Linear Predictive Coding Coefficients Research Articles

Linear predictive coding

The basic principles of linear predictive coding (LPC) are presented. Least-squares methods for obtaining the LPC coefficients characterizing the all-pole filter are described. Computational factors, instantaneous updating, and spectral estimation are discussed. >

A Vector Quantizer Combining Energy and LPC Parameters and Its Application to Isolated Word Recognition

The theory of vector quantization (VQ) of linear predictive coding (LPC) coefficients has established a wide variety of techniques for quantizing LPC spectral shape to minimize overall spectral distortion. Such vector quantizers have been widely used in the areas of speech coding and speech recognition. The conventional vector quantizer utilizes only spectral shape information and essentially disregards the energy or gain term associated with the optimal LPC fit to the signal being modeled. In this paper we present a method of incorporating LPC spectral shape and energy into the code-book entries of the vector quantizer. To do this, we postulate a distortion measure for comparing two LPC vectors that uses the weighted sum of an LPC shape distortion and a log energy distortion. Based on this combined distortion measure, we have designed and studied vector quantizers of several sizes for use in isolated word speech recognition experiments. We found that a fairly significant correlation exists between LPC shape and signal energy. Hence, an LPC shape combined with energy vector quantizer with a given distortion requires far fewer code-book entries than one in which LPC shape and energy are quantized separately. Based on isolated word recognition tests on both a 10-digit and a 129-word airlines vocabulary, we found improvements in recognition accuracy by using the VQ with both LPC shape and energy over that obtained using a VQ with LPC shape alone.

Note on the Properties of a Vector Quantizer for LPC Coefficients

Vector quantization has been used in coding applications for several years. Recently, quantization of linear predictive coding (LPC) vectors has been used for speech coding and recognition. In these latter applications, the only method that has been used for deriving the vector quantizer code book from a set of training vectors is the one described by Linde, Buzo, and Gray. In this paper, we compare this algorithm to several alternative algorithms and also study the properties of the resulting code books. Our conclusion is that the various algorithms that we tried gave essentially identical code books.

An 800 bit/s vector quantization LPC vocoder

An 800 bit/s vector quantization linear predictive coding (LPC) vocoder has been developed. The recently developed LPC vector quantization theory is applied to reduce the bit rate for LPC coefficients coding by a factor of four. Branch search techniques and separation of voiced and unvoiced codebooks are applied for better algorithm efficiency. Differential coding is applied to reduce the bit rate for the pitch and gain parameters by one third. Formal subjective evaluation shows that the 800 bit/s vocoder preserves most of the intelligibility of an LPC system. It is also robust under different transmission error and acoustic conditions. Informal listening comparisons show the quality to be acceptable and sometimes very close to 2400 bit/s LPC speech. The computational cost of the 800 bit/s vocoder is equivalent to or even lower than the 2400 bit/s LPC-10. Compatibility with any LPC-10 vocoder is guaranteed because the 800 bit/s design only differs in the quantization and encoding algorithms. Further bit rate reduction can be achieved by removing frame to frame redundancy in the code.

A statical decision approach to the recognition of connected digits

A statistical decision approach to the recognition of connected digits is described in this paper. The method can be either speaker dependent (i.e., each new speaker must first train the system on representative digit strings before he can successfully use the system) or speaker independent. Multiple repetitions of each digit (spoken in connected strings) are used in the training sequence. Repetitions of the same digit are combined by linearly warping the individual reference patterns to the speakers' average length for the digit. Statistics of the mean and covariance of the recognition parameters between repetitions of the same digit are computed and are used in the recognition phase of the system. Once a spoken digit string has been segmented, the recognition of each digit within the string is achieved using a distance measure based on an expanded form of the principle of minimum residual error. In cases where a great deal of coarticulation can be anticipated between adjacent digits (i.e., between digits bounded by voiced regions) a second distance metric is employed. This metric includes both the effects of the analysis estimation error and the effects of coarticulation. The analysis parameters used in this system are the linear prediction coefficients (LPC's) of a 10-pole LPC analysis. For stability purposes, the linear predictive coding (LPC) coefficients are converted to parcor or reflection coefficients prior to the linear warping, and then the warped parcor coefficients are converted back to LPC coefficients for recognition purposes. The recognition system was tested on six speakers in the speaker-dependent mode with recognition accuracies of from 97 to 100 percent. It was also tested with 10 new speakers in the speaker-independent mode, with a digit recognition accuracy of 95 percent.

Wide-hand noise reduction of noisy speech

We present a new technique that reduces wide-band noise in noisy speech. The concept of the approach is based on residual or voice excited linear predictive coding (LPC). The technique can be used in either a digital or an analog communication link that is degraded by quantization or receiver thermal noise, or both. The technique capitalizes on two effects. (1) Linear predictive coefficients representing an all-pole vocal tract transfer function can be calculated unaffected by noise, if the noise present in speech is additive and white. (2) The signal-to-noise ratio of voiced sound in the fundamental frequency region of the voice band is much higher than that in the high-frequency region. The basic procedure for conditioning noisy speech is as follows. When a voice signal that has been distorted because of quantization or thermal noise, or both, is received at a receiver, the decoded signal is analyzed to determine a set of LPC coefficients by the autocorrelation approach. The noise affects only the diagonal elements of the autocorrelation matrix, since it is not correlated with speech. Hence, to find the coefficients that are unaffected by noise, the noise power is subtracted from the diagonal elements of the matrix. The power of white noise is measured from a speech pause. The residual signal or noisy speech is low-pass filtered at a cutoff frequency of 600 Hz and then passed through a nonlinear distortion device to broaden its bandwidth, This nonlinearly processed signal is used as the input to excite the filter which has been formed by the LPC coefficients. Initial testing of the conditioner by computer simulation has yielded promising results.

Hard-wired real-time LPC synthesizer

A hard-wired real-time linear predictive coding (LPC) synthesizer has been built. The synthesizer computes the speech signal from data which it receives pitch synchronously from a computer: the data includes a pitch period, sum of squares amplitude, and 12 LPC coefficients. In computing the speech signal for one pitch period, the overhang energy from previous pitch periods contributes to the energy of the signal; this factor must be taken into account in order to eliminate roughness in the resultant speech. The device computes a scaling factor for the excitation function which allows for this overhang energy. This requires substantial additional computation; therefore, fast arithmetic processors were designed so that speech can be synthesized in real time. All computation is performed in floating-point arithmetic to eliminate round-off errors. The resulting device can synthesize speech in real time, and the speech quality is the same as that of the software version.11