Split Vector Quantization Research Articles

Low bit-rate speech coding based on an improved sinusoidal model

This paper addresses the design, implementation and evaluation of efficient low bit-rate speech coding algorithms based on an improved sinusoidal model. A series of algorithms were developed for speech classification and pitch frequency determination, modeling of sinusoidal amplitudes and phases, and frame interpolation. An improved paradigm for sinusoidal phase coding is presented, where short-time sinusoidal phases are modeled using a combination of linear prediction, spectral sampling, linear phase alignment and all-pass phase error correction components. A class-dependent split vector quantization scheme is used to encode the sinusoidal amplitudes. The masking properties of the human auditory system are effectively exploited in the algorithms. The algorithms were successfully integrated into a 2.4 kbps sinusoidal coder. The performance of the 2.4 kbps coder was evaluated in terms of informal subjective tests such as the mean opinion score (MOS) and the diagnostic rhyme test (DRT), as well as some perceptually motivated objective distortion measures. Performance analysis on a large speech database indicates considerable improvement in short-time signal matching both in the time and the spectral domains. In addition, subjective quality of the reproduced speech is considerably improved.

Safety-net pyramid VQ of LSF parameters for widebandspeech codecs

The authors propose a predictive pyramid VQ (PPVQ) that uses the safety-net concept to quantise the LSF parameters of wideband speech. A first-order auto-regressive (AR) predictor is used to predict the LSF parameters. To quantise the prediction error signal effectively, a pyramid VQ (PVQ) is used while a memoryless PVQ is used to encode low-correlation frames. By combining the PPVQ and the memoryless PVQ, denoted ‘safety-net pyramid VQ’, the advantages of both quantisation methods can be utilised. The average SD value of a safety-net pyramid VQ is 0.35 dB less than that of a predictive split VQ (PSVQ). The safety-net pyramid VQ also requires low complexity and does not require any memory for the codebook storage.

A fast-search algorithm for the predictive split VQ of LPC parameters

A fast-search method is introduced to reduce the searching range of a codebook using the ordering property of LSF parameters. We apply the proposed fast-search method to the G.723.1 coder with the predictive split VQ structure. Simulation results indicate that using this fast-search method, the average searching range of the codebook can be reduced by 20.1% without degradation of the spectral distortion (SD) and additional increases of memories. The reduction of 20.1% in the average searching range results in the decrease of 19.1%, 20.1%, 19.4%, and 12.2% in the number of additions, subtractions, multiplies, and comparisons, respectively.

Temporal decomposition for low rate wideband speechcompression

An investigation into low bit rate wideband speech coding for applications such as unicast streaming is presented. Wideband spectral parameters are quantised below 1 kbit/s using temporal decomposition (TD) applied to the line spectral frequencies. Quantisation using TD performs significantly better than split vector quantisation at an equivalent bit rate.

Multiple transform domain split vector quantisation

A novel split vector quantisation scheme where each signal vector is projected into multiple transform domains is proposed. The coder selects codes from the domain that best represents the signal vector. Sample results using one-dimensional speech signals confirm the superior performance of the proposed scheme.

Low delay coder (< 25 ms) of wideband audio (20 Hz-15 kHz) scalable from 64 to 32 kbit/s

A low delay coder for speech and music signals sampled at 32kHz is described. Its algorithmic delay does not exceed 25 ms which enables audioconferencing applications without echo cancellation. Its bit rate is scalable between 64 and 32 kbit/s by steps of 8 kbit/s. The transmitter issues the binary code at 64 kbit/s with lower bit rate codes embedded in it. The receiver may operate at lower bit rates with gradual loss of quality. The proposed coder is based on a mixed scheme : the adopted solution contains elements from the CELP speech coder and frequency domain music coders. The perceptual signal is obtained in the time domain, then transformed to the frequency domain where bit allocation is calculated and transform coefficients are quantized. A first solution based on the dft is discussed, then a second solution based on a mdct with small overlap is applied. The quantization of these coefficients is done in the following way. First, a prediction of the whole spectrum is applied. Then, a mean- removed gain- shape split vq is used for amplitude spectrum quantization and a hierarchical 2- dimensional vq is used for phase spectrum quantization with amplitude correction. At the phase quantization stage, each codeword describing the selected vector index is split into parts corresponding to different bit rates. Due to the hierarchical codebook structure, truncated indices may be used, without much affecting the signal quality. Simulation results are presented and the robustness of the proposed coder is examined.

Split vector quantization of LSF parameters with minimum of dLSF constraint

In speech coding, the spectral envelope of an analysis frame is often represented by line spectral frequencies (LSFs). LSFs are estimated from given linear predictive coefficients (LPCs) and can be transformed back to corresponding LPCs without loss of information. The authors present two improved split vector quantization (SVQ) methods for line spectral frequency (LSF) parameters. By using these methods jointly, the codewords and quantization results conserve a given minimum difference LSF (dLSF), although they are trained and quantized with a weighted distance measure. Experimental results show that the proposed methods are more effective than conventional SVQ methods, because the total training error and number of outliers due to quantization are all reduced.

Fast Search Methods for Spectral Quantization

In this paper, we examine the computational requirements for the split-vector class of vector quantizers when applied to low-rate speech spectrum quantization. The split-vector quantization techniques are able to reduce the complexity and storage requirements of the 24-bit per frame spectral quantizer to manageable proportions. However, further dramatic reductions in computational complexity are possible, as will be demonstrated. As the fast-search algorithms reported in the literature are somewhat data dependent, it has been necessary to carefully evaluate several methods specifically for the speech coding problem. A total of six methods have been evaluated for their effectiveness in this task, and we show that a so-called “geometric” fast-search method results in a reduction in the average search time of an order of magnitude.

Split matrix quantization of LPC parameters

This paper examines in detail the design issues and performance characteristics of linear predictive coding (LPC) split matrix quantization (SMQ). This efficient LPC quantization method which was proposed by Xydeas and Papanastasiou (1995) can be viewed as an extension of the conventional split vector quantization (SVQ) process. SMQ removes existing interframe/intraframe line spectral frequency (LSF) redundancy by applying VQ principles on trajectories of smoothly evolving, with time, LSF coefficients. Using a 20 ms LPC analysis frame size, transparent quantization is achieved at 900 b/s, whereas high quality LSF quantization is easily obtained at 650 b/s, Furthermore, the SMQ methodology offers valuable flexibility in the way quantization of LPC coefficients is performed and leads into several schemes of varying computational complexity/storage characteristics.

Vector quantization of speech line spectrum pair parameters and reflection coefficients

Methods are presented for intraframe encoding of line spectrum pairs. A two-stage vector quantizer-lattice vector quantizer (VQ-LVQ) has a 2-3 b/vector performance advantage over the split vector quantizer, and can save 1-3 b/vector, compared to multistage vector quantization, depending on the number of stages and the number of survivors retained at each stage. The vector quantizer-lattice vector quantizer has a moderate computational complexity. Among three trellis codes considered, trellis-coded vector quantization with nonlinear prediction provides line spectrum pair encoding performance similar to the split vector quantizer. Hybrid transformations are developed for encoding the reflection coefficients using vector quantization, and offer a small encoding advantage over the commonly used arcsine transformation and log area ratios.

Spectrum quantisation using two-dimensional multipletypeframe segmentation

An efficient spectrum quantisation method using two-dimensional multiple type frame segmentation (2D-MTFS) is proposed. Conventional methods generate redundant spectral information or spectral distortion using a fixed update rate for LSP parameters, independent of the order of coefficients and the phonetic context. The proposed method optimises the update rate of each LSP parameter by various types of two-dimentional frame segmentation. Simulation results show that the proposed method generates a less distorted spectrum with fewer bits than with the conventional method. Transparent quantisation is achieved at 1275 bit/s using split vector quantisation.

A classified split vector quantization of LFS parameters

Split vector quantization (SVQ) of LSF parameter suffers from huge complexity and storage requirements. Increasing the number of subvectors M results in a considerable decrease in both complexity and storage, but at the expense of rapid degradation in performance as M increases. To alleviate this problem, we propose classified SVQ (CSVQ), which uses class-dependent splitting and bit allocation schemes combined with a classified VQ structure. For practical applications, we designed two CSVQ structures. Experimental results have shown that both of the CSVQ schemes achieve nearly transparent quantization at 28 bits/frame while requiring much less complexity than the conventional SVQ.

Robust LPC spectrum quantization-vector quantization by a linear mapping of a block code

In this paper, we address the use of a vector quantizer (VQ) defined as a linear mapping of a block code (LMBC-VQ) in robust LPC spectrum quantization. Split VQ and multistage VQ of LSF's are considered. We demonstrate several properties for the LMBC-VQ including design methods and channel robustness theory. The LMBC-VQ codebooks are, guided by the robustness theory, directly designed with inherent robustness against channel errors. We demonstrate results where the resulting LMBC-VQ's outperform conventional codebooks designed with the generalized Lloyd algorithm in the presence of channel errors even when postprocessing index assignment is applied to the latter. Listening tests confirmed that the increased robustness is significant for the perceptual quality. We show that short block codes can be used to build the codebooks with no or small losses in quantization performance. This is advantageous for channel error robustness, lower memory requirement, and lower complexity. We also describe the structure of split VQ and multistage VQ using the LMBC framework. This allows us to assess the structural constraints and to generalize these VQ schemes, resulting in improved performance.

Performance of LSF vector quantizers for VSELP coders in noisy channels

AbstractEfficient quantization of synthesis filter coefficients for CELP (Code Excited Linear Prediction) coders is essential to achieve high quality speech at low rates. Three vector quantizers with good potential for utilization in low rate coders are studied. Each of them is implemented inside the structure of the VSELP (Vector Sum Excited Linear Prediction) coder, an important member in the class of CELP coders. These three vector quantizers are used to encode LSF (Line Spectral Frequency) parameters and are compared in terms of robustness to channel errors, complexity and quality of synthesized speech. Performance of synthesized speech is evaluated considering the objective measure of frequency weighted signal to noise ratio and subjective results obtained from listening tests. With the purpose of improving the robustness to channel errors, the application of simulated annealing to assign binary indices to the output levels of the quantizer is also investigated. A split vector quantization scheme which employs interframe prediction has shown to be an attractive approach to encode the synthesis filter parameters. It provides a performance comparable to the IS‐54 scheme and uses 10 bits less for each LSF frame.

Split-dimension vector quantization of Parcor coefficients for low bit rate speech coding

A novel split vector quantization (SVQ) scheme for low bit rate coding of speech signals is proposed. In this scheme, the LPC parameter vector, which is represented by Parcor coefficients, is split into small-dimension subvectors, and each subvector is sequentially quantized according to a multistage structure that resembles a segmented lattice filter. The forward and backward prediction residuals in the segmented filter are coupled across VQ stages. The quantizer in each stage operates on the principle of minimizing the forward and backward prediction error energies similar to linear predictive analysis. Simulation results show that the new split VQ scheme can achieve transparent quantization of LPC parameters at 25 b/frame.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

An LSF quantizer for the North-American half-rate speech coder

The design of the half-rate coder for the North-American cellular communication standard poses a challenging problem. The desired speech quality is that of the full-rate coder, but the total bit rate specified for the half-rate is only 6.5 kb/s. Since the mobile communication channel is very noisy, the use of error-correcting codes is necessary, which leaves only about 4 kb/s for the actual speech information. This very restricted bit budget requires that fewer than 30 b be allotted to the quantization of the LSF parameters, which precludes the use of scalar quantization. In this work, a quantizer design procedure which is tailored for the transmission frame structure of the North-American (TIA) cellular communication standard is presented. It is basically a split vector quantizer which makes use of interframe prediction to lower the number of bits required for quantization. Since the prediction is performed only on every other frame, the error propagation is limited. The quantization technique yields good performance in the 26-27 b/frame range, and the complexity of its implementation is less than 0.5 Mip.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Efficient vector quantization of LPC parameters at 24 bits/frame

For low bit rate speech coding applications, it is important to quantize the LPC parameters accurately using as few bits as possible. Though vector quantizers are more efficient than scalar quantizers, their use for accurate quantization of linear predictive coding (LPC) information (using 24-26 bits/frame) is impeded by their prohibitively high complexity. A split vector quantization approach is used here to overcome the complexity problem. An LPC vector consisting of 10 line spectral frequencies (LSFs) is divided into two parts, and each part is quantized separately using vector quantization. Using the localized spectral sensitivity property of the LSF parameters, a weighted LSF distance measure is proposed. With this distance measure, it is shown that the split vector quantizer can quantize LPC information in 24 bits/frame with an average spectral distortion of 1 dB and less than 2% of the frames having spectral distortion greater than 2 dB. The effect of channel errors on the performance of this quantizer is also investigated and results are reported.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Efficient vector quantization of LPC parameters at 24 bits/frame

Linear prediction coding (LPC) parameters are widely used in various speech processing applications for representing the spectral envelope information of speech. For low-bit-rate speech coding application, it is important to quantize these parameters accurately using as few bits as possible without sacrificing the speech quality. Though the vector quantizers are more efficient than the scalar quantizers, their use for fine quantization of LPC information (using 24–26 bits/frames) is impeded due to their prohibitively high complexity. In this paper, a split vector quantization approach is used to overcome the complexity problem. Here, the LPC vector is divided into two parts and each part is vector-quantized separately. The splitting of LPC vector is studied in the following three domains: (1) line spectral-pair frequency (LSF), (2) arc-sine reflection coefficient, and (3) log area ratio. Splitting in LSF domain is found to be the best. Using the localized spectral properties of the LSF parameters, a weighted LSF distance measure is proposed. Using this distance measure, it is shown that the split vector quantizer can quantize LPC information in 24 bits/frame with 1-dB average spectral distortion. In terms of average spectral distortion and number of outliers, its performance is better than the 32 bits/frame scalar quantizer.