Abstract: A portion of an audio signal is separated into multiple frames from which one or more different features are extracted. These different features are used, in combination with a set of rules, to classify the portion of the audio signal into one of multiple different classifications (for example, speech, non-speech, music, environment sound, silence, etc.). In one embodiment, these different features include one or more of line spectrum pairs (LSPs), a noise frame ratio, periodicity of particular bands, spectrum flux features, and energy distribution in one or more of the bands. The line spectrum pairs are also optionally used to segment the audio signal, identifying audio classification changes as well as speaker changes when the audio signal is speech.

Abstract: According to the invention, quantization encoding is conducted using the probability density function of the source, enabling fixed, variable and adaptive rate encoding. To achieve adaptive encoding, an update is conducted with a new observation of the data source, preferably with each new observation of the data source, preferably with each new observation of the data source. The current probability density function of the source is then estimated to produce codepoints to vector quantize the observation of the data source.

Abstract: A system for performing a computationally efficient method of searching through N Vector Quantization (VQ) codevectors for a preferred one of the N VQ codevectors predicts a speech signal to derive a residual signal, derives a ZERO-INPUT response error vector common to each of the N VQ codevectors, derives N ZERO-STATE response error vectors each based on a corresponding one of the N VQ codevectors, and selects the preferred one of the N VQ codevectors based on the N ZERO-STATE response error vectors and the ZERO-INPUT response error vector.

Abstract: In a Noise Feedback Coding (NFC) system operable in a ZERO-STATE condition and a ZERO-INPUT condition, the NFC system including at least one filter having a filter memory, a method of updating the filter memory. The method comprises: (a) producing a ZERO-STATE contribution to the filter memory when the NFC system is in the ZERO-STATE condition; (b) producing a ZERO-INPUT contribution to the filter memory when the NFC system is in the ZERO-INPUT condition; and (c) updating the filter memory as a function of both the ZERO-STATE contribution and the ZERO-INPUT contribution.

Abstract: This disclosure describes a flexible digital transmission system that improves upon the ATSC A/53 HDTV signal transmission standard. The system includes a digital television signal transmitter for generating a first Advanced Television Systems Committee (ATSC) standard 8-VSB bit stream and, for generating an encoded new bit stream capable of transmitting high priority information bits, wherein symbols of the new bit stream are capable of being transmitted according to a transmission mode selected from group comprising: a 2-VSB mode, a 4-VSB mode, and a hierarchical-VSB (H-VSB) transmission mode. Each respective 2-VSB, 4-VSB, and H-VSB mode is characterized as having symbols mapped according to possible symbol values from an alphabet comprising respectively, {?7, ?5, 5, 7}, {7, 3, ?3, ?7}, and {7, 5, 3, ?3, ?5, ?7}. The standard 8-VSB bit stream and new bit stream may be simultaneously transmitted over a terrestrial channel according to a broadcaster defined bit-rate ratio.

Abstract: A digital signal of which input data has been segmented as block each having a predetermined data amount and highly efficiently encoded along with an adjacent block is decoded, edited, and then highly efficiently encoded. A delay that takes place in such signal processes is compensated. Thus, part of a digital signal that has been highly efficiently encoded digital signal can be edited.

Abstract: A method of coding an audio signal comprises receiving an audio signal x to be coded and transforming the received signal from the time to the frequency domain. A quantised audio signal {tilde over (x)} is generated from the transformed audio signal x together with a set of long-term prediction coefficients A which can be used to predict a current time frame of the received audio signal directly from one or more previous time frames of the quantised audio signal {tilde over (x)}. A predicted audio signal {circumflex over (x)} is generated using the prediction coefficients A. The predicted audio signal {circumflex over (x)} is then transformed from the time to the frequency domain and the resulting frequency domain signal compared with that of the received audio signal x to generate an error signal E(k) for each of a plurality of frequency sub-bands.

Abstract: A low-bit-rate coding technique for unvoiced segments of speech, without loss of quality compared to the conventional Code Excited Linear Prediction (CELP) method operating at a much higher bit rate. A set of gains are derived from a residual signal after whitening the speech signal by a linear prediction filter. These gains are then quantized and applied to a randomly generated sparse excitation. The excitation is filtered, and its spectral characteristics are analyzed and compared to the spectral characteristics of the original residual signal. Based on this analysis, a filter is chosen to shape the spectral characteristics of the excitation to achieve optimal performance.

Abstract: An Adaptive Sound Source Vector Generator (ASSVG) 103 sets preceding and succeeding pitch cycles centered on an integral-accuracy pitch cycle T0 selected in the previous subframe as a range for searching for a fractional-accuracy pitch frequency, and extracts an adaptive sound source vector P(T-frac) that has fractional-accuracy pitch cycle T-frac within this range from an Adaptive Code Book (ACB) 102. A Last Sub Frame Integral Pitch Cycle Storage (LSFIPCS) 108 stores integral component T0 of the optimal pitch cycle selected by a Distortion Comparator (DC) 107, and when a pitch cycle of the next subframe is searched for, outputs this optimal pitch cycle integral component T0 to the Adaptive Sound Source Vector Generator (ASSVG) 103. An Optimal Pitch Cycle Accuracy Judge Section (OPCAJS) 109 judges whether the optimal pitch cycle is of integral accuracy or fractional accuracy.

Abstract: An encoding device (200) includes an MDCT unit (202) that transforms an input signal in a time domain into a frequency spectrum including a lower frequency spectrum, a BWE encoding unit (204) that generates extension data which specifies a higher frequency spectrum at a higher frequency than the lower frequency spectrum, and an encoded data stream generating unit (205) that encodes to output the lower frequency spectrum obtained by the MDCT unit (202) and the extension data obtained by the BWE encoding unit (204). The BWE encoding unit (204) generates as the extension data (i) a first parameter which specifies a lower subband which is to be copied as the higher frequency spectrum from among a plurality of the lower subbands which form the lower frequency spectrum obtained by the MDCT unit (202) and (ii) a second parameter which specifies a gain of the lower subband after being copied.

Abstract: A method of performing an excitation Vector Quantization (VQ) in a Noise Feedback Coding environment involves reorganizing a calculation of an energy of an error vector for each of a plurality of candidate excitation vectors of a codebook. The energy of the error vector is a cost function that is minimized during a search of the codebook for a best candidate excitation VQ vector. The reorganization includes expanding a Mean Squared Error (MSE) term of the error vector, excluding an energy term that is invariant to the candidate excitation vector, and pre-computing energy terms of ZERO-STATE responses of the candidate excitation vectors that are invariant to sub-vectors of a subframe. Another method searches a signed codebook. Both methods use correlation techniques.

Abstract: A method for searching a codebook which predicts a residual element of an input voice signal includes combining each track of the input signal, forming track units including at least two tracks, and determining a pulse code for each track. The method further includes calculating energy for each track using an energy formula including a vector dot product, arranging or selecting codewords in a small track energy order, and searching or selecting an optimal pulse for a single- or double-pulse track of the selected codeword.

Abstract: Method and apparatus for subsampling phase spectrum information by analyzing and reconstructing a prototype of a frame. The prototype is analyzed by correlating phase parameters generated from the prototype with phase parameters generated from a reference prototype in multiple frequency bands. The prototype is reconstructed using linear phase shift values by producing a set of phase parameters of the reference prototype, generating a set of linear phase shift values associated with the prototype, and composing a phase vector from the set of phase parameters and the set of linear phase shift values across multiple frequency bands. The prototype is reconstructed using circular rotation values by producing a set of circular rotation values associated with the prototype, generating a set of bandpass waveforms associated with the phase parameters of the reference prototype in multiple frequency bands, and modifying the set of bandpass waveforms based upon the circular rotation values.

Abstract: A method for processing speech in a spread spectrum communication system uses CELP speech encoded signals. A speech input receives samples of a speech signal and a codebook analysis block for selects an index of a code from each of a plurality of codebooks. A weighted synthesis filter is used in the generation of a prediction error between a predicted current sample and a current sample of the speech samples. The index is transmitted to the receiver to enable reconstruction of the speech signal at the receiver.

Abstract: An audio coder that improves audio quality by reducing a quantization error. When a code corresponding to a sampled value of an audio signal is determined, a candidate code storage section stores all combinations of candidate codes in a neighborhood interval of the sampled value. A local decoder generates reproduced signals by decoding the codes stored in the candidate code storage section. An error evaluation section calculates, for each candidate code, a sum of squares of differentials between input sampled values and reproduced signals, detects a combination of candidate codes by which a smallest sum is obtained, that is to say, which minimizes a quantization error, and outputs a code included in the detected combination of candidate codes.

Abstract: With respect to each of codes corresponding to code vectors in a code book stored in a code book storage section, an expectation degree storage section stores an expectation degree at which observation is expected when an integrated parameter with respect to a word as a recognition target is inputted. A vector quantization section vector-quantizes the integrated parameter and outputs a series of codes of a code vector which has a shortest distance to the integrated parameter.

Abstract: In coding and decoding an acoustic parameter, a weighted vector is generated by multiplying a code vector output in a past frame and a code vector selected in a present frame by weighting factors respectively selected from a factor code book and adding the products to each other.

Abstract: A speech analyzing stage (12) and a method for analyzing a speech signal is described. The speech analyzing stage (12) is part of an automatic speech recognition system (10) and is adapted for analyzing in the spectral domain a speech signal sampled at one of at least two different system sampling rates. The speech analyzing stage (12) comprises a first spectral analyzer (18a) for analyzing the speech signal up to a first frequency (flowest) which is preferably derived from the lowest system sampling rate (2×flowest) and a second spectral analyzer (18b) for analyzing the speech signal at least above the first frequency (flowest).

Abstract: Sound source separation, without permutation, using convolutional mixing independent component analysis based on a priori knowledge of the target sound source is disclosed. The target sound source can be a human speaker. The reconstruction filters used in the sound source separation take into account the a priori knowledge of the target sound source, such as an estimate the spectra of the target sound source. The filters may be generally constructed based on a speech recognition system. Matching the words of the dictionary of the speech recognition system to a reconstructed signal indicates whether proper separation has occurred. More specifically, the filters may be constructed based on a vector quantization codebook of vectors representing typical sound source patterns. Matching the vectors of the codebook to a reconstructed signal indicates whether proper separation has occurred. The vectors may be linear prediction vectors, among others.

Abstract: A dispersed vector generator used in an excitation vector generator for a speech coder/decoder is provided. The dispersed vector generator includes a pulse vector generating section that generates a pulse vector having a signed unit pulse on one element of a vector axis. The dispersed vector generator also includes a dispersion pattern storing section that stores a plurality of dispersion patterns, a switch that selects a dispersion pattern out of the plurality of dispersion patterns stored in the dispersion pattern storing section and a pulse vector dispersion section that generates a dispersed vector by convoluting the selected dispersion pattern and the pulse vector.

Abstract: The present invention is a method for determining linear predictive coding filter parameters for encoding a voice signal. The method includes sampling a voice signal, grouping the samples into a plurality of frames, generating a plurality of reflection coefficients for each frame of samples, quantizing the reflection coefficients, generating spectral coefficients from the quantized reflection coefficients, selecting a quantized reflection coefficient having the smallest log-spectral distance between a quantized spectrum, and an unquantized spectrum and, converting the selected quantized reflection coefficient to linear predictive coding (LPC) filter coefficient.

Abstract: A system and method is provided that employs a frequency domain interpolative CODEC system for low bit rate coding of speech which comprises a linear prediction (LP) front end adapted to process an input signal providing LP parameters which are quantized and encoded over predetermined intervals and used to compute a LP residual signal. An open loop pitch estimator adapted to process the LP residual signal, a pitch quantizer, and a pitch interpolator also provides a pitch contour within the predetermined intervals.

Abstract: An enhanced analysis-by-synthesis waveform interpolative speech coder able to operate at 2.8 kbps. Novel features include dual-predictive analysis-by-synthesis quantization of the slowly-evolving waveform, efficient parametrization of the rapidly-evolving waveform magnitude, and analysis-by-synthesis vector quantization of the rapidly evolving waveform parameter. Subjective quality tests indicate that it exceeds G.723.1 at 5.3 kbps, and of G.723.1 at 6.3 kbps.

Abstract: A method and system for quantizing LSF vectors in a speech coder, wherein predicted LSF values based on previously decoded output values are used to estimate spectral distortion, along with the residual codebook vectors and the LSF coefficients. The method comprises the steps of obtaining a plurality of quantized LSF coefficients from the respective predicted LSF values and the residual codebook vectors; rearranging the quantized LSF coefficients in the frequency domain in an orderly fashion; obtaining the spectral distortion from the rearranged quantized LSF coefficients and the respective LSF coefficients; and an optimal code vector is selected based on the spectral distortion.

Abstract: A system and method is provided that employs a frequency domain interpolative CODEC system for low bit rate coding of speech which comprises a linear prediction (LP) front end adapted to process an input signal that provides LP parameters which are quantized and encoded over predetermined intervals and used to compute a LP residual signal. An open loop pitch estimator adapted to process the LP residual signal, a pitch quantizer, and a pitch interpolator and provide a pitch contour within the predetermined intervals is also provided.

Abstract: The input signal can be quickly and accurately classified and a descriptor can be generated according to the result of classification. Then, the input signal can be retrieved on the basis of the result of classification or the descriptor. A signal processing apparatus comprises a time block splitting section 3 for splitting an audio signal into blocks that are typically 1 second long, a feature extracting section 4 for extracting a characteristic quantity of 18 degrees on the signal attribute from the audio signal in each block and a vector quantizing section 5 for carrying out an operation of categorical classification for the audio signal of each block by means of a vector quantization technique that uses a VQ code book 8 and a characteristic vector formed from the characteristic quantity of 18 degrees. The vector quantizing section 5 outputs a classification label obtained as a result of the categorical classification and a descriptor indicating the reliability of the label.

Abstract: The LSF quantizer for a wideband speech coder comprises a subtracter for receiving an input LSF coefficient vector and removing a DC component from it; a memory-based vector quantizer and a memoryless vector quantizer for respectively receiving the DC-component-removed LSF coefficient vector and independently quantizing the same; a switch for receiving quantized vectors respectively quantized by the memory-based vector quantizer and the memoryless vector quantizer, selecting a quantized vector that has less quantized error that is a difference between the received quantized vector and the input LSF coefficent vector from among the received quantized vectors, and outputting the same; and an adder for adding the quantized vector selected by the switch to the DC component of the LSF coefficient vector.

Abstract: A multiple description coder generates a number of different descriptions of a given portion of a signal in a wireless communication system, using multiple description scalar quantization (MDSQ) or another type of multiple description coding. The different descriptions of the given portion of the signal are then arranged into packets such that at least a first description of the given portion is placed in a first packet and a second description is placed in a second packet. Each of the packets are then transmitted using a frequency hopping modulator, and the hopping rate of the modulator is selected or otherwise configured based at least in part on the number of descriptions generated for the different portions of the signal.

Abstract: A method of searching a plurality of Vector Quantization (VQ) codevectors for a preferred one of the VQ codevectors to be used as an output of a vector quantizer for encoding a speech signal, includes determining a quantized prediction residual vector, and calculating a corresponding unquantized prediction residual vector and the energy of the difference between these two vectors (that is, a VQ error vector).

Abstract: According to one embodiment of the invention, a multistage vector list quantizer comprises a first stage quantizer to select candidate first stage codewords from a plurality of first stage codewords, a reference table memory storing a set of second stage codewords for each first stage codeword, and a second stage codebook constructor to generate a reduced complexity second stage codebook that is the union of sets corresponding to the candidate first stage codewords selected by the first stage quantizer.

Abstract: In a coding procedure, a spectral content of a speech signal is estimated. A preferential coding algorithm or preferential value of at least one coding parameter is selected based on the estimated spectral content of the speech signal. The speech signal is coded in accordance with the selected coding algorithm or the selected coding parameter to control the operation of one or more of the following: a pre-processing filter, a post-processing filter, a coding control coefficient, a weighting filter, a synthesis filter, and a quantization table.

Abstract: A system and method is provided that employs a frequency domain interpolative CODEC system for low bit rate coding of speech which comprises a linear prediction (LP) front end adapted to process an input signal that provides LP parameters which are quantized and encoded over predetermined intervals and used to compute a LP residual signal. An open loop pitch estimator adapted to process the LP residual signal, a pitch quantizer, and a pitch interpolator and provide a pitch contour within the predetermined intervals is also provided.

Abstract: The total number of entries of an algebraic codebook is decreased by liming a random code vector generated from the algebraic codebook, and entries of a random codebook with a large number of pulses are assigned to a decreased portion. Further, the number of entries of the decreased portion is adaptively switched according to a mode.

Abstract: Speech data containing waveform data is extracted from an existing speech waveform dictionary and input. A part used for speech synthesis in the waveform data is specified, and a starting point and an ending point for compression are set before and after the part. The waveform data is compressed with respect to a compression interval specified by the starting point and the ending point for compression. The compressed waveform data is expanded, and the compression interval, in which an expansion result of the compressed waveform data has highest quality, is determined as a compression/expansion position. The compressed waveform data, and the starting point and the ending point for compression are registered in a database as waveform data used for speech synthesis.

Abstract: A music data distribution system for distributing music data to an external device connected to a network, comprises: a storage device that stores first music data; a receiver that receives a music data distribution request from the external device connected to the network, the music data distribution request comprising at least music data identification information and music data quality information; a reading device that reads the first music data from said storage device in accordance with the music data identification information; a quality converter that converts the first music data into second music data having a quality different from the first music data in accordance with the music data quality information; and a transmitter that transmits the first or the second music data to the external device in accordance with contents of the music data distribution request.

Abstract: An encoding apparatus includes a band gain encoding section for calculating an average amplitude of a frequency spectrum stream corresponding to each of a plurality of frequency bands so as to generate a first code representing the average amplitude of the frequency spectrum stream; an encoding band determination section for determining at least one frequency band, for which the corresponding frequency spectrum stream is to be quantized and encoded from among the plurality of frequency bands; a spectrum encoding section for quantizing and encoding the frequency spectrum stream of each of the at least one frequency band determined by the encoding band determination section so as to generate a second code; and an encoded stream generation section for generating an encoded stream based on the first code and the second code.

Abstract: An apparatus for providing at least first and second representations of an audio signal for use in a communications system is described. The apparatus comprises a first quantizer for quantizing at least a portion of the signal in accordance with a first multidimensional lattice to generate a first representation. The apparatus further comprises a second quantizer for quantizing at least a portion of the signal in accordance with a second, different multidimensional lattice to generate a second representation. In an illustrative embodiment, the first representation is a core representation containing core audio information. The second representation is an enhancement representation containing enhancement audio information. The core representation is necessary for recovering the audio signal with minimal acceptable quality. Audio quality is enhanced when the core representation, together with the enhancement representation, is used to recover the audio signal.

Abstract: With respect to audio signal coding and decoding apparatuses, there is provided a coding apparatus that enables a decoding apparatus to reproduce an audio signal even through it does not use all of data from the coding apparatus, and a decoding apparatus corresponding to the coding apparatus. A quantization unit constituting a coding apparatus includes a first sub-quantization unit comprising sub-quantization units for low-band, intermediate-band, and high-band; a second sub-quantization unit for quantizing quantization errors from the first sub-quantization unit; and a third sub-quantization unit for quantizing quantization errors which have been processed by the first sub-quantization unit and the second sub-quantization unit.

Abstract: A new method for quantization of the LPC coefficients in a speech coder includes a new weighted error measure including every frame sampling an impulse response from LPC filter 21 of said coder, filtering the samples using a perceptual weighting filter 39 and processing in a computer 39 to calculate autocorrelation function of the weighted impulse response, computing Jacobian matrix for LSF (Line Spectral Frequency), computing correlation of rows of Jacobian matrix and calculating LSF weights by multiplying correlation matrices.

Abstract: Compressing the digitized time-domain continuous input signal typically includes formatting the input signal into a plurality of time-domain blocks having boundaries, forming an overlapping time-domain block by prepending a fraction of a previous time-domain block to a current time-domain block, transforming each overlapping time-domain block to a transform domain block including a plurality of coefficients, partitioning the coefficients of each transform domain block into signal coefficients and residue coefficients, quantizing the signal coefficients for each transformed domain block and generating signal quantization indices indicative of such quantization, modeling the residue coefficients for each transform domain block as stochastic noise and generating residue quantization indices indicative of such quantization, and formatting the signal quantization indices and the residue quantization indices for each transform domain block as an output bit-stream. The continuous data may include audio data.

Abstract: Sound source separation, without permutation, using convolutional mixing independent component analysis based on a priori knowledge of the target sound source is disclosed. The target sound source can be a human speaker. The reconstruction filters used in the sound source separation take into account the a priori knowledge of the target sound source, such as an estimate the spectra of the target sound source. The filters may be generally constructed based on a speech recognition system. Matching the words of the dictionary of the speech recognition system to a reconstructed signal indicates whether proper separation has occurred. More specifically, the filters may be constructed based on a vector quantization codebook of vectors representing typical sound source patterns. Matching the vectors of the codebook to a reconstructed signal indicates whether proper separation has occurred. The vectors may be linear prediction vectors, among others.

Abstract: An audio signal coding apparatus includes a first-stage encoder for quantizing the time-to-frequency transformed audio signal and second-and-subsequent-stages of encoders each for quantizing a quantization error output from the previous-stage encoder A characteristic decision unit is provided which decides the frequency band of an audio signal to be quantized by each encoder of multiple-stage encoders, and a coding band control unit receives the frequency band decided by the characteristic decision unit and the time-to-frequency transformed audio signal, decides the order of connecting the respective encoders, and transforms the quantization bands of the encoders and the connecting order to code sequences. Therefore, it is possible to provide an audio signal coding apparatus performing adaptive scalable coding, which exhibits sufficient performance when coding various audio signals.

Abstract: A low bit rate phase excited linear prediction type speech encoder filters a speech signal to limit its bandwidth and then fragments the filtered speech signal into speech segments. The speech segments are decomposed into a spectral envelope and an LP residual signal. The spectral envelope is represented by LP filter coefficients. The LP filter coefficients are converted into line spectral frequencies (LSF). Each speech segment is also classified as one of a voiced segment and an unvoiced segment based on a pitch of the segment. Parameters are extracted from the LP residual signal, where for an unvoiced segment the extracted parameters include pitch and gain and for a voiced segment the extracted parameters include pitch, gain and excitation level. The extracted parameters are then quantized.

Abstract: A method and apparatus for reducing the complexity of linear prediction analysis-by-synthesis (LPAS) speech coders. The speech coder includes a multi-tap pitch predictor having various parameters and utilizing an adaptive codebook subdivided into at least a first vector codebook and a second vector codebook. The pitch predictor removes certain redundancies in a subject speech signal and vector quantizes the pitch predictor parameters. Further included is a source excitation (fixed) codebook that indicates pulses in the subject speech signal by deriving corresponding vector values. Serial optimization of the adaptive codebook first and then the fixed codebook produces a low complexity LPAS speech coder of the present invention.

Abstract: An impulse detector which can detect both low and high levels of impulse noise in a CDMA system is comprised of circuitry to calculate the background noise level in unused codes. Another circuit calculates the average noise power in the unused codes of each spreading interval to output the noise power per spreading interval. This average is continuously averaged over spreading intervals by another circuit which outputs the average background noise power. A comparator compares the noise power in the current spreading interval with the background noise power plus a programmable threshold and generates an erasure indication if the background noise power plus a discrimination threshold is exceeded.

Abstract: A voice coding apparatus which can obtain preferable sound quality at a low bit rate is provided. In a mode decision circuit 800 of the voice coding apparatus, a mode is decided from an input voice signal by using a characteristic amount every sub-frame. In a sound source quantization circuit 350, in case of a predetermined mode, the amplitude or polarity of a non-zero pulse is calculated in advance. Further, combinations of a plurality of shift amounts by which the position of a predetermined pulse is time-shifted and a gain code vector for quantizing a gain are searched. Finally, the combination which minimizes distortion between a reproduced voice and an input voice is selected.

Abstract: Code-excited linear prediction speech encoders/decoders with excitation including an algebraic codebook contribution encoded with a single sign bit for each track of pulses by inferring pulse amplitude signs from the pulse position code ordering within a codeword.

Abstract: A fast search method for searching for an optimum codeword for nearest neighbor vector quantization. An upper boundary value and a lower boundary value between which an optimum codeword will exist in a codebook are calculated using a distortion of a designated element in an input vector and an experimentally determined threshold. Further, a start point and an end point for codebook search are determined using a binary search method from a codebook rearranged in descending order, and a full search scheme is applied only within a search range calculated by the determined start point and end point, thereby determining an optimum codeword for nearest neighbor vector quantization.

Abstract: A speech encoder that analyzes and classifies each frame of speech as being periodic-like speech or non-periodic like speech where the speech encoder performs a different gain quantization process depending if the speech is periodic or not. If the speech is periodic, the improved speech encoder obtains the pitch gains from the unquantized weighted speech signal and performs a pre- vector quantization of the adaptive codebook gain Gp for each subframe of the frame before subframe processing begins and a closed-loop delayed decision vector quantization of the fixed codebook gain GC. If the frame of speech is non-periodic, the speech encoder may use any known method of gain quantization.

Abstract: Speech signal information is formatted, processed and transported in accordance with a format adapted for TCP/IP protocols used on the Internet and other communications networks. NULL characters are used for indicating the end of a voice segment. The method is useful for distributed speech recognition systems such as a client-server system, typically implemented on an intranet or over the Internet based on user queries at his/her computer, a PDA, or a workstation using a speech input interface.