JPH0540480A - Electronic musical instrument - Google Patents

Abstract

PURPOSE:To reproduce a formant sound having a proper formant without taking a time for an input of formant data, etc., by generating the formant data by an inter-polating operation, and reproducing the formant sound. CONSTITUTION:A panel 3 is provided with a liquid crystal display LCD 4 for displaying various information, a tone color selection switch 5 for selecting a tone color, and formant setting/correcting switches 6, 7, 8 and 9 for setting or correcting formant data. In such a state, with respect to musical performance information generated by a fact that a performer executes a musical performance operation, several sample points are read out, and from those sample points, the formant data corresponding to its musical performance information can be obtained by an interpolating operation. A musical tone having a formant based on this formant data is synthesized. In such a way, trouble is saved, and also, a sound of a new tone color can easily be made.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic musical instrument capable of producing formant data and reproducing formant sound by using interpolation calculation.

[0002]

2. Description of the Related Art Generally, it is said that a voice having a fixed formant does not change in formant even if the basic pitch changes. However, in actuality, when the low pitched sound and the high pitched sound are compared, it is observed that the position and shape of the formant are changed.

FIG. 12 is a graph showing changes in formant frequency when a professional soprano singer produces four vowels "a, i, u, e" at various fundamental frequencies (pitch). "A, i, u, e" represent phonetic symbols of vowels. The solid line F1 is these four vowels "a, i, u, e"
5 is a graph showing changes in the first formant frequency when a sound is generated at various fundamental frequencies. For example, for the "a" sound, the fundamental frequency is 200 Hz to 620 Hz.
The first formant frequency is substantially constant in the range up to the level, and the first formant frequency gradually increases at the fundamental frequency of about 620 Hz or higher. Similarly, the broken line F2, the alternate long and short dash line F3 and the alternate long and two short dashes line F4 have four types of vowels "a,
5 is a graph showing changes in the second, third and fourth formant frequencies when "i, u, e" is pronounced at various fundamental frequencies. From this figure, it can be seen that the formant frequency changes as the fundamental frequency changes.

FIG. 13 shows an amplitude spectrum when the "a" sound is produced with a voice volume that is large enough to make a normal conversation. G2 indicates the spectrum envelope of the amplitude spectrum G1. FIG. 14 shows an amplitude spectrum when a strong "a" sound is produced. G4 indicates the spectrum envelope of the amplitude spectrum G3. It can be seen from FIGS. 13 and 14 that the formant frequency and shape are changed depending on the loudness of the voice.

In addition, there is a sound having a moving formant in which the position of the formant changes according to the change of the fundamental frequency.

In order to reproduce such various sounds with an electronic keyboard instrument, for example, formant data corresponding to each key of the keyboard (that is, corresponding to a pitch) is stored in advance in a memory, When operating the keyboard, it is necessary to read the formant data corresponding to the pitch and supply it to the formant synthesis sound source. Also, the strength of the sound to be pronounced is raised or lowered in response to the touch of a key on the keyboard, for example, but in order to produce a formant according to the strength of the sound, all formant data corresponding to each touch value is stored. You need to do it.

[0007]

However, in order to store the formant data corresponding to each key of the keyboard or each touch value, it is necessary to input all the formant data to be stored in advance. The operation is very time-consuming.

In view of the above circumstances, an object of the present invention is to provide an electronic musical instrument capable of reproducing a formant sound having an appropriate formant, without the trouble of inputting formant data.

[0009]

In order to achieve the above object, an electronic musical instrument according to the present invention can be used as performance information generating means for generating performance information according to a performance operation of a performer, and the performance information. Storage means for storing formant data corresponding to some performance information values within a range of values, and a plurality of formant data from the storage means based on the performance information values generated by the performance information generating means. The formant data corresponding to the value of the performance information generated by the performance information generating means is obtained by performing an interpolation operation using the read control means for selectively reading and a plurality of formant data read by the read control means. The interpolating means for calculating and outputting and the musical tone signal of the musical tone having the formant based on the formant data output from the interpolating means are combined. Characterized by comprising the formant sound synthesizing means for and output.

The performance information includes, for example, key code and touch information as pitch information output from the keyboard. The formant data includes, for example, formant frequency, formant level, formant bandwidth, and the like. A plurality of sets of performance information and corresponding formant data are stored in the storage means. The sound input means for inputting an external sound and the sound analysis means for analyzing the sound input by the sound input means to extract performance information and formant data of the sound are further provided. It is advisable to input or input the set data of the performance information and the corresponding formant data and store it. As the interpolation method, in addition to linear interpolation, curve interpolation using a function such as a spline function may be used.

Further, in the electronic musical instrument according to the present invention, performance information generating means for generating performance information in response to a performance operation of a performer, sound input means for inputting external sound, and the sound input means are used. Acoustic analysis means for extracting the performance information of the sound by analyzing the sound, and a set of a plurality of sets of performance information and formant data obtained by analyzing a plurality of sounds by the sound analysis means. Interpolation means for calculating and outputting formant data corresponding to all performance information values other than the performance information included in the set data by performing interpolation calculation using the data, and formant data extracted by the acoustic analysis means. And storage means for storing the formant data calculated by the interpolating means, and the performance information generated by the performance information generating means. It reads the formant data corresponding to the value of the information from the storage means, characterized by comprising the formant sound synthesizing means for outputting a musical tone signal of a tone having formants based on the formant data synthesis to.

[0012]

The formant data corresponding to some performance information values are stored, and these set data serve as sample points. That is, some sample points are read out from the performance information generated by the performer performing a performance operation, and formant data corresponding to the performance information can be obtained from these sample points by interpolation calculation. Then, a musical sound having a formant based on the formant data is synthesized.

[0013]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an electronic musical instrument according to an embodiment of the present invention. In this figure, 1 is a keyboard having a large number of keys operated by a player, 2 is an interface of the keyboard 1, and 3 is a panel having various panel switches. The panel 3 includes a liquid crystal display (LCD) 4 for displaying various information, a tone color selection switch 5 for selecting a tone color, and formant data setting / correction switches 6, 7, 8 for setting or correcting the formant data. , 9 are provided. Tone selection switch 5
Consists of 10 switches for selecting 10 kinds of tones. Formant data setting / correction switch 6
It consists of four switches for setting or modifying the formant data for the first formant. That is, the first formant data setting / correction switch 6
Is a switch 6-1 for increasing the formant level for the first formant, a switch 6-2 for decreasing the formant level, a switch 6-3 for increasing the formant frequency, and a switch 6-4 for decreasing the formant frequency. Consists of. Formant data setting / correcting switches 7, 8 and 9 are setting / correcting switches for the second, third and fourth formant data, respectively. Each of these switches 7, 8 and 9 is also composed of four switches similar to the switch 6.

Reference numeral 10 is an interface of various panel switches of the panel 3, 11 is an interface of the LCD 4 of the panel 3, and 12 is a digital signal processor (DSP). The DSP 12 is for performing at high speed an operation for extracting performance information such as a pitch from an input sound and an operation for synthesizing a formant sound.
13 is a read only memory (ROM) that stores programs or parameters that define various operating procedures of this electronic musical instrument, 14 is a random access memory (RAM) used as a work register, and 15 is a voice or the like. Sound input device for inputting external sound, 1
Reference numeral 6 is an analog / digital converter (A / D converter) for converting analog data output from the sound input device 15 into digital data, 17 is a central processing unit (CPU) for controlling the operation of the entire electronic musical instrument, and 18 is a controller. A formant synthesis sound source for synthesizing a formant sound based on the formed formant data, a sound system 19 for generating a music sound based on a tone signal output from the formant synthesis sound source 18, and a bidirectional bus line 20.

This electronic musical instrument holds formant data of several sample points in advance.
That is, some sets of set data including a key code (MIDI code) indicating a pitch, touch data of the keyboard, and formant data corresponding to these key code and touch data are stored. At the time of actual performance, formant data corresponding to the key code and touch data obtained by the performance operation is interpolated from the formant data at the sample points to generate a formant sound. The formant data of the sample points can be manually set by using the formant data setting / correction switches 6, 7, 8 and 9 on the panel 3 of FIG. 1, or by inputting voice or other sound from the outside. It is also possible to extract acoustic performance information and formant data and use them as sample point data.

Table 1 shows an example of sample point data stored and held in the electronic musical instrument of this embodiment. However,
Here, touch data is omitted, and only the key code and formant data are shown. Further, the sample regarding the tone color "e" is omitted.

[0018]

[Table 1] In the electronic musical instrument of this embodiment, a voice is input in each tone color of "a, i, u, e" to acquire sample point data,
These tones can be used to generate musical tones when the keyboard is played. For example, when the "a" sound is selected as the tone color, when a key on the keyboard is pressed, the "a" sound is reproduced at the pitch corresponding to the key. The formant is proper according to the pitch. In order to generate such a formant sound, data of sample points regarding the first to fourth formants of the "a" sound in Table 1 is stored.
That is, for the "a" sound, the formant data F1a1 and L1a1 corresponding to the note name D5 ^# with the key code "75" is used as the sample point for the first formant.
(First data) and the formant data F1a2, L1a2 corresponding to the note name A5 whose key code is "81"
It stores (second data). Further, as sample points regarding the second formant for the "a" sound, the formant data F2a1, L2a1 corresponding to the note name C5 having the key code "72" and the key code "8".
Formant data F2a2 corresponding to the note name A5 of "1",
It stores L2a2. Thus, "a, i, u,
Data of two sample points are stored for each of the first to fourth formants of each tone color "e".
In addition, among the formant data, FXyy is shown as the formant frequency data of the Xth formant (X = 1 to 4). yy is a lowercase letter a, i, u,
It consists of a combination of e and the numbers 1 and 2, but this letter "a"
The tone, the “i” tone, the “u” tone, and the “e” tone are distinguished, and the numbers indicate the first data or the second data. Similarly, in the formant data, LXyy is the formant level data of the Xth formant (X = 1 to 4). The meaning of yy is the same as the frequency data FXyy.

Next, key codes "75" and "8" are provided as sample points for the first formant of the "a" sound.
The reason why the formant data corresponding to "1" is stored will be described.

In the electronic musical instrument of this embodiment, the formant sound is generated by embodying the change of the formant frequency with respect to the fundamental frequency shown in FIG. Here, the graph of the first formant F1 of the "a" sound in FIG. 12 shows a substantially constant formant frequency in the range of the fundamental frequency 200 Hz to 620 Hz, and linearly increases with a substantially constant slope in the range of 620 Hz to 900 Hz. The formant frequency is shown. Therefore, for example, by storing the formant data at the positions B5 and B6 in FIG. 12, the formant data to be used can be determined based on the fundamental frequency of the sound to be generated as follows.

If the fundamental frequency is in the range of 200 Hz to 620 Hz, the formant data at the position B5 is used.

If the fundamental frequency is in the range of 620 Hz to 900 Hz, the formant data at the position B5 and the position B
The formant data obtained by interpolation from the formant data in 6 is used.

If the basic frequency is lower than 200 Hz, the formant data at the position B5 is used, and if the basic frequency is higher than 900 Hz, the formant data at the position B6 is used (or the straight line from B5 to B6 is extended and interpolated). Good).

As described above, in the electronic musical instrument of this embodiment, the key code "7" near the fundamental frequency 620 Hz at the position B5 is used.
Formant data F1a1, L1a1 corresponding to "5",
Key code "8" near the fundamental frequency of 900 Hz at position B6
The formant data F1a2 and L1a2 corresponding to "1" are stored as the data of the first formant of the "a" sound. The other sample points listed in Table 1 are similarly selected based on FIG. Each graph in FIG. 12 is either a line segment having a predetermined slope or a line segment having a slope “0” and a line segment having a predetermined slope connected to each other. Therefore, if formant data at two predetermined positions in each graph is stored as sample point data, corresponding formant data can be calculated for any value of the fundamental frequency. In Table 1, "a, i, u,
This is the reason why the data of two sample points are stored for each of the first to fourth formants of each timbre of "e". Next, in the electronic musical instrument of this embodiment, a procedure for acquiring formant data of sample points as shown in Table 1 by external acoustic input will be described. FIG. 2 is a block flow diagram showing the flow of processing when sound is input from the outside to extract formant data of sample points. In this figure, in order to extract the formant data, first, in FIG.
A sound such as a voice is input by the sound input device 15 (block 21). Next, the input analog sound data is A / D converted by the A / D converter 16 (block 22). Formant data 28 can be obtained by performing a formant analysis (block 23) using the obtained digital acoustic data. The formant analysis may be performed by using a general method in the field of voice processing, such as a method of obtaining a spectrum by using a linear prediction method and a method of using a complex sine wave model.
In this embodiment, the center frequency (formant frequency) and level (formant level) of the four formants of the input sound are detected.

The key code data 29 is obtained by extracting the pitch (pitch) using the digital audio data obtained by the A / D conversion (block 24) and further converting the key code (block 25). .. Further, the corresponding sound data 30 can be obtained by detecting the sound volume using digital sound data (block 26) and converting it into touch data (block 27). In the sound volume detection, the loudness of the sound is obtained by detecting the maximum amplitude of the input waveform and the sound power. The touch data is converted into MIDI data in the range of "0" to "127" based on the sound volume detection result. The key code conversion and the touch data conversion may be performed by calculation or by a table reference method.

In this manner, one sample point of data, that is, one set of key code, touch data, and formant data corresponding to these is obtained. Further, the above-described processing is performed on the input sounds of several pitches and volumes, and the data of a plurality of sets of sample points is acquired. In this way, the sample point data as shown in Table 1 can be acquired and stored.

Formant analysis (block 2
3), pitch extraction (block 24), key code conversion (block 25), volume detection (block 26), and touch data conversion (block 27) are mainly performed by the CPU 17.
And the DSP 12. Further, the formant data 28, the key code 29, and the touch data 30 obtained by inputting the sound from the outside can be corrected by using the formant data setting / correcting switches 6, 7, 8, 9 of the panel 3. ..

Next, using the formant data of Table 1 thus obtained and stored, the operation for generating a formant sound having an appropriate formant based on the pitch will be described.

FIG. 3 is a flow chart for explaining the main routine of the electronic musical instrument of this embodiment. When the processing of this electronic musical instrument starts, first, in step S1, initial settings such as registers are performed, and in step S2, panel switch processing is performed. In the panel switch processing, the pressing of each switch on the panel 3 in FIG. 1 is detected and the processing corresponding to the pressing is performed. After step S2, it is determined in step S3 whether there is a key depression event of the keyboard. If there is a key depression event, the key depression process is performed in step S4, and if there is no key depression event, the process proceeds to step S5. Step S
In step 5, it is determined whether or not there is a key release event on the keyboard. If there is a key release event, the key release process is performed in step S6, and if there is no key release event, the process proceeds to step S7.
Proceed to. After performing other processing in step S7, the process returns to step S2 again. The loop processing from step S2 above is repeated.

FIG. 4 is a flow chart for explaining the key pressing process of step S4 of FIG. In the key depression process, first, formant data interpolation calculation is executed in step S11. By this formant data interpolation calculation,
From this, formant data of the sound to be pronounced is obtained. Next, in step S12, a channel to be sounded is selected, and in step S13 the formant data obtained by the interpolation calculation is sent to the formant synthesis sound source. And
In step S14, the key-on signal is sent to the formant synthesis sound source to generate a formant sound.

FIG. 5 is a flow chart for explaining the formant data interpolation calculation processing of step S11 of FIG. In the formant data interpolation calculation, first, in step S21, it is determined what the timbre is. If it is the "a" tone, the process branches to step S22, and if it is another tone color, the process branches to each tone color.

In step S22, it is determined whether the key code KC of the depressed key is "75" or less. If the key code KC is "75" or less, it means that the pitch of the note to be produced is D5 ^# note or less. Therefore, as described in Table 1, the formant data F1a1 and L1a1 at the position B5 in FIG. 12 may be used. Therefore, in step S23, the register F1a
The first data F1a1 of the first formant frequency data of the "a" sound is substituted into, and the first data L1a1 of the first formant level data of the "a" sound is substituted into the register L1a. Then, it progresses to step S28.

If the key code KC of the key pressed in step S22 is not "75" or less, the process proceeds to step S24, and it is determined whether the key code KC is between "75" and "81". If the key code KC is not between "75" and "81", it means that the key code KC is "81" or more. Therefore, the formant data to be used are the formant data F1a2 and L1a2 at the position B6 in FIG. Therefore, in step S25, the first "a" sound is input to the register F1a.
The second data F1a2 of the formant frequency data is substituted, and the second data L1a2 of the "a" tone first formant level data is substituted into the register L1a. Then, it progresses to step S28.

In step S24, the key code KC is "7".
If it is between "5" and "81", it means that the fundamental frequency of the sound to be pronounced is between the positions B5 and B6 of FIG. 12, and therefore the "a" sound is selected based on the key code KC in step S26. Interpolation calculation is performed between the first data F1a1 and the second data F1a2 of 1 formant frequency data,
The result is stored in the register F1a. Similarly, in step S27, an interpolation operation is performed between the first data L1a1 and the second data L1a2 of the "a" sound first formant level data based on the key code KC, and the result is stored in the register L1a. .. Next, in step S28, the second to fourth formant data are similarly obtained, and the registers F2a,
The data is stored in F3a, F4a, L2a, L3a, and L4a, respectively, and the process returns.

Although the processing when the tone color is other than the "a" tone is omitted in step S21, the "i" other than the "a" tone is omitted.
The formant data is similarly calculated for the sound, the "u" sound, and the "e" sound. That is, by comparing the key code of the sound to be pronounced with the key code of the sample point in Table 1, it is determined which range is included, and predetermined formant data is set according to the determination result or formant is calculated by interpolation calculation. Calculate the data.

The formant data thus obtained is sent to the formant synthesis sound source in step S13 of FIG. 4 described above. As a result, a formant sound having an appropriate formant according to the pitch is generated.

Next, a second embodiment using another formant data calculation method will be described. The second embodiment is an example in which the data of the sample points shown in Table 1 in the above embodiment and the interpolation calculation processing shown in FIG. 5 are replaced with different data and processing. Since the other configurations are the same, only the data of the sample points and the interpolation processing will be described below.

FIG. 6 shows the contents of data of sample points in the second embodiment. In this example, all four formant data regarding the fundamental frequency at the position where the way of changing the formant is changed based on the graph of FIG. 12 (that is, the position where the slope of the graph is changed) are stored in the RAM 14. There is. For example, "a"
Focusing on the formant frequency data of the sound, the first formant changes the inclination at the position B5, the second formant changes the inclination at the position B4, the third formant changes the inclination at the position B2, and the fourth formant changes the inclination at the position B3. Is changing. The basic frequencies at these positions and the lowest frequency (fundamental frequency of the position B1) and highest frequency (fundamental frequency of the position B6) of the measurement range of this measurement data are arranged from the lowest frequency as follows. However, K
Cn (n = 1 to 5) represents the key code (closest one) of the frequency.

KC1: fundamental frequency of position B1 KC2: fundamental frequency of positions B2 and B3 KC3: fundamental frequency of position B4 KC4: fundamental frequency of position B5 KC5: fundamental frequency of position B6 The sample point data shown in FIG. These are first to fourth formant data in these key codes KCn. Further, in this embodiment, since the interpolation related to the touch is also performed, the first value for each of the plurality of touch values is
~ The fourth formant data is stored as sample point data. For example, the key code K for the "a" sound
For C1, formant data is stored for each of the two types of touch values T1 and T2. Touch value T
1 is a relatively small value, and the touch value T2 is a relatively large value. From the above, for example, the following 16 formant data are stored for the key code KC1 of the sound "a".

<Formant data with touch T1> 1st formant frequency data: F1KC1T1 2nd formant frequency data: F2KC1T1 3rd formant frequency data: F3KC1T1 4th formant frequency data: F4KC1T1 1st formant level data: 1st L1K1 formant level data Data: L2KC1T1 3rd formant level data: L3KC1T1 4th formant level data: L4KC1T1 <Formant data where touch is T2> 1st formant frequency data: F1KC1T2 2nd formant frequency data: F2KC1T2 3rd formant frequency data F2KC1T2 3rd formant frequency data Data: F4KC1T2 1st formant level data: L1KC1T2 2nd formant level data: L2KC1T2 3rd formant level data: L3KC1T2 Fourth Formant Level Data: L4KC1T2 Further, similar to the above, the formant data of the sample points is stored for each of the other sounds "i, u, e". These data may be input by the procedure of FIG. 2 as in the first embodiment described above. The interpolation of the touch value is performed to reproduce the change in the formant data according to the strength of the sound, but the rate of change of the change in the formant data corresponding to the strength of the sound is not always constant. , The number of sample points regarding the touch value is not limited to two. For example, although the sample points for the two touch values T1 and T2 are stored for the key code KC1 of the “a” sound in FIG. 6, three touch values T11 and T1 are stored for the key code KC6 of the “i” sound.
2, sample points for T13 are stored.
What kind of touch value should be used as the sample point may be determined according to the state of change in the formant data corresponding to the sound intensity.

FIG. 7 shows an interpolation calculation execution process when the formant data of the sample points as shown in FIG. 6 is stored. In this interpolation calculation execution process, first, step S3
In step 1, the key code KCn of the sample point of FIG. 6 stored in the RAM 14 is read to detect which sample the key code KC of the note to be sounded is between. For example, if the tone color is the "a" tone, it is detected which of the above KC1 to KC5 lies. Next, in step S32, the formant data in the touch direction is interpolated, and then in step S33.
Press to interpolate the formant data in the key direction and return.

FIG. 8 is a conceptual diagram for specifically explaining the interpolation procedure of FIG. FIG. 8 shows the interpolation procedure when the key code KC of the depressed key is between the above key codes KC1 and KC2 and the touch value is "50". Key code KC and touch value "5" at this time
First formant frequency data F1KC 5 corresponding to "0"
0 is calculated as follows. First, with the key code KC1 fixed, interpolation regarding touch is performed between F1KC1T1 and F1KC1T2 to obtain F1KC1 50. Also, similarly, interpolation regarding touch is performed between F1KC2T3 and F1KC2T4 to obtain F1KC2 50. The above interpolation of the touch direction is performed in step S32 of FIG. Next, F1KC1 50
Interpolate between F1KC250 and F1KC250,
Calculate F1KC 50. This key direction interpolation is performed in step S33 of FIG. Similarly, F2KC 50, F
3KC 50, F4KC 50, L1KC 50, L2KC 50, L3
Calculate KC 50 and L4 KC 50.

The formant data thus obtained is
In step S13 of FIG. 4, the sound is transmitted to the formant synthesis sound source. As a result, a formant sound having an appropriate formant according to the pitch is generated.

In the first and second embodiments described above, it is not necessary to store the formant data for all performance information, so the memory capacity is saved as compared with the case where all formant data is stored. be able to.

In the first embodiment, the value of the key code of the sample data is incorporated as a fixed value in the processing flow (eg, step S22 or step S2 in FIG. 5).
4). Therefore, it is inconvenient when you want to change the key code of the sample point. According to the second embodiment, the data of each sample point can be arbitrarily input by the user as shown in FIG. 2 and can be corrected by the switches 6, 7, 8 and 9 in FIG. 1, so that more flexible setting is possible. It is easy to create various sounds.

Next, a third embodiment will be described in which the key direction and the touch direction are interpolated in advance and all the results are stored and used during reproduction.

FIG. 9 shows the contents of a table storing the formant data of the electronic musical instrument of this embodiment. This table is set in the RAM 14 and stores the first to fourth formant data corresponding to the respective key codes in correspondence with the touch values "0" to "127".

FIG. 10 is a flow chart for explaining the procedure for setting data in the table of FIG. To set data in the table, first, in step S41, formant data is extracted at a plurality of sample points,
Store in the corresponding position in the table. The extraction of the formant data may be performed in the same manner as the procedure of FIG. 2 of the first embodiment described above. Next, in step S42, 12 is applied in the touch direction.
Interpolation is performed for all keys in the key direction in eight steps and stored in the table of FIG. Such interpolation processing may be performed in the same manner as described in the second embodiment with reference to FIGS. By the above procedure, the formant data corresponding to all the key codes and all the touch values are stored in the table of FIG.

FIG. 11 is a flow chart showing a procedure for reproducing a formant sound by pressing a key on the keyboard. It should be noted that at the time of reproduction, the main routine of FIG. 3 may be used and the key pressing process of step S4 may be replaced with the procedure of FIG.

In the key depression process for reproducing the formant sound,
First, in step S51, formant data is read from the table of FIG. 9 according to the touch value and key code when the key is pressed. Next, in step S52, a tone generation channel is selected, the read formant data and the key-on signal are sent to the formant synthesis sound source 18, and the process returns. As a result, a musical tone having a formant corresponding to the touch value and the key code is generated.

By previously calculating and storing the formant data corresponding to all the key codes and all the touch values in this way, the processing time up to the execution of sound generation can be shortened, and the processing at the time of reproduction is performed at high speed. be able to. In addition, the formant data stored in the table is calculated by extracting the formant data at some sample points and the rest by interpolation calculation. Therefore, the setting of the table is easy and the sound is easily made.

Although the linear interpolation is used as the formant data interpolation method in the above-mentioned three embodiments, the invention is not limited to this, and an interpolation method using a function curve or other interpolation methods may be used. In that case, the number of sample points of the formant data necessary for interpolation may be three or more.

Since the sound range of the voice is narrower than that of the keyboard, it is almost impossible to pronounce the sounds corresponding to the keys with the highest and lowest pitches. Therefore, the formant data of the sound corresponding to the key with the highest pitch and the key with the lowest pitch cannot be extracted from the voice. In the above-described embodiment, as can be seen from steps S22 to S25 in FIG. 5, the sound outside the predetermined range uses the formant data of the sample point at the closest position. However, the present invention is not limited to this, and fictitious formant data may be created for sounds in a range other than the range in which samples can be taken, or interpolation curves (including straight lines) may be extended for calculation.

In the above embodiment, the interpolation between the touch direction and the key direction is shown, but the interpolation in the time direction may be performed. That is, the formant data at different times may be further stored in the memory, and the interpolation in the time direction may be executed in addition to the interpolation in the touch direction and the key direction depending on the time from the start of key depression. Further, the interpolation may be one-dimensional interpolation such as only the touch direction, only the key direction, or only the time direction, or two-dimensional interpolation by appropriately combining these. The number of sample points is not limited to that in the above embodiment. For example, for touch data, three or more sample points may be taken and stored.

The present invention can also be applied to a preset type electronic musical instrument. In this case, the formant data may be stored in the ROM in advance.

In the above embodiment, the formant frequency and the formant level are interpolated, but other formant data such as the formant bandwidth can be interpolated in the same manner. Further, in the above embodiment, the formant data is interpolated between the tones of the same tone color (for example, "a" tone) of different pitches, but the interpolation may be performed between the tones of different tone colors. This makes it possible to easily create a new tone color.

In the above embodiment, the formant frequency and the formant level change at positions of substantially the same fundamental frequency, but the formant frequency and the formant level may be processed independently.

[0058]

As described above, according to the present invention, the formant data is created by the interpolation calculation and the formant sound is reproduced, so that the appropriate formant corresponding to the performance information such as the pitch and the volume is produced. Since it is possible to synthesize a formant sound having "," and it is not necessary to input formant data corresponding to all performance information, it is possible to save time and effort. Also, by interpolating between formants of sounds having completely different tones, it is possible to easily create a new tone color.

[Brief description of drawings]

FIG. 1 is a block diagram of an electronic musical instrument according to an embodiment of the present invention.

FIG. 2 is a block diagram showing the flow of processing when a sound is input from the outside to extract formant data of sample points.

FIG. 3 is a flowchart for explaining a main routine of the electronic musical instrument of this embodiment.

FIG. 4 is a flowchart for explaining the key depression process of step S4 of FIG.

FIG. 5 shows an example of formant data interpolation calculation in step S11 of FIG.

FIG. 6 shows the contents of the formant data memory in another example of the interpolation calculation of the present invention.

FIG. 7 is a flowchart of interpolation calculation execution processing when the formant data as described above is stored.

FIG. 8 is a graph for specifically explaining the procedure of FIG.

FIG. 9 shows the contents of a table storing formant data of the electronic musical instrument of this embodiment.

FIG. 10 is a flowchart for explaining a procedure for setting data in the table of FIG.

FIG. 11 is a flowchart showing a procedure for playing a formant sound by pressing a key on the keyboard.

FIG. 12 is a graph showing changes in formant frequency when a professional soprano singer pronounces four vowels “a, i, u, e” at various fundamental frequencies.

FIG. 13: Amplitude spectrum when the “a” sound is pronounced with a voice volume that is large enough for a normal conversation

FIG. 14: Amplitude spectrum when a strong “a” sound is pronounced

[Explanation of symbols]

1 ... Keyboard, 2 ... Interface, 3 ... Panel, 4 ... Liquid crystal display, 5 ... Tone selection switch, 6, 7, 8,
9 ... Formant data setting / correction switch, 10 ... Panel switch interface, 11 ... LCD interface, 12 ... Digital signal processor (DS
P), 13 ... ROM, 14 ... RAM, 15 ... Acoustic input device, 16 ... A / D converter, 17 ... Central processing unit (C)
PU), 18 ... Formant synthesis sound source, 19 ... Sound system, 20 ... Bus line.

Claims (2)

[Claims] 1. Performance information generating means for generating performance information in response to a performance operation of a performer, and formant data corresponding to some performance information values within a range of possible values of the performance information. Storage means, read control means for selecting and reading a plurality of formant data from the storage means based on the value of the performance information generated by the performance information generating means, and a plurality of formants read by the read control means. Interpolation means for calculating and outputting formant data corresponding to the value of the performance information generated from the performance information generating means by performing interpolation calculation using the data, and formant based on the formant data output from the interpolation means. Formant sound synthesizing means for synthesizing and outputting a tone signal of a musical sound having Musical instruments. 2. A performance information generating means for generating performance information in response to a performance operation of a performer, a sound input means for inputting an external sound, and a sound input by the sound input means to analyze the performance information. Along with the extraction, the acoustic analysis means for extracting the formant data of the sound, and the interpolation calculation using the set data of a plurality of sets of performance information and formant data obtained by analyzing the plurality of sounds by the sound analysis means. Thus, the interpolation means for calculating and outputting the formant data corresponding to all the performance information values other than the performance information included in the set data, the formant data extracted by the acoustic analysis means and the formant calculated by the interpolation means. Storage means for storing data, and formant corresponding to the value of the performance information generated by the performance information generating means Read data from said storage means,
An electronic musical instrument comprising: a formant sound synthesizing means for synthesizing and outputting a musical tone signal of a musical tone having a formant based on the formant data.