Voice synthesis system utilizing a transfer function

• US 5,806,037 A
• Filed: 03/29/1995
• Issued: 09/08/1998
• Est. Priority Date: 03/29/1994
• Status: Expired due to Term

• Current Assignee: Yamaha Corporation
• Original Assignee: Yamaha Corporation
• Inventors: Sogo, Akira
• Primary Examiner(s): Dorvil, Richemond
• Application Number:
  US08/411,909
• Time in Patent Office: 1,259 Days
• Field of Search: 395/2.67, 395/2.16, 395/2.77, 395/2.7, 704/207, 704/208, 704/209, 704/258, 704/261, 704/264, 704/266, 704/268, 704/269
• US Class Current: 704/268
• CPC Class Codes: G10H 1/366 with means for modifying or correcting the external signal, e.g. pitch correction, reverberation, changing a singer's voice
• CPC Class Codes: G10H 2250/061 Allpass filters
• CPC Class Codes: G10H 2250/455 Gensound singing voices, i.e. generation of human voices for musical applications, vocal singing sounds or intelligible words at a desired pitch or with desired vocal effects, e.g. by phoneme synthesis
• CPC Class Codes: G10H 2250/581 Codebook-based waveform compression
• CPC Class Codes: G10L 13/033 Voice editing, e.g. manipulating the voice of the synthesiser
• CPC Class Codes: G10L 19/12 the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
• CPC Class Codes: G10L 21/013 Adapting to target pitch

Abstract

A voice synthesis system is fundamentally configured by a sound-source model, which simulates human voices and the like, and a voice-path model which simulates properties of voice paths between vocal cords and lips. The sound-source model is embodied by a code book which stores a plurality of code words, representative of waveform patterns, with respect to each of the voices. Each of the code words is selected by an information index. The voice-path model is embodied by a full-pole synthesis filter whose characteristic curve provides multiple poles, each of which is represented by polar coordinates. There is further provided a pitch filter and an all-pass filter. Data representative of the code word selected is supplied to the pitch filter, in which a first delay time, set by a number of delay-time units, is imparted to the data. Then, the all-pass filter imparts a second delay time, which is smaller than the delay-time unit, to the data in response to pitch-variation information. Those filters are provided to perform a fine adjustment of the pitch of the data. Thereafter, the full-pole synthesis filter performs filtering processing on the data in accordance with a coefficient which is set in response to the polar coordinates and pitch-variation information. Thus, signals indicative of synthesized sounds are produced by the full-pole synthesis filter.

Claims

1. A voice synthesis comprising:

means for providing voice information which is obtained by analyzing a voice signal, the voice information at least containing polar coordinates of a transfer function

means for converting the polar coordinates to filter coefficients; and

voice source means, having a synthesis filter with the transfer function and responsive to the filter coefficients, for reproducing the voice signal based on the voice information,

wherein the means for converting is responsive to pitch-variation information which is independent of the voice information so that the reproduced voice signal is changeable in pitch in response to the pitch-variation information independently of the voice information.

2. The voice synthesis system as defined in claim 1, wherein the voice source means includes code-book means for storing a plurality of code words representative of waveform patterns with respect to the voice signal, so that at least one code word is selected in response to an information index contained in the voice information.

3. The voice synthesis system as defined in claim 2, wherein the voice source means include pitch adjusting means for adjusting a pitch of data representative of the code word selected, in response to the pitch-variation information.

4. The voice synthesis system as defined in claim 3, wherein the pitch adjusting means includes:

a pitch filter for delaying the data by a first delay time, which is set by changing a number of delay-time units, in response to pitch information contained in the voice information; and

an all-pass filter for further delaying the data by a second delay time, which is smaller than the delay-time unit, in response to the pitch-variation information.

5. The voice synthesis system as defined in claim 3, wherein the pitch adjusting means includes:

a pitch filter for delaying the data by a first delay time, which is set by changing a number of delay-time units, in response to pitch information contained in the voice information; and

FIR filters, each of which performs filtering processing on the data in response to a FIR coefficient, which is set responsive to the pitch-variation information, so that the FIR filters as a whole further delay the data by a second delay time which is smaller than the delay-time unit.

6. The voice synthesis system as defined in claim 2, wherein the the synthesis filter is a full-pole synthesis filter for effecting full-pole-filtering processing on the code word, so as to produce a signal representative of a synthesized sound which corresponds to the voice signal.

7. The voice synthesis system as defined in claim 2, wherein the code-book means stores the code word which is represented by a time function.

8. The voice synthesis system as defined in claim 1,

wherein the means for providing voice information is part of a transmitting station,

the voice source means is part of a receiving station, and

the pitch-variation information is not received from the transmitting station, but is set at the receiving station.

9. In a voice synthesis system which comprises voice source means for reproducing a voice signal based on voice information which is obtained by analyzing the voice signal, the voice source means comprising:

code-book means for storing a plurality of code words representative of waveform patterns with respect to the voice signal, so that at least one code word is selected in response to an information index contained in the voice information;

pitch adjusting means for adjusting a pitch of data representative of the code word selected, in response to pitch variation information;

coefficient computing means for computing a coefficient based on polar coordinates and the pitch-variation information, the polar coordinates including a parameter representative of a formant frequency of a transfer function, the format frequency being varied in accordance with the pitch variation information; and

full-pole synthesis filter means, having a transfer function, for effecting full-pole-filtering processing, using the coefficient, on the code word, whose pitch has been adjusted by the pitch adjusting means, so as to produce a signal representative of a synthesized sound which corresponds to the voice signal.

10. A voice synthesis system according to claim 9 wherein the code-book means stores the code word which is represented by a time function.

11. A voice synthesis system according to claim 9 wherein the pitch adjusting means comprises:

a pitch filter for delaying the data by a first delay time, which is set by changing a number of delay-time units, in response to pitch information contained in the voice information; and

an all-pass filter for further delaying the data by a second delay time, which is smaller than the delay-time unit, in response to the pitch-variation information.

12. A voice synthesis system according to claim 9 wherein the pitch adjusting means comprises:

a pitch filter for delaying the data by a first delay time, which is set by changing a number of delay-time units, in response to pitch information contained in the voice information; and

FIR filters, each of which performs filtering processing on the data in response to an FIR coefficient, which is set responsive to the pitch-variation information, so that the FIR filters as a whole further delay the data by a second delay time which is smaller than the delay-time unit.

13. A voice synthesis system comprising:

a voice analysis device for analyzing a voice signal to generate signals representative of polar coordinates for pole locations of a transfer function of a synthesis filter, code-book information and pitch information; and

a voice source device, the voice source device including;

a pitch adjuster for providing pitch-variation information;

a code-book for storing a plurality of code words representative of waveform patterns for the voice signal, at least one of the code words being selected in response to the code book information;

a pitch filter, responsive to the pitch information and to the pitch-variation information, for adjusting a pitch of data representative of the selected code word;

a coefficient computing portion for computing filter coefficients based on the polar coordinates, the filter coefficients being varied in accordance with the pitch-variation information; and

a synthesis filter, having the transfer function and responsive to the filter coefficients, for filtering the pitch adjusted data representative of the selected code word to produce a synthesized sound signal corresponding to the voice signal.