JP3039232B2 - Modulation signal generator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for generating a modulation signal for causing fluctuations in the tone of an electronic musical instrument, and more particularly to a modulation signal for generating a modulation signal which causes a natural change in tone in terms of hearing. It concerns the device.

[0002]

2. Description of the Related Art The tone of a musical tone output from a natural musical instrument has a natural fluctuation, and the tone of an electronic musical instrument is modulated in order to give a natural feeling.
Generally, a signal having regularity such as a sine wave is conventionally used as the modulation signal. FIG. 12 shows a conventional configuration of an electronic musical instrument using such a regular signal as a modulated wave.
As shown in this figure, the one shown in this figure generates a tone of a wind instrument.

In FIG. 12, a wind instrument signal forming apparatus 100 for forming a wind instrument signal includes a mouth pressure signal PRES indicating a mouth pressure (blowing pressure) at the time of playing a wind instrument, and a lip posture, a tightening and the like at the time of playing a wind instrument. The embouchure signal EMBS is applied to form a wind instrument signal. Then, the depth and speed parameters are changed to LFO (Low Frequen
cy Oscillator) The LFO is applied to the oscillator 101.
The sine wave signal generated from the oscillator 101 is added to the adder 1
03, the vibrato is applied to the embouchure signal EMBS by applying the parameters depth and speed to the LFO oscillator 102,
The sine wave signal generated from the O-oscillator 102 is added to the intraoral pressure signal PRES by the adder 104 to apply tremolo.

FIG. 14 is a block diagram of the wind instrument signal forming apparatus 100. In this figure, the tone control signal input unit 120 includes a subtractor 121 and a nonlinear conversion circuit 122. The subtractor 121 subtracts the intraoral pressure signal PRES from the waveform signal from the signal line L2 forming the return path of the waveform signal, and outputs the subtracted signal. The nonlinear conversion circuit 122 performs, for example, the nonlinear conversion of the subtraction result according to the input / output characteristics shown in FIG. Thus, a waveform signal is output to a line L1 forming an outward path. As a result, the formation of the incident wave W1 by the signal of the lead 142 fixed to the end of the mouthpiece 141 shown in FIG. 14 is simulated by the subtraction and the non-linear conversion.

That is, the subtractor 121 subtracts the lead 142 in accordance with the pressure difference Q between the intraoral pressure PRES and the pressure of the reflected wave W2 propagating from the resonance tube into the mouthpiece 141.
Shows a state in which the incident wave W1 is formed at this displacement. The nonlinear conversion circuit 122 converts the nonlinearity of the bending to the force of the lead 142 and the nonlinearity of the air flow and air pressure passing through the mouthpiece 141. The characteristics are shown. An embouchure signal EMBS is input to the nonlinear conversion circuit 122, and the input / output characteristics shown in FIG. 15 are changed by this signal.

Next, the waveform signal loop section 200 includes adders 201 and 202 inserted in the signal lines L1 and L2. The adder 201 adds the waveform signal supplied from the signal line L1 and the waveform signal supplied from the signal line L2, and outputs the added signal to the signal line L1.
Numeral 2 is for adding the waveform signal supplied from the signal line L2 and the waveform signal supplied from the signal line L1, and outputting the added signal to the signal line L2. As a result, as shown in FIG. 16, the generation state of the pressure Q is simulated at the exit of the gap formed by the mouthpiece 141 and the lead 142 as a combination of the incident wave W1 due to the input flow velocity and the reflected wave W2 from the resonance tube. .

Further, the waveform signal transmission section 300 feeds a waveform signal back to the signal line L2 on the signal line L1, and a low-pass filter 301 and a delay circuit 302 are cascaded on this feedback path. The low-pass filter 301 simulates the shape of the resonance tube, and the delay circuit 302 converts the incident wave incident from the mouthpiece 141 into a reflected wave corresponding to the length from the end of the resonance tube to the tone hole. This simulates the state of returning to the piece 141. The delay circuit 30
Reference numeral 2 indicates that the delay time is variably controlled by the pitch signal PIT so that the pitch of the generated musical tone is mainly determined. In this way, the generated signal is extracted from the line L1.

However, since the output signals of the LFO oscillators 101 and 102 for generating the modulation signal shown in FIG. 12 are regular sinusoidal signals, it is necessary to give the timbre a natural timbre change that a person plays. Can not. Therefore, it has been proposed to use random white noise as a modulation signal, as in the configuration shown in FIG. In the conventional modulation signal generator shown in FIG.
The white noise generated from the noise generator 110 is band-limited by the band-pass filter 111 and output as a modulated signal. The reason for limiting the band is that if white noise is used as it is as a modulation signal, the timbre changes will be too random and will not be natural timbre changes, so the randomness will be reduced by limiting the white noise band. It is.

[0009]

When a tone is modulated by using a modulation signal generated from the modulation signal generator shown in FIG. 13, a tone close to a natural tone can be obtained. Since the frequency characteristics of the filter are fixed, it cannot be adjusted to an intermediate state between the state where modulation is performed using white noise and the state where modulation is performed without using white noise. was there. Therefore, an object of the present invention is to provide a modulation signal generator capable of continuously controlling from a case where a modulation frequency is constant to a state where randomness due to white noise is added around the frequency.

[0010]

In order to achieve the above object, the present invention provides a filter capable of continuously lowering the quality factor (Q) of a digital bandpass filter for limiting the band of white noise from an almost infinite oscillation state. Thus, the ratio of white noise can be continuously increased. Further, in the present invention,
The noise generating means to a predetermined state according to the sounding instruction.
It is obtained by the reset to so that.

[0011]

According to the present invention, since the Q of the band-pass filter for limiting the band of white noise can be continuously reduced from infinity, an intermediate level from a state including no noise to a state including noise is obtained. It can be continuously changed to a proper state. Further, according to the present invention, since the coefficient change ratio of the filter and the ratio of noise included in the modulation signal can be changed linearly, the modulation is simple in configuration and the frequency of the modulation signal can be changed continuously. It can be a signal generator.

[0012]

FIG. 1 is a block diagram showing an electronic musical instrument equipped with a modulation signal generator according to the present invention. In this figure, the control unit 1
Is equipped with a CPU. When an event of the operator 2 or the keyboard 3 is detected by the controller 1, the wind instrument model sound source 6 is controlled to perform sound generation processing, and the parameter changed by the operator 2 is changed. The control stored in the storage unit 4 is performed. The operation element 2 includes a slider, a wheel, a pedal, and the like. When the operation of the operation element is detected by the control section 1, the data is supplied to the effect applying section 7 to change the tone. . Further, the modulation signal generator 5 can continuously change the modulation signal from a state where the modulation frequency is constant to a state where randomness due to noise is provided around the modulation frequency.

The modulated signal 10 having the randomness
Is supplied to the wind instrument model sound source 6 as an intraoral pressure signal PRES and an embouchure EMBS signal, the output of which is fluctuated, an effect is given by an effect imparting unit 7, and the left and right speakers 8 and 9 generate sound. In this case, the operator 2 is assigned as the operator 2 for changing the parameter of the modulation signal generator 5,
You may make it set speed or randomness. As the wind instrument model sound source 6, a wind instrument signal forming device as shown in FIG. 14 can be used.

Next, a first example of the modulation signal generator 5 of the present invention will be described.
An example is shown in FIG. In this figure, the noise generator 2
The white noise generated from 1 is band-limited by a digital band-pass filter 22 and output. The output from the filter 22 is multiplied by a multiplication coefficient in a multiplier 24 and output as a modulated signal. The coefficients read from the coefficient table 23 are supplied to the band-pass filter 22.
The parameter of eed and the parameter of randomness are given, and the coefficient of the band-pass filter 22 is read by the speed parameter and the randomness parameter. The depth parameter is given to the multiplier 24.

Among these parameters, the speed parameter is a parameter for changing the frequency of the modulated signal.
The randomness parameter is a parameter of a ratio of adding randomness due to noise, and the depth is a parameter of a modulation depth. In addition, the noise generator 21 and the band-pass filter 22 are reset by the note-on signal to be in an initial state so that the same tone is always generated at the time of note-on. The output frequency is changed by changing the coefficient of the band-pass filter 22 according to the speed parameter and changing the cutoff frequency. Also, by changing the coefficient of the band-pass filter 22 with the randomness parameter,
By changing the bandwidth, the noise generator 21
The ratio of the noise passing through is changed.

The band-pass filter will be further described. When the Q of the band-pass filter 22 approaches infinity, the band-pass filter 22 oscillates. FIG. 5A shows the pass characteristics of the band-pass filter 22 at this time. As shown in this figure, in this case, a steep pass characteristic is obtained, so that only a certain frequency can pass through the band-pass filter 22 and be output. Then, when the coefficient of the band-pass filter 22 is set to a different set value and its Q is reduced, as shown in FIG. 5 (b), the pass characteristics are expanded and a part of the noise signal can be passed. Therefore, a modulated signal including noise can be obtained.
The pass characteristic of the band-pass filter 22 can be continuously changed by changing its coefficient, so that the ratio including noise can be continuously changed. The depth parameter is applied to a multiplier 24 that determines the output level of the modulation signal.
The modulation depth is set by adjusting the amplitude level with the depth parameter.

Next, the digital bandpass filter 22
FIG. 3 shows an example. The digital bandpass filter shown in this figure is a known digital filter called an IIR (Infinite Impulse Response) filter, and the detailed description thereof is omitted, but the configuration is as follows. An input signal and cascaded delay elements (D0, D1) 30- each delaying the input signal by one sample time;
1 and 30-2 are used as coefficients a0, a1, and a1 by coefficient units 31-1, 31-2, and 31-3, respectively.
a2 is multiplied and applied to the adder 32. Further, cascaded delay elements (D2, D3) 33-1 and 33-, which delay the output signal of the adder 32 by one sample time, respectively.
2 are output to the coefficient units 34-1 and 34-2.
Are multiplied by the coefficients b1 and b2 and applied to the adder 32.

The frequency characteristic of this filter is determined by the coefficient unit 3
The coefficients are determined by coefficients a0 to a2 and coefficients b1 and b2 set in 1-1 to 31-3 and coefficient units 34-1 and 34-2. These coefficient groups are read from the coefficient table as shown in FIG.
The ness and speed parameters are input, and a coefficient group is read out according to these parameters. Then, by setting the read coefficient group in the digital filter shown in FIG. 3, a desired frequency characteristic can be obtained. In this coefficient table, coefficient groups 1 to n are shown in FIG.
(B), the coefficient group 1 to the coefficient group n are read from the coefficient table using the speed parameter as a horizontal address and the randomness parameter as a vertical address. .

FIG. 4 shows an example of the noise generator 21. The noise generator shown in FIG. 3A uses an M-sequence (maximum period sequence) generator, and includes an m-stage shift register 40 and an exclusive-OR circuit (EX-OR) 41.
The EX-OR 41 receives the output 43 of the m-th stage and the output 42 of the k-th stage of the shift register 40, performs an exclusive OR operation on the outputs, and returns the result to the first stage of the shift register 40. Generates an M-sequence code from the shift register 40. Since the frequency spectrum of this M-sequence code is close to the frequency spectrum of white noise, the M-sequence generator can be used as a noise generator. Note that the shift register 40 is reset to the initial value by the note on signal so that the same tone is always obtained when the note is turned on.

The noise generator shown in FIG. 4B is a noise generator called a chaos type, and a non-linear circuit 48 having a limiter characteristic for converting the output of the shift circuit 47 and an output of the non-linear circuit 48 are used. 1 to delay
A stage delay element 49, a multiplier 45 for multiplying the output of the delay element 49 by a multiplication coefficient, an adder 46 for adding an addition coefficient to the output of the multiplier 45, and shifting up the output of the adder 46 by four digits. The chaotic noise output from the non-linear circuit 48 is determined by the multiplication coefficient, the addition coefficient, and the initial value of the delay circuit 49.
Note that the delay element 49 of this chaotic noise generator is
reset to initial value by on signal
Sometimes the same tone is always pronounced.

By the way, in the modulation signal generator 5 configured as described above, the speed parameter and the randomness
It is known that when a parameter is changed linearly, a coefficient group for changing the frequency characteristic of the digital filter linearly does not change linearly. Therefore, FIG.
The coefficient group 1 to coefficient group n for linearly changing the frequency characteristic of the digital filter stored in the coefficient table shown in (1) are discrete data. Therefore, the frequency characteristic obtained by the discrete coefficient group 1 to coefficient group n An intermediate frequency characteristic cannot be obtained by calculating an interpolated value between the read coefficient groups. Therefore, since the frequency characteristic of the digital filter is a discrete characteristic due to the stored coefficient group, the transition to the timbre by the parameter changed when the parameter is changed cannot be smoothly transitioned, and a step-like change occurs. It will be migrated while doing so.

Furthermore, in order to make the Q of the digital filter almost infinite, the precision of the multiplier constituting the filter must be increased, and the number of digits handled by the high-precision multiplier becomes enormous. When the multiplier is configured by using the software, it becomes expensive. When the multiplier is configured by the software, it takes a long time to perform the calculation, and there is a possibility that the multiplier cannot cope with the real-time processing. FIG. 7 shows a second embodiment of the modulated signal generator according to the present invention in which these points are improved.

The modulation signal generator shown in FIG. 1 has two noise generators 51 and 63, and the white noise output from these noise generators 51 and 63 is sp.
The noise band is controlled by low-pass filters (LPFs) 52 and 64 whose cutoff frequencies are controlled in accordance with the eed parameter. And LPF52
The noise band-limited by the random
A multiplication coefficient corresponding to the parameter is multiplied and input to one input of the adder 56. The other input of the adder 56 receives the output converted by the converter 57 so that the frequency becomes 0.2 Hz to 200 Hz when the speed parameter is changed to 0 to 127 by an operator.
This frequency is the frequency of the signal read from the waveform memory 60 that generates the modulation signal provided at the subsequent stage.

The added output of the adder 56 is converted into input data exponentially by an accumulation data converter 58 which performs exponential conversion, and is converted into accumulated data. Applied. Then, the sine wave data stored in the waveform memory 60 is read based on the phase data output from the accumulator 59, and a modulation waveform is created. Here, the addition output is subjected to exponential function conversion because the frequency is doubled for each octave.

The noise band-limited by the LPF 64 is multiplied by a multiplier coefficient according to a randomness parameter by a multiplier 66 and input to one input of an adder 68. A constant 1.0 is input to the other end of the adder 68, and the output of the multiplier 66 is subtracted from the constant 1.0 by the adder 68 and output from the adder 68. The addition output from the adder 66 is applied to the multiplier 61 as a multiplication coefficient, and the amplitude of the waveform data read from the waveform memory 60 is fluctuated by noise. The adder 68 is provided so as to add a constant of 1.0 in order to prevent the multiplier coefficient of the multiplier 61 from being zero when the randomness parameter is set to zero and the multiplier coefficient of the multiplier 66 is set to zero. It is what is being done. The LPFs 52 and 64 are well-known primary digital LPFs.
The coefficients set in the LPFs 52 and 64 and their cutoff frequencies have a linear relationship.

Next, the operation of the modulation signal generator will be described. The speed parameter is a parameter for controlling the frequency of the waveform data read from the waveform memory 60. When the speed parameter is increased, the accumulator 59 The accumulated data to be inputted is increased, and the waveform memory 60
, A sine wave is read at a high frequency. At this time, the frequency of the read sine wave fluctuates with a system including the noise generator 51, the LPF 52, and the multiplier 54. By inputting the output of the multiplier 54 to the adder 56, the waveform memory The fluctuation is given to the phase data for reading out 60. For this reason, the modulation signal frequency read from the waveform memory 60 fluctuates.

In this case, even if the frequency read from the waveform memory 60 changes according to the speed parameter, the speed parameter is set to, for example, a conversion table 53 having an exponential characteristic in order to keep the fluctuation ratio at that frequency constant.
Is applied to the LPF 52 as a coefficient. Therefore, the change of the cutoff frequency of the LPF 52 is controlled so as to be exponential with respect to the speed parameter. Further, the frequency change of the fluctuation in human hearing is perceived to be large when the fluctuation is small, and it is perceived to be small when the fluctuation is large.
The multiplication coefficient is 4. Thus, the coefficient table 5
According to 3,55, the parameter and the change of the modulation signal obtained from the audibility are in a linear relationship.

The system composed of the noise generator 63, the LPF 64 and the multiplier 64 is a system for giving fluctuation to the amplitude of the modulation waveform. When the frequency read from the waveform memory 60 is low, the amplitude change speed is small. Since the amplitude change speed of the fluctuation is also reduced and the amplitude change speed is high when the read frequency is high, the speed change speed of the fluctuation is also exponentially converted by the conversion table 65 in order to increase the amplitude change speed of the LPF 64. Controls the cutoff frequency.

Further, since the sense of hearing with respect to the sound pressure has a logarithmic characteristic, the randomness parameter is
7 is exponentially converted into a multiplication coefficient of the multiplier 66. In this way, the coefficient tables 65 and 67 have a linear relationship between these parameters and the change in the modulated signal obtained from the audibility. The depth parameter is applied as a multiplication coefficient of a multiplier 62 provided after the multiplier 61 to control the modulation depth. By applying the note-on signal to the noise generators 51 and 63 and resetting it to the initial state, the tone changes whenever the key is turned on.

In this case, since the initial values of the noise generator 51 and the noise generator 63 in the initial state are set differently, the noises output from the noise generators 51 and 63 are slightly different noises. You. As the noise generators 51 and 63, any of the M-sequence generator and the chaotic noise generator shown in FIG. 4 can be used. Although the modulation signal generators described in the first and second embodiments described above are applied to the electronic musical instrument shown in FIG. 1, they can be applied not only to wind instrument models but also to various electronic musical instruments.

Next, a flowchart of each block of the electronic musical instrument shown in FIG. 1 will be described. First, when the power is turned on, an initial state is set in step 61 (initial setting) shown in FIG. 8, and after the initial state, the process proceeds to the panel processing in step 62. As shown in FIG. 9, the flow chart of the panel process is as follows: speed, rand set by operators such as slider, wheel, pedal, etc. in step 71.
A setting process of parameters such as omness and depth and a process of assigning an operator to these parameters are performed.

Then, in step 72, it is detected whether or not there is an ON event of an initialization switch for controlling whether or not to initialize the modulation signal generator 5, and if an ON event is detected, the INIT flag is set in step 73. Is reversed, and the process proceeds to step 74 where the process proceeds when no on-event is detected. In this step 74, setting processing relating to the timbre and the like is performed, and the process returns to the keyboard processing step 63 shown in FIG.

In this keyboard processing, as shown in the flowchart of FIG. 10, it is determined whether or not there is a key-on of the keyboard in step 8.
1, if key-on is detected, it is detected in step 82 whether or not the INIT flag is "1". If the INIT flag is "1", the modulation signal generator 5
The noise generator shown in the first and second embodiments and the accumulator shown in the second embodiment are reset to the initial state. Next, a sound generation assignment process is performed in step 84, and a sound generation process is performed in step 85. When the INIT flag is "0", the process proceeds to the sound generation assignment processing step 84 without performing the initialization processing. Further, when the key-on is not detected, it is detected whether or not there is a key-off in a step 86, and when the key-off is detected, a silencing process is performed in a step 87. Thereafter, the process returns as in the case where no key-off is detected, and shifts to the operation processing step 64 shown in FIG.

In this operation process, as shown in FIG. 11, an event of the operation is detected in step 91, and the spee
When an operator such as a slider, a wheel, or a pedal assigned to the parameters of d, randomness, and depth is operated, the operation amount is detected in step 92, and the assigned parameter is changed. Then, the operation is returned in the same way as when the event of the operator is not detected,
Other processing is performed in step 65 shown in FIG. 8, and the process returns to panel processing step 62. in this way,
The panel processing step 62, the keyboard processing 63, the operation processing step 64, and the other processing steps 65 constitute a loop, and the speed at which the loop makes one cycle is at least such that the sound is not delayed after the key is turned on.

[0035]

According to the present invention, as described above, the Q of the band-pass filter for limiting the band of white noise can be continuously reduced from infinity, so that noise is included. It can be continuously changed from a non-existing state to an intermediate state including noise.
For this reason, it is possible to provide a modulation signal generator that can smoothly change the modulation signal. . Further, according to the second embodiment of the present invention, since the coefficient change ratio of the filter and the noise-containing ratio can be linearly changed, the modulation is simple and the frequency can be continuously changed. It can be a signal generator.

[Brief description of the drawings]

FIG. 1 is a block diagram of an electronic musical instrument including a modulation signal generation device according to the present invention.

FIG. 2 is a first embodiment of a modulated wave generator according to the present invention.

FIG. 3 is a diagram showing a digital bandpass filter.

FIG. 4 is a block diagram of a noise generator.

FIG. 5 is a diagram illustrating frequency characteristics of a digital bandpass filter.

FIG. 6 is a diagram showing a coefficient table and a coefficient group stored therein.

FIG. 7 is a block diagram of a second embodiment of the modulated signal generator according to the present invention.

FIG. 8 is a view showing a flowchart of the electronic musical instrument.

FIG. 9 is a diagram showing a flowchart of panel processing.

FIG. 10 is a diagram showing a flowchart of keyboard processing.

FIG. 11 is a diagram showing a flowchart of an operator process.

FIG. 12 is a block diagram of a conventional electronic musical instrument.

FIG. 13 is a diagram showing a conventional modulation signal generation method.

FIG. 14 is a diagram showing a wind instrument signal forming device.

FIG. 15 shows input / output characteristics of a nonlinear circuit.

FIG. 16 is a diagram showing a mouthpiece.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Control part 2 Operator 3 Keyboard part 4 Storage part 5 Modulation signal generator 6 Wind instrument model sound source 7 Effect giving part 8, 9 Speaker 10, 11, 25, 112 Output modulation signal 21, 51, 63, 110 Noise generator 22 , 111 bandpass filter 23 coefficient table 24, 45, 54, 66, 61, 62 multiplier 30-1, 30-2, 33-1, 33-2, 49 delay element 31-1 to 31-3, 34- 1,34-2 Coefficient unit 32,46,56,68,103,104,201,2
02 Adder 40 Shift register 41 EX-OR 42, 43 Shift register output 44 EX-OR output 47 Shift circuit 48 Nonlinear circuit 52, 64, 301 LPF 53, 55, 65, 67 Coefficient table 57, 58 Converter 59 Accumulation Instrument 60 Waveform memory 61-65 Each step of electronic musical instrument flow 71-74 Each step of panel processing 81-87 Each step of keyboard processing 91, 92 Each step of operator processing 100 Wind instrument signal forming device 101, 102 LFO oscillator 120 Tone control signal input unit 121 Subtractor 122 Nonlinear conversion circuit 200 Waveform signal loop unit 300 Waveform signal transmission unit 302 Delay circuit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G10H 1/043 G10H 7/00 G10H 7/08

Claims (2)

(57) [Claims] And 1. A noise generator, a bandwidth of the generated noise by the noise generating means vans
A filter means for limiting the de-pass filter, a coefficient setting means for supplying said filter means coefficients for changing the quality factor (Q) of said filter means, changes the setting of the pre-Symbol coefficients for changing the quality factor (Q) changing means for controls the quality factor (Q) of the filter means in accordance with the coefficients said changing means has changed, the click
Modulation signal generating device and a control means for controlling said filtering means to the oscillation state if the previous SL coefficient is a predetermined coefficient for changing the O utility factor (Q). 2. The noise generating means according to a sounding instruction.
Is reset to a predetermined state by
The modulated signal generator according to claim 1.