LPC Vocoder Research Articles

Correction to "A Compact, Flexible LPC Vocoder Based on a Commercial Signal Processing Microcomputer"

Correction to "A Compact, Flexible LPC Vocoder Based on a Commercial Signal Processing Microcomputer"

Changing pitch and duration in LPC synthesized speech using multipulse excitation

Pitch and duration changes are fairly simple to realize in pitch-excited LPC vocoders. Recent work [Atal and Remde, Proc. ICASSP 82, pp. 614–617) has shown that use of multi-pulse excitation provides a significant improvement in the quality of synthetic speech. However, procedures for changing pitch with multi-pulse excitation are not known. We discuss methods for introducing pitch and duration changes in speech synthesized with multi-pulse excitation. Pitch is changed by modifying the length of individual pitch periods. Two methods of adjusting the period length were investigated: one by linear scaling of the time axis of the multi-pulse excitation, the other, by adding or deleting zeros in the excitation. The second method produced very little distortion in the synthetic speech, except in those cases where there was a significant interaction between the first formant and the pitch frequency. The scaling of the time axis produced significantly more distortion. Duration changes are accomplished by adding or removing pitch periods. The LPC area parameters and the amplitude of the major excitmion pulse are interpolated as necessary in creating additional pitch periods.

A family of special purpose microprogrammable digital signal processor IC's in an LPC vocoder system

This paper discusses a family of digital signal processor integrated circuits designed to be the fundamental building blocks of an LPC vocoder, specifically LPC analysis, AMDF pitch extraction, and LPC synthesis. The IC's are custom designs intended to minimize silicon real estate for extremely high performance, novel DSP architectures, and high computational demand algorithms. Each IC shares a common architecture, that of a single internal data bus connecting input and output FIFO's, multiple arithmetic logic units, registers, and a microcode control bus. Each application is microprogrammed in register transfer language providing considerable flexibility. The CMOS design of each IC allows it to implement the high performance digital signal processing algorithms with minimal power consumption. The LPC synthesis IC allows various common filter topologies and residual excitation. The AMDF IC extracts the commonly used features for pitch and voicing. The LPC analyzer IC is microprogrammed to implement a flow form PARCOR LPC analysis. Algorithm architecture tradeoffs between computational load and memory are discussed.

A Family of Special Purpose Microprogrammable Digital Signal Processor IC's in an LPC Vocoder System

This paper discusses a family of digital signal processor integrated circuits designed to be the fundamental buitding blocks of an LPC vocoder, specifically LPC analysis, AMDF pitch extraction, and LPC synthesis. The IC's are custom designs intended to minimize silicon real estate for extremely high performance, novel DSP architectures, and high computational demand algorithms. Each IC shares a common architecture, that of a single internal data bus connecting input and output FIFO's, multiple arithmetic logic units, registers, and a microcode control bus. Each application is microprogrammed in register transfer language providing considerable flexibility. The CMOS design of each IC allows it to implement the high performance digital signal processing algorithms with minimal power consumption. The LPC synthesis IC allows various common filter topologies and residual excitation. The AMDF IC extracts the commonly used features for pitch and voicing. The LPC analyzer IC is microprogrammed to implement a flow form PARCOR LPC analysis. Algorithm architecture tradeoffs between computational load and memory are discussed.

A Compact, Flexible LPC Vocoder Based on a Commercial Signal Processing Microcomputer

A very small, flexible, high-quality, full-duplex 2.4-kbit/s linear predictive vocoder has been implemented with commercially available integrated circuits. This fully digital realization is based on a distributed signal processing architecture employing three Nippon Electric Company (NEC) µPD7720 signal processing interface (SPI) single-chip microcomputers. One SPI implements the LPC analyzer, a second implements the Gold pitch and voicing decision algorithm, white the third µPD7720 implements the excitation generator and synthesizer. An Intel 8085-based 8-bit microcomputer is used for data transfer, control and multiplexing functions, and communications with the host terminal. The LPC chip set achieves high flexibility by accepting run time initialization options from the Intel 8085. These parameters include choice of linear predictive model (<= 15), analysis and synthesis frame size, and speech sampling frequency, as well as choice of speech input and output coding formats (linear or µ-255 law) and choice of analog or digjtal pre- and deemphasis. A total of 16 integrated circuits is used in the LPC vocoder with a power disipation of 5.5 W and occupying 18 in/sup 2/ of circuit area.

A compact, flexible LPC vocoder based on a commercial signal processing microcomputer

A very small, flexible, high-quality, full-duplex 2.4-kbit/s linear predictive vocoder has been implemented with commercially available integrated circuits. This fully digital realization is based on a distributed signal processing architecture employing three Nippon Electric Company (NEC) μPD7720 signal processing interface (SPI) single-chip microcomputers. One SPI implements the LPC analyzer, a second implements the Gold pitch and voicing decision algorithm, while the third μPD7720 implements the excitation generator and synthesizer. An Intel 8085-based 8-bit microcomputer is used for data transfer, control and multiplexing functions, and communications with the host terminal. The LPC chip set achieves high flexibility by accepting run time initialization options from the Intel 8085. These parameters include choice of linear predictive model ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">\leq 15</tex> ), analysis and synthesis frame size, and speech sampling frequency, as well as choice of speech input and output coding formats (linear or μ-255 law) and choice of analog or digital pre- and deemphasis. A total of 16 integrated circuits is used in the LPC vocoder with a power dissipation of 5.5 W and occupying 18 in <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of circuit area.

Effect of channel errors on the performance of LPC vocoders

The effect of channel errors on the performance of LPC vocoders has been investigated. The vocoder performance is fairly robust when the error rate is 10-3or less. Among the various parameters, the first two reflection coefficients and pitch are most sensitive to channel errors. When these parameters are coded with an error protection scheme, the performance is robust even when the error rate is as high as 10-2.

Variable Frame Rate Transmission: A Review of Methodology and Application to Narrow-Band LPC Speech Coding

We review the variable frame rate (VFR) transmission methodology that we developed, implemented, and tested during the period 1973-1978 for efficiently transmitting LPC vocoder parameters extracted from the input speech at a fixed frame rate. In the VFR method, parameters are transmitted only when their values have changed sufficiently over the interval since their preceding transmission. We explored two distinct approaches to automatic implementation of the VFR method. The first approach bases the transmission decisions on comparisons of the parameter values of the present frame and the last transmitted frame. The second approach, which is based on a functional perceptual model of speech, compares the parameter values of all the frames that lie in the interval between the present frame and the last transmitted frame against a linear model of parameter variation over that interval. The application of VFR transmission to the design of narrow-band LPC speech coders with average bit rates of 2000-2400 bits/s is also considered. The transmission decisions are made separately for the three sets of LPC parameters, pitch, gain, and spectral parameters, using separate VFR schemes. A formal subjective spccch quality test of six selected LPC coders is described, and the results are presented and analyzed in detail. It is shown that a 2075 bit/s VFR coder produces speech quality equal to or better than that of a 5700 bit/s fixed frame rate coder.

Recursive windowing for generating autocorrelation coefficients for LPC analysis

A method for recursively computing the autocorrelation estimates needed for LPC analysis in a vocoder environment has been developed theoretically and studied experimentally. The method has three specific advantages: 1) it requires very little memory for its implementation; 2) it is realized by a structure consisting of several identical modules; and 3) the effective window length may be changed without varying the structure. Experimental results showed the speech quality to be comparable to and slightly better than that produced by an auto-correlation LPC vocoder using a Hamming window.

F0 rule for discourse

Rules for the fundamental frequency (F0) contour for discourse have been formulated. The rules are based on F0 data that consider boundaries and segmental factors in a number of natural readings. The rules give F0 within a word according to stress and boundary values, and a concatenation rule is applied to obtain the continuous F0 across two adjacent words. The rules have been tested on a naturally read story using an LPC vocoder synthesis. A demonstration tape will be played.

Real-time simulation of adaptive transform coding

Adaptive transform coding (ATC) has recently been proposed as a technique for speech coding at bit rates in the range of 9.6- 16 kbits/s. In this paper we report on two new developments: 1) the use of a homomorphic vocoder model for the "side-information" channel in ATC and 2) a real-time simulation of ATC on an array processing computer. It is shown that the choice of the homomorphic "side-information" model leads to a convenient form of the ATC algorithm for real-time block processing using array processing techniques. It is also shown that the log spectrum output of the homomorphic model is in a convenient form for input to both the bit assignment algorithm in ATC (which becomes a straightforward quantization operation) and the quantization of the transform coefficients (which may be done in the log domain). An array processor simulation of this form of the algorithm has been implemented and it serves as a highly useful and convenient tool for studying the ATC algorithm in real time in a Fortran programming environment. It has allowed us, for the first time, to perform actual telephone conversations over a transform coder. The quality of this ATC algorithm was found to be essentially equivalent to that of a previous version using an LPC vocoder model for the side information.

Simple algorithm to improve LPC vocoder speech with noisy input

The letter suggests a simple method for improving the quality of linear prediction vocoder speech with noisy input. The improvement has been substantiated by the results of an experiment using the absolute isopreference method for speech quality testing.

An 800 bps speech compression system based on vector quantization

Recent developments in vector quantization as applied to speech compression have allowed significant reductions in the bit rate of LPC vocoders. A 2400 bps LPC vocoder is being modified to operate at 800 bps while retaining acceptable intelligibility and naturalness in quality. The design of such a speech compression system is discussed and a demonstration tape played.

A real‐time “homomorphic‐vocoder driven” transform coder for speech

Adaptive transform coding (ATC) has recently been proposed as a technique for good quality, robust coding of speech at bit rates from 9.6 to 16 kb/s. In this paper we report on two new developments: (1) the use of a homomorphic vocoder model for the “side‐information” channel in ATC and (2) the application of these techniques to a real‐time implementation of ATC on an array processing computer. In the ATC coder speech is buffered into blocks and transformed into a frequency domain representation. The transform coefficients are then quantized for transmission. The discrete cosine transform (DCT) is generally used because it minimizes “end effects” in the blocks. The step sizes and number of bits in the quantizers are adaptively controlled to track the dynamic spectral variations of speech. The control of the step sizes and bit assignment is based on a local estimate of the spectrum which is parameterized, encoded, and transmitted as “side‐information” to the receiver. In previous work an LPC vocoder model was used to efficiently model the spectrum in order to take advantage of known properties of speech. In this paper we investigate the use of a homomorphic vocoder model which accomplishes essentially the same thing. The properties of the “pseudo‐cepstrum” in the homomorphic model (which is based on a DCT analysis) are explored in order to define the side information needed to characterize the DCT spectrum. It is also shown that the choice of the homomorphic model leads to a particularly convenient form of the ATC algorithm for a real‐time block processing implementation on an array processing computer. Also it is shown that the log spectrum output of the homomorphic model is in a convenient form for both the bit assignment algorithm (which now becomes a straightforward quantization operation) and the quantization of the transform coefficients (which is done in the log domain). Details of the algorithm will be discussed and examples of real‐time coded speech will be played. Good quality is achieved at 9.6 kb/s and nearly transparent quality is achieved at 16 kb/s.

Objective and Subjective Performance of Tandem Connections of Waveform Coders with an LPC Vocoder

In a recently proposed communication system, there would be tandem connections of 16 kb/s delta modulators and 2.4 kb/s vocoders. Preliminary work has indicated that such tandem links would be of substantially lower quality than either the delta modulator link or the vocoder link alone. The present study, which includes an elaborate subjective speech quality experiment, confirms this preliminary conclusion. It also shows that two other differential waveform coders are no better than the proposed delta modulator in tandem links. On the other hand, a 5-band sub-band coder does offer substantially higher quality than the delta modulator. Still, its performance in tandem with the vocoder is poorer than that of the vocoder or the sub-band coder alone and is probably of only marginal value for practical communication. We have obtained several objective measures of speech quality which, for the most part, show relatively little correlation with subjective quality. The most successful objective predictor of subjective ratings is a linear combination of linear predictive coding distances.

Subjective quality testing of a perceptual-model-based variable frame rate (VFR) LPC vocoder

Subjects judged the degradation of quality caused by processing speech through three LPC vocoder systems, one of which was the VFR system based on our functional perceptual model of speech [Proc. 1977 IEEE Int. Conf. ASSP, p. 219 (1977)], and the other two were fixed-rate systems, one of which had its transmission frame rate equal to the average frame rate of the VFR system (about 30/sec), and the other had its frame rate equal to the peak frame rate of the VFR system (100/sec). Stimulus materials were six phoneme-specific sentences read by each of six speakers, three male and three female [see Proc. 1977 Int. Conf. ASSP, p. 413 (1977)]. Unprocessed speech was also included as an anchor. The results show that the VFR system's quality approximates that of the 100-frames-per-second fixed-rate system, and in some cases even surpasses it. Reasons for this rather surprising result will be discussed. [Work sponsored by ARPA-IPTO.)

Tandem Connections of Wideband and Narrowband Speech Communication Systems Part 2-Wideband-to-Narrowband Link

In this paper the tandem link of a 16 kb/s Continuously Variable Slope Delta modulator (cvsd) waveform coder and a 2.4 kb/s Linear Predictive Coding (lpc) vocoder is studied. Of prime concern are the effects of the cvsd coder on the lpc vocoder analyzer. In particular the problems involved in making a reliable voiced-unvoiced decision, estimating pitch period, and estimating lpc coefficients from the coder output are studied. It is shown that lpc coefficient estimation from the cvsd output is highly inaccurate. An analytical distortion measure (an lpc distance) is used to show the magnitude of the distortion introduced by the coder as a function of the signal gain into the cvsd coder. Although the remainder of the lpc analysis (i.e., pitch detection, voiced-unvoiced decision, and gain calculation) can be performed reasonably accurately, the magnitude of the distortions in estimating the lpc coefficients is sufficiently large to make the vocoded speech barely intelligible and of poor quality.

Tandem Connections of Wideband and Narrowband Speech Communications Systems: Part 1-Narrowband-to-Wideband Link

The performance of a tandem connection of narrowband and wideband speech communication systems is evaluated. Specifically, the narrowband system consists of a conventional linear Predictive Coding (lpc) vocoder operating at a bit-rate of 2.4 kb/s and the wideband system consists of a Continuously Variable Slope Delta modulator cvsd operating at a bit rate of 16 kb/s. In Part 1 of this paper the properties of the narrowband-to-wideband link are investigated and in Part 2 the properties of the wideband-to-narrowband link are investigated. In part 1 the snr (signal-to-quantizing noise ratio) of the cvsd coder is analyzed over a 50-dB variation of the input signal levels and for a variety of source excitations for the lpc synthesizer. It is shown that snr improvements in the cvsd coder of 2 to 2.5 dB are possible in the slope overload region of the coder by modifying the source excitation of the lpc synthesizer and by preprocessing the input signal to the coder with an allpass filter. Both methods aid in reducing the peak factor (peak-to-rms level) of the input speech to the coder. Subjectively, however, only slight improvements in quality, if any, were observed with these modifications.

Speech‐quality testing of some variable‐frame‐rate (VFR) linear‐predictive (LPC) vocoders

VFR transmission of LPC vocoder coefficients is a technique developed to reduce the average transmission rate without appreciable loss of quality. The technique transmits parameters at a variable rate in accordance with the changing characteristics of the speech signal. In order to assess the effectiveness of VFR transmission, we performed an experiment to compare it with two other methods for reducing the bit rate: (a) reducing the number of poles, and (b) increasing the quantization step size of the LPC coefficients (log–area ratios). Thirty‐two stimulus sentences were prepared by passing four utterances (2 sentences × 2 speakers) through eight vocoder systems in a 2×2×2 factorial design; two values were assigned to each of the three parameters: average frame rate, number of poles, and quantization step size. Eight listeners made seven‐point category ratings of quality degradation. The results of the experiment show that, of the three methods studied, the VFR technique produced the highest quality at any given transmission rate (or, equivalently, yielded the lowest bit rate for a fixed level of speech quality).

Analysis of the performance of a tandem connection of a 16-Kbps CVSD coder to a 2.4-Kbps LPC vocoder

In this paper the tandem link of a 16-Kbps CVSD waveform coder and a 2.4-Kbps LPC vocoder are studied. Of prime concern are the effects of the CVSD coder on the LPC vocoder analyzer. In particular the problems involved in making a reliable voiced-unvoiced decision, estimating pitch period, and estimating LPC coefficients from the coder output are studied. It is shown that LPC coefficients estimation from the CVSD output is highly inaccurate. An analytical distortion measure (an LPC distance) is used to show the magnitude of the distortion introduced by the coder as a function of the signal gain into the CVSD coder. Although the remainder of the LPC analysis (i.e., pitch detection, voiced-unvoiced decision, and gain calculation) can be performed reasonably accurately, the magnitude of the distortions in estimating the LPC coefficients is sufficiently large to make the vocoded speech barely intelligible and of poor quality.