WO2010092827A1 - ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 - Google Patents
ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 Download PDFInfo
- Publication number
- WO2010092827A1 WO2010092827A1 PCT/JP2010/000881 JP2010000881W WO2010092827A1 WO 2010092827 A1 WO2010092827 A1 WO 2010092827A1 JP 2010000881 W JP2010000881 W JP 2010000881W WO 2010092827 A1 WO2010092827 A1 WO 2010092827A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- vector
- code
- quantization
- matrix
- lsp
- Prior art date
Links
- 239000013598 vector Substances 0.000 title claims abstract description 559
- 238000013139 quantization Methods 0.000 title claims abstract description 268
- 238000000034 method Methods 0.000 title claims description 42
- 239000011159 matrix material Substances 0.000 claims abstract description 245
- 238000013507 mapping Methods 0.000 description 79
- 230000005284 excitation Effects 0.000 description 69
- 238000004364 calculation method Methods 0.000 description 57
- 230000003044 adaptive effect Effects 0.000 description 35
- 238000009826 distribution Methods 0.000 description 21
- 238000000926 separation method Methods 0.000 description 17
- 230000015572 biosynthetic process Effects 0.000 description 13
- 238000003786 synthesis reaction Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 11
- 238000012545 processing Methods 0.000 description 11
- 101100119135 Mus musculus Esrrb gene Proteins 0.000 description 7
- 238000004891 communication Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 7
- 230000005236 sound signal Effects 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 230000003595 spectral effect Effects 0.000 description 5
- 238000007781 pre-processing Methods 0.000 description 4
- 230000010354 integration Effects 0.000 description 3
- 229940050561 matrix product Drugs 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 238000010295 mobile communication Methods 0.000 description 2
- 238000012805 post-processing Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000007493 shaping process Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; 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
- G10L19/038—Vector quantisation, e.g. TwinVQ audio
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; 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
- G10L19/06—Determination or coding of the spectral characteristics, e.g. of the short-term prediction coefficients
- G10L19/07—Line spectrum pair [LSP] vocoders
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; 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
- G10L19/08—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
- G10L19/12—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
- H03M7/3082—Vector coding
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; 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
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; 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
- G10L2019/0001—Codebooks
- G10L2019/0004—Design or structure of the codebook
- G10L2019/0005—Multi-stage vector quantisation
Definitions
- the present invention relates to a vector quantization apparatus, a vector inverse quantization apparatus, and a method for vector quantization of LSP (Line Spectral Pairs) parameters, and particularly to a packet communication system represented by Internet communication, a mobile communication system, and the like.
- the present invention relates to a vector quantization apparatus, a vector inverse quantization apparatus, and a method thereof that perform vector quantization of LSP parameters used in a speech encoding / decoding apparatus that transmits speech signals.
- audio signal encoding / decoding technology is indispensable for effective use of transmission path capacity such as radio waves and storage media. is there.
- CELP Code Excited Linear Prediction
- a CELP speech encoding apparatus encodes input speech based on a speech model stored in advance. Specifically, the CELP speech coding apparatus divides a digitized speech signal into frames with a constant time interval of about 10 to 20 ms, and performs linear prediction analysis on the speech signal in each frame to perform linear prediction. A coefficient (LPC: Linear Prediction Coefficient) and a linear prediction residual vector are obtained, and the linear prediction coefficient and the linear prediction residual vector are individually encoded.
- LPC Linear Prediction Coefficient
- LSP Line Spectral Pairs
- vector quantization is often performed on the LSP parameter.
- a code vector that is closest to the vector to be quantized is selected from a code book (code book) having a plurality of representative vectors (code vectors), and assigned to the selected code vector. This is a method of outputting a current index (code) as a quantization result.
- Multi-stage vector quantization is a method in which a vector is quantized once and then the quantization error is further vector-quantized.
- Divided vector quantization is a method in which each divided vector obtained by dividing a vector is quantized. It is a method to convert.
- the LSP can be switched by appropriately switching the codebook used for vector quantization in accordance with speech characteristics (eg, voice voiced, unvoiced, mode information, etc.) having a correlation with the LSP to be quantized.
- speech characteristics eg, voice voiced, unvoiced, mode information, etc.
- classification vector quantization classification VQ
- a narrowband LSP is classified according to characteristics using the interrelationship between a wideband LSP (LSP obtained from a wideband signal) and a narrowband LSP (LSP obtained from a narrowband signal).
- the first-stage codebook of multistage vector quantization is switched according to the type of narrowband LSP feature (hereinafter referred to as the type of narrowband LSP), and the wideband LSP is vector quantized.
- first-stage vector quantization is performed using a codebook corresponding to the type of narrowband LSP.
- the deviation of the distribution of quantization errors differs depending on the type of narrowband LSP.
- the second and subsequent vector quantization uses one common codebook regardless of the type of narrowband LSP, there is a problem that the second and subsequent vector quantization accuracy is insufficient.
- FIG. 1 is a diagram for explaining the problems in the multistage vector quantization described above.
- a black circle indicates a two-dimensional vector
- a broken-line circle schematically indicates the magnitude of variance of the vector set
- the center of the circle indicates the average of the vector set.
- CBa1, CBa2,..., CBab correspond to each type of narrowband LSP and indicate a plurality of codebooks used for the first stage vector quantization.
- CBb indicates a code book used for second-stage vector quantization.
- An object of the present invention is to improve the quantization accuracy of vector quantization in the second and subsequent stages in multi-stage vector quantization in which the first stage codebook is switched according to the type of feature having a correlation with the quantization target vector. It is to provide a vector quantization device, a vector dequantization device, and methods thereof that can be used.
- the vector quantization apparatus includes: a first selection unit that selects a classification code vector indicating a type of feature having a correlation with a quantization target vector from a plurality of classification code vectors; Using a second selection means for selecting a first codebook corresponding to the selected classification code vector from among the codebooks, and a plurality of first code vectors constituting the selected first codebook First quantization means for quantizing the quantization target vector to obtain a first code; and third selection means for selecting a first matrix corresponding to the selected classification code vector from a plurality of matrices; , Using a plurality of second code vectors and the selected first matrix, a first residual vector which is a difference between the first code vector indicated by the first code and the vector to be quantized.
- the torque quantized employs a configuration comprising a second quantizing means for obtaining a second code, the.
- the vector inverse quantization apparatus includes a first code obtained by quantizing a vector to be quantized in the vector quantization apparatus, a second code obtained by further quantizing the quantization error of the quantization, , Receiving means for selecting, a first selecting means for selecting a classification code vector indicating a type of feature having a correlation with the quantization target vector from among a plurality of classification code vectors, and a plurality of first codes Second selection means for selecting a first code book corresponding to the selected code vector for classification from among the code books, and a plurality of first code vectors constituting the selected first code book A first inverse quantization means for designating a first code vector corresponding to the first code; and a matrix corresponding to the selected classification code vector is selected from a plurality of matrices.
- a third selecting unit designating a second code vector corresponding to the second code from a plurality of second code vectors; designating the designated second code vector; selecting the matrix; And a second inverse quantization unit that obtains a quantization vector using the first code vector.
- the vector quantization method of the present invention includes a step of selecting a classification code vector indicating a type of feature having a correlation with a quantization target vector from a plurality of classification code vectors, and a plurality of first codebooks. Selecting a first codebook corresponding to the selected classification code vector, and using the plurality of first code vectors constituting the selected first codebook to determine the quantization target vector Quantizing to obtain a first code; selecting a first matrix corresponding to the selected classification code vector from a plurality of matrices; a plurality of second code vectors and the selected first code 1 matrix is used to quantize the first residual vector, which is the difference between the first code vector indicated by the first code and the vector to be quantized, Obtaining a second code, and to have.
- the vector inverse quantization method of the present invention includes a first code obtained by quantizing a vector to be quantized in a vector quantization apparatus, a second code obtained by further quantizing the quantization error of the quantization, , A step of selecting a classification code vector indicating a type of feature having a correlation with the quantization target vector from among a plurality of classification code vectors, and a plurality of first codebooks Selecting a first codebook corresponding to the selected classification code vector, and corresponding to the first code from among a plurality of first code vectors constituting the selected first codebook Designating a first code vector to be selected, selecting a matrix corresponding to the selected classification code vector from a plurality of matrices, A second code vector corresponding to the second code is designated from among the second code vectors, and the designated second code vector, the selected matrix, and the designated first code vector And a step of obtaining a quantization vector.
- the second-stage and subsequent vectors using the matrix corresponding to the type.
- the wideband LSP is set as a vector quantization target, and the type of narrowband LSP having a correlation with the vector quantization target is used.
- the code book used for the quantization of the eyes is switched will be described as an example.
- a codebook used for the first-stage quantization may be switched using a quantized narrowband LSP (a narrowband LSP pre-quantized by a not-shown narrowband LSP quantizer).
- the quantized narrowband LSP may be converted into a wideband form, and the codebook used for the first-stage quantization may be switched using the converted quantized narrowband LSP.
- mapping matrix a matrix for moving the bias of the distribution of code vectors by performing matrix operation on all code vectors constituting the code book.
- the mapping matrix is used to perform a matrix operation on a vector to be quantized as in the embodiment of the present invention, rather than being used to perform a matrix operation on a code vector. There are many cases.
- FIG. 2 is a block diagram showing a main configuration of LSP vector quantization apparatus 100 according to the embodiment of the present invention.
- the LSP vector quantization apparatus 100 quantizes an input LSP vector by three-stage multi-level vector quantization.
- the LSP vector quantization apparatus 100 includes a classifier 101, a switch 102, a first code book 103, an adder 104, an error minimizing unit 105, a matrix determining unit 106, a matrix calculating unit 107, and a second code book 108. , An adder 109, a matrix calculation unit 110, a third codebook 111, and an adder 112.
- the classifier 101 stores in advance a classification code book composed of a plurality of pieces of classification information indicating a plurality of types of narrowband LSP vectors.
- the classifier 101 selects classification information indicating the type of the wideband LSP vector that is the vector quantization target from the classification codebook, and outputs the classification information to the switch 102 and the matrix determination unit 106.
- the classifier 101 has a built-in classification code book composed of code vectors (classification code vectors) corresponding to each type of narrowband LSP vector.
- the classifier 101 searches the codebook for classification to obtain a code vector that minimizes the square error with the input narrowband LSP vector.
- the classifier 101 uses the index of the code vector obtained by the search as classification information indicating the type of the LSP vector.
- the switch 102 selects one subcodebook corresponding to the classification information input from the classifier 101 from the first codebook 103 and connects the output terminal of the subcodebook to the adder 104.
- the first codebook 103 stores in advance subcodebooks (CBa1 to CBan) corresponding to each type of narrowband LSP. That is, for example, when the total number of types of narrowband LSP is n, the number of subcodebooks constituting the first codebook 103 is also n.
- the first code book 103 outputs, to the switch 102, the first code vector designated by the instruction from the error minimizing unit 105 among the plurality of first code vectors constituting the sub code book selected by the switch 102. To do.
- the adder 104 calculates a difference between the wideband LSP vector input as a vector quantization target and the first code vector input from the switch 102, and outputs the difference to the error minimizing unit 105 as a first residual vector. To do. Further, adder 104 outputs to matrix calculation unit 107 one of the first residual vectors corresponding to each of the first code vectors, which is found to be minimum by the search of error minimizing unit 105.
- the error minimizing unit 105 sets the squared error of the wideband LSP vector and the first code vector as a result of squaring the first residual vector input from the adder 104, and searches for the first codebook to find this square error. Find the first code vector that minimizes. Similarly, the error minimizing unit 105 searches the second codebook using the squared error of the first residual vector and the second code vector as a result of squaring the second residual vector input from the adder 109. Thus, a second code vector that minimizes the square error is obtained. Similarly, error minimizing section 105 uses the result of squaring the third residual vector input from adder 112 as the square error between the second residual vector and the third code vector, and searches for the third codebook. Thus, a third code vector that minimizes this square error is obtained. The error minimizing unit 105 collectively encodes the indexes assigned to the three code vectors obtained by the search, and outputs the encoded data.
- the matrix determination unit 106 stores in advance a mapping matrix codebook including mapping matrices corresponding to each type of narrowband LSP vector.
- the mapping matrix codebook is composed of two types of mapping matrix codebooks, a first mapping matrix codebook and a second mapping matrix codebook.
- the first mapping matrix codebook includes a first mapping matrix used for performing matrix operation on the first code vector
- the second mapping matrix codebook performs matrix operation on the second code vector. It consists of a second mapping matrix used to do.
- the matrix determination unit 106 selects a first mapping matrix and a second mapping matrix corresponding to the classification information input from the classifier 101 from the mapping matrix codebook. Then, the matrix determination unit 106 outputs the inverse matrix of the selected first mapping matrix to the matrix operation unit 107, and outputs the inverse matrix of the selected second mapping matrix to the matrix operation unit 110.
- the matrix calculation unit 107 performs matrix calculation on the first residual vector input from the adder 104 using the inverse matrix of the first mapping matrix input from the matrix determination unit 106, and the vector after matrix calculation Is output to the adder 109.
- the second code book (CBb) 108 is composed of a plurality of second code vectors, and outputs the second code vector designated by the instruction from the error minimizing unit 105 to the adder 109.
- the adder 109 obtains a difference between the vector after the matrix calculation input from the matrix calculation unit 107 and the second code vector input from the second codebook 108, and uses the difference as a second residual vector as an error. Output to the minimizing unit 105. Further, adder 109 outputs one of the second residual vectors corresponding to each of the second code vectors, which is found to be the minimum by the search of error minimizing section 105, to matrix computing section 110.
- the matrix calculation unit 110 performs a matrix calculation on the second residual vector input from the adder 109 using an inverse matrix of the second mapping matrix input from the matrix determination unit 106, and a vector after the matrix calculation Is output to the adder 112.
- the third code book 111 (CBc) is composed of a plurality of third code vectors, and outputs the third code vector designated by the instruction from the error minimizing unit 105 to the adder 112.
- the adder 112 obtains a difference between the vector after matrix calculation input from the matrix calculation unit 110 and the third code vector input from the third codebook 111, and uses this difference as a third residual vector to minimize the error. To the conversion unit 105.
- the operation performed by the LSP vector quantization apparatus 100 will be described by taking as an example the case where the order of the wideband LSP vector to be quantized is the R order.
- the classifier 101 has a built-in classification codebook composed of n code vectors (classification code vectors) corresponding to each of the n types of narrowband LSP vectors.
- the classifier 101 searches the code vector to obtain the mth code vector that minimizes the square error with the input narrowband LSP vector.
- the classifier 101 outputs m (1 ⁇ m ⁇ n) as classification information to the switch 102 and the matrix determination unit 106.
- the switch 102 selects the sub code book CBam corresponding to the classification information m from the first code book 103, and connects the output terminal of the sub code book to the adder 104.
- D1 is the total number of code vectors of the first codebook
- d1 is the index of the first code vector.
- the error minimizing unit 105 stores the first code vector index d1 'that minimizes the square error Err as the first index d1_min.
- the matrix determination unit 106 selects the inverse matrix MM_1 ⁇ 1 (m) of the first mapping matrix corresponding to the classification information m from the first mapping matrix codebook, and outputs it to the matrix calculation unit 107. Further, the matrix determination unit 106 selects the inverse matrix MM_2 ⁇ 1 (m) of the second mapping matrix corresponding to the classification information m from the second mapping matrix codebook, and outputs it to the matrix calculation unit 110.
- D2 is the total number of code vectors of the second codebook
- d2 is the code vector index.
- the error minimizing unit 105 stores the second code vector index d2 'that minimizes the square error Err as the second index d2_min.
- D3 is the total number of code vectors of the third codebook
- d3 is the code vector index.
- the error minimizing unit 105 stores the index d3 'of the third code vector that minimizes the square error Err as the third index d3_min. Then, the error minimizing unit 105 collectively encodes the first index d1_min, the second index d2_min, and the third index d3_min, and outputs the encoded data.
- FIG. 3 is a block diagram showing the main configuration of LSP vector inverse quantization apparatus 200 according to the present embodiment.
- the LSP vector inverse quantization apparatus 200 decodes the encoded data output from the LSP vector quantization apparatus 100, and generates a quantized LSP vector.
- the LSP vector inverse quantization apparatus 200 includes a classifier 201, a code separation unit 202, a switch 203, a first code book 204, a matrix determination unit 205, a second code book (CBb) 206, a matrix calculation unit 207, an adder 208, A third code book (CBc) 209, a matrix calculation unit 210, and an adder 211 are provided.
- the first codebook 204 includes a subcodebook having the same contents as the subcodebooks (CBa1 to CBan) included in the first codebook 103
- the matrix determination unit 205 includes a mapping matrix codebook included in the matrix determination unit 106.
- the second code book 206 includes a code book having the same contents as the code book included in the second code book 108
- the third code book 209 includes a code book having the same contents as the code book included in the third code book 111.
- the classifier 201 stores in advance a classification codebook including a plurality of classification information indicating a plurality of types of narrowband LSP vectors.
- the classifier 201 selects classification information indicating the type of the wideband LSP vector that is a vector quantization target from the classification codebook, and outputs the classification information to the switch 203 and the matrix determination unit 205.
- the classifier 201 has a built-in classification code book composed of code vectors (classification code vectors) corresponding to each type of narrowband LSP vector.
- the classifier 201 searches the classification codebook to obtain a code vector that minimizes the square error with the quantized narrowband LSP vector input from a narrowband LSP quantizer (not shown).
- the classifier 201 uses the code vector index obtained by the search as classification information indicating the type of the LSP vector.
- the code separation unit 202 separates the encoded data transmitted from the LSP vector quantization apparatus 100 into a first index, a second index, and a third index.
- the code separation unit 202 instructs the first code book 204 for the first index, instructs the second code book 206 for the second index, and instructs the third code book 209 for the third index.
- the switch 203 selects one subcodebook (CBam) corresponding to the classification information input from the classifier 201 from the first codebook 204, and connects the output terminal of the subcodebook to the adder 208.
- CBam subcodebook
- the first code book 204 corresponds to the first index designated by the code separation unit 202 from among a plurality of first code vectors constituting the first code book (subcode book) selected by the switch 203. Two first code vectors are output to the switch 203.
- the matrix determination unit 205 selects a first mapping matrix corresponding to the classification information input from the classifier 201 from the first mapping matrix codebook, and selects the selected first mapping matrix as a matrix calculation unit 207 and a matrix calculation unit. Output to 210.
- the matrix determination unit 205 selects a second mapping matrix corresponding to the classification information input from the classifier 201 from the second mapping matrix codebook, and outputs the selected second mapping matrix to the matrix calculation unit 210. To do.
- the second code book 206 outputs one second code vector corresponding to the second index designated by the code separation unit 202 to the matrix calculation unit 207.
- the matrix calculation unit 207 performs matrix calculation on the second code vector input from the second codebook 206 using the first mapping matrix input from the matrix determination unit 205, and the matrix calculation result is added to the adder 208. Output to.
- the adder 208 adds the first code vector input from the switch 203 and the matrix calculation result input from the matrix calculation unit 207, and outputs the obtained addition result to the adder 211.
- the third code book 209 outputs one third code vector corresponding to the third index specified by the code separation unit 202 to the matrix calculation unit 210.
- the matrix calculation unit 210 performs matrix calculation on the third code vector input from the third codebook 209 using the second mapping matrix input from the matrix determination unit 205, and the matrix calculation result is added to the adder 211. Output to.
- the adder 211 adds the addition result input from the adder 208 and the matrix calculation result input from the matrix calculation unit 210, and outputs the obtained addition result as a quantized broadband LSP vector.
- the classifier 201 has a built-in classification codebook composed of n code vectors (classification code vectors) corresponding to each of the n types of narrowband LSP vectors.
- the classifier 201 searches the code vector to obtain the mth code vector that minimizes the square error with the quantized narrowband LSP vector input from a narrowband LSP quantizer (not shown).
- the classifier 201 outputs m (1 ⁇ m ⁇ n) as classification information to the switch 203 and the matrix determination unit 205.
- the code separation unit 202 separates the encoded data transmitted from the LSP vector quantization apparatus 100 into the first index d1_min, the second index d2_min, and the third index d3_min.
- the code separation unit 202 instructs the first code book 204 for the first index d1_min, instructs the second code book 206 for the second index d2_min, and instructs the third code book 209 for the third index d3_min.
- the switch 203 selects the sub code book CBam corresponding to the classification information m input from the classifier 201 from the first code book 204, and connects the output terminal of the sub code book to the adder 208.
- the matrix determination unit 205 selects the first mapping matrix MM_1 (m) corresponding to the classification information m input from the classifier 201 from the first mapping matrix codebook, and sends it to the matrix calculation unit 207 and the matrix calculation unit 210. Output. Further, the matrix determination unit 205 selects the second mapping matrix MM_2 (m) corresponding to the classification information m input from the classifier 201 from the second mapping matrix codebook, and outputs it to the matrix calculation unit 210.
- the first codebook, mapping matrix codebook, second codebook, and third codebook used in the LSP vector quantization apparatus 100 and the LSP vector inverse quantization apparatus 200 are created by learning in advance, A method for learning these codebooks will be described.
- V LSP vectors obtained from a large number of learning speech data are prepared.
- V LSP vectors are grouped for each type (n types), and D1 first code vectors CODE_1 (d1 ) are used according to a learning algorithm such as an LBG (Linde Buzo Gray) algorithm using the LSP vectors belonging to each group.
- LBG Longde Buzo Gray
- the first code book obtained by the above method using the V LSP vectors described above is used.
- the first codebook obtained by the above method using the V LSP vectors described above is used.
- the obtained V first residual vectors are grouped for each type (n types), and a centroid (average) C1 of the first residual vector set belonging to each group is obtained.
- centroid D1 is obtained by calculating the average of all the second code vectors in the second code book.
- centroid C1 of the first codebook and the centroid D1 of the second codebook are obtained for each type (n types). Then, a matrix operation is performed on the centroid D1 corresponding to the type m to obtain a matrix in which the centroid C1 and the centroid D1 match, and the obtained matrix is set as a first mapping matrix corresponding to the type m.
- a first mapping matrix is obtained for each type (n types), and a first mapping matrix codebook is generated by assigning a serial number to each type of first mapping matrix and storing it.
- the first mapping matrix MM_1 (d2) of the type m is obtained by solving the simultaneous equations of the following equation (14).
- the first mapping matrix MM_1 (m) can be generated by obtaining the mapping matrix in two-dimensional units and obtaining the matrix product of the obtained mapping matrix.
- a matrix TMP_1 that can map vector elements corresponding to the first and second orders is obtained by solving the simultaneous equations of the following equation (15).
- a first mapping matrix MM_1 (m) is generated by obtaining a matrix product of the matrix TMP_1 and the matrix TMP_2 according to the following equation (17).
- the matrix is obtained in units of two dimensions, and finally the mapping matrix is obtained by obtaining the matrix product of all the matrices.
- the first codebook obtained by the above method using the V LSP vectors described above is used.
- the obtained V second residual vectors are grouped for each type (n types), and the centroid C2 of the second residual vector set belonging to each group is obtained.
- centroid D2 is obtained by calculating the average of all the third code vectors.
- centroid C2 and the centroid D2 of the third codebook are obtained. Then, a matrix operation is performed on the centroid D2 corresponding to the type m to obtain a matrix in which the centroid C2 and the centroid D2 match, and the obtained matrix is set as a second mapping matrix corresponding to the type m.
- a second mapping matrix is obtained for each type (n types), and a second mapping matrix codebook is generated by assigning a serial number to each type of second mapping matrix and storing it.
- the second mapping matrix MM_2 (m) is the same procedure after the centroid C1 is replaced with the centroid C2 and the centroid D1 is replaced with the centroid D2 in the method for obtaining the first mapping matrix MM_1 (m). Can be obtained by performing
- each codebook may be generated by a method other than the above method.
- the first stage vector quantization codebook is switched according to the type of the narrowband LSP vector having a correlation with the wideband LSP vector.
- a matrix operation is performed on the first residual vector using the first mapping matrix corresponding to the classification result of the narrowband LSP vector.
- the bias in the distribution of the second-stage code vector can be adapted to the statistical distribution bias in the first-stage vector quantization error, and hence the quantization accuracy of the wideband LSP vector can be improved.
- matrix calculation is performed on the second residual vector using the second mapping matrix corresponding to the classification result of the narrowband LSP vector.
- the bias in the distribution of the third-stage code vector can be adapted to the bias in the distribution of the second-stage vector quantization error, and hence the quantization accuracy of the wideband LSP vector can be improved.
- FIG. 4 is a diagram for conceptually explaining the effect of LSP vector quantization according to the present embodiment.
- an arrow written as “rotation” indicates a process of performing a matrix operation on the second code vector set using a mapping matrix.
- matrix calculation is performed using a mapping matrix corresponding to this type.
- the distribution bias of the set of quantization error vectors after the matrix operation is matched with the distribution bias of the set of second code vectors constituting the common second codebook CBb used for the second stage vector quantization. be able to. Accordingly, the quantization accuracy of the second stage vector quantization can be improved.
- the statistical distribution bias of the first stage vector quantization error is changed according to the bias of the second stage code vector distribution, and the second stage vector quantization error is changed.
- the case where the statistical distribution bias is changed according to the third-stage code vector distribution has been described as an example.
- the present invention is not limited to this, and the bias of the distribution of the code vector used for the second-stage vector quantization target is changed in accordance with the statistical distribution of the first-stage vector quantization error.
- the bias of the distribution of the code vector used for the stage vector quantization target may be changed according to the statistical distribution bias of the second stage vector quantization error. This also provides the effect of improving the quantization accuracy of the wideband LSP vector, as in the present embodiment.
- mapping matrix constituting the mapping matrix codebook included in the matrix determination unit 106 and the matrix determination unit 205 corresponds to the type of narrowband LSP vector
- the present invention is not limited to this, and the mapping matrix constituting the mapping matrix codebook included in the matrix determination unit 106 and the matrix determination unit 205 may correspond to each type in which the features of speech are classified.
- the classifier 101 inputs not the narrowband LSP vector but a parameter representing the voice feature as the voice feature information, and determines the switch 102 and the matrix using the voice feature type corresponding to the inputted voice feature information as the classification information. To the unit 106.
- VMR-WB Very-Rate Multimode Wideband Speech Codec
- the type information may be used as it is as a voice feature amount.
- the present invention is not limited to this, and the two-stage vector quantization or four or more stages are performed.
- the present invention can also be applied when performing vector quantization.
- the wideband LSP vector is described as an example of the quantization target.
- the quantization target is not limited to this and may be a vector other than the wideband LSP vector.
- the case where the bias of the distribution of the code vector is moved by performing a matrix operation using the mapping matrix has been described as an example.
- the present invention is not limited to this, and the bias of the distribution of the code vector may be moved by performing a matrix operation using a rotation matrix.
- LSP vector inverse quantization apparatus 200 decodes encoded data output from LSP vector quantization apparatus 100.
- the present invention is not limited to this, and any encoded data in a format that can be decoded by the LSP vector inverse quantization device 200 can be received and decoded by the LSP vector inverse quantization device. Needless to say.
- the LSP vector quantization apparatus 100 and the LSP vector inverse quantization apparatus 200 have been described with reference to an example in which a matrix operation is performed on an R-dimensional vector using an R ⁇ R mapping matrix.
- the present invention is not limited to this, and the LSP vector quantization apparatus 100 and the LSP vector inverse quantization apparatus 200 prepare, for example, a plurality of 2 ⁇ 2 mapping matrices, and each second-order vector element of an R-dimensional vector.
- a matrix operation may be performed using a plurality of 2 ⁇ 2 mapping matrices. According to this configuration, it is possible to reduce the memory required for storing the mapping matrix, and further reduce the amount of calculation required for the matrix calculation.
- Equation (18) the above equation (8) is represented by the following equation (19).
- mapping matrix (MMA — 1 (m) , MMB — 1 (m), and MMC — 1 (m) ) by learning
- a learning method when the above-described vector dimension number R is two-dimensional (the above formula (14 )) May be performed for each secondary vector element.
- the vector quantization apparatus and the vector inverse quantization apparatus according to the present embodiment can be used in a CELP encoding apparatus / CELP decoding apparatus that encodes / decodes a speech signal, a musical sound signal, and the like.
- the CELP encoding apparatus an LSP converted from a linear prediction coefficient obtained by linear prediction analysis of an input signal is input, quantization processing is performed, and the quantized quantized LSP is output to a synthesis filter.
- the LSP vector quantization apparatus 100 according to the present embodiment is applied to a CELP speech coding apparatus, the LSP quantization unit that outputs a quantized LSP code representing the quantized LSP as encoded data is provided.
- the LSP vector quantization apparatus 100 according to the present embodiment is arranged.
- the quantized LSP is decoded from the quantized LSP code obtained by separating the received multiplexed code data.
- the LSP vector inverse quantization apparatus 200 When the LSP vector inverse quantization apparatus 200 according to the present invention is applied to a CELP speech decoding apparatus, the LSP inverse quantization unit that outputs the decoded quantized LSP to the synthesis filter is connected to the present embodiment.
- the LSP vector inverse quantization apparatus 200 may be disposed, and the same effect as described above can be obtained.
- CELP encoding apparatus 400 and CELP decoding apparatus 450 including LSP vector quantization apparatus 100 and LSP vector inverse quantization apparatus 200 according to the present embodiment will be described using FIG. 5 and FIG.
- FIG. 5 is a block diagram showing a main configuration of CELP encoding apparatus 400 including LSP vector quantization apparatus 100 according to the present embodiment.
- the CELP encoding apparatus 400 divides the input voice / musical sound signal into a plurality of samples, and encodes each frame with the plurality of samples as one frame.
- the pre-processing unit 401 performs high-pass filter processing for removing DC components on the input audio signal or musical tone signal, and performs waveform shaping processing or pre-emphasis processing for improving the performance of subsequent encoding processing. Then, the preprocessing unit 401 outputs the signal Xin obtained by these processes to the LSP analysis unit 402 and the adder 405.
- the LSP analysis unit 402 performs linear prediction analysis using the signal Xin input from the preprocessing unit 401, converts the obtained LPC into an LSP vector, and outputs the LSP vector to the LSP vector quantization unit 403.
- the LSP vector quantization unit 403 performs quantization on the LSP vector input from the LSP analysis unit 402.
- the LSP vector quantization unit 403 outputs the obtained quantized LSP vector as a filter coefficient to the synthesis filter 404, and outputs the quantized LSP code (L) to the multiplexing unit 414.
- LSP vector quantization section 403 LSP vector quantization apparatus 100 according to the present embodiment is applied. That is, the specific configuration and operation of LSP vector quantization section 403 are the same as those of LSP vector quantization apparatus 100.
- the wideband LSP vector input to the LSP vector quantization apparatus 100 corresponds to the LSP vector input to the LSP vector quantization unit 403.
- the encoded data output from the LSP vector quantization apparatus 100 corresponds to the quantized LSP code (L) output from the LSP vector quantization unit 403.
- the filter coefficient input to the synthesis filter 404 is a quantized LSP vector obtained by inverse quantization using a quantized LSP code (L) in the LSP vector quantizing unit 403.
- the narrowband LSP vector input to the LSP vector quantization apparatus 100 is input from the outside of the CELP encoding apparatus 400, for example.
- the LSP vector quantization apparatus 100 when the LSP vector quantization apparatus 100 is applied to a scalable encoding apparatus (not shown) having a wideband CELP encoding section (corresponding to the CELP encoding apparatus 400) and a narrowband CELP encoding section, A narrowband LSP vector output from the narrowband CELP encoding unit is input to the LSP vector quantization apparatus 100.
- the synthesis filter 404 uses the filter coefficient based on the quantized LSP vector input from the LSP vector quantization unit 403 to perform synthesis processing on a driving sound source input from an adder 411 described later, and generates the generated synthesis.
- the signal is output to the adder 405.
- the adder 405 calculates the error signal by inverting the polarity of the combined signal input from the combining filter 404 and adding the signal to the signal Xin input from the preprocessing unit 401, and outputs the error signal to the auditory weighting unit 412. To do.
- the adaptive excitation codebook 406 stores in the buffer the driving excitation input from the adder 411 in the past, and one frame from the cut-out position specified by the adaptive excitation lag code (A) input from the parameter determination unit 413. Min samples are extracted from the buffer and output to the multiplier 409 as adaptive sound source vectors.
- adaptive excitation codebook 406 updates the contents of the buffer each time a driving excitation is input from adder 411.
- the quantization gain generation unit 407 determines the quantization adaptive excitation gain and the quantization fixed excitation gain based on the quantized excitation gain code (G) input from the parameter determination unit 413, and multiplies the multiplier 409 and the multiplier respectively. 410.
- Fixed excitation codebook 408 outputs a vector having a shape specified by fixed excitation vector code (F) input from parameter determination section 413 to multiplier 410 as a fixed excitation vector.
- Multiplier 409 multiplies the adaptive excitation vector input from adaptive excitation codebook 406 by the quantized adaptive excitation gain input from quantization gain generation section 407 and outputs the result to adder 411.
- Multiplier 410 multiplies the quantized fixed excitation gain input from quantization gain generating section 407 by the fixed excitation vector input from fixed excitation codebook 408 and outputs the result to adder 411.
- the adder 411 adds the adaptive excitation vector after gain multiplication input from the multiplier 409 and the fixed excitation vector after gain multiplication input from the multiplier 410, and uses the addition result as a driving sound source for the synthesis filter 404 and Output to adaptive excitation codebook 406.
- the driving excitation input to adaptive excitation codebook 406 is stored in the buffer of adaptive excitation codebook 406.
- the auditory weighting unit 412 performs auditory weighting processing on the error signal input from the adder 405 and outputs the result to the parameter determining unit 413 as coding distortion.
- the parameter determination unit 413 selects an adaptive excitation lag that minimizes the coding distortion input from the auditory weighting unit 412 from the adaptive excitation codebook 406, and selects an adaptive excitation lag code (A) indicating the selection result. It outputs to 406 and the multiplexing part 414.
- the adaptive sound source lag is a parameter indicating the position where the adaptive sound source vector is cut out.
- the parameter determination unit 413 selects a fixed excitation vector that minimizes the coding distortion output from the perceptual weighting unit 412 from the fixed excitation codebook 408, and selects a fixed excitation vector code (F) indicating the selection result as the fixed excitation.
- the data is output to the code book 408 and the multiplexing unit 414.
- the parameter determination unit 413 selects the quantization adaptive excitation gain and the quantization fixed excitation gain that minimize the coding distortion output from the auditory weighting unit 412 from the quantization gain generation unit 407, and shows the selection result.
- the quantized excitation gain code (G) is output to the quantization gain generation unit 407 and the multiplexing unit 414.
- the multiplexing unit 414 is a quantized LSP code (L) input from the LSP vector quantization unit 403, an adaptive excitation lag code (A) input from the parameter determination unit 413, a fixed excitation vector code (F), and a quantum
- the encoded excitation gain code (G) is multiplexed and encoded information is output.
- FIG. 6 is a block diagram showing a main configuration of CELP decoding apparatus 450 including LSP vector inverse quantization apparatus 200 according to the present embodiment.
- the separation unit 451 performs separation processing on the encoded information transmitted from the CELP encoding apparatus 400, and performs quantization LSP code (L), adaptive excitation lag code (A), and quantization excitation gain code.
- L quantization LSP code
- A adaptive excitation lag code
- G quantization excitation gain code
- Separation section 451 outputs the quantized LSP code (L) to LSP vector inverse quantization section 452, outputs the adaptive excitation lag code (A) to adaptive excitation codebook 453, and outputs the quantized excitation gain code (G). It outputs to quantization gain generation section 454 and outputs fixed excitation vector code (F) to fixed excitation codebook 455.
- the LSP vector inverse quantization unit 452 decodes the quantized LSP vector from the quantized LSP code (L) input from the separating unit 451, and outputs the quantized LSP vector as a filter coefficient to the synthesis filter 459.
- LSP vector dequantization section 452 LSP vector dequantization apparatus 200 according to the present embodiment is applied. That is, the specific configuration and operation of the LSP vector inverse quantization unit 452 are the same as those of the LSP vector inverse quantization apparatus 200. In this case, the encoded data input to the LSP vector inverse quantization apparatus 200 and the quantized LSP code (L) input to the LSP vector inverse quantization unit 452 correspond to each other.
- the quantized broadband LSP vector output from the LSP vector inverse quantization apparatus 200 corresponds to the quantized LSP vector output from the LSP vector inverse quantization unit 452.
- the narrowband LSP vector input to the LSP vector inverse quantization apparatus 200 is input from the outside of the CELP decoding apparatus 450, for example.
- the LSP vector inverse quantization device 200 is applied to a scalable decoding device (not shown) having a wideband CELP decoding unit (corresponding to the CELP decoding device 450) and a narrowband CELP decoding unit
- the narrowband CELP The narrowband LSP vector output from the decoding unit is input to the LSP vector inverse quantization apparatus 200.
- the adaptive excitation codebook 453 extracts a sample for one frame from the extraction position specified by the adaptive excitation lag code (A) input from the separation unit 451 from the buffer, and uses the extracted vector as an adaptive excitation vector to the multiplier 456. Output.
- adaptive excitation codebook 453 updates the contents of the buffer each time a driving excitation is input from adder 458.
- the quantization gain generation unit 454 decodes the quantization adaptive excitation gain and the quantization fixed excitation gain indicated by the quantization excitation gain code (G) input from the separation unit 451, and multiplies the quantization adaptive excitation gain by the multiplier 456.
- the quantized fixed sound source gain is output to the multiplier 457.
- the fixed excitation codebook 455 generates a fixed excitation vector indicated by the fixed excitation vector code (F) input from the separation unit 451 and outputs it to the multiplier 457.
- Multiplier 456 multiplies the adaptive excitation vector input from adaptive excitation codebook 453 by the quantized adaptive excitation gain input from quantization gain generating section 454 and outputs the result to adder 458.
- Multiplier 457 multiplies the fixed excitation vector input from fixed excitation codebook 455 by the quantized fixed excitation gain input from quantization gain generation section 454 and outputs the result to adder 458.
- the adder 458 adds the adaptive excitation vector after gain multiplication input from the multiplier 456 and the fixed excitation vector after gain multiplication input from the multiplier 457 to generate a driving excitation, and the generated driving
- the sound source is output to synthesis filter 459 and adaptive excitation codebook 453.
- the driving excitation input to adaptive excitation codebook 453 is stored in the buffer of adaptive excitation codebook 453.
- the synthesis filter 459 performs synthesis processing using the driving sound source input from the adder 458 and the filter coefficient decoded by the LSP vector inverse quantization unit 452, and outputs the generated synthesized signal to the post-processing unit 460. To do.
- the post-processing unit 460 performs processing for improving the subjective quality of speech, such as formant enhancement and pitch enhancement, and processing for improving the subjective quality of stationary noise, with respect to the synthesized signal input from the synthesis filter 459.
- the obtained audio signal or musical sound signal is output.
- the vector quantization accuracy / vector inverse quantization device is used to improve the vector quantization accuracy during encoding. Since it becomes possible to improve, the audio
- CELP decoding apparatus 450 decodes encoded data output from CELP encoding apparatus 400 has been described.
- the present invention is not limited to this, and it goes without saying that any encoded data in a format that can be decoded by the CELP decoding device 450 can be received and decoded by the CELP decoding device.
- vector quantization apparatus the vector inverse quantization apparatus, and these methods according to the present invention are not limited to the above-described embodiments, and can be implemented with various modifications.
- the vector quantization device the vector inverse quantization device, and these methods have been described for a speech signal or a musical sound signal, but may be applied to other possible signals.
- the LSP is sometimes called LSF (Line Spectral Frequency), and the LSP may be read as LSF.
- LSF Line Spectral Frequency
- the present embodiment can be used as an ISP quantizing / inverse quantizing device by replacing LSP with ISP.
- ISF InterferenceittSpectrum Frequency
- the present embodiment can be used as an ISF quantization / inverse quantization apparatus by replacing LSP with ISF.
- the vector quantization apparatus and vector inverse quantization apparatus can be mounted on a communication terminal apparatus or base station apparatus in a mobile communication system that transmits voice, music, or the like. Thereby, it is possible to provide a communication terminal apparatus and a base station apparatus having the same effects as described above.
- the present invention can also be realized by software.
- the vector quantization method and the vector inverse quantization method algorithm according to the present invention are described in a programming language, and the program is stored in a memory and executed by an information processing means, whereby the vector quantization method according to the present invention is performed. Functions similar to those of the quantization device and the vector inverse quantization device can be realized.
- each functional block used in the description of each of the above embodiments is typically realized as an LSI which is an integrated circuit. These may be individually made into one chip, or may be made into one chip so as to include a part or all of them.
- LSI LSI
- IC system LSI
- super LSI ultra LSI
- the method of circuit integration is not limited to LSI, and implementation with a dedicated circuit or a general-purpose processor is also possible.
- An FPGA Field Programmable Gate Array
- a reconfigurable processor that can reconfigure the connection or setting of circuit cells inside the LSI may be used.
- the vector quantization apparatus, the vector inverse quantization apparatus, and these methods according to the present invention can be applied to applications such as speech encoding and speech decoding.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Theoretical Computer Science (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Description
Claims (9)
- 複数の分類用コードベクトルの中から、量子化対象ベクトルとの相関を有する特徴の種類を示す分類用コードベクトルを選択する第1選択手段と、
複数の第1コードブックの中から、選択された前記分類用コードベクトルに対応する第1コードブックを選択する第2選択手段と、
選択された前記第1コードブックを構成する複数の第1コードベクトルを用いて前記量子化対象ベクトルを量子化し、第1符号を得る第1量子化手段と、
複数の行列の中から、選択された前記分類用コードベクトルに対応する第1行列を選択する第3選択手段と、
複数の第2コードベクトルおよび選択された前記第1行列を用い、前記第1符号が示す前記第1コードベクトルと、前記量子化対象ベクトルとの差である第1残差ベクトルを量子化し、第2符号を得る第2量子化手段と、
を具備するベクトル量子化装置。 - 前記第2量子化手段は、
前記複数の第2コードベクトルそれぞれに対して、選択された前記第1行列を用いた行列演算を行い、前記行列演算後の前記複数の第2コードベクトルを用いて前記第1残差ベクトルを量子化する、
請求項1記載のベクトル量子化装置。 - 前記第2量子化手段は、
前記第1残差ベクトルに対して、選択された前記第1行列の逆行列を用いた行列演算を行い、前記複数の第2コードベクトルを用いて前記行列演算後の前記第1残差ベクトルを量子化する、
請求項1記載のベクトル量子化装置。 - 前記第1行列は回転行列である、
請求項1記載のベクトル量子化装置。 - 前記選択手段は、さらに、前記複数の行列の中から、選択された前記分類用コードベクトルに対応する第2行列を選択し、
複数の第3コードベクトルおよび選択された前記第2行列を用い、前記第2符号が示す前記第1残差ベクトルと、前記第2コードベクトルとの差である第2残差ベクトルを量子化し、第3符号を得る第3量子化手段、をさらに具備する、
請求項1記載のベクトル量子化装置。 - 前記第2行列は回転行列である、
請求項5記載のベクトル量子化装置。 - ベクトル量子化装置において量子化対象ベクトルを量子化して得られた第1符号と、前記量子化の量子化誤差をさらに量子化して得られた第2符号と、を受信する受信手段と、
複数の分類用コードベクトルの中から、前記量子化対象ベクトルとの相関を有する特徴の種類を示す分類用コードベクトルを選択する第1選択手段と、
複数の第1コードブックの中から、選択された前記分類用コードベクトルに対応する第1コードブックを選択する第2選択手段と、
選択された前記第1コードブックを構成する複数の第1コードベクトルの中から、前記第1符号に対応する第1コードベクトルを指定する第1逆量子化手段と、
複数の行列の中から、選択された前記分類用コードベクトルに対応する行列を選択する第3選択手段と、
複数の第2コードベクトルの中から前記第2符号に対応する第2コードベクトルを指定し、指定された前記第2コードベクトルと、選択された前記行列と、指定された前記第1コードベクトルとを用い、量子化ベクトルを得る第2逆量子化手段と、
を具備するベクトル逆量子化装置。 - 複数の分類用コードベクトルの中から、量子化対象ベクトルとの相関を有する特徴の種類を示す分類用コードベクトルを選択するステップと、
複数の第1コードブックの中から、選択された前記分類用コードベクトルに対応する第1コードブックを選択するステップと、
選択された前記第1コードブックを構成する複数の第1コードベクトルを用いて前記量子化対象ベクトルを量子化し、第1符号を得るステップと、
複数の行列の中から、選択された前記分類用コードベクトルに対応する第1行列を選択するステップと、
複数の第2コードベクトルおよび選択された前記第1行列を用い、前記第1符号が示す前記第1コードベクトルと、前記量子化対象ベクトルとの差である第1残差ベクトルを量子化し、第2符号を得るステップと、
を有するベクトル量子化方法。 - ベクトル量子化装置において量子化対象ベクトルを量子化して得られた第1符号と、前記量子化の量子化誤差をさらに量子化して得られた第2符号と、を受信するステップと、
複数の分類用コードベクトルの中から、前記量子化対象ベクトルとの相関を有する特徴の種類を示す分類用コードベクトルを選択するステップと、
複数の第1コードブックの中から、選択された前記分類用コードベクトルに対応する第1コードブックを選択するステップと、
選択された前記第1コードブックを構成する複数の第1コードベクトルの中から、前記第1符号に対応する第1コードベクトルを指定するステップと、
複数の行列の中から、選択された前記分類用コードベクトルに対応する行列を選択するステップと、
複数の第2コードベクトルの中から前記第2符号に対応する第2コードベクトルを指定し、指定された前記第2コードベクトルと、選択された前記行列と、指定された前記第1コードベクトルとを用い、量子化ベクトルを得るステップと、
を有するベクトル逆量子化方法。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP10741107.6A EP2398149B1 (en) | 2009-02-13 | 2010-02-12 | Vector quantization device, vector inverse-quantization device, and associated methods |
US13/148,892 US8493244B2 (en) | 2009-02-13 | 2010-02-12 | Vector quantization device, vector inverse-quantization device, and methods of same |
JP2010550471A JP5335004B2 (ja) | 2009-02-13 | 2010-02-12 | ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 |
RU2011134054/08A RU2519027C2 (ru) | 2009-02-13 | 2010-02-12 | Устройство векторного квантования, устройство векторного обратного квантования и способы для этого |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2009031651 | 2009-02-13 | ||
JP2009-031651 | 2009-02-13 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010092827A1 true WO2010092827A1 (ja) | 2010-08-19 |
Family
ID=42561675
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2010/000881 WO2010092827A1 (ja) | 2009-02-13 | 2010-02-12 | ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 |
Country Status (5)
Country | Link |
---|---|
US (1) | US8493244B2 (ja) |
EP (1) | EP2398149B1 (ja) |
JP (1) | JP5335004B2 (ja) |
RU (1) | RU2519027C2 (ja) |
WO (1) | WO2010092827A1 (ja) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
RU2541168C2 (ru) * | 2010-09-02 | 2015-02-10 | Майкрософт Корпорейшн | Формирование и применение кодовой подкниги кодовой книги кодирования с контролем ошибок |
JP2017509926A (ja) * | 2014-03-28 | 2017-04-06 | サムスン エレクトロニクス カンパニー リミテッド | 線形予測係数量子化方法及びその装置、並びに線形予測係数逆量子化方法及びその装置 |
US10504532B2 (en) | 2014-05-07 | 2019-12-10 | Samsung Electronics Co., Ltd. | Method and device for quantizing linear predictive coefficient, and method and device for dequantizing same |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9236063B2 (en) | 2010-07-30 | 2016-01-12 | Qualcomm Incorporated | Systems, methods, apparatus, and computer-readable media for dynamic bit allocation |
US9208792B2 (en) | 2010-08-17 | 2015-12-08 | Qualcomm Incorporated | Systems, methods, apparatus, and computer-readable media for noise injection |
EP3547261B1 (en) * | 2012-03-29 | 2023-08-09 | Telefonaktiebolaget LM Ericsson (publ) | Vector quantizer |
KR101821532B1 (ko) | 2012-07-12 | 2018-03-08 | 노키아 테크놀로지스 오와이 | 벡터 양자화 |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH04351018A (ja) * | 1991-05-28 | 1992-12-04 | Fujitsu Ltd | ゲイン・シェープベクトル量子化方式 |
JPH08275163A (ja) * | 1995-03-31 | 1996-10-18 | Canon Inc | 画像処理装置およびその方法 |
JPH09120300A (ja) * | 1995-10-25 | 1997-05-06 | Fujitsu Ltd | ベクトル量子化装置 |
JPH09258794A (ja) * | 1996-03-18 | 1997-10-03 | Fujitsu Ltd | ベクトル量子化装置 |
JP2956473B2 (ja) * | 1994-04-21 | 1999-10-04 | 日本電気株式会社 | ベクトル量子化装置 |
JP3257386B2 (ja) * | 1996-02-01 | 2002-02-18 | 松下電器産業株式会社 | ベクトル量子化方法 |
WO2006030865A1 (ja) | 2004-09-17 | 2006-03-23 | Matsushita Electric Industrial Co., Ltd. | スケーラブル符号化装置、スケーラブル復号化装置、スケーラブル符号化方法、スケーラブル復号化方法、通信端末装置および基地局装置 |
WO2006062202A1 (ja) * | 2004-12-10 | 2006-06-15 | Matsushita Electric Industrial Co., Ltd. | 広帯域符号化装置、広帯域lsp予測装置、帯域スケーラブル符号化装置及び広帯域符号化方法 |
WO2008047795A1 (fr) * | 2006-10-17 | 2008-04-24 | Panasonic Corporation | Dispositif de quantification vectorielle, dispositif de quantification vectorielle inverse et procédé associé |
JP2009031651A (ja) | 2007-07-30 | 2009-02-12 | Canon Inc | 画像形成装置、その制御方法、及びプログラム |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH03257386A (ja) | 1990-03-07 | 1991-11-15 | Toshiba Corp | 発電機脱調予測の誤判定防止方法 |
US5765127A (en) * | 1992-03-18 | 1998-06-09 | Sony Corp | High efficiency encoding method |
JP3224955B2 (ja) * | 1994-05-27 | 2001-11-05 | 株式会社東芝 | ベクトル量子化装置およびベクトル量子化方法 |
DE69628083T2 (de) | 1995-03-31 | 2003-12-04 | Canon K.K., Tokio/Tokyo | Bildverarbeitungsgerät und Methode |
JPH09311732A (ja) * | 1996-05-23 | 1997-12-02 | Olympus Optical Co Ltd | 光学的手法を用いたベクトル照合装置とそれを用いたベクトル量子化演算装置 |
WO1999016050A1 (en) | 1997-09-23 | 1999-04-01 | Voxware, Inc. | Scalable and embedded codec for speech and audio signals |
US6067515A (en) * | 1997-10-27 | 2000-05-23 | Advanced Micro Devices, Inc. | Split matrix quantization with split vector quantization error compensation and selective enhanced processing for robust speech recognition |
CA2429832C (en) * | 2000-11-30 | 2011-05-17 | Matsushita Electric Industrial Co., Ltd. | Lpc vector quantization apparatus |
US7003454B2 (en) * | 2001-05-16 | 2006-02-21 | Nokia Corporation | Method and system for line spectral frequency vector quantization in speech codec |
CA2415105A1 (en) * | 2002-12-24 | 2004-06-24 | Voiceage Corporation | A method and device for robust predictive vector quantization of linear prediction parameters in variable bit rate speech coding |
-
2010
- 2010-02-12 RU RU2011134054/08A patent/RU2519027C2/ru not_active IP Right Cessation
- 2010-02-12 WO PCT/JP2010/000881 patent/WO2010092827A1/ja active Application Filing
- 2010-02-12 US US13/148,892 patent/US8493244B2/en active Active
- 2010-02-12 JP JP2010550471A patent/JP5335004B2/ja not_active Expired - Fee Related
- 2010-02-12 EP EP10741107.6A patent/EP2398149B1/en not_active Not-in-force
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH04351018A (ja) * | 1991-05-28 | 1992-12-04 | Fujitsu Ltd | ゲイン・シェープベクトル量子化方式 |
JP2956473B2 (ja) * | 1994-04-21 | 1999-10-04 | 日本電気株式会社 | ベクトル量子化装置 |
JPH08275163A (ja) * | 1995-03-31 | 1996-10-18 | Canon Inc | 画像処理装置およびその方法 |
JPH09120300A (ja) * | 1995-10-25 | 1997-05-06 | Fujitsu Ltd | ベクトル量子化装置 |
JP3257386B2 (ja) * | 1996-02-01 | 2002-02-18 | 松下電器産業株式会社 | ベクトル量子化方法 |
JPH09258794A (ja) * | 1996-03-18 | 1997-10-03 | Fujitsu Ltd | ベクトル量子化装置 |
WO2006030865A1 (ja) | 2004-09-17 | 2006-03-23 | Matsushita Electric Industrial Co., Ltd. | スケーラブル符号化装置、スケーラブル復号化装置、スケーラブル符号化方法、スケーラブル復号化方法、通信端末装置および基地局装置 |
WO2006062202A1 (ja) * | 2004-12-10 | 2006-06-15 | Matsushita Electric Industrial Co., Ltd. | 広帯域符号化装置、広帯域lsp予測装置、帯域スケーラブル符号化装置及び広帯域符号化方法 |
WO2008047795A1 (fr) * | 2006-10-17 | 2008-04-24 | Panasonic Corporation | Dispositif de quantification vectorielle, dispositif de quantification vectorielle inverse et procédé associé |
JP2009031651A (ja) | 2007-07-30 | 2009-02-12 | Canon Inc | 画像形成装置、その制御方法、及びプログラム |
Non-Patent Citations (3)
Title |
---|
ALLEN GERSHO, ROBERT M. GRAY: "Vector quantization and information compression", 10 November 1998, CORONA PUBLISHING CO.,LTD. CORONA PUBLISHING CO.,LTD., pages: 491 - 493 |
ALLEN GERSHO, ROBERT M. GRAY: "Vector quantization and information compression", 10 November 1998, CORONA PUBLISHING CO.,LTD., pages: 506,524 - 531 |
See also references of EP2398149A4 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
RU2541168C2 (ru) * | 2010-09-02 | 2015-02-10 | Майкрософт Корпорейшн | Формирование и применение кодовой подкниги кодовой книги кодирования с контролем ошибок |
US9363043B2 (en) | 2010-09-02 | 2016-06-07 | Microsoft Technology Licensing, Llc | Generation and application of a sub-codebook of an error control coding codebook |
RU2668988C2 (ru) * | 2010-09-02 | 2018-10-05 | Майкрософт Корпорейшн | Формирование и применение кодовой подкниги кодовой книги кодирования с контролем ошибок |
JP2017509926A (ja) * | 2014-03-28 | 2017-04-06 | サムスン エレクトロニクス カンパニー リミテッド | 線形予測係数量子化方法及びその装置、並びに線形予測係数逆量子化方法及びその装置 |
US10515646B2 (en) | 2014-03-28 | 2019-12-24 | Samsung Electronics Co., Ltd. | Method and device for quantization of linear prediction coefficient and method and device for inverse quantization |
US11450329B2 (en) | 2014-03-28 | 2022-09-20 | Samsung Electronics Co., Ltd. | Method and device for quantization of linear prediction coefficient and method and device for inverse quantization |
US10504532B2 (en) | 2014-05-07 | 2019-12-10 | Samsung Electronics Co., Ltd. | Method and device for quantizing linear predictive coefficient, and method and device for dequantizing same |
US11238878B2 (en) | 2014-05-07 | 2022-02-01 | Samsung Electronics Co., Ltd. | Method and device for quantizing linear predictive coefficient, and method and device for dequantizing same |
US11922960B2 (en) | 2014-05-07 | 2024-03-05 | Samsung Electronics Co., Ltd. | Method and device for quantizing linear predictive coefficient, and method and device for dequantizing same |
Also Published As
Publication number | Publication date |
---|---|
EP2398149A1 (en) | 2011-12-21 |
US8493244B2 (en) | 2013-07-23 |
JPWO2010092827A1 (ja) | 2012-08-16 |
EP2398149A4 (en) | 2012-11-28 |
RU2519027C2 (ru) | 2014-06-10 |
RU2011134054A (ru) | 2013-03-20 |
US20110316732A1 (en) | 2011-12-29 |
JP5335004B2 (ja) | 2013-11-06 |
EP2398149B1 (en) | 2014-05-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5419714B2 (ja) | ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 | |
JP5300733B2 (ja) | ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 | |
JP5335004B2 (ja) | ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 | |
JPWO2008047795A1 (ja) | ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 | |
JP5190445B2 (ja) | 符号化装置および符号化方法 | |
WO2009090875A1 (ja) | ベクトル量子化装置、ベクトル逆量子化装置、およびこれらの方法 | |
JP5687706B2 (ja) | 量子化装置及び量子化方法 | |
JP6195138B2 (ja) | 音声符号化装置及び音声符号化方法 | |
JPH0764599A (ja) | 線スペクトル対パラメータのベクトル量子化方法とクラスタリング方法および音声符号化方法並びにそれらの装置 | |
JP2003345392A (ja) | 分割型スケーリング因子を用いたスペクトル包絡パラメータのベクトル量子化器 | |
WO2012053149A1 (ja) | 音声分析装置、量子化装置、逆量子化装置、及びこれらの方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 10741107 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2010550471 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010741107 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13148892 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2011134054 Country of ref document: RU |