James L. Flanagan

Speech analysis, synthesis, and perception

J. L. Flanagan Berlin: Springer-Verlag. 1965. Pp. viii + 327.

Self-oscillating source for vocal-tract synthesizers

14.50 J. L. Flanagan is as knowledgeable on the subject of speech as any man in the world. He has filled important research roles in the field both at M.I.T. and in the Research Laboratories of the Bell Telephone Laboratories.

A Difference Limen for Vowel Formant Frequency

A self-oscillating model of the human vocal cords is derived and simulated on a digital computer. The model is used as a source of excitation for a vocal-tract synthesizer, also programmed on the computer. Synthetic speech from the simulation is used to study the influence of glottal parameters upon signal features. The cord model produces glottal volume velocity functions which reflect the acoustic interaction between source and tract. Voice pitch and irregularities in excitation are generated intrinsically from specification of subglottal pressure, cord tension, and tract configuration. Pitch produced by the cord model is a monotone increasing function of subglottal pressure and tension. Mean air flow and glottal duty factor depend upon a combination of parameters, but primarily upon the properties of the contacting surfaces during cord closure.

Synthetic voices for computers

Experiments have been conducted to determine difference limens (DLs) for vowel formant frequency. The DLs are obtained from quality judgments on synthetic vowel sounds. The results indicate the maximum accuracy necessary in analyzing the formant structure of spoken vowels and in synthesizing the sounds from the resulting formant data.

Synthesis of speech from a dynamic model of the vocal cords and vocal tract

The two methods described for giving voices to computers recognize the importance of economical storage of speech information and extensive vocabularies, and consequently are based on principles of speech synthesis. The first, formant synthesis, generates connected speech from low-bit-rate representations of spoken words. The second, text synthesis, produces connected speech solely from printed English text. For both methods the machine must contain stored knowledge of fundamental rules of language and acoustic constraints of human speech. Formant synthesis from an input information rate of about 1000 bits per second is demonstrated, as is text synthesis from a rate of about 75 bits per second. To give the reader an opportunity to evaluate some of the results described, a sample recording is available; see Appendix A for details.

Spatially selective sound capture for speech and audio processing

We describe a computer model of the human vocal cords and vocal tract that is amenable to dynamic control by parameters directly identified in the human physiology. The control format consequently provides an efficient, parsimonious description of speech information. The control parameters represent subglottal lung pressure, vocal-cord tension and rest opening, vocal-tract shape, and nasal coupling. Using these inputs, we synthesize vowel-consonant-vowel syllables to demonstrate the dynamic behavior of the cord/tract model. We show that inherent properties of the model duplicate phenomena observed in human speech; in particular, cord/tract acoustic interaction, cord vibration, and tract-wall radiation during occlusion, and voicing onset-offset behavior. Finally, we describe an approach to deriving the physiological controls automatically from printed text, and we present sentence-length synthesis obtained from a preliminary system.

Method and apparatus including microphone arrays and neural networks for speech/speaker recognition systems

Abstract Advances in transducer technology, signal processing and computing make possible high-quality sound capture from designated spatial volumes under adverse acoustic conditions. The techniques of multiple beamforming and matched filtering are applied to two- and three-dimensional arrays of sensors. Array performance is assessed in a preliminary way from computer simulations of rooms and from image characterization of the multipath environment. The results suggest that high-quality signals can be retrieved from spatially-selected volumes in severely reverberant enclosures. Reciprocally, the same techniques can be applied to spatially-selecteve sound projection.

The Huge Microphone Array. 2

A neural network is trained to transform distant-talking cepstrum coefficients, derived from a microphone array receiving speech from a speaker distant therefrom, into a form substantially similar to close-talking cepstrum coefficients that would be derived from a microphone close to the speaker, for providing robust hands-free speech and speaker recognition in adverse practical environments with existing speech and speaker recognition systems which have been trained on close-talking speech.

Computers that talk and listen: Man-machine communication by voice

The Huge Microphone Array project began in February 1994 to design, construct, debug, and test a real-time 512-microphone array system and to develop algorithms for it. Analysis of known algorithms indicated that signal-processing performance of over 6 Gflops would be required, while the need for portability-fitting it into a small van-also set an upper limit to the power required. These trade-offs and many others have led to a unique design in both hardware and software. This two-part article presents the full design and its justifications. The authors also discuss performance for a few important algorithms relative to the use of processing capability, response latency, and difficulty of programming.

On the Pitch of Periodic Pulses

Computer techniques now emerging in the laboratory promise new capabilities for voice communication between man and machine. Three modes of interaction are of special interest: computer voice readout of stored information, automatic verification of a callers identity by means of his voice signal, and automatic recognition of spoken commands. Applications extend to: voice-directed installation of telephone equipment, authentication by voice of a credit customer or of an individual requesting readout of privileged information, and voice-controlled services such as repertory dialing or automatic booking of travel reservations.