JP3339273B2 - Musical sound wave forming device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a musical tone waveform forming apparatus provided with means for circulating a waveform signal.

[0002]

2. Description of the Related Art Conventionally, in order to simulate a sounding mechanism of a natural musical instrument, a musical tone waveform forming apparatus having means for circulating a waveform signal has been proposed in Japanese Patent Publication No. 7-23999. FIG. 12 shows the configuration of this musical tone waveform forming apparatus.
Speed information VE of controls such as bows supplied according to performance information
The excitation signal of the nonlinear function is read from a table stored in advance according to L and the pressing force (bow pressure) information PRES on the string of the operating element.

On the other hand, a delay circuit 23, a low-pass filter (LPF) 25, a multiplier 27, an adder 29, and a delay circuit 2
2, low-pass filter (LPF) 24, multiplier 26,
A cyclic loop is formed from the adder 28 and the excitation signal output from the non-linear table 43.
Is applied to this circulation loop to supply an excitation signal. Then, the signal circulated through the circulation loop is supplied to an adder 32 via an adder 31, where the speed information VEL of the operating element is added, and further divided by a pressing force (bow pressure) information PRES in a divider 42. The result is input to the non-linear table 43.

The excitation signal read from the non-linear table 43 in this manner is subjected to a pressing force (bow pressure) by a multiplier 44.
The information PRES is multiplied and supplied to the circulation loop. The excitation signal supplied to the circulating loop is delayed by the delay circuits 22 and 23, given predetermined frequency characteristics by the low-pass filters 24 and 25, and circulates in the circulating loop while being attenuated. The pitch of the circulating signal in this case is mainly determined by the delay circuits 22 and 23. Then, the circulating signal is extracted as a tone signal. Note that "-
The multipliers 26 and 27 multiplied by "1" simulate that the string vibration is reflected in the opposite end at the fixed end of the string.

The non-linear table 43 stores, for example, a saw-toothed positive / negative non-linear function as shown in FIG. 12 (b). The vertical axis represents the amount of movement of a string of a musical instrument such as a violin. In other words, the linearly increasing or descending portion indicates a portion where the friction between the two strings is pulled and moved as the bow moves, and the friction between them is static friction, and the preceding string resists the bow movement. The portion where the amount of movement of the string returns to a small value and becomes a substantially constant value is a portion where the friction between the two is regarded as dynamic friction. And such static friction and dynamic friction repeatedly occur on the strings,
The sawtooth-shaped non-linear function as shown in FIG.

[0006]

As described above, since the transition portion of the nonlinear function has a steep characteristic, the tone waveform output from the circulation loop includes a signal in a frequency range that does not satisfy the sampling theorem. There is a possibility that aliasing noise may occur. In particular, when the pitch is high, there is a high possibility that this phenomenon occurs. To solve this problem, as described in JP-A-4-303896, when the pitch of the musical tone is high, the characteristic of the nonlinear function changes smoothly as shown in FIG. As
It has been proposed to limit the frequency band of musical sounds. However, if the characteristic of the nonlinear function has a waveform that transitions smoothly as shown in FIG. 12C, there is a problem that the reality of the musical tone is lost because the frequency band is limited.

[0007] The pitch of the musical tone waveform output from the circulating loop is the time from the transition point of the nonlinear function to the original point and to the next transition. Therefore, when looking at the waveform of the signal output from the non-linear table, this signal is a periodic sawtooth waveform signal, and the period from the transition point of the sawtooth waveform to the next transition point corresponds to the pitch. Become like In this case, since the tone signal forming apparatus shown in FIG. 12 is configured as a digital system, the signal is a digital signal sampled at a predetermined cycle. That is, since the sampling is not always performed at the transition point when viewed as an analog signal, the apparent pitch of the digital tone signal is not changed until the transition point exceeds the next sampling point. . That is, there is a problem that a pitch lock corresponding to the sample period occurs. The effect of the pitch lock is particularly remarkable in a high frequency range.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a musical sound waveform forming apparatus which does not generate aliasing noise and does not lock the pitch while maintaining the reality of the musical sound.

[0009]

In order to achieve the above object, a musical sound waveform forming apparatus of the present invention comprises a signal circulating means for circulating a waveform signal to obtain a resonance frequency corresponding to a pitch of a musical tone to be produced. And a non-linear control means for converting the waveform signal circulating in the signal circulating means in accordance with a predetermined non-linear characteristic and outputting the converted signal to the signal circulating means. It consists of filter means for limiting the frequency band of the output of the conversion means.
Sampling value at the previous sampling timing
The last time when the value changed from the ring value after passing a predetermined level
This time of sampling Thailand from the sampling timing
Depending on the time proportion of up timing which passes a predetermined level in the timing, by the filter coefficient of said filter means is controlled, in response to said proportion the time
The digital sample signal delayed by the filter means is output from the non-linear conversion means.

According to the present invention, the non-linear control means can limit the necessary minimum frequency band, thereby preventing the generation of aliasing noise. Further, since the filter coefficient is set in accordance with the timing of the transition point approximated in an analog manner, the delay time of the filter can be changed in accordance with the timing of this transition point. Therefore, digital data in which pitch lock is prevented can be output.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First, the principle of preventing pitch locking in the musical sound waveform forming apparatus of the present invention is shown in FIG.
This will be described with reference to FIG. Here, it is assumed that the non-linear function is a step function as shown in FIG. In this step function, the input x is xs ≧
This function is “1” when x and “0” when x> xs.
That is, xs is the threshold. Consider that this step function is read out at two discrete times. The first example is shown as A, but in this example the time (ti
me) At the sampling timing of t0, the input x is x00, the step function on the y-axis read at the input x is “1”, and at the next sampling timing t1 after the time T, the input x becomes x01. As a result, the step function read by the input x becomes “0”.

The second example is shown as B. In this example, the input x is x10 at the sampling timing of the time (time) t0, and the step function read by this input x is "1". The input x becomes x11 at the next sampling timing t1 after the time T, and the step function read out by this input x becomes “0”. As described above, although the input x of the read timing is different between the first example and the second example, the read step functions are both “1” at the timing t0 and “1” at the timing t0 as shown in FIG. It becomes "0" at t1. In this way, the transition point differs from the original time position due to the sampling timing. The inaccuracy of this transition point causes aliasing noise and pitch lock.

In this case, in the first example, FIG.
A step function that transits at a timing t02 as shown in A of FIG. 7 is read out, and in the second example, B of FIG.
When the step function that transits at the timing t12 is read out as shown in FIG.
In order to represent the case where the transition point is slightly shifted from the sampling timing in a digital signal, a configuration as shown in FIG. 2 is considered. In this figure, an analog low-pass filter 1 is a filter for removing a frequency component higher than the Nyquist frequency of an input analog signal, and an analog / digital converter (A / D) 2 takes in the signal from the filter 1 The digital signal is converted at every sampling period, and the converted digital data is output.

As shown in FIG. 3, the impulse response of the analog low-pass filter 1 has a shape of y = sin (x) / x and is an infinitely continuous function. This function is called a zinc function. Accordingly, the analog low-pass filter 1 outputs a function waveform obtained by convolving the step function with the zinc function shown in FIG. The falling part of this function waveform is shown in FIG.
And the filter coefficient of the analog low-pass filter 1 in the second example B is set to be different. That is, in the first example A, the filter coefficient of the analog low-pass filter 1 has a small delay time, but in the second example B, the filter coefficient of the analog low-pass filter 1 has a large delay time. It is assumed.

As a result, the analog low-pass filter 1
From the case of the first example A, as shown in FIG.
The output value y00 is obtained at the sampling timing t0, and the output value y0 is obtained at the sampling timing t1.
1 is obtained. Further, in the case of the second example, as shown in FIG.
0 is obtained, and the output value y11 is obtained at the sampling timing t1. As described above, signals that transition at different timings between the sampling timings can be represented by digital signals corresponding to the transition timings. The present invention thus prevents pitch locking.

In the above description, the step function in which the nonlinear function falls is described. The case where the nonlinear function rises will be described with reference to FIG. FIG. 4A shows A and B having different transition timings after conversion into a step function.
And the two step function waveforms are represented as analog signals. FIG. 3B shows the timing of the sampling clock, and the sampling period is T. FIG. 4C shows the output digital data as sample points indicated by black circles on the analog waveform. FIG.
The first transition at a timing close to the timing t0 shown in FIG.
In the example A, the processing of the configuration shown in FIG. 2 is performed, and the digital data value y00 is output at the timing t0.
A digital data value y01 is output at timing t1, and a digital data value y02 is output at timing t2. Further, in the second example B in which the transition is made at a timing close to the timing t1, the processing of the configuration shown in FIG.
11. At timing t2, the digital data value y1
2 is output. In this way, a digital data value corresponding to the transition timing can be obtained.

The tone signal forming apparatus according to the present invention is based on the above principle. However, in order to control the filter coefficient of the low-pass filter in accordance with the transition timing, as shown in FIG. The ratio of the transition point time is calculated. In FIG. 5, the transition point is a moment when the input x exceeds the threshold xs, and the input x being sampled is the input x0 at the timing t0 and the timing t1 after the next sampling period T.
Is assumed to be input x1. When the timing Txs when the interpolated value obtained by linearly interpolating the input x0 and the input x1 at this time becomes the threshold xs is obtained, Txs = (xs−x0) T / (x1−x0). By setting the filter coefficient according to the timing Txs, a digital data value corresponding to the transition timing can be obtained as shown in FIG. 1 (e) and FIG. 4 (c).

FIG. 6 is a functional block diagram showing the configuration of an embodiment of the musical sound waveform forming apparatus according to the present invention. The musical sound waveform forming apparatus shown in FIG. 1 includes a performance information generator 11, a timbre information generator 12, and a tone control signal generator 13, and the performance information from the performance information generator 11 and the timbre from the timbre information generator. Based on the information, a tone control signal generated from the tone control signal generator 13 is supplied to a tone waveform signal generator 20 to form a tone waveform signal of a bowed instrument such as a violin or viola. .

Here, the performance information generating section 11 includes a keyboard composed of a plurality of keys corresponding to the scale, a key depression detecting circuit for detecting the presence or absence of a key depression operation of each key of the keyboard, and a key depression detection circuit. There are provided various circuits associated with the keyboard such as an initial touch detection circuit for detecting a key pressing operation speed and an after touch detection circuit for detecting a key pressing pressure or a key pressing depth of a pressed key. Then, performance information such as presence / absence of key depression and key depression of each key, initial touch information, and after touch information is supplied to the musical tone control signal generation unit 13.
The timbre information generation unit 12 includes a timbre selection switch and an operation detection circuit for the selection switch, and outputs timbre information representing the selected timbre to the tone control signal generation unit 13.

Further, the tone control signal generator 13 comprises, for example, a microcomputer, a tone control parameter storage table, etc., and refers to the table in accordance with the performance information and tone color information to obtain various tone control signals which do not change over time. A signal and various time-varying tone control signals are output to the tone waveform signal forming unit 20. These musical tone control signals are, for example, a first value in which the pitch of a generated musical tone determined by a key pressed on a keyboard is represented by a sum value.
A pitch signal PIT1 and a second pitch signal PIT2; a bow speed signal VEL representing a bow moving speed of a bowed instrument determined by initial touch performance information, after touch performance information and timbre information; bow pressure signal PRES which represents the pressure applied to, while being primarily determined by the tone color information, consisting of tone color control signals TC ₁ to Tc ₅ serving as a tone color control elements determined in consideration of the performance information.

When the electronic musical instrument has other performance operators such as wheels and pedals operated by a player, performance information relating to the operation of each of these performance operators is stored in the initial touch and after-touch. It may be handled as performance information similar to a touch. Further, as the performance information generating section 11 and the timbre information generating section 12, other musical instruments, automatic performance devices and the like are adopted, and the performance information and the timbre information are supplied to the musical tone control signal generating section 13 from them. Further, the various musical tone control signals may be formed in another musical instrument or an automatic performance device, and the formed musical tone control signals may be directly supplied to the musical tone waveform signal forming unit 20.

The tone waveform signal forming section 20 has a closed loop circulating loop 21 for circulating the tone waveform signal corresponding to the strings of the bowed musical instrument. In the circulating loop 21, delay circuits 22, 23 and a low-pass filter are provided. (LPF) 24, 2
5, multipliers 26 and 27 and adders 28 and 29 are connected in cascade. The delay circuits 22 and 23 use the pitch signal PIT
The delay time is variably controlled by l and PIT2. By the variable control of the delay time, the resonance frequency of the circulating loop 21 can be made to correspond to the pitch of a pressed key. The pitch of the musical tone is almost determined. The low-pass filters 24 and 25 simulate the vibration characteristics of various strings by changing the frequency characteristics of the circulating waveform signal. The characteristics are switched and controlled by the timbre control signals TC ₁ and TC _2. ing.

The multipliers 26 and 27 invert the phase of the circulating waveform signal by multiplying the circulating waveform signal by "-1", and simulate the reflection of the vibration wave at the fixed ends on both sides of the string. are doing. By the way, the circulation loop 21
The upper waveform signal is supplied to the adder 31 at each output side of the multipliers 26 and 27. This adder 31 is an adder 3
2 together with an input means, one of the inputs of the adder 32 is supplied with the output of the adder 31, and the other input of the adder 32 is used for starting and maintaining the generation of the tone waveform signal. A bow speed signal VEL is supplied as an activation control signal to be controlled. The adders 31 and 32 simulate that the contact portion of the string with the bow is displaced by the movement of the bow, and that the contact portion is displaced by the vibration wave traveling on the string.

The output of the adder 32 is
1. The signal is input to the nonlinear controller 100 via the divider 42, and the output of the nonlinear controller 100 is supplied to the circulation loop 21 via the multiplier 44 and the adders 28 and 29. The non-linear control unit 100 simulates a string displacement state due to a movement of a bow, which is an operator, by non-linearly converting an output from the adder 32. The principle of preventing the pitch lock described above is applied to the nonlinear control unit 100, so that digital data without pitch lock is output to the multiplier 44.

The bow pressure signal PRES is input to both the divider 42 and the multiplier 44. The divider 42 divides the signal input to the non-linear control unit 100 by the bow pressure signal PRES. Multiplies the output signal of the nonlinear control unit 100 by the bow pressure signal PRES. The divider 42 and the multiplier 44 simulate that the coefficient of friction changes due to the bow pressure applied to the string, thereby changing the nonlinear characteristic. This change is such that the chord speed with respect to the bow speed is similarly increased or decreased in accordance with the bow pressure. The output of the multiplier 44 is fed back to the adder 41 as a synthesizing means via a low-pass filter (LPF) 45 and a multiplier 46, and the feedback causes the nonlinear control unit including the divider 42 and the multiplier 44 to return. Hysteresis characteristics are given to the non-linear conversion of the signal by 100.

The circulating signal is supplied to the adder 28 and the delay circuit 2
3 and is connected to a formant filter 51. The formant filter 51 simulates the acoustic characteristics of the dynamic of the rubbed string instrument, the frequency characteristics are switched controlled by tone color control signal TC _5. The output from the formant filter 51 is supplied to a sound system 52, where the tone signal is converted into an analog signal, and then amplified and emitted from a speaker.

In the musical sound waveform forming apparatus thus constructed, the performance information generating section 11 supplies performance information such as key information of a performance operator, initial touch information, and after touch information to the musical tone control signal generating section 13. The tone control signal generator 13 supplies various control signals to the tone waveform signal generator 20 based on the performance information and the tone color information supplied from the tone color information generator. That is, the bow speed signal VEL is supplied to the adder 32, and the bow pressure signal PRES is supplied to the divider 42,
By being supplied to the multiplier 44, the nonlinear controller 10
0 supplies the bow speed signal VEL to the adders 28 and 29 in the circulating loop after performing non-linear processing so that pitch lock does not occur. As a result, the start signal starts to circulate in the circulation loop, but the delay time of the delay circuits 22 and 23 is changed by the pitch signal PIT
1, controlled by PIT2, the sum of these delay times is set to a value corresponding to the pitch period to be generated.

Therefore, the cycle of the circulating signal in the circulating loop coincides with the cycle of the pitch information to be generated supplied from the performance information generator 11. The circulating signal includes an LPF ₂ controlled by timbre information TC ₁ and TC _2.
4, 25, a frequency characteristic corresponding to the characteristic of the string is given. Such a circulating signal is based on the tone control information.
Formant filter 51 controlled by TC ₅
And the acoustic characteristics of the body of the bowed instrument are given to the sound system 52. As a result, the sound signal of the bowed instrument is emitted from the sound system.

Next, FIG. 7 shows an example of the configuration of the nonlinear control section 100 which performs nonlinear processing so that pitch lock does not occur. In this figure, a signal from a divider 42 is input to a non-linear function table 101, subjected to non-linear conversion, and input to a filter unit 200. This filter unit 200
Is controlled in accordance with the transition timing of the nonlinearly converted signal, so that the filter unit 20
From 0, digital data free of pitch lock as shown in FIGS. 1 (e) and 4 (c) is output,
It is output to the multiplier 44.

Here, the filter unit 200 has a FIR (Fi
delay means (D) 104-1, 104-2,... which are delayed by one sample period T connected in cascade with n stages.
, 104-n and respective delay means 104-1, 1
04-2,..., 104-n (n + 1) coefficient multipliers 105-
, 105- (n + 1) and coefficient multipliers 105-1, 105-2, ..., 105- (n
+1) to add all the outputs from the adder 106−
, 106-n.

The filter coefficients of the filter section 200 are controlled by a filter coefficient control section 103 having a filter coefficient table. Then, as described later, in the filter coefficient control unit 103, the time ratio up to the transition point with respect to the sampling period is determined from the two consecutive sampling signals input from the divider 42 in FIG.
(Xs-x0) T /
Based on (x1-x0), a predetermined filter coefficient is read from the coefficient table. Since the timing at which the filter coefficient is changed is a timing at which a discontinuous point as a transition point occurs, the timing at which the discontinuous point is generated is calculated from two consecutive sampling signals input from the divider 42 as described later. Continuous point detection unit 1
02, the output of the detection unit 102 is used as the filter coefficient change timing, and the filter coefficient control unit 1
03 is applied as a trigger signal.

An example of the function of the non-linear function table 101 is shown in the block, but this function characteristic simulates the state of displacement of the string due to the movement of the bow. That is, when the bow is rubbed against the string, when the bow speed is low, the frictional force between the bow and the string is mainly governed by the static friction coefficient, and the string speed becomes almost the same as the bow speed. As the frictional force becomes larger, the frictional force is mainly controlled by the dynamic friction coefficient, and the string speed becomes lower than the bow speed.
1 is realized by nonlinear conversion. The characteristic of the non-linear function table 101 is to be changed controlled by tone color control signal TC _3.

Next, FIG. 8 shows a detailed configuration of the discontinuous point detecting unit 102. A discontinuous point detection unit 102 shown in FIG.
When the level of a signal input to the nonlinear control unit 100 (hereinafter, referred to as a signal x) becomes a threshold xs as shown in FIG.
The discontinuous point is detected by utilizing the fact that the discontinuous point is generated. The sampled signal x1 is input to the discontinuous point detection unit 102.
1 is delayed by one sample period by the delay means (D) 110 having one sample period. Therefore, the signal x0 one sample before the signal x1 is output from the delay unit 110. The delay means 110 has a sampling clock φ
s is supplied.

Two consecutive sampling signals x1, x taken from the input side and the output side of the delay means 110
0 indicates a comparator 111 (comparator 1) and a comparator 112, respectively.
(Comparator 2). The comparator 1 compares the sampling signal x0 with the threshold xs, and when the sampling signal x0 becomes equal to or larger than the threshold xs,
Outputs a "1" signal. The comparator 2 compares the sampling signal x1 with the threshold xs, and when the sampling signal x1 becomes equal to or larger than the threshold xs,
Outputs a "1" signal. The output of the comparator 1 is an AND circuit 1
16 and one of the AND circuits 114 via the inverter 113. The output of the comparator 2 is input to the other terminal of the AND circuit 114 and also input to the other terminal of the AND circuit 116 via the inverter 115.

Therefore, the AND circuit 114 outputs x0 <xs · x1 ≧ xs. That is, "1" is output when the input signal increases and becomes equal to or larger than the threshold xs. Further, the AND circuit 116 outputs x0 ≧ xs · x1 <xs. That is, "1" is output when the input signal decreases and falls below the threshold xs.

And circuit 114 and AND circuit 1
The OR of the outputs of the OR circuit 117 is obtained by the OR circuit 117, and the trigger signal generator 118 is triggered by the rising edge of the output of the OR circuit 117, and the impulse-like trigger signal (TR
IGGER) is generated. That is, the trigger signal generator 1
A rising edge is detected by 18. As described above, in the discontinuous point detection unit 102, when the input sampling signal increases and reaches a level equal to or higher than the threshold xs, and when the input sampling signal decreases, the threshold xs When the following level is reached, a trigger signal is generated on the assumption that a discontinuous point occurs.

Next, the operation of the discontinuous point detecting section 102 will be described with reference to the flowchart shown in FIG.
Since the discontinuous point detection unit 102 requires two sampling signals, the process starts when the sampling signal x1 is input, and it is determined in step S10 whether x0 <xs and x1 ≧ xs. (Equivalent to AND circuit 114). Here, when the increased sampling signal becomes equal to or larger than the threshold xs, it is determined to be “YES” and the process proceeds to step S20 to output a trigger signal. Next, in step S25, the sampling signal x1 is set as the sampling signal x0, and the process returns. When the sampling signal x1 is input at the next sampling timing, the process is started again.

If "NO" is determined in step 10, x0≥xs and x1 <x in step S15.
s is determined (corresponding to AND circuit 116). Here, when the decreasing sampling signal becomes equal to or smaller than the threshold xs, it is determined to be “YES” and the step S is performed.
Proceeding to 20, a trigger signal is output. Next, in step S25, the sampling signal x1 is
The processing is returned to 0, and the processing is started again when the sampling signal x1 is input at the next sampling timing. As described above, the process of detecting a discontinuous point is performed at each sampling timing. If "NO" is determined in the step S15, the sampling signal x1 is set as the sampling signal x0 in the step S25, and the process returns.

The trigger signal from the discontinuous point detecting section 102 is supplied to a filter coefficient control section 103, which is used as a timing for changing the filter coefficient. FIG. 10 shows a detailed configuration of the filter coefficient control unit 103.
The filter coefficient control unit 103 calculates the time ratio of the transition point (discontinuous point) to the sampling period T based on the principle shown in FIG. In FIG. 10, the input sampling signal x is delayed by a one-sample delay unit (D) 120, and the delay unit 120 outputs a sampling signal x0 one sample cycle earlier.

Then, in the subtractor 121, (x1-x
0) is calculated and output, and the calculated output is the absolute value circuit (A
BS) 122 and outputs the signal of the absolute value signal
Input to the input terminal. Further, (xs-x0) is output from the subtractor 123, and the output is converted into an absolute value signal by an absolute value circuit (ABS) 124 and input to the B input terminal of the divider 125. In the divider 125, the signal at the B input terminal is divided by the signal at the A input terminal, and the result is output. That is, (xs-x0)
/ (X1-x0) is output from the divider 125. This output is multiplied by the sampling period T in the multiplier (NA) 126, and the time ratio (xs-x0) T / (x1-x0) of the transition point to the sampling period T is output from the multiplier 126. .

The output of this time ratio (xs-x0) T /
(X1-x0) is input to the coefficient table 127, and a filter coefficient corresponding to the time ratio is calculated in the coefficient table 127.
And supplied to the latch circuit 128. Then, in the latch circuit 128, the filter coefficient read from the coefficient table 127 is latched by a trigger signal from the discontinuous point detection unit 102 described above. The latched filter coefficient is shown in FIG.
Are supplied to the coefficient multipliers 105-1 to 105- (n + 1), and the sampling signals without pitch lock as shown in FIGS. 1 (e) and 4 (c) can be obtained.

An example of the filter coefficients stored in the coefficient table 127 will be described with reference to FIG. 11. Here, the FIR filter has the configuration shown in FIG. 11B and is supplied to the coefficient multiplier. It is assumed that the filter coefficients to be performed are four coefficients of 1, 2, 3, and 4. FIG. 11A shows an impulse response obtained by multiplying the impulse response of the FIR filter by a window function and setting the response to a finite length.
Are defined as the first set of filter coefficients, the first set of filter coefficients is symmetric with respect to the impulse response, and the center of the sampling period T That is, the time delay is a coefficient of 50%. If the points of △ 1, △ 2, △ 3, and △ 4 with △ are used as the second set of filter coefficients, the time delay is shortened by 0.5T, and the time delay is a coefficient of 0%. Further, when points □ 1, □ 2, □ 3, □ 4 with □ are set as the third set of filter coefficients, the time delay becomes 0.5T longer and the time delay becomes a coefficient of 100%.

This second filter coefficient of 0% and 10
By storing the filter coefficients between the third filter coefficients of 0% in the multiple coefficient table 127, 0 to 3
Coefficient table 1 with filter coefficient of 100% delay time
27, and digital data in which pitch lock has been prevented can be output from the filter unit 200.

Although the musical tone shaping apparatus of the present invention has been described as simulating a bowed musical instrument, the present invention is not limited to this, and simulates a wind instrument by performing non-linear characteristics and controlling and adjusting each part. And unrealistic tone synthesis is also possible. Also, the tone synthesis algorithm in the tone waveform forming apparatus is a DSP (Di
gitalSignal Processor), MPU (Micro Processing)
Unit) or the like, and may be described and executed by a microprogram or the like. Moreover, you may comprise with hardware. Further, as the performance control, an existing keyboard or the like or a control imitating a violin can be used.

[0045]

According to the present invention, the frequency band can be limited to the minimum necessary in the nonlinear control means, thereby preventing the generation of aliasing noise. Further, since the filter coefficient is set in accordance with the timing of the transition point approximated in an analog manner, the delay time of the filter can be changed in accordance with the timing of this transition point. For this reason, pitch lock of the output digital data can be prevented.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of a musical sound waveform forming apparatus according to the present invention.

FIG. 2 is a diagram showing a principle configuration of a musical sound wave forming apparatus of the present invention.

FIG. 3 is a diagram showing an impulse response of the analog filter having the configuration shown in FIG. 2;

FIG. 4 is a diagram showing an output with respect to a step input for explaining an output that is not pitch-locked in the musical sound waveform forming apparatus of the present invention.

FIG. 5 is a diagram for explaining calculation of a time ratio from a sampling input to a transition point with respect to a sampling cycle in the musical sound waveform forming apparatus of the present invention.

FIG. 6 is a functional block diagram illustrating a configuration of an embodiment of a musical sound waveform forming apparatus according to the present invention.

FIG. 7 is a block diagram showing a detailed configuration of a non-linear control unit in the musical sound waveform forming apparatus of the present invention.

FIG. 8 is a block diagram showing a detailed configuration of a discontinuous point detecting unit in the musical sound waveform forming apparatus of the present invention.

FIG. 9 is a flowchart showing an operation of a discontinuous point detecting unit in the musical sound waveform forming apparatus of the present invention.

FIG. 10 is a block diagram showing a detailed configuration of a filter coefficient control unit in the musical sound waveform forming apparatus of the present invention.

FIG. 11 is a diagram showing an impulse response for explaining a filter coefficient stored in a filter coefficient control unit and a configuration of a filter in the musical sound waveform forming apparatus of the present invention.

FIG. 12 is a functional block diagram showing a configuration of a conventional musical sound waveform forming apparatus.

[Explanation of symbols]

1 analog low-pass filter, 2 A / D converter,
21 circulating loop, 22, 23 delay circuit, 24, 25
LPF, 26, 27 multiplier, 100 nonlinear control unit, 101 nonlinear function table, 102 discontinuous point detection unit, 103 filter coefficient control unit, 200 filter unit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G10H 7/00 G10H 7/08

Claims (1)

(57) [Claims] 1. A signal circulating means for obtaining a resonance frequency corresponding to a pitch of a musical tone to be generated by circulating a waveform signal, and converting the circulating waveform signal according to a predetermined nonlinear characteristic by the signal circulating means. Te, and a non-linear control means for outputting said signal circulation means, said nonlinear control means comprises filter means for limiting the non-linear conversion means, the frequency band of the output of the nonlinear conversion means, sump at the current sampling time
The ring value is
When it changes from the sampling value through a predetermined level,
Current sampling from previous sampling timing
Timing passes a predetermined level at the timing or
Depending on the time rate at, by the filter coefficient of said filter means is controlled, depending on the percentage the time
A musical sound waveform forming apparatus wherein the digital sample signal delayed by the filter means is output from the non-linear conversion means.