CA1170370A - Sound synthesizer - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
A sound synthesizer in which pulses of a period indicated by a fundamental period parameter are produced by a fundamental period sound source, the output from the fundamental period sound source or the output from a noise source is selected depending on whether a sound to be synthesized is a voiced or unvoiced sound, and the selected pulses are applied to a sound synthesis filter section to synthesize the sound. The sound synthesis filter section is composed of second-order filter means which serves as a second-order filter having the zero on a unit circle in a complex plane, means for cascade-operating second-order filter means of different coef-ficients, and feedback means for feeding back the output from the synthesis filter section to the input side thereof through two kinds of such cascade-operating means, and coefficients of the second-order filters are controlled by control parameters.

Description

SOUND SYNTHESIZER

This application is a division of Serial No.
360,865 filed September 23, 1980.
BACKGROUND OF THE INVENTION
The present invention relates to a sound synthesizer with which it is possible to reconstruct a sound of substantially t~e same quality as an original sound from its features transmitted or stored in a memory in a small amount of information.
For example, in the case of reconstructing speech from feature parameters of original speech, according to the prior art the output of a pulse generator simulating the vibration of the vocal cord and the output of a noise generator simulating turbulence are changed over or mixed together depending on whether the speech is ~oiced or unvoiced and the resulting output is amplitude-modulated in accorda~ce with the speech amplitude to produce an excitation source si~nal which is applied to a filter simulating the resonance characteristics of the vocal tract to obtain synthesized speech. A synthesis system using partial auto correlation (PARCO~) coefficients and a formant synthesis system are examples of such speech synthesis system employing the featurc parameters. The former is set forth, for example, in J. D. Markel et al., "Linear Prediction of Speech", pa~es 92-128, Springer-Verlag, 1976, in which the partial auto correlation coefficients or the so-called PARCO~ cocfricicnts of a speech ~aveform are used as the feature paramcters. If the absolute values ~' 1 of the PA~COR coefficients are all smaller than unity, the speech synthesizing filter is stable. The PARCOR
coefficients may be relatively small in the amount of information for speech synthesis and the automatic extraction of the coefficients is relatively easy, but the individual parameters differ widely in the spectral sensitivity. Accordingly, if all the parameters are quantized usin~ the same number of bits, spectral dis-tortions caused by quantization errors for the respective parameters largely differ from each other. ~urther, the PARCOR coefficients are poor in their interpolation characteristics and, by the interpolation of the parame-; ters, there are produced noises, resulting in an indis-tinct speech. Especially at a low bit rate, the speech ! 15 quality is deteriorated by the spectral distortion and no satisfactory synthesized speech q~ality is obtainable.
In addition, the PARCOR coefficients do not directly correspond to spectral properties such as formant frequen-cies, and hence the PARCOR coefficients are not suitable I`or speech synthesis by rule.
The formant synthesis system is disclosed, for example, in J. L. Flanagan, "Speech Analysis, Synthesis and Perception", pa~es 339-347, Sprin~er-Verlag, 1972.
This system is one which synthesizes speech using the formant frequencies and their intensity as parameters and which is advanta~eous in that the amount Or information for the parameters may bc small and in that the ~ 170370 1 correspondence of the parameters to spectral quantities is easy to obtain. For the extraction of the formant frequency and the intensity thereof, however, it is necessary to make use of general dynamic characteristics and statistical properties of the parameters, and complete automatic extraction of the formant frequency and the intensity thereof is difficult. Accordingly, it is difficult to automatically obtain synthesized speech of high quality and it is likely to mar~edly degrade the quality of the synthesized speech by an error in the extraction of the parameters.
It is an object of the present invention to provide a sound synthesizer which is able to synthesize a sound of high quality using a small amount of infor-mation.
Another object of the present invention is toprovide a sound synthesizer which permits relatively easy extraction of the feature parameters and operates stably and in which differences in the spectral sensitivity among the parameters are small and the quantization accuracy of the parameters is the same in the case of the same quantization bits.
Another object of the present invention is to provide a sound synthesizer which is excellent in interpolation characteristics for parameters used and hence is able to obtain a synthesized sound Or high quality with a small amount of information.

_ 4 --1 Yet another object of the present invention is to provide a sound synthesizer which can be produced in a relatively simple structure.

SUMMAnY 0~ THE INVENTION
In a linear predictive analysis, the speech spectral envelope is approximated by a transfer function of ~ all-pole filter which is given by the following express~on (1):

H(Z) = ~ = -- - 2 a zP (1) p 1 + alZ + a2Z + -- + P

where Z = e ~, ~ is a normalized angular frequency

2~f~T, aT is a sampling period, f i5 a sampling frequency, p is the degree of analysis, ai (i = 1, 2, ... p) are predictor coefficients which are parameters for controlling the resonance characteristic of the rilter and o is the gain Or the filter. Here, Ap(Z) is represented ~y the sum of two polynomials which can be expressed as follows:

Ap(Z) = 1/2(P(Z) ~ Q(Z)~ (2) P(Z) = Ap(Z) _ z-zPAp(z ) (3) Q(Z) = Ap(Z) ~ Z-ZPAp(Z ) . (4) 5 (a) When the degree of analysis p is even, the expressions

(3) and (4) are factorized as follows:

(Z) = (l - z) /~ 2~os ~iz + z ) (5) p/2 2 i=l (b) When the degree of analysis p is odd, the expressions (3) and (4) are factorized as follows:

p(Z) = (1 _ z2) ll (1 - 2cos~iZ + Z ) (6) ( ) ~lll (1 -2cos iZ + Z2) ~i and i in the expressions (5) and (6) are called a line spectrum pair (hereinafter referred to as LSP) and in the present invention, they are used as parameters for rep-resenting spectral envelope ~nforma~ion.
Expresfiing Ap(Z) as given by the expression (2), the transfer function H(Z) becomes as follo~s:

H(Z) = ~ ~ p~ ) ) 2{P(Z3 - 1 + Q(Z) - 11 (7) As will be sêen from expression (7~, the transfer function H(Z3 is also formed as a filter having two feedback loops whose transfer functions are P(Z) -1 and Q(Z) -1, respectively. The transfer 1 functions P(Z) and Q(Z) are ant~-resonance circuits and thcir output become O at ~i and ~l. The frequency characteristic of A (Z) becomes as follows:

¦Ap(Z)¦ = 2P~cos 2 ~ (cos~ - cos~i)2 + sin22 n (Cs~ - cos~i) ) (8) where Z = e i~. It app~ars from the above expression (8) that in a region where adjacent line spectral frequencies are close to each other, ¦A (Z)l2 is small and the transfer function H(Z) exhibits a strong resonance characteristic. ~y changing the values of the LSP
parameters ~i and ~i describing the resonance character-istic of the transfer functions, an arbitrary speechspectral envelope can be obtained.

In the first step a speech wave of an input voice signal is A-D converte~ at sampling intevals of, for example, 0.1-0.125 m sec (8 - 10 ~Hz) to produce a sequence of samples S(n) which are passed througn a window func.ion w(n) to obtain a data sequence S'(n) + w(n) S(n). The interYal of the window function is 10 to 20 m 52C, for e~ample. Secondly, the ~ acorrelation of S'(n~ is calculated in accordance with Vp = - ~P S'(i)x S'(i+ p), N being a number of samples in the window. ~n th~:
third stop the predict~r coefficients ~p~ are calcul2ted in accordance with the followinq matrix eauation:

1 17~370 - 6a -Vl ' ' ' Vp_l ~ V~`
~ l vo ~ !~ J

The abovementioned procedure is described in "Linear Prediction: A Tutorial Review~ Proc. IEEE, vol. 63, April 1975, pages 126-129 in which expressions (17) and (37) correspond to abovesaid expression of V and the matrix equation, respectively. Thus, the predictor coefficients l~iJ i ~ 1, 2, ... p of the polynomial Ap(Z) = 1 + lZ + 2Z
+ ... + ~pZP are determined. In the fourth step the roots of the polynomials P(Z) and Q(Z) defined by the expressions ~3) and (4) respectively assuming zero are computed using the Newton method. The roots which render the polynomials + j~i + j~i P(Z) and Q(Z)to zero are represented by Z = e- and Z= e~
respectively, and the sets of angular frequencies r~i¦and are referred to as the LSP parameters.

1 utilization of the parameters representin~ the speech spectral envelope information, there can be obtained a filter whose transfer function H(Z) is equivalent to the speech spectral envelope. The transfer function of the reedback loop in the synthesis filter is provided in the form of a cascade connection of second-order filters, whose zeros are on a unit circle in a plane Z, as indicated by the expressions (5) and (6). Since these two second-order filters are identical in construction, the con-struction can also be simplified by multiple utilizationof one second-order filter using time shared operation or what is called a pipeline operation. It is also possible to perform the filter operation by the process~ng of an electronic computer without forming the second-order filters as circuits.
As described above, in the present invention the characteristics of the synthesis filter are controlled by the aforesaid parameters ~i and i but, in addition to these LSP parameters ~i and ~i~ a fundamental frequency parameter and an amplitude parameter are employed as is the case with this kind of speech synthesizers heretofore used. The fundamental frequency parameter controls a voiced sound source to generate a pulse or a group of pulses of the frequency indicated by the parameter;
the output from the voiced sound source or the output from a noise source is selected depending on ~hether the sound to bc reconstructed is voiccd or unvoiced; the selected 1 output is applied to the sound synthesis filter; and the magnitude of a signal on the input or output side of the synthesis filter is controlled by the amplitude parameters. The LSP parameters ~j and ~i are subjected to cosine transformation by parameter transformin~ means to obtain -2cos~j and -2cos~i, which are used as control parameters for controlling the coefficients of the second-order filters of the sound synthesis filter re-spectively corresponding to the parameters. The control parameters are interpolated by interpolating means in the form of the cosine-transformed LSP parameters -2cos ~i and -2cos~i. Also the interpolating means may be employed for the interpolation of the amplitude parameter. The LSP parameters ~i and ~i are excellent in interpolatability and the interpolation is conducted at time intervals equal to or twice the sampling period of the original sound for producing the parameter~; for example, the LSP parameters ~i and ~j are updated e~ery frame of 20 msec and the parameters in each frame are further interpolated every 125 ~sec in the case of 8 K~z sampling. rt is also possible to effect the interpolation in the state of the LSP parameters ~i and ~i and convert them to the control parameters.

The LSP parameters ~j and ~j are small in the amount of information per frame as compared with the control ~rameters used in a prior art synthesis filter for speech synthesis, and are excellent in interpolation g l characteristic. Accordingly, it is suitable to transmit or store the LSP parameters ~j and ~i as they are and it is also possible to convert the received or recon-structcd LSP parameters ~i and ~i to the control parameters for the synthesis filter employed in other speech synthesiz-ing systems, i.e. the PARCOR coefficients or linear pre-dictor coefficients. In this way, the LSP parameters ~i and ~i can also be used in existing speechsynthesizers. The sound synthesizer of the present in-vention is applicable to the synthesis of not only ordinaryspeech but also sounds such as a time signal tone, an alarm tone, a musical instrument sound and so forth.

BRIEF DESC~IPTION OF THE DRAWINGS
~g. 1 is a block diagram showing the funda-mental construction of an embodiment of the sound synthesizer of the present invention;
Fig. 2 is a block diagram showing a specific operat~ve ex~mple of the sound synthesizer of the present invention;
Figs. 3A, 3B and 3C are circuit diagrams respectively showing an example of a first-order or second-order filter forming a synthesis filter section;
Fig. 4A is a dia6ram illustrating an example 2~ Of the synthesis filter section where the deqree of analysis is even;
Fig. 4n is a dia6ram illustratinF an examplc 1 Of the synthesis filter section where the degree of analysis is odd;
Fig. 5 is a diagram showing the relationship between the LSP parameters ~i and ~i and the speech spectral envelope;
Fig. 6 is a circuit diagram illustrating a specific operative example of the synthesis filter section in the case of the degree of analysis being 4;
Fig. 7 is a circuit diagram illustrating a specific operative example of the synthesis filter section obtained by an equivalent conversion of the circuit shown in Fig. 6;
Fig. 8 is a circuit diagram showing a specific example of the synthesis filter section in the case of the degree of analysis being 5;
Fig. 9 is a circuit diagram showing a specific operative example of the synthesis filter section obtained by an equivalent conversion of the circuit shown in Fig. 8;
Fig. 10 is a block diagram illustrating an example of the synthesis filter section employing the pipelinc calculation system;
Figs. llA to llI, inclusive, are timing charts showing the variations of signals appearing at respective parts during the operation of the filtcr section ~cpicted in Fig. 10;
Fig. 12 is a circuit diaFram showing the case in 1 which the filtcr operation achieved by èhe operation shown in Fig. 11 is provided by a series connection of filters;
Fig. 13 is a block diagram illustrating an example of the synthesis filter using a microcomputer;
Fig. 14A is a diagram showing the variations of power with the lapse of time in the case where a speech "ba ku o N ga" was made;
Fig. 14B is a diagram showing the fluctuations in the LSP parameters ~j and ~i with the lapse of time in the case where the speech "ba ku o N ga" was made;
Fig. 15 is a diagram showing the relative frequency distributions of the LSP parameters ~i and ~i to frequency;
Fig. 16 is a diagram showing the relationship between the number of quantizing bits per frame and the spectral distortion by quantization;
Fig. 17 is a diagram showing the relationship of the spcctral distortion by inter~polation to the frame length in the case of the parameters having been inter-polated; and Fig. 18 is a diagram showing an example of synthesizing specch by converting the LSP parameters C~i and ~j to a paramcters.

~ESC~IPTION OF Tl~ P~EEE~ED EMBODIMENTS
Rcferring first to Fig. 1, thc feature 1 parameters of a speech to be synthesized are applied from an input terminal 11 to an interface section 12 every constant period of time (hereinafter referred to as the frame period), for example, every 20 msec and latched in the interface section 12. Of the parameters thus input, the LSP parameters ~i and ~i indicating spectral envelope information are provided to a parameter transforming section 13; and, of parameters indicating sound source information, amplitude information is applied to a parameter interpolating section 14 and the other parameters, ~hat is, information indicating the fundamental period (pitch) of the speech and information indicating whether the speech i5 a voiced or unvoiced sound are applied to a sound source signal generating section 15.
In the parameter transforming section 13, the input LSP parameters ~i and ~i are transformed into control parameters -2cos ~i and -2cos ~i for a synthesis filter section 16, which parameters are provided to the parameter interpolating section 14. In the parameter interpolating section 14, interpolation values for the control parameters and the sound source amplitude parameter are respectively calculated at re~ular time intervals so that the spectral envelope may un~crgo a smooth change. Thc control parame-ters thus interpolatcd .~rc supplied to thc synthesis rilter section 16, and thc sound source amplitudc parameter is applied to the sound sourcc signal gencratin~ section 15. In thc sound sourcc si~nal ~cncrating section 15, ~ 170370 1 a sound source signal depending on the features of speech is produced on the basis of the pitch information and the voiced or unvoiced sound information, and the sound source signal thus obtained is applied to the synthesis filter section 16 together with the interpolated sound source amplitude parameter. In the synthesis filter section 16, a synthesized speech is produced from the sound source signal and the control parameters. The output from the synthesis filter section 16 is provided to a digital-analog converting section 17 and derived therefrom as ananalog signal at its output terminal 18. A control section 19 generates various clocks for activating the speech synthesizer correctly and supplies them to the respective sections.

Fig. 2 illustrates in greater detail each section of Fig. 1. Every fr~me period the information on the voiced or unvoiced sound of speech is applied from the interface section 12 to a voiced sound register 23 and an unvoiced sound register 24, and a voice frequency parameter indicating the voice pitch is stored in a pitch register 25. The content of the pitch register 25 is preset in a down counter 27. The down counter 27 counts down pulses of a sampling frequency from a terminal 26 and every time its content becomes ~ero, the counter 27 presets therein thc content of the pitch register 25 and, at the same time, supplies one pulse to a gate 31.
To the gate 31 are also applied the output ~rom the voiced _ 14 -1 sound register 23 and an output pulse or pulses from a pulse generator 28, and when these inputs coincide, the content of a sound source amplitude register 34 is provided via the gate 31 to an adder 32. In other words, when the speech to be synthesized is a voiced sound, the amplitude information is applied to the adder 32 from the sound source amplitude register 34 every period of fundamental voice frequency of the pitch register 25, the amplitude information from the sound source amplitude register 34 being preset therein from the interpolating section 14.
In the case where the speech to be synthesized is an unvoiced sound, the output from the unvoiced sound register 24 and a pseudo random series pulse from a pseudo random signal generator 36 are provided to a gate 37, and upon every coincidence of both inputs, the amplitude information in the sound source amplitude register 34 is provided via the gate 37 to the adder 32. A sound source signal thus derived from the adder 32 is amplified, if necessary, by an amplifier 39 and then applied to the speech synthesis filter section 16.

In the parameter transforming section 13, the LSP parameters ~i and i and the amplitudc parameter are set in a register 21 f`rom thc interfacc section 12 evcry f`rame pcriod. The LSP parametcrs ~i and i arc applied to a parameter converter 22, wherein thcy arc transfor~ed to control parametcrs -2cos(~i and -2cosOi. Thc parametcr 1 convertcr 22 is formed, for example, by a conversion table of a read only memory (~OM), which is arran~ed so that when accessed with addresses corresponding to and ~i' -2cos~i and -2cos ~i are read out. A shift register 20 receives alternately the output from the parameter converter 22 and the arnplitude parameter stored in the register 21 and converts them to a series signal, which is applied to the parameter interpolating section 14.
In the illustrated example, the parameter interpolating section 14 is shown to perform a linear interpolation. Upon turning ON a suitch 29, the parameters of one frame are supplied to a subtractor 30, wherein a difference is detected between the parameter and that of the previous frame from an adder 33. The difference is stored in a difference value register 38 via a switch 91. Thereafter, the switch 91 is changed over to the output side of the difference value register 38 and the content thereof is circulated. At this time, the content of the difference value register 38 is taken out from bit positions higher than a predetermined bit position and supplied to the adder 33, wherein it is added to the content of an interpolation result register 92. For example, in the case of the parameter update period being 16 msec, if it is necessary to provide interpolation parameters 12~ timcs during a frame update period, then the interpolation step width is a value obtained by 1 dividing tile difference value by 128 and this is obtained by shifting the difference value in the difference value register 38 towards the lower order side by seYen bits.
The result of addition by the adder 33 is provided to the interpolation result register 92 and, at the same time, it is used as the output from the parameter inter-polating section 14. In this way, there are derived from the adder 33 the values that are obtained by sequentially adding values once, twice, three times, ... the shifted value of the difference register 38 to the parameter of the previous frame in the interpolation result register 92 every circulation of the difference value register 38.
In this example, the parameter interpolating section 14 is used for the control parameter and the amplitude parameter on a time-shared basis, so that, though not shown, the control parameter and the amplitude parameter are alternately interpolated and the inter-polation result register 92 is used in common to the both parameters. The amplitude parameter interpolated in the parameter interpolating section 14 is proYided to the amplitude information register 34 in the sound source signal generating section 15, whereas the control parame-ter interpolated as mentioned above is applied to the speech s~nthesis f`ilter section 16 as information for controlling its filter coefficient. The parameter update period, that is, thc frame period, is selected to be in the range Or 10 to 20 msec, and the interpolation period ~ 170370 is selected to range from one to two sampling intervals.
The interpolation method is not limited specifically to linear interpolation but may be other types of interpolation. The point is to ensure smooth variations of the interpolated parameters.
The synthesis filter section 16 is provided with a loop for feeding back the output through filter circuits 41 and 42 parallelly connected to each other. The filter circuits 41 and 42 are supplied with the interpolated control parameter from an input terminal 44 and the outputs from the filter circuits 41 and 42 are added together by an adder 43, the output from which is, in turn, added to the input to the filter section 16 in an adder 45. The added output therefrom is fed back to the filter circuits 41 and 42 and, at the same time, derived at an output terminal 55.
As each of the filter circuits 41 and 42, use is made of a circuit including cascade connected second-order filters each having zeros on a unit circle in a complex plane. The filter circuits 41 and 42 can be both formed by a multi-stage cascade connection of first-order and/or second-order filters. In the case of forming the filter circuits as digital filters, use can be made of a first-order filter such, for example, as shown in Fig. 3A
which is composed of a delay circuit Sl having a delay of one sample period and an adder 52 for adding the delayed output and a non-delayed input. A second-order filter such as shown in Fig. 3B can also be used 1 which is composed of two stages of delay circuits 51 and the adder 52 for adding the delayed output and the non-delayed input; and/or a second-order filter such as shown in Fig.
3C can be used in which the output from a multiplier 53 for multip~ying the delayed output from one stage of delay circuit 51 by -2cos~i, the delayed output from two stages of delay circuits 51 and the non-delayed input are added together by the adder 52. The transfer functions of the filters shown in ~igs. 3A, 3B and 3C are l+Z, l_z2 and 1-2cos~iZ+Z2, respectively. It is also possible to employ higher order filters.
The combination and the number of such filters depend on the degree of analysis; and are selected as shown in Fig. 4A or 4B depending on whether the degree of analysis is even or odd. In Fig. 4A, the degree of analysis is 10, namely, an even number and the filter circuit 41 is constituted by a series connection of a first-order filter 56 having the transfer function l-Z and second-order filters 57 to 61 each having the transfer function 1-2cos~jZ+Z2, and the output at the output terminal 55 is multiplied by +1/2 in a multiplier 63 and applied to the series circuit 56-61. The output from the second-order filter 61 of the last stage and the output from the multiplier 63 are added together by an adder 62 and the added output therefrom is provided to the adder 43. In the filter circuit 42, the output from the multiplier 63 is supplied to another series circuit consisting o~ a first-order filber 64 having the 1 1~0370 1 transfer function l+Z and second-order filters 65 to 69 each having the transfer function 1-2cos~iZIZ , and the output from the series circuit 65-73 and the output from the multiplier 63 are added together in an adder 71, the added output from which is applied to the adder 43. The multi-pliers 53 of the second-order filters 57 to 61 are re-spectively given control parameters al = -2cos~tto a5 = -2cos~5and the multipliers 53 of the second-order filters 65 to 69 are respectively given control parameters bl = -2cos ~1to b5 = -2cos~5.
Fig. 4B shows the case where the degree of analysis is 11, namely, an odd number. In the filter circuit 41, the first-order filter 56 employed in the case of ~ig. 4A is omitted but instead a second-order filter 72 having a transfer function 1_z2 is used. In the filter circuit 42, the first-order filter 64 is omitted but instead a second-order filter 73 given a parameter b6 = -2cos ~6 is used.
In the filter circuits 41 and 42 the control parameters ~i and ~i represent anti-resonance frequencies, at which the outputs from the filter circuits 41 and 42 become 0.5. Accordingly, in the case where the anti-resonance frequencies applied to the filter circuits 41 and 42 are close to each other, the output from the adder 43 becomes close to unity and the feedback loop gain approaches unity. As a consequencc, a high resonance characteristic appcars at the output terminal 55. Ilcrc, - 20 _ 1 ~1 to ~5 and ~1 to 05 are anti-resonance frequencies, which are characteristic of the speech spectral envelope information. These parameters and the spectral envelope characteristic bear a relationship of the type depicted in ~ig. 5, from which it appears that the resonance char-acteristic of the spectrum can be expressed by the spacing between adjacent parameters. These parameters have the following relationship of order:

< ~1 < ~1 < ~2 < ~2 < ~i < i < ~

The synthesizing filter has the feature that it is stable when the above condition is fulfilled.
;, Next, a description will be given of a specific operative example of the synthesis filter section 16.
Corresponding to the term in the braces of the denominator in the expression (7), the following identical equations are obtained from the expression (5):

(z) - 1 = (1 - Z) n (1 - 2 cos~iZ + z2) i=l = Z {(al + Z-) + ~ (ai+l f Z) 1l (1 + ajZ f Z ) p/2 2 - j/71(1 + ajZ + Z )} (9) p/2-l i i=l i+1 + Z) I/ (1 + b z + z 2 p/2 2 + // (L + b.Z + Z )~ (10) ,i=l J

1 1~0370 1 ai = -2cos ~i bi = -2cos~i ~

~i ' i ~ ~ (11) A digital filter is formed which has an all pole transfer function approximating the speech spectral envelope given by the expression (1) using the relationships given by the expressions (7), (9) and (10). ~ig. 6 shows the case ~here P = 4. In ~ig. 6, parts corresponding to those in Fig. 4A are identified by the same reference numerals. The input from the terminal 54 is added by the adder 45 to the output from the adder 43, and the added r output is provided to the output terminal 55 and, at the same time, multiplied by +1/2 in the multiplier 63. This 1/2 multiplication corresponds to that in the denominator in the expression (7~. The output from the multiplier 63 is applied to delay means 74 whose delay time is one sampling period, i.e. the unit time. The delayed output is applied as the input to each of the second-order filters 57 and 65, in which it is applied to the delay means Sl, the mu~tipliers 53 and the adders S2. In the two multipliers 53, the inputs thereto are respectively multi-plied by al and bl, and the multiplied outputs are each applied to an adder 94 for addition with the output from the delay mcans 51 in each Or thc rilters 57 and 65. The outputs f rom the two adders 94 are provided to a common adder 81 and, at the same time, applied to the ~dder 52 1 via delay means having a delay time of one samplin~
period in each of the filters 57 and 65. The outputs ~rom the two adders 52 are respectively applied ~s the outputs from the filters 57 and 65 to the second-order filters 58 and 66 of the next stage. The filters 58 and 66 are identical in construction with the filters 57 and 65, but the coefficients for the multipliers 53 are a2 and b2, respectively. The output from the adder 94 of each filter is applied to an adder 82 for addition with the output from the adder 81. The outputs from the adders 52 of the filters 58 and 66 are supplied to the adder 43 for subtraction from each other, and the adder 43 is further supplied with the output from the adder 82.
,.
The delay means 74 corresponds to Z outside the braces in the expressions (9) ~nd (10), and the filters 57 and 58 each constitute a second-order filter having a transfer function 1 + Z(aj + Z), and similarly the filters 65 and 66 each constitute a second-order filter having a transfer function 1 + Z(bj + Z). Accordin~ly, the series connection of the second-order filters 57 and 58 realizes the third term in the braces in the expression (9), and the delay means 51, the multiplier 53 and the adder 94 in the filter 58 realize (ai 1 + Z); consequently, by this circuit and the second-order filter 57, the second term in the braces in the expression (9) is realized, and the output is provided via the adder 82 to the adder 43. The dclay means 51, the multi.plier 53 and the adder gl~ in the /

1 second-order filter 57 realize (al + Z) and the output is supplied to the adder 43 via the adders 81 and 82.
In this way, the terms in the braces in the expression (9) are realized by the second-order filters 57 and 58 and the adders 43, 81 and 82. Likewise, the terms in the braces in the expression (10) are realized by the second-order filters 65 and 66 and the adders 43, 81 and 82. The expressions (9) and (10) differ in form only in that the signs of the third terms in the braces are different from each other, and on account of this difference, the sign of the input to the adder 43 differs. Accordin~ly, the adder 43, the second-order filters 57, 58, 65 and 66, the multiplier 63 and the delay means 74 realize the expression (2), and the circuit arrangement of Fig. 6 materializes the expression (13 as a whole. In this circuit arrangement, the expressions (9) and (10) are materialized by forming the filter circuit 41 with a series connection of (P/2)~s second-order filters 57 and 58 and the filter circuit 42 with a series con-ZO nection of (P/2)~s second-order filters 65 and 66 in the feedback loop, by taking out the nodes of the second-order filters of the filter circuit 41, that is, taps 96 and 97, from the output sides of the adders 94 to obtain the total sums with the adders 81, 82 and 83. The arrangemcnt for taking out outputs from the taps of the filter circuits will hereinafter referred to as the tap output type.
In ~ig. 6, the sccond-order f`iltcrs are arrange~

1 towards the adder 43 in an increasing order of the value i but they may also be arranged in a decreasing order of the value j. In such a case, for example, as shown in Fig. 7, the output from the delay means 74 is provided to the second-order filters 58 and 66, the outputs from which are applied ~ia the second-order filters 57 and 65 to the adder 43. In Fig. 7, the preceding stage of each second-order filter in Fig. 6 is exchanged with the succeeding stage; namely, the circuit 94 for addin~
to~ether the outputs from the delay means 51 and the multiplier 53 is exchanged with the delay means 95. The output from the delay means 74 is provided via the taps y 96 and 97 to the nodes of the second-order filters 57 and 58. In other words, the circuit arrangement of Fi~. 6 is the tap output type, whereas the circuit arrangement of ~ig. 7 is a tap input type. The circuit beginning with the tap 96 and ending with the adder 43 constitutes the first term in the braces of the expression (9), and the circuit from the tap 97 to the adder 43 constitutes the second term in the braces of the expression (9). The second-order filters 65 and 66 of the filter circuit 41 are also similarly formed. In connection with the filter circuit 41, thc output from the delay means 74 is multiplied by -1 in a multiplier 98 to nlaterialize the minus si~n for the third tcrm in the braces of the cxpression (9).
In the case where p is odd, the followin~
idcntical equation is obtained from thc cxprc~sion (8) 1 corresponding to the term in the braces of the denominator in the cxpression (7).

p(Z) 1 Z(( ) (p-3)/2( l 2 (p-1)/2 - Z 1l (1 + ajz + z2)} (12) Q(Z) - 1 = Z{(bl + Z) + ~ (bi+l + Z) i=l x n (1 + bjZ + Z )~ (13) ai = -2cos ~i~
, bi = -2cos ~i~
~ i< ~ J (14) As in the case of p being even described above, when p is odd two types of digital filters respectively called the tap output type and the tap input type are materialized in such forms as shown in Figs. 8 and 9 from the relations of the expressions (7), (12) and (13). In ~igs. 8 and 9, it is assumed that p is 5. In Figs. 8 and 9, the first-order filter 72 corrcsponds to Z in the third term in the braces of thc expression (13) and the sccond-ordcr filter 73 is to obtain such a characteristic that the pro~ucts of the transfer functions (1 + blZ + z2~ ~nd (1 + b2Z + Z ) of the filters 65 and 66 is multiplicd by (b3 + Z).
As will be understood f`rom Figs. 6 to 9, tl-e ~ 170370 1 ~1/2 multiplier 63 and the delay means 74 may also be disposed at any places in the feedback loop. Since the second-order filters are of the same type, it is possible to simplify hardware by forming the circuit arrangement so that the so-called pipeline operation is effected by using, on a time-division multiplex basis, one multiplier 53, the plurality of adders 52 and 94 and the plurality of delay means 51 and 95 making up one second-order filter.
~ig. 10 illustrates the case where the example Or the filter shown in ~ig. 12 is arranged to conduct the pipeline operation. In this example, p = 10, and an operation of a set of parameters applied from the interpolating section is i,~ completed with a period of 176 clocks. In ~ig. 10, parts corresponding to those in Fig. 12 are marked with the same reference numerals. The input side of a 16-bit static shift register 74, which performs the function Or the delay means 74 is changed over by a switch Sl between the output side of the shift register itself and the output side of the adder 45. A multiplicand input side of the multiplier 53 and the input side of the adder 52 are changed over by a switch S2 to the output side of the shift register 74, thc output side of a (27-d)th shift stage counted from the input of the shift re~ister 74 and the output sidc of a 31-bit shift register 101, ~ bein6 an operation delay of thc multiplier 53. The multiplier 53 is connected at one end to the output terminal 55 and the input side of the adder ~3~1 and derives at the other output cnd t~le 1 multiplicand input delayed by 22 clocks, which is provided to the (154 + d)-bit shift register 51. The output from an adder 81 is fed back to the input side thereof via a gate 102 and a 16-bit shift register 103, performing a cumulative addition through the adders 81 and 82 in Fig. 10. The gate 102 is opened only in the time interval between d+2 and 145+d. One input side of the adder 43 is changed over by a switch S3 between the output sides of the adders 52 and 81, and the other input side of the adder 43 is changed over by a switch S4 between the output sides of a 16th and a (d~l)th shift stages of the shift register 101. The input side of the shift register 101 is changed over by a switch S5 between the output sides of the adders 43 and 52.
The switches Sl to S5 are each connected to the fixed contact side, during one operation period, that is, 176 clocks, for a clock period indicated by numerals labelled at the fixed contact. The shift registers 51, 95, 101 and 103 are respectively of the (154+d)-bit, (175-d)-bit, 31-bit and 16-bit dynamic type are al~ays supplied witll shift clocks. The broken line input to each of the adders 43, 45, 52, 81 and 94 indicates the timing of the operation boundary Or each parameter; for example, 00 indicates a repetition every 16 clocks and an operation delay of each adder is selcctc-d to bc one clock. Fig. 11 is a timing chart of the opcration of eaC~I part in Fig. 10, Fi~. llA showin~ thc timinl~ o~` the clock, Fig. 11B the interpolated inputs of the 1 coefficients ai, bi and the interpolated amplitude A to the ~ultiplier 53 from the input terminal 44, Fig. llC the multiplicand of the multiplier 53, Fig. llD
one input to the adder 94 from the multiplier 53, Fig.
llE the other input to the adder 94, Fig. llF the output from the adder 94, ~ig. llG the output from the adder 81, and consequently the content of the register 103, ~ig.
llH the input to the adder 52 from the shift register 95, and Fig. llI the output from the adder 52. Fig. 12 shows t~ese inputs and outputs in the form of signals appearing at the respective parts in the case where the second-order filters are cascade-connected.
As shown in Fig. 11, in the period between clocks 0 and 16, a coefficient al(t) and a multiplicand xl(t) are multiplied in the multiplier 53 to effect the multiplication in the second-order filter 57 in Fig. 12, and the result of multiplication is obtained from a dth clock. In the period between clocks 16 and 32, as shown in ~igs. llB and llC, a coefficient bl(t3 and a multi-plicand yl(t) are multiplied to perform the multiplicationin the second-order filter 65. The multiplicand xl(t) is de~ayed by the shif`t register 51 along with 22 bits of the multiplier 53 by (176+d) clocks, so that as shown in Fig. 11E, a multiplicand xl(t-l) is applied to thc adder ~4 I`rom the dth clock and added with tlle output alxl derived f`rom the multiplier 53 at that time, and the added output x1~(t) is provided via the adder 81 to the 1 shift register 103 for accumulation. That is, the output from the adder 81 is supplied to the signal system of the adders 81, 82, -- in Fig. 12.
The output from the adder 94 is also provided to the (175-d)-bit shift register 95, as shown in Fig.
llH. Accordingly, in the period between the clocks 0 and 16, the output from the shift register is x1~(t-1), as shown in Fig. llH, and this output is added with the multiplicand xl(t) in the adder 52, the output x2(t) from which is applied as the input to the second-order filter 58 in Fig. 12. The output x2(t) from the adder 52 is provided via the shift register 101 to the multiplier 53. As shown in ~ig. llC, the output x2(t) is multiplied by the coefficient a2(t) in the multiplier 53 in the period between clocks 32 to 48. Prior to this multipli-cation, bl(t) and yl(t) are multiplied, as described previously, and the multiplied output is similarly processed, thereby to obtain the output y2(t) from the second-order filter 65 in the period between clocks 48 and 64. In this way, the multiplication of the coefficient a and the multiplicand x and the multiplication of the coefficient b and the multiplicand y are carricd out alternately every 16 clocks, and thc multiplied rcsults are applicd to the shift rc~istcr 51, as indicated by 1 1' lYl' a2X2' b2Y2~ -- in ~ig. llD. ~urther, thc second-order filters 57, 58, 59, 60 and 61 rcspectively dcrivc thercfrom xl'(t), x2t(t), x3'(t), x4~(t), ~5~(t) 1 and x2(t), x3(t), x4(t), x5(t)~ ~6(t), which are provided to the shift registers 95 and 101. Similarly, Yl (t) to y5'(t) and y2(t) to y6(t) are respectively obtained from the second-order filters 65 to 69, and these outputs are applied to the shift registers g5 and 101 alternately with x~(t) and x(t), respecti~ely. In the period between clocks 145 and 161, the output Y6 derived from the adder 52 at that time and X6 in the shift register provided previously are subtracted one from the other in the adder 43, and (x6-y6) is supplied via the switch S5 to the shift register 101, wherein it is delayed by (d+l) clocks. The delayed output is taken out from the switch S4 for input to the adder 43 in the period between clocks 147~d and 163+d. The output yielded from the shift register 103 at that time is provided to the adder 43 via the adder 81 and the switch S3. The output from the adder 43 at that time becomes the output from the adder 43 in Fig. 12 and this output is applied to the adder 45, wherein it is added with the input at the terminal 54 to provide Z(t). The added output Z(t) is supplied to the register 74, wherein it is delayed by the delay means 74 in ~i~. 12. The delayed output is applied to thc multi-plicr 53 and at that time the coefficient A is provided as an amplitude interpolation output at the terminal 44 and ~Z~t) is derivcd from thc multiplier 53 at the output tcrmin~l 55. This multiplication is perform~d in thc casc wherc the output rrom the synthesis ~`iltcr s~ction lG is 1 multiplied by the c~lplitude information A in a multiplier 104 in ~ig. 12. From the Shirt register 74 is taken out an output Z(t)/2 having shifted down by one bit and this is taken out via the switch S2 to the multiplier 53 as Z(t-1)/2, that is, x(t) and y(t), in the next subsequent operation period for a new set of the parameters. The output at the output terminal 55 can also be obtained as parallel outputs through an output buffer 105 of a static shift register.
The pipeline operation described above is also applicable to other types of synthesis filter section 16.
~urthermore, as will be appreciated from the arrangement of Fig. 10, the filter operation can be achieved by addition, multiplication and delay, so that this filter processing can also be effected using a microcomputer.
~or example, in Fig. 13, by successively reading out, interpreting an~ executing programs in a prograrn memory 107, a central processor unit 106 loads therein from an input port 111 a sound source signal and control parameters respectively applied from the sound source signal generating section 15 and the interpolating section 14 to terminals 108 and 109, and the central processor 106 sequentially performs the operations described previously with regard to lig. 11.
A read-writc memory 112 is used instead of the rcgisters 51, 74, 95, 101, 103 and 105 in ~ig. 10. Thc results of tlle opcrations are written in the rcad-writc memory ~12 ~d read out theref`rom at suita~lc timing to p~rlorm 1 operations. The output thus obtained is applied from an output port 113 to the output tcrminal 55. The central processor 106, the memories 107 and 112 and the ports 111 and 113 are connected to a bus 114.
By any one of the abovesaid methods the output from the synthesis filter section 16 is obtained. The output is converted by the D-A converting section 17 in Fig. 2 to an analog signal to provide a speech output.
In the D-A con~erting section 17, if the input thereto is a serial signal, then it is applied to a shift register 115 and the content of the shift register 115 is converted by a D-A converter 116 to analog form.
As described previously, the LSP parameters ~i and a i in the speech feature parameters used in the present invention can be obtained by obtaining the solutions of the expressions (5) and (6). In ~igs.
14A and 14B there are shown the results of analysis of a speech "bakuoNga" using the LSP paramcters ~i and ~i.
In ~igs. 14A and 14B, the abscissa represents time t, in Fig. 14A the ordinate represents power, and in ~ig. 14B
the ordinate represents normalized angular frequency.
Seeing instantaneous points in Fig. 14n, the frcquency rises in thc order of parametcrs 1~ 2 ~2' 05,~5 , this order does not change and thc paramctcrs i and ~i do not coincide with each othcr in one f`rame.
Accordin~ly, it is ~uaranteccl that the synthesizing filtcr sccti~n 16 is ~lways st.~l~le. The fre~luc~ncy di~tri~utions 1 of the LSP parameters i andc~i are shoun in Fig. 15, in wlliCh the abscissa represents normalized an~ular frequency f and the ordinate the relative frequency D.
As shown in Fig. 15, each parameter is no~ distributed over a wide frequency band but is restricted to a relatively narrow frequency band, so that the LSP parameters ~i and ~i can be quantized in connection with the frequency range in which they are distributed.

The LSP parameters ~i and ~i are small in quantizing distortion. Fig. 16 shows a spectral dis-tortion DS of a synthesized speech when various parameters were quantized variously, the abscissa representing the ; number of quantizing bits B per frame and the ordinate the spectral distortion Ds. The line 117 shows the case where in consideration of only the parameter distribution, the PARCOR coefficient is quantized linearly only in the coe~ficient that was distributed; the line 118 shows the case where the number of quantizing bits for the PARCOR coef-ficient was increased in consideration of the spectral sensitivity in addition to the parameter distribution in the case of the line 117, especially in the case of markedly arfectin~ the spectrum; the line 119 shows the case where thc LSP parameters ~i an~ O; were quantized in consideration of only the parameter distribution; and the line 121 shows the case wherc the LSP parameters ~i and ~j were (luantizecl in considcration of the pararncter <listri~utiol~ arl~l tl-- spectral sensitivity.

-- 3'~ --1 It will be seen from ~ig. 16 that in the case of using thc samc number of quantizing bits, the spectral dis-tortion becomes smaller in the order of the lines 117, 118, 119 and 121. Since the lines 119 and 121 are close to each other, the LSP parameters ~i and ~i are not so much affected in spectral distortion even if the spectral sensitivity is not taken into account. Accordingly, since it is sufficient to perform the quantization taking into consideration the parameter distribution range alone, the quantization is easy. ~he value that the number of quantizing bits per frame at which the spectral distortion is 1 dB in the case of the line 119 is divided by that number of quantizing bits in the case of the line 117 is 0.7. Similarly, the ratio of the number of quantizing bits per frame at which the spectral distortion is 1 dB
between the lines 118 and 121 is 0.8. ~rom this, it will be understood that the LSP parameters ~i and ~i are excellent. One dB is a difference limen of the spectral distortion of a synthesized speech.
~ig. 17 shows interpolation characteristics, the abscissa representing a frame length Tf and the ordinate the spectral distortion Ds. ~ig. 17 shows the spectral distortion oI a synthesizec~ speech iJI thc case whcre a I`ramc in which an original spcech was ~nalyz-:d in 10 mscc 2~ was used as thc refcrellce, the framc 1engtl-l was incre.~se~
to 20 to 70 Inscc and parametcrs wer~ intcrpol~ted cvcry 10 IIISCC. Thc line 1 ~ shows thc c~s~ whcr~ ~lSC was madc 1 17~370 1 of the PARCOR coefficients, and the line 123 sho~s the casc where use was made of the LSP parameters ~i and ~j.
As will be seen from Eig. 17, in the case of the same distortion, the frame length Tf can be made longer by the LSP parameters than the frame length Tf by the PARCOR coefficients, that is, the parameter update period can be increased, so that the entire amount of infor-mation can be reduced by that. In addition, since the LSP parameters are smaller than the PARCOR coefficients in the number of bits per frame, as seen from Fig. 16, the amount of information for the same distortion may be reduced by the product of the reduction ratios in Figs. 16 1 and 17; namely, in the case of the LSP parameters, the amount of information may be about 60~ of that in the case of the PARCOR coefficients.
In the case of employing the LSP parameters, it is meaningless as in the cases of other parameters that they are interpolated with a shorter period than the sample period of the original speech used in the making of the parameters. Experiments revealed that the interpolation period might be about twice or lcss the sample period of the ori~inal speech, but that when the I`ormer was about four times the latter, noises were introciuced to ma~e the synthesi~e~ speech indistinct. Accordin~ly, it is prel`crred that the interpolation pcriod be cclua1 to or twicc t~le ori~inal speech samplin~ pcriod.
As has becll dcscribcd in t~lC ~`orcgoill~, t~c 1 LSP parameters arc relativcly casy to automatically extract, and consequently can be extracted on a real time basis. Furthermore, the LSP parameters are excellent in the interpolation characteristic and small in deviation of the quantizing characteristic and permits transmission and storage of speech in a small amount of information.
In the speech syr.thesis, speech of high quality can be reconstructed and synthesized with a small amount of information, and as long as the relationship of the expression (87 holds true, the stability of the synthe-sizing filter is guaranteed.
In ~ig. 2, it is also possible to widen the spectrum by generating from the pulse generating section 28 a train of pulse groups, such as the Barker series, instead of the pulse train. The interpolating section 14 may also be provided at the preceding stage of the parameter transforming section 13. Namely, the LSP parameters from the interface section 12 may also be subjected to the cosine transformation in the parameter transforming section 13 after being interpolated. In this case, the use of a read only memory is uneconomical since its memory capacity must be enormous, accordingly, it is prefcrred to perform parannctcr conversion ~sing an approximation operation oI thc cosine ratiler than using thc reacl only memory as describcd in the example Or Fi~. ~. In Fig. 2, ttle inl`ormation indicating wlletlleI
s~eccll is a voic~d or UllVOiCC'd soullci is elltcred ancl loadccl l in the voiced sound re~ister 23 ancl the unvoiced sound re~ister 24~ but this information need not always be provided. That is, a detector circuit is provided for detectin~ ~hether the fundamental period parameter applied to the pitch register 25 is zero or not; in the case of detecting zero, the sound is considered to be an unvoiced sound and the gate 37 is opened; and in the case of other values than zero, the sound is considered to be a voiced sound and the gate 31 is opened. The control by the amplitude parameter may also be effected in connection with the output from the filter section 16, as described previously with respect to the embodiment of ~ig. 12.
In the foregoing, as the synthesis filter, use is made of a filter which includes in the feedback circuit the means for connecting in series a plurality of first-order and second-order filters of different coefficients, each having the zero on a unit circle, through utilization of the LSP parameters. ~{owever, the synthesis filter need not always be limited specifically to such a filter and the speech synthesis may also be effected by transforming the LSP parameters to some other types of parameters ~Id using other filters. For ex~nple, as shown in ~i6. 18 in which parts corrcsponding to those in li~. l are identificcl by the same refcrencc numerals, the funclamcntal period par~mctcr in tl-e l`clturc parameters applied to the interfacc section 12 is provid<cl to the S ~uncl S OUI-C`C' si6llal ~cncratin~ sectioll 15, al-cJ the 1` amplitude parameter is supplied to the interpolating section 14. The amplitude parameter thus interpolated is applied to the sound source signal generating section 15, in which it is processed as described previously in respect of Fig. 2, providing a sound source signal to the synthesis filter section 16. The LSP parameters are supplied to an LSP parameter transforming section 124, in which they are transformed to other types of parame-ters, such as an parameter, PARCOR parameter or the like. For example, from the LSP parameters are obtained polynomials P(Z) and Q(Z) using the expression (5) or (6), and fcom the polynomials the predictor coefficients a. of the transfer function H(Z) are obtained using the expressions (1) and (2). By interpolating the thus obtained predictor coefficients ai in the interpolatin~
section 14 as required, the characteristics of the sound synthesis filter section 16 are controlled. The filter section 16 is formed, for example, as a cyclic filter, in which, as shown in ~ig. 1~, the sound source signal from the sound source si~nal generating section 15 is made ~-fold by a multiplier 125 and applied to an adder 126 for subtraction from the output of an addcr 127 and the output from the adder 126 is provided to the output terminal 5;. Tlle output thus derived at thc output tcrminal 55 is applicd to a scrics circuit of` d-lay circuits Dl to D , eac}l having a dclay timc of one salllple pcriod. Thc outputs from tllc d~lay citcuit~

l Dl to D are respectively multiplied by coefficients ~1 to a from the interpolating section 14 in multi-pliers Ml to ~1 . The multiplied outputs are sequentially added and then added together in the adder 127.
It will be apparent that many modifications and variations may be effected without departing from the scope of the novel concepts of this invention.

Claims (5)

Claims:

1. A method of synthesizing sound comprising the steps of:
extracting, from an original sound signal, sound source signal parameters representing a sound source signal for energizing a synthesis filter means having a transfer function H(Z) defined by H(Z) = extracting feature parameters {.omega.i} and {.theta.i} representing the feature of the original sound signal;
transforming the feature parameters {.omega.i} and {.theta.i} to control parameters {ai} and {bi} in accordance with ai = -2 cos .omega.i and bi = -2cos .theta.i;
generating a sound source signal in accordance with the sound source signal parameters;
supplying the sound source signal to said synthesis filter means to energize it; and controlling the characteristics of said synthesis filter means by the control parameters {ai} and {bi} thereby producing a synthesized sound signal from said synthesis filter means;
the step of extracting feature parameters {.omega.i} and {.theta.i}
comprising the steps of:
sampling the original sound signal at a predetermined period .DELTA.T;

obtaining auto correlation coefficients from the samples in a predetermined time interval;
calculating predictor coefficients {.alpha.i} (i = 1, 2, ... p) from the auto correlation coefficients to determine the function Ap(Z);
calculating the feature parameters {.omega.i} and {.theta.i} which render polynomials P(Z) and Q(Z) to zero, the polynomials P(Z) and Q(Z) being defined by Ap(Z) = ?{P(Z) + Q(Z)}
P(Z) = Ap(Z) - Zp+1Ap(Z-1) Q(Z) = Ap(Z) + Zp+1Ap(Z-1) where Z = e-j.omega., .omega. is a normalized angular frequency 2.pi.f.DELTA.T
and p is the degree of analysis, and said synthesis filter means comprises two feedback paths respectively having transfer functions P(Z) - 1 and Q(Z) -1 including cascade operating second-order filter means represented by (1 + aiZ + Z2) and (1 + biZ + Z2), respectively.

2. A sound synthesizer comprising:
a sound source signal source for generating a sound source signal;
a LSP (line spectrum pair) parameter source for generating LSP parameters {.omega.i} and {.theta.i} expressed in angular frequencies respectively allowing the roots of polynomials P(Z) and Q(Z) assuming zero and defined by the following expressions:

P(Z) = Ap(Z) - Zp+1Ap(Z-1) Q(Z) = Ap(Z) + Zp+1Ap(Z-1) Ap(Z) = 1 + .alpha.iZ + .alpha.2Z2 + ? ? ? + .alpha.pZp where {.alpha.i} are predictor coefficients determined from each predetermined number of samples of soundwave signal;
parameter transforming means for transforming the LSP
parameters to control parameters different from the LSP
parameters; and a sound synthesis filter means having a transfer function H(Z) = .alpha./Ap(Z), said sound synthesis filter means being supplied with the sound source signal from said sound source signal source while the character-istics thereof are controlled by the transformed control parameters.

3. A sound synthesizer according to claim 2, wherein the sound source signal source is composed of a fundamental period sound source controlled by a fundamental period parameter to generate a pulse or a pulse group of the period indicated by the parameter, a noise source for generating random pulses, and select means for selectively taking out the output from the fundamental period sound source or the output from the noise source depending on whether a sound to be synthesized is a voice or unvoiced sound.

4. A sound synthesizer according to claim 2 or 3, further comprising amplitude control means for controlling the magnitude of a signal at the input or output side of the sound synthesis filter means by an amplitude parameter.

5. A sound synthesizer according to claim 2 or 3, wherein said parameter transforming means is means for transforming said LSP parameters .omega.i, .theta.i to predictor coefficients .alpha., and wherein said sound synthesis filter means comprises: first adder means one input of which is supplied with said sound source signal; a cascade connection of a plurality of unit time delay means, the input of said cascade connection being connected to the output of said first adder means; a plurality of multiplier means each for multiplying the output of one of said unit time delay means and corresponding one of said predictor coefficients .alpha.1;
and second adder means for summing up all the outputs from said multiplier means and supplying the sum to another input of said first adder means.