A method of determining pitch and voicing information from speech signals is presented. The algorithm, which employs time-domain analysis and pattern recognition techniques, is fast and yields accurate pitch and voicing estimates. A search routine is employed to find periodicity in each of four signals derived from the speech waveform and the results are combined to form a pitch estimate. The voicing decision uses linear discriminant analysis, and declares speech frames voiced or unvoiced based on a weighted sum of 13 parameters. Performance comparisons with other pitch detectors are reported.

This paper presents a new method of modeling the LPC residual during unvoiced speech for voice coding at 4.8 kb/s. With this method, speech is synthesized using one of three excitation types: periodic pitch pulses, random noise, or multipulse. By using multipulse excitation it is possible to accurately produce speech which is difficult to model using noise and pitch pulses alone [1]. Since multipulse is only used where appropriate, efficient, sub-optimal methods of calculating the pulse amplitudes and positions are adequate, simplifying the implementation into a real-time system. The synthetic speech may be coded at 4.8 kb/s since multipulse, used only where appropriate, suffers little quality loss when quantized. A method of determining which excitation type is to be used is discussed. Formal listening test results are also presented.

This paper constitutes a mathematical analysis of the effects of random channel errors when the transmitted signal is in adaptive delta modulation (ADM) format. This study not only permits us to model the behavior of an ADM signal, with respect to errors, but also provides a method for controlling excessive signal distortion, particularly for high bit error rates. Our analysis shows that the decoded error signal exhibits both a dc and an ac component. The dc error is controlled with a leaky accumulator and an appropriate output filter. The ac error is shown to be proportional to the original signal and we are able to control its amplitude by artificial error injection. This yields an effective way of improving performance in a noisy environment. We present computer simulation results to verify our analysis and error control with a variety of test signals.

This paper considers the realization of an all-digital system for direct filtering of a delta modulation (DM) encoded signal. We have shown [1] that this system can be realized fairly easily for any type of filter function when the signals are encoded with a linear DM. We now address the problem of digitally filtering an adaptive DM (ADM) encoded signal. We show that we can again realize any type of filter function and our system will operate directly on the ADM bit stream. In addition, we present a circuit realization that does not require complex digital hardware. Performance has been tested via simulation on a digital computer, using a variety of test signals; both temporal and spectral results are presented.