Speech synthesizer

• US 6,029,128 A
• Filed: 06/13/1996
• Issued: 02/22/2000
• Est. Priority Date: 06/16/1995
• Status: Expired due to Term

• Current Assignee: Nokia Technologies Oy (Nokia Corporation)
• Original Assignee: Nokia Mobile Phones UK Limited (Nokia Corporation)
• Inventors: Jarvinen, Kari, Honkanen, Tero
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): RICHARDSON, SCO
• Application Number:
  US08/662,991
• Time in Patent Office: 1,349 Days
• Field of Search: 704/264, 704/207, 704/210, 704/222, 704/224, 704/262, 395/216, 395/217, 395/229, 395/273, 395/231, 395/271
• US Class Current: 704/220
• CPC Class Codes: G10L 19/04 using predictive techniques
• CPC Class Codes: G10L 19/26 Pre-filtering or post-filtering

Abstract

A post-processor 317 and method substantially for enhancing synthesised speech is disclosed. The post-processor 317 operates on a signal ex(n) derived from an excitation generator 211 typically comprising a fixed code book 203 and an adaptive code book 204, the signal ex(n) being formed from the addition of scaled outputs from the fixed code book 203 and adaptive code book 204. The post-processor operates on ex(n) by adding to it a scaled signal pv(n) derived from the adaptive code book 204. A gain or scale factor p is determined by the speech coefficients input to the excitation generator 211. The combined signal ex(n)+pv(n) is normalised by unit 316 and input to an LPC or speech synthesis filter 208, prior to being input to an audio processing unit 209.

Claims

1. A synthesiser for speech synthesis, comprising:

an excitation source; and

a post-processing means coupled to said excitation source for operating on a first signal including speech periodicity information derived from said excitation source, wherein the post-processing means modifies the speech periodicity information content of the first signal in accordance with a second signal derivable from said excitation source in order to produce an enhanced synthesised speech signal;

wherein the post-processing means comprises gain control means for scaling the second signal in accordance with a first scaling factor (p) derivable from pitch information associated with the first signal;

wherein the excitation source comprises a fixed code book and an adaptive code book, the first signal comprising a combination of first and second partial excitation signals respectively originating from the fixed and adaptive code books, the second signal being substantially the same as the second partial excitation signal and originating from the adaptive code book, the first signal being modified by combining the second signal with the first signal, and the first scaling factor (p) being derivable from an adaptive code book gain factor (b) in accordance with the following relationship, ##EQU13## where TH represents threshold values, b is the adaptive code book gain factor, p is the first post-processing means scale factor, a_enh is a linear scaler and f(b) is a function of the adaptive code book gain factor b, and

wherein the post-processing means further comprises an adaptive energy control means adapted to scale a modified first signal in accordance with the following relationship, ##EQU14## where N is a suitably chosen adaption period, ex(n) is the first signal, ew'"'"' (n) is a modified first signal and k is an energy scale factor.

2. A synthesiser for speech synthesis, comprising:

an excitation source; and

a post-processing means coupled to said excitation source for operating on a first signal including speech periodicity information derived from said excitation source, wherein the post-processing means modifies the speech periodicity information content of the first signal in accordance with a second signal derivable from said excitation source in order to produce an enhanced synthesised speech signal;

wherein the post-processing means comprises gain control means for scaling the second signal in accordance with a first scaling factor (p) derivable from pitch information associated with the first signal;

wherein the excitation source comprises a fixed code book and an adaptive code book, the first signal comprising a combination of first and second partial excitation signals respectively originating from the fixed and adaptive code books, the second signal being substantially the same as the first impartial excitation signal and originating from the fixed code book, the first signal being modified by combining the second signal with the first signal, and the first scaling factor (p) being derivable from an adaptive code book gain factor (b) in accordance with the following relationship, ##EQU15## where TH represents threshold values, b is the adaptive code book gain factor, p is the first post-processing means scale factor, a_enh is a linear scaler and f(b) is a function of the adaptive code book gain factor b, and

wherein the post-processing means further comprises an adaptive energy control means adapted to scale a modified first signal in accordance with the following relationship, ##EQU16## where N is a suitably chosen adaption period, ex(n) is the first signal, ew'"'"' (n) is a modified first signal and k is an energy scale factor.

3. A method for enhancing synthesised speech, comprising steps of:

deriving a first signal including speech periodicity information from an excitation source,

deriving a second signal from the excitation source, and

modifying the speech perodicity information content of the first signal in accordance with the second signal in order to produce an enhanced synthesised speech signal;

the method further comprising, scaling the second signal in accordance with a first scaling factor (p) derived from pitch information associated with the first signal;

wherein the excitation source comprises a fixed code book and an adaptive code book, the first signal comprising a combination of first and second partial excitation signals respectively originating from the fixed and adaptive code books;

wherein the first scaling factor (p) is derivable from a gain factor (b) for, the pitch information of the first signal; and

wherein the scaling factor (p) is derivable in accordance with ##EQU17## where a_enh is a constant that controls the strength of the enhancement operation, b is the gain factor for the pitch information of the first signal, TH are threshold values and p is the first scaling factor.

4. A method for enhancing synthesised speech, comprising steps of:

deriving a first signal including speech periodicity information from an excitation source, comprising a fixed code book and an adaptive code book,

the first signal comprising a combination of first and second partial excitation signals respectively originating from the fixed and adaptive code books,

deriving a second signal from the excitation source, and

modifying the speech periodicity information content of the first signal in accordance with the second signal in order to produce an enhanced synthesised speech signal,

the second signal being substantially the same as the second partial excitation signal and originating from the adaptive code book, the first signal being modified by combining the second signal with the first signal, and a first scaling factor (p) being derivable from an adaptive code book scaling factor (b) in accordance with the following relationship, ##EQU18## where TH represents threshold values, a_enh is a linear scaler and f(b) is a function of b,

wherein the modified first signal is normalised in accordance with the following relationship, ##EQU19## where N is a suitably chosen adaption period, ex(n) is the first signal, ew'"'"'(n) is a modified first signal and k is an energy scale factor.

5. A method for enhancing synthesised speech, comprising steps of:

deriving a first signal including speech periodicity information from an excitation source, comprising a fixed code book and an adaptive code book,

the first signal comprising a combination of first and second partial excitation signals respectively originating from the fixed and adaptive code books,

deriving a second signal from the excitation source, and

modifying the speech periodicity information content of the first signal in accordance with the second signal in order to produce an enhanced synthesised speech signal,

the second signal being substantially the same as the first partial excitation signal and originating from the fixed code book, the first signal being modified by combining the second signal with the first signal, and a first scaling factor (p) being derivable from an adaptive code book scaling factor (b) in accordance with the following relationship, ##EQU20## where TH represents threshold values, a_enh is a linear scaler and f(b) is a function of b,

wherein the modified first signal is normalised in accordance with the following relationship, ##EQU21## where N is a suitably chosen adaption period, ex(n) is the first signal, ew'"'"' (n) is a modified first signal and k is an energy scale factor.

6. A synthesiser for speech synthesis, comprising first and second excitation sources for respectively generating first and second excitation signals, and modifying means for modifying the second excitation signal in accordance with a scaling factor derivable from pitch information associated with the first excitation signal in order to produce an enhanced synthesised speech signal, wherein the modifying means scales the second excitation signal in accordance with a scaling factor (a'"'"') derivable from pitch information associated with the first signal, wherein the first excitation source is an adaptive code book and the second equitation source is a fixed code book, and wherein the scaling factor (a'"'"') satisfies the following relationship;

##EQU22## where g is a fixed code book gain factor, b is an adaptive code gain factor and p is a perceptual enhancement gain factor, wherein the perceptual enhancement gain factor p is derivable in accordance with;

##EQU23## where a_enh is a constant that controls the strength of the enhancement operation and TH are threshold values.

7. A synthesiser according to claim 6, wherein the first and second excitation signals are combined after modification.

8. A synthesiser for speech synthesis, comprising first and second excitation sources for respectively generating first and second excitation signals, and modifying means for modifying the first excitation signal in accordance with a scaling factor derivable from pitch information associated with the first excitation signal in order to produce an enhanced synthesised speech signal, wherein the modifying means scales the first excitation signal in accordance with a scaling factor (a) derivable from pitch information associated with the first signal, wherein the first excitation source is an adaptive code book and the second excitation source is a fixed code book, wherein the scaling factor (a) is of the form a=b+p, where b is an adaptive code book gain and p is a perceptual enhancement gain factor derivable in accordance with the following relationships;

##EQU24## where TH represents threshold values, a_enh is a linear scaler and f(b) is a function of gain b,

wherein the first and second excitation signals are combined after modification, and

further comprising an adaptive energy control means for modifying combined scaled first and second signals in accordance with the following relationship;

##EQU25## where N is a suitable adaption period, ex(n) is the combined first and second signals, ew'"'"' (n) is the combined scaled first and second signals and K is an energy scale factor.

9. A synthesiser for speech synthesis, comprising first and second excitation sources for respectively generating first and second excitation signals, and modifying means for modifying the second excitation signal in accordance with a scaling factor derivable from pitch information associated with the first excitation signal in order to produce an enhanced synthesised speech signal, wherein the modifying means scales the second excitation signal in accordance with a scaling factor (a'"'"') derivable from pitch information associated with the first signal, wherein the first excitation source is an adaptive code book and the second excitation source is a fixed code book,

wherein the scaling factor (a'"'"') satisfies the following relationship;

##EQU26## where g is a fixed code book gain factor, b is an adaptive code gain factor and p is a perceptual enhancement gain factor, wherein the perceptual enhancement gain factor p is derivable in accordance with;

##EQU27## where a_enh is a constant that controls the strength of the enhancement operation and TH are threshold values,

wherein the first and second excitation signals are combined after modification, and

further comprising an adaptive energy control means for modifying combined scaled first and second signals in accordance with the following relationship;

##EQU28## where N is a suitable adaption period, ex(n) is the combined first and second signals, ew'"'"' (n) is the combined scaled first and second signals and K is an energy scale factor.

10. A synthesiser for speech synthesis, comprising:

an excitation source; and

a post-processing means coupled to said excitation source for operating on a first signal including speech periodicity information derived from said excitation source, wherein the post-processing means modifies the speech periodicity information content of the first signal in accordance with a second signal derivable from said excitation source in order to produce an enhanced synthesised speech signal;

wherein the post-processing means comprises gain control means for scaling the second signal in accordance with a first scaling factor (p) derivable from pitch information associated with the first signal;

wherein the excitation source comprises a fixed code book and an adaptive code book, the first signal comprising a combination of first and second partial excitation signals respectively originating from the fixed and adaptive code books;

wherein the first scaling factor (p) is derivable from an adaptive code book gain factor (b);

and wherein the scaling factor (p) is derivable in accordance with the relationships, ##EQU29## where a_enh is a constant that controls the strength of the enhancement operation, b is the adaptive code book gain factor, TH are threshold values and p is the first post-processing means scale factor.

11. A synthesiser for speech synthesis, comprising:

first and second excitation sources for respectively generating first and second excitation signals, and

modifying means for modifying the first excitation signal in accordance with a scaling factor derivable from pitch information associated with the first excitation signal in order to produce an enhanced synthesised speech signal,

wherein the modifying means scales the first excitation signal in accordance with a scaling factor (a) derivable from pitch information associated with the first signal,

wherein the first excitation source is an adaptive code book and the second excitation source is a fixed code book,

wherein the scaling factor (a) is of the form a=b+p, where b is an adaptive code book gain and p is a perceptual enhancement gain factor, and wherein the perceptual enhancement gain factor p is derivable in accordance with the relationships;

##EQU30## where a_enh is a constant that controls the strength of the enhancement operation and TH are threshold values.

12. A synthesiser for speech synthesis, comprising;

an input unit for inputting a signal and for extracting coded information from said signal, the coded information comprising fixed codebook and adaptive codebook parameters, including an adaptive codebook gain factor;

an excitation source comprising a fixed codebook and an adaptive codebook and having inputs coupled to outputs of said input unit for receiving extracted coded information therefrom, said excitation source being responsive to the received extracted coded information for outputting a first partial excitation signal from said fixed codebook and a second partial excitation signal from said adaptive codebook, said excitation source further comprising means for combining said first and second partial excitation signals into a composite excitation signal; and

a perceptual enhancement post-processor coupled to said excitation source for operating on said composite excitation signal by combining said composite excitation signal with a scaled version of said second partial excitation signal, wherein an amount of scaling of said second partial excitation signal is controlled by a scalincg factor having a value that is function of a value of said adaptive codebook gain factor;

wherein said scaling factor (p) is derived from said adaptive code book gain factor (b) in accordance with the relationships, ##EQU31## where a_enh is a constant that controls a strength of perceptual enhancement and TH are threshold values.