US 6,188,980 B1Grant

Synchronized encoder-decoder frame concealment using speech coding parameters including line spectral frequencies and filter coefficients

Issue Date:2001-02-13

•20 Claims

•11 Drawing Sheets

Abstract

A multi-rate speech codec supports a plurality of encoding bit rate modes by adaptively selecting encoding bit rate modes to match communication channel restrictions. In higher bit rate encoding modes, an accurate representation of speech through CELP (code excited linear prediction) and other associated modeling parameters are generated for higher quality decoding and reproduction. The encoder produces a series of LSF (line spectral frequencies) vectors. For filter stability, each LSF vector comprises an ascending sequence of LSF values. Occasionally, pairs of LSF values are produced (or become through an introduction of channel error) out of ascending order. In response, the encoder performs frame erasure, LSF concealment or pair flipping. With a relatively large number of out of order pairs, frame erasure is applied. With a single out of order pair, within the LSF vector, the pair are flipped. Likewise, with two pairs out of order, the previous LSF vector is used to generate the current LSF vector using concealment. The performance of frame erasure, LSF concealment or pair flipping is also performed in the decoder as well as in the encoder in some embodiments.

Metadata

Assignee

Conexant Systems, Inc.

Inventor

Jes Thyssen

Application Information

Application Number:US 09/154,653

Filing Date:1998-09-18

Priority Date:1998-08-24

Art Unit:7

Classifications

IPC:

G10L 1100

Field of Search:

705221705203705230705201

Patent Drawings (11 sheets)

Description

[0002] The present application is based on U.S. Provisional Application Serial No. 60/097,569, filed Aug. 24, 1998.

Microfiche Appendix

[0003] A microfiche appendix is included in the application.

Incorporation by Reference

[0004] The following applications are hereby incorporated herein by reference in their entirety and made part of the present application:

[0005] 1) U.S. Provisional Application Serial No. 60/097,569, entitled “Adaptive Rate Speech Codec,” filed Aug. 24, 1998;

[0006] 2) U.S. patent application Ser. No. 09/154,675, entitled “Speech Encoder Using Continuous Warping In Long Term Preprocessing,” filed Sep. 18, 1998;

[0007] 3) U.S. patent application Ser. No. 09/156,814, entitled “Completed Fixed Codebook For Speech Encoder,” filed Sep. 18, 1998;

[0008] 4) U.S. patent application Ser. No. 09/156,649, entitled “Comb Codebook Structure,” filed Sep. 18, 1998;

[0009] 5) U.S. patent application Ser. No. 09/156,648, entitled “Low Complexity Random Codebook Structure,” filed Sep. 18, 1998;

[0010] 6) U.S. patent application Ser. No. 09/156,650, entitled “Speech Encoder Using Gain Normalization That Combines Open And Closed Loop Gains,” filed Sep. 18, 1998;

[0011] 7) U.S. patent application Ser. No. 09/156,832, entitled “Speech Encoder Using Voice Activity Detection In Coding Noise,” filed Sep. 18, 1998;

[0012] 8) U.S. patent application Ser. No. 09/154,654, entitled “Pitch Determination Using Speech Classification And Prior Pitch Estimation,” filed Sep. 18, 1998;

[0013] 9) U.S. patent application Ser. No. 09/154,657, entitled “Speech Encoder Using A Classifier For Smoothing Noise Coding,” filed Sep. 18, 1998;

[0014] 10) U.S. patent application Ser. No. 09/156,826, entitled “Adaptive Tilt Compensation For Synthesized Speech Residual,” filed Sep. 18, 1998;

[0015] 11) U.S. patent application Ser. No. 09/154,662, entitled “Speech Classification And Parameter Weighting Used In Codebook Search,” filed Sep. 18, 1998;

[0016] 12) U.S. patent application Ser. No. 09/154,663, entitled “Adaptive Gain Reduction To Produce Fixed Codebook Target Signal,” filed Sep. 18, 1998;

[0017] 13) U.S. patent application Ser. No. 09/154,660, entitled “Speech Encoder Adaptively Applying Pitch Long-Term Prediction and Pitch Preprocessing With Continuous Warping,” filed Sep. 18, 1998.

Background

[0018] 1. Technical Field

[0019] The present invention relates generally to speech encoding and decoding in voice communication systems; and, more particularly, it relates to various techniques used with code-excited linear prediction coding to obtain high quality speech reproduction through a limited bit rate communication channel.

[0020] 2. Related Art

[0021] Signal modeling and parameter estimation play significant roles in communicating voice information with limited bandwidth constraints. To model basic speech sounds, speech signals are sampled as a discrete waveform to be digitally processed. In one type of signal coding technique called LPC (linear predictive coding), the signal value at any particular time index is modeled as a linear function of previous values. A subsequent signal is thus linearly predictable according to an earlier value. As a result, efficient signal representations can be determined by estimating and applying certain prediction parameters to represent the signal.

[0022] Applying LPC techniques, a conventional source encoder operates on speech signals to extract modeling and parameter information for communication to a conventional source decoder via a communication channel. Once received, the decoder attempts to reconstruct a counterpart signal for playback that sounds to a human ear like the original speech.

[0023] A certain amount of communication channel bandwidth is required to communicate the modeling and parameter information to the decoder. In embodiments, for example where the channel bandwidth is shared and real-time reconstruction is necessary, a reduction in the required bandwidth proves beneficial. However, using conventional modeling techniques, the quality requirements in the reproduced speech limit the reduction of such bandwidth below certain levels.

[0024] In conventional speech codecs, linear predictive coding (LPC) produces LSF (Line Spectral Frequencies) vectors that occasionally have LSF entries that are out of ascending order. LSF vectors may also become out of ascending order when errors are introduced, for example, through a communication channel. Flipped LSFs at a decoder produce unwanted clicking or popping sounds due to filter instability.

[0025] Further limitations and disadvantages of conventional systems will become apparent to one of skill in the art after reviewing the remainder of the present application with reference to the drawings.

Summary of the Invention

[0026] Various aspects of the present invention can be found in a speech encoding system using an analysis by synthesis approach on a speech signal. The speech encoding system comprises an encoder that generates a series of line spectral frequencies that are occasionally produced out of order. If so, the encoder selects from a plurality of correction techniques to accommodate the out of order series. One of the correction techniques involves reordering the series of line spectral frequencies that are produced out of order.

[0027] Depending on the embodiment, there may be several types of other correction techniques employed. For example, line spectral frequency concealment, application of at least a portion of one previous series of line spectral frequencies, or frame erasure might be applied. Similarly, a further technique might involve disregarding at least in part the series of line spectral frequencies that are produced out of order.

[0028] The encoder may select one of the plurality of correction techniques by examining the series of line spectral frequencies that are produced out of order, for example, by considering a number of line spectral frequencies out of order.

[0029] Other aspects of the present invention may be found in a method used by a speech encoder that operates on a speech signal. The method involves producing from the speech signal a series of line spectral frequencies that are occasionally produced out of order. When out of order, the method also involves deciding whether to reorder the series of line spectral frequencies that are produced out of order, or to replace at least a portion of the series of line spectral frequencies that are produced out of order using at least a portion of one previous series of line spectral frequencies.

[0030] The method may also involve determining a number of line spectral frequencies out of order in the series of line spectral frequencies that are produced out of order. Therein, the number of line spectral frequencies out of order can be used in the decision to either reorder or replace.

[0031] A speech decoder may also be included which reorders a received series of line spectral frequencies that are out of order. The speech decoder may also or alternatively replace at least a portion of a received series of line spectral frequencies that are out of order using at least a portion of one previously received series of filter coefficients.

[0032] Other aspects, advantages and novel features of the present invention will become apparent from the following detailed description of the invention when considered in conjunction with the accompanying drawings.

Brief Description of the Drawings

[0033] FIG. 1ais a schematic block diagram of a speech communication system illustrating the use of source encoding and decoding in accordance with the present invention.

[0034] FIG. 1bis a schematic block diagram illustrating an exemplary communication device utilizing the source encoding and decoding functionality of FIG. 1a.

[0035] FIGS. 2-4 are functional block diagrams illustrating a multi-step encoding approach used by one embodiment of the speech encoder illustrated in FIGS. 1aand 1b. In particular, FIG. 2 is a functional block diagram illustrating of a first stage of operations performed by one embodiment of the speech encoder of FIGS. 1aand 1b. FIG. 3 is a functional block diagram of a second stage of operations, while FIG. 4 illustrates a third stage.

[0036] FIG. 5 is a block diagram of one embodiment of the speech decoder shown in FIGS. 1aand 1bhaving corresponding functionality to that illustrated in FIGS. 2-4.

[0037] FIG. 6 is a block diagram of an alternate embodiment of a speech encoder that is built in accordance with the present invention.

[0038] FIG. 7 is a block diagram of an embodiment of a speech decoder having corresponding functionality to that of the speech encoder of FIG. 6.

[0039] FIG. 8 is a flow diagram illustrating a procedure used by a speech encoder built in accordance with the present invention to correct LPC filter coefficients.

[0040] FIG. 9 is a flow diagram illustrating a procedure similar to that illustrated in FIG. 8 but used by a speech decoder built in accordance with the present invention to correct LPC filter coefficients.

[0041] FIG. 10 provides a more detailed specific embodiment of the functionality of FIGS. 8 and 9 performed by either an encoder or a decoder, yet without using frame erasure.

Detailed Description

[0042] FIG. 1ais a schematic block diagram of a speech communication system illustrating the use of source encoding and decoding in accordance with the present invention. Therein, a speech communication system 100 supports communication and reproduction of speech across a communication channel 103. Although it may comprise for example a wire, fiber or optical link, the communication channel 103 typically comprises, at least in part, a radio frequency link that often must support multiple, simultaneous speech exchanges requiring shared bandwidth resources such as may be found with cellular telephony embodiments.

[0043] Although not shown, a storage device may be coupled to the communication channel 103 to temporarily store speech information for delayed reproduction or playback, e.g., to perform answering machine functionality, voiced email, etc. Likewise, the communication channel 103 might be replaced by such a storage device in a single device embodiment of the communication system 100 that, for example, merely records and stores speech for subsequent playback.

[0044] In particular, a microphone 111 produces a speech signal in real time. The microphone 111 delivers the speech signal to an A/D (analog to digital) converter 115. The A/D converter 115 converts the speech signal to a digital form then delivers the digitized speech signal to a speech encoder 117.

[0045] The speech encoder 117 encodes the digitized speech by using a selected one of a plurality of encoding modes. Each of the plurality of encoding modes utilizes particular techniques that attempt to optimize quality of resultant reproduced speech. While operating in any of the plurality of modes, the speech encoder 117 produces a series of modeling and parameter information (hereinafter “speech indices”), and delivers the speech indices to a channel encoder 119.

[0046] The channel encoder 119 coordinates with a channel decoder 131 to deliver the speech indices across the communication channel 103. The channel decoder 131 forwards the speech indices to a speech decoder 133. While operating in a mode that corresponds to that of the speech encoder 117, the speech decoder 133 attempts to recreate the original speech from the speech indices as accurately as possible at a speaker 137 via a D/A (digital to analog) converter 135.

[0047] The speech encoder 117 adaptively selects one of the plurality of operating modes based on the data rate restrictions through the communication channel 103. The communication channel 103 comprises a bandwidth allocation between the channel encoder 119 and the channel decoder 131. The allocation is established, for example, by telephone switching networks wherein many such channels are allocated and reallocated as need arises. In one such embodiment, either a 22.8 kbps (kilobits per second) channel bandwidth, i.e., a full rate channel, or a 11.4 kbps channel bandwidth, i.e., a half rate channel, may be allocated.

[0048] With the full rate channel bandwidth allocation, the speech encoder 117 may adaptively select an encoding mode that supports a bit rate of 11.0, 8.0, 6.65 or 5.8 kbps. The speech encoder 117 adaptively selects an either 8.0, 6.65, 5.8 or 4.5 kbps encoding bit rate mode when only the half rate channel has been allocated. Of course these encoding bit rates and the aforementioned channel allocations are only representative of the present embodiment. Other variations to meet the goals of alternate embodiments are contemplated.

[0049] With either the full or half rate allocation, the speech encoder 117 attempts to communicate using the highest encoding bit rate mode that the allocated channel will support. If the allocated channel is or becomes noisy or otherwise restrictive to the highest or higher encoding bit rates, the speech encoder 117 adapts by selecting a lower bit rate encoding mode. Similarly, when the communication channel 103 becomes more favorable, the speech encoder 117 adapts by switching to a higher bit rate encoding mode.

[0050] With lower bit rate encoding, the speech encoder 117 incorporates various techniques to generate better low bit rate speech reproduction. Many of the techniques applied are based on characteristics of the speech itself. For example, with lower bit rate encoding, the speech encoder 117 classifies noise, unvoiced speech, and voiced speech so that an appropriate modeling scheme corresponding to a particular classification can be selected and implemented. Thus, the speech encoder 117 adaptively selects from among a plurality of modeling schemes those most suited for the current speech. The speech encoder 117 also applies various other techniques to optimize the modeling as set forth in more detail below.

[0051] FIG. 1bis a schematic block diagram illustrating several variations of an exemplary communication device employing the functionality of FIG. 1a. A communication device 151 comprises both a speech encoder and decoder for simultaneous capture and reproduction of speech. Typically within a single housing, the communication device 151 might, for example, comprise a cellular telephone, portable telephone, computing system, etc. Alternatively, with some modification to include for example a memory element to store encoded speech information the communication device 151 might comprise an answering machine, a recorder, voice mail system, etc.

[0052] A microphone 155 and an A/D converter 157 coordinate to deliver a digital voice signal to an encoding system 159. The encoding system 159 performs speech and channel encoding and delivers resultant speech information to the channel. The delivered speech information may be destined for another communication device (not shown) at a remote location.

[0053] As speech information is received, a decoding system 165 performs channel and speech decoding then coordinates with a D/A converter 167 and a speaker 169 to reproduce something that sounds like the originally captured speech.

[0054] The encoding system 159 comprises both a speech processing circuit 185 that performs speech encoding, and a channel processing circuit 187 that performs channel encoding. Similarly, the decoding system 165 comprises a speech processing circuit 189 that performs speech decoding, and a channel processing circuit 191 that performs channel decoding.

[0055] Although the speech processing circuit 185 and the channel processing circuit 187 are separately illustrated, they might be combined in part or in total into a single unit. For example, the speech processing circuit 185 and the channel processing circuitry 187 might share a single DSP (digital signal processor) and/or other processing circuitry. Similarly, the speech processing circuit 189 and the channel processing circuit 191 might be entirely separate or combined in part or in whole. Moreover, combinations in whole or in part might be applied to the speech processing circuits 185 and 189, the channel processing circuits 187 and 191, the processing circuits 185, 187, 189 and 191, or otherwise.

[0056] The encoding system 159 and the decoding system 165 both utilize a memory 161. The speech processing circuit 185 utilizes a fixed codebook 181 and an adaptive codebook 183 of a speech memory 177 in the source encoding process. The channel processing circuit 187 utilizes a channel memory 175 to perform channel encoding. Similarly, the speech processing circuit 189 utilizes the fixed codebook 181 and the adaptive codebook 183 in the source decoding process. The channel processing circuit 187 utilizes the channel memory 175 to perform channel decoding.

[0057] Although the speech memory 177 is shared as illustrated, separate copies thereof can be assigned for the processing circuits 185 and 189. Likewise, separate channel memory can be allocated to both the processing circuits 187 and 191. The memory 161 also contains software utilized by the processing circuits 185,187,189 and 191 to perform various functionality required in the source and channel encoding and decoding processes.

[0058] FIGS. 2-4 are functional block diagrams illustrating a multi-step encoding approach used by one embodiment of the speech encoder illustrated in FIGS. 1aand 1b. In particular, FIG. 2 is a functional block diagram illustrating of a first stage of operations performed by one embodiment of the speech encoder shown in FIGS. 1aand 1b. The speech encoder, which comprises encoder processing circuitry, typically operates pursuant to software instruction carrying out the following functionality.

[0059] At a block 215, source encoder processing circuitry performs high pass filtering of a speech signal 211. The filter uses a cutoff frequency of around 80 Hz to remove, for example, 60 Hz power line noise and other lower frequency signals. After such filtering, the source encoder processing circuitry applies a perceptual weighting filter as represented by a block 219. The perceptual weighting filter operates to emphasize the valley areas of the filtered speech signal.

[0060] If the encoder processing circuitry selects operation in a pitch preprocessing (PP) mode as indicated at a control block 245, a pitch preprocessing operation is performed on the weighted speech signal at a block 225. The pitch preprocessing operation involves warping the weighted speech signal to match interpolated pitch values that will be generated by the decoder processing circuitry. When pitch preprocessing is applied, the warped speech signal is designated a first target signal 229. If pitch preprocessing is not selected the control block 245, the weighted speech signal passes through the block 225 without pitch preprocessing and is designated the first target signal 229.

[0061] As represented by a block 255, the encoder processing circuitry applies a process wherein a contribution from an adaptive codebook 257 is selected along with a corresponding gain 257 which minimize a first error signal 253. The first error signal 253 comprises the difference between the first target signal 229 and a weighted, synthesized contribution from the adaptive codebook 257.

[0062] At blocks 247, 249 and 251, the resultant excitation vector is applied after adaptive gain reduction to both a synthesis and a weighting filter to generate a modeled signal that best matches the first target signal 229. The encoder processing circuitry uses LPC (linear predictive coding) analysis, as indicated by a block 239, to generate filter parameters for the synthesis and weighting filters. The weighting filters 219 and 251 are equivalent in functionality.

[0063] Next, the encoder processing circuitry designates the first error signal 253 as a second target signal for matching using contributions from a fixed codebook 261. The encoder processing circuitry searches through at least one of the plurality of subcodebooks within the fixed codebook 261 in an attempt to select a most appropriate contribution while generally attempting to match the second target signal.

[0064] More specifically, the encoder processing circuitry selects an excitation vector, its corresponding subcodebook and gain based on a variety of factors. For example, the encoding bit rate, the degree of minimization, and characteristics of the speech itself as represented by a block 279 are considered by the encoder processing circuitry at control block 275. Although many other factors may be considered, exemplary characteristics include speech classification, noise level, sharpness, periodicity, etc. Thus, by considering other such factors, a first subcodebook with its best excitation vector may be selected rather than a second subcodebook's best excitation vector even though the second subcodebook's better minimizes the second target signal 265.

[0065] FIG. 3 is a functional block diagram depicting of a second stage of operations performed by the embodiment of the speech encoder illustrated in FIG. 2. In the second stage, the speech encoding circuitry simultaneously uses both the adaptive the fixed codebook vectors found in the first stage of operations to minimize a third error signal 311.

[0066] The speech encoding circuitry searches for optimum gain values for the previously identified excitation vectors (in the first stage) from both the adaptive and fixed codebooks 257 and 261. As indicated by blocks 307 and 309, the speech encoding circuitry identifies the optimum gain by generating a synthesized and weighted signal, i.e., via a block 301 and 303, that best matches the first target signal 229 (which minimizes the third error signal 311). Of course if processing capabilities permit, the first and second stages could be combined wherein joint optimization of both gain and adaptive and fixed codebook rector selection could be used.

[0067] FIG. 4 is a functional block diagram depicting of a third stage of operations performed by the embodiment of the speech encoder illustrated in FIGS. 2 and 3. The encoder processing circuitry applies gain normalization, smoothing and quantization, as represented by blocks 401, 403 and 405, respectively, to the jointly optimized gains identified in the second stage of encoder processing. Again, the adaptive and fixed codebook vectors used are those identified in the first stage processing.

[0068] With normalization, smoothing and quantization functionally applied, the encoder processing circuitry has completed the modeling process. Therefore, the modeling parameters identified are communicated to the decoder. In particular, the encoder processing circuitry delivers an index to the selected adaptive codebook vector to the channel encoder via a multiplexor 419. Similarly, the encoder processing circuitry delivers the index to the selected fixed codebook vector, resultant gains, synthesis filter parameters, etc., to the muliplexor 419. The multiplexor 419 generates a bit stream 421 of such information for delivery to the channel encoder for communication to the channel and speech decoder of receiving device.

[0069] FIG. 5 is a block diagram of an embodiment illustrating functionality of speech decoder having corresponding functionality to that illustrated in FIGS. 2-4. As with the speech encoder, the speech decoder, which comprises decoder processing circuitry, typically operates pursuant to software instruction carrying out the following functionality.

[0070] A demultiplexor 511 receives a bit stream 513 of speech modeling indices from an often remote encoder via a channel decoder. As previously discussed, the encoder selected each index value during the multi-stage encoding process described above in reference to FIGS. 2-4. The decoder processing circuitry utilizes indices, for example, to select excitation vectors from an adaptive codebook 515 and a fixed codebook 519, set the adaptive and fixed codebook gains at a block 521, and set the parameters for a synthesis filter 531.

[0071] With such parameters and vectors selected or set, the decoder processing circuitry generates a reproduced speech signal 539. In particular, the codebooks 515 and 519 generate excitation vectors identified by the indices from the demultiplexor 511. The decoder processing circuitry applies the indexed gains at the block 521 to the vectors which are summed. At a block 527, the decoder processing circuitry modifies the gains to emphasize the contribution of vector from the adaptive codebook 515. At a block 529, adaptive tilt compensation is applied to the combined vectors with a goal of flattening the excitation spectrum. The decoder processing circuitry performs synthesis filtering at the block 531 using the flattened excitation signal. Finally, to generate the reproduced speech signal 539, post filtering is applied at a block 535 deemphasizing the valley areas of the reproduced speech signal 539 to reduce the effect of distortion.

[0072] In the exemplary cellular telephony embodiment of the present invention, the A/D converter 115 (FIG. 1a) will generally involve analog to uniform digital PCM including: 1) an input level adjustment device; 2) an input anti-aliasing filter; 3) a sample-hold device sampling at 8 kHz; and 4) analog to uniform digital conversion to 13-bit representation.

[0073] Similarly, the D/A converter 135 will generally involve uniform digital PCM to analog including: 1) conversion from 13-bit/8 kHz uniform PCM to analog; 2) a hold device; 3) reconstruction filter including x/sin(x) correction; and 4) an output level adjustment device.

[0074] In terminal equipment, the A/D function may be achieved by direct conversion to 13-bit uniform PCM format, or by conversion to 8-bit/A-law compounded format. For the D/A operation, the inverse operations take place.

[0075] The encoder 117 receives data samples with a resolution of 13 bits left justified in a 16-bit word. The three least significant bits are set to zero. The decoder 133 outputs data in the same format. Outside the speech codec, further processing can be applied to accommodate traffic data having a different representation.

[0076] A specific embodiment of an AMR (adaptive multi-rate) codec with the operational functionality illustrated in FIGS. 2-5 uses five source codecs with bit-rates 11.0, 8.0, 6.65, 5.8 and 4.55 kbps. Four of the highest source coding bit-rates are used in the full rate channel and the four lowest bit-rates in the half rate channel.

[0077] All five source codecs within the AMR codec are generally based on a code-excited linear predictive (CELP) coding model. A 10th order linear prediction (LP), or short-term, synthesis filter, e.g., used at the blocks 249, 267, 301, 407 and 531 (of FIGS. 2-5), is used which is given by:

\begin{matrix} H (z) = \frac{1}{\hat{A} (z)} = \frac{1}{1 + \sum_{i = 1}^{m} {\hat{a}}_{i} z^{- i}}, & (1) \end{matrix}

[0078] where â_i,i=1, . . . , m, are the (quantized) linear prediction (LP) parameters.

[0079] A long-term filter, i.e., the pitch synthesis filter, is implemented using the either an adaptive codebook approach or a pitch pre-processing approach. The pitch synthesis filter is given by:

\begin{matrix} \frac{1}{B (z)} = \frac{1}{1 - g_{p} z^{- T}}, & (2) \end{matrix}

[0080] where T is the pitch delay and g_pis the pitch gain.

[0081] With reference to FIG. 2, the excitation signal at the input of the short-term LP synthesis filter at the block 249 is constructed by adding two excitation vectors from the adaptive and the fixed codebooks 257 and 261, respectively. The speech is synthesized by feeding the two properly chosen vectors from these codebooks through the short-term synthesis filter at the block 249 and 267, respectively.

[0082] The optimum excitation sequence in a codebook is chosen using an analysis-by-synthesis search procedure in which the error between the original and synthesized speech is minimized according to a perceptually weighted distortion measure. The perceptual weighting filter, e.g., at the blocks 251 and 268, used in the analysis-by-synthesis search technique is given by:

\begin{matrix} W (z) = \frac{A (z / γ_{1})}{A (z / γ_{2})}, & (3) \end{matrix}

[0083] where A(z) is the unquantized LP filter and 0<γ₂<γ₁<1 are the perceptual weighting factors. The values γ₁=[0.9, 0.94] and γ₂=0.6 are used. The weighting filter, e.g., at the blocks 251 and 268, uses the unquantized LP parameters while the formant synthesis filter, e.g., at the blocks 249 and 267, uses the quantized LP parameters. Both the unquantized and quantized LP parameters are generated at the block 239.

[0084] The present encoder embodiment operates on 20 ms (millisecond) speech frames corresponding to 160 samples at the sampling frequency of 8000 samples per second. At each 160 speech samples, the speech signal is analyzed to extract the parameters of the CELP model, i.e., the LP filter coefficients, adaptive and fixed codebook indices and gains. These parameters are encoded and transmitted. At the decoder, these parameters are decoded and speech is synthesized by filtering the reconstructed excitation signal through the LP synthesis filter.

[0085] More specifically, LP analysis at the block 239 is performed twice per frame but only a single set of LP parameters is converted to line spectrum frequencies (LSF) and vector quantized using predictive multi-stage quantization (PMVQ). The speech frame is divided into subframes. Parameters from the adaptive and fixed codebooks 257 and 261 are transmitted every subframe. The quantized and unquantized LP parameters or their interpolated versions are used depending on the subframe. An open-loop pitch lag is estimated at the block 241 once or twice per frame for PP mode or LTP mode, respectively.

[0086] Each subframe, at least the following operations are repeated. First, the encoder processing circuitry (operating pursuant to software instruction) computes x(n), the first target signal 229, by filtering the LP residual through the weighted synthesis filter W(z)H(z) with the initial states of the filters having been updated by filtering the error between LP residual and excitation. This is equivalent to an alternate approach of subtracting the zero input response of the weighted synthesis filter from the weighted speech signal.

[0087] Second, the encoder processing circuitry computes the impulse response, h(n), of the weighted synthesis filter. Third, in the LTP mode, closed-loop pitch analysis is performed to find the pitch lag and gain, using the first target signal 229, x(n), and impulse response, h(n), by searching around the open-loop pitch lag. Fractional pitch with various sample resolutions are used.

[0088] In the PP mode, the input original signal has been pitch-preprocessed to match the interpolated pitch contour, so no closed-loop search is needed. The LTP excitation vector is computed using the interpolated pitch contour and the past synthesized excitation.

[0089] Fourth, the encoder processing circuitry generates a new target signal x₂(n), the second target signal 253, by removing the adaptive codebook contribution (filtered adaptive code vector) from x(n). The encoder processing circuitry uses the second target signal 253 in the fixed codebook search to find the optimum innovation.

[0090] Fifth, for the 11.0 kbps bit rate mode, the gains of the adaptive and fixed codebook are scalar quantized with 4 and 5 bits respectively (with moving average prediction applied to the fixed codebook gain). For the other modes the gains of the adaptive and fixed codebook are vector quantized (with moving average prediction applied to the fixed codebook gain).

[0091] Finally, the filter memories are updated using the determined excitation signal for finding the first target signal in the next subframe.

[0092] The bit allocation of the AMR codec modes is shown in table 1. For example, for each 20 ms speech frame, 220, 160, 133, 116 or 91 bits are produced, corresponding to bit rates of 11.0, 8.0, 6.65, 5.8 or 4.55 kbps, respectively.

[0093] With reference to FIG. 5, the decoder processing circuitry, pursuant to software control, reconstructs the speech signal using the transmitted modeling indices extracted from the received bit stream by the demultiplexor 511. The decoder processing circuitry decodes the indices to obtain the coder parameters at each transmission frame. These parameters are the LSF vectors, the fractional pitch lags, the innovative code vectors, and the two gains.

[0094] The LSF vectors are converted to the LP filter coefficients and interpolated to obtain LP filters at each subframe. At each subframe, the decoder processing circuitry constructs the excitation signal by: 1) identifying the adaptive and innovative code vectors from the codebooks 515 and 519; 2) scaling the contributions by their respective gains at the block 521; 3) summing the scaled contributions; and 3) modifying and applying adaptive tilt compensation at the blocks 527 and 529. The speech signal is also reconstructed on a subframe basis by filtering the excitation through the LP synthesis at the block 531. Finally, the speech signal is passed through an adaptive post filter at the block 535 to generate the reproduced speech signal 539.

[0095] The AMR encoder will produce the speech modeling information in a unique sequence and format, and the AMR decoder receives the same information in the same way. The different parameters of the encoded speech and their individual bits have unequal importance with respect to subjective quality. Before being submitted to the channel encoding function the bits are rearranged in the sequence of importance.

[0096] Two pre-processing functions are applied prior to the encoding process: high-pass filtering and signal down-scaling. Down-scaling consists of dividing the input by a factor of 2 to reduce the possibility of overflows in the fixed point implementation. The high-pass filtering at the block 215 (FIG. 2) serves as a precaution against undesired low frequency components. A filter with cut off frequency of 80 Hz is used, and it is given by:

H_{hl} (z) = \frac{0.92727435 - 1.8544941 z^{- 1} + 0.92727435 z^{- 2}}{1 - 1.9059465 z^{- 1} + 0.9114024 z^{- 2}}

[0097] Down scaling and high-pass filtering are combined by dividing the coefficients of the numerator of H_hl(z) by 2.

[0098] Short-term prediction, or linear prediction (LP) analysis is performed twice per speech frame using the autocorrelation approach with 30 ms windows. Specifically, two LP analyses are performed twice per frame using two different windows. In the first LP analysis (LP_analysis_—1), a hybrid window is used which has its weight concentrated at the fourth subframe. The hybrid window consists of two parts. The first part is half a Hamming window, and the second part is a quarter of a cosine cycle. The window is given by:

w_{1} (n) = {\begin{matrix} 0.54 - 0.46 \cos (\frac{π n}{L}), & n = 0 to 214, L = 215 \\ \cos (\frac{0.49 (n - L) π}{25}), & n = 215 to 239 \end{matrix}

[0099] In the second LP analysis (LP_analysis_—2), a symmetric Hamming window is used.

w_{2} (n) = {\begin{matrix} 0.54 - 0.46 \cos (\frac{π n}{L}) & n = 0 to 119, L = 120 \\ 0.54 + 0.46 \cos (\frac{(n - L) π}{120}), & n = 120 to 239 \end{matrix}

[0100] In either LP analysis, the autocorrelations of the windowed speech s(n),n=0,239 are computed by:

r (k) = \sum_{n = k}^{239} s^{'} (n) s^{'} (n - k), k = 0, 10.

[0101] A 60 Hz bandwidth expansion is used by lag windowing, the autocorrelations using the window:

w_{lag} (i) = \exp [- \frac{1}{2} {(\frac{2 π 60 i}{8000})}^{2}], i = 1, 10.

[0102] Moreover, r(0) is multiplied by a white noise correction factor 1.0001 which is equivalent to adding a noise floor at −40 dB.

[0103] The modified autocorrelations r (0)=1.0001r(0) and r (k)=r(k)w_lag(k),k=1,10 are used to obtain the reflection coefficients k_iand LP filter coefficients a_i,i=1,10 using the Levinson-Durbin algorithm. Furthermore, the LP filter coefficients a_iare used to obtain the Line Spectral Frequencies (LSFs).

[0104] The interpolated unquantized LP parameters are obtained by interpolating the LSF coefficients obtained from the LP analysis_—1 and those from LP_analysis_—2 as:

[0105] q₁(n)=0.5q₄(n−1)+0.5q₂(n)

[0106] q₃(n)=0.5q₂(n)+0.5q₄(n)

[0107] where q₁(n) is the interpolated LSF for subframe 1, q₂(n) is the LSF of subframe 2 obtained from LP_analysis₁₃2 of current frame, q₃(n) is the interpolated LSF for subframe 3, q₄(n−1) is the LSF (cosine domain) from LP_analysis_—1 of previous frame, and q₄(n) is the LSF for subframe 4 obtained from LP_analysis_—1 of current frame. The interpolation is carried out in the cosine domain.

[0108] A VAD (Voice Activity Detection) algorithm is used to classify input speech frames into either active voice or inactive voice frame (background noise or silence) at a block 235 (FIG. 2).

[0109] The input speech s(n) is used to obtain a weighted speech signal s_w(n) by passing s(n) through a filter:

W (z) = \frac{A (z / γ1)}{A (z / γ2)} .

[0110] That is, in a subframe of size L_SF, the weighted speech is given by:

s_{w} (n) = s (n) + \sum_{i = 1}^{10} a_{i} γ_{1}^{i} s (n - i) - \sum_{i = 1}^{10} a_{i} γ_{2}^{i} s_{w} (n - i), n = 0, L_{—} SF - 1.

[0111] A voiced/unvoiced classification and mode decision within the block 279 using the input speech s(n) and the residual r_w(n) is derived where:

r_{w} (n) = s (n) + \sum_{i = 1}^{10} a_{i} γ_{1}^{i} s (n - i), n = 0, L_{—} SF - 1.

[0112] The classification is based on four measures: 1) speech sharpness P1_SHP;2) normalized one delay correlation P2_R1; 3) normalized zero-crossing rate P3_ZC; and4) normalized LP residual energy P4_RE.

[0113] The speech sharpness is given by:

{P1}_{—} SHP = \frac{\sum_{n = 0}^{L} abs (r_{w} (n))}{MaxL},

[0114] where Max is the maximum of abs(r_w(n)) over the specified interval of length L. The normalized one delay correlation and normalized zero-crossing rate are given by:

{P2}_{—} R1 = \frac{\sum_{n = 0}^{L - 1} s (n) s (n + 1)}{\sqrt{\sum_{n = 0}^{L - 1} s (n) s (n) \sum_{n = 0}^{L - 1} s (n + 1) s (n + 1)}}

{P3}_{—} ZC = \frac{1}{2 L} \sum_{i = 0}^{L - 1} [| sgn [s (i)] - sgn [s (i - 1)] |],

[0115] where sgn is the sign function whose output is either 1 or −1 depending that the input sample is positive or negative. Finally, the normalized LP residual energy is given by:

[0116] P4_—RE=1−{square root over (lpc_gain)}

[0117] where

{lpc}_{—} gain = \prod_{i = 1}^{10} (1 - k_{i}^{2}),

[0118] where k_iare the reflection coefficients obtained from LP analysis_—1.

[0119] The voiced/unvoiced decision is derived if the following conditions are met:

[0120] if P2_R1<0.6 and P1_SHP>0.2 set mode=2,

[0121] if P3_ZC>0.4 and P1_SHP>0.18 set mode=2,

[0122] if P4_RE<0.4 and P1_SHP>0.2 set mode=2,

[0123] if (P2_R1<−1.2+3.2P1_SHP) set VUV=−3

[0124] if (P4_RE<−0.21+1.4286P1_SHP) set VUV=−3

[0125] if (P3_ZC>0.8−0.6P1_SHP) set VUV=−3

[0126] if (P4_RE<0.1) set VUV=−3

[0127] Open loop pitch analysis is performed once or twice (each 10 ms) per frame depending on the coding rate in order to find estimates of the pitch lag at the block 241 (FIG. 2). It is based on the weighted speech signal s_w(n+n_m),n=0,1, . . . ,79, in which n_mdefines the location of this signal on the first half frame or the last half frame. In the first step, four maxima of the correlation:

C_{k} = \sum_{n = 0}^{79} s_{w} (n_{m} + n) s_{w} (n_{m} + n - k)

[0128] are found in the four ranges 17 . . . 33, 34 . . . 67, 68 . . . 135, 136 . . . 145, respectively. The retained maxima C_k, i=1,2,3,4, are normalized by dividing by:

[0129] {square root over (Σ_ns_w²+L (n_m+n−k+L ))}, i=1, . . . 4,

[0130] respectively.

[0131] The normalized maxima and corresponding delays are denoted by (R_i,k_i),i=1,2,3,4.

[0132] In the second step, a delay, k_I, among the four candidates, is selected by maximizing the four normalized correlations. In the third step, k_Iis probably corrected to k_i(i<I) by favoring the lower ranges. That is, k_i(i<I) is selected if k_iis within [k_I/m−4, k_I/m+4],m=2,3,4,5, and if k_i>k_I0.95^I−iD, i<I, where D is 1.0, 0.85, or 0.65, depending on whether the previous frame is unvoiced, the previous frame is voiced and k_iis in the neighborhood (specified by ±8) of the previous pitch lag, or the previous two frames are voiced and k_iis in the neighborhood of the previous two pitch lags. The final selected pitch lag is denoted by T_op.

[0133] A decision is made every frame to either operate the LTP (long-term prediction) as the traditional CELP approach (LTP_mode=1), or as a modified time warping approach (LTP_mode=0) herein referred to as PP (pitch preprocessing). For 4.55 and 5.8 kbps encoding bit rates, LTP_mode is set to 0 at all times. For 8.0 and 11.0 kbps, LTP_mode is set to 1 all of the time. Whereas, for a 6.65 kbps encoding bit rate, the encoder decides whether to operate in the LTP or PP mode. During the PP mode, only one pitch lag is transmitted per coding frame.

[0134] For 6.65 kbps, the decision algorithm is as follows. First, at the block 241, a prediction of the pitch lag pit for the current frame is determined as follows:

[0135] if (LTP_MODE_m=1)

[0136] pit=lagl1+2.4*(lag_—f[3]−lagl1);

[0137] else

[0138] pit=lag_—f[1]+2.75*(lag_—f[3]−lag_—f[1]);

[0139] where LTP_mode_m is previous frame LTP_mode, lag_f[1],lag_f[3] are the past closed loop pitch lags for second and fourth subframes respectively, lagl is the current frame open-loop pitch lag at the second half of the frame, and, lagl1 is the previous frame open-loop pitch lag at the first half of the frame.

[0140] Second, a normalized spectrum difference between the Line Spectrum Frequencies (LSF) of current and previous frame is computed as:

e_lsf = \frac{1}{10} \sum_{i = 0}^{9} abs (LSF (i) - LSF_m (i)),

[0141] if (abs(pit-lagl)<TH and abs(lag_f[3]−lagl)<lagl*0.2) if (Rp>0.5 && pgain_past>0.7 and e_lsf<0.5/30) LTP_mode=0;

[0142] else LTP_mode=1;

[0143] where Rp is current frame normalized pitch correlation, pgain_past is the quantized pitch gain from the fourth subframe of the past frame, TH=MIN(lagl*0.1, 5), and TH=MAX(2.0, TH).

[0144] The estimation of the precise pitch lag at the end of the frame is based on the normalized correlation:

R_{k} = \frac{\sum_{n = 0}^{L} s_{w} (n + n1) s_{w} (n + n1 - k)}{\sqrt{\sum_{n = 0}^{L} s_{w}^{2} (n + n1 - k)}},

[0145] where s_w(n+n1), n=0,1 . . . , L−1, represents the last segment of the weighted speech signal including the look-ahead (the look-ahead length is 25 samples), and the size L is defined according to the open-loop pitch lag T_opwith the corresponding normalized correlation C_T:

[0146] if (C_T>0.6)

[0147] L=max {50, T_op}

[0148] L=min {80, L}

[0149] else

[0150] L=80

[0151] In the first step, one integer lag k is selected maximizing the R_kin the range k ε[T_op−10, T_op+10] bounded by [17, 145]. Then, the precise pitch lag P_mand the corresponding index I_mfor the current frame is searched around the integer lag, [k−1, k+1], by up-sampling R_k.

[0152] The possible candidates of the precise pitch lag are obtained from the table named as PitLagTab8b[i], i=0,1, . . . , 127. In the last step, the precise pitch lag P_m=PitLagTab8b[I_m] is possibly modified by checking the accumulated delay τ_accdue to the modification of the speech signal:

[0153] if (τ_acc>5) I_m<=min{I_m+1, 127}, and

[0154] if (τ_acc<−5) I_m<=max{I_m−1,0}.

[0155] The precise pitch lag could be modified again:

[0156] if (τ_acc>10) I_m<=min{I_m+1, 127}, and

[0157] if (τ_acc<−10) I_m<=max{I_m−1, 0}.

[0158] The obtained index I_mwill be sent to the decoder.

[0159] The pitch lag contour, τ_c(n), is defined using both the current lag P_mand the previous lag P_m−1:

[0160] if (|P_m−P_m−1|<0.2 min{P_m, P_m−1})

[0161] τ_c(n)=P_m−1+n(P_m−P_m−1)/L_f,

[0162] n=0,1, . . . , L_f−1

[0163] τ_c(n)=P_m, n=L_f, . . . ,170

[0164] else

[0165] τ_c(n)=P_m−1, n=0,1, . . . ,39;

[0166] τ_c(n)=P_m, n=40, . . . ,170

[0167] where L_f=160 is the frame size.

[0168] One frame is divided into 3 subframes for the long-term preprocessing. For the first two subframes, the subframe size, L_s, is 53, and the subframe size for searching, L_sr, is 70. For the last subframe, L_sis 54 and L_sris:

[0169] L_sr=min {70, L_s+L_khd−10−τ_acc},

[0170] where L_khd=25 is the look-ahead and the maximum of the accumulated delay τ_accis limited to 14.

[0171] The target for the modification process of the weighted speech temporally memorized in {ŝ_w(m0+n), n=0,1, . . . , L_sr−1} is calculated by warping the past modified weighted speech buffer, ŝ_w(m0+n), n<0, with the pitch lag contour, τ_c(n+m·L_s), m=0,1,2,

{\hat{s}}_{w} (m0 + n) = \sum_{i = - f_{l}}^{f_{l}} {\hat{s}}_{w} (m0 + n - T_{c} (n) + i) I_{s} (i, T_{IC} (n)), n = 0, 1, \dots, L_{sr} - 1,

[0172] where T_C(n) and T_IC(n) are calculated by:

[0173] T_c(n)=trunc{τ_c(n+m·L_s)},

[0174] T_IC(n)=τ_c(n)−T_C(n),

[0175] m is subframe number, I_s(i, T_IC(n)) is a set of interpolation coefficients, and f_lis 10. Then, the target for matching, ŝ_t(n), n=0,1, . . . , L_sr−1, is calculated by weighting ŝ_w(m0+n), n=0,1, . . . , L_sr−1, in the time domain:

[0176] ŝ_t(n)=n·ŝ_w(m0+n)/L_s,

[0177] n=0,1, . . . , L_s−1,

[0178] ŝ_t(n)=ŝ_w(m0+n),

[0179] n=L_s, . . . , L_sr−1

[0180] The local integer shifting range [SR0, SR1] for searching for the best local delay is computed as the following:

[0181] if speech is unvoiced

[0182] SR0=−1,

[0183] SR1=1,

[0184] else

[0185] SR0=round{−4 min {1.0, max {0.0, 1−0.4 (P_sh−0.2)}}},

[0186] SR1=round{4 min {1.0, max {0.0, 1−0.4 (P_sh−0.2)}}},

[0187] where P_sh=max{P_sh1, P_sh2}, P_sh1is the average to peak ratio (i.e., sharpness) from the target signal:

P_{sh1} = \frac{\sum_{n = 0}^{L_{sr} - 1} | {\hat{s}}_{w} (m0 + n) |}{L_{sr} \max {| {\hat{s}}_{w} (m0 + n) |, n = 0, 1, \dots, L_{sr} - 1}}

[0188] and P_sh2is the sharpness from the weighted speech signal:

P_{sh2} = \frac{\sum_{n = 0}^{L_{sr} - L_{s} / 2 - 1} | s_{w} (n + n0 + L_{s} / 2) |}{\begin{matrix} (L_{sr} - L_{s} / 2) \max {| s_{w} (n + n0 + L_{s} / 2) |, \\ n = 0, 1, \dots, L_{sr} - L_{s} / 2 - 1} \end{matrix}}

[0189] where

[0190] n0=trunc{m0+τ_acc+0.5}

[0191] (here, m is subframe number and τ_accis the previous accumulated delay).

[0192] In order to find the best local delay, τ_opt, at the end of the current processing subframe, a normalized correlation vector between the original weighted speech signal and the modified matching target is defined as:

R_{I} (k) = \frac{\sum_{n = 0}^{L_{sr} - 1} s_{w} (n0 + n + k) {\hat{s}}_{t} (n)}{\sqrt{\sum_{n = 0}^{L_{sr} - 1} s_{w}^{2} (n0 + n + k) \sum_{n = 0}^{L_{sr} - 1} {\hat{s}}_{t}^{2} (n)}}

[0193] A best local delay in the integer domain, k_opt, is selected by maximizing R_I(k) in the range of k ε[SR0, SR1], which is corresponding to the real delay:

[0194] k_r=k_opt+n0−m0−τ_acc

[0195] If R_I(k_opt)<0.5, k_ris set to zero.

[0196] In order to get a more precise local delay in the range {k_r−0.75+0.1j, j=0,1 . . . 15} around k_r, R_I(k) is interpolated to obtain the fractional correlation vector, R_f(j), by:

R_{f} (j) = \sum_{i = - 7}^{8} R_{I} (k_{opt} + I_{j} + i) I_{f} (i, j), j = 0, 1, \dots, 15,

[0197] where {I_f(i,j)} is a set of interpolation coefficients. The optimal fractional delay index, j_opt, is selected by maximizing R_f(j). Finally, the best local delay, τ_opt, at the end of the current processing subframe, is given by,

[0198] τ_opt=k_r−0.75+0.1 j_opt

[0199] The local delay is then adjusted by:

τ_{opt} = {\begin{matrix} 0, & if τ_{acc} + τ_{opt} > 14 \\ τ_{opt}, & otherwise \end{matrix}

[0200] The modified weighted speech of the current subframe, memorized in {ŝ_w(m0+n), n=0,1, . . . , L_s−1} to update the buffer and produce the second target signal 253 for searching the fixed codebook 261, is generated by warping the original weighted speech {s_w(n)} from the original time region,

[0201] [m0+τ_acc, m0+τ_acc+L_s+τ_opt],

[0202] to the modified time region,

[0203] [m0, m0+L_s]:

{\hat{s}}_{w} (m0 + n) = \sum_{i = - f_{l} + 1}^{f_{l}} s_{w} (m0 + n + T_{W} (n) + i) I_{s} (i, T_{IW} (n)), n = 0, 1, \dots, L_{s} - 1,

[0204] where T_W(n) and T_IW(n) are calculated by:

[0205] T_W(n)=trunc{τ_acc+n·τ_opt/L_s},

[0206] T_IW(n)=τ_acc+n·τ_opt/L_s−T_W(n),

[0207] {I_s(i, T_IW(n))} is a set of interpolation coefficients.

[0208] After having completed the modification of the weighted speech for the current subframe, the modified target weighted speech buffer is updated as follows:

[0209] ŝ_w(n)<=ŝ_w(n+L_s),

[0210] n=0,1, . . . ,n_m−1.

[0211] The accumulated delay at the end of the current subframe is renewed by:

[0212] τ_acc<=τ_acc+τ_opt.

[0213] Prior to quantization the LSFs are smoothed in order to improve the perceptual quality. In principle, no smoothing is applied during speech and segments with rapid variations in the spectral envelope. During non-speech with slow variations in the spectral envelope, smoothing is applied to reduce unwanted spectral variations. Unwanted spectral variations could typically occur due to the estimation of the LPC parameters and LSF quantization. As an example, in stationary noise-like signals with constant spectral envelope introducing even very small variations in the spectral envelope is picked up easily by the human ear and perceived as an annoying modulation.

[0214] The smoothing of the LSFs is done as a running mean according to:

[0215] lsf_i(n)=β(n)·lsf_i(n−1)+(1−β(n))·lsf_—est_i(n),

[0216] i=1, . . . ,10

[0217] where lsf_est_i(n) is the i^thestimated LSF of frame n, and lsf_i(n) is the i^thLSF for quantization of frame n. The parameter β(n) controls the amount of smoothing, e.g. if β(n) is zero no smoothing is applied.

[0218] β(n) is calculated from the VAD information (generated at the block 235) and two estimates of the evolution of the spectral envelope. The two estimates of the evolution are defined as:

Δ SP = \sum_{i = 1}^{10} {({lsf_est}_{i} (n) - {lsf_est}_{i} (n - 1))}^{2}

{Δ SP}_{int} = \sum_{i = 1}^{10} {({lsf_est}_{i} (n) - {ma_lsf}_{i} (n - 1))}^{2}

ma_—lsf_i(n)=β(n)·ma_—lsf_i(n−1)+(1−β(n))·lsf_—est_i(n),

[0219] i=1, . . . ,10

[0220] The parameter β(n) is controlled by the following logic:

[0221] Step 1:

[0222] if (Vad=1|PastVad=1|k₁>0.5)

[0223] N_mode^frm(n−1)=0

[0224] β(n)=0.0

[0225] elseif (N_mode^frm(n−1)>0 & (ΔSP>0.0015|ΔSP_int>0.0024))

[0226] N_mode^frm(n−1)=0

[0227] β(n)=0.0

[0228] elseif (N_mode^frm(n−1)>1 & ΔSP>0.0025)

[0229] N_mode^frm(n−1)=1

[0230] endif

[0231] Step 2:

[0232] if (Vad=0 & PastVad=0)

[0233] N_mode^frm(n)=N_mode^frm(n−1)+1

[0234] if (N_mode^frm(n)>5)

[0235] N_mode^frm(n)=5

[0236] endif

[0237] β(n)=0.9/16·(N_mode^frm(n)−1)²

[0238] else

[0239] N_mode^frm(n)=N_mode^frm(n−1)

[0240] endif

[0241] where k_iis the first reflection coefficient.

[0242] In step 1, the encoder processing circuitry checks the VAD and the evolution of the spectral envelope, and performs a full or partial reset of the smoothing if required. In step 2, the encoder processing circuitry updates the counter, N_mode^frm(n), and calculates the smoothing parameter, β(n). The parameter β(n) varies between 0.0 and 0.9, being 0.0 for speech, music, tonal-like signals, and non-stationary background noise and ramping up towards 0.9 when stationary background noise occurs.

[0243] The LSFs are quantized once per 20 ms frame using a predictive multi-stage vector quantization. A minimal spacing of 50 Hz is ensured between each two neighboring LSFs before quantization. A set of weights is calculated from the LSFs, given by w_i=K|P(f_i)|^0.4where f_iis the i^thLSF value and P(f_i) is the LPC power spectrum at f_i(K is an irrelevant multiplicative constant). The reciprocal of the power spectrum is obtained by (up to a multiplicative constant):

{P (f_{i})}^{- 1} ~ {\begin{matrix} (1 - \cos (2 π f_{i}) \prod_{odd j} {[\cos (2 π f_{i}) - \cos (2 π f_{j})]}^{2} & even i \\ (1 + \cos (2 π f_{i}) \prod_{even j} {[\cos (2 π f_{i}) - \cos (2 π f_{j})]}^{2} & odd i \end{matrix}

[0244] and the power of −0.4 is then calculated using a lookup table and cubic-spline interpolation between table entries.

[0245] A vector of mean values is subtracted from the LSFs, and a vector of prediction error vector fe is calculated from the mean removed LSFs vector, using a full-matrix AR(2) predictor. A single predictor is used for the rates 5.8, 6.65, 8.0, and 11.0 kbps coders, and two sets of prediction coefficients are tested as possible predictors for the 4.55 kbps coder.

[0246] The vector of prediction error is quantized using a multi-stage VQ, with multi-surviving candidates from each stage to the next stage. The two possible sets of prediction error vectors generated for the 4.55 kbps coder are considered as surviving candidates for the first stage.

[0247] The first 4 stages have 64 entries each, and the fifth and last table have 16 entries. The first 3 stages are used for the 4.55 kbps coder, the first 4 stages are used for the 5.8, 6.65 and 8.0 kbps coders, and all 5 stages are used for the 11.0 kbps coder. The following table summarizes the number of bits used for the quantization of the LSFs for each rate.

[0248] The number of surviving candidates for each stage is summarized in the following table.

[0249] The quantization in each stage is done by minimizing the weighted distortion measure given by:

ϵ_{k} = \sum_{i = 0}^{9} {w_{i} (f e_{i} - C_{i}^{k})}^{2} .

[0250] The code vector with index k_minwhich minimizes ε_ksuch that ε_k<ε_kfor all k, is chosen to represent the prediction/quantization error (fe represents in this equation both the initial prediction error to the first stage and the successive quantization error from each stage to the next one).

[0251] The final choice of vectors from all of the surviving candidates (and for the 4.55 kbps coder—also the predictor) is done at the end, after the last stage is searched, by choosing a combined set of vectors (and predictor) which minimizes the total error. The contribution from all of the stages is summed to form the quantized prediction error vector, and the quantized prediction error is added to the prediction states and the mean LSFs value to generate the quantized LSFs vector.

[0252] For the 4.55 kbps coder, the number of order flips of the LSFs as the result of the quantization if counted, and if the number of flips is more than 1, the LSFs vector is replaced with 0.9·(LSFs of previous frame)+0.1·(mean LSFs value). For all the rates, the quantized LSFs are ordered and spaced with a minimal spacing of 50 Hz.

[0253] The interpolation of the quantized LSF is performed in the cosine domain in two ways depending on the LTP_mode. If the LTP_mode is 0, a linear interpolation between the quantized LSF set of the current frame and the quantized LSF set of the previous frame is performed to get the LSF set for the first, second and third subframes as:

[0254] {overscore (q)}₁(n)=0.75{overscore (q)}₄(n−1)+0.25{overscore (q)}₄(n)

[0255] {overscore (q)}₂(n)=0.5{overscore (q)}₄(n−1)+0.5{overscore (q)}₄(n)

[0256] {overscore (q)}₃(n)=0.25{overscore (q)}₄(n−1)+0.75{overscore (q)}₄(n)

[0257] where {overscore (q)}₄(n−1) and {overscore (q)}₄(n) are the cosines of the quantized LSF sets of the previous and current frames, respectively, and {overscore (q)}₁(n), {overscore (q)}₂(n) and {overscore (q)}₃(n) are the interpolated LSF sets in cosine domain for the first, second and third subframes respectively.

[0258] If the LTP_mode is 1, a search of the best interpolation path is performed in order to get the interpolated LSF sets. The search is based on a weighted mean absolute difference between a reference LSF set r{overscore (l)}(n) and the LSF set obtained from LP analysis_—2 {overscore (l)}(n). The weights {overscore (w)} are computed as follows:

[0259] w(0)=(1−l(0))(1−l(1)+l(0))

[0260] w(9)=(1−l(9))(1−l(9)+l(8))

[0261] for i=1 to 9

[0262] w(i)=(1−l(i))(1−Min (l(i+1)−l(i),l(i)−l(i−1)))

[0263] where Min(a,b) returns the smallest of a and b.

[0264] There are four different interpolation paths. For each path, a reference LSF set r{overscore (q)}(n) in cosine domain is obtained as follows:

[0265] r{overscore (q)}(n)=α(k){overscore (q)}₄(n)+(1−α(k)){overscore (q)}₄(n−1), k=1 to 4

[0266] {overscore (α)}={0.4,0.5,0.6, 0.7} for each path respectively. Then the following distance measure is computed for each path as:

[0267] D=|r{overscore (l)}(n)−{overscore (l)}(n)|^T{overscore (w)}

[0268] The path leading to the minimum distance D is chosen and the corresponding reference LSF set r{overscore (q)}(n) is obtained as:

[0269] r{overscore (q)}(n)=α_opt{overscore (q)}₄(n)+(1−α_opt){overscore (q)}₄(n−1)

[0270] The interpolated LSF sets in the cosine domain are then given by:

[0271] {overscore (q)}₁(n)=0.5{overscore (q)}₄(n−1)+0.5r{overscore (q)}(n)

[0272] {overscore (q)}₂(n)=r{overscore (q)}(n)

[0273] {overscore (q)}₃(n)=0.5r{overscore (q)}(n)+0.5{overscore (q)}₄(n)

[0274] The impulse response, h(n), of the weighted synthesis filter H(z)W(z)=A(z/γ₁)/[{overscore (A)}(z)A(z/γ₂)] is computed each subframe. This impulse response is needed for the search of adaptive and fixed codebooks 257 and 261. The impulse response h(n) is computed by filtering the vector of coefficients of the filter A(z/γ₁) extended by zeros through the two filters 1/{overscore (A)}(z) and 1/A(z/γ₂).

[0275] The target signal for the search of the adaptive codebook 257 is usually computed by subtracting the zero input response of the weighted synthesis filter H(z)W(z) from the weighted speech signal s_w(n). This operation is performed on a frame basis. An equivalent procedure for computing the target signal is the filtering of the LP residual signal r(n) through the combination of the synthesis filter 1/{overscore (A)}(z) and the weighting filter W(z).

[0276] After determining the excitation for the subframe, the initial states of these filters are updated by filtering the difference between the LP residual and the excitation. The LP residual is given by:

r (n) = s (n) + \sum_{i = 1}^{10} {\overline{a}}_{i} s (n - i), n = 0, L_SF - 1

[0277] The residual signal r(n) which is needed for finding the target vector is also used in the adaptive codebook search to extend the past excitation buffer. This simplifies the adaptive codebook search procedure for delays less than the subframe size of 40 samples.

[0278] In the present embodiment, there are two ways to produce an LTP contribution. One uses pitch preprocessing (PP) when the PP-mode is selected, and another is computed like the traditional LTP when the LTP-mode is chosen. With the PP-mode, there is no need to do the adaptive codebook search, and LTP excitation is directly computed according to past synthesized excitation because the interpolated pitch contour is set for each frame. When the AMR coder operates with LTP-mode, the pitch lag is constant within one subframe, and searched and coded on a subframe basis.

[0279] Suppose the past synthesized excitation is memorized in {ext(MAX_LAG+n), n<0}, which is also called adaptive codebook. The LTP excitation codevector, temporally memorized in {ext(MAX_LAG+n), 0<=n<L_SF}, is calculated by interpolating the past excitation (adaptive codebook) with the pitch lag contour, τ_c(n+m·L_SF), m=0,1,2,3. The interpolation is performed using an FIR filter (Hamming windowed sinc functions):

ext (MAX_LAG + n) = \sum_{i = - f_{l}}^{f_{l}} ext (MAX_LAG + n - T_{c} (n) + i) \cdot I_{s} (i, T_{IC} (n)), n = 0, 1, \dots, L_SF - 1, \dots

[0280] where T_C(n) and T_IC(n) are calculated by

[0281] T_c(n)=trunc{τ_c(n+m·L_—SF)},

[0282] T_IC(n)=τ_c(n)−T_C(n),

[0283] m is subframe number, {I_s(i,T_IC(n))} is a set of interpolation coefficients, f_lis 10, MAX_LAG is 145+11, and L_SF=40 is the subframe size. Note that the interpolated values {ext(MAX_LAG+n), 0<=n<L_SF−17+11} might be used again to do the interpolation when the pitch lag is small. Once the interpolation is finished, the adaptive codevector Va={v_a(n),n=0 to 39} is obtained by copying the interpolated values:

[0284] v_a(n)=ext(MAX_LAG+n),

[0285] 0<=n<L_SF

[0286] Adaptive codebook searching is performed on a subframe basis. It consists of performing closed-loop pitch lag search, and then computing the adaptive code vector by interpolating the past excitation at the selected fractional pitch lag. The LTP parameters (or the adaptive codebook parameters) are the pitch lag (or the delay) and gain of the pitch filter. In the search stage, the excitation is extended by the LP residual to simplify the closed-loop search.

[0287] For the bit rate of 11.0 kbps, the pitch delay is encoded with 9 bits for the 1^stand 3^rdsubframes and the relative delay of the other subframes is encoded with 6 bits. A fractional pitch delay is used in the first and third subframes with resolutions: ⅙ in the range [17,93{fraction (4/6)}], and integers only in the range [95,145]. For the second and fourth subframes, a pitch resolution of ⅙ is always used for the rate 11.0 kbps in the range [T₁−5{fraction (3/6)},T₁+4{fraction (3/6)}], where T₁is the pitch lag of the previous (1^stor 3^rd) subframe.

[0288] The close-loop pitch search is performed by minimizing the mean-square weighted error between the original and synthesized speech. This is achieved by maximizing the term:

R (k) = \frac{\sum_{n = 0}^{39} T_{gs} (n) y_{k} (n)}{\sqrt{\sum_{n = 0}^{39} y_{k} (n) y_{k} (n)}},

[0289] where T_gs(n) is the target signal and y_k(n) is the past filtered excitation at delay k (past excitation convoluted with h(n)). The convolution y_k(n) is computed for the first delay t_minin the search range, and for the other delays in the search range k=t_min+1, . . . ,t_max, it is updated using the recursive relation:

[0290] y_k(n)=y_k−1(n−1)+u(−)h(n),

[0291] where u(n),n=−(143+11) to 39 is the excitation buffer.

[0292] Note that in the search stage, the samples u(n),n=0 to 39, are not available and are needed for pitch delays less than 40. To simplify the search, the LP residual is copied to u(n) to make the relation in the calculations valid for all delays. Once the optimum integer pitch delay is determined, the fractions, as defined above, around that integor are tested. The fractional pitch search is performed by interpolating the normalized correlation and searching for its maximum.

[0293] Once the fractional pitch lag is determined, the adaptive codebook vector, v(n), is computed by interpolating the past excitation u(n) at the given phase (fraction). The interpolations are performed using two FIR filters (Hamming windowed sinc functions), one for interpolating the term in the calculations to find the fractional pitch lag and the other for interpolating the past excitation as previously described. The adaptive codebook gain, g_p, is temporally given then by:

g_{p} = \frac{\sum_{n = 0}^{39} T_{gs} (n) y (n)}{\sum_{n = 0}^{39} y (n) y (n)},

[0294] bounded by 0<g_p<1.2, where y(n)=v(n)*h(n) is the filtered adaptive codebook vector (zero state response of H(z)W(z) to v(n)). The adaptive codebook gain could be modified again due to joint optimization of the gains, gain normalization and smoothing. The term y(n) is also referred to herein as C_p(n).

[0295] With conventional approaches, pitch lag maximizing correlation might result in two or more times the correct one. Thus, with such conventional approaches, the candidate of shorter pitch lag is favored by weighting the correlations of different candidates with constant weighting coefficients. At times this approach does not correct the double or treble pitch lag because the weighting coefficients are not aggressive enough or could result in halving the pitch lag due to the strong weighting coefficients.

[0296] In the present embodiment, these weighting coefficients become adaptive by checking if the present candidate is in the neighborhood of the previous pitch lags (when the previous frames are voiced) and if the candidate of shorter lag is in the neighborhood of the value obtained by dividing the longer lag (which maximizes the correlation) with an integer.

[0297] In order to improve the perceptual quality, a speech classifier is used to direct the searching procedure of the fixed codebook (as indicated by the blocks 275 and 279) and to-control gain normalization (as indicated in the block 401 of FIG. 4). The speech classifier serves to improve the background noise performance for the lower rate coders, and to get a quick start-up of the noise level estimation. The speech classifier distinguishes stationary noise-like segments from segments of speech, music, tonal-like signals, non-stationary noise, etc.

[0298] The speech classification is performed in two steps. An initial classification (speech_mode) is obtained based on the modified input signal. The final classification (exc_mode) is obtained from the initial classification and the residual signal after the pitch contribution has been removed. The two outputs from the speech classification are the excitation mode, exc_mode, and the parameter β_sub(n), used to control the subframe based smoothing of the gains.

[0299] The speech classification is used to direct the encoder according to the characteristics of the input signal and need not be transmitted to the decoder. Thus, the bit allocation, codebooks, and decoding remain the same regardless of the classification. The encoder emphasizes the perceptually important features of the input signal on a subframe basis by adapting the encoding in response to such features. It is important to notice that misclassification will not result in disastrous speech quality degradations. Thus, as opposed to the VAD 235, the speech classifier identified within the block 279 (FIG. 2) is designed to be somewhat more aggressive for optimal perceptual quality.

[0300] The initial classifier (speech_classifier) has adaptive thresholds and is performed in six steps:

[0301] 1. Adapt thresholds:

[0302] if (updates_noise≧30 & updates_speech≧30)

SNR_max = \min (\frac{ma_max_speech}{ma_max_noise}, 32)

[0303] else

[0304] SNR_max=3.5

[0305] endif

[0306] if (SNR_max<1.75)

[0307] deci_max_mes=1.30

[0308] deci_ma_cp=0.70

[0309] update_max_mes=1.10

[0310] update_ma_cp_speech=0.72

[0311] elseif (SNR_max<2.50)

[0312] deci_max_mes=1.65

[0313] deci_ma_cp=0.73

[0314] update_max_mes=1.30

[0315] update_ma_cp_speech=0.72

[0316] else

[0317] deci_max_mes=1.75

[0318] deci_ma_cp=0.77

[0319] update_max_mes=1.30

[0320] update_ma_cp_speech=0.77

[0321] endif

[0322] 2. Calculate parameters:

[0323] Pitch correlation:

cp = \frac{\sum_{i = 0}^{L_SF - 1} \tilde{s} (i) \cdot \tilde{s} (i - lag)}{\sqrt{(\sum_{i = 0}^{L_SF - 1} \tilde{s} (i) \cdot \tilde{s} (i)) \cdot (\sum_{i = 0}^{L_SF - 1} \tilde{s} (i - lag) \cdot \tilde{s} (i - lag))}}

[0324] Running mean of pitch correlation:

[0325] ma_cp(n)=0.9·ma_cp(n−1)+0.1·cp

[0326] Maximum of signal amplitude in current pitch cycle:

[0327] max(n)=max{|{tilde over (s)}(i)|,i=start, . . . , L_SF−1}

[0328] where:

[0329] start=min{L_SF−lag,0}

[0330] Sum of signal amplitudes in current pitch cycle:

mean (n) = \sum_{i = start}^{L_SF - 1} | \tilde{s} (i) |

[0331] Measure of relative maximum:

max_res = \frac{\max (n)}{ma_max_noise (n - 1)}

[0332] Maximum to long-term sum:

max2sum = \frac{\max (n)}{\sum_{k = 1}^{14} mean (n - k)}

[0333] Maximum in groups of 3 subframes for past 15 subframes:

[0334] max_group(n,k)=max{max(n−3·(4−k)−j), j=0, . . . ,2}, k=0, . . . ,4

[0335] Group-maximum to minimum of previous 4 group-maxima:

endmax2minmax = \frac{max_group (n, 4)}{\min {max_group (n, k), k = 0, \dots, 3}}

[0336] Slope of 5 group maxima:

slope = 0.1 \cdot \sum_{k = 0}^{4} (k - 2) \cdot max_group (n, k)

[0337] 3. Classify subframe:

[0338] if (((max_mes<deci_max_mes & ma_cp<deci_ma_cp)|(VAD=0)) &

[0339] (LTP_MODE=1|5.8 kbit/s|4.55 kbit/s))

[0340] speech_mode=0/* class1*/

[0341] else

[0342] speech_mode=1/* class2*/

[0343] endif

[0344] 4. Check for change in background noise level, i.e. reset required:

[0345] Check for decrease in level:

[0346] if (updates_noise=31 & max_mes<=0.3)

[0347] if (consec_low<15)

[0348] consec_low++

[0349] endif

[0350] else

[0351] consec_low=0

[0352] endif

[0353] if (consec_low=15)

[0354] updates_noise=0

[0355] lev_reset=−1 /*low level reset*/

[0356] endif

[0357] Check for increase in level:

[0358] if ((updates_noise>=30|lev_reset=−1) & max_mes>1.5 & ma_cp<0.70 & cp<0.85 & k1<−0.4 & endmax2minmax<50 & max2sum<35 & slope>−100 & slope<120)

[0359] if (consec_high<15)

[0360] consec_high++

[0361] endif

[0362] else

[0363] consec_high=0

[0364] endif

[0365] if (consec_high=15 & endmax2minmax<6 & max2sum<5))

[0366] updates_noise=30

[0367] lev_reset=1 /*high level reset*/

[0368] endif

[0369] 5. Update running mean of maximum of class 1 segments, i.e. stationary noise:

[0370] if(

[0371] /*1.condition:regular update*/

[0372] (max_mes<update_max_mes & ma_cp<0.6 & cp<0.65 & max_mes>0.3)|

[0373] /*2.condition:VAD continued update*/

[0374] (consec_vad_—0=8)|

[0375] /*3.condition:start-up/reset update*/

[0376] (updates_noise≦30 & ma_cp<0.7 & cp<0.75 & k₁<−0.4 & endmax2minmax<5 &

[0377] (lev_reset≠−1|(lev_reset=−1 & max_mes<2)))

[0378] )

[0379] ma_max_noise(n)=0.9·ma_max_(n−1)+0.1·max(n)

[0380] if(updates_noise≦30)

[0381] updates_noise++

[0382] else

[0383] lev_reset=0

[0384] endif

[0385] .

[0386] .

[0387] .

[0388] where k₁is the first reflection coefficient.

[0389] 6. Update running mean of maximum of class 2 segments, i.e. speech, music, tonal-like signals, non-stationary noise, etc, continued from above:

[0390] .

[0391] .

[0392] .

[0393] elseif(ma_cp>update_ma_cp_speech)

[0394] if(updates_speech≦80)

[0395] α_speech=0.95

[0396] else

[0397] α_speech=0.999

[0398] endif

[0399] ma_max_speech(n)=α_speech·ma_max_speech(n−1)+(1−α_speech)·max(n)

[0400] if(updates_speech≦80)

[0401] updates_speech++

[0402] endif

[0403] The final classifier (exc_preselect) provides the final class, exc_mode, and the subframe based smoothing parameter, β_sub(n). It has three steps:

[0404] 1. Calculate parameters:

[0405] Maximum amplitude of ideal excitation in current subframe:

[0406] max_res2(n)=max{|res2(i)|,i=0, . . . ,L_SF−1}

[0407] Measure of relative maximum:

{max_mes}_{res2} = \frac{\max_{res2} (n)}{{ma_max}_{res2} (n - 1)}

[0408] 2. Classify subframe and calculate smoothing:

[0409] if (speech_mode=1|max_mes_res2≧1.75)

[0410] exc_mode=1 /*class2*/

[0411] β_sub(n)=0

[0412] N_mode_sub(n)=−4

[0413] else

[0414] exc_mode=0 /*class1*/

[0415] N_mode_sub(n)=N_mode_sub(n−1)+1

[0416] if (N_mode_sub(n)>4)

[0417] N_mode_sub(n)=4

[0418] endif

[0419] if (N_mode_sub(n)>0)

[0420] β_sub(n)=0.7/9·(N_mode_sub(n)−1)²

[0421] else

[0422] β_sub(n)=0

[0423] endif

[0424] endif

[0425] 3. Update running mean of maximum:

[0426] if (max_mes_res2≦0.5)

[0427] if (consec<51)

[0428] consec++

[0429] endif

[0430] else

[0431] consec=0

[0432] endif

[0433] if ((exc_mode=0 & (max_mes_res2>0.5|consec>50))|

[0434] (updates≦30 & ma_cp<0.6 & cp<0.65))

[0435] ma_max(n)=0.9 ma_max(n−1)+0.1·max_res2(n)

[0436] if (updates≦30)

[0437] updates++

[0438] endif

[0439] endif

[0440] When this process is completed, the final subframe based classification, exc_mode, and the smoothing parameter, β_sub(n), are available.

[0441] To enhance the quality of the search of the fixed codebook 261, the target signal, T_g(n), is produced by temporally reducing the LTP contribution with a gain factor, G_r:

[0442] T_g(n)=T_gs(n)−G_r*g_p*Y_a(n),

[0443] n=0,1, . . . ,39

[0444] where T_gs(n) is the original target signal 253, Y_a(n) is the filtered signal from the adaptive codebook, g_pis the LTP gain for the selected adaptive codebook vector, and the gain factor is determined according to the normalized LTP gain, R_p, and the bit rate:

[0445] if (rate<=0) /* for 4.45 kbps and 5.8 kbps*/

[0446] G_r=0.7 R_p+0.3;

[0447] if (rate==1) /* for 6.65 kbps*/

[0448] G_r=0.6 R_p+0.4;

[0449] if (rate==2) /* for 8.0 kbps*/

[0450] G_r=0.3 R_p+0.7;

[0451] if (rate==3) /* for 11.0 kbps*/

[0452] G_r=0.95;

[0453] if (T_op>L_SF & g_p>0.5 & rate<=2)

[0454] G_r<=G_r·(0.3{circumflex over ( )}R_p{circumflex over ( )}+{circumflex over ( )}0.7); and

[0455] where normalized LTP gain, R_p, is defined as:

R_{p} = \frac{\sum_{n = 0}^{39} T_{gs} (n) Y_{a} (n)}{\sqrt{\sum_{n = 0}^{39} T_{gs} (n) T_{gs} (n)} \sqrt{\sum_{n = 0}^{39} Y_{a} (n) Y_{a} (n)}}

[0456] Another factor considered at the control block 275 in conducting the fixed codebook search and at the block 401 (FIG. 4) during gain normalization is the noise level+“)” which is given by:

P_{NSR} = \sqrt{\frac{\max {(E_{n} - 100), 0.0}}{E_{s}}}

[0457] where E_sis the energy of the current input signal including background noise, and E_nis a running average energy of the background noise. E_nis updated only when the input signal is detected to be background noise as follows:

[0458] if (first background noise frame is true)

[0459] E_n=0.75 E_s;

[0460] else if (background noise frame is true)

[0461] E_n=0.75 E_n^m+0.25 E_s;

[0462] where E_n^mis the last estimation of the background noise energy.

[0463] For each bit rate mode, the fixed codebook 261 (FIG. 2) consists of two or more subcodebooks which are constructed with different structure. For example, in the present embodiment at higher rates, all the subcodebooks only contain pulses. At lower bit rates, one of the subcodebooks is populated with Gaussian noise. For the lower bit-rates (e.g., 6.65, 5.8, 4.55 kbps), the speech classifier forces the encoder to choose from the Gaussian subcodebook in case of stationary noise-like subframes, exc_mode=0. For exc_mode=1 all subcodebooks are searched using adaptive weighting.

[0464] For the pulse subcodebooks, a fast searching approach is used to choose a subcodebook and select the code word for the current subframe. The same searching routine is used for all the bit rate modes with different input parameters.

[0465] In particular, the long-term enhancement filter, F_p(z), is used to filter through the selected pulse excitation. The filter is defined as F_p(z)=1/_{(1−β z}₎, where T is the integer part of pitch lag at the center of the current subframe, and β is the pitch gain of previous subframe, bounded by [0.2, 1.0]. Prior to the codebook search, the impulsive response h(n) includes the filter F_p(z).

[0466] For the Gaussian subcodebooks, a special structure is used in order to bring down the storage requirement and the computational complexity. Furthermore, no pitch enhancement is applied to the Gaussian subcodebooks.

[0467] There are two kinds of pulse subcodebooks in the present AMR coder embodiment. All pulses have the amplitudes of +1 or −1. Each pulse has 0, 1, 2, 3 or 4 bits to code the pulse position. The signs of some pulses are transmitted to the decoder with one bit coding one sign. The signs of other pulses are determined in a way related to the coded signs and their pulse positions.

[0468] In the first kind of pulse subcodebook, each pulse has 3 or 4 bits to code the pulse position. The possible locations of individual pulses are defined by two basic non-regular tracks and initial phases:

[0469] POS(n_p,i)=TRACK(m_p,i)+PHAS(n_p, phas_mode),

[0470] where i=0,1, . . . ,7 or 15 (corresponding to 3 or 4 bits to code the position), is the possible position index, n_p=0, . . . ,N_p−1 (N_pis the total number of pulses), distinguishes different pulses, m_p=0 or defines two tracks, and phase_mode=0 or 1, specifies two phase modes.

[0471] For 3 bits to code the pulse position, the two basic tracks are:

[0472] {TRACK(0,i)}={0, 4, 8, 12, 18, 24, 30, 36}, and

[0473] {TRACK(1,i)}={0, 6, 12, 18, 22, 26, 30, 34}.

[0474] If the position of each pulse is coded with 4 bits, the basic tracks are:

[0475] {TRACK(0,i)}={0, 2, 4, 6, 8, 10, 12, 14, 17, 20, 23, 26, 29, 32, 35, 38}, and

[0476] {TRACK(1,i)}={0, 3, 6, 9, 12, 15, 18, 21, 23, 25, 27, 29, 31, 33, 35, 37}.

[0477] The initial phase of each pulse is fixed as:

[0478] PHAS(n_p,0)=modulus(n_p/MAXPHAS)

[0479] PHAS(n_p,1)=PHAS(N_p−1−n_p, 0)

[0480] where MAXPHAS is the maximum phase value.

[0481] For any pulse subcodebook, at least the first sign for the first pulse, SIGN(n_p), n_p=0, is encoded because the gain sign is embedded. Suppose N_signis the number of pulses with encoded signs; that is, SIGN(n_p), for n_p<N_sign,<=N_p, is encoded while SIGN(n_p), for n_p>=N_sign, is not encoded. Generally, all the signs can be determined in the following way:

[0482] SIGN(n_p)=−SIGN(n_p−1),

[0483] for n_p>=N_sign,

[0484] due to that the pulse positions are sequentially searched from n_p=0 to n_p=N_p−1 using an iteration approach. If two pulses are located in the same track while only the sign of the first pulse in the track is encoded, the sign of the second pulse depends on its position relative to the first pulse. If the position of the second pulse is smaller, then it has opposite sign, otherwise it has the same sign as the first pulse.

[0485] In the second kind of pulse subcodebook, the innovation vector contains 10 signed pulses. Each pulse has 0, 1, or 2 bits to code the pulse position. One subframe with the size of 40 samples is divided into 10 small segments with the length of 4 samples. 10 pulses are respectively located into 10 segments. Since the position of each pulse is limited into one segment, the possible locations for the pulse numbered with n_pare, {4n_p}, {4n_p, 4n_p+2}, or {4n_p, 4n_p30 1, 4n_p+2, 4n_p+3}, respectively for 0, 1, or 2 bits to code the pulse position. All the signs for all the 10 pulses are encoded.

[0486] The fixed codebook 261 is searched by minimizing the mean square error between the weighted input speech and the weighted synthesized speech. The target signal used for the LTP excitation is updated by subtracting the adaptive codebook contribution. That is:

[0487] x₂(n)=x(n)−ĝ_py(n),

[0488] n=0, . . . ,39,

[0489] where y(n)=v(n)*h(n) is the filtered adaptive codebook vector and ĝ_pis the modified (reduced) LTP gain.

[0490] If c_kis the code vector at index k from the fixed codebook, then the pulse codebook is searched by maximizing the term:

A_{k} = \frac{{(C_{k})}^{2}}{E_{D_{k}}} = \frac{{(d^{t} c_{k})}^{2}}{c_{k}^{t} Φ c_{k}},

[0491] where d=H^tx₂is the correlation between the target signal x₂(n) and the impulse response h(n), H is a the lower triangular Toepliz convolution matrix with diagonal h(0) and lower diagonals h(1), . . . ,h(39), and Φ=H^tH is the matrix of correlations of h(n). The vector d (backward filtered target) and the matrix Φ are computed prior to the codebook search. The elements of the vector d are computed by:

d (n) = \sum_{i = n}^{39} x_{2} (i) h (i - n), n = 0, \dots, 39,

[0492] and the elements of the symmetric matrix Φ are computed by:

φ (i, j) = \sum_{n = j}^{39} h (n - i) h (n - j), (j \geq i) .

[0493] The correlation in the numerator is given by:

C = \sum_{i = 0}^{N_{p} - 1} ϑ_{i} d (m_{i}),

[0494] where m_iis the position of the i th pulse and υ_iis its amplitude. For the complexity reason, all the amplitudes {υ_i} are set to +1 or −1; that is,

[0495] υ_i=SIGN(i),

[0496] i=n_p=0, . . . ,N_p−1.

[0497] The energy in the denominator is given by:

E_{D} \sum_{i = 0}^{N_{p} - 1} φ (m_{i}, m_{i}) + 2 \sum_{i = 0}^{N_{p} - 2} \sum_{j = i + 1}^{N_{p} - 1} ϑ_{i} ϑ_{j} φ (m_{i}, m_{j}) .

[0498] To simplify the search procedure, the pulse signs are preset by using the signal b(n), which is a weighted sum of the normalized d(n) vector and the normalized target signal of x₂(n) in the residual domain res₂(n):

b (n) = \frac{{res}_{2} (n)}{\sqrt{\sum_{i = 0}^{39} {res}_{2} (i) {res}_{2} (i)}} + \frac{2 d (n)}{\sqrt{\sum_{i = 0}^{39} d (i) d (i)}}, n = 0, 1, \dots, 39

[0499] If the sign of the i th (i=n_p) pulse located at m_iis encoded, it is set to the sign of signal b(n) at that position, i.e., SIGN(i)=sign[b(m_i)].

[0500] In the present embodiment, the fixed codebook 261 has 2 or 3 subcodebooks for each of the encoding bit rates. Of course many more might be used in other embodiments. Even with several subcodebooks, however, the searching of the fixed codebook 261 is very fast using the following procedure. In a first searching turn, the encoder processing circuitry searches the pulse positions sequentially from the first pulse (n_p=0) to the last pulse (n_p=N_p−1) by considering the influence of all the existing pulses.

[0501] In a second searching turn, the encoder processing circuitry corrects each pulse position sequentially from the first pulse to the last pulse by checking the criterion value A_kcontributed from all the pulses for all possible locations of the current pulse. In a third turn, the functionality of the second searching turn is repeated a final time. Of course further turns may be utilized if the added complexity is not prohibitive.

[0502] The above searching approach proves very efficient, because only one position of one pulse is changed leading to changes in only one term in the criterion numerator C and few terms in the criterion denominator E_Dfor each computation of the A_k. As an example, suppose a pulse subcodebook is constructed with 4 pulses and 3 bits per pulse to encode the position. Only 96 (4 pulses×2³positions per pulse×3turns=96) simplified computations of the criterion A_kneed be performed.

[0503] Moreover, to save the complexity, usually one of the subcodebooks in the fixed codebook 261 is chosen after finishing the first searching turn. Further searching turns are done only with the chosen subcodebook. In other embodiments, one of the subcodebooks might be chosen only after the second searching turn or thereafter should processing resources so permit.

[0504] The Gaussian codebook is structured to reduce the storage requirement and the computational complexity. A comb-structure with two basis vectors is used. In the comb-structure, the basis vectors are orthogonal, facilitating a low complexity search. In the AMR coder, the first basis vector occupies the even sample positions, (0,2, . . . ,38), and the second basis vector occupies the odd sample positions, (1,3, . . . ,39).

[0505] The same codebook is used for both basis vectors, and the length of the codebook vectors is 20 samples (half the subframe size).

[0506] All rates (6.65, 5.8 and 4.55 kbps) use the same Gaussian codebook. The Gaussian codebook, CB_Gauss, has only 10 entries, and thus the storage requirement is 10·20=200 16-bit words. From the 10 entries, as many as 32 code vectors are generated. An index, idx_δ, to one basis vector 22 populates the corresponding part of a code vector, c_idx, in the following way:

[0507] c_idx(2(i−τ)+δ)=CB_Gauss(l,i)

[0508] i=τ,τ+1, . . . ,19

[0509] c_idx(2·(i+20−τ)+δ)=CB_Gauss(l,i)

[0510] i=0,1, . . . , τ−1

[0511] where the table entry, l, and the shift, iτ, are calculated from the index, dx_δ, according to:

[0512] τ=trunc{idx_δ/10}

[0513] l=idx_δ−10·τ

[0514] and δ is 0 for the first basis vector and 1 for the second basis vector. In addition, a sign is applied to each basis vector.

[0515] Basically, each entry in the Gaussian table can produce as many as 20 unique vectors, all with the same energy due to the circular shift. The 10 entries are all normalized to have identical energy of 0.5, i.e.,

\sum_{i = 0}^{19} {({CB}_{Gauss} (l, i))}^{2} = 0.5, l = 0, 1, \dots, 9

[0516] That means that when both basis vectors have been selected, the combined code vector, c_idx_,idx, will have unity energy, and thus the final excitation vector from the Gaussian subcodebook will have unity energy since no pitch enhancement is applied to candidate vectors from the Gaussian subcodebook.

[0517] The search of the Gaussian codebook utilizes the structure of the codebook to facilitate a low complexity search. Initially, the candidates for the two basis vectors are searched independently based on the ideal excitation, res₂. For each basis vector, the two best candidates, along with the respective signs, are found according to the mean squared error. This is exemplified by the equations to find the best candidate, index idx_δ, and its sign, s_idx:

{idx}_{δ} = \max_{k = 0, 1, \dots, N_{Gauss}} {| \sum_{i = 0}^{19} {res}_{2} (2 \cdot i + δ) \cdot c_{k} (2 \cdot i + δ) |}

s_{{idx}_{δ}} = sign (\sum_{i = 0}^{19} {res}_{2} (2 \cdot i + δ) \cdot c_{{idx}_{δ}} (2 \cdot i + δ))

[0518] where N_Gaussis the number of candidate entries for the basis vector. The remaining parameters are explained above. The total number of entries in the Gaussian codebook is 2·2·N_Gauss². The fine search minimizes the error between the weighted speech and the weighted synthesized speech considering the possible combination of candidates for the two basis vectors from the pre-selection. If c_k0,k1is the Gaussian code vector from the candidate vectors represented by the indices k₀and k₁and the respective signs for the two basis vectors, then the final Gaussian code vector is selected by maximizing the term:

A_{k_{0}, k_{1}} = \frac{{(C_{k_{0}, k_{1}})}^{2}}{E_{D_{k_{0}, k_{1}}}} = \frac{{(d^{t} c_{k_{0}, k_{1}})}^{2}}{c_{k_{0}, k_{1}}^{t} Φ c_{k_{0}, k_{1}}}

[0519] over the candidate vectors. d=H^tx₂is the correlation between the target signal x₂(n) and the impulse response h(n) (without the pitch enhancement), and H is a the lower triangular Toepliz convolution matrix with diagonal h(0) and lower diagonals h(1), . . . ,h(39), and Φ=H^tH is the matrix of correlations of h(n).

[0520] More particularly, in the present embodiment, two subcodebooks are included (or utilized) in the fixed codebook 261 with 31 bits in the 11 kbps encoding mode. In the first subcodebook, the innovation vector contains 8 pulses. Each pulse has 3 bits to code the pulse position. The signs of 6 pulses are transmitted to the decoder with 6 bits. The second subcodebook contains innovation vectors comprising 10 pulses. Two bits for each pulse are assigned to code the pulse position which is limited in one of the 10 segments. Ten bits are spent for 10 signs of the 10 pulses. The bit allocation for the subcodebooks used in the fixed codebook 261 can be summarized as follows:

[0521] Subcodebook1: 8 pulses×3 bits/pulse+6 signs=30 bits

[0522] Subcodebook2: 10 pulses×2 bits/pulse+10 signs=30 bits

[0523] One of the two subcodebooks is chosen at the block 275 (FIG. 2) by favoring the second subcodebook using adaptive weighting applied when comparing the criterion value F1 from the first subcodebook to the criterion value F2 from the second subcodebook:

[0524] if (W_c·F1>F2),

[0525] the first subcodebook is chosen, else, the second subcodebook is chosen,

[0526] where the weighting, 0<W_c<=1, is defined as:

W_{c} = {\begin{matrix} 1.0, & if P_{NSR} < 0.5, \\ 1.0 - 0.3 P_{NSR} (1.0 - 0.5 R_{p}) \cdot \min {P_{sharp} + 0.5, 1.0}, \end{matrix}

[0527] P_NSRis the background noise to speech signal ratio (i.e., the “noise level” in the block 279), R_pis the normalized LTP gain, and P_sharpis the sharpness parameter of the ideal excitation res₂(n) (i.e., the “sharpness” in the block 279).

[0528] In the 8 kbps mode, two subcodebooks are included in the fixed codebook 261 with 20 bits. In the first subcodebook, the innovation vector contains 4 pulses. Each pulse has 4 bits to code the pulse position. The signs of 3 pulses are transmitted to the decoder with 3 bits. The second subcodebook contains innovation vectors having 10 pulses. One bit for each of 9 pulses is assigned to code the pulse position which is limited in one of the 10 segments. Ten bits are spent for 10 signs of the 10 pulses. The bit allocation for the subcodebook can be summarized as the following:

[0529] Subcodebook1: 4 pulses×4 bits/pulse+3 signs=19 bits

[0530] Subcodebook2: 9 pulses×1 bits/pulse+1 pulse×0 bit+10 signs=19 bits

[0531] One of the two subcodebooks is chosen by favoring the second subcodebook using adaptive weighting applied when comparing the criterion value F1 from the first subcodebook to the criterion value F2 from the second subcodebook as in the 11 kbps mode. The weighting, 0<W_c<=1, is defined as:

[0532] W_c=1.0−0.6 P_NSR(1.0−0.5 R_p)·min {P_sharp+0.5, 1.0}.

[0533] The 6.65 kbps mode operates using the long-term preprocessing (PP) or the traditional LTP. A pulse subcodebook of 18 bits is used when in the PP-mode. A total of 13 bits are allocated for three subcodebooks when operating in the LTP-mode. The bit allocation for the subcodebooks can be summarized as follows:

[0534] PP-mode:

[0535] Subcodebook: 5 pulses×3 bits/pulse+3 signs=18 bits

[0536] LTP-mode:

[0537] Subcodebook1: 3 pulses×3 bits/pulse+3 signs=12 bits, phase_mode=1,

[0538] Subcodebook2: 3 pulses×3 bits/pulse+2 signs=11 bits, phase_mode=0,

[0539] Subcodebook3: Gaussian subcodebook of 11 bits.

[0540] One of the 3 subcodebooks is chosen by favoring the Gaussian subcodebook when searching with LTP-mode. Adaptive weighting is applied when comparing the criterion value from the two pulse subcodebooks to the criterion value from the Gaussian subcodebook. The weighting, 0<W_c<=1, is defined as:

[0541] W_c=1.0−0.9 P_NSR(1.0−0.5 R_p)·min {P_sharp+0.5, 1.0},

[0542] if (noise−like unvoiced),

[0543] W_c<=W_c·(0.2 R_p(1.0−P_sharp)+0.8)

[0544] The 5.8 kbps encoding mode works only with the long-term preprocessing (PP). Total 14 bits are allocated for three subcodebooks. The bit allocation for the subcodebooks can be summarized as the following:

[0545] Subcodebook1: 4 pulses×3 bits/pulse+1 signs=13 bits, phase_mode=1,

[0546] Subcodebook2: 3 pulses×3 bits/pulse+3 signs=12 bits, phase_mode=0,

[0547] Subcodebook3: Gaussian subcodebook of 12 bits.

[0548] One of the 3 subcodebooks is chosen favoring the Gaussian subcodebook with adaptive weighting applied when comparing the criterion value from the two pulse subcodebooks to the criterion value from the Gaussian subcodebook. The weighting, 0<W_c<=1, is defined as:

[0549] W_c=1.0−P_NSR(1.0−0.5R_p)·min {P_sharp+0.6,1.0},

[0550] if (noise−likeunvoiced),

[0551] W_c<=W_c·(0.3R_p(1.0−P_sharp)+0.7).

[0552] The 4.55 kbps bit rate mode works only with the long-term preprocessing (PP). Total 10 bits are allocated for three subcodebooks. The bit allocation for the subcodebooks can be summarized as the following:

[0553] Subcodebook1: 2 pulses×4 bits/pulse+1 signs=9 bits, phase_mode=1,

[0554] Subcodebook2: 2 pulses×3 bits/pulse+2 signs=8 bits, phase_mode=0,

[0555] Subcodebook3: Gaussian subcodebook of 8 bits.

[0556] One of the 3 subcodebooks is chosen by favoring the Gaussian subcodebook with weighting applied when comparing the criterion value from the two pulse subcodebooks to the criterion value from the Gaussian subcodebook. The weighting, 0<W_c<=1, is defined as:

[0557] W_c=1.0−1.2 P_NSR(1.0−0.5 R_p)·min {P_sharp+0.6, 1.0},

[0558] if (noise−like unvoiced),

[0559] W_c<=W_c·(0.6 R_p(1.0−P_sharp)+0.4).

[0560] For 4.55, 5.8, 6.65 and 8.0 kbps bit rate encoding modes, a gain re-optimization procedure is performed to jointly optimize the adaptive and fixed codebook gains, g_pand g_c, respectively, as indicated in FIG. 3. The optimal gains are obtained from the following correlations given by:

g_{p} = \frac{R_{1} R_{2} - R_{3} R_{4}}{R_{5} R_{2} - R_{3} R_{3}}

g_{c} = \frac{R_{4} - g_{p} R_{3}}{R_{2}},

[0561] where R₁=<{overscore (C)}_p,{overscore (T)}_gs>, R₂=<{overscore (C)}_c,{overscore (C)}_c>, R₃=<{overscore (C)}_p, {overscore (C)}_c>R₄=<{overscore (C)}_c,{overscore (T)}_gs>, and R₅=<{overscore (C)}_p,{overscore (C)}_p>·{overscore (C)}_c, {overscore (C)}_p, and {overscore (T)}_gsare filtered fixed codebook excitation, filtered adaptive codebook excitation and the target signal for the adaptive codebook search.

[0562] For 11 kbps bit rate encoding, the adaptive codebook gain, g_p, remains the same as that computed in the closeloop pitch search. The fixed codebook gain, g_c, is obtained as:

g_{c} = \frac{R_{6}}{R_{2}},

[0563] where R₆=<{overscore (C)}_c,{overscore (T)}_g> and {overscore (T)}_g={overscore (T)}_gs−g_p{overscore (C)}_p.

[0564] Original CELP algorithm is based on the concept of analysis by synthesis (waveform matching). At low bit rate or when coding noisy speech, the waveform matching becomes difficult so that the gains are up-down, frequently resulting in unnatural sounds. To compensate for this problem, the gains obtained in the analysis by synthesis close-loop sometimes need to be modified or normalized.

[0565] There are two basic gain normalization approaches. One is called open-loop approach which normalizes the energy of the synthesized excitation to the energy of the unquantized residual signal. Another one is close-loop approach with which the normalization is done considering the perceptual weighting. The gain normalization factor is a linear combination of the one from the close-loop approach and the one from the open-loop approach; the weighting coefficients used for the combination are controlled according to the LPC gain.

[0566] The decision to do the gain normalization is made if one of the following conditions is met: (a) the bit rate is 8.0 or 6.65 kbps, and noise-like unvoiced speech is true; (b) the noise level P_NSRis larger than 0.5; (c) the bit rate is 6.65 kbps, and the noise level P_NSRis larger than 0.2; and (d) the bit rate is 5.8 or 4.45 kbps.

[0567] The residual energy, E_res, and the target signal energy, E_Tgs, are defined respectively as:

E_{res} = \sum_{n = 0}^{L_SF - 1} {res}^{2} (n)

E_{Tgs} = \sum_{n = 0}^{L_SF - 1} T_{gs}^{2} (n)

[0568] Then the smoothed open-loop energy and the smoothed closed-loop energy are evaluated by:

[0569] if (first subframe is true)

[0570] Ol_Eg=E_res

[0571] else

[0572] Ol_—Eg<=β_sub·Ol_—Eg+(1−β_sub)E_res

[0573] if (first subframe is true)

[0574] Cl_Eg=E_Tgs

[0575] else

[0576] Cl_—Eg<=β_sub·Cl_—Eg+(1−β_sub)E_Tgs

[0577] where β_subis the smoothing coefficient which is determined according to the classification. After having the reference energy, the open-loop gain normalization factor is calculated:

ol_g = MIN {C_{ol} \sqrt{\frac{Ol_Eg}{\sum_{n = 0}^{L_SF - 1} v^{2} (n)}}, \frac{1.2}{g_{p}}}

[0578] where C_olis 0.8 for the bit rate 11.0 kbps, for the other rates C_olis 0.7, and v(n) is the excitation:

[0579] v(n)=v_a(n)g_p+v_c(n)g_c,

[0580] n=0,1, . . . ,L_SF−1.

[0581] where g_pand g_care unquantized gains. Similarly, the closed-loop gain normalization factor is:

Cl_g = MIN {C_{cl} \sqrt{\frac{Cl_Eg}{\sum_{n = 0}^{L_SF - 1} y^{2} (n)}}, \frac{1.2}{g_{p}}}

[0582] where C_clis 0.9 for the bit rate 11.0 kbps, for the other rates C_clis 0.8, and y(n) is the filtered signal (y(n)=v(n)*h(n)):

[0583] y(n)=y_a(n)g_p+y_c(n)g_c,

[0584] n=0,1, . . . ,L_SF−1.

[0585] The final gain normalization factor, g_f, is a combination of Cl_g and Ol_g, controlled in terms of an LPC gain parameter, C_LPC,

[0586] if (speech is true or the rate is 11 kbps)

[0587] g_f=C_LPCOl_g+(1−C_LPC) Cl_g

[0588] g_f=MAX(1.0, g_f)

[0589] g_f=MIN(g_f, 1+C_LPC)

[0590] if (background noise is true and the rate is smaller than 11 kbps)

[0591] g_f=1.2 MIN{Cl_g, Ol_g}

[0592] where C_LPCis defined as:

[0593] C_LPC=MIN{sqrt(E_res/E_Tgs), 0.8}/0.8

[0594] Once the gain normalization factor is determined, the unquantized gains are modified:

[0595] g_p<=g_p·g_f

[0596] For 4.55 ,5.8, 6.65 and 8.0 kbps bit rate encoding, the adaptive codebook gain and the fixed codebook gain are vector quantized using 6 bits for rate 4.55 kbps and 7 bits for the other rates. The gain codebook search is done by minimizing the mean squared weighted error, Err, between the original and reconstructed speech signals:

[0597] Err=∥{overscore (T)}_gs−g_p{overscore (C)}_p−g_c{overscore (C)}_c∥².

[0598] For rate 11.0 kbps, scalar quantization is performed to quantize both the adaptive codebook gain, g_p, using 4 bits and the fixed codebook gain, g_c, using 5 bits each.

[0599] The fixed codebook gain, g_c, is obtained by MA prediction of the energy of the scaled fixed codebook excitation in the following manner. Let E(n) be the mean removed energy of the scaled fixed codebook excitation in (dB) at subframe n be given by:

E (n) = 10 \log (\frac{1}{40} g_{c}^{2} \sum_{i = 0}^{39} c^{2} (i)) - \overline{E},

[0600] where c(i) is the unscaled fixed codebook excitation, and {overscore (E)}=30 dB is the mean energy of scaled fixed codebook excitation.

[0601] The predicted energy is given by:

\tilde{E} (n) = \sum_{i = 1}^{4} b_{i} \hat{R} (n - i)

[0602] where [b₁b₂b₃b₄]=[0.68 0.58 0.34 0.19] are the MA prediction coefficients and {circumflex over (R)}(n) is the quantized prediction error at subframe n.

[0603] The predicted energy is used to compute a predicted fixed codebook gain g_c(by substituting E(n) by {tilde over (E)}(n) and g_cby g_c). This is done as follows. First, the mean energy of the unscaled fixed codebook excitation is computed as:

E_{i} = 10 \log (\frac{1}{40} \sum_{i = 0}^{39} c^{2} (i)),

[0604] and then the predicted gain g_cis obtained as:

[0605] g_c=10^{(0.05({tilde over (E)}(n)+{overscore (E)}−E}⁾.

[0606] A correction factor between the gain, g_c, and the estimated one, g_c, is given by:

[0607] γ=^g/_g.

[0608] It is also related to the prediction error as:

[0609] R(n)=E(n)−{tilde over (E)}(n)=20 log γ.

[0610] The codebook search for 4.55, 5.8, 6.65 and 8.0 kbps encoding bit rates consists of two steps. In the first step, a binary search of a single entry table representing the quantized prediction error is performed. In the second step, the index Index_—1 of the optimum entry that is closest to the unquantized prediction error in mean square error sense is used to limit the search of the two-dimensional VQ table representing the adaptive codebook gain and the prediction error. Taking advantage of the particular arrangement and ordering of the VQ table, a fast search using few candidates around the entry pointed by Index_—1 is performed. In fact, only about half of the VQ table entries are tested to lead to the optimum entry with Index_—2. Only Index_—2 is transmitted.

[0611] For 11.0 kbps bit rate encoding mode, a full search of both scalar gain codebooks are used to quantize g_pand g_c. For g_p, the search is performed by minimizing the error Err=abs(g_p−{overscore (g)}_p). Whereas for g_c, the search is performed by minimizing the error Err=∥{overscore (T)}_gs−{overscore (g)}_p{overscore (C)}_p−g_c{overscore (C)}_c∥².

[0612] An update of the states of the synthesis and weighting filters is needed in order to compute the target signal for the next subframe. After the two gains are quantized, the excitation signal, u(n), in the present subframe is computed as:

[0613] u(n)={overscore (g)}_pv(n)+{overscore (g)}_cc(n),

[0614] n=0,39,

[0615] where {overscore (g)}_pand {overscore (g)}_care the quantized adaptive and fixed codebook gains respectively, v(n) the adaptive codebook excitation (interpolated past excitation), and c(n) is the fixed codebook excitation. The state of the filters can be updated by filtering the signal r(n)−u(n) through the filters 1/{overscore (A)}(z) and W(z) for the 40-sample subframe and saving the states of the filters. This would normally require 3 filterings.

[0616] A simpler approach which requires only one filtering is as follows. The local synthesized speech at the encoder, ŝ(n), is computed by filtering the excitation signal through 1/{overscore (A)}(z). The output of the filter due to the input r(n)−u(n) is equivalent to e(n)=s(n)−ŝ(n), so the states of the synthesis filter 1/{overscore (A)}(z) are given by e(n),n=0,39. Updating the states of the filter W(z) can be done by filtering the error signal e(n) through this filter to find the perceptually weighted error e_w(n). However, the signal e_w(n) can be equivalently found by:

[0617] e_w(n)=T_gs(n)−{overscore (g)}_pC_p(n)−{overscore (g)}_cC_c(n).

[0618] The states of the weighting filter are updated by computing e_w(n) for n=30 to 39.

[0619] The function of the decoder consists of decoding the transmitted parameters (dLP parameters, adaptive codebook vector and its gain, fixed codebook vector and its gain) and performing synthesis to obtain the reconstructed speech. The reconstructed speech is then postfiltered and upscaled.

[0620] The decoding process is performed in the following order. First, the LP filter parameters are encoded. The received indices of LSF quantization are used to reconstruct the quantized LSF vector. Interpolation is performed to obtain 4 interpolated LSF vectors (corresponding to 4 subframes). For each subframe, the interpolated LSF vector is converted to LP filter coefficient domain, a_k, which is used for synthesizing the reconstructed speech in the subframe.

[0621] For rates 4.55, 5.8 and 6.65 (during PP_mode) kbps bit rate encoding modes, the received pitch index is used to interpolate the pitch lag across the entire subframe. The following three steps are repeated for each subframe:

[0622] 1) Decoding of the gains: for bit rates of 4.55, 5.8, 6.65 and 8.0 kbps, the received index is used to find the quantized adaptive codebook gain, {overscore (g)}_p, from the 2-dimensional VQ table. The same index is used to get the fixed codebook gain correction factor {overscore (γ)} from the same quantization table. The quantized fixed codebook gain, {overscore (g)}_c, is obtained following these steps:

[0623] the predicted energy is computed

\tilde{E} (n) = \sum_{i = 1}^{4} b_{i} \hat{R} (n - i);

[0624] the energy of the unscaled fixed codebook excitation is calculated as

E_{i} = 10 \log (\frac{1}{40} \sum_{i = 0}^{39} c^{2} (i));

[0625] and

[0626] the predicted gain g_cis obtained as g_c=10^{(0.05({tilde over (E)}(n)+{overscore (E)}−E}⁾.

[0627] The quantized fixed codebook gain is given as {overscore (g)}_c={overscore (γ)}g_c. For 11 kbps bit rate, the received adaptive codebook gain index is used to readily find the quantized adaptive gain, {overscore (g)}_pfrom the quantization table. The received fixed codebook gain index gives the fixed codebook gain correction factor γ. The calculation of the quantized fixed codebook gain, {overscore (g)}_cfollows the same steps as the other rates.

[0628] 2) Decoding of adaptive codebook vector: for 8.0 , 1.0 and 6.65 (during LTP_mode=1) kbps bit rate encoding modes, the received pitch index (adaptive codebook index) is used to find the integer and fractional parts of the pitch lag. The adaptive codebook v(n) is found by interpolating the past excitation u(n) (at the pitch delay) using the FIR filters.

[0629] 3) Decoding of fixed codebook vector: the received codebook indices are used to extract the type of the codebook (pulse or Gaussian) and either the amplitudes and positions of the excitation pulses or the bases and signs of the Gaussian excitation. In either case, the reconstructed fixed codebook excitation is given as c(n). If the integer part of the pitch lag is less than the subframe size 40 and the chosen excitation is pulse type, the pitch sharpening is applied. This translates into modifying c(n) as c(n)=c(n)+βc(n−T), where β is the decoded pitch gain {overscore (g)}_pfrom the previous subframe bounded by [0.2,1.0].

[0630] The excitation at the input of the synthesis filter is given by u(n)={overscore (g)}_pv(n)+{overscore (g)}_cc(n),n=0,39. Before the speech synthesis, a post-processing of the excitation elements is performed. This means that the total excitation is modified by emphasizing the contribution of the adaptive codebook vector:

\overline{u} (n) = {\begin{matrix} u (n) + 0.25 β {\overline{g}}_{p} v (n), & {\overline{g}}_{p} > 0.5 \\ u (n), & {\overline{g}}_{p} <= 0.5 \end{matrix}

[0631] Adaptive gain control (AGC) is used to compensate for the gain difference between the unemphasized excitation u(n) and emphasized excitation {overscore (u)}(n). The gain scaling factor η for the emphasized excitation is computed by:

η = {\begin{matrix} \sqrt{\frac{\sum_{n = 0}^{39} u^{2} (n)}{\sum_{n = 0}^{39} {\overline{u}}^{2} (n)}} & {\overline{g}}_{p} > 0.5 \\ 1.0 & {\overline{g}}_{p} <= 0.5 \end{matrix}

[0632] The gain-scaled emphasized excitation {overscore (u)}(n) is given by:

[0633] {overscore (u)}(n)=η{overscore (u)}(n).

[0634] The reconstructed speech is given by:

\overline{s} (n) = \overline{u} (n) - \sum_{i = 1}^{10} {\overline{a}}_{i} \overline{s} (n - i), n = 0 to 39,

[0635] where {overscore (a)}_iare the interpolated LP filter coefficients. The synthesized speech {overscore (s)}(n) is then passed through an adaptive postfilter.

[0636] Post-processing consists of two functions: adaptive postfiltering and signal up-scaling. The adaptive postfilter is the cascade of three filters: a formant postfilter and two tilt compensation filters. The postfilter is updated every subframe of 5 ms. The formant postfilter is given by:

H_{f} (z) = \frac{\overline{A} (\frac{z}{γ_{n}})}{\overline{A} (\frac{z}{γ_{d}})}

[0637] where {overscore (A)}(z) is the received quantized and interpolated LP inverse filter and γ_nand γ_dcontrol the amount of the formant postfiltering.

[0638] The first tilt compensation filter H_t1(z) compensates for the tilt in the formant postfilter H_f(z) and is given by:

[0639] H_t1(z)=(1−μz⁻¹)

[0640] where μ=γ_t1k₁is a tilt factor, with k₁being the first reflection coefficient calculated on the truncated impulse response h_f(n), of the formant postfilter

k_{1} = \frac{r_{h} (1)}{r_{h} (0)}

[0641] with:

r_{h} (i) = \sum_{j = 0}^{L_{h} - i - 1} h_{f} (j) h_{f} (j + i), (L_{h} = 22) .

[0642] The postfiltering process is performed as follows. First, the synthesized speech {overscore (s)}(n) is inverse filtered through {overscore (A)}(z/γ_n) to produce the residual signal {overscore (r)}(n). The signal {overscore (r)}(n) is filtered by the synthesis filter 1/{overscore (A)}(z/γ_d) is passed to the first tilt compensation filter h_t1(z) resulting in the postfiltered speech signal {overscore (s)}_f(n).

[0643] Adaptive gain control (AGC) is used to compensate for the gain difference between the synthesized speech signal {overscore (s)}(n) and the postfiltered signal {overscore (s)}_f(n). The gain scaling factor γ for the present subframe is computed by:

γ = \sqrt{\frac{\sum_{n = 0}^{39} {\overline{s}}^{2} (n)}{\sum_{n = 0}^{39} {\overline{s}}_{f}^{2} (n)}}

[0644] The gain-scaled postfiltered signal {overscore (s)}(n) is given by:

[0645] {overscore (s)}(n)=β(n){overscore (s)}_f(n)

[0646] where β(n) is updated in sample by sample basis and given by:

[0647] β(n)=αβ(n−1)+(1−α)γ

[0648] where α is an AGC factor with value 0.9. Finally, up-scaling consists of multiplying the postfiltered speech by a factor 2 to undo the down scaling by 2 which is applied to the input signal.

[0649] FIGS. 6 and 7 are drawings of an alternate embodiment of a 4 kbps speech codec that also illustrates various aspects of the present invention. In particular, FIG. 6 is a block diagram of a speech encoder 601 that is built in accordance with the present invention. The speech encoder 601 is based on the analysis-by-synthesis principle. To achieve toll quality at 4 kbps, the speech encoder 601 departs from the strict waveform-matching criterion of regular CELP coders and strives to catch the perceptual important features of the input signal.

[0650] The speech encoder 601 operates on a frame size of 20 ms with three subframes (two of 6.625 ms and one of 6.75 ms). A look-ahead of 15 ms is used. The one-way coding delay of the codec adds up to 55 ms.

[0651] At a block 615, the spectral envelope is represented by a 10^thorder LPC analysis for each frame. The prediction coefficients are transformed to the Line Spectrum Frequencies (LSFs) for quantization. The input signal is modified to better fit the coding model without loss of quality. This processing is denoted “signal modification” as indicated by a block 621. In order to improve the quality of the reconstructed signal, perceptual important features are estimated and emphasized during encoding.

[0652] The excitation signal for an LPC synthesis filter 625 is build from the two traditional components: 1) the pitch contribution; and 2) the innovation contribution. The pitch contribution is provided through use of an adaptive codebook 627. An innovation codebook 629 has several subcodebooks in order to provide robustness against a wide range of input signals. To each of the two contributions a gain is applied which, multiplied with their respective codebook vectors and summed, provide the excitation signal.

[0653] The LSFs and pitch lag are coded on a frame basis, and the remaining parameters (the innovation codebook index, the pitch gain, and the innovation codebook gain) are coded for every subframe. The LSF vector is coded using predictive vector quantization. The pitch lag has an integer part and a fractional part constituting the pitch period. The quantized pitch period has a non-uniform resolution with higher density of quantized values at lower delays. The bit allocation for the parameters is shown in the following table.

[0654] When the quantization of all parameters for a frame is complete the indices are multiplexed to form the 80 bits for the serial bit-stream.

[0655] FIG. 7 is a block diagram of a decoder 701 with corresponding functionality to that of the encoder of FIG. 6. The decoder 701 receives the 80 bits on a frame basis from a demultiplexor 711. Upon receipt of the bits, the decoder 701 checks the sync-word for a bad frame indication, and decides whether the entire 80 bits should be disregarded and frame erasure concealment applied. If the frame is not declared a frame erasure, the 80 bits are mapped to the parameter indices of the codec, and the parameters are decoded from the indices using the inverse quantization schemes of the encoder of FIG. 6.

[0656] When the LSFs, pitch lag, pitch gains, innovation vectors, and gains for the innovation vectors are decoded, the excitation signal is reconstructed via a block 715. The output signal is synthesized by passing the reconstructed excitation signal through an LPC synthesis filter 721. To enhance the perceptual quality of the reconstructed signal both short-term and long-term post-processing are applied at a block 731.

[0657] Regarding the bit allocation of the 4 kbps codec (as shown in the prior table), the LSFs and pitch lag are quantized with 21 and 8 bits per 20 ms, respectively. Although the three subframes are of different size the remaining bits are allocated evenly among them. Thus, the innovation vector is quantized with 13 bits per subframe. This adds up to a total of 80 bits per 20 ms, equivalent to 4 kbps.

[0658] The estimated complexity numbers for the proposed 4 kbps codec are listed in the following table. All numbers are under the assumption that the codec is implemented on commercially available 16-bit fixed point DSPs in full duplex mode. All storage numbers are under the assumption of 16-bit words, and the complexity estimates are based on the floating point C-source code of the codec.

[0659] The decoder 701 comprises decode processing circuitry that generally operates pursuant to software control. Similarly, the encoder 601 (FIG. 6) comprises encoder processing circuitry also operating pursuant to software control. Such processing circuitry may coexists, at least in part, within a single processing unit such as a single DSP.

[0660] FIG. 8 is a flow diagram illustrating a procedure used by a speech encoder built in accordance with the present invention to correct LPC filter coefficients. In particular, encoder processing circuitry, operating pursuant to software control, first generates a series of LSFs (Line Spectral Frequencies) via linear predictive coding (LPC) as represented by a block 811. The encoder processing circuitry performs quantization of the LSFs at a block 815.

[0661] The LPC synthesis filter is stable when all LSFs occur in ascending order. Occasionally, however, one or more sequential pairs of LSFs are flipped. If a relatively large number, M, of LSFs are flipped, as indicated at the block 819, the encoder processing circuitry concludes that all of the frame parameters will probably be incorrect. Thus, at a block 823, the encoder processing circuitry erases the frame and uses all previous frame parameters for the current frame. An exemplary value for “M” might comprise 4 flips.

[0662] If M flips have not occurred yet at least N flips have occurred, as determined at a block 827, the encoder processing circuitry performs LSF concealment at a block 831. LSF concealment involves discarding the calculated LSFs and using the LSFs from the previous frame as (or to generate) the LSFs for the current frame. An exemplary value for “N” might comprise 2 flips.

[0663] If N flips have not occurred yet at least P flips have occurred, as determined at a block 835, the encoder processing circuitry corrects the LSF sequence to make it ascending by flipping out of order pair(s). An exemplary value for “P” might comprise 1.

[0664] Thus, for example, using M, N and P values of 4, 2 and 1, respectively, if the quantized sequence of LSFs has one set of LSF values out of order (i.e., one flipped pair), the encoder processing circuit branches at the block 835 to flip the pair at the block 839. If the sequence has two or three sets of LSFs out of order, the encoder processing circuitry branches at the block 827 to use the LSFs from the previous frame as (or to generate) the LSFs for the current frame. Likewise, if the sequence has four or more sets of LSFs out of order, the encoder processing circuitry branches at the block 819 to replace all frame parameters with (or based on) those of the previous frame.

[0665] FIG. 9 is a flow diagram illustrating a procedure similar to that illustrated in FIG. 8 but used by a speech decoder built in accordance with the present invention to correct LPC filter coefficients. If an encoder employing the functionality described in reference to FIG. 8 is used to correct out of order LSF pairs, a corresponding decoder will typically not need to engage in such correction unless errors are introduced in the communication channel. Thus, the decoder functionality illustrated can be used to correct such errors, and may otherwise be used with encoders not employing the functionality of FIG. 8 or only employing a portion thereof.

[0666] In particular, at blocks 911, 917 and 923, the decoder, or decoder processing circuitry typically operating pursuant to software control, identifies whether at least the greater of H, J or K flipped pairs exists in the present LSF sequence (or vector) received from an encoder. If at least H flips are detected, the decoder processing circuitry branches by performing frame erasure at a block 915. If at least J flipped pairs are detected, LSF concealment is applied at the block 919. Lastly, if at least K flipped pairs are detected, the ordering is corrected by flipping the pairs back. Of course, only one such branching is performed per LSF vector at issue as illustrated. As before, exemplary values for H, J and K are 4, 2 and 1, although they need not be the same as those used by the encoder.

[0667] FIG. 10 provides a more detailed specific embodiment of the functionality of FIGS. 8 and 9 performed by either or both of an encoder and a decoder, yet without using frame erasure. In particular, if two or more LSF pairs are flipped as considered at a block 1011, a substitute LSF vector is generated. In the present embodiment, the substitute LSF vector is generated by combining the previous LSF vector with the mean LSF values at a ratio of 0.9 to 0.1, respectively. If only one pair appears out of order, as determined at a block 1019, the pair is merely flipped at a block 1023.

[0668] Although frame erasure when the LSF vector exhibits even higher numbers of out of sequence pairs might be employed, because of the relatively low occurrence thereof, frame erasure need not be employed. Similarly, some embodiments may operate without employing any one or two of LSF concealment, frame erasure or pair flipping so long as stable filter coefficients are produced. Moreover, LSF concealment, frame erasure and/or pair flipping need not be duplicated at both the decoder and encoder. For example, only part of such functionality might be duplicated while other of such functionality might only be assigned to either the decoder or encoder.

[0669] Of course, many other modifications and variations are also possible. In view of the above detailed description of the present invention and associated drawings, such other modifications and variations will now become apparent to those skilled in the art. It should also be apparent that such other modifications and variations may be effected without departing from the spirit and scope of the present invention.

[0670] In addition, the following Appendix A provides a list of many of the definitions, symbols and abbreviations used in this application. Appendices B and C respectively provide source and channel bit ordering information at various encoding bit rates used in one embodiment of the present invention. Appendices A, B and C comprise part of the detailed description of the present application, and, otherwise, are hereby incorporated herein by reference in its entirety.

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

Claims

I claim:

1. A speech encoding system using an analysis by synthesis approach on a speech signal, the speech encoding system comprising:

an encoder that generates a series of line spectral frequencies, some of the line spectral frequencies are produced out of order;

the encoder determines the number of line spectral frequencies that are produced out of order;

the encoder selectively applies a first of a plurality of correction techniques to process the series of line spectral frequencies that are produced out of order if the number of line spectral frequencies that is produced out of order exceeds a first predetermined threshold;

the encoder selectively applies a second of the plurality of correction techniques to process the series of line spectral frequencies that are produced out of order if the number of line spectral frequencies that is produced out of order exceeds a second predetermined threshold; and

the second of the plurality of correction techniques comprises reordering the series of line spectral frequencies that are produced out of order.

2. The speech encoding system of claim 1 wherein a third of the plurality of correction techniques comprises line spectral frequency concealment.

3. The speech encoding system of claim 1 wherein a third of the plurality of correction techniques comprises selection of a portion of one previous series of line spectral frequencies.

4. The speech encoding system of claim 1 wherein a third of the plurality of correction techniques comprises frame erasure.

5. The speech encoding system of claim 1 wherein a third of the plurality of correction techniques comprises disregarding at least in part the series of line spectral frequencies that are produced out of order.

6. The speech encoding system of claim 1 wherein the encoder selects the one of the plurality of correction techniques by examining the series of line spectral frequencies that are produced out of order.

7. The speech encoding system of claim 6 further comprising a decoder selecting one of a plurality of correction techniques to accommodate errors in the series of line spectral frequencies that are received out of order.

8. A speech codec using an analysis by synthesis approach on a speech signal, the speech encoding system comprising:

an encoder processing circuit that generates a series of filter coefficients, some of the filter coefficients is produced out of order;

the encoder processing circuit determines the number of filter coefficients that are produced out of order; and

the encoder processing circuit selectively applies one of a plurality of techniques to accommodate the production of the series of filter coefficients that are produced out of order if the number of filter coefficients that is produced out of order exceeds a predetermined threshold.

9. The speech codec of claim 8 wherein a first of the plurality of techniques comprises selection of a portion of one previous series of filter coefficients.

10. The speech codec of claim 8 wherein a first of the plurality of techniques comprises reordering the series of filter coefficients that are produced out of order.

11. The speech codec of claim 8 wherein a first of the plurality of techniques comprises disregarding at least in part the series of filter coefficients that are produced out of order.

12. The speech codec of claim 8 wherein the encoder processing circuit selects the one of the plurality of techniques by examining the series of filter coefficients that are produced out of order.

13. The speech codec of claim 12 wherein the examination comprises considering a number of filter coefficients out of order in the series of filter coefficients.

14. The speech codec of claim 8 further comprising a decoder processing circuit that accommodates a received series of filter coefficients that are out of order.

15. The speech codec of claim 14 wherein the decoder processing circuit selects one of a plurality of techniques to accommodate the received series of filter coefficients that are out of order.

16. A method used by a speech encoder that operates on a speech signal, the method comprising:

producing from the speech signal a series of line spectral frequencies, at least one of the line spectral frequencies being produced out of order;

determining the number of line spectral frequencies that are produced out of order; and

deciding either to reorder the series of line spectral frequencies that are produced out of order, or to replace at least a portion of the series of line spectral frequencies that are produced out of order using at least a portion of one previous series of line spectral frequencies if the number of line spectral frequencies that is produced out of order exceeds a predetermined threshold.

17. The method of claim 16 further comprising determining a number of line spectral frequencies out of order in the series of line spectral frequencies that are produced out of order.

18. The method of claim 17 wherein the number of line spectral frequencies out of order is used by the speech encoder in the decision to either reorder or replace.

19. The method of claim 16 wherein a speech decoder reordering a received series of line spectral frequencies that are out of order.

20. The method of claim 16 wherein a speech decoder replacing at least a portion of a received series of line spectral frequencies that are out of order using at least a portion of one previously received series of filter coefficients.

Patent Citations (9)

Patent	Date	Inventor	Cited By
US4969192	1990-11-01	Chen et al.	Examiner
US5502713	1996-03-01	Lagerqvist et al.	Applicant
US5636231	1997-06-01	Huang et al.	Examiner
US5657420	1997-08-01	Jacobs et al.	Examiner
US5732188	1998-03-01	Moriya	Examiner
US5732389	1998-03-01	Kroon et al.	Examiner
US5778338	1998-07-01	Jacobs et al.	Examiner
EP751493A2	1997-01-01		Applicant
WO93/15502	1993-08-01		Applicant

Non-Patent Literature (12)

W. Bastiaan Kleijn and Peter Kroon, “The RCELP Speech-Coding Algorithm,” vol. 5, No. 5, Sep.-Oct. 1994, pp. 39/573-47/581.Applicant
C. Laflamme, J-P. Adoul, H.Y. Su, and S. Morissette, “On Reducing Computational Complexity of Codebook Search in CELP Coder Through the Use of Algebraic Codes,” 1990, pp. 177-180.Applicant
Chih-Chung Kuo, Fu-Rong Jean, and Hsiao-Chuan Wang, “Speech Classification Embedded in Adaptive Codebook Search for Low Bit-Rate CELP Coding,” IEEE Transactions on Speech and Audio Processing, vol. 3, No. 1, Jan. 1995, pp. 1-5.Applicant
Erdal Paksoy, Alan McCree, and Vish Viswanathan, “A Variable-Rate Multimodal Speech Coder with Gain-Matched Analysis-By-Synthesis,” 1997, pp. 751-754.Applicant
Gerhard Schroeder, “International Telecommunication Union Telecommunications Standardization Sector,” Jun. 1995, pp. i-iv, 1-42.Applicant
“Digital Cellular Telecommunications System; Comfort Noise Aspects for Enhanced Full Rate (EFR) Speech Traffic Channels (GSM 06.62),” May 1996, pp. 1-16.Applicant
W. B. Kleijn and K.K. Paliwal (Editors), Speech Coding and Synthesis, Elseveier Science B.V.; Kroon and W.B. Kleijn (Authors), Chapter 3: “Linear-Prediction Based on Analysis-by-Synthesis Coding”, 1995, pp. 81-113.Applicant
W. B. Kleijn and K.K. Paliwal (Editors), Speech Coding and Synthesis, Elsveier Science B.V.; A. Das, E. Paskoy and A. Gersho (Authors), Chapter 7: “Multimode and Variable-Rate Coding of Speech,” 1995, pp. 257-288.Applicant
B.S. Atal, V. Cuperman, and A. Gersho (Editors), Speech and Audio Coding for Wireless and Network Applications, Kluwer Academic Publishers; T. Taniguchi, Y. Tanaka and Y. Ohta (Authors), Chapter 27: “Structured Stochastic Codebook and Codebook Adaptation for CELP,” 1993, pp. 217-224.Applicant
B.S. Atal, V. Cuperman, and A. Gersho (Editors), Advances in Speech Coding, Kluwer Academic Publishers; I. A. Gerson and M.A. Jasiuk (Authors), Chapter 7: “Vector Sum Excited Linear Prediction (VSELP),” 1991, pp. 69-79.Applicant
B.S. Atal, V. Cuperman, and A. Gersho (Editors), Advances in Speech Coding, Kluwer Academic Publishers; J.P. Campbell, Jr., T.E. Tremain, and V.C. Welch (Authors), Chapter 12: “The DOD 4.8 KBPS Standard (Proposed Federal Standard 1016),” 1991, pp. 121-133.Applicant
B.S. Atal, V. Cuperman, and A. Gersho (Editors), Advances in Speech Coding, Kluwer Academic Publishers; R.A. Salami (Author), Chapter 14: “Binary Pulse Excitation: A Novel Approach to Low Complexity CELP Coding,” 1991, pp. 145-157.Applicant

Abstract