JP2661599B2 - Music signal processor - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention
Music signal processing device using
Abundant filters by creating filter coefficients
It relates to the fact that characteristics can be realized. 2. Description of the Related Art A digital audio signal for controlling a tone signal is provided.
By changing the filter coefficient of the filter
An electronic musical instrument that realizes color change over time is disclosed in
No. 60-52895. Where,
Filter coefficients stored in memory for a given time frame
And stores the corresponding file from the memory for each time frame.
Read the filter coefficients and input them to the digital filter
In this way, the filter coefficient can be switched in frame units.
And change the filter characteristics on a frame-by-frame basis.
I'm trying. Also, for key touch, key scaling, etc.
Let's variably control the filter characteristics for appropriate tone control
In the past, for each filter characteristic that can be realized,
Should filter coefficients be stored in memory in advance
Was. [0003] However, the prior art described above
In surgery, the filter coefficients change during one time frame
Not so, the time length of one time frame is relatively long
In that case, the tone change during that time cannot be obtained and it becomes monotonous.
There was a problem that. Also the change of time frame
Tone change due to sudden change of filter coefficient by eyes
There was also a problem that it became unnatural. One time frame
If you shorten the time length of the
Although it is possible to do so,
There is a problem that the capacity of the moly increases. Also,
Perform tone control according to key touch, key scaling, etc.
If you try to perform rich tone change control,
A large number of sets of filter coefficients are stored in memory in advance.
And the capacity of the filter coefficient memory increases.
Problem arises. The present invention has been made in view of the above points.
To increase the amount of filter coefficients prepared in advance.
Abundant filter characteristics are realized.
Tone signal processing device capable of realizing accurate tone change control
To provide a device. [0004] To solve the above problems,
Music signal processing apparatus according to the present invention according to the first aspect of the present invention
Inputs the tone signal and the filter coefficient, and
Data for controlling the input tone signal in accordance with the characteristic corresponding to the coefficient.
Digital filter means and desired filter characteristics
Filter coefficient storage means storing a plurality of sets of filter coefficients
Stage and the filter to be read from the filter coefficient storage means.
Criteria for setting reference value data for specifying the filter coefficient
Value data setting means and control data for variably controlling the tone color.
Control data generating means for generating data;
Data and control data, and as a result of the
Calculating means for obtaining filter coefficient calculation data;
Filtering a plurality of filter coefficients according to operation data
It reads out from the coefficient storage means and the plurality of read out
Filter coefficients are interpolated according to the filter coefficient calculation data
And the filter coefficients obtained by interpolation are
Filter coefficient interpolating means to be applied to the filter means.
Things. According to the tone signal processing apparatus having the above-mentioned structure,
For example, a plurality of sets of filter
The number is stored in the filter coefficient storage means.
The filter coefficients to be read from the filter coefficient storage means are based on
Basically, the reference value set by the reference value data setting means
Specified by value data. Separate from this reference value data
The control data for variably controlling the tone is
Generated by the data generating means. In the calculation means, the reference
Calculates the value data and control data, and calculates
Thus, filter coefficient calculation data is obtained. Thus, the reference value
The filter coefficient is the data changed by the control data.
Obtained as operation data. This filter coefficient calculation data
A plurality of filter coefficients are stored according to the filter coefficient storage means.
From the filter coefficient calculation data.
As the inter-parameter, the plurality of read filter
The filter coefficients obtained by interpolation are
Digital filter means. Thus, the reference value
By calculating the data and control data, in other words, both
, The filter coefficients to be read and
Filter coefficients to determine the interpolation parameters
Calculation data and generate filter coefficients based on interpolation calculation
I am trying to do it. Therefore, according to the present invention, the interpolation calculation
Generates precise filter coefficients according to the tone control content.
Can be stored in the filter coefficient storage means in advance.
Without increasing the amount of filter coefficients to keep in mind,
Filter characteristics, which allows for rich tone change control.
It can be realized. Also, filter coefficients
Basically, filter coefficients to be read from the storage means
Is generally specified by the reference value data.
Various variations are available depending on the data settings.
The storage contents of the filter coefficient storage means can be used,
This also enables rich tone change control.
I will be able to. According to a second aspect for solving the above problems,
The musical sound signal processing device according to the present invention includes a musical sound signal and a filter.
Input the filter coefficient and obtain the characteristic according to the filter coefficient.
Therefore, a digital filter means for controlling the input tone signal is provided.
And a filter section for storing a plurality of sets of filter coefficients, respectively.
Number storage means, and for each tone of the tone signal,
Filter for at least one of range and touch
Filter coefficient specification information that stores specification information for specifying coefficients
Information storage means, wherein the designated information includes different timbres,
Common filter coefficient can be specified for area or touch
Timbre information indicating the tone of the musical tone to be generated
Range information indicating the range of the tone signal to be generated and
At least one of touch information indicating a touch of a musical tone to be generated
The information generating means to generate, and the information generating means
At least two sets of filter coefficients according to the information generated
The designation information for designating the filter coefficient designation information
Read from the storage means and filter based on this specified information
A filter for reading coefficients from the filter coefficient storage means
A coefficient reading means and a control signal for controlling a tone
From the control signal generating means,
Using the generated control signal as an interpolation parameter,
The at least two sets read by the filter coefficient reading means.
Filter coefficients obtained by interpolation
A filter section for providing coefficients to the digital filter means
It is provided with number interpolation means. According to the tone signal processing apparatus having the above-mentioned structure,
For example, a filter coefficient that stores a plurality of sets of filter coefficients
Designation information that specifies not only storage means but also filter coefficients
Also has filter coefficient designation information storage means for storing
I have. This filter coefficient designation information storage means stores the tone signal
Of the tone signal and the range of the touch
Specification information that specifies the filter coefficient for at least one of each
Where the specified information is a different
Common filter for tone, range or touch
A number can be specified. From the information generation means,
Tone information indicating the tone of the musical tone to be generated,
Range information indicating the range of the sound signal and the type of musical tone to be generated.
At least one of the touch information indicating the switch is generated.
You. According to the information generated by this information generating means
In both cases, the specification information for specifying two sets of filter coefficients is not
Read from the filter coefficient specification information storage means, and
Filter coefficients are stored from the filter coefficient storage means based on the information.
Is read. At least two sets of
Filter coefficient interpolates control signal for tone change control
Interpolated parameters obtained as parameters
The filter coefficients are used in the digital filter means. Therefore, also in this case, by the interpolation operation,
Generates precise filter coefficients according to the tone control
Can be stored in the filter coefficient storage means in advance.
Without having to increase the amount of filter coefficients
Filter characteristics, which enables rich tone change control.
Will be able to manifest itself. Also, depending on the timbre information
Instead of reading the filter coefficient storage means directly,
Read the specified information for specifying the filter coefficient, and
Read the filter coefficient storage means according to the specified information
The indirect read method is used. This allows
Filter coefficients stored in the filter coefficient storage means
It can be read and used indirectly in common with tone information.
Can be relatively stored in the filter coefficient storage means.
The number of filter coefficients can be reduced.
It is said that the storage capacity of the Luta coefficient can be relatively reduced.
It has excellent effects. This effect is due to the filter coefficient interpolation
Coupled with the relative reduction of filter coefficient storage capacity due to
To further reduce the filter coefficient storage capacity
You. Further, the filter coefficient designation information storage means stores
The specified information stored for each tone color of the tone signal is stored.
And at least one of the range and touch of the tone signal.
Are stored for each person, and
Some of the specified information includes different tones,
Includes those that specify common filter coefficients for
It is also characterized by. That is, one tone
Each of the plurality of ranges and / or each of the plurality of touches
Correspondingly, a plurality of designated information is stored,
Are stored for each tone.
However, many such designations are all completely separate files.
It is not necessary to specify the filter coefficient.
Data factor can be specified.
The information may specify common filter coefficients. this
The number of independent variables for reading out the filter coefficients
Can reduce the filter coefficient storage capacity relatively
That is, for each tone of the tone signal and
Designated for at least one of musical range and touch
Increased versatility of tone control by storing information
(Independent variable for reading out the filter coefficients)
While the filter coefficient storage capacity is
It can be said that it does not increase so much
It has the effect. The control signal generated by the control signal generating means is
Signals include time elapsed, manual control output, envelope
Waveform data, low frequency modulation signal, and key touch and pitch
Or, do not include it in the readout
Tone control based on some tone control elements
It may be your information. For example, filter characteristics
Try to achieve time-varying timbre by changing
The filter coefficient supply means in different time frames.
It is also possible to supply each filter coefficient correspondingly.
No. In that case, the control signal generation means
A control signal whose value changes may be generated.
As a result, the filter coefficient interpolation means
Filling of different time frames supplied by feeding means
Control, which is a time-varying interpolation parameter
Interpolate according to the signal. As a result, the interpolation output
The filter coefficient supplied to the digital filter is a certain time
Another time frame from the filter coefficients corresponding to the frame
Up to the filter coefficient corresponding to
The value of the crab changes, resulting in a smooth temporal change in tone.
You. Further, the filter characteristic is changed with time.
Use the filter coefficient interpolation method even when it is not necessary.
Can be used. For example, fill
By variably controlling the sound characteristics, the sound
When performing color change control, filter coefficient supply means
Filter coefficients prepared in advance by
Filter coefficient corresponding to a key
That there are only at least two sets of filter coefficients
Even the control signal according to the degree of multi-step key touch during that
A control signal is generated from the generating means, and this control signal is interpolated.
By performing filter coefficient interpolation as a parameter.
To obtain precise filter coefficients according to the multi-step key touch.
Rich tone change control according to key touch
Can be realized. The control signal is pitch or range,
Manual control output, envelope waveform data, low frequency modulation
Signals generated based on other tone control elements such as signals
The same applies to the case. BRIEF DESCRIPTION OF THE DRAWINGS FIG.
The embodiments of the present invention will be described in detail. FIG. 1 illustrates the present invention.
Shows the hardware configuration of one embodiment of an applied electronic musical instrument
In this embodiment, the filter corresponding to different time frames is used.
Fill by interpolating the filter coefficients over time
The time characteristic of the data characteristic is changed. This
In the electronic musical instrument of the embodiment, the CPU (central processing unit) is used.
) 11, program ROM (read only memo)
12) and data and working RAM (random
Microcomputer unit 10 including access memory 13
Various operations including interpolation of filter coefficients
Processing is controlled. The basics of filter coefficient interpolation calculation
Timer counter 14 for setting the timing
It is provided in the microcomputer section 10. This Thailand
The counter 14 is activated at a predetermined time by a predetermined clock pulse.
It is incremented every interval (for example, every 2 ms). The keyboard 15 designates a pitch of a musical tone to be generated.
Have multiple keys to do. Touch detection circuit 16
Is used to detect the touch of a key pressed on the keyboard 15.
For example, a key touch is detected with a 32-step resolution.
You. The type of touch to be detected is the initial touch or
May be either aftertouch. Filter modus
Switch 17 has one of two filter coefficient interpolation modes.
Or to choose. In this example,
The interpolation of the filter coefficient is performed in one of the following two modes.
Can be performed selectively,
Mode selection is performed by the filter mode switch 17.
U. The first mode (this is called mode "0")
), Two files corresponding to adjacent time frames
An interpolation operation is performed based on the difference between the filter coefficients. Second mode
(This is called mode “1”).
At least one of the two filter coefficients corresponding to the
An interpolation operation is performed based on the coefficient difference value prepared in advance.
Details of these modes will become clear later. Operation panel
A tone selection switch 19 and various other music
A switch for sound setting / control and an operator are provided. The filter parameter memory 20 has a plurality of
A set of filter coefficients corresponding to a time frame is stored in advance.
The set of such filter coefficients is
Timbre, pitch, key touch, etc.
I remember a lot. Selected by the tone selection switch 19
The selected tone type, the pitch of the key pressed on the keyboard 15, and
Key touch of the pressed key detected by the touch detection circuit 16
Filter coefficients corresponding to multiple time frames
When a set of filter coefficients consisting of
Filter coefficients corresponding to each time frame as time passes
Are sequentially read from the memory 20. Each read
The filter coefficient of the frame is calculated by the microcomputer unit 10.
Given to. In the microcomputer unit 10, the adjacent
Calculation to interpolate the filter coefficients of the time frame
Therefore, the interpolation output file obtained at each interpolation step
Filter coefficients via the data and address bus 21
The signal is supplied to the digital filter 23 in the signal generation circuit 22.
You. This filter parameter memory 20 is used for an electronic musical instrument.
The internal ROM may be built-in,
Removable by ROM or other external storage media
It may be a thing. The tone signal generating circuit 22 includes data and add
Information on the pressed key given via the
Key and key-on signal) to determine the pitch corresponding to the pressed key.
Generates a digital tone signal having
The signal is input to the digital filter 23 and the filter
It performs tone control according to characteristics. Tone signal generation
What type of tone signal generation method is used
May be. For example, the sound specified by the key code
Responds to phase address data that changes at a rate corresponding to high
The tone waveform sample value data stored in the waveform memory
Sequential reading method (memory reading method)
Predetermined phase address data as phase angle parameter data
Performs the frequency modulation operation of the tone waveform sample value data
(FM method) or the phase address data
Data as phase angle parameter data
Calculation of musical sound waveform sample value data by performing arithmetic
(AM system), or any other known system may be used.
No. When the memory reading method is used,
The tone waveform stored in the memory may be only a one-cycle waveform.
However, a multi-period waveform is preferable because sound quality can be improved.
Good. Store multiple cycle waveforms in waveform memory and read them
For example, Japanese Patent Laid-Open Publication No.
And store the entire waveform from the start to the end
One time reading method or Japanese Patent Application Laid-Open No. 58-14239.
As shown in No. 6, the multi-period waveform of the attack part and its duration
One or more periodic waveforms of the part are stored, and the waveform of the attack part is stored.
A method of repeatedly reading the waveform of the sustained part after reading once,
Alternatively, as disclosed in JP-A-60-147793,
Stores and reads multiple discretely sampled waveforms
The waveform to be switched in time and
Various methods such as a method of repeatedly reading out waveforms are known.
And these may be appropriately adopted. Furthermore, such multiple
By reading the waveform memory that stores the periodic waveform,
As an example of a method for generating a digital sound signal of high quality
Is the one shown in Japanese Patent Application No. 61-86835.
Where the waveform data stored in the waveform memory is L
Data compression using a method such as PC (linear prediction)
It is disclosed to save storage space. The digital filter 23 is provided with the digital
Input the tone signal and the filter coefficient obtained by interpolation,
According to the filter characteristics set by this filter coefficient
Thus, the input tone signal is controlled. This digital fill
The format of the data 23 may be any. Digital
The filter operation format is basically a finite impulse
Response (FIR) filter and infinite impulse response (II
R) filter, but if an FIR filter is used,
Such as ease of design, stability, and tone control.
It is convenient for a reason. In that case, set the filter coefficient to
Japanese Patent Application No. 60-2675 discloses a type that switches between the two.
No. 42, use the one as shown there
May be. The digital signal output from the tone signal generation circuit 22
The digital tone signal is converted by the digital / analog converter 24.
It is converted to a analog signal and given to the sound system 25.
It is. FIG. 2 shows the configuration as a 32nd-order FIR filter.
FIG. 4 is a basic circuit configuration diagram of a digital filter 23, in which x
(n) is an arbitrary n-th sample point in the input tone signal
Is digital tone waveform sample value data. (Equation 1) Is the unit time delay element, the time of one sampling cycle
This is to set a delay. Therefore, x (n-1) becomes n-1
Digital tone waveform sample value data at the th sample point
X (n-31) is the n-31st sample point
This is digital tone waveform sample value data. h (0) ~
h (31) is a 32nd-order filter coefficient. This filter
The triangular block with the input coefficients is a multiplying element,
Data x (n) to x of each sample point delayed by the delay element
Filter coefficients h (0) corresponding to (n-31) respectively
To h (31). The + sign to which the multiplication output is input
The blocks with additions are addition elements, and each multiplication output is added together.
To obtain an output signal y (n). One feature of such an FIR filter
Means that the phase characteristics can be linear.
You. Assuming a linear phase, there is a gap between the input and output waveforms of the filter.
Phase is completely linear, and the output waveform is distorted.
Does not occur. Therefore, signals of musical sounds, voices, audios, etc.
It is suitable for the filter processing of. Such linear phase characteristics
The necessary and sufficient conditions for an FIR filter with
Response is symmetric and impulse response is
Having the nomenclature means that the filter coefficients h (0) to h (3
1) has symmetry. That is, fill
By setting the data coefficient with symmetrical characteristics,
Phase characteristics can be realized. In this case,
The order at the nominal position has the same filter coefficient.
To prepare filter coefficients of all orders N (= 32).
There is no need, half of that is fine. For details, see this example
If N is an even number, 16 flags from the 0th to the 15th order
Filter coefficients need only be prepared.
The filter coefficients have a symmetrical positional relationship from the 0th order to the 15th order.
A filter coefficient may be used. Therefore, this embodiment
Then, actually 16 filters from the 0th to the 15th order
A coefficient is generated, and this is directly used as a 0th to 15th order coefficient.
Used as filter coefficients and folded back to be symmetric
The filter coefficients from the 16th order to the 31st order in the positional relationship
Even try to use. Other features of the FIR filter include:
Stability is good because there is no feedback loop
There are features. That is, like a IIR filter,
If there is a feedback loop, problems such as oscillation will occur.
Problems such as oscillation do not occur in the IR filter, and the design is easy.
is there. Also, as in the present invention, the filter characteristics are temporally changed.
The FIR filter is advantageous even when changing
You. In this case, in the past, it was
You have to prepare a set of filter coefficients for each
However, if you do so, the time variation of the filter characteristics
Requires a large set of filter coefficients to
As already mentioned. To solve this problem, the present invention
Are two sets of (two-frame)
Filter coefficients are prepared and complemented between the two sets of filter coefficients.
By doing so, you can filter over time
Generate a set of coefficients densely and thus generate by interpolation
Filter characteristics that fluctuate over time due to the filtered filter coefficients
Is set. In this way, interpolation of filter coefficients
When realizing time-varying filter characteristics while performing in real time
If the filter is stable like an FIR filter,
There is no need to devise filter coefficients in consideration of qualitative
It is very advantageous. FIG. 3 is a diagram of the filter parameter memory 20.
Shows an example of the format
Re 20a, voice bank 20b, parameter bank 20
c. The actual filter coefficients are
And stored in the voice bank 20b.
Is the timbre determinant consisting of timbre type, pitch and key touch
The filter coefficient determined corresponding to the combination of
Address data and a file to be read from the bank 20c.
Parameter data for the Luther coefficient interpolation calculation
I have. The voice directory 20a has a plurality of types of tones.
(This is called voice 0 to N)
In this case, the packet address VOA is stored. Tone selection switch
19 in accordance with the tone code representing the tone selected in step 19.
The voice offset address VOA corresponding to the tone is
It is read from the chair directory 20a. Voice off
The set address VOA is stored in the voice bank 20b.
Start address of the storage bank corresponding to each voice 0 to N
Is the data indicating The voice bank 20b stores each voice
From the storage banks (voice banks) corresponding to
You. The voice bank corresponding to each voice is, for example, a voice
0 as shown in the figure.
And a plurality of key groups (these are key groups 0 to M
And a key bank corresponding to the same. The key bank offset table corresponds to the keyboard 1
5 corresponds to the key bank offset address KO
A is stored. Pressing according to the key code of the pressing key
Key bank offset corresponding to the key group to which the key belongs
Is set in the key bank offset table.
Is read from. Key bank offset address KO
A corresponds to each key group 0-M in the key bank
Data indicating the start address of the storage bank to be executed.
It consists of relative address data in the voice bank. Ki
-Bank is a storage bank corresponding to each of key groups 0 to M.
(Key bank). Corresponding to each key group
The key bank is shown for key group 0, for example.
The touch bank offset table and multiple touch
Group (hereinafter referred to as touch groups 0 to L).
And a corresponding touch bank. The touch bank offset table has a
Touch detection circuit 16
Touch bank offset address corresponding to each stage of data
Store the TOA. Detected by the touch detection circuit 16
Touch data to which the touch data belongs
Touch bank offset address corresponding to the touch group
The TOA reads from the touch bank offset table.
Will be issued. The touch bank offset address TOA is
Corresponding to each touch group 0 to L in the touch bank
This is the data indicating the start address of the storage bank.
It consists of relative address data in the key bank. touch
The bank is a storage bank corresponding to each of the touch groups 0 to L.
(Touch bank). For each touch group
The touch bank to perform is, for example, for touch group 0.
As shown, as shown in FIG.
To K). Frame vans corresponding to frames 0 to K
As shown for frame 0, for example,
The parameter data FT for interpolation calculation for the frame,
Stores IN, IT, II and coefficient address data CAD
doing. Also, the coefficient difference for the mode “1” described above.
The value data DCF (0) to DCF (15) are further
It is stored in the frame bank. For details,
Different memory contents for mode “0” and mode “1”
Parameter memory 20 consisting of
It is. Then, the filter parameter menu for mode "0"
In the memory 20, one frame bank has eight addresses.
Parameter data for interpolation calculation F
T, IN, IT, II and coefficient address data CAD
I remember. Also, the filter parameter for mode "1"
In the data memory 20, one frame bank is 4
It consists of 0 address, and there is a parameter for the above interpolation calculation
Data FT, IN, IT, II and coefficient address data C
AD and coefficient difference value data DCF (0) to DCF (15)
I remember. Such a filter for mode "0"
Parameter memory 20 and filter parameters for mode "1"
Data memory 20 with factory set
Attached to an electronic musical instrument, and depending on the memory mode
The filter mode switch 17 may be switched.
Or the filter parameter menu for mode "0".
Memory 20 and filter parameter memory for mode "1"
20 is selected and selected by the filter mode switch 17.
Depending on the selected mode, one of the memories 20 is read.
It may be protruded. Each parameter data for interpolation calculation
Next, a description will be given. The frame time data FT is stored in the frame
The duration of the last frame
“0”, appropriate duration other than “0” for other frames
This is data indicating time. The final frame is the pronunciation of the musical tone
Until the end of the frame.
The time is unspecified, so FT = "0"
You. The interpolation shift number data IN is obtained by
Data indicating the shift amount when shifting the filter coefficient difference
It is. This interpolation shift number data IN is calculated by the interpolation
This is the logarithm of the count data IT. Interpolation count data IT is
Interpolation times when interpolating filter coefficients of adjacent frames
This is data indicating the number (the number of interpolation steps). Interpolation interval
Data II is data indicating the time of one interpolation step.
You. The coefficient address data CAD corresponds to the frame
Filter coefficients to be read from the parameter bank 20c
Address data. Parameter bank 20
c is a set of plural types of filter coefficients (this is
(Sets 0 to R). One set of filter staff
The numbers are, for example, as shown for parameter set 0
And 16 filter coefficients CF (0) from the 0th to the 15th order
~ CF (15). Read from frame bank
The parameter address data CAD
Of the parameter sets 0 to R stored in the link 20c
One of the sets is specified and the specified parameter
Set of filter coefficients CF (0) to C corresponding to the data set
F (15) is read. Each frame 0 to K in one touch bank
The relative address of the frame bank corresponding to
It is specified according to the number FN. Each frame 0-K is hour
At the start of musical sound
Yes, frames 1, 2, 3
The frames are switched in this order. The frame number FN is
This is data indicating the current frame. As mentioned above
The number of addresses in one frame bank is an example.
In the case of mode "0", it is 8 and in the case of mode "1"
40. Therefore, in detail, for each frame 0 to K
The relative address of the corresponding frame bank is the mode “0”.
In the case of, it is “FN × 8”, and in the case of mode “1”,
“FN × 40”. A place in the filter parameter memory 20
The absolute address of the desired frame bank (this is the frame address
Dress FAD) is voiced as described above.
Directory 20a, key bank offset table and
Buttons read from the switch bank offset table
Chair offset address VOA, key bank offset
Address KOA and touch bank offset address T
It is determined by the following arithmetic expression using OA. In the case of mode "0": FAD = VOA + KOA + TOA + FN.times.8 In the case of mode "1": FAD = VOA + KOA + TOA + FN.times.40 That is, the end of the voice bank to be read out
For the voice offset address VOA indicating the head address,
Key bank off, which is a relative address to
The set address KOA is added, and the relative
Touch bank offset address TOA
And add a frame that is a relative address to it.
Bank relative address “FN × 8” or “FN × 40”
Is added. Note that by design,
For each offset address VOA, KOA, TOA
The data expression is VOA 1/16, K
Assume that OA is 1/8 and TOA is 1/8.
You. In that case, lower the value of each offset address data.
Determine the frame address FAD by modifying as described above.
Shall be. In the case of mode “0”: FAD = VOA × 16 + KOA × 8 + TOA × 8 + FN
× 8 mode “1”: FAD = VOA × 16 + KOA × 8 + TOA × 8 + FN
× 40 Hierarchical filter as shown in FIG.
The structure of the parameter memory 20 saves memory capacity
This is advantageous because Without doing this, the tone
All combinations of types, key groups, and touch groups
If the filter coefficients are stored separately for
Although a certain amount of storage capacity is required, using FIG.
Not much storage capacity is required. In other words, the tone type,
The combination of key group and touch group is different
Even if the filter coefficients are common,
Therefore, the parameters stored in the parameter bank 20c of FIG.
The number of data sets 0 to R depends on the timbre type, key group and
Much more than all numbers of touch group combinations
This saves memory capacity. Next, there are two types of filter coefficient interpolation modes.
An outline will be described. Mode “0”: In this mode, adjacent time frames
Interpolation operation based on the difference between two filter coefficients corresponding to
Do. Referring to FIG. 4, in adjacent time frames 0 and 1,
Think about it. The filter coefficient corresponding to frame 0 is F
C0 and the filter coefficient corresponding to frame 1 is FC1
And Find the difference (FC1-FC0) between the two filter coefficients
This is shifted according to the interpolation shift number data IN.
You. In other words, in terms of the arithmetic expression, the interpolation count data IT
It is equivalent to dividing. This gives the coefficient
A difference value DCF is obtained. (FC1-FC0) / IT = DCF 1 set by interpolation interval data II
This coefficient difference value DCF is framed for each interpolation step time.
Is accumulated for the filter coefficient FC0 corresponding to
Thus, an interpolation operation is performed. Thus, the interpolation performance
The filter coefficient obtained by the calculation starts from FC0
Gradually increases by the coefficient difference value DCF at each interpolation step
(Or decrease), and finally the interpolation count data IT
When the same number of accumulations are completed, it corresponds to the next frame 1.
The value reaches the value corresponding to the filter coefficient FC1. And
The frame switches and the next frame 1 and 2
The same interpolation operation as described above is performed. Note that such interpolation
The calculation is performed for each order. Mode "1": In this mode, adjacent modes
At least two filter coefficients corresponding to the time frame
Interpolation calculation based on one and the coefficient difference value prepared in advance
I do. That is, in the mode “0”, the coefficient difference value
DCF is obtained by calculation. In this mode “1”,
Coefficient difference value data DCF (0) to DCF corresponding to each order
(15) is the frame bar in the filter parameter memory 20.
This coefficient difference value DCF is stored in advance in the link.
There is no need to perform the required calculations. Therefore, time frame 0
If we consider the interpolation in
The coefficient difference value DCF read from the interpolation interval data II
0 every 1 interpolation step time set by
Of the filter coefficient FC0 corresponding to
Performs the interpolation operation. And the interpolation count data
Time frames 0 and 1 when the same number of accumulations as
Is completed. And the frame switches
And the same interpolation calculation as described above for the next frames 1 and 2.
Is performed. Executed by microcomputer unit 10
Of the processing related to the present invention among the processing performed
One example of a chart is shown in FIGS. In this process
Related data and working RAM 13
5 is shown in FIG. TIMER
Is a timer count value, and at a predetermined timing (example
Timer counter 14 incremented (for example, every 2 ms) (FIG. 1)
Is stored. TCODE is
This is the tone code, which can be selected with the tone selection switch 19 (FIG. 1)
This shows the selected tone. KCODE is the key code
Indicates a key pressed on the keyboard 15 (FIG. 1).
You. TDATA is the touch data, and the touch detection
2 shows a key touch detected on the road 16 (FIG. 1). MODE is mode data.
Mode selected by the mode switch 17 (FIG. 1).
Is shown. FT, IN, IT, II are the above interpolations
The operation parameter and CAD are the above-mentioned coefficient addresses,
Read from the filter parameter memory 20 as described above.
The one corresponding to the current frame is stored in the RAM 13.
Stored in FN is the frame number described above.
Indicates the current frame. RTIME is
Time at the beginning of one interpolation step.
This is to store the ima count value TIMER. C
IT is the current number of interpolations, and the current number of interpolations (interpolation step
Number of taps). ACF (0) to ACF (15) are 0
Next to fifteenth current filter coefficient data CF (0) to CF (1
5) and changes with the progress of interpolation. As mentioned above,
This is also used as the 16th to 31st order filter coefficient data.
It is. That is, the 0th to 15th order filters in FIG.
The current filter coefficient data as numbers h (0) to h (15)
ACF (0) to ACF (15) are supplied, and the 16th to 31st
This current filter is also used as the filter coefficients h (16) to h (31).
Coefficient data ACF (15) to ACF (0) are symmetrically folded
Supplied. NCF (0) to NCF (15) represent the next frame
0th to 15th order filter coefficient data CF (0) to
CF (15). DCF (0) to DCF (15) are the aforementioned
Numerical difference value data corresponding to each of the 0th to 15th orders
It is. As described above, when the mode is “0”, the calculation is performed.
When the mode is “1”, the filter parameter
Data from the data memory 20 corresponding to the current frame.
It is. Record data or signals as described above.
The storage area is the data and working RAM 13
It is provided within. Data and working RA
The key data (key code and key
Area for storing ON signals, etc.) and the operation panel
18 operation detection data such as switches and LED
Area for storing on / off data, other work
A king area is provided. FIG. 6 shows the main routine.
In the “processing”, the on / off of each key of the keyboard 15 is detected,
If a key press is detected, the "New key on process" in FIG.
Execute the "control". Note that for simplicity of explanation,
In the example, the tone corresponding to the new pressed key is
It shall be pronounced. "Tone switch scanning"
Scans the color selection switch 19 and selects the tone color of the selected tone.
While storing the code as the above tone code TCODE.
The signal is sent to the tone signal generation circuit 22. "Mode switch running
In the inspection process, the filter mode switch 17 is scanned and the
Check which mode “0” or “1” is selected
You. If mode "0" is selected, the mode data
When MODE is set to “0” and mode “1” is selected
For example, the mode data MODE is set to "1". Mainle
Other switches on the operation panel 18
And controls are scanned to perform necessary processing. It is detected that a new key is pressed on the keyboard 15.
When it is issued, the new key on process of FIG. 7 is executed.
You. Here, first, the key code of the newly pressed key
Is stored as the key code KCODE. Next,
Detected by the touch detection circuit 16 corresponding to the new pressing key
Touch data is recorded as the above touch data TDATA.
Remember Next, the contents of various registers are initialized.
For example, a timer count is used as the start time RTIME.
Set the default value TIMER (open the first interpolation step)
Start time is set to RTIME), frame number F
N is reset to "0".
Processing such as resetting to “0” is performed. Next steps
In 30, the tone code TCODE and the key code KCODE
E, Filter parameter according to touch data TDATA
Offset addresses VOA, KO from the data memory 20.
A, TOA are read, and these offset addresses V
Previous according to OA, KOA, TOA and frame number FN
The frame address is calculated as described above. Initially a frame
Since the number FN = "0", the frame of frame 0
The address FAD (0) is obtained. In the next step 31, the calculated frame
Filter parameter memory 2 according to the program address FAD
Various interpolation calculation parameters described above from frame bank 0
Read FT, IN, IT, II and coefficient address CAD
Then, these are stored in a register in the RAM 13. Next
In step 32, according to the coefficient address CAD,
Parameter bank 20 of filter parameter memory 20
read a set of filter coefficients from c and replace them with the current
RA is used as filter coefficient data ACF (0) to ACF (15).
Store in the register in M13. In the next step 33
Are the current filter coefficient data ACF (0) to ACF (15)
To the digital filter 23 in the tone signal generation circuit 22.
Send out. In step 34, the key code KCODE,
Generates tone signal for touch data TDATA and key-on signal
It is sent to the raw circuit 22. Processing of steps 33 and 34
Is performed before the processing of the next steps 35 to 38,
Since the processing of steps 35 to 38 takes time for calculation,
Before sending data to the tone signal generation circuit 22,
In order not to delay the onset of musical tone
You. In the next step 35, the mode data MO
Check whether DE is "0". YES, that is, mode
If "0", the process proceeds to steps 36, 37 and 38,
Find the coefficient difference value data DCF (0) to DCF (15) for the interval
Performs an arithmetic operation. In step 36, the next frame of frame 0
Calculate frame address FAD (1) for frame 1
And a filter according to the frame address FAD (1).
Frame 1 from the frame bank of the parameter memory 20
Read the coefficient address CAD related to. Step 37
Now, add the coefficient for frame 1 found in the previous step.
Of the filter parameter memory 20 in accordance with the
Read a set of filter coefficients from parameter bank 20c
These are used as filter coefficient data NCF for the next frame.
(0) to NCF (15) are stored in the registers in the RAM 13.
A. In step 38, for each order, the next frame
The filter coefficient data NCF (0) to NCF (15) and the current filter
The difference between the filter coefficient data ACF (0) to ACF (15) (for example,
In the case of the zero order, “NCF (0) −ACF (0)” is obtained.
Is shifted according to the interpolation shift number data IN.
The coefficient difference value data DCF (0) to DCF (15) are obtained from
The obtained coefficient difference value data DCF (0) to DCF (15) are
Store in the register in M13. If the mode is "1", from step 35
Proceeding to step 39, as described above,
Coefficient difference stored in the frame bank of the data memory 20
Value data DCF (0) to DCF (15) are converted to frame address F
Read from the frame bank corresponding to AD (0) and read
The coefficient difference value data DCF (0) to DCF (15)
Store in the register in M13. New Kio mentioned above
After the timer process, the timer count value TIMER counts
Each time the timer is interrupted (every 2 ms), the timer interrupt of FIG.
A routine is executed, and the interpolation calculation is performed by this routine.
Will be In FIG. 8, in step 40, the frame
Is the frame time data FT equal to "0", that is, the current frame
Is the last frame. In the last frame
No return time, so return immediately if the last frame
Go to step 41 if it is not the last frame.
In step 41, the timer count value TIMER and the
The difference from the auto time RTIME is based on the interpolation interval data II.
Is the set interpolation step time?
Find out. That is, whether one interpolation step time has elapsed
Find out what. If not, go to return, but
If so, proceed to step 42. In step 42,
Timer count value TI as start time RTIME
Set MER. Thus, the next one interpolation step
Is set to RTIME. Next steps
At 43, the current number of times of interpolation CIT is increased by one. Next
At step 44, the current number of interpolations CIT increased by 1
Whether the number of times matches the interpolation count data IT, that is, the interpolation is completed
Investigate. Step 4 if interpolation is not completed
5 and for each order, the current filter coefficient data AC
The coefficient difference value data DCF (0) to D (F) are added to F (0) to ACF (15).
Add CF (15) and add the result to the new current filter
Perform a process to convert the number data ACF (0) to ACF (15) (example
For example, in the case of the 0th order, “ACF (0) ← ACF (0) + DCF
(0) "). Thus, the calculation of the next one interpolation step is performed.
It is. In step 46, the new interpolation output, the current
Filter coefficient data ACF (0) to ACF (15)
The signal is sent to the digital filter 23 in the generation circuit 22. The routine of steps 41 to 46 is performed according to a predetermined
When the number of times of interpolation IT is repeated, step 44 becomes YES.
No, go to step 47. In step 47,
Increment the Mnumber FN by 1 and increase the current number of interpolations CIT
Reset to "0". The frame switches in this way
You. In the next step 48, the mode data MODE is
Check whether it is "0". YES, that is, in mode “0”
If there is, the process proceeds to step 49 and NO, that is, mode “1”
If so, proceed to step 50. In step 49,
Described offset addresses VOA, KOA, TOA and new
New frame address according to the new frame number FN
Calculate FAD. For example, frame number FN = "1"
If the frame address FAD (1) of the frame 1 is
Desired. In this case, as described above,
Since one frame bank consists of eight addresses,
If 8 is added to the frame address FAD of the frame, a new
Frame address FAD can be obtained. Soshi
And fills in accordance with the calculated frame address FAD.
From the frame bank of the parameter memory 20.
Seed interpolation calculation parameters FT, IN, IT, II and coefficient
Reads the dress CAD and stores them in a register in the RAM 13.
Store in the Star. In step 50, as in step 49,
The offset addresses VOA, KOA, TOA
And a new frame number according to the new frame number FN.
The dress FAD is calculated, and the calculated frame address F
Frame of filter parameter memory 20 according to AD
From the bank, the various interpolation calculation parameters FT, IN,
Read IT, II and coefficient address CAD, and read them
The data is stored in a register in the RAM 13. However,
In frame mode "1", one frame bank has 40 addresses
Address, the frame address F of the previous frame
By adding 40 to AD, a new frame address FAD is obtained.
You can ask. If the mode is "0", step 4
After 9 the process proceeds to step 51. Steps 51 and 52
7, steps 32 and 33, the coefficient address CAD
Of the filter parameter memory 20 according to the
Read a set of filter coefficients from the
With current filter coefficient data ACF (0) to ACF (15)
And store it in a register in the RAM 13
Tones of existing filter coefficient data ACF (0) to ACF (15)
Send to digital filter 23 in signal generation circuit 22
You. Then, in step 53, the frame time data FT is
"0", that is, the new frame is the last frame
Find out if No interpolation in the last frame
If it is the last frame, go to the return, but the last frame
If not, proceed to step 54. Steps 54, 55 and 56 correspond to the steps in FIG.
The processing is almost the same as steps 36, 37, and 38.
For obtaining coefficient difference value data DCF (0) to DCF (15)
Perform the calculation. In step 54, the next of the new frame FN
Address FAD for frame FN + 1
(FN + 1) is calculated, and this frame address FAD (F
N + 1) in the filter parameter memory 20.
Address from frame bank to frame FN + 1
Read CAD. In step 55, the value obtained in the previous step
Address CAD for the frame FN + 1
Parameter bank of the filter parameter memory 20
20c, a set of filter coefficients is read out,
Frame filter coefficient data NCF (0) to NCF (15)
Is stored in a register in the RAM 13. Steps
At 56, the filter coefficient data of the next frame is set for each order.
NCF (0) to NCF (15) and current filter coefficient data
The difference between ACF (0) to ACF (15) (for example, “N
CF (0) -ACF (0) "), and calculates the number of interpolation shifts.
By shifting according to the data IN, the coefficient difference value
Data DCF (0) to DCF (15), and the obtained coefficient difference value
The data DCF (0) to DCF (15) are stored in the
Store in In the case of the mode "1", the scan after step 50
Proceed to step 57. In step 57, the frame time data
FT is "0", that is, the new frame is the last
Check if it is a frame. If not the last frame
Go to step 58, and almost as in step 39 of FIG.
Already stored in the frame bank of the filter parameter memory 20
Stored coefficient difference value data DCF (0) to DCF (15)
To the frame corresponding to the new frame address FAD (FN).
From the bank and read out the coefficient difference value data
DCF (0) to DCF (15) are stored in the registers in the RAM 13.
To torre. In the next steps 59 and 60, the aforementioned steps
In almost the same way as steps 45 and 46, the current fill
New coefficient difference between the coefficient data ACF (0) to ACF (15).
Add the value data DCF (0) to DCF (15) and add
To the new current filter coefficient data ACF (0) to AC
F (15) is processed and the new interpolation output
The filter coefficient data ACF (0) to ACF (15)
To the digital filter 23 in the signal generation circuit 22
You. In this case, even if the frame is switched, the new coefficient
Reading of filter coefficient corresponding to address CAD and it
Current filter coefficient data ACF (0) to ACF (1
5) is not updated, and the old current filter coefficient data AC
Coefficient difference corresponding to a new frame from F (0) to ACF (15)
Accumulate the value data DCF (0) to DCF (15).
ing. In the case of the last frame of mode "1",
If YES in step 57, the process goes to step 61. Step 6
In steps 1 and 62, the same processing as steps 51 and 52 is performed.
U. In this case, the coefficient address C corresponding to the last frame
The parameter of the filter parameter memory 20 according to AD
A set of filter coefficients is read out from the bank 20c.
These are converted to the current filter coefficient data ACF (0) to ACF (1
5) Store in the register in RAM 13 as
Of the current filter coefficient data ACF (0) to ACF (15)
Is transmitted to the digital filter 23 in the tone signal generation circuit 22.
Put out. It should be noted that instead of the processing of steps 61 and 62,
The same processing as steps 59 and 60 may be performed.
No. Note that the mode “1” has a larger coefficient than the mode “0”.
An operation for obtaining the difference value data DCF (0) to DCF (15)
Advantages of quick data access because no calculation is required
There is. There is no remaining memory for storing coefficient difference value data.
Minute, but the coefficient difference value data itself is small.
Since the value is a value, the number of bits is small and the capacity is small. Ma
In addition, the coefficient difference value data for mode “1” is
In the embodiment, each is stored for each frame bank.
As with the filter coefficients, each frame bank has an address.
The coefficient difference value data itself is stored in the parameter bank.
20c may be stored. In the above embodiment, the interpolation operation is performed by software.
Performed by processing, but dedicated hardware circuit
Therefore, it may be executed. In the above embodiment,
Switching between frames is based on the output of the timer counter
Is performed and is unrelated to the pitch of the musical tone, but
Time frame in relation to the phase address signal of the power tone
Switching (and interpolation timing)
May be. In the above embodiment, the switching order of each frame is
The order is predetermined, but it switches randomly
It may be. Example of the method of interpolation calculation of filter coefficients
Are not limited to those shown in FIG. Also
Not only linear interpolation but also any interpolation characteristics may be used. example
For example, interpolation through a digital low-pass filter
You may. Alternatively, for example, a file corresponding to frame 0
The initial value of the filter coefficient is stored.
The difference between the frame and the filter coefficient is stored, and the difference
Data may be shifted to obtain a coefficient difference value.
The present invention can be applied not only to single music instruments but also to double music instruments.
Of course. Also, the tone signal processing device of the present invention
On the other hand, the keyboard and sound source circuit may be separate units
No. Next, another embodiment of the present invention will be described with reference to FIGS.
An embodiment will be described. FIG. 9 shows an electronic device to which the present invention is applied.
FIG. 4 shows a hardware configuration of an embodiment of a musical instrument.
Of the electronic musical instrument, the CPU 111 and the program memo
Including the memory 112 and the data and working RAM 113
Readout of filter coefficients by microcomputer
Various operations and processes including filter coefficient interpolation calculation are controlled.
Is controlled. The program memory 112 is composed of RAM,
By rewriting the control program stored there,
Program specifications can be changed easily
Has become. The filter coefficient memory 114 stores a desired filter.
Store multiple sets of filter coefficients to realize filter characteristics
Each set contains at least two axes of coordinate data.
Each is assigned as dress data. Learn more
To explain an example, for example, the filter coefficient memory 114
Has multiple tables, with multiple tables at each table.
Several sets of filter coefficients are stored in each table.
X and Y coordinate data for each filter coefficient set
Are assigned as address data. 1
Format of filter coefficients in two tables
FIG. 10 schematically shows an example of
In this case, the coordinate position on the X axis is n + 1 points of 0, 1, 2,.
The coordinate position of the axis is the n + 1 point of 0, 1, 2,.
"(N + 1) specified by the intersection of each coordinate position
Of (n + 1) squared pairs of addresses
Filter coefficients F00 to Fnn are assigned. One set of
The filter coefficient realizes one desired filter characteristic
And is composed of multiple order filter coefficients. The filter coefficient memory 114 is a RAM.
And rewrite the contents of the filter coefficients stored there.
It is possible to obtain. External magnetic disk drive
Unit 115 is attached to the electronic musical instrument.
Flexible magnetic disk storing several sets of filter coefficients
Disk cartridge (floppy disk) FD1
Set it on the screen drive unit 115 and store it there.
The set of filter coefficients transferred to the filter coefficient memory 114 is recorded.
You can remember. In addition, the control professional
Floppy disk FD2 (that is,
Stem disk) to the disk drive unit 115
Set and memorize the control program stored there
Can be transferred and stored in the memory 112
I have. The keyboard 116 specifies a pitch of a musical tone to be generated.
With multiple keys for The touch detection circuit 117
It is also used to detect a touch of a key pressed on the keyboard 116.
It is. The type of touch to be detected is the initial touch
Alternatively, it may be either aftertouch. This touch
The key touch data detected by the detection circuit 117 includes a tone color,
It is used to control musical tone elements such as volume and pitch.
When using key touch data as tone control information,
Control the filter coefficients to change the characteristics of the digital filter.
Used to modify. The modulation wheel 118 has a timbre control.
This is a manual operator for generating control information.
The control information is used for controlling the filter coefficients. Low circumference
The wave oscillator 119 converts the low-frequency signal into a modulated signal of a filter coefficient.
It occurs as a signal. Key scaling table
Key 120 is a key scale corresponding to the pitch of the musical tone to be generated.
Ring data is generated by this key scale
Data is used as timbre control information to control the filter coefficients.
Used for Filter envelope generator 121
Are special features such as attack, decay, sustain, release, etc.
Generates envelope waveform data consisting of characteristics in response to key presses
This envelope waveform data is
It is used for controlling filter coefficients as control information.
The operation panel unit 122 includes an operation panel 1 related to a filter.
23 and other various tone setting / control switches and operations
A child 124 is provided. Filter-related operation panel 1
23, the display 230 and the cursor
Switches 231 and 232, filter edit switch 2
33, including a numeric key switch 234, a filter
Used to perform operations related to coefficient selection and setting.
You. The interrupt timer 125 stores various setting data.
Basic timing of data acquisition and filter coefficient interpolation calculation
Is set every predetermined time (for example, 1
The microcomputer sends interrupt signals every 0 ms.
Give to. The tone signal generation circuit 127 includes data and add
Information on the pressed key (key
Pitch corresponding to the pressed key based on the chord and key-on signal)
Generates a digital tone signal having
The sound signal is input to the digital filter 128 and
It performs tone control according to the filter characteristics. Sound signal emission
As for the raw method, as described above,
A sound signal generation method may be used. Note that the tone signal generation method
When the memory read method is used as the expression,
External sound can be freely sampled through the
External sound waveform data stored in the shape memory and sampled
Is read from the waveform memory,
Signal. The digital filter 128 is the digital filter of FIG.
Like the digital filter, the digital tone signal
Enter the filter coefficient and set by this filter coefficient.
The input tone signal is controlled according to the filter characteristics.
The format of the digital filter 128 is also as described above.
Anything may be used, such as the 32nd order shown in FIG.
Using an FIR filter makes it easy to design, stable, and easy to use.
It is convenient because it is suitable for sound control.
Therefore, the following description will be made assuming that this is used. What
In the above, the orders at symmetric positions are the filter coefficients
Are the same value, so that the fill of all orders N (= 32)
Without preparing the coefficient, 16
The filter coefficients are prepared and the 16th to 31st
Filter coefficients are symmetrical from 0th order to 15th order
Although the explanation has been made assuming that the filter coefficient is used, the following
In the embodiment, for convenience, all the filters from the 0th order to the 31st order are included.
The description will be made assuming that the filter coefficient is prepared. In addition, FIR
Adding to the benefits of the filter,
When the filter coefficients are generated densely by interpolation,
The filter characteristics corresponding to the interpolated filter coefficients
This is advantageous because the property can be obtained as it is. Tone signal generation
The digital tone signal output from the circuit 127 is digitized.
Analog / digital converter 130
And given to the sound system 131. Next, the filter-related operation panel 123 is used.
An outline of an example of selecting and setting the filter coefficients
explain. First, set the filter edit switch 233
When turned on, the display 230 is turned on as shown in FIG.
Displays the filter edit menu with appropriate contents. Use
The user can select a desired number among the edit modes 1-4.
Select with the numeric key switch 234. 1st parameter
When set mode (Parameter) is selected,
The screen of the display 230 changes as shown in FIG.
You. In this mode, set the digital filter characteristics.
From the filter coefficient memory 114
The conditions for reading and the filter coefficients are variably controlled.
Various conditions for setting. On display 230
Column setting items “tbl”, “Dyn-axis”, “d
yn ”and“ fix ”with the cursor switch.
231 and 232 to select and
The corresponding desired selection / setting data is stored in numeric key switch 2.
Selection and setting are made by 34. Display 230
"Tbl", "Dyn-axis",
Under “dyn” and “fix”, corresponding to the item,
The selected and set data is displayed. For each setting item
And will be described next. Tbl: stored in the filter coefficient memory 114
Select the filter coefficient table to be used. Selected here
The data consists of filter coefficient table data FTABLE and
To register in the data and working RAM 113
It is registered (see FIG. 13). Dyn-axis: X in the filter coefficient table
-Coordinates of any of the Y coordinate axes (see FIG. 10)
Select whether data is variably controlled according to tone control information
I do. The coordinate axes that are variably controlled according to the tone control information
It will be called the dynamic axis. For example, dyna
When setting as a mic axis, use the numeric key switch 23
Enter “0” in 4 and set the Y axis as the dynamic axis
Input "1" using the numeric key switch 234.
You. The data set here is the dynamic axis data D
Data and working RAM 113 as YNXS
(See FIG. 13). dyn: Specifies the coordinate data of the dynamic axis. This
The data specified here is the dynamic axis reference coordinate data.
The data and the data in the working RAM 113 as DYN
Registered in the register (see FIG. 13). fix: coordinate axis that is not a dynamic axis (fixed axis and
) Is specified. Specify here
The data obtained is data and data as fixed axis coordinate data FIX.
Registered in a register in the working RAM 113
(See FIG. 13). The dynamic axis coordinate data and the fixed axis
Specifying the coordinate data is performed using the numeric key switch 234 as desired.
This is performed by inputting coordinate values. Using FIG.
As described above, each text in the filter coefficient memory 114 is
X and Y axis coordinates are stored in each filter coefficient set stored in the table.
The mark data is assigned respectively, and the dyn and fix
Dynamic axis coordinate data and fixed axis specified in the item
One set of filter coefficients is specified by the combination of coordinate data
Is determined. Note that each filter coefficient group is assigned
If the coordinate data value is an integer,
Value of coordinate data of dynamic axis and coordinate data of fixed axis
Is not limited to an integer value and may include a decimal value. Specify
If the coordinate data contains decimal values,
The corresponding filter coefficients are stored in the filter coefficient memory 114.
Since it is not performed, find the filter coefficient by interpolation
Like that. If you add an explanation about this point,
If the specified coordinate data is an integer part like (1,1)
And the coordinates as shown in FIG.
A set of filter coefficients F11 corresponding to (1,1) is filtered.
By reading from the filter coefficient memory 114,
Get the specified filter coefficient without interpolation
be able to. However, if the specified coordinate data is (1,
As shown in FIG. 12, when a minority part is included as in 1.5),
A set of filter coefficients F corresponding to the coordinates (1, 1)
11 and another set of filters corresponding to coordinates (1, 2)
The number F12 is read from the filter coefficient memory 114,
Decimal value of each coordinate between both filter coefficient sets (0, 0.5)
By performing interpolation at a ratio according to the
Set of filters not stored in the coefficient memory 114
Create the coefficients. Returning to FIG. 11A, the second envelope
When the set mode (Envelope) is selected,
The screen of the spray 230 changes as shown in FIG.
In this mode, the envelope for variable filter coefficient control
Various conditions for forming a waveform are set. Display
The setting items “R1”, “R2”, “R
3 "," R4 "," L1 "," L2 "," L3 "," L
4 "is a desired item.
2 is selected and the desired item corresponding to the selected item is selected.
Set the setting data using the numeric key switch 234.
You. Each setting item “R1” on the display 230,
"R2", "R3", "R4", "L1", "L2",
The lower part of “L3” and “L4” are set corresponding to the item.
The displayed data is displayed. The following describes each setting item.
I will tell. Each element of a commonly known envelope waveform
Related to attack, decay, sustain, release
"R1" sets the attack rate, "R2"
Is for setting the decay rate, "R3" is for sustain
"R4" sets the release rate
What to set, "L1" sets the attack level
"L2" sets the decay level, and "L2"
“3” sets the sustain level, “L4” sets the level.
This sets the lease level. Each set here
The data of items R1 to L4 are the filter envelope settings
Data FENV (0) to FENV (7)
Registered in the registers in the working RAM 113
(See FIG. 13). The filter envelope generator 1 of FIG.
21, the filter envelope setting data FEN
Based on V (0) to FENV (7), key
Generates envelope waveform data. Returning to FIG. 11A, the third low frequency modulation signal
When the number set mode (Lfo) is selected, the display
The screen 230 changes as shown in FIG. This mode
The low frequency modulation signal generated from the low frequency oscillator 119
Set various conditions. The upper setting of the display 230
Fixed items "wave", "speed", "dela"
y ”and“ depth ”with the cursor switch.
Items selected by operating the switches 231 and 232
The desired selection / setting data corresponding to
234 is selected and set. On display 230
Items "wave", "speed", "del"
ay ”and“ depth ”are shown below
The set data is displayed. The following describes each item.
I will tell. wave: low frequency change generated from the low frequency oscillator 119
Select the waveform of the tuning signal. For example, sine, triangle, saw
Select desired waveform from tooth, inverted sawtooth or square wave
I do. The data selected here is the low-frequency waveform data WA
Data and register in working RAM 113 as VE
(See FIG. 13). speed: low frequency generated from the low frequency oscillator 119
Set the speed (frequency) of the modulation signal. Set here
The data obtained is used as low-frequency speed data SPEED.
Register in data and register in working RAM 113
(See FIG. 13). delay: The low-frequency oscillator 119 is activated from the start of key press.
Set the delay time before the shake starts. Set here
The obtained data is low-frequency oscillation delay time data DELAY
To register in the data and working RAM 113
It is registered (see FIG. 13). depth: low frequency generated from the low frequency oscillator 119
Set modulation signal amplitude (ie modulation depth or depth)
I do. The data set here is the low-frequency amplitude data DE
The data in the working RAM 113 as the PTH
Registered in the register (see FIG. 13). The low-frequency oscillator 119 shown in FIG.
Based on each setting data WAVE-DEPTH for the tone signal
Generate a low frequency modulation signal. Returning to FIG.
Select key scaling set mode (Scale)
Then, the screen of the display 230 changes as shown in FIG.
It changes. In this mode, the key scaling table
120 among several key scaling curves
Choose the one you want. The data selected here is Keith
Calling curve select data KYSCV
Data and registers in the working RAM 113.
(See FIG. 13). This key scaling curve select
Key scaling curve corresponding to data KYSCV is
Selectively read in key scaling table 120
Pressing in this key scaling curve
Key scaling corresponding to the pitch (or range) of the keypress
Data is read from the table 120. Executed by the microcomputer unit
Of processing related to the present invention,
14 to 17 show an example of this. In this process
Data and working RAM 113 used in connection with
FIG. 13 shows an example of the contents stored in the table. KCOD
E is a key code, which is pressed on the keyboard 116 (FIG. 9).
It shows the key that was lost. TDATA is touch data
And the key detected by the touch detection circuit 117 (FIG. 9).
It shows a touch. WHEELD is a modular
Modulation wheel data
Data corresponding to the operation amount in the file 118 (FIG. 9).
You. Each data FTABLE to KYSCV is as described above.
belongs to. EGDATA is the envelope waveform data
And the filter envelope generator 121 (FIG. 9)
Envelope waveform data for filter coefficient control generated from
It shows the current value of the data. LFODAT is low-frequency modulated signal data.
Yes, low frequency generated from low frequency oscillator 119 (FIG. 9)
This shows the current value of the modulation signal. KYSDAT is
Key scaling data.
From the bull 120 (FIG. 9) to the pitch (or range) of the pressed key
The current value of the corresponding read key scaling data is
It is shown. DYNDAT is the dynamic axis coordinate
Data, indicating the current coordinate value of the dynamic axis
It is. XAXIS is X-axis coordinate data, and X-axis
Indicates the current coordinate value of the. YAXIS is the Y axis
Coordinate data, indicating the current coordinate value of the Y axis
You. One of the X or Y axis corresponding to the dynamic axis
Standard data XAXIS or YAXIS is a dynamic axis
It has the same contents as the coordinate data DYNDAT, and
The corresponding coordinate data XAXIS or YAXIS is fixed
The contents are the same as those of the axis coordinate data FIX. X-axis coordinate data XAXIS and Y-axis coordinate data
In combination with YAXIS, filter coefficient memory 1
The filter coefficient to be read out from 4 is specified. On the spot
In this case, the coordinate data XAXIS or YAXIS of one axis is
As described above, this is the same as the fixed axis coordinate data FIX.
This uses the filter-related operation panel 123 as described above.
Of the parameter set mode (see FIG. 11B)
This is specified by the processing of Meanwhile, Dyna
Of the other axis corresponding to the mic axis coordinate data DYNDAT.
The coordinate data XAXIS or YAXIS is as follows:
It is changed and controlled by various tone color control information.
The finger is arbitrarily operated by operating the filter-related operation panel 123.
Touch the specified dynamic axis reference coordinate data DYN
Data TDATA, modulation wheel data W
HEELD, envelope waveform data EGDATA, low
Frequency modulation signal data LFODAT and key scaling
Performs an operation of modulating with the data KYSDAT.
The calculation result is defined as dynamic axis coordinate data DYNDAT.
You. Each data TDATA, WHEELD, EGDAT
A, LFODAT and KYSDAT include decimal parts
The die obtained by the modulation operation
Numeric axis coordinate data DYNDAT also contains decimal part
It may be. As mentioned earlier, filter coefficient memory
In step 114, the coordinate value directly including the decimal part
Filter coefficients have not been assigned,
By performing interpolation calculation, it is possible to handle coordinate values including decimals.
Calculate the filter coefficient. In addition, dynamic shaft reference seat
The target data DYN is converted to each data TD as tone color control information.
ATA, WHEELD, EGDATA, LFODAT and
For modulating or changing with KYSDAT
Is any operation such as addition, subtraction, multiplication, division, etc.
Is also good. For example, addition and subtraction are used. COEFA and COEFB are interpolation calculation-related components.
Number register, two files to be interpolated
Data coefficients are stored in each register COEFA, COEFB
I do. The interpolation operation consists of two sets of filters to be interpolated.
It is done between coefficients of the same order in the number. Register C
For OEFA and COEFB, perform the interpolation operation at that time.
Is stored. COEFD
(0) to COEFD (31) are filter coefficient registers.
Reading from the filter coefficient memory 114 and interpolation
For each order finally obtained by arithmetic or the like (from 0th order to 3rd
This stores the (primary) filter coefficient data. This
Filter coefficient registers COEFD (0) to COEFD (3)
The filter coefficient data for each order stored in 1) is
Digital filter 128, and adjusts the characteristics of the filter.
decide. FIG. 14 shows the main routine.
Filter-related operation panel scan processing ”
Scans on / off detection of each switch of the operation panel 123
Further, according to this scanning result, referring to FIG.
Related to the filter coefficient selection and setting
Data FTABL selected and set
E, DYNXS, DYS, FIX, FENV (0) to F
ENV (7), WAVE, SPEED, DELAY, DE
PTH, KYSCV data and working RAM 11
3 is registered in the register. In the next process, the data and
The low frequency modulation signal setting registered in the working RAM 113
Fixed data WAVE, SPEED, DELAY, DEP
TH to the low frequency oscillator 119 (abbreviated as LFO in the figure)
Send out and filter envelope setting data FENV
(0) to FENV (7) are converted to the filter envelope generator 12
1 (abbreviated as filter EG in the figure). In the next process, the keyboard 116 and other operations are performed.
Performs on / off detection scan processing of the cropper 124, and the result
The data necessary for forming the obtained tone is converted to a tone signal generating circuit 127.
(Abbreviated as TG in the figure). Each key of the keyboard 116
A new key is pressed during the on / off detection scanning process
When a new key event is detected as shown in FIG.
Is executed. For simplicity of explanation,
In the embodiment shown in FIG.
Assuming that musical tones are to be pronounced in the
It is assumed that the system is assembled. In the main routine
In the interrupt signal from the interrupt timer 125
Each time is given, the "timer in" shown in FIG.
A "rupture routine" is executed. The "New Key On Event" shown in FIG.
First, the key code of the new press key is
Register as KCODE (step 140). Next,
Is detected by the touch detection circuit 117 corresponding to a new key depression of
The registered key touch data is registered as TDATA (step
141). In the next step 142, the envelope
Set initial value "0" as waveform data EGDATA
And the filter envelope generator 121
"1" indicating the key press is transmitted as the control signal KON. What
The initial value of the envelope waveform data EGDATA is
The value is not limited to “0” and may be another value. Filter Envelope
In the loop generator 121, as the key-on signal KON
When “1” is given, generation of envelope waveform data
To start. In the next step 143, the low frequency modulation signal
When the initial value "0" is set as the data LFODAT
In both cases, a start command is given to the low frequency oscillator 119. Low
The frequency oscillator 119 oscillates in response to the start command.
To start. However, it is set by the above-mentioned data DELAY.
The actual oscillation operation starts after the elapse of the set delay time.
In the next step 144, a key scaling curve selection is performed.
Key scaling curve according to the target data KYSCV
Is selected in the key scaling table 120, and this selection
Code KC in the key scaling curve obtained
Key scaling data corresponding to the pitch of the ODE
-Read from the scaling table 120. reading
Key scaling data is registered as KYSDAT.
You. In the next step 145, the modulation
Captures the operation output of the
It is registered as application wheel data WHEELD. next
In step 146, the dynamic axis reference coordinate data D
Touch data TDATA, modulation for YN
Wheel data and key scaling
Calculates the data KYSDAT and obtains the coordinate data DYN.
Control and change the calculation result to dynamic axis coordinate data D
Register as YNDAT. At this time, the envelope
Waveform data EGDATA and low frequency modulation signal data LF
The reason that ODAT is not added to the calculation is that the initial values of both data are
This is because "0" was set. For an initial value other than "0"
May add this data to the operation. Next, "Phil
Data coefficient calculation subroutine ". An example of “filter coefficient calculation subroutine”
Is shown in FIG. Here, first, dynamic
Check whether the axis data DYNAXS is "1"
(Step 147). If “1”, dynamic
The axis is the Y axis, go to step 148,
The dynamic axis coordinate data DYNDAT obtained in step 46 is
Register as Y-axis coordinate data YAXIS
When the “Filter-related operation panel scan process”
The fixed axis coordinate data FIX that has been set is converted to X-axis coordinate data XAX.
Register as IS. Dynamic if not "1"
The axis is the X axis, go to step 149,
The dynamic axis coordinate data DYNDAT obtained in step 46 is
Register as X-axis coordinate data XAXIS and fix
Axis coordinate data FIX as Y axis coordinate data YAXIS
register. In the next step 150, the order register C
Set the content to "0". Contents of this order register C
Is the order of the filter coefficient to be obtained by the current operation
Instruct. In the next step 151, the filter coefficient
Of the filter coefficient table in the memory 114
Set for "Filter-related operation panel scan processing"
The specified filter coefficient table data FTABLE
Select the indicated filter coefficient table and select this selected
-Axis coordinate data XAXIS and Y in the table
Indicated by each integer part of axis coordinate data YAXIS
A set of filter coefficients corresponding to the specified X and Y coordinate values
Then, the order register C in this set of filter coefficients
Therefore, the filter coefficient of the designated order is read. like this
The filter coefficients of a certain order read out as
FCOEF @ FTABLE, C, (XAXIS) I, (Y
AXIS) I}. Thus, a certain degree of
The filter coefficient is stored in one of the interpolation calculation coefficient registers COE.
Store in FA. In the next step 152, step 147
Similarly, the dynamic axis data DYNXS is "1".
Is checked, and if it is “1”, step 153 is executed.
If it is “0”, go to step 154. Step
In step 153, the file specified by the data
In the filter coefficient table, the X-axis coordinate data XAXI
Integer part of X (XAXIS) I and Y-axis coordinate data YAXIS
Of the Y-axis coordinate value (YA
XIS) I + 1 to identify a set of filter coefficients,
The order register C for this set of filter coefficients
The filter coefficient of the specified order is read. In this way
The filter coefficient of a certain order to be read is represented by FC in the figure.
OEF @ FTABLE, C, (XAXIS) I, (YAX
IS) I + 1}. Thus, a certain degree of
The filter coefficient is stored in the other interpolation calculation coefficient register COE.
Store in FB. In this case, the Y axis is the dynamic axis.
Therefore, interpolation along the Y axis is performed. On the other hand, in step 154, data FTA
In the filter coefficient table specified by BLE
The integer part (XAXIS) I of the X-axis coordinate data XAXIS
X-axis coordinate value (XAXIS) I + 1 greater than 1 and Y-axis coordinate
One set with the integer part (YAXIS) I of the data YAXIS
Of the filter coefficients, and the set of filter coefficients
Filter of order indicated by order register C
Read the coefficient. Thus, a certain order
The Ruta coefficient is shown in the figure as FCOEF ， FTABLE, C,
(XAXIS) I + 1, (YAXIS) I}. Like this
The filter coefficient of a certain order read out by
It is stored in the arithmetic coefficient register COEFB. In this case, X
Interpolates along the X axis because the axis is a dynamic axis
It is. In the next step 155, the X-axis coordinate data X
AXIS and Y-axis coordinate data YAXIS
The decimal part of the coordinate value corresponding to the axis
And the file stored in the registers COEFA and COEFB.
Performs an operation to interpolate between data coefficients. By this interpolation calculation
The obtained filter coefficient is designated by the order register C.
Filter coefficient register COEFD (C) corresponding to order
(One of COEFD (0) to COEFD (31))
I do. The function of this interpolation calculation is linear interpolation.
Or other appropriate curves (for example, a quadratic curve
Or any interpolation curve previously stored in the interpolation function memory)
There may be. Next, the content of the order register C is incremented by one.
(Step 156) If the value is greater than the maximum degree 31,
It is checked whether there is (step 157). Still maximum order 31
If not, the process returns to step 151, and
Is performed for the next filter order. In this way
By repeating the processing of steps 151 to 57, all orders
Is completed, the step 157 returns to YE
It becomes S, and it goes to step 158. In step 158
Are the filter coefficient registers COEFD (0) to COEFD
The filter coefficient data of all orders stored in (31) is
It is sent to the digital filter 128. Returning to FIG.
After completing the filter coefficient calculation subroutine, step 15 is executed.
Go to 9 and press key code KCODE and key
An ON signal KON is sent to the tone signal generation circuit 127. "Timer Interrupt Routine" in FIG.
First, at step 160, the current
Check if it is sounding. If it is sounding, step 161
Go to the return, if not sounding. Steps
At 161, the current envelope is generated by the filter envelope generator 121.
Imports the generated envelope waveform data and performs EGD
Register as ATA. In the next step 162, the low
Low-frequency modulation signal data currently generated by the wave oscillator 119.
Data and register it as LFODAT. Next step
In step 163, similar to step 145 described above, the module
Operation output of the simulation wheel 118, and
Registered as wheel data WHEELD
I do. In the next step 164, the dynamic axis reference
Touch data TDATA and mode are
DURATION wheel data WHEELD, Envelo
Waveform data EGDATA, low frequency modulation signal data L
FODAT and key scaling data KYSDAT
The coordinate data DYN is variably controlled by calculation, and the calculation result is obtained.
Result is registered as dynamic axis coordinate data DYNDAT
I do. Next, in FIG.
Chin ", which allows filtering by interpolation
Find the coefficient data and digitize the found filter coefficient data.
To the tall filter 128. File as in Fig. 12
Filter table, assuming that
An example of coefficient selection and interpolation calculation will be described. In addition,
In FIG. 12, in order to facilitate understanding, each filter coefficient
The coordinate positions of sets F00 to F22 are set by the filter coefficient set.
The diagram of the filter characteristic realized by the above is arranged in a schematic diagram. This
, The horizontal axis f is frequency, and the vertical axis L is
Level. Dynamic axis data DYNXS
If it is “1” (the Y axis is a dynamic axis),
Dynamic axis reference coordinate data DYN is "0", fixed shaft seat
It is assumed that the target data FIX is set to “2”. This
, The fixed axis coordinate data FIX is used as the X axis coordinate data.
Is specified as “2”, and
"0" corresponding to the dynamic axis reference coordinate data DYN is
Initially specified. By this, X, Y seat
Filter coefficient set F assigned to the target position (2,0)
20 is selected. Touch data TDAT as tone control information
A, modulation wheel data WHEELD, d
Envelope waveform data EGDATA, low frequency modulation signal data
Data LFODAT and key scaling data KYSD
AT changes dynamic axis reference coordinate data DYN.
Adjustments, and the dyna
Mic axis coordinate data DYNDAT is, for example, "0.3"
If there is, the Y-axis coordinate data YAXIS is "0.3"
Is changed to However, the Y-axis coordinate data YAXIS
A few parts (YAXIS) I are not changed to "0",
The filter coefficient set selected from the number table is the X and Y coordinates
F20 assigned to position (2,0), initial
It is not different from what you set. However, the Y-axis coordinate data Y
Fill in the AXIS decimal part "0.3" as a parameter
An interpolation operation is performed between the data coefficient sets F20 and F21. That is, the position is divided into the X and Y coordinate positions (2, 0).
The filter coefficient of the assigned filter coefficient set F20 is
Stored in register COEFA, Y-axis coordinate data YAXI
Y-axis coordinate value that is 1 greater than the integer part (YAXIS) I of S
(YAXIS) I + 1 = “1” and X-axis coordinate data XAX
X, Y coordinate position (2, 2) specified by IS “2”
The filter coefficient set F21 assigned to 1) is filed.
Select from the filter coefficient table and register the filter coefficient
Data in the registers COEFB and COEFA.
Y-axis coordinate data between filter coefficients stored in FB
Interpolation using YAXIS decimal part "0.3" as a parameter
Perform the operation. Is created by interpolation
The filter characteristics realized by the filter coefficients
Corresponding to the filter coefficient sets F20 and F21
The two filter characteristics in the Y-axis coordinate data YAXIS.
It is a characteristic that is interpolated and synthesized at a ratio corresponding to the value of several
is there. For example, if the filter coefficient set F20 is a low-pass filter
F21 realizes high-pass filter characteristics
Each data that is timbre control information.
TDATA, WHEELD, EGDATA, LFODA
Depending on the magnitude of the combined value of T and KYSDAT,
Because of the characteristics closer to the low-pass filter characteristics,
Filter characteristics up to characteristics closer to pass filter characteristics
It can be variably controlled. As described above, the X, Y coordinate position (2, 0)
The assigned filter coefficient set F20 is initially specified.
Data TDAT, which is timbre control information,
A, WHEELD, EGDATA, LFODAT, KY
Dynamic axis reference coordinate data based on SDAT composite value
Performs an operation to modulate the data DYN and obtains the operation result.
The dynamic axis coordinate data DYNDAT obtained is, for example,
If it changes to the integer part like “1.2”, the Y axis
The coordinate data YAXIS has been changed to "1.2"
Accordingly, the integer part of the Y-axis coordinate data YAXIS (YAXI
S) I changes to "1" and the filter is selected from the filter coefficient table.
The selected filter coefficient set is located at the X, Y coordinate position (2, 1).
It changes to the assigned F21. This filter coefficient set
The filter coefficient of F21 is read from the filter coefficient table.
And stored in the register COEFA. Also this
, The integer part of the Y-axis coordinate data YAXIS (YAXI
S) Y-axis coordinate value (YAXIS) I + 1 greater than I by 1 =
Characteristic is "2" and X-axis coordinate data XAXIS "2".
Assigned to the specified X, Y coordinate position (2, 2)
The filter coefficient set F22 is read from the filter coefficient table.
And the filter coefficient is stored in the register COEFB.
Is done. Thus, the filter coefficient set F21 and
Minor part of Y-axis coordinate data YAXIS between F22
An interpolation operation is performed using “0.2” as a parameter. This
Thus, by the filter coefficient created by the interpolation operation,
The realized filter characteristics depend on the filter that was interpolated.
Two filter characteristics corresponding to the coefficient sets F21 and F22
According to the value of the decimal part "0.2" of the Y-axis coordinate data YAXIS.
This is a characteristic obtained by interpolation synthesis at the same ratio. For example, the filter
The coefficient set F21 realizes a high-pass filter characteristic,
Is a high-pass filter whose cutoff frequency is higher than the above.
If the sound quality control information is realized,
Each data TDATA, WHEELD, EGDATA, L
Depending on the magnitude of the combined value of FODAT and KYSDAT,
High-pass with relatively low cutoff frequency corresponding to F21
Relatively high cutoff corresponding to F22 from the filter characteristics
Filter characteristics up to high-pass filter characteristics of frequency
In addition, steps 152, 153 and FIG.
In 154, the coordinate value of the axis corresponding to the dynamic axis is set to 1
Increase, the coordinate value of the axis corresponding to the fixed axis does not increase by 1,
The filter coefficient of the corresponding X, Y coordinate position is stored in the register CO.
It is stored in EFB, but not limited to this.
The coordinate values of both the X and Y axes are each increased by 1, and the corresponding X and Y axes are increased.
Store the filter coefficient of the coordinate position in the register COEFB
You may make it. That is, FCOEF @ FTAB
LE, C, (XAXIS) I + 1, (YAXIS) I + 1}
The filter coefficient of a certain order represented by
You may make it memorize | store in the register COEFB. That
In the case of "Filter related operation panel scan processing"
Data including decimal part as fixed axis coordinate data FIX
Is set, the filter along the fixed axis corresponds to the decimal part.
The filter coefficients can be interpolated. In the above embodiment, control is performed according to the tone color control information.
Coordinate axis (dynamic axis) is only one axis,
This may be two axes of X and Y. Then, both axes
When the coordinate data of the
Filter coefficients to be selected on table coordinates
The direction of movement is not a direction along only one axis.
Arbitrarily, such as the direction of drawing
it can. In the above embodiment, the dynamic axis is a key touch.
Which sound, such as pitch, key scaling, envelope waveform, etc.
Although common to control elements, the type of each tone control element
Dynamic axis can be selected independently for each class
You may do it. In that case, for each type of tone control element
Separately for each of the X and Y axes selected as dynamic axes
The synthesis operation of the corresponding tone color control information and the
It is assumed that a coordinate data change operation is performed. Do so
Thus, more complicated filter control becomes possible. The above embodiment
Now, the filter coefficient table is arbitrarily selected by the performer.
The appropriate file is selected according to the tone selection operation.
The filter coefficient table may be automatically selected.
In addition, dynamic axis selection and dynamic axis coordinate data
Performer settings and fixed axis coordinate data
The X, Y coordinate data is specified by the player
Arbitrarily, but this is also a tone
Appropriate data is automatically selected according to the selection operation etc.
It may be. In the embodiment shown in FIG.
Is executed by software processing, which is also
Even if it is executed by a dedicated hardware circuit,
Good. In the above-described embodiment, the low-frequency oscillator 119 and
The filter envelope generator 121 has dedicated hardware.
It is composed of a
It may be realized by processing. The above embodiment
Then key scaling data generation is key scaling
By reading from the table 120
However, the present invention is not limited to this.
Key scaling corresponding to the pressed key
Data may be generated. Also, in the embodiment of FIG.
It is needless to say that the present invention can be applied not only to compound music instruments. So
, The filter envelope generator 121
Generates envelope waveform data for each sound generation channel
So that Also, in the embodiment of FIG.
Keyboard 116 and touch panel for the unit with
Switch detection circuit 117 may be a separate unit,
In that case, data is exchanged according to the MIDI standard
And In addition, the present invention is limited to musical tone signal processing of scale sounds.
To apply to signal processing of rhythm sounds and other sounds
Can be. In the above embodiment, the filter coefficient table
Are two axes (two-dimensional coordinates),
It may be three axes (three-dimensional coordinates) or one axis (1
As described above, according to the present invention, the filter
Data coefficients are generated by interpolation.
Timbre control content without increasing the amount of filter coefficients
So that filter coefficients can be generated densely
Timbre control using digital filters
Abundant filter characteristics with relatively simple equipment configuration
Performance and rich tone change control.
The effect is excellent. Also this
According to the invention, the data should be read from the filter coefficient storage means.
Basically, the filter coefficient is roughly determined by the reference value data.
Since it is specified, any setting of this reference value data
In the memory of the filter coefficient storage means with various variations
Can be used, which also provides a rich
To be able to realize control
It has the effect. That is, change the selection of filter coefficients
Can be performed easily in a rich manner.
As a result, more complicated digital filter characteristic control
Is relatively easy to do.
The effect is excellent. Further, according to the present invention, the filter coefficient is
Read the specified information for specifying
Indirect reading to read filter coefficient storage means more
And stored in the filter coefficient storage means.
Indirectly by using different timbre information
Can be read and used in common for
Reduce the number of filter coefficients stored in the
And the filter coefficient storage capacity can be
It has an excellent effect that it can be reduced. Ma
Stored in the filter coefficient designation information storage means.
The designation information is for each tone of a tone signal, and
At least one of the range of the signal and the touch
Is remembered, and in these designated information
Have a common timbre, range or touch
Since the one that specifies the filter coefficient is included,
The number of independent variables for reading out the filter coefficients
Can reduce the filter coefficient storage capacity relatively
That is, for each tone of the tone signal and
Designated for at least one of musical range and touch
Increased versatility of tone control by storing information
(Independent variable for reading out the filter coefficients)
While the filter coefficient storage capacity is
It can be said that it does not increase so much
It has the effect.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a hardware configuration showing an embodiment of an electronic musical instrument to which the present invention is applied. FIG. 2 is a block diagram showing an example of a case where the digital filter of FIG. 1 is configured by a 32nd-order FIR filter. FIG. 3 is a view showing an example of a storage format of a filter parameter memory of FIG. 1; FIG. 4 is a view for explaining an example of interpolation of filter coefficients. FIG. 5 is a view showing an example of data in FIG. 1 and contents stored in a working RAM; FIG. 6 is a flowchart showing an example of processing executed by the microcomputer unit of FIG. 1, and is a diagram showing a main routine. FIG. 7 is a flowchart illustrating an example of a new key-on processing routine. FIG. 8 is a flowchart illustrating an example of a timer interrupt routine. FIG. 9 is a block diagram of a hardware configuration showing another embodiment of the electronic musical instrument to which the present invention is applied. FIG. 10 is a graph showing an example of coordinate data assigned to each set of filter coefficients in one filter coefficient table in the filter coefficient memory of FIG. 9; FIG. 11 is a view showing a display example of a display on the filter-related operation panel in FIG. 9; FIG. 12 is a graph showing an example of filter characteristics realized by each filter coefficient set in one filter coefficient table, corresponding to a coordinate arrangement to which the filter coefficient set is assigned; FIG. 13 is a view showing an example of the data in FIG. 9 and the contents stored in a working RAM; FIG. 14 is a flowchart showing an example of processing executed by the microcomputer unit of FIG. 9 and showing a main routine. FIG. 15 is a flowchart illustrating an example of a new key-on event routine. FIG. 16 is a flowchart illustrating an example of a filter coefficient calculation subroutine. FIG. 17 is a flowchart illustrating an example of a timer interrupt routine. DESCRIPTION OF SYMBOLS 10 Microcomputer unit 14 Timer counter 15 Keyboard 16 Touch detection circuit 17 Filter mode switch 20 Filter parameter memory 22 Music signal generation circuit 23 Digital filter 114 Filter coefficient memory 116 Keyboard 117 Touch detection circuit 118 Modulation wheel 119 Low frequency Oscillator 120 Key scaling table 121 Filter envelope generator 122 Operation panel unit 123 Filter operation panel 230 Display 127 Musical tone signal generation circuit 128 Digital filter

Claims (1)

(57) [Claims] Digital filter means for inputting a tone signal and a filter coefficient and controlling the input tone signal in accordance with characteristics according to the filter coefficient; filter coefficient storage means storing a plurality of sets of filter coefficients for realizing desired filter characteristics; Reference value data setting means for setting reference value data for designating a filter coefficient to be read from the filter coefficient storage means; control data generating means for generating control data for variably controlling a timbre; Calculating means for calculating data and control data to obtain filter coefficient calculation data as a result of the calculation; and a plurality of filters read out from the filter coefficient storage means in accordance with the filter coefficient calculation data and a plurality of the read out filters. The coefficients are interpolated according to the filter coefficient calculation data, and are obtained by interpolation. Tone signal processing device with a filter coefficient interpolation means for providing a filter coefficient to the digital filter means has. 2. Digital filter means for inputting a tone signal and a filter coefficient and controlling the input tone signal in accordance with characteristics according to the filter coefficient; filter coefficient storage means for storing a plurality of sets of filter coefficients; and for each tone color of the tone signal And a filter coefficient designation information storage unit for storing designation information for designating a filter coefficient for at least one of a range and a touch of a musical tone signal, wherein the designation information is common to different timbres, ranges or touches. , And timbre information indicating the timbre of the musical tone to be generated, and at least one of the gamut information indicating the range of the musical tone signal to be generated and the touch information indicating the touch of the musical tone to be generated Generating information generating means, and at least two sets of filter coefficients according to information generated from the information generating means. Reading out the designation information for designating the filter coefficient from the filter coefficient designation information storage means, generating filter coefficient reading means for reading out the filter coefficient from the filter coefficient storage means based on the designation information, and generating a control signal for controlling the timbre. Control signal generating means; and using the control signal generated from the control signal generating means as an interpolation parameter, interpolating the at least two sets of filter coefficients read by the filter coefficient reading means, and obtaining a filter coefficient obtained by interpolation. And a filter coefficient interpolating means for providing the digital sound signal to the digital filter means.