-frequency signal coding that records on it, said method comprising the steps of: (a) input signal is transformed into frequency domain from time domain; (b) input signal with conversion is separated into low band signal and high-frequency band signals; (c) to low band signal execution quantification with based on contextual Bit-Plane Encoding; (d) by using low band signal to produce the bandwidth extend information of the characteristic of representing high-frequency band signals and described bandwidth extend information being encoded; (e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
According to a further aspect in the invention, provide a kind of method, said method comprising the steps of: (a) sound signal of received code audio signal decoding; (b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal; (c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from low band signal by the bandwidth extend information of using decoding; (d), in low band signal and the high-frequency band signals each is transformed into time domain from frequency domain by in low band signal and the high-frequency band signals each is carried out the contrary discrete cosine transform of revising (MDCT); (e) with the low band signal of conversion and the high-frequency band signals combination of conversion.
According to a further aspect in the invention, provide a kind of computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio signal decoding that records on it, said method comprising the steps of: (a) sound signal of received code; (b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal; (c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals (MDCT) from low band signal by the bandwidth extend information of using decoding; (d), in low band signal and the high-frequency band signals each is transformed into time domain from frequency domain by in low band signal and the high-frequency band signals each is carried out the contrary discrete cosine transform of revising; (e) with the low band signal of conversion and the high-frequency band signals combination of conversion.
According to a further aspect in the invention, provide a kind of method, said method comprising the steps of: (a) sound signal of received code audio signal decoding; (b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal; (c) by low band signal being carried out the contrary discrete cosine transform of revising (MDCT), low band signal is transformed into time domain from frequency domain; (d) low band signal that will carry out contrary MDCT is transformed into frequency domain or time/frequency domain from time domain; (e) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from the low band signal that is transformed into frequency domain or time/frequency domain by the bandwidth extend information of using decoding; (f) with the high-frequency band signals inverse conversion to time domain; (g) low band signal of conversion and the high-frequency band signals of inverse conversion are made up.
According to a further aspect in the invention, provide a kind of computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio signal decoding that records on it, said method comprising the steps of: (a) sound signal of received code; (b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal; (c) by low band signal being carried out the contrary discrete cosine transform of revising (MDCT), low band signal is transformed into time domain from frequency domain; (d) low band signal that will carry out contrary MDCT is transformed into frequency domain or time/frequency domain from time domain; (e) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from the low band signal that is transformed into frequency domain or time/frequency domain by the bandwidth extend information of using decoding; (f) with the high-frequency band signals inverse conversion to time domain; (g) low band signal of conversion and the high-frequency band signals of inverse conversion are made up.
According to a further aspect in the invention, provide a kind of method, said method comprising the steps of: (a) sound signal of received code audio signal decoding; (b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal; (c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from low band signal by the bandwidth extend information of using decoding; (d) with low band signal and high-frequency band signals combination; (e) pass through the signal of combination is carried out the contrary discrete cosine transform of revising (MDCT), the signal that makes up is transformed into time domain from frequency domain.
According to a further aspect in the invention, provide a kind of computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio signal decoding that records on it, said method comprising the steps of: (a) sound signal of received code; (b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal; (c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from low band signal by the bandwidth extend information of using decoding; (d) with low band signal and high-frequency band signals combination; (e) pass through the signal of combination is carried out the contrary discrete cosine transform of revising (MDCT), the signal that makes up is transformed into time domain from frequency domain.
According to a further aspect in the invention, provide a kind of equipment that is used for audio-frequency signal coding, described equipment comprises: the frequency band separative element is separated into low band signal and high-frequency band signals with input signal; Converting unit is transformed into frequency domain with in low band signal and the high-frequency band signals each from time domain; The low-frequency band coding unit is to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding; The band spread coding unit, by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding.
According to a further aspect in the invention, provide a kind of equipment that is used for audio-frequency signal coding, described equipment comprises: the frequency band separative element is separated into low band signal and high-frequency band signals with input signal; Discrete cosine transform (MDCT) applying unit of revising by low band signal is carried out MDCT, is transformed into frequency domain with low band signal from time domain; The low-frequency band coding unit carry out to quantize and based on contextual Bit-Plane Encoding the low band signal of having carried out MDCT; Converting unit is transformed into frequency domain or time/frequency domain with in low band signal and the high-frequency band signals each from time domain; Bandwidth extended coding unit, by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding.
According to a further aspect in the invention, provide a kind of equipment that is used for audio-frequency signal coding, described equipment comprises: converting unit is transformed into frequency domain with input signal from time domain; The frequency band separative element is separated into low band signal and high-frequency band signals with the input signal of changing; The low-frequency band coding unit is to low band signal execution quantification with based on contextual Bit-Plane Encoding; Bandwidth extended coding unit is by using low band signal to produce the bandwidth extend information of the characteristic of representing high-frequency band signals and described bandwidth extend information being encoded.
According to a further aspect in the invention, provide a kind of equipment that is used for audio signal decoding, described equipment comprises: the low-frequency band decoding unit by the bit plane of coding is carried out based on contextual decoding and re-quantization, produces low band signal; Bandwidth expansion decoding unit, to the bandwidth extend information decoding of coding, and by using the bandwidth extend information of decoding to produce high-frequency band signals from low band signal; Contrary discrete cosine transform (MDCT) applying unit of revising by in low band signal and the high-frequency band signals each is carried out contrary MDCT, is transformed into time domain with in low band signal and the high-frequency band signals each from frequency domain; The combinations of bands unit is with the low band signal of conversion and the high-frequency band signals combination of conversion.
According to a further aspect in the invention, provide a kind of equipment that is used for audio signal decoding, described equipment comprises: the low-frequency band decoding unit by the bit plane of coding is carried out based on contextual decoding and re-quantization, produces low band signal; Contrary discrete cosine transform (MDCT) applying unit of revising by low band signal being carried out contrary MDCT, is transformed into time domain with low band signal from frequency domain; Converting unit is transformed into frequency domain or time/frequency domain with the low band signal of having carried out contrary MDCT from time domain; Bandwidth expansion decoding unit, to the bandwidth extend information decoding of coding, and by using the bandwidth extend information of decoding to produce high-frequency band signals from the low band signal that is transformed into frequency domain or time/frequency domain; The inverse conversion unit, with the high-frequency band signals inverse conversion to time domain; The combinations of bands unit makes up the low band signal of conversion and the high-frequency band signals of inverse conversion.
According to a further aspect in the invention, provide a kind of equipment that is used for audio signal decoding, described equipment comprises: the low-frequency band decoding unit by the bit plane of coding is carried out based on contextual decoding and re-quantization, produces low band signal; Bandwidth expansion decoding unit, to the bandwidth extend information decoding of coding, and by using the bandwidth extend information of decoding to produce high-frequency band signals from low band signal; The combinations of bands unit is with low band signal and high-frequency band signals combination; Against discrete cosine transform (MDCT) applying unit of revising,, the signal that makes up is transformed into time domain from frequency domain by the signal of combination is carried out contrary MDCT.
Description of drawings
To the detailed description of exemplary embodiment, above-mentioned and other characteristics of the present invention and advantage will become more obvious by with reference to the accompanying drawings, wherein:
Fig. 1 is the block diagram to the equipment of coding audio signal according to the embodiment of the invention;
Fig. 2 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal;
Fig. 3 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal;
Fig. 4 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal;
Fig. 5 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal;
Fig. 6 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal;
Fig. 7 is the block diagram according to the equipment that sound signal is decoded of the embodiment of the invention;
Fig. 8 is the block diagram of the equipment of according to another embodiment of the present invention sound signal being decoded;
Fig. 9 is the block diagram of the equipment of according to another embodiment of the present invention sound signal being decoded;
Figure 10 is the block diagram of the equipment of according to another embodiment of the present invention sound signal being decoded;
Figure 11 is the block diagram of the equipment of according to another embodiment of the present invention sound signal being decoded;
Figure 12 is the block diagram of the equipment of according to another embodiment of the present invention sound signal being decoded;
Figure 13 is the process flow diagram to the method for coding audio signal according to the embodiment of the invention;
Figure 14 is according to another embodiment of the present invention to the process flow diagram of the method for coding audio signal;
Figure 15 is according to another embodiment of the present invention to the process flow diagram of the method for coding audio signal;
Figure 16 is the process flow diagram according to the method that sound signal is decoded of the embodiment of the invention;
Figure 17 is the process flow diagram of the method for according to another embodiment of the present invention sound signal being decoded;
Figure 18 is the process flow diagram of the method for according to another embodiment of the present invention sound signal being decoded.
Embodiment
Owing to provide structure or functional description describing exemplary embodiment of the present invention, so the present invention can implement and the embodiment that should not be construed as limited to here to be set forth with many different forms.
Now, describe the present invention with reference to the accompanying drawings more fully, exemplary embodiment of the present invention has been shown in the accompanying drawing.These exemplary embodiments only should be considered to for descriptive purpose, and non-limiting purpose, all differences in the scope all will be interpreted as comprising in the present invention.Same numeral in the accompanying drawing is represented identical parts.
Unless otherwise defined, employed all terms all have (comprising technical term and scientific terminology) implication of those of ordinary skills' common sense otherwise in describing.The term that defines in the common dictionary should be interpreted as having with background of related in the identical implication of implication of term, unless and in description, define, otherwise described term can not or be interpreted as having the formalization implication terrifically by idealized ground.
Below, describe exemplary embodiment of the present invention with reference to the accompanying drawings in detail.Same numeral in the accompanying drawing is represented same parts, therefore will omit the description that repeats.
Fig. 1 is the block diagram to the equipment of coding audio signal according to the embodiment of the invention.
With reference to Fig. 1, this equipment comprises frequency band separative element 100, first discrete cosine transform (MDCT) applying unit of revising 110, frequency linearity prediction execution unit 120, many resolution decomposition (multi-resolution analysis) unit 130, quantifying unit 140, the back quantizes (PQ-SPSC) module 150 of rectangular coordinate polar coordinates stereo codings (post-quantization square polar stereo coding), based on contextual bit plane (context-dependent bitplane) coding unit 160, the 2nd MDCT applying unit 170, bandwidth extended coding unit 180 and Multiplexing Unit 190.
Frequency band separative element 100 is separated into low band signal LB and high-frequency band signals HB with input signal IN.Here, input signal IN can be that wherein analog voice or sound signal are modulated to the pulse code modulation (pcm) signal of digital signal, low band signal LB can be the frequency signal that is lower than predetermined threshold, and high-frequency band signals HB can be the frequency signal that is higher than described predetermined threshold.
110 pairs of the one MDCT applying units are carried out MDCT by frequency band separative element 100 isolated low band signal LB, thereby low band signal LB is transformed into frequency domain from time domain.
120 pairs of low band signal LB execution frequency linearity predictions that are transformed into frequency domain by a MDCT applying unit 110 of frequency linearity prediction execution unit.Here, the frequency linearity prediction is approximately current frequency signal the linear combination of previous frequency signal.More particularly, frequency linearity prediction execution unit 120 calculate the coefficient of linear prediction filters so that predicated error (signal of linear prediction and current frequency signal poor) minimum, and the low band signal LB that is switched to frequency domain is carried out linear prediction filtering according to the coefficient that calculates.Here, frequency linearity prediction execution unit 120 can be passed through the analog value of the coefficient of linear prediction filter is carried out vector quantization (vector quantization) so that with vector index (vector index) expression analog value, improve code efficiency.
In more detail, if the low band signal LB that is transformed into frequency domain by a MDCT applying unit 110 is voice or fundamental tone (pitched) signal, then frequency linearity prediction execution unit 120 can be carried out the frequency linearity prediction to this voice signal or pitch signal.That is, frequency linearity prediction execution unit 120 can improve code efficiency by predicting according to the characteristic execution frequency linearity of the signal that receives.
Many resolution decomposition unit 130 receives and is transformed into the low band signal LB of frequency domain by a MDCT applying unit 110 or from the result of frequency linearity prediction execution unit 120 outputs, and audible spectrum (audio spectrum) coefficient of the signal that receives of transient change is carried out many resolution decomposition.More particularly, many resolution decomposition unit 130 can by the intensity that changes according to audible spectrum will by the audible spectrum of frequency linearity prediction execution unit 120 filtering be divided into two types (as, stability types and short type), this audible spectrum is carried out many resolution decomposition.
More particularly, if be transformed into the low band signal LB of frequency domain by a MDCT applying unit 110 or be transient signal from the result of frequency linearity prediction execution unit 120 output, then many resolution decomposition can be carried out to this transient signal in many resolution decomposition unit 130.That is, many resolution decomposition unit 130 can improve code efficiency by carrying out many resolution decomposition according to the characteristic of the signal that receives.
Quantifying unit is 140 couples of results from frequency linearity prediction execution unit 120 or 130 outputs of many resolution decomposition unit quantize.
150 pairs of frequency spectrums from the result of quantifying unit 140 outputs of PQ-SPSC module are carried out limit polar coordinates (side-polar coordination) stereo coding.
160 couples of results from 150 outputs of PQ-SPSC module carry out based on contextual Bit-Plane Encoding based on contextual Bit-Plane Encoding unit.Here, it is described based on contextual Bit-Plane Encoding to use the Huffman coding method to carry out based on contextual Bit-Plane Encoding unit 160.
Therefore frequency linearity prediction execution unit 120, many resolution decomposition unit 130, quantifying unit 140, PQ-SPSC module 150 and encode based on the 160 couples of low band signal LB from a MDCT applying unit 110 outputs in contextual Bit-Plane Encoding unit can be called the low-frequency band coding unit jointly.
170 pairs of the 2nd MDCT applying units are carried out MDCT by frequency band separative element 100 isolated high-frequency band signals HB, thereby high-frequency band signals HB is transformed into frequency domain from time domain.
180 uses of bandwidth extended coding unit are transformed into the low band signal LB generation bandwidth extend information of frequency domain by a MDCT applying unit 110 and this information are encoded, and described bandwidth extend information represents to be transformed into by the 2nd MDCT applying unit 170 characteristic of the high-frequency band signals HB of frequency domain.The bandwidth extend information can comprise the multiple information of high-frequency band signals HB, as energy level and envelope.More particularly, bandwidth extended coding unit 180 can use low band signal LB to produce the bandwidth extend information based on the fact that has strong correlation between low band signal LB and the high-frequency band signals HB.
Multiplexing Unit 190 is by to frequency linearity prediction execution unit 120, PQ-SPSC module 150, carry out the multiplexing bit stream that produces based on the coding result of contextual Bit-Plane Encoding unit 160 and bandwidth extended coding unit 180, thereby this bit stream is exported as output signal OUT.
Fig. 2 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal.
With reference to Fig. 2, this equipment comprises frequency band separative element 200, MDCT applying unit 210, frequency linearity prediction execution unit 220, many resolution decomposition unit 230, quantifying unit 240, PQ-SPSC module 250, based on contextual Bit-Plane Encoding unit 260, low-frequency band converting unit 270, high frequency band converting unit 275, bandwidth extended coding unit 280 and Multiplexing Unit 290.
Frequency band separative element 200 is separated into low band signal LB and high-frequency band signals HB with input signal IN.Here, input signal IN can be that wherein analog voice or sound signal are modulated to the PCM signal of digital signal, and low band signal LB can be the frequency signal that is lower than predetermined threshold, and high-frequency band signals HB can be the frequency signal that is higher than described predetermined threshold.
210 pairs of MDCT applying units are carried out MDCT by frequency band separative element 200 isolated low band signal LB, thereby low band signal LB is transformed into frequency domain from time domain.
220 pairs of low band signal LB execution frequency linearity predictions that are transformed into frequency domain by MDCT applying unit 210 of frequency linearity prediction execution unit.Here, the frequency linearity prediction is approximately current frequency signal the linear combination of previous frequency signal.More particularly, frequency linearity prediction execution unit 220 calculate the coefficient of linear prediction filters so that predicated error (signal of linear prediction and current frequency signal poor) minimum, and the low band signal LB that is switched to frequency domain is carried out linear prediction filtering according to the coefficient that calculates.Here, frequency linearity prediction execution unit 220 can improve code efficiency by the analog value of the coefficient of linear prediction filter is carried out vector quantization so that represent analog value with vector index.
More particularly, if the low band signal LB that is transformed into frequency domain by MDCT applying unit 210 is voice or pitch signal, then frequency linearity prediction execution unit 220 can be carried out the frequency linearity prediction to this voice signal or pitch signal.That is, frequency linearity prediction execution unit 220 can improve code efficiency by predicting according to the characteristic execution frequency linearity of the signal that receives.
Many resolution decomposition unit 230 receives from the result of frequency linearity prediction execution unit 220 outputs, and the audible spectrum coefficient of the signal that receives of transient change is carried out many resolution decomposition.More particularly, many resolution decomposition unit 230 can by the intensity that changes according to audible spectrum will by the audible spectrum of frequency linearity prediction execution unit 220 filtering be divided into two types (as, stability types and short type), this audible spectrum is carried out many resolution decomposition.
More particularly, if be transformed into the low band signal LB of frequency domain by MDCT applying unit 210 or be transient signal from the result of frequency linearity prediction execution unit 220 output, then many resolution decomposition can be carried out to this transient signal in many resolution decomposition unit 230.That is, many resolution decomposition unit 230 can improve code efficiency by carrying out many resolution decomposition according to the characteristic of the signal that receives.
Quantifying unit is 240 couples of results from frequency linearity prediction execution unit 220 or 230 outputs of many resolution decomposition unit quantize.
250 pairs of frequency spectrums from the result of quantifying unit 240 outputs of PQ-SPSC module are carried out limit polar coordinates stereo coding.
260 couples of results from 250 outputs of PQ-SPSC module carry out based on contextual Bit-Plane Encoding based on contextual Bit-Plane Encoding unit.Here, it is described based on contextual Bit-Plane Encoding to use the Huffman coding method to carry out based on contextual Bit-Plane Encoding unit 260.
Therefore frequency linearity prediction execution unit 220, many resolution decomposition unit 230, quantifying unit 240, PQ-SPSC module 250 and encode based on the 260 couples of low band signal LB from 210 outputs of MDCT applying unit in contextual Bit-Plane Encoding unit can be called the low-frequency band coding unit jointly.
The conversion method that low-frequency band converting unit 270 is used except the MDCT method will be transformed into frequency domain or time/frequency domain from time domain by frequency band separative element 200 isolated low band signal LB.For example, low-frequency band converting unit 270 can use discrete sine transform (MDST) method, fast Fourier transform (FFT) method or quadrature mirror filter (QMF) method of correction that low band signal LB is transformed into frequency domain or time/frequency domain from time domain.
The conversion method that high frequency band converting unit 275 is used except the MDCT method will be transformed into frequency domain or time/frequency domain from time domain by frequency band separative element 200 isolated high-frequency band signals HB.Here, high frequency band converting unit 275 is used identical conversion method with low-frequency band converting unit 270.For example, high frequency band converting unit 275 can be used MDST method, FFT method or QMF method.
280 uses of bandwidth extended coding unit are transformed into the low band signal LB generation bandwidth extend information of frequency domain by low-frequency band converting unit 270 and this information are encoded, and described bandwidth extend information represents to be transformed into by high frequency band converting unit 275 characteristic of the high-frequency band signals HB of frequency domain.The bandwidth extend information can comprise the multiple information of high-frequency band signals HB, as energy level and the envelope of high-frequency band signals HB.More particularly, bandwidth extended coding unit 280 can use low band signal LB to produce the bandwidth extend information based on the fact that has strong correlation between low band signal LB and the high-frequency band signals HB.
Multiplexing Unit 290 is by to frequency linearity prediction execution unit 220, PQ-SPSC module 250, carry out the multiplexing bit stream that produces based on the coding result of contextual Bit-Plane Encoding unit 260 and bandwidth extended coding unit 280, thereby this bit stream is exported as output signal OUT.
Fig. 3 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal.
With reference to Fig. 3, this equipment comprises MDCT applying unit 300, frequency band separative element 310, frequency linearity prediction execution unit 320, many resolution decomposition unit 330, quantifying unit 340, PQ-SPSC module 350, based on contextual Bit-Plane Encoding unit 360, bandwidth extended coding unit 370 and Multiplexing Unit 380.
300 couples of input signal IN of MDCT applying unit carry out MDCT, so that input signal IN is transformed into frequency domain from time domain.Here, input signal can be that wherein analog voice or sound signal are modulated to the PCM signal of digital signal.
Frequency band separative element 310 will be separated into low band signal LB and high-frequency band signals HB by the input signal IN that MDCT applying unit 300 is transformed into frequency domain.Here, low band signal LB can be the frequency signal that is lower than predetermined threshold, and high-frequency band signals HB can be the frequency signal that is higher than described predetermined threshold.
320 pairs of frequency linearity prediction execution unit are carried out frequency linearity by frequency band separative element 310 isolated low band signal LB and are predicted.Here, the frequency linearity prediction is approximately current frequency signal the linear combination of previous frequency signal.More particularly, frequency linearity prediction execution unit 320 calculate the coefficient of linear prediction filters so that predicated error (signal of linear prediction and current frequency signal poor) minimum, and the low band signal LB that is switched to frequency domain is carried out linear prediction filtering according to the coefficient that calculates.Here, frequency linearity prediction execution unit 320 can improve code efficiency by the analog value of the coefficient of linear prediction filter is carried out vector quantization so that represent analog value with vector index.
More particularly, if be voice or pitch signal by frequency band separative element 310 isolated low band signal LB, then frequency linearity prediction execution unit 320 can be carried out the frequency linearity prediction to this voice signal or pitch signal.That is, frequency linearity prediction execution unit 320 can improve code efficiency by predicting according to the characteristic execution frequency linearity of the signal that receives.
Many resolution decomposition unit 330 receives from the result of frequency linearity prediction execution unit 320 outputs, and the audible spectrum coefficient of the signal that receives of transient change is carried out many resolution decomposition.More particularly, many resolution decomposition unit 330 can by the intensity that changes according to audible spectrum will by the audible spectrum of frequency linearity prediction execution unit 320 filtering be divided into two types (as, stability types and short type), this audible spectrum is carried out many resolution decomposition.
More particularly, if be transient signal by frequency band separative element 310 isolated low band signal LB or from the result of frequency linearity prediction execution unit 320 output, then many resolution decomposition can be carried out to this transient signal in many resolution decomposition unit 330.That is, many resolution decomposition unit 330 can improve code efficiency by carrying out many resolution decomposition according to the characteristic of the signal that receives.
Quantifying unit is 340 couples of results from frequency linearity prediction execution unit 320 or 330 outputs of many resolution decomposition unit quantize.
350 pairs of frequency spectrums from the result of quantifying unit 340 outputs of PQ-SPSC module are carried out limit polar coordinates stereo coding.
360 couples of results from 350 outputs of PQ-SPSC module carry out based on contextual Bit-Plane Encoding based on contextual Bit-Plane Encoding unit.Here, it is described based on contextual Bit-Plane Encoding to use the Huffman coding method to carry out based on contextual Bit-Plane Encoding unit 360.
Therefore frequency linearity prediction execution unit 320, many resolution decomposition unit 330, quantifying unit 340, PQ-SPSC module 350 and encode based on the 360 couples of low band signal LB from 310 outputs of frequency band separative element in contextual Bit-Plane Encoding unit can be called the low-frequency band coding unit jointly.
Bandwidth extended coding unit 370 uses by frequency band separative element 310 isolated low band signal LB generation bandwidth extend informations and to this information and encodes, and described bandwidth extend information is represented the characteristic by frequency band separative element 310 isolated high-frequency band signals HB.The bandwidth extend information can comprise the multiple information of high-frequency band signals HB, as energy level and envelope.More particularly, bandwidth extended coding unit 370 can use low band signal LB to produce the bandwidth extend information based on the fact that has strong correlation between low band signal LB and the high-frequency band signals HB.
Multiplexing Unit 380 is by to frequency linearity prediction execution unit 320, PQ-SPSC module 350, carry out the multiplexing bit stream that produces based on the coding result of contextual Bit-Plane Encoding unit 360 and bandwidth extended coding unit 370, thereby this bit stream is exported as output signal OUT.
Fig. 4 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal.
With reference to Fig. 4, this equipment comprises frequency band separative element 400, a MDCT applying unit 410, frequency linearity prediction execution unit 420, many resolution decomposition unit 430, quantifying unit 440, based on contextual Bit-Plane Encoding unit 450, the 2nd MDCT applying unit 460, bandwidth extended coding unit 470 and Multiplexing Unit 480.
Frequency band separative element 400 is separated into low band signal LB and high-frequency band signals HB with input signal IN.Here, input signal IN can be that wherein analog voice or sound signal are modulated to the PCM signal of digital signal, and low band signal LB can be the frequency signal that is lower than predetermined threshold, and high-frequency band signals HB can be the frequency signal that is higher than described predetermined threshold.
410 pairs of the one MDCT applying units are carried out MDCT by frequency band separative element 400 isolated low band signal LB, thereby low band signal LB is transformed into frequency domain from time domain.Here, the amplitude of time-domain representation input signal IN (as energy or acoustic pressure) over time, and the amplitude of frequency domain representation input signal IN is with the variation of frequency.
420 pairs of low band signal LB execution frequency linearity predictions that are transformed into frequency domain by a MDCT applying unit 410 of frequency linearity prediction execution unit.Here, the frequency linearity prediction is approximately current frequency signal the linear combination of previous frequency signal.More particularly, frequency linearity prediction execution unit 420 calculate the coefficient of linear prediction filters so that predicated error (signal of linear prediction and current frequency signal poor) minimum, and the low band signal LB that is switched to frequency domain is carried out linear prediction filtering according to the coefficient that calculates.Here, frequency linearity prediction execution unit 420 can improve code efficiency by the analog value of the coefficient of linear prediction filter is carried out vector quantization so that represent analog value with vector index.
More particularly, if the low band signal LB that is transformed into frequency domain by a MDCT applying unit 410 is voice or pitch signal, then frequency linearity prediction execution unit 420 can be carried out the frequency linearity prediction to this voice signal or pitch signal.That is, frequency linearity prediction execution unit 420 can improve code efficiency by predicting according to the characteristic execution frequency linearity of the signal that receives.
Many resolution decomposition unit 430 receives from the result of frequency linearity prediction execution unit 420 outputs, and the audible spectrum coefficient of the signal that receives of transient change is carried out many resolution decomposition.More particularly, many resolution decomposition unit 430 can by the intensity that changes according to audible spectrum will by the audible spectrum of frequency linearity prediction execution unit 420 filtering be divided into two types (as, stability types and short type), this audible spectrum is carried out many resolution decomposition.
More particularly, if be transformed into the low band signal LB of frequency domain by a MDCT applying unit 410 or be transient signal from the result of frequency linearity prediction execution unit 420 output, then many resolution decomposition can be carried out to this transient signal in many resolution decomposition unit 430.That is, many resolution decomposition unit 430 can improve code efficiency by carrying out many resolution decomposition according to the characteristic of the signal that receives.
Quantifying unit is 440 couples of results from frequency linearity prediction execution unit 420 or 430 outputs of many resolution decomposition unit quantize.
450 couples of results from quantifying unit 440 outputs carry out based on contextual Bit-Plane Encoding based on contextual Bit-Plane Encoding unit.Here, it is described based on contextual Bit-Plane Encoding to use the Huffman coding method to carry out based on contextual Bit-Plane Encoding unit 450.
Therefore frequency linearity prediction execution unit 420, many resolution decomposition unit 430, quantifying unit 440 and encode based on the 450 couples of low band signal LB from a MDCT applying unit 410 outputs in contextual Bit-Plane Encoding unit can be called the low-frequency band coding unit jointly.
460 pairs of the 2nd MDCT applying units are carried out MDCT by frequency band separative element 400 isolated high-frequency band signals HB, thereby high-frequency band signals HB is transformed into frequency domain from time domain.
470 uses of bandwidth extended coding unit are transformed into the low band signal LB generation bandwidth extend information of frequency domain by a MDCT applying unit 410 and this information are encoded, and described bandwidth extend information represents to be transformed into by the 2nd MDCT applying unit 460 characteristic of the high-frequency band signals HB of frequency domain.The bandwidth extend information can comprise the various information of high-frequency band signals HB, as energy level and envelope.More particularly, bandwidth extended coding unit 470 can use low band signal LB to produce the bandwidth extend information based on the fact that has strong correlation between low band signal LB and the high-frequency band signals HB.
Multiplexing Unit 480 is by to frequency linearity prediction execution unit 420, carry out the multiplexing bit stream that produces based on the coding result of contextual Bit-Plane Encoding unit 450 and bandwidth extended coding unit 470, thereby this bit stream is exported as output signal OUT.
Fig. 5 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal.
With reference to Fig. 5, this equipment comprises frequency band separative element 500, MDCT applying unit 510, frequency linearity prediction execution unit 520, many resolution decomposition unit 530, quantifying unit 540, based on contextual Bit-Plane Encoding unit 550, low-frequency band converting unit 560, high frequency band converting unit 570, bandwidth extended coding unit 580 and Multiplexing Unit 590.
Frequency band separative element 500 is separated into low band signal LB and high-frequency band signals HB with input signal IN.Here, input signal IN can be that wherein analog voice or sound signal are modulated to the PCM signal of digital signal, and low band signal LB can be the frequency signal that is lower than predetermined threshold, and high-frequency band signals HB can be the frequency signal that is higher than described predetermined threshold.
510 pairs of MDCT applying units are carried out MDCT by frequency band separative element 500 isolated low band signal LB, thereby low band signal LB is transformed into frequency domain from time domain.
520 pairs of low band signal LB execution frequency linearity predictions that are transformed into frequency domain by MDCT applying unit 510 of frequency linearity prediction execution unit.Here, the frequency linearity prediction is approximately current frequency signal the linear combination of previous frequency signal.More particularly, frequency linearity prediction execution unit 520 calculate the coefficient of linear prediction filters so that predicated error (signal of linear prediction and current frequency signal poor) minimum, and the low band signal LB that is switched to frequency domain is carried out linear prediction filtering according to the coefficient that calculates.Here, frequency linearity prediction execution unit 520 can improve code efficiency by the analog value of the coefficient of linear prediction filter is carried out vector quantization so that represent analog value with vector index.
More particularly, if the low band signal LB that is transformed into frequency domain by MDCT applying unit 510 is voice or pitch signal, then frequency linearity prediction execution unit 520 can be carried out the frequency linearity prediction to this voice signal or pitch signal.That is, frequency linearity prediction execution unit 520 can improve code efficiency by predicting according to the characteristic execution frequency linearity of the signal that receives.
Many resolution decomposition unit 530 receives from the result of frequency linearity prediction execution unit 520 outputs, and the audible spectrum coefficient of the signal that receives of transient change is carried out many resolution decomposition.More particularly, many resolution decomposition unit 530 can by the intensity that changes according to audible spectrum will by the audible spectrum of frequency linearity prediction execution unit 520 filtering be divided into two types (as, stability types and short type), this audible spectrum is carried out many resolution decomposition.
More particularly, if be transformed into the low band signal LB of frequency domain by MDCT applying unit 510 or be transient signal from the result of frequency linearity prediction execution unit 520 output, then many resolution decomposition can be carried out to this transient signal in many resolution decomposition unit 530.That is, many resolution decomposition unit 530 can improve code efficiency by carrying out many resolution decomposition according to the characteristic of the signal that receives.
Quantifying unit is 540 couples of results from frequency linearity prediction execution unit 520 or 530 outputs of many resolution decomposition unit quantize.
550 couples of results from quantifying unit 540 outputs carry out based on contextual Bit-Plane Encoding based on contextual Bit-Plane Encoding unit.Here, it is described based on contextual Bit-Plane Encoding to use the Huffman coding method to carry out based on contextual Bit-Plane Encoding unit 560.
Therefore frequency linearity prediction execution unit 520, many resolution decomposition unit 530, quantifying unit 540 and encode based on the 550 couples of low band signal LB from 510 outputs of MDCT applying unit in contextual Bit-Plane Encoding unit can be called the low-frequency band coding unit jointly.
The conversion method that low-frequency band converting unit 560 is used except the MDCT method will be transformed into frequency domain or time/frequency domain from time domain by frequency band separative element 500 isolated low band signal LB.For example, low-frequency band converting unit 560 can use MDST method, FFT method or QMF method that low band signal LB is transformed into frequency domain or time/frequency domain from time domain.Here, the amplitude of time-domain representation low band signal LB (as energy or acoustic pressure) over time, the amplitude of frequency domain representation low band signal LB is with the variation of frequency, and the time/frequency domain representation low band signal LB amplitude in time with the variation of frequency.
The conversion method that high frequency band converting unit 570 is used except the MDCT method will be transformed into frequency domain or time/frequency domain from time domain by frequency band separative element 500 isolated high-frequency band signals HB.Here, high frequency band converting unit 570 is used identical conversion method with low-frequency band converting unit 560.For example, high frequency band converting unit 570 can be used MDST method, FFT method or QMF method.
580 uses of bandwidth extended coding unit are transformed into the low band signal LB generation bandwidth extend information of frequency domain by low-frequency band converting unit 560 and this information are encoded, and described bandwidth extend information represents to be transformed into by high frequency band converting unit 570 characteristic of the high-frequency band signals HB of frequency domain.The bandwidth extend information can comprise the multiple information of high-frequency band signals HB, as energy level and envelope.More particularly, bandwidth extended coding unit 580 can use low band signal LB to produce the bandwidth extend information based on the fact that has strong correlation between low band signal LB and the high-frequency band signals HB.
Multiplexing Unit 590 is by to frequency linearity prediction execution unit 520, carry out the multiplexing bit stream that produces based on the coding result of contextual Bit-Plane Encoding unit 550 and bandwidth extended coding unit 580, thereby this bit stream is exported as output signal OUT.
Fig. 6 is according to another embodiment of the present invention to the block diagram of the equipment of coding audio signal.
With reference to Fig. 6, this equipment comprises MDCT applying unit 600, frequency band separative element 610, frequency linearity prediction execution unit 620, many resolution decomposition unit 630, quantifying unit 640, based on contextual Bit-Plane Encoding unit 650, bandwidth extended coding unit 660 and Multiplexing Unit 670.
600 couples of input signal IN of MDCT applying unit carry out MDCT, so that input signal IN is transformed into frequency domain from time domain.Here, input signal can be that wherein analog voice or sound signal are modulated to the PCM signal of digital signal.
Frequency band separative element 610 will be separated into low band signal LB and high-frequency band signals HB by the input signal IN that MDCT applying unit 600 is transformed into frequency domain.Here, low band signal LB can be the frequency signal that is lower than predetermined threshold, and high-frequency band signals HB can be the frequency signal that is higher than described predetermined threshold.
620 pairs of frequency linearity prediction execution unit are carried out frequency linearity by frequency band separative element 610 isolated low band signal LB and are predicted.Here, the frequency linearity prediction is approximately current frequency signal the linear combination of previous frequency signal.More particularly, frequency linearity prediction execution unit 620 calculate the coefficient of linear prediction filters so that predicated error (signal of linear prediction and current frequency signal poor) minimum, and the low band signal LB that is switched to frequency domain is carried out linear prediction filtering according to the coefficient that calculates.Here, frequency linearity prediction execution unit 620 can improve code efficiency by the analog value of the coefficient of linear prediction filter is carried out vector quantization so that represent analog value with vector index.
More particularly, if be voice or pitch signal by frequency band separative element 610 isolated low band signal LB, then frequency linearity prediction execution unit 620 can be carried out the frequency linearity prediction to this voice signal or pitch signal.That is, frequency linearity prediction execution unit 620 can improve code efficiency by predicting according to the characteristic execution frequency linearity of the signal that receives.
Many resolution decomposition unit 630 receives from the result of frequency linearity prediction execution unit 620 outputs, and the audible spectrum coefficient of the signal that receives of transient change is carried out many resolution decomposition.More particularly, many resolution decomposition unit 630 can by the intensity that changes according to audible spectrum will by the audible spectrum of frequency linearity prediction execution unit 620 filtering be divided into two types (as, stability types and short type), this audible spectrum is carried out many resolution decomposition.
More particularly, if be transient signal by frequency band separative element 610 isolated low band signal LB or from the result of frequency linearity prediction execution unit 620 output, then many resolution decomposition can be carried out to this transient signal in many resolution decomposition unit 630.That is, many resolution decomposition unit 630 can improve code efficiency by carrying out many resolution decomposition according to the characteristic of the signal that receives.
Quantifying unit is 640 couples of results from frequency linearity prediction execution unit 620 or 630 outputs of many resolution decomposition unit quantize.
650 couples of results from quantifying unit 640 outputs carry out based on contextual Bit-Plane Encoding based on contextual Bit-Plane Encoding unit.Here, it is described based on contextual Bit-Plane Encoding to use the Huffman coding method to carry out based on contextual Bit-Plane Encoding unit 650.
Therefore frequency linearity prediction execution unit 620, many resolution decomposition unit 630, quantifying unit 640 and encode based on the 650 couples of low band signal LB from 610 outputs of frequency band separative element in contextual Bit-Plane Encoding unit can be called the low-frequency band coding unit jointly.
Bandwidth extended coding unit 660 uses by frequency band separative element 610 isolated low band signal LB generation bandwidth extend informations and to this information and encodes, and described bandwidth extend information is represented the characteristic by frequency band separative element 610 isolated high-frequency band signals HB.The bandwidth extend information can comprise the multiple information of high-frequency band signals HB, as energy level and the envelope of high-frequency band signals HB.More particularly, bandwidth extended coding unit 660 can use low band signal LB to produce the bandwidth extend information based on the fact that has strong correlation between low band signal LB and the high-frequency band signals HB.
Multiplexing Unit 670 is by to frequency linearity prediction execution unit 620, carry out the multiplexing bit stream that produces based on the coding result of contextual Bit-Plane Encoding unit 650 and bandwidth extended coding unit 660, thereby this bit stream is exported as output signal OUT.
Fig. 7 is the block diagram according to the equipment that sound signal is decoded of the embodiment of the invention.
With reference to Fig. 7, this equipment comprises demultiplexing unit 700, based on contextual bit plane decoding unit 710, PQ-SPSC module 720, inverse quantization unit 730, differentiate synthesis unit 740, frequency inverse linear prediction performance element 750, the first contrary MDCT applying unit 760, bandwidth expansion decoding unit 770, second contrary MDCT applying unit 780 and the combinations of bands unit 790 more.
Demultiplexing unit 700 receives from the bit stream of coding side output and to this bit stream and carries out demultiplexing.More particularly, demultiplexing unit 700 is separated into the data block (data piece) corresponding with several data level (data level) with this bit stream, and decomposes the also information of output bit flow at data block.Here, the information from demultiplexing unit 700 output comprises: about decomposed information, quantized value and other reconfiguration information of the audible spectrum that will use by PQ-SPSC module 720, quantize frequency spectrum reconfiguration information, about based on the information of contextual bit plane decoding, about quantizing wide polar coordinates (wide-polar coordination) stereo decoding information, signal type information, about frequency linearity prediction and the information of vector quantization and the bandwidth extend information of coding.
Bit plane based on 710 pairs of codings of contextual bit plane decoding unit is carried out based on contextual decoding.Here, receive from the information of demultiplexing unit 700 outputs, and use Huffman coding method reconstructed spectrum, code frequency band model information and scale factor (scale factor) based on contextual bit plane decoding unit 710.Receive lossy coding (prejudicecoding) frequency band pattern information, the scale factor of lossy coding and the frequency spectrum of lossy coding based on contextual bit plane decoding unit 710, and output encoder frequency band mode value, the decoding decoration indication (cosmetic indication) of scale factor and the quantized value of frequency spectrum.
PQ-SPSC module 720 receives from the result based on contextual bit plane decoding unit 710 outputs, and this result's frequency spectrum is carried out limit polar coordinates stereo decoding.Here, PQ-SPSC module 720 is carried out limit polar coordinates stereo decoding by the coupling information (coupling information) between received spectrum and the limit polar coordinates stereophonic signal, and output quantizes frequency spectrum then.
730 couples of results from 720 outputs of PQ-SPSC module of inverse quantization unit carry out re-quantization.
The synthesis units 740 of differentiating receive from the results of inverse quantization unit 730 outputs more, and to the audible spectrum coefficient of the signal that receives of transient change carry out differentiate synthetic.In more detail, if the sound signal that receives from coding side has been carried out many resolution decomposition, then synthesis units 740 of differentiating can improve decoding efficiency by the result from inverse quantization unit 730 outputs is carried out to differentiate to synthesize more more.Here, differentiate synthesis unit 740 more and receive re-quantization frequency spectrum/difference frequency spectrum, and output reconstructed spectrum/difference frequency spectrum.
Frequency inverse linear prediction performance element 750 will predict the outcome combined from the frequency linearity of inverse quantization unit 730 or the result who differentiates synthesis unit 740 output and the coding side that receives from demultiplexing unit 700 more.In more detail, if the sound signal that receives from coding side has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 750 can by frequency linearity is predicted the outcome with from inverse quantization unit 730 or differentiate the combined decoding efficiency that improves of result of synthesis unit 740 outputs more.Here, frequency inverse linear prediction performance element 750 has improved decoding efficiency effectively by the vector quantization technology that adopts frequency domain forecasting techniques and predictive coefficient.Frequency inverse linear prediction performance element 750 receives difference spectral coefficient and vector index, and output MDCT spectral coefficient.
Decode based on contextual bit plane decoding unit 710, PQ-SPSC module 720, inverse quantization unit 730, the low band signal of differentiating synthesis unit 740 and 750 pairs of codings of frequency inverse linear prediction performance element more, therefore can be called the low-frequency band decoding unit jointly.
The first contrary MDCT applying unit 760 is carried out the inverse operation of the conversion of being carried out by coding side.760 pairs of low band signal from many resolutions synthesis unit 740 and 750 outputs of frequency inverse linear prediction performance element of the first contrary MDCT applying unit are carried out contrary MDCT, so that low band signal is transformed into time domain from frequency domain.Here, the first contrary MDCT applying unit 760 receives the spectral coefficient that obtains from the re-quantization result of many resolutions synthesis unit 740 or frequency inverse linear prediction performance element 750, and output is corresponding to the voice data of the reconstruct of low-frequency band.
The bandwidth extend information of 770 pairs of codings of bandwidth expansion decoding unit is decoded, and the low band signal that the bandwidth extend information of using decoding is exported from many resolutions synthesis unit 740 or frequency inverse linear prediction performance element 750 produces high-frequency band signals.Here, bandwidth expansion decoding unit 770 is applied to low band signal by the bandwidth extend information with decoding and produces high-frequency band signals based on the fact that has strong correlation between low band signal and the high-frequency band signals.Here, the bandwidth extend information is represented the characteristic of high-frequency band signals, comprises the multiple information of high-frequency band signals, as energy level and envelope.
780 pairs of high-frequency band signals by 770 decodings of bandwidth expansion decoding unit of the second contrary MDCT applying unit are carried out contrary MDCT, so that high-frequency band signals is transformed into time domain from frequency domain.
The high-frequency band signals that combinations of bands unit 790 will be transformed into the low band signal of time domain and be transformed into time domain by the second contrary MDCT applying unit 780 by the first contrary MDCT applying unit 760 is combined, so that the result is exported as output signal OUT.
Fig. 8 is the block diagram of the equipment of according to another embodiment of the present invention sound signal being decoded.
With reference to Fig. 8, this equipment comprises demultiplexing unit 800, based on contextual bit plane decoding unit 810, PQ-SPSC module 820, inverse quantization unit 830, differentiate synthesis unit 840, frequency inverse linear prediction performance element 850, contrary MDCT applying unit 860, converting unit 865, bandwidth expansion decoding unit 870, inverse conversion unit 880 and combinations of bands unit 890 more.
Demultiplexing unit 800 receives from the bit stream of coding side output and to this bit stream and carries out demultiplexing.More particularly, demultiplexing unit 800 is separated into the data block corresponding with a plurality of data level with this bit stream, and decomposes the also information of output bit flow at data block.Here, the information from demultiplexing unit 800 output comprises: about decomposed information, quantized value and other reconfiguration information of the audible spectrum that will use by PQ-SPSC module 820, quantize frequency spectrum reconfiguration information, about based on the information of contextual bit plane decoding, about quantizing wide polar coordinates stereo decoding information, signal type information, about frequency linearity prediction and the information of vector quantization and the bandwidth extend information of coding.
Bit plane based on 810 pairs of codings of contextual bit plane decoding unit is carried out based on contextual decoding.Here, receive from the information of demultiplexing unit 800 outputs, and use Huffman coding method reconstructed spectrum, code frequency band model information and scale factor based on contextual bit plane decoding unit 810.Receive lossy coding frequency band pattern information, the scale factor of lossy coding and the frequency spectrum of lossy coding based on contextual bit plane decoding unit 810, and output encoder frequency band mode value, the decoding decoration indication of scale factor and the quantized value of frequency spectrum.
PQ-SPSC module 820 receives from the result based on contextual bit plane decoding unit 810 outputs, and this result's frequency spectrum is carried out limit polar coordinates stereo decoding.Here, PQ-SPSC module 820 is carried out limit polar coordinates stereo decoding by the coupling information between received spectrum and the limit polar coordinates stereophonic signal, and output quantizes frequency spectrum then.
830 couples of results from 820 outputs of PQ-SPSC module of inverse quantization unit carry out re-quantization.
The synthesis units 840 of differentiating receive from the results of inverse quantization unit 830 outputs more, and to the audible spectrum coefficient of the signal that receives of transient change carry out differentiate synthetic.In more detail, if the sound signal that receives from coding side has been carried out many resolution decomposition, then synthesis units 840 of differentiating can improve decoding efficiency by the result from inverse quantization unit 830 outputs is carried out to differentiate to synthesize more more.Here, differentiate synthesis unit 840 more and receive re-quantization frequency spectrum/difference frequency spectrum, and output reconstructed spectrum/difference frequency spectrum.
Frequency inverse linear prediction performance element 850 will predict the outcome combined from the frequency linearity of inverse quantization unit 830 or the result who differentiates synthesis unit 840 output and the coding side that receives from demultiplexing unit 800 more, and the result of combination is carried out inverse vector quantize.In more detail, if the sound signal that receives from coding side has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 850 can by frequency linearity is predicted the outcome with from inverse quantization unit 830 or differentiate the combined decoding efficiency that improves of result of synthesis unit 840 outputs more.Here, frequency inverse linear prediction performance element 850 has improved decoding efficiency effectively by the vector quantization technology that adopts frequency domain forecasting techniques and predictive coefficient.Frequency inverse linear prediction performance element 850 receives difference spectral coefficient and vector index, and output MDCT spectral coefficient.
Decode based on contextual bit plane decoding unit 810, PQ-SPSC module 820, inverse quantization unit 830, the low band signal of differentiating synthesis unit 840 and 850 pairs of codings of frequency inverse linear prediction performance element more, therefore can be called the low-frequency band decoding unit jointly.
Contrary MDCT applying unit 860 is carried out the inverse operation of the conversion of being carried out by coding side.860 pairs of low band signal from many resolutions synthesis unit 840 and 850 outputs of frequency inverse linear prediction performance element of contrary MDCT applying unit are carried out contrary MDCT, so that low band signal is transformed into time domain from frequency domain.Here, contrary MDCT applying unit 860 receives the spectral coefficient that obtains from the re-quantization result of many resolutions synthesis unit 840 or frequency inverse linear prediction performance element 850, and output is corresponding to the voice data of the reconstruct of low-frequency band.
Converting unit 865 utilizes conversion method to be transformed into frequency domain or time/frequency domain from time domain by the low band signal that contrary MDCT applying unit 860 is transformed into time domain.For example, converting unit 865 can be used MDST method, FFT method or QMF method conversion low band signal.Here, also can use the MDCT method.Yet if use the MDCT method, the embodiment among Fig. 7 of front is more effective than current embodiment.
The bandwidth extend information of 870 pairs of codings of bandwidth expansion decoding unit is decoded, and the low band signal that the bandwidth extend information of use decoding is exported from converting unit 865 produces high-frequency band signals.Here, bandwidth expansion decoding unit 870 is applied to low band signal by the bandwidth extend information with decoding and produces high-frequency band signals based on the fact that has strong correlation between low band signal and the high-frequency band signals.Here, the bandwidth extend information is represented the characteristic of high-frequency band signals, comprises the multiple information of high-frequency band signals, as energy level and envelope.
Inverse conversion unit 880 use high-frequency band signals that the conversion method except the MDCT method will be expanded decoding unit 870 decodings by bandwidth from frequency domain or the time/the frequency domain inverse conversion is to time domain.Here, converting unit 865 is used identical conversion method with inverse conversion unit 880.For example, inverse conversion unit 880 can use MDST method, FFT method or QMF method.
Combinations of bands unit 890 will be transformed into the low band signal of time domain by contrary MDCT applying unit 860 and be reversed and change unit 880 to be transformed into the high-frequency band signals of time domain combined, so that the result is exported as output signal OUT.
Fig. 9 is according to another embodiment of the present invention to the block diagram of the equipment of audio signal decoding.
With reference to Fig. 9, described equipment comprises demultiplexing unit 900, based on contextual bit plane decoding unit 910, PQ-SPSC module 920, inverse quantization unit 930, differentiate synthesis unit 940, frequency inverse linear prediction performance element 950, bandwidth expansion decoding unit 960, combinations of bands unit 970 and contrary MDCT applying unit 980 more.
Demultiplexing unit 900 receives from the bit stream of coding side output and to this bit stream and carries out demultiplexing.More particularly, demultiplexing unit 900 is separated into bit stream and a plurality of data level corresponding data block, and decomposes and the information of output bit flow at described data block.Here, from the information of demultiplexing unit 900 output comprise decomposed information, quantized value and other reconfiguration information about the audible spectrum that will use by PQ-SPSC module 920, quantize frequency spectrum reconfiguration information, about based on the information of contextual bit plane decoding, about quantizing wide polar coordinates stereo decoding information, signal type information, about frequency linearity prediction and the information of vector quantization and the bandwidth extend information of coding.
Bit plane based on 910 pairs of codings of contextual bit plane decoding unit is carried out based on contextual decoding.Here, receive from the information of demultiplexing unit 900 outputs based on contextual bit plane decoding unit 910, and by using the Huffman coding method to come reconstructed spectrum, code frequency band model information and scale factor.Receive lossy coding frequency band pattern information, the scale factor of lossy coding and the frequency spectrum of lossy coding based on contextual bit plane decoding unit 910, and output encoder frequency band mode value, the decoding decoration indication of scale factor and the quantized value of frequency spectrum.
PQ-SPSC module 920 receives from the result based on contextual bit plane decoding unit 910 outputs, and result's frequency spectrum is carried out limit polar coordinates stereo decoding.Here, PQ-SPSC module 920 is carried out limit polar coordinates stereo decoding by the coupling information between received spectrum and the limit polar coordinates stereophonic signal, and output quantizes frequency spectrum then.
930 couples of results from 920 outputs of PQ-SPSC module of inverse quantization unit carry out re-quantization.
The synthesis units 940 of differentiating receive from the results of inverse quantization unit 930 outputs more, and to the audible spectrum coefficient of the signal that receives of transient change carry out differentiate synthetic.More particularly, if the sound signal that receives from coding side has been carried out many resolution decomposition, then synthesis units 940 of differentiating can improve decoding efficiency by the result from inverse quantization unit 930 outputs is carried out to differentiate to synthesize more more.Here, differentiate synthesis unit 940 receives re-quantization frequency spectrum/difference frequency spectrum and exports reconstructed spectrum/difference frequency spectrum more.
Frequency inverse linear prediction performance element 950 combination is passed through the frequency linearity prediction result that coding side is carried out from the result of many resolutions synthesis unit 940 outputs with from what demultiplexing unit 900 received, and the result of combination is carried out inverse vector quantizes.More particularly, if the sound signal that receives from coding side has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 950 can improve decoding efficiency by combination frequency linear prediction result with from inverse quantization unit 930 or the result that differentiates synthesis unit 940 output more.Here, frequency inverse linear prediction performance element 950 has improved decoding efficiency effectively by the vector quantization technology that utilizes frequency domain forecasting techniques and predictive coefficient.Frequency inverse linear prediction performance element 950 receives difference spectral coefficient and vector index, and output MDCT spectral coefficient.
Decode based on contextual bit plane decoding unit 910, PQ-SPSC module 920, inverse quantization unit 930, the low band signal of differentiating synthesis unit 940 and 950 pairs of codings of frequency inverse linear prediction performance element more, thereby can be called the low-frequency band decoding unit jointly.
The bandwidth extend information of 960 pairs of codings of bandwidth expansion decoding unit is decoded, and comes to produce high-frequency band signals from the low band signal that many resolutions synthesis unit 940 or frequency inverse linear prediction performance element 950 are exported by the bandwidth extend information of using decoding.Here, based on the fact that has strong correlation between low band signal and the high-frequency band signals, bandwidth expansion decoding unit 960 is applied to low band signal by the bandwidth extend information with decoding and produces high-frequency band signals.Here, the bandwidth extend information is represented the characteristic of high-frequency band signals, and it comprises the multiple information (such as energy level and envelope) of high-frequency band signals.
970 combinations of combinations of bands unit are from the low band signal of many resolutions synthesis unit 940 or 950 outputs of frequency inverse linear prediction performance element and the high-frequency band signals of decoding by bandwidth expansion decoding unit 960.
Contrary MDCT applying unit 980 comes the result of 970 outputs from the combinations of bands unit is carried out inverse conversion by carrying out contrary MDCT, so that this result is exported as output signal OUT.Here, contrary MDCT applying unit 980 receives the spectral coefficient that obtains from the re-quantization result of frequency inverse linear prediction performance element 950, and the voice data of output and the corresponding reconstruct of low-frequency band.
Figure 10 is according to another embodiment of the present invention to the block diagram of the equipment of audio signal decoding.
With reference to Figure 10, described equipment comprises demultiplexing unit 1000, based on contextual bit plane decoding unit 1010, inverse quantization unit 1020, differentiate synthesis unit 1030, frequency inverse linear prediction performance element 1040, bandwidth expansion decoding unit 1050, the first contrary MDCT applying unit 1060, second contrary MDCT applying unit 1070 and the combinations of bands unit 1080 more.
Demultiplexing unit 1000 receives from the bit stream of coding side output and to it and carries out demultiplexing.More particularly, demultiplexing unit 1000 is separated into bit stream and a plurality of data level corresponding data block, and decomposes and the information of output bit flow at described data block.Here, from the information of demultiplexing unit 1000 output comprise decomposed information, quantized value and other reconfiguration information, quantize the reconfiguration information of frequency spectrum about audible spectrum, about based on the information of contextual bit plane decoding, signal type information, about frequency linearity prediction and the information of vector quantization and the bandwidth extend information of coding.
Bit plane based on 1010 pairs of codings of contextual bit plane decoding unit is carried out based on contextual decoding.Here, receive from the information of demultiplexing unit 1000 outputs based on contextual bit plane decoding unit 1010, and by using the Huffman coding method to come reconstructed spectrum, code frequency band model information and scale factor.Receive lossy coding frequency band pattern information, the scale factor of lossy coding and the frequency spectrum of lossy coding based on contextual bit plane decoding unit 1010, and output encoder frequency band mode value, the decoding decoration indication of scale factor and the quantized value of frequency spectrum.
1020 pairs of inverse quantization unit are carried out re-quantization from the result based on contextual bit plane decoding unit 1010 outputs.
The synthesis units 1030 of differentiating receive from the results of inverse quantization unit 1020 outputs more, and to the audible spectrum coefficient of the signal that receives of transient change carry out differentiate synthetic.More particularly, if the sound signal that receives from coding side has been carried out many resolution decomposition, then synthesis units 1030 of differentiating can improve decoding efficiency by the result from inverse quantization unit 1020 outputs is carried out to differentiate to synthesize more more.Here, differentiate synthesis unit 1030 receives re-quantization frequency spectrum/difference frequency spectrum and exports reconstructed spectrum/difference frequency spectrum more.
1040 combinations of frequency inverse linear prediction performance element are from the result of many resolutions synthesis unit 1030 outputs and the frequency linearity prediction result of passing through the coding side execution that receives from demultiplexing unit 1000.More particularly, if the sound signal that receives from coding side has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 1040 can improve decoding efficiency by combination frequency linear prediction result with from inverse quantization unit 1020 or the result that differentiates synthesis unit 1030 output more.Here, frequency inverse linear prediction performance element 1040 has improved decoding efficiency effectively by the vector quantization technology that utilizes frequency domain forecasting techniques and predictive coefficient.Frequency inverse linear prediction performance element 1040 receives difference spectral coefficient and vector index, and output MDCT spectral coefficient.
Decode based on contextual bit plane decoding unit 1010, inverse quantization unit 1020, the low band signal of differentiating synthesis unit 1030 and 1040 pairs of codings of frequency inverse linear prediction performance element more, thereby can be known as the low-frequency band decoding unit jointly.
The bandwidth extend information of 1050 pairs of codings of bandwidth expansion decoding unit is decoded and is come to produce high-frequency band signals from the low band signal that many resolutions synthesis unit 1030 or frequency inverse linear prediction performance element 1040 are exported by the bandwidth extend information of using decoding.Here, based on the fact that has strong correlation between low band signal and the high-frequency band signals, bandwidth expansion decoding unit 1050 is applied to low band signal by the bandwidth extend information with decoding and produces high-frequency band signals.Here, the bandwidth extend information is represented the characteristic of high-frequency band signals, and it comprises the multiple information (such as energy level and envelope) of high-frequency band signals.
The first contrary MDCT applying unit 1060 is carried out the inverse operation of the conversion of being carried out by coding side.1060 pairs of low band signal from many resolutions synthesis unit 1030 and 1040 outputs of frequency inverse linear prediction performance element of the first contrary MDCT applying unit are carried out contrary MDCT, so that low band signal is transformed into time domain from frequency domain.Here, the first contrary MDCT applying unit 1060 receives the spectral coefficient that obtains from the result of the re-quantization of many resolutions synthesis unit 1030 or frequency inverse linear prediction performance element 1040, and the voice data of output and the corresponding reconstruct of low-frequency band.
1070 pairs of high-frequency band signals by 1050 decodings of bandwidth expansion decoding unit of the second contrary MDCT applying unit are carried out contrary MDCT, so that high-frequency band signals is transformed into time domain from frequency domain.
1080 combinations of combinations of bands unit are switched to the low band signal of time domain and are switched to the high-frequency band signals of time domain by the second contrary MDCT applying unit 1070 by the first contrary MDCT applying unit 1060, so that the result is exported as output signal OUT.
Figure 11 is according to another embodiment of the present invention to the block diagram of the equipment of audio signal decoding.
With reference to Figure 11, described equipment comprises demultiplexing unit 1100, based on contextual bit plane decoding unit 1110, inverse quantization unit 1120, differentiate synthesis unit 1130, frequency inverse linear prediction performance element 1140, contrary MDCT applying unit 1150, converting unit 1160, bandwidth expansion decoding unit 1170, inverse conversion unit 1180 and combinations of bands unit 1190 more.
Demultiplexing unit 1100 receives from the bit stream of coding side output and to this bit stream and carries out demultiplexing.More particularly, demultiplexing unit 1100 is separated into bit stream and a plurality of data level corresponding data block, and decomposes and the information of output bit flow at described data block.Here, from the information of demultiplexing unit 1100 output comprise about with decomposed information, quantized value and other reconfiguration information of audible spectrum, quantize the reconfiguration information of frequency spectrum, about based on the information of contextual bit plane decoding, signal type information, about frequency linearity prediction and the information of vector quantization and the bandwidth extend information of coding.
Bit plane based on 1110 pairs of codings of contextual bit plane decoding unit is carried out based on contextual decoding.Here, receive from the information of demultiplexing unit 1100 outputs based on contextual bit plane decoding unit 1110, and by using the Huffman coding method to come reconstructed spectrum, code frequency band model information and scale factor.Receive lossy coding frequency band pattern information, the scale factor of lossy coding and the frequency spectrum of lossy coding based on contextual bit plane decoding unit 1110, and output encoder frequency band mode value, the decoding decoration indication of scale factor and the quantized value of frequency spectrum.
1120 pairs of inverse quantization unit are carried out re-quantization from the result based on contextual bit plane decoding unit 1110 outputs.
The synthesis units 1130 of differentiating receive from the results of inverse quantization unit 1120 outputs more, and to the audible spectrum coefficient of the signal that receives of transient change carry out differentiate synthetic.More particularly, if the sound signal that receives from coding side has been carried out many resolution decomposition, then synthesis units 1130 of differentiating can improve decoding efficiency by the result from inverse quantization unit 1120 outputs is carried out to differentiate to synthesize more more.Here, differentiate synthesis unit 1130 receives re-quantization frequency spectrum/difference frequency spectrum and exports reconstructed spectrum/difference frequency spectrum more.
Frequency inverse linear prediction performance element 1140 combination is passed through the frequency linearity prediction result that coding side is carried out from the result of many resolutions synthesis unit 1130 outputs with from what demultiplexing unit 1100 received, and the result of combination is carried out inverse vector quantizes.More particularly, if the sound signal that receives from coding side has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 1140 can improve decoding efficiency by combination frequency linear prediction result with from inverse quantization unit 1120 or the result that differentiates synthesis unit 1130 output more.Here, frequency inverse linear prediction performance element 1140 has improved decoding efficiency effectively by the vector quantization technology that utilizes frequency domain forecasting techniques and predictive coefficient.Frequency inverse linear prediction performance element 1140 receives difference spectral coefficient and vector index, and output MDCT spectral coefficient.
Decode based on contextual bit plane decoding unit 1110, inverse quantization unit 1120, the low band signal of differentiating synthesis unit 1130 and 1140 pairs of codings of frequency inverse linear prediction performance element more, thereby can be called the low-frequency band decoding unit jointly.
Contrary MDCT applying unit 1150 is carried out the inverse operation of the conversion of being carried out by coding side.1150 pairs of low band signal from many resolutions synthesis unit 1130 and 1140 outputs of frequency inverse linear prediction performance element of contrary MDCT applying unit are carried out contrary MDCT, so that low band signal is transformed into time domain from frequency domain.Here, contrary MDCT applying unit 1150 receives the spectral coefficient that obtains from the result of the re-quantization of many resolutions synthesis unit 1130 or frequency inverse linear prediction performance element 1140, and the voice data of output and the corresponding reconstruct of low-frequency band.
Converting unit 1160 will be transformed into frequency domain or time/frequency domain from time domain by the low band signal that contrary MDCT applying unit 1150 is transformed into time domain by using conversion method.For example, converting unit 1160 can be changed low band signal by using MDST method, FFT method or QMF method.Here, also can use the MDCT method.Yet if use the MDCT method, the embodiment of Figure 10 of front is more effective than current embodiment.
The bandwidth extend information of 1170 pairs of codings of bandwidth expansion decoding unit is decoded and the low band signal exported from converting unit 1160 by the bandwidth extend information of using decoding produces high-frequency band signals.Here, based on the fact that has strong correlation between low band signal and the high-frequency band signals, bandwidth expansion decoding unit 1170 is applied to low band signal by the bandwidth extend information with decoding and produces high-frequency band signals.Here, the bandwidth extend information is represented the characteristic of high-frequency band signals, and it comprises the multiple information (such as energy level and envelope) of high-frequency band signals.
Inverse conversion unit 1180 by the high-frequency band signals that uses conversion method except the MDCT method and bandwidth is expanded decoding unit 1170 decodings from frequency domain or the time/the frequency domain inverse conversion is to time domain.Here converting unit 1160 is used identical conversion method with inverse conversion unit 1180.For example, inverse conversion unit 1180 can use MDST method, FFT method or QMF method.
1190 combinations of combinations of bands unit are transformed into the low band signal of time domain and are transformed into the high-frequency band signals of time domain by inverse conversion unit 1180 by contrary MDCT applying unit 1150, so that the result is exported as output signal OUT.
Figure 12 is according to another embodiment of the present invention to the block diagram of the equipment of audio signal decoding.
With reference to Figure 12, described equipment comprises demultiplexing unit 1200, based on contextual bit plane decoding unit 1210, inverse quantization unit 1220, differentiate synthesis unit 1230, frequency inverse linear prediction performance element 1240, bandwidth expansion decoding unit 1250, combinations of bands unit 1260 and contrary MDCT applying unit 1270 more.
Demultiplexing unit 1200 receives from the bit stream of coding side output and to this bit stream and carries out demultiplexing.More particularly, demultiplexing unit 1200 is separated into bit stream and a plurality of data level corresponding data block, and decomposes and the information of output bit flow at described data block.Here, from the information of demultiplexing unit 1200 output comprise about with decomposed information, quantized value and other reconfiguration information of audible spectrum, quantize the reconfiguration information of frequency spectrum, about based on the information of contextual bit plane decoding, signal type information, about frequency linearity prediction and the information of vector quantization and the bandwidth extend information of coding.
Bit plane based on 1210 pairs of codings of contextual bit plane decoding unit is carried out based on contextual decoding.Here, receive from the information of demultiplexing unit 1200 outputs based on contextual bit plane decoding unit 1210, and by using the Huffman coding method to come reconstructed spectrum, code frequency band model information and scale factor.Receive lossy coding frequency band pattern information, the scale factor of lossy coding and the frequency spectrum of lossy coding based on contextual bit plane decoding unit 1210, and output encoder frequency band mode value, the decoding decoration indication of scale factor and the quantized value of frequency spectrum.
1220 pairs of inverse quantization unit are carried out re-quantization from the result based on contextual bit plane decoding unit 1210 outputs.
The synthesis units 1230 of differentiating receive from the results of inverse quantization unit 1220 outputs more, and to the audible spectrum coefficient of the signal that receives of transient change carry out differentiate synthetic.More particularly, if the sound signal that receives from coding side has been carried out many resolution decomposition, then synthesis units 1230 of differentiating can improve decoding efficiency by the result from inverse quantization unit 1220 outputs is carried out to differentiate to synthesize more more.Here, differentiate synthesis unit 1230 receives re-quantization frequency spectrum/difference frequency spectrum and exports reconstructed spectrum/difference frequency spectrum more.
Frequency inverse linear prediction performance element 1240 combination is passed through the frequency linearity prediction result that coding side is carried out from the result of many resolutions synthesis unit 1230 outputs with from what demultiplexing unit 1200 received, and the result of combination is carried out inverse vector quantizes.More particularly, if the sound signal that receives from coding side has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 1240 can improve decoding efficiency by combination frequency linear prediction result with from inverse quantization unit 1220 or the result that differentiates synthesis unit 1230 output more.Here, frequency inverse linear prediction performance element 1240 has improved decoding efficiency effectively by the vector quantization technology that utilizes frequency domain forecasting techniques and predictive coefficient.Frequency inverse linear prediction performance element 1240 receives difference spectral coefficient and vector index, and output MDCT spectral coefficient.
Decode based on contextual bit plane decoding unit 1210, inverse quantization unit 1220, the low band signal of differentiating synthesis unit 1230 and 1240 pairs of codings of frequency inverse linear prediction performance element more, thereby can be called the low-frequency band decoding unit jointly.
The bandwidth extend information of 1250 pairs of codings of bandwidth expansion decoding unit is decoded, and comes to produce high-frequency band signals from the low band signal that many resolutions synthesis unit 1230 or frequency inverse linear prediction performance element 1240 are exported by the bandwidth extend information of using decoding.Here, based on the fact that has strong correlation between low band signal and the high-frequency band signals, bandwidth expansion decoding unit 1250 is applied to low band signal by the bandwidth extend information with decoding and produces high-frequency band signals.Here, the bandwidth extend information is represented the characteristic of high-frequency band signals, and it comprises the multiple information (such as energy level and envelope) of high-frequency band signals.
1260 combinations of combinations of bands unit are by the low band signal of many resolutions synthesis unit 1230 or 1240 outputs of frequency inverse linear prediction performance element and the high-frequency band signals of being decoded by bandwidth expansion decoding unit 1250.
Contrary MDCT applying unit 1270 comes the result of 1260 outputs from the combinations of bands unit is carried out inverse conversion by carrying out contrary MDCT, thereby the result is exported as output signal OUT.Here, contrary MDCT applying unit 1270 receives the spectral coefficient that obtains from the re-quantization result of frequency inverse linear prediction performance element 1240, and the voice data of output and the corresponding reconstruct of low-frequency band.
Figure 13 is the process flow diagram to the method for audio-frequency signal coding according to the embodiment of the invention.
According to the method for current embodiment sequential processes corresponding to equipment shown in Figure 4.Therefore, will describe this method, and will omit the description that repeats in conjunction with Fig. 4.
With reference to Figure 13, in operation 1300, frequency band separative element 400 is separated into low band signal and high-frequency band signals with input signal.
Respectively low band signal and high-frequency band signals are transformed into frequency domain from time domain at operation 1310, the one MDCT applying units 410 and the 2nd MDCT applying unit 460.
In operation 1320, the low-frequency band coding unit is to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding.Here, the low-frequency band coding unit can comprise frequency linearity prediction execution unit 420, many resolution decomposition unit 430, quantifying unit 440 and based on contextual Bit-Plane Encoding unit 450.More particularly, frequency linearity prediction execution unit 420 comes the low band signal of conversion is carried out filtering by carrying out the frequency linearity prediction according to the characteristic of low band signal.Many resolution decomposition are carried out according to the characteristic of low band signal low band signal that change or filtering in 430 pairs of many resolution decomposition unit.440 pairs of quantifying unit have quantized its low band signal of having carried out many resolution decomposition, carry out based on contextual Bit-Plane Encoding based on the low band signal of the 450 pairs of quantifications in contextual Bit-Plane Encoding unit.
In operation 1330, bandwidth extended coding unit 470 use the low band signal of conversion produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and it is encoded.
In operation 1340, Multiplexing Unit 480 carries out the bit plane of coding and the bandwidth extend information of coding multiplexing, and with the coding result output as input signal of the bandwidth extend information of the bit plane of coding and coding.
Figure 14 is according to another embodiment of the present invention to the process flow diagram of the method for audio-frequency signal coding.
According to the method for current embodiment sequential processes corresponding to equipment shown in Figure 5.Therefore, will describe this method, and will omit the description that repeats in conjunction with Fig. 5.
With reference to Figure 14, in operation 1400, frequency band separative element 500 is separated into low band signal and high-frequency band signals with input signal.
In operation 1410,510 pairs of low band signal of MDCT applying unit are carried out MDCT so that low band signal is transformed into frequency domain from time domain.
In operation 1420, the low-frequency band coding unit is to carry out quantizing its low band signal of having carried out MDCT and based on contextual Bit-Plane Encoding.Here, the low-frequency band coding unit can comprise frequency linearity prediction execution unit 520, many resolution decomposition unit 530, quantifying unit 540 and based on contextual Bit-Plane Encoding unit 550.More particularly, frequency linearity prediction execution unit 520 comes the low band signal of conversion is carried out filtering by carrying out the frequency linearity prediction according to the characteristic of low band signal.Many resolution decomposition are carried out according to the characteristic of low band signal low band signal that change or filtering in 530 pairs of many resolution decomposition unit.540 pairs of quantifying unit have quantized its low band signal of having carried out many resolution decomposition, carry out based on contextual Bit-Plane Encoding based on the low band signal of the 550 pairs of quantifications in contextual Bit-Plane Encoding unit.
In operation 1430, low-frequency band converting unit 560 and high frequency band converting unit 570 are transformed into frequency domain or time/frequency domain with low band signal and high-frequency band signals from time domain respectively.
In operation 1440, encode by the bandwidth extend information of the characteristic of the high-frequency band signals of the low band signal generation expression conversion of use conversion and to it in bandwidth extended coding unit 580.
In operation 1450, Multiplexing Unit 590 is with the bit plane of coding and the bandwidth extend information of coding is multiplexing and with the coding result output as input signal of the bandwidth extend information of the bit plane of coding and coding.
Figure 15 is according to another embodiment of the present invention to the process flow diagram of the method for audio-frequency signal coding.
According to the method for current embodiment sequential processes corresponding to the equipment shown in Fig. 6.Therefore, will describe this method, and omit the description that repeats in conjunction with Fig. 6.
With reference to Figure 15, in operation 1500, MDCT applying unit 600 is transformed into frequency domain with input signal from time domain.
In operation 1510, frequency band separative element 610 is separated into low band signal and high-frequency band signals with the input signal of conversion.
In operation 1520, the low-frequency band coding unit is to low band signal execution quantification with based on contextual Bit-Plane Encoding.Here, the low-frequency band coding unit can comprise frequency linearity prediction execution unit 620, many resolution decomposition unit 630, quantifying unit 640 and based on contextual Bit-Plane Encoding unit 650.More particularly, frequency linearity prediction execution unit 620 comes the low band signal of conversion is carried out filtering by carrying out the frequency linearity prediction according to the characteristic of low band signal.Many resolution decomposition are carried out according to the characteristic of low band signal low band signal that change or filtering in 630 pairs of many resolution decomposition unit.640 pairs of quantifying unit have quantized its low band signal of having carried out many resolution decomposition, carry out based on contextual Bit-Plane Encoding based on the low band signal of the 650 pairs of quantifications in contextual Bit-Plane Encoding unit.
In operation 1530, bandwidth extended coding unit 660 is by using low band signal generation bandwidth extend information and it being encoded, and described bandwidth extend information is represented the characteristic of high-frequency band signals.
In operation 1540, Multiplexing Unit 670 is multiplexing with the bit plane and the bandwidth extend information of coding of coding, and with the coding result output as input signal of the bandwidth extend information of the bit plane of coding and coding.
Figure 16 is the process flow diagram to the method for audio signal decoding according to the embodiment of the invention.
According to the method for current embodiment sequential processes corresponding to the equipment shown in Figure 10.Therefore, can describe this method, and will omit the description that repeats in conjunction with Figure 10.
In operation 1600, the sound signal of demultiplexing unit 1000 received codes.Here, the sound signal of coding comprises the bandwidth extend information of the low-frequency band of the bit plane of coding and coding.
In operation 1610, the low-frequency band decoding unit is carried out based on contextual decoding and re-quantization the bit plane of coding, so that produce low band signal.Here, the low-frequency band decoding unit can comprise based on contextual bit plane decoding unit 1010, inverse quantization unit 1020, differentiate synthesis unit 1030 and frequency inverse linear prediction performance element 1040 more.More particularly, carry out based on contextual decoding based on the bit plane of 1010 pairs of codings of contextual bit plane decoding unit.The bit plane of 1020 pairs of decodings of inverse quantization unit is carried out re-quantization.If the sound signal of the coding that received by demultiplexing unit 1000 has been carried out many resolution decomposition, then bit planes of differentiating 1030 pairs of re-quantizations of synthesis units more carry out differentiate synthetic.If the sound signal of the coding that demultiplexing unit 1000 is received has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 1040 comes the bit plane of combination frequency linear prediction result and re-quantization or it has been carried out the synthetic bit plane of many resolutions by using vector index, so that produce low band signal.
In operation 1620, the bandwidth extend information of 1050 pairs of codings of bandwidth expansion decoding unit is decoded, and produces high-frequency band signals by the bandwidth extend information of using decoding from low band signal.
Respectively low band signal and high-frequency band signals are carried out contrary MDCT at operation 1630, the first contrary MDCT applying units 1060 and the second contrary MDCT applying unit 1070, so that low band signal and high-frequency band signals are transformed into time domain from frequency domain.
In operation 1640, the low band signal of combinations of bands unit 1080 combination conversions and the high-frequency band signals of conversion.
Figure 17 is according to another embodiment of the present invention to the process flow diagram of the method for audio signal decoding.
According to the method for current embodiment sequential processes corresponding to the equipment shown in Figure 11.Therefore, can describe this method, and will omit the description that repeats in conjunction with Figure 11.
In operation 1700, the sound signal of demultiplexing unit 1100 received codes.Here, the sound signal of coding comprises the bandwidth extend information of the low-frequency band of the bit plane of coding and coding.
In operation 1710, the low-frequency band decoding unit is carried out based on contextual decoding and re-quantization the bit plane of coding, so that produce low band signal.Here, the low-frequency band decoding unit can comprise based on contextual bit plane decoding unit 1110, inverse quantization unit 1120, differentiate synthesis unit 1130 and frequency inverse linear prediction performance element 1140 more.More particularly, carry out based on contextual decoding based on the bit plane of 1110 pairs of codings of contextual bit plane decoding unit.The bit plane of 1120 pairs of decodings of inverse quantization unit is carried out re-quantization.If the sound signal of the coding that received by demultiplexing unit 1100 has been carried out many resolution decomposition, then bit planes of differentiating 1130 pairs of re-quantizations of synthesis units more carry out differentiate synthetic.If the sound signal of the coding that demultiplexing unit 1100 is received has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 1140 comes the bit plane of combination frequency linear prediction result and re-quantization or it has been carried out the synthetic bit plane of many resolutions by using vector index, so that produce low band signal.
In operation 1720,1150 pairs of low band signal of contrary MDCT applying unit are carried out contrary MDCT, so that low band signal is transformed into time domain from frequency domain.
In operation 1730, converting unit 1160 will be transformed into frequency domain or time/frequency domain to its low band signal of having carried out contrary MDCT from time domain.
In operation 1740, the bandwidth extend information of 1170 pairs of codings of bandwidth expansion decoding unit is decoded and is produced high-frequency band signals by the bandwidth extend information of using decoding from the low band signal that is switched to frequency domain or time/frequency domain.
The operation 1750, inverse conversion unit 1180 with the high-frequency band signals inverse conversion to time domain.
In operation 1760, the low band signal of combinations of bands unit 1190 combination conversions and the high-frequency band signals of inverse conversion.
Figure 18 is according to another embodiment of the present invention to the process flow diagram of the method for audio signal decoding.
According to the method for current embodiment sequential processes corresponding to the equipment shown in Figure 12.Therefore, can describe this method, and will omit the description that repeats in conjunction with Figure 12.
In operation 1800, the sound signal of demultiplexing unit 1200 received codes.Here, the sound signal of coding comprises the bandwidth extend information of the low-frequency band of the bit plane of coding and coding.
In operation 1810, the low-frequency band decoding unit is carried out based on contextual decoding and re-quantization the bit plane of coding, so that produce low band signal.Here, the low-frequency band decoding unit can comprise based on contextual bit plane decoding unit 1210, inverse quantization unit 1220, differentiate synthesis unit 1230 and frequency inverse linear prediction performance element 1240 more.More particularly, carry out based on contextual decoding based on the bit plane of 1210 pairs of codings of contextual bit plane decoding unit.The bit plane of 1220 pairs of decodings of inverse quantization unit is carried out re-quantization.If the sound signal of the coding that received by demultiplexing unit 1200 has been carried out many resolution decomposition, then bit planes of differentiating 1230 pairs of re-quantizations of synthesis units more carry out differentiate synthetic.If the sound signal of the coding that demultiplexing unit 1200 is received has been carried out the frequency linearity prediction, then frequency inverse linear prediction performance element 1240 is by using vector index and come the bit plane of combination frequency linear prediction result and re-quantization or it has been carried out the bit plane of many resolution decomposition, so that produce low band signal.
In operation 1820, the bandwidth extend information of 1250 pairs of codings of bandwidth expansion decoding unit is decoded and is produced high-frequency band signals by the bandwidth extend information of using decoding from low band signal.
In operation 1830, combinations of bands unit 1260 combination low band signal and high-frequency band signals.
In operation 1840, the signal of contrary 1270 pairs of combinations of MDCT applying unit is carried out contrary MDCT, so that the signal of combination is transformed into time domain from frequency domain.
The present invention also can be implemented as the computer-readable code on the computer readable recording medium storing program for performing.
Computer readable recording medium storing program for performing is any data storage device that can store thereafter the data that can be read by computer system.The example of computer readable recording medium storing program for performing comprises ROM (read-only memory) (ROM), random-access memory (ram), CD-ROM, tape, floppy disk, optical data storage device and carrier wave (such as the data transmission by the internet).Computer readable recording medium storing program for performing also can be distributed on the computer system of networking, thereby stores and computer readable code executed with distributed way.
As mentioned above, according to the present invention, by input signal being separated into low band signal and high-frequency band signals, each of low band signal and high-frequency band signals is transformed into frequency domain from time domain, to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and it is encoded, the bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal, can carry out efficient coding to high fdrequency component with limited bit rate, thereby improve the quality of sound signal.
In addition, sound signal by received code, the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization so that produce low band signal, bandwidth extend information in the sound signal that is included in coding is decoded, produce high-frequency band signals by the bandwidth extend information of using decoding from low band signal, low band signal and high-frequency band signals are carried out contrary MDCT so that low band signal and high-frequency band signals are transformed into time domain from frequency domain, and the low band signal of combination conversion and the high-frequency band signals of conversion, can be from effectively high fdrequency component being decoded with the bit stream of limited bit rate coding.
Although show and described the present invention with reference to exemplary embodiment of the present, but will be understood by those skilled in the art that, under the situation that does not break away from the spirit and scope of the present invention that are defined by the claims, can carry out various changes in form and details.Exemplary embodiment should be understood that just for describing significance not as the restriction purpose.Therefore, scope of the present invention be can't help detailed description of the present invention and is limited, and is limited by claim, and all differences in described scope should be understood to include in the present invention.

Claims (43)

1. method to audio-frequency signal coding said method comprising the steps of:
(a) input signal is separated into low band signal and high-frequency band signals;
(b) in low band signal and the high-frequency band signals each is transformed into frequency domain from time domain;
(c) to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding;
(d) by using the low band signal of conversion, produce the bandwidth extend information of the characteristic of the high-frequency band signals of representing conversion, and to described bandwidth extend information coding;
(e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
2. the method for claim 1, wherein step (b) comprising: by to each the discrete cosine transform that carry out to revise in low band signal and the high-frequency band signals, in low band signal and the high-frequency band signals each is transformed into frequency domain from time domain.
3. the method for claim 1 also comprises following at least one step:
(f) by the low band signal of conversion is carried out the frequency linearity prediction, the low band signal of conversion is carried out filtering;
(g) low band signal of conversion is carried out many resolution decomposition,
Wherein, step (c) comprising: the low band signal of filtering or the low band signal of having carried out many resolution decomposition are carried out quantized and based on contextual Bit-Plane Encoding.
4. method as claimed in claim 3, wherein, step (f) comprising: by being carried out frequency linearity, the low band signal of conversion predicts the coefficient that calculates linear prediction filter, and by using vector index to represent the analog value of coefficient,
Wherein, step (e) comprising: the bit plane of coding, the bandwidth extend information and the vector index of coding are exported as the coding result of input signal.
5. one kind records the computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio-frequency signal coding on it, said method comprising the steps of:
(a) input signal is separated into low band signal and high-frequency band signals;
(b) in low band signal and the high-frequency band signals each is transformed into frequency domain from time domain;
(c) to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding;
(d) by using the low band signal of conversion, produce the bandwidth extend information of the characteristic of the high-frequency band signals of representing conversion, and to described bandwidth extend information coding;
(e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
6. method to audio-frequency signal coding said method comprising the steps of:
(a) input signal is separated into low band signal and high-frequency band signals;
(b) by low band signal is carried out the discrete cosine transform of revising, low band signal is transformed into frequency domain from time domain;
(c) low band signal of having carried out the discrete cosine transform of revising is carried out quantized and based on contextual Bit-Plane Encoding;
(d) in low band signal and the high-frequency band signals each is transformed into frequency domain or time/frequency domain from time domain;
(e) by using the low band signal of conversion, produce the bandwidth extend information of the characteristic of the high-frequency band signals of representing conversion, and to described bandwidth extend information coding;
(f) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
7. method as claimed in claim 6 also comprises following at least one step:
(g) by low band signal being carried out the frequency linearity prediction, the low band signal of having carried out the discrete cosine transform of revising is carried out filtering;
(h) low band signal of having carried out the discrete cosine transform of revising is carried out many resolution decomposition,
Wherein, step (c) comprising: the low band signal of filtering or the low band signal of having carried out many resolution decomposition are carried out quantized and based on contextual Bit-Plane Encoding.
8. method as claimed in claim 7, wherein, step (g) comprises by the low band signal of conversion is carried out frequency linearity predicts the coefficient that calculates linear prediction filter, and by using vector index to represent the analog value of coefficient,
Wherein, step (f) comprising: the bit plane of coding, the bandwidth extend information and the vector index of coding are exported as the coding result of input signal.
9. one kind records the computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio-frequency signal coding on it, said method comprising the steps of:
(a) input signal is separated into low band signal and high-frequency band signals;
(b) by low band signal is carried out the discrete cosine transform of revising, low band signal is transformed into frequency domain from time domain;
(c) low band signal of having carried out the discrete cosine transform of revising is carried out quantized and based on contextual Bit-Plane Encoding;
(d) in low band signal and the high-frequency band signals each is transformed into frequency domain or time/frequency domain from time domain;
(e) by using the low band signal of conversion, produce the bandwidth extend information of the characteristic of the high-frequency band signals of representing conversion, and to described bandwidth extend information coding;
(f) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
10. method to audio-frequency signal coding said method comprising the steps of:
(a) input signal is transformed into frequency domain from time domain;
(b) input signal with conversion is separated into low band signal and high-frequency band signals;
(c) to low band signal execution quantification with based on contextual Bit-Plane Encoding;
(d) pass through to use low band signal to produce the bandwidth extend information of the characteristic of expression high-frequency band signals, and to described bandwidth extend information coding;
(e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
11. method as claimed in claim 10, wherein, step (b) comprising: by input signal is carried out the discrete cosine transform of revising, input signal is transformed into frequency domain from time domain.
12. method as claimed in claim 10 also comprises following at least one step:
(f) by low band signal being carried out the frequency linearity prediction, low band signal is carried out filtering;
(g) low band signal is carried out many resolution decomposition,
Wherein, step (c) comprising: the low band signal of filtering or the low band signal of having carried out many resolution decomposition are carried out quantized and based on contextual Bit-Plane Encoding.
13. method as claimed in claim 12, wherein, step (f) comprising: by being carried out frequency linearity, low band signal predicts the coefficient that calculates linear prediction filter, and by using vector index to represent the analog value of coefficient,
Wherein, step (e) comprising: the bit plane of coding, the bandwidth extend information and the vector index of coding are exported as the coding result of input signal.
14. one kind records the computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio-frequency signal coding on it, said method comprising the steps of:
(a) input signal is transformed into frequency domain from time domain;
(b) input signal with conversion is separated into low band signal and high-frequency band signals;
(c) to low band signal execution quantification with based on contextual Bit-Plane Encoding;
(d) by using low band signal to produce the bandwidth extend information of the characteristic of representing high-frequency band signals and described bandwidth extend information being encoded;
(e) bit plane of coding and the bandwidth extend information of coding are exported as the coding result of input signal.
15. the method to audio signal decoding said method comprising the steps of:
(a) sound signal of received code;
(b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal;
(c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from low band signal by the bandwidth extend information of using decoding;
(d), in low band signal and the high-frequency band signals each is transformed into time domain from frequency domain by in low band signal and the high-frequency band signals each is carried out the contrary discrete cosine transform of revising;
(e) with the low band signal of conversion and the high-frequency band signals combination of conversion.
16. method as claimed in claim 15, wherein, step (b) also comprises following at least one step:
The bit plane of re-quantization is carried out differentiate more to be synthesized;
Be included in the vector index in the sound signal of coding by use, the frequency linearity prediction result that will be undertaken by coding side and the bitplane combinations of re-quantization.
17. one kind records the computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio signal decoding on it, said method comprising the steps of:
(a) sound signal of received code;
(b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal;
(c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from low band signal by the bandwidth extend information of using decoding;
(d), in low band signal and the high-frequency band signals each is transformed into time domain from frequency domain by in low band signal and the high-frequency band signals each is carried out the contrary discrete cosine transform of revising;
(e) with the low band signal of conversion and the high-frequency band signals combination of conversion.
18. the method to audio signal decoding said method comprising the steps of:
(a) sound signal of received code;
(b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal;
(c) by low band signal being carried out the contrary discrete cosine transform of revising, low band signal is transformed into time domain from frequency domain;
(d) low band signal that will carry out the discrete cosine transform of contrary correction is transformed into frequency domain or time/frequency domain from time domain;
(e) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from the low band signal that is transformed into frequency domain or time/frequency domain by the bandwidth extend information of using decoding;
(f) with the high-frequency band signals inverse conversion to time domain;
(g) low band signal of conversion and the high-frequency band signals of inverse conversion are made up.
19. method as claimed in claim 18, wherein, step (b) also comprises following at least one step:
The bit plane of re-quantization is carried out differentiate more to be synthesized;
Be included in the vector index in the sound signal of coding by use, the frequency linearity prediction result that will be undertaken by coding side and the bitplane combinations of re-quantization.
20. one kind records the computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio signal decoding on it, said method comprising the steps of:
(a) sound signal of received code;
(b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal;
(c) by low band signal being carried out the contrary discrete cosine transform of revising, low band signal is transformed into time domain from frequency domain;
(d) low band signal that will carry out the discrete cosine transform of contrary correction is transformed into frequency domain or time/frequency domain from time domain;
(e) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from the low band signal that is transformed into frequency domain or time/frequency domain by the bandwidth extend information of using decoding;
(f) with the high-frequency band signals inverse conversion to time domain;
(g) low band signal of conversion and the high-frequency band signals of inverse conversion are made up.
21. the method to audio signal decoding said method comprising the steps of:
(a) sound signal of received code;
(b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal;
(c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from low band signal by the bandwidth extend information of using decoding;
(d) with low band signal and high-frequency band signals combination;
(e) pass through the signal of combination is carried out the contrary discrete cosine transform of revising, the signal that makes up is transformed into time domain from frequency domain.
22. method as claimed in claim 21, wherein, step (b) also comprises following at least one step:
The bit plane of re-quantization is carried out differentiate more to be synthesized;
Be included in the vector index in the sound signal of coding by use, the frequency linearity prediction result that will be undertaken by coding side and the bitplane combinations of re-quantization.
23. one kind records the computer readable recording medium storing program for performing that is used to carry out to the computer program of the method for audio signal decoding on it, said method comprising the steps of:
(a) sound signal of received code;
(b) by the bit plane of the coding in the sound signal that is included in coding is carried out based on contextual decoding and re-quantization, produce low band signal;
(c) the bandwidth extend information of the coding in the sound signal that is included in coding is decoded, and produce high-frequency band signals from low band signal by the bandwidth extend information of using decoding;
(d) with low band signal and high-frequency band signals combination;
(e) pass through the signal of combination is carried out the contrary discrete cosine transform of revising, the signal that makes up is transformed into time domain from frequency domain.
24. an equipment that is used for audio-frequency signal coding, described equipment comprises:
The frequency band separative element is separated into low band signal and high-frequency band signals with input signal;
Converting unit is transformed into frequency domain with in low band signal and the high-frequency band signals each from time domain;
The low-frequency band coding unit is to the low band signal execution quantification of conversion with based on contextual Bit-Plane Encoding;
The band spread coding unit, by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding.
25. equipment as claimed in claim 24, wherein, converting unit comprises:
The first discrete cosine transform applying unit of revising by low band signal is carried out the discrete cosine transform of revising, is transformed into frequency domain with low band signal from time domain;
The second discrete cosine transform applying unit of revising by high-frequency band signals is carried out the discrete cosine transform of revising, is transformed into frequency domain with high-frequency band signals from time domain.
26. equipment as claimed in claim 24, wherein, the low-frequency band coding unit comprises following at least one unit:
The frequency linearity prediction execution unit by the low band signal of conversion is carried out the frequency linearity prediction, is carried out filtering to the low band signal of conversion;
Many resolution decomposition are carried out to the low band signal of conversion in many resolution decomposition unit,
Wherein, the low band signal of filtering or the low band signal of having carried out many resolution decomposition are carried out quantized and based on contextual Bit-Plane Encoding.
27. equipment as claimed in claim 26, wherein, the frequency linearity prediction execution unit is predicted the coefficient that calculates linear prediction filter by the low band signal of conversion is carried out frequency linearity, and by using vector index to represent the analog value of coefficient.
28. equipment as claimed in claim 27 also comprises: Multiplexing Unit, the bandwidth extend information and the vector index of bit plane, the coding of coding is multiplexing.
29. an equipment that is used for audio-frequency signal coding, described equipment comprises:
The frequency band separative element is separated into low band signal and high-frequency band signals with input signal;
The discrete cosine transform applying unit of revising by low band signal is carried out the discrete cosine transform of revising, is transformed into frequency domain with low band signal from time domain;
The low-frequency band coding unit is carried out the low band signal of having carried out the discrete cosine transform of revising and to be quantized and based on contextual Bit-Plane Encoding;
Converting unit is transformed into frequency domain or time/frequency domain with in low band signal and the high-frequency band signals each from time domain;
Bandwidth extended coding unit, by using the low band signal of conversion, produce the expression conversion high-frequency band signals characteristic the bandwidth extend information and to described bandwidth extend information coding.
30. equipment as claimed in claim 29, wherein, the low-frequency band coding unit comprises following at least one unit:
The frequency linearity prediction execution unit by the low band signal of having carried out the discrete cosine transform of revising is carried out the frequency linearity prediction, is carried out filtering to the low band signal of having carried out the discrete cosine transform of revising;
Many resolution decomposition are carried out to the low band signal of having carried out the discrete cosine transform of revising in many resolution decomposition unit,
Wherein, the low band signal of filtering or the low band signal of having carried out many resolution decomposition are carried out quantized and based on contextual Bit-Plane Encoding.
31. equipment as claimed in claim 30, wherein, the frequency linearity prediction execution unit is predicted the coefficient that calculates linear prediction filter by the low band signal of having carried out the discrete cosine transform of revising is carried out frequency linearity, and by using vector index to represent the analog value of coefficient.
32. equipment as claimed in claim 31 also comprises: Multiplexing Unit, the bandwidth extend information and the vector index of bit plane, the coding of coding is multiplexing.
33. an equipment that is used for audio-frequency signal coding, described equipment comprises:
Converting unit is transformed into frequency domain with input signal from time domain;
The frequency band separative element is separated into low band signal and high-frequency band signals with the input signal of changing;
The low-frequency band coding unit is to low band signal execution quantification with based on contextual Bit-Plane Encoding;
Bandwidth extended coding unit is by using low band signal to produce the bandwidth extend information of the characteristic of representing high-frequency band signals and described bandwidth extend information being encoded.
34. equipment as claimed in claim 33, wherein, converting unit is transformed into frequency domain with input signal from time domain by input signal is carried out the discrete cosine transform of revising.
35. equipment as claimed in claim 33, wherein, the low-frequency band coding unit comprises following at least one unit:
The frequency linearity prediction execution unit by low band signal being carried out the frequency linearity prediction, is carried out filtering to low band signal;
Many resolution decomposition are carried out to low band signal in many resolution decomposition unit,
Wherein, the low band signal of filtering or the low band signal of having carried out many resolution decomposition are carried out quantized and based on contextual Bit-Plane Encoding.
36. equipment as claimed in claim 35, wherein, the frequency linearity prediction execution unit is predicted the coefficient that calculates linear prediction filter by low band signal is carried out frequency linearity, and by using vector index to represent the analog value of coefficient.
37. equipment as claimed in claim 36 also comprises: Multiplexing Unit, the bandwidth extend information and the vector index of bit plane, the coding of coding is multiplexing.
38. an equipment that is used for audio signal decoding, described equipment comprises:
The low-frequency band decoding unit by the bit plane of coding is carried out based on contextual decoding and re-quantization, produces low band signal;
Bandwidth expansion decoding unit, to the bandwidth extend information decoding of coding, and by using the bandwidth extend information of decoding to produce high-frequency band signals from low band signal;
The contrary discrete cosine transform applying unit of revising by in low band signal and the high-frequency band signals each is carried out the contrary discrete cosine transform of revising, is transformed into time domain with in low band signal and the high-frequency band signals each from frequency domain;
The combinations of bands unit is with the low band signal of conversion and the high-frequency band signals combination of conversion.
39. equipment as claimed in claim 38, wherein, the low-frequency band decoding unit comprises following at least one unit:
The synthesis units of differentiating more, it is synthetic that the bit plane of re-quantization is carried out many resolutions;
Frequency inverse linear prediction performance element, by using vector index, the frequency linearity prediction result that will be undertaken by coding side and the bitplane combinations of re-quantization.
40. an equipment that is used for audio signal decoding, described equipment comprises:
The low-frequency band decoding unit by the bit plane of coding is carried out based on contextual decoding and re-quantization, produces low band signal;
The contrary discrete cosine transform applying unit of revising by low band signal being carried out the contrary discrete cosine transform of revising, is transformed into time domain with low band signal from frequency domain;
Converting unit is transformed into frequency domain or time/frequency domain with the low band signal of having carried out the discrete cosine transform of contrary correction from time domain;
Bandwidth expansion decoding unit, to the bandwidth extend information decoding of coding, and by using the bandwidth extend information of decoding to produce high-frequency band signals from the low band signal that is transformed into frequency domain or time/frequency domain;
The inverse conversion unit, with the high-frequency band signals inverse conversion to time domain;
The combinations of bands unit makes up the low band signal of conversion and the high-frequency band signals of inverse conversion.
41. equipment as claimed in claim 40, wherein, the low-frequency band decoding unit comprises following at least one unit:
The synthesis units of differentiating more, it is synthetic that the bit plane of re-quantization is carried out many resolutions;
Frequency inverse linear prediction performance element, by using vector index, the frequency linearity prediction result that will be undertaken by coding side and the bitplane combinations of re-quantization.
42. an equipment that is used for audio signal decoding, described equipment comprises:
The low-frequency band decoding unit by the bit plane of coding is carried out based on contextual decoding and re-quantization, produces low band signal;
Bandwidth expansion decoding unit, to the bandwidth extend information decoding of coding, and by using the bandwidth extend information of decoding to produce high-frequency band signals from low band signal;
The combinations of bands unit is with low band signal and high-frequency band signals combination;
The discrete cosine transform applying unit of contrary correction by the signal of combination is carried out the contrary discrete cosine transform of revising, is transformed into time domain with the signal that makes up from frequency domain.
43. equipment as claimed in claim 42, wherein, the low-frequency band decoding unit comprises following at least one unit:
The synthesis units of differentiating more, it is synthetic that the bit plane of re-quantization is carried out many resolutions;
Frequency inverse linear prediction performance element, by using vector index, the frequency linearity prediction result that will be undertaken by coding side and the bitplane combinations of re-quantization.
CNA2007101533639A 2006-09-18 2007-09-18 Method and apparatus to encode and decode audio signal by using bandwidth extension technique Pending CN101162584A (en)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
KR1020060090152 2006-09-18
KR20060090152 2006-09-18
KR1020070079781 2007-08-08

Publications (1)

Publication Number Publication Date
CN101162584A true CN101162584A (en) 2008-04-16

Family

ID=39297493

Family Applications (1)

Application Number Title Priority Date Filing Date
CNA2007101533639A Pending CN101162584A (en) 2006-09-18 2007-09-18 Method and apparatus to encode and decode audio signal by using bandwidth extension technique

Country Status (2)

Country Link
KR (1) KR101346358B1 (en)
CN (1) CN101162584A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103210443A (en) * 2010-09-15 2013-07-17 三星电子株式会社 Apparatus and method for encoding and decoding signal for high frequency bandwidth extension
CN102089813B (en) * 2008-07-11 2013-11-20 弗劳恩霍夫应用研究促进协会 Audio encoder and audio decoder
CN104715756A (en) * 2015-02-10 2015-06-17 百度在线网络技术（北京）有限公司 Audio data processing method and device
CN104903956A (en) * 2012-10-10 2015-09-09 弗兰霍菲尔运输应用研究公司 Apparatus and method for efficient synthesis of sinusoids and sweeps by employing spectral patterns
CN103119646B (en) * 2010-07-20 2016-09-07 弗劳恩霍夫应用研究促进协会 Audio coder, audio decoder, the method for codes audio information and the method for decoded audio information
CN106463135A (en) * 2014-06-26 2017-02-22 高通股份有限公司 High-band signal coding using mismatched frequency ranges
CN106782575A (en) * 2011-06-01 2017-05-31 三星电子株式会社 Audio coding method and equipment, audio-frequency decoding method and equipment
CN106941003A (en) * 2011-10-21 2017-07-11 三星电子株式会社 Energy lossless coding method and equipment and energy losslessly encoding method and equipment

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3115991A4 (en) 2014-03-03 2017-08-02 Samsung Electronics Co., Ltd. Method and apparatus for high frequency decoding for bandwidth extension
SG10201808274UA (en) 2014-03-24 2018-10-30 Samsung Electronics Co Ltd High-band encoding method and device, and high-band decoding method and device
CN112309408A (en) * 2020-11-10 2021-02-02 北京百瑞互联技术有限公司 Method, device and storage medium for expanding LC3 audio encoding and decoding bandwidth

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2388502A (en) 2002-05-10 2003-11-12 Chris Dunn Compression of frequency domain audio signals
KR100561869B1 (en) 2004-03-10 2006-03-17 삼성전자주식회사 Lossless audio decoding/encoding method and apparatus

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102089813B (en) * 2008-07-11 2013-11-20 弗劳恩霍夫应用研究促进协会 Audio encoder and audio decoder
CN103119646B (en) * 2010-07-20 2016-09-07 弗劳恩霍夫应用研究促进协会 Audio coder, audio decoder, the method for codes audio information and the method for decoded audio information
CN105719655B (en) * 2010-09-15 2020-03-27 三星电子株式会社 Apparatus and method for encoding and decoding signal for high frequency bandwidth extension
CN103210443A (en) * 2010-09-15 2013-07-17 三星电子株式会社 Apparatus and method for encoding and decoding signal for high frequency bandwidth extension
US9183847B2 (en) 2010-09-15 2015-11-10 Samsung Electronics Co., Ltd. Apparatus and method for encoding and decoding signal for high frequency bandwidth extension
CN105719655A (en) * 2010-09-15 2016-06-29 三星电子株式会社 Apparatus and method for encoding and decoding signal for high frequency bandwidth extension
US10418043B2 (en) 2010-09-15 2019-09-17 Samsung Electronics Co., Ltd. Apparatus and method for encoding and decoding signal for high frequency bandwidth extension
US9837090B2 (en) 2010-09-15 2017-12-05 Samsung Electronics Co., Ltd. Apparatus and method for encoding and decoding signal for high frequency bandwidth extension
CN106782575A (en) * 2011-06-01 2017-05-31 三星电子株式会社 Audio coding method and equipment, audio-frequency decoding method and equipment
CN106782575B (en) * 2011-06-01 2020-12-18 三星电子株式会社 Audio encoding method and apparatus, audio decoding method and apparatus
CN107025909B (en) * 2011-10-21 2020-12-29 三星电子株式会社 Energy lossless encoding method and apparatus, and energy lossless decoding method and apparatus
CN106941003A (en) * 2011-10-21 2017-07-11 三星电子株式会社 Energy lossless coding method and equipment and energy losslessly encoding method and equipment
CN107025909A (en) * 2011-10-21 2017-08-08 三星电子株式会社 Energy lossless coding method and equipment and energy losslessly encoding method and equipment
US11355129B2 (en) 2011-10-21 2022-06-07 Samsung Electronics Co., Ltd. Energy lossless-encoding method and apparatus, audio encoding method and apparatus, energy lossless-decoding method and apparatus, and audio decoding method and apparatus
CN106941003B (en) * 2011-10-21 2021-01-26 三星电子株式会社 Energy lossless encoding method and apparatus, and energy lossless decoding method and apparatus
US10878827B2 (en) 2011-10-21 2020-12-29 Samsung Electronics Co.. Ltd. Energy lossless-encoding method and apparatus, audio encoding method and apparatus, energy lossless-decoding method and apparatus, and audio decoding method and apparatus
CN104903956B (en) * 2012-10-10 2018-11-16 弗劳恩霍夫应用研究促进协会 For being effectively synthesized the device and method of sine curve and scanning by using spectrum mode
CN104903956A (en) * 2012-10-10 2015-09-09 弗兰霍菲尔运输应用研究公司 Apparatus and method for efficient synthesis of sinusoids and sweeps by employing spectral patterns
CN106463135A (en) * 2014-06-26 2017-02-22 高通股份有限公司 High-band signal coding using mismatched frequency ranges
CN106463135B (en) * 2014-06-26 2019-11-12 高通股份有限公司 It is decoded using the high-frequency band signals of mismatch frequency range
CN104715756A (en) * 2015-02-10 2015-06-17 百度在线网络技术（北京）有限公司 Audio data processing method and device

Also Published As

Publication number Publication date
KR101346358B1 (en) 2013-12-31
KR20080025636A (en) 2008-03-21

Similar Documents

Publication Publication Date Title
CN101162584A (en) Method and apparatus to encode and decode audio signal by using bandwidth extension technique
CN101518083B (en) Method, medium, and system encoding and/or decoding audio signals by using bandwidth extension and stereo coding
JP4506039B2 (en) Encoding apparatus and method, decoding apparatus and method, and encoding program and decoding program
CN100454389C (en) Sound encoding apparatus and sound encoding method
CN101223576B (en) Method and apparatus to extract important spectral component from audio signal and low bit-rate audio signal coding and/or decoding method and apparatus using the same
KR100958144B1 (en) Audio Compression
KR101238239B1 (en) An encoder
US9037454B2 (en) Efficient coding of overcomplete representations of audio using the modulated complex lapped transform (MCLT)
CN101371447A (en) Complex-transform channel coding with extended-band frequency coding
RU97122037A (en) METHOD AND DEVICE FOR SCALABLE ENCODING / DECODING OF THE AUDIO SIGNAL
JP2006011456A (en) Method and device for coding/decoding low-bit rate and computer-readable medium
KR20080005325A (en) Method and apparatus for adaptive encoding/decoding
CN102511062A (en) Allocation of bits in an enhancement coding/decoding for improving a hierarchical coding/decoding of digital audio signals
US20080071550A1 (en) Method and apparatus to encode and decode audio signal by using bandwidth extension technique
EP0919989A1 (en) Audio signal encoder, audio signal decoder, and method for encoding and decoding audio signal
JP2003110429A (en) Coding method and device, decoding method and device, transmission method and device, and storage medium
US20040083094A1 (en) Wavelet-based compression and decompression of audio sample sets
JP3297749B2 (en) Encoding method
JP3475985B2 (en) Information encoding apparatus and method, information decoding apparatus and method
US20120123788A1 (en) Coding method, decoding method, and device and program using the methods
WO1999044291A1 (en) Coding device and coding method, decoding device and decoding method, program recording medium, and data recording medium
JP3557164B2 (en) Audio signal encoding method and program storage medium for executing the method
JP2004246038A (en) Speech or musical sound signal encoding method, decoding method, encoding device, decoding device, encoding program, and decoding program
JP2016035501A (en) Voice encoding device, voice encoding method, voice encoding computer program, voice decoding device, voice decoding method, and voice decoding computer program
James et al. A comparative study of speech compression using different transform techniques

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C12 Rejection of a patent application after its publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20080416