论文信息 - Design of a real-time French text-to-speech system

Design of a real-time French text-to-speech system

Abstract A microprocessor system has been designed to convert ordinary French text into audible speech in real-time. The highly-modular Pascal software translates an input text into phonemes, and assigns duration and pitch via a simple syntactic analysis. Simulation of a human vocal tract models changes (with 10 ms updates) in spectral resonances, which are converted into multiplier coefficients for second-order digital filters. The speech is generated at 10 000 samples/sec with a programmable DSP integrated circuit. Of two printed circuit cards in the system, one contains an Intel 8086 microprocessor and memory, and the other has the DSP chip and an associated interface, D/A converter and amplifier. An earlier version of the system, for a VAX-11/780 entirely in software with floating-point arithmetic, required eight times real-time calculation. The current real-time system shows the practicability of generating high quality French synthetic speech on a printed circuit board.

Douglas D. O'Shaughnessy

[1] R. Crochiere,et al. Current perspectives in digital speech , 1983, IEEE Communications Magazine.

[2] S. Hertz. From text to speech with SRS , 1982 .

[3] D. O'Shaughnessy. A study of French vowel and consonant durations , 1981 .

[4] Douglas O’Shaughnessy. A study of French spectral patterns for synthesis , 1982 .

[5] Françoise Emerard,et al. Sparte: A text-to-speech machine using synthesis by diphones , 1982, ICASSP.

[6] M. Divay,et al. Grapheme-to-phoneme transcription for French , 1977 .

[7] Sheri Hunnicutt,et al. A multi-language text-to-speech module , 1982, ICASSP.

[8] Rolf Carlson,et al. Swedish Speech Researchers Team Up with Electronic Ventrure Capitalists , 1984 .

[9] D. O'Shaughnessy. Automatic speech synthesis , 1983, IEEE Communications Magazine.

[10] Rolf Carlson,et al. MITalk‐79: The 1979 MIT text‐to‐speech system , 1979 .

[11] Dennis H. Klatt,et al. Software for a cascade/parallel formant synthesizer , 1980 .

[12] David B. Pisoni,et al. Unlimited text-to-speech system: Description and evaluation of a microprocessor based device , 1980, ICASSP.

[13] J. Makhoul,et al. Linear prediction: A tutorial review , 1975, Proceedings of the IEEE.

[14] Shinji Maeda. Better acoustic model of the vocal tract , 1981 .