EP0718822A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
A low rate multi-mode CELP CODEC that uses backward prediction
Assignee: The AT&T MVPD Group LLC
•
Filing: 1995-12-18
•
Granted: —
•
Expires: 2015-12-18
Abstract
Abstract The present invention provides a multi-mode CELP encoding and decoding method and device for digitized speech signals providing improvements over prior art codecs and coding methods by selectively utilizes backward prediction for the short-term predictor parameters and fixed codebook gain of a speech signal. In order to achieve these improvements, the present invention provides a coding method comprising the steps of classifying a segment of the digitized speech signal as one of a plurality of predetermined modes, determining a set of unquantized line spectral frequencies to represent the short term predictor parameters for that segment, and quantizing the determined set of unquantized line spectral frequencies using a mode-specific combination of scalar quantization and vector quantization, which utilizes backward prediction for modes with voiced speech signals. Furthermore, backward prediction is selectively applied to the fixed codebook gain in the modes that are free of transients so that it may be used in the fixed codebook search and fixed codebook gain quantization in those modes.
Inventors
- 👤 Kumar Swaminathan
- 👤 Murthy Vemuganti
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-03 22:06
CPC Classifications
G; G10; G10L; G10L19/00; G10L19/04; G10L19/06; G10L19/16; G10L19/18; G10L2019/0001; G10L2019/0013; G10L25/00; G10L25/03; G10L25/24; G10L25/93
Claims Analysis
Total Claims: 11
Claim 1
Independent Claim
121 words
A method of coding a digitized speech signal comprising the steps of: analyzing the digitized speech signal in discrete segments; classifying a segment of the digitized speech signal in one of a plurality of predetermined modes (2); determining a set of unquantized line spectral frequencies to represent the short term predictor parameters for the digitized speech signal segment; and quantizing the determined set of unquantized line spectral frequencies, wherein the quantization step further comprises the steps of scalar quantizing a first subset of the unquantized line spectral frequencies and vector quantizing a second subset of the unquantized line spectral frequencies, wherein the scalar quantizing and vector quantizing steps depend on the mode in which the digitized speech signal segment is classified.
Claim 2
Dependent Claim
44 words
The coding method according to claim 1 further comprising the step of providing at least one set of scalar quantizers, for example, scalar quantizers trained on IRS-filtered speech or unfiltered flat speech, to scalar quantize the first subset of the unquantized line spectral frequencies.
Claim 3
Dependent Claim
106 words
The coding method according to claim 2, wherein the plurality of predetermined modes includes a voiced mode for classifying digitized speech signal segments containing voiced speech, and wherein, for each set of scalar quantizers, the scalar quantizing step for each of the first subset of unquantized line spectral frequencies for a digitized speech signal segment classified in the voiced mode further comprises the steps of: predicting each line spectral frequency as a weighted sum of neighboring line spectral frequencies scalar quantized for a preceding digitized speech signal segment (41); and encoding the unquantized line spectral frequency as an offset from its quantized predicted line spectral frequency.
Claim 4
Dependent Claim
135 words
The coding method according to claim 1, 2 or 3 wherein the plurality of predetermined modes includes a voiced mode for classifying digitized speech signal segments containing voiced speech, and wherein, for each vector quantization table, the vector quantizing step for a digitized speech signal segment classified in the voiced mode further comprises the steps of: determining a range of indices representing vectors in the vector quantization table for vector quantizing a second subset of unquantized line spectral frequencies, depending on line spectral frequencies vector quantized for a preceding digitized speech signal segment (53); selecting a vector having an index in the determined range of indices for vector quantizing the second subset of the determined set of unquantized line spectral frequencies; and encoding the selected vector as an offset within the determined range of indices.
Claim 5
Dependent Claim
277 words
The coding method according to claim 1, 2, 3 or 4 wherein the plurality of predetermined modes includes a non-voiced mode for classifying digitized speech signals not primarily containing voiced speech, and wherein the coding method further comprises the steps of: providing a first set of scalar quantizers trained on filtered speech and a second set of scalar quantizers trained on unfiltered flat speech to scalar quantize a first subset of the unquantized line spectral frequencies; providing a first vector quantization table trained on filtered speech and a second vector quantization table trained on unfiltered flat speech, wherein each vector quantization table has entries of vectors for vector quantizing the second subset of the unquantized line spectral frequencies; determining a first set of quantized line spectral frequencies by scalar quantizing the first subset of unquantized line spectral frequencies with the first set of scalar quantizers and vector quantizing the second subset of unquantized line spectral frequencies with the first vector quantization table; determining a second set of quantized line spectral frequencies by scalar quantizing the first subset of unquantized line spectral frequencies with the second set of scalar quantizers and vector quantizing the second subset of unquantized line spectral frequencies with the second vector quantization table; measuring the cepstral distortion between the first set of quantized line spectral frequencies and the determined set of unquantized line spectral frequencies (30), and between the second set of quantized line spectral frequencies and the determined set of unquantized line spectral frequencies (31); selecting the set of quantized line spectral frequencies having the smaller measured cepstral distortion for representing the short term predictor parameters for the digitized speech signal segment (32).
Claim 6
Dependent Claim
90 words
The coding method according to claim 1, 2, 3, 4 or 5 further comprising the step of analyzing at least one of: a spectral stationarity in the digitized speech signal segment (57); a pitch stationarity in the digitized speech signal segment (58); a zero crossing rate in the digitized speech signal segment (59); a short term level gradient in the digitized speech signal segment (60); and a short term energy in the digitized speech signal segment (61); wherein the classifying step depends on the results of the analyzing step (62).
Claim 7
Independent Claim
139 words
A method of decoding a data bitstream containing encoded parameters for a segment of a digitized speech signal comprising the steps of: extracting from the data bitstream: a mode parameter encoding a mode of the digitized speech signal segment, a set of scalar quantizer parameters, and a vector quantizer parameter; classifying the digitized speech signal segment in one of a plurality of predetermined modes based on the extracted mode parameter (6); determining a set of inverse quantized line spectral frequencies for the digitized speech signal segment by determining a first subset of inverse quantized line spectral frequencies based on the extracted set of scalar quantizer parameters, and determining a second subset of inverse quantized line spectral frequencies based on the extracted vector quantizer parameter, wherein the determining steps depend on the classified mode of the digitized speech signal segment.
Claim 8
Independent Claim
195 words
A method of encoding the excitation of a digitized speech signal, comprising the steps of: partitioning the digitized speech signal into discrete segments; classifying a segment of the digitized speech signal in one of a plurality of predetermined modes, wherein the plurality of predetermined modes includes at least one non-transient mode for classifying a digitized speech signal segment not containing transients (2); further partitioning the digitized speech signal segment into a plurality of subframes for analyzing the excitation of the digitized speech signal segment; providing an adaptive codebook for deriving an adaptive codebook vector and an adaptive codebook gain; providing a fixed codebook for deriving a fixed codebook vector and fixed codebook gain, wherein the fixed codebook comprises at least one section; and modeling the excitation of each digitized speech signal subframe as a vector sum of an adaptive codebook vector scaled by an adaptive codebook gain, both derived from the adaptive codebook, and a fixed codebook vector scaled by a fixed codebook gain, both derived from the fixed codebook, and wherein, for a digitized speech signal segment classified in any non-transient mode, the step of deriving the fixed codebook gain comprises backward predictive analysis.
Claim 9
Dependent Claim
35 words
The excitation encoding method according to claim 8, wherein the number of subframes into which the digitized speech signal segment is partitioned depends on the mode in which the digitized speech signal segment is classified.
Claim 10
Dependent Claim
90 words
The excitation encoding method according to claim 8, wherein the step of deriving the fixed codebook gain by backward predictive analysis further comprises the steps of: determining an actual fixed codebook gain; predicting an rms value for the fixed codebook vector based on at least one rms value of a fixed codebook vector modeling an excitation of a preceding subframe; predicting the fixed codebook gain based on the predicted rms value (69); and quantizing the determined actual fixed codebook gain as an offset from the predicted fixed codebook gain (70).
Claim 11
Independent Claim
169 words
A method of decoding a data bitstream containing encoded parameters for an excitation of a segment of a digitized speech signal comprising the steps of: providing an adaptive codebook and a fixed codebook; extracting from the data bitstream: a mode parameter encoding a mode of the digitized speech signal segment, an adaptive codebook parameter, and a fixed codebook parameter; classifying the digitized speech signal segment in one of a plurality of predetermined modes based on the extracted mode parameter (6); modeling the excitation of each of a plurality of subframes for the digitized speech signal segment as a vector sum of an adaptive codebook vector scaled by an adaptive codebook gain, both derived from the adaptive codebook and the adaptive codebook parameter, and a fixed codebook vector scaled by a fixed codebook gain, both derived from the fixed codebook and the fixed codebook parameter, and wherein, for a digitized speech signal segment classified in any non-transient mode, the step of deriving the fixed codebook gain comprises backward predictive analysis.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.