Method of and device for synthesis of speech from printed text
First Claim
1. A method for synthesizing speech, comprising the steps of:
- analyzing a printed text grammatically and phonetically for sequences of phonemes, for the placement of accents, for the placement and duration of pauses and intonations to form frequency magnitude and amplitude characteristics of a sentence to be synthesized;
reading out from a read-only memory, according to said frequency magnitude characteristics, binary signals coding at least in part successive magnitudes of voice-frequency functions;
converting said binary signals at the output of said read-only memory into an analog signal;
modulating said analog signal according to said amplitude characteristics;
introducing quasirandom changes in said frequency magnitude and amplitude characteristics to facilitate the production of natural-sounding speech, the quasirandom changes introduced in said frequency magnitude and amplitude characteristics being within the limits of ±
3% for the frequency and ±
30% for the amplitude; and
feeding the modulated analog signal to the loudspeaker.
0 Assignments
0 Petitions
Accused Products
Abstract
Upon analyzing grammatically and phonetically a printed text for accents, pauses, intonations and influences of adjacent voice elements in a sentence to be synthesized, a computer loads a plurality of registers including an address counter with instructions for addressing a read-only memory, these instructions specifying rates of counting, numbers or counts, whether counting is to be decremental or incremental and initial addresses of sequences of binary bits coding successive magnitudes of noise signals or of voice-frequency functions. The output of the read-only memory is fed to a loudspeaker via a digital/analog converter and an amplifier whose output is modulated by a signal transmitted from the computer through another d/a converter. The durations of noise and voice-frequency speech elements read out from the memory and the modulation of their amplitudes by the amplifier are randomly modified within ±3% for the frequency and ±30% for the amplitude by the computer to obtain natural-sounding speech from the loudspeaker, while smooth transitions between phonemes or voice elements are attained via the insertion of noise or voice-frequency elements ensuring an even formant or frequency distribution.
-
Citations
4 Claims
-
1. A method for synthesizing speech, comprising the steps of:
-
analyzing a printed text grammatically and phonetically for sequences of phonemes, for the placement of accents, for the placement and duration of pauses and intonations to form frequency magnitude and amplitude characteristics of a sentence to be synthesized; reading out from a read-only memory, according to said frequency magnitude characteristics, binary signals coding at least in part successive magnitudes of voice-frequency functions; converting said binary signals at the output of said read-only memory into an analog signal; modulating said analog signal according to said amplitude characteristics; introducing quasirandom changes in said frequency magnitude and amplitude characteristics to facilitate the production of natural-sounding speech, the quasirandom changes introduced in said frequency magnitude and amplitude characteristics being within the limits of ±
3% for the frequency and ±
30% for the amplitude; andfeeding the modulated analog signal to the loudspeaker. - View Dependent Claims (2, 3)
-
-
4. A speech synthesizer comprising:
-
a computer for analyzing a printed text for sequences of phonemes, the placement of accents, the placement and duration of pauses and intonations fo form frequency and amplitude characteristics of a sentence to be synthesized; means for introducing into the analysis quasirandom changes in said frequency characteristics ±
3% and in said amplitude characteristics of ±
30%;a read-only memory storing binary signals coding at least in part successive amplitudes of voice-frequency functions; counting means coupled with an input of said read-only memory for addressing same, said counting means being connected to said computer for receiving therefrom according to said frequency characteritics initial addresses, rates of counting, and numbers of counts; a digital/analog converter connected to an output of said read-only memory for converting into an analog signal binary signals read from said read-only memory by said counting means; an amplifier having an input extending from said computer and another input extending from said digital/analog converter for modulating said analog signal according to said amplitude characteristics; and a loudspeaker connected to an output of said amplifier for transducing into acoustic energy the modulated signal from said amplifier.
-
Specification