JP2002304175A - Waveform-generating method, performance data processing method and waveform-selecting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate music sound having full of expressions. SOLUTION: Partial waveforms and their attribute information are stored in an HDD 23. When music sound waveforms are to be generated, based on sounding control information, attribute information corresponding to the sounding control information is detected by a reproducing circuit 24, and partial waveforms having the attribute information are read from the HDD 23. Then, a comparison is made for the sounding control information and the attribute information and a processing is made for the read partial waveforms according to the difference between the two information to generate waveforms having full of expression. These waveforms are outputted as sounds from a sound system 19 via a mixer 17 and a DAC 18.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a waveform generation method, a performance data processing method, a waveform selection device, and a waveform selection method when performing using recorded musical instrument sounds or singing sounds.

[0002]

2. Description of the Related Art Conventionally, a waveform memory sound source has been known.
The waveform memory sound source reads out the waveform data stored in the waveform memory according to event data such as note-on and note-off to generate a musical tone. At this time, different waveform data is selected from the waveform memory according to the pitch information and the touch information included in the note-on. On the other hand, an electronic musical instrument called a sampler has been conventionally known. In the sampler, the waveform of a played single tone is sampled (for example, the pitch of C3 is set to 5 seconds) and recorded, and a musical tone is generated using the recorded waveform data. In this case, for each sampled waveform data, an original key indicating the pitch of the waveform data is set by the user, and based on the original key, a sound range in which the waveform data is used or a predetermined pitch is generated. The pitch shift amount and the like are determined. With respect to the method of determining the original key, a pitch is extracted from the sampled waveform data, and an original key corresponding to the extracted pitch is automatically set for the waveform data (for example, Japanese Patent Laid-Open No. 325579). Further, the played phrase waveform is sampled, and the sampled phrase waveform data is divided into a plurality of partial waveform data based on an envelope level extracted from the phrase waveform data, and the timing and pitch of each partial waveform data are determined. A phrase sampler that performs while changing has also been proposed.

[0003]

As in a conventional sampler, the sound actually played is sampled and recorded.
Even if a musical tone is generated using the recorded waveform data, there is a problem that only a sound with a poorer expression than an actual musical instrument can be obtained. In this case, even if waveform data is selected based on pitch information or touch information, a real musical tone cannot be obtained.

Accordingly, the present invention provides a waveform generation method, a performance data processing method, and a performance data generation method capable of generating a musical tone rich in expression.
It is an object of the present invention to provide a waveform selection device and a method thereof.

[0005]

In order to achieve the above object, a waveform generating method according to the present invention stores a plurality of partial waveforms in a partial waveform memory, and stores a plurality of partial waveforms stored in the partial waveform memory. Attribute information is stored in an attribute information memory, and a partial waveform having attribute information corresponding to the tone generation control information is read out from the partial waveform memory by referring to the attribute information memory based on the input tone generation control information. By processing the obtained partial waveform based on the attribute information and the sound control information, a waveform corresponding to the sound control information is generated.

Further, according to the performance data processing method of the present invention which can achieve the above object, attribute information of each partial waveform is stored in an attribute information memory, and the characteristic of each note included in the performance data to be processed. And the attribute information stored in the attribute information memory, thereby detecting a partial waveform that is optimal for the feature of the note, and specifying data specifying the detected partial waveform as the note in the performance data. , And the performance data to which the designated data is added is stored. Further, another waveform generating method of the present invention capable of achieving the above object is a waveform generating method for reproducing performance data to which designated data stored by the performance data processing method is added,
A plurality of partial waveforms are stored in a partial waveform memory, and when a note to which designated data is assigned based on the performance data is reproduced, the partial waveform is stored based on the designated data assigned to the note to be reproduced. By reading a partial waveform from the waveform memory, a waveform corresponding to the musical note is generated. According to still another waveform generating method of the present invention capable of achieving the above object, performance information corresponding to a real-time performance is stored, and automatic performance or automatic accompaniment is performed according to a generated tempo clock. In addition to generating the accompaniment sound, the performance information is reproduced in accordance with the tempo clock, and the performance event corresponding to the performance event is performed based on the performance event played in real time in accordance with the accompaniment sound and the reproduced performance information. To generate a waveform.

Next, a waveform selecting apparatus of the present invention capable of achieving the above object is a waveform selecting apparatus for selecting one or more partial waveforms from a plurality of partial waveforms stored in a partial waveform memory. A database for storing, for each partial waveform, attribute data indicating characteristics of two sounds, a sound corresponding to the partial waveform and a sound preceding the sound, a sound to be played and a sound to be played before the inputted sound. By the characteristic data of two sounds,
A search unit that extracts one or a plurality of partial waveforms having attribute data close to the input characteristic data of the two sounds as a partial waveform for generating the sound to be played, with reference to the database. Have.

According to another aspect of the present invention, there is provided a waveform selecting apparatus for selecting one or more partial waveforms from a plurality of partial waveforms stored in a partial waveform memory. For each partial waveform, a database storing attribute data indicating characteristics of two sounds, a sound corresponding to the partial waveform and the next sound, a sound to be played and a sound to play before With reference to the database, one or more partial waveforms having attribute data close to the input characteristic data of two sounds are extracted as partial waveforms for generating the sound to be played, by referring to the database. Search means for performing the search.

According to such a waveform generation method of the present invention, a waveform can be generated using an optimum partial waveform in accordance with the tone generation control information, and a musical tone rich in expression can be generated. become. Further, according to the performance data processing method of the present invention, it is possible to select in advance the partial waveform data most suitable for the characteristics of each note in the performance data and to add the partial waveform data to each note of the performance data. Therefore, by using the performance data, it is possible to select the optimum partial waveform data in the waveform generation in the automatic performance without increasing the load on the processing device when performing the automatic performance. Further, according to the waveform generation method of the present invention, even when a musical tone is generated by a real-time performance, the musical tone generated by the current performance according to a performance to be performed in the future by using performance information corresponding to the real-time performance. Characteristics can be controlled. Furthermore, if one or a plurality of partial waveforms having attribute data close to the characteristic data of the sound to be played and the sound to be played before or the sound to be played next are selected, the sound to be played is An optimal partial waveform can be selected as a partial waveform for generating a musical tone.

[0010]

FIG. 1 shows a hardware configuration of a waveform data recording / reproducing apparatus provided with a waveform generating method and a performance data processing method according to the present invention. In the waveform data recording / reproducing apparatus 1 shown in FIG. 1, the CPU 10 executes various programs to control the recording and reproducing operations in the waveform data recording / reproducing apparatus 1 (Central Processes).
sing unit), and the timer 11 is a timer that indicates the elapsed time during operation or generates a timer interrupt at a specific interval, and is used for time management of automatic performance and the like. ROM12
Is a ROM that stores a program for a dividing process and an attribute information adding process executed by the CPU 10 at the time of recording, a program for performance data processing and a note-on event process executed at the time of reproduction, and various data. (Read Only Memory). The RAM 13 is a main memory in the waveform data recording / reproducing apparatus 1 and has a C
RAM (Random) where the work area of PU10 is set
Access Memory).

The display 14 is a display on which various information is displayed at the time of recording or reproduction, and the operation device 15 is an operation device operated at the time of performing various operations at the time of recording or reproduction. May be included. The sound source 16 is a sound source that generates a musical tone based on performance data of a plurality of phrases that are automatically played at the time of recording, and may be any sound source such as an FM sound source, a PCM sound source, or a harmonic sound source. The mixer 17 is a mixer that mixes a musical tone generated by the sound source 16 and a musical tone reproduced by the reproducing circuit 24 to be described later and sends it to the DAC 18. The DAC 18 converts musical tone data supplied from the mixer 17 into an analog musical tone signal. A digital-to-analog converter that sends the digital signal to the sound system 19. The sound system 19 amplifies and emits the tone signal supplied from the DAC 18.

The microphone 20 receives an instrumental sound played by a player or a singing sound sung by the player, and a phrase waveform signal of the input instrumental sound or singing sound is sent to the ADC 21.
The ADC 21 is an analog-digital converter that converts the supplied sound waveform signal into digital phrase waveform data. The converted phrase waveform data is transmitted to an HDD (Hard Disk Drive) 23 under the control of the recording circuit 22. Will be recorded. The recorded phrase waveform data is divided into partial waveform data based on performance data reproduced at the time of recording by the sound source 16 as described later, and the divided partial waveform data is provided with attribute information. The division processing and the attribute information provision processing are performed by the CPU 10 executing a program of the division processing and the attribute information provision processing. Further, partial waveform data to be used is selected from the partial waveform data, and a tone color set for performance is created from the attribute information added to the selected partial waveform data. The reproduction circuit 24 searches the partial waveform management database for the designated timbre of the timbre set, retrieves the optimal partial waveform data from the attribute information, and generates performance waveform data. The generated performance waveform data is supplied to the mixer 17, the DAC
The sound is emitted from the sound system 19 via.

The MIDI interface 25 has an MID
A MIDI signal output from a MIDI-compatible device such as an I-key is input, and the reproduction circuit 24 generates performance waveform data based on the event data in the MIDI signal as described above, and the performance sound is emitted from the sound system 19. It has been made to be. Other interfaces 26 are L
A communication interface for a communication network such as an AN (Local Area Network), a public telephone network, or the Internet. The communication interface is connected to a server computer via the communication network, and can download desired performance data and the like from the server computer. Reference numeral 27 denotes a bus line for transmitting and receiving signals between the devices.

Next, the process of recording phrase waveform data and dividing it into partial waveform data in the waveform data recording / reproducing apparatus 1 according to the present invention will be described below with reference to FIGS. When recording the phrase waveform data in the waveform data recording / reproducing apparatus 1 according to the present invention, as shown in FIG. 3 (a), the performance data 30 composed of a plurality of phrases is supplied to the automatic performance means 31 for automatic performance. .
The phrase includes a note having characteristic information such as a pitch to be recorded. The automatic performance means 31 is mainly constituted by the CPU 10 and the sound source 16 shown in FIG.
Then, the CPU 10 executes the automatic performance program to generate sound source parameters based on the performance data read from the ROM 12 or the RAM 13, and supplies the sound source parameters to the sound source 16, whereby the performance waveform generated by the sound source 16 is generated. Data is in mixer 1
7. Sound is output from a headphone 36 connected to the sound system 19 via the DAC 18. The headphone 36 is worn by the player 32, and the player 32 can hear the performance sound. It should be noted that sequence data of one or more parts in the form of, for example, SMF (Standard MIDI File), including performance data of a part to be recorded and performance data of other parts, is created in advance. Is stored in the RAM 13 or the like.

The player 32 plays a musical instrument (not shown) or sings according to the performance sound while listening to the performance sound based on the performance data. This player 32
Is performed while listening to the performance sound, so that the performance sound or singing sound becomes a natural sound. The phrase waveform of the performance sound or the singing sound is supplied to the waveform recording means 33 via the microphone 20.
The phrase waveform is sampled to be digital phrase waveform data 34 and recorded for each phrase. In this case, the automatic performance means 31 and the waveform recording means 3
3 is synchronized by a clock synchronization line 35, and the phrase waveform data is recorded in synchronization with the performance timing. In this synchronization, the sampling clock of the waveform recording means 33 and the operation clock of the automatic performance means 31 are made common, and the timing of the start of the automatic performance and the timing of the recording start are matched, or
The difference between the two timings is stored. As a result, the synchronization of the recorded phrase waveform data 34 and the performance data to be automatically played over the entire section is guaranteed. Therefore, even if the tempo of the performance data is changed in the middle, synchronization is not affected at all.

FIG. 2 schematically shows the relationship between the performance data to be automatically played and the phrase waveform data. The performance data in this case is represented by a musical score on a staff as shown in FIG. 2A. While listening to the automatic performance sound based on the performance data from the headphones 36, the performance or singing is performed in accordance with the performance sound. The phrase waveform recorded as described above has a waveform as shown in FIG. In this case, the quarter note "E (mi)" of the performance data corresponds to the waveform a, and the quarter note "A (la)" of the performance data similarly corresponds to the waveform b, and the eighth note "E" of the performance data. F (fa) "corresponds to waveform c, eighth note" E (mi) "of performance data corresponds to waveform d, quarter note" C (do) "of performance data corresponds to waveform e, The musical note "F (fa)" of the performance data corresponds to the waveform f. Since the synchronization is performed as described above, each waveform matches the performance timing of the performance data and has a sound length corresponding to the note. Further, although it is not known from the waveform, since the performance is performed while listening to the performance sound based on the performance data, the timbre is also a natural timbre.

Further, the waveform recording means 33 is provided with the C shown in FIG.
It comprises a PU 10, an ADC 21, a recording circuit 22, and an HDD 23. Under the control of the CPU 10, the ADC 21 samples the phrase waveform, which is a performance sound or a singing sound input from the microphone 20, and converts the sampled waveform into digital phrase waveform data. The phrase waveform data 34 is written and recorded in a predetermined storage area of the HDD 23 by the recording circuit 22. The mode of the automatic performance may be any of the following modes. The first automatic performance mode is a solo part performance including performance data of only the part (recording part) of the phrase that the user wants to record. The second automatic performance mode is an all-part performance composed of performance data of a plurality of parts including the part of the phrase to be recorded. The third automatic performance mode is a minus one performance composed of performance data of one or a plurality of parts excluding the part of the phrase desired to be recorded, or performance data of the other parts.

Waveform data recording / reproducing apparatus 1 according to the present invention
In recording the phrase waveform in the above, the recording may be performed as shown in FIG. In the configuration shown in FIG. 3B, the player 32 listens to the automatic performance sound from the speaker 37 instead of the headphones 36. Here, FIG.
In the configuration shown in FIG. 3B, only the portions different from the configuration shown in FIG. 3A will be described. In the configuration shown in FIG.
The sound is emitted from 7 and the performer 32 listens. In this case, the automatic performance sound from the speaker 37 is the microphone 2
0, the automatic performance sound picked up by the microphone 20 is removed by the speaker sound removal processing means 38. The automatic performance sound is supplied from the automatic performance means 31 to the speaker sound removal processing means 38, and the automatic performance sound is removed by inverting the phase of the automatic performance sound and adjusting the time and amplitude. Like that.

In the automatic performance means 31, the automatic performance is performed in any of the above-described modes, and the sampling of the performance sound or singing sound of the performer 32 is performed by the waveform recording means 3.
3, the performance data (corresponding performance data) for waveform division corresponding to the part (recording part) to be recorded always exists in the recorded phrase waveform data. Therefore, as shown in FIG. 4, the division processing means 41 divides the phrase waveform data 40 into partial waveform data 43 of a note corresponding to each note of the performance data 44 in accordance with the performance data 44 automatically played. Next, the attribute information adding processing means 42 refers to the characteristic information of each note in the performance data 44 corresponding to the divided partial waveform data 43 with the attribute information indicating the attributes such as the pitch, the pitch and the intensity. Give. Thus, the division processing means 41 for dividing the phrase waveform data 40 into the partial waveform data 43, and the divided partial waveform data 43
Attribute information processing means 42 for giving attribute information to
This is performed by the PU 10 executing the division processing program and the attribute information addition processing program.

In this case, since the phrase waveform data 40 is synchronized with the performance data 44, the phrase data can be divided into partial waveform data 43 only by the sounding timing of the performance data 44. Alternatively, the phrase waveform data 40 may be subjected to frequency analysis (formant analysis) to correct the division position. That is, based on the temporary division position provisionally determined according to the performance data 44, a search is made for a formant start position of the partial waveform data 43 in the vicinity thereof,
The detected start position can be determined as the division position. In this way, it is possible to perform division at a musically accurate position, as compared to division based on only the analysis result of the phrase waveform data 40.

More specifically, in the division method using only the performance data 44, the performance data 44 to be automatically played and the phrase waveform data 40 are divided using the fact that they are completely synchronized. . Therefore, according to the data of each note included in the performance data 44 of the recording part, the time range corresponding to each note (from the start timing (note on) of the note corresponding to the note to the release of the sound after the start of release (note off) attenuates). , Or a time range up to the start of the sound of the next note) can be extracted as partial waveform data 43. However, since the phrase waveform data 40 is obtained by sampling a performance performed by a human, the start timing of the note in the performance data 44 of the recording part and the start timing of the waveform corresponding to the note in the phrase waveform data 40 are not always present. There are times when it is not. Therefore, based on the start timing of each note in the performance data 44 of the recording part, the waveform of the section before and after the phrase waveform data 40 (for example, several seconds before and after the start timing) is analyzed, and the start timing of the note in the phrase waveform data 40 is determined. If the division position is detected and corrected, the partial waveform data 43 can be cut out accurately.

As a waveform analysis technique used for correcting the division position, a formant analysis and an FFT analysis can be adopted. In this formant analysis,
On the basis of the start timing of each note in the performance data 44 of the recording part, an LPC (Linear Prediction Code) is obtained from the correlation function of the waveform data before and after the phrase waveform data 40.
g) Calculate the coefficient. Next, the LPC coefficient is converted into a formant parameter, and a formant in the phrase waveform data 40 is obtained. Then, the start timing of the note is detected from the rising position of the formant corresponding to the note. Thus, the partial waveform data 43 can be cut out at musically accurate positions and divided into partial waveform data 43 for each note. In the FFT analysis, first, the phrase waveform data 4 is moved while moving the time window for the section before and after the phrase waveform data 40 based on the start timing of each note in the performance data 44 of the recording part.
Perform a fast Fourier transform of 0. Next, the trajectory of the fundamental tone and the plurality of overtones corresponding to the note are detected, and the start timing is detected from the rising position in the detected trajectories of the fundamental tone and the overtones. According to this method as well, the partial waveform data 43 can be cut out at musically accurate positions and divided into partial waveform data 43 for each note. Here, the formant analysis and the FFT analysis have been described, but other analysis methods such as pitch analysis and envelope analysis may be used for the analysis for correcting the division position.

When the division process is completed, an attribute information adding process for adding attribute information to each of the divided partial waveform data is performed. In this attribute information adding process, attribute information corresponding to the characteristic information of the note is added to the partial waveform data 43 corresponding to each note in the performance data 44. The attribute information includes pitch information used as an original key, intensity information used as an original intensity, pitch information used as an original pitch, and preceding sound information used to select partial waveform data. And one or more pieces of post-sound information. That is, the pitch information, the intensity information, and the pitch information are respectively the pitch, the intensity, and the pitch of the event of the note (referred to as the current tone) corresponding to the position where the partial waveform data of the performance data 44 is cut out. Has a value. The preceding sound information is information on whether the pitch of the preceding sound is higher or lower than the current sound, or information on the pitch difference from the preceding sound, information on whether the intensity of the preceding sound is higher or lower than the present sound, Or information on the intensity difference from the previous sound, and information on whether the length of the previous sound is longer or shorter than the current sound, or information on the length difference from the previous sound, and other relationships between the previous sound and the current sound Etc.
In this case, the values of these pieces of information can be determined based on the events of the current sound in the performance data 44 and one or more notes before it. Further, the post-sound information is information on the relationship between the post-sound and the current sound when the pre-sound in the pre-sound information is replaced by the post-sound. Similarly, the values of these pieces of information can be determined based on the current sound in the performance data 44 and one or more subsequent note events.
Further, when the performance data 44 is performance data including music symbols such as slurs, trills, and crescendos, the music symbols may be included in the attribute information. This attribute information is used as a reference for selecting the partial waveform data or as information used as a parameter when processing the partial waveform data. It should be noted that, by recording a musical note corresponding to the desired attribute information in the performance data, partial waveform data corresponding to the desired attribute information can be obtained.

As shown in FIG. 4, the division processing means 41 divides the data into partial waveform data 43, and the divided partial waveform data 43 is given attribute information by the attribute information provision processing means 42. The divided partial waveform data 43 and the attribute information provided are stored in the HDD 23 together with the performance data and the like as one piece of recorded data. Since the phrase waveform data 40 is recorded for each instrument type and vocal timbre, the recorded data is recorded for each timbre. FIG. 5 shows the data structure of the recorded data. As shown in FIG. 5, the recorded data manages performance data, which is a plurality of phrases automatically played, and where in the storage area the phrase waveform data corresponding to the performance data of each phrase is stored. Phrase waveform management information,
Partial waveform management information including a start position and an end position in a partial waveform data storage area obtained by dividing the phrase waveform data and attribute information of the partial waveform data, and a phrase waveform 1, a phrase waveform 2, and a phrase waveform 2 corresponding to performance data of each phrase. ... Note that the number i assigned to the phrase waveform i is a phrase waveform ID described later. The data of each phrase waveform of each phrase waveform is divided into a plurality of partial waveform data by the method described above.
Is exemplified. That is, the phrase waveform 1 has a partial waveform 1-1, a partial waveform 1-2,..., A partial waveform 1-6.
Are divided into six partial waveform data. The numbers ij given to the partial waveform data ij are as follows:
This is a partial waveform ID to be described later.

The recorded data shown in FIG. 5 is data obtained as a result of dividing the phrase waveform data recorded for each piece of performance data and performing an attribute assigning process, and the phrase waveform data is composed of an instrument type and a vocal tone. Recorded every time. That is, the recorded data is created for each timbre, so that by reading and using the corresponding partial waveform data from the recorded data in accordance with the performance event data, the performance waveform data of the timbre can be obtained. But,
In a plurality of partial waveform data in the recorded data shown in FIG. 5, partial waveform data in which the pitch of attribute information and the like overlap and unnecessary partial waveform data are also recorded. Therefore, the partial waveform data to be used is selected from the recorded data, and a tone color set for performance is created based on the selected partial waveform data. When creating a timbre set, selection information as to which partial waveform data to select, timbre parameters, and the like are recorded in the recorded data as timbre management information, and the timbre set is created according to the timbre management information. become. For this reason, the recorded data shown in FIG. 5 can be called a tone color set creation tool. The timbre management information includes selection information for selecting the partial waveform data selected as desired by the user, timbre parameters set by the user, and the like.

FIG. 6 shows the data structure of this tone color set. The tone color set for performance shown in FIG. 6 is composed of tone color data for each tone color such as “violin 1”, “male voice 4”, “trumpet 2”,. The data structure of each tone color data is the same, and the data structure of “male voice 4” is shown in FIG. 6 as an example. As shown in FIG.
The timbre data includes a header, a partial waveform management database, a timbre parameter, and a partial waveform pool. The partial waveform pool stores partial waveform data selected from a plurality of phrase waveform data in accordance with the tone color management information in the recorded data shown in FIG. 5, as shown in FIG. ... pooled. In this case, when partial waveform data corresponding to many pitches is selected according to the tone color management information, the range to which the partial waveform data is assigned can be finely set. In addition, even when the pitch information is the same, the partial waveform data in which the pitch information and the intensity information are different, and the partial waveform data in which the information such as the pitch difference from the preceding sound and the pitch of the preceding sound is higher or lower are different are also selected. Timbre management information.

The partial waveform management database in the tone color set thus created includes the start position and the end position of the storage area in the partial waveform pool and the partial waveform in the partial waveform data selected according to the tone color management information. Data attribute information is recorded. The timbre parameters included in the timbre management information are stored in the timbre parameter storage area as they are, and the timbre parameters record the envelope parameters and filter coefficients of the timbre. For the tone color data created for each tone color in the tone color set, a plurality of tone color data can be created from one piece of recorded data by selecting partial waveform data in a different combination from the recorded data shown in FIG. In the waveform data recording / reproducing apparatus 1 according to the present invention, an automatic performance or a real-time performance can be performed using the tone color set thus created.

FIG. 7 shows a data structure of a partial waveform management database in which attribute information and the like of a partial waveform selected according to the tone color management information stored in the partial waveform pool are recorded. The partial waveform management database stores a partial waveform 1 record, a partial waveform 2 record, and a partial waveform showing characteristics of each partial waveform of the partial waveform 1, the partial waveform 2, the partial waveform 3, the partial waveform 4,... It is composed of record information of three partial waveform records, four partial waveform records,.... The data structure of each record information is the same, and the details of the data structure of three partial waveform records are shown in FIG. That is, the record information represented by the three partial waveform records is the ID of the partial waveform (PWI
D), the partial waveform name (PWNAM), the partial waveform creation date (PWDAT), and the partial waveform creator (PWAUT). The IDs of the partial waveforms are the phrase waveform ID (upper 10 bits) and the partial waveform ID as shown in FIG.
(Lower 6 bits). Note that the phrase waveform ID is the phrase waveform ID of the phrase waveform from which the partial waveform was cut out, and the partial waveform number is a number indicating the order of the partial waveform cut out from the phrase waveform. By using the ID (PWID) of the partial waveform having such a configuration, the phrase waveform from which the partial waveform is extracted can be specified. When the user selects the partial waveform, the source of the partial waveform is selected. By listening to the phrase waveform, it becomes possible to confirm what kind of characteristic partial waveform is.

Further, each record information includes a pitch (PWPIT) and an intensity (PWI) of the sound in the partial waveform data.
NT), pitch (PWLEN), the pitch ratio of the next note (PWFPR), the intensity ratio of the next note (PWFIR), and the pitch of the next note indicating the relationship between the two notes of the note and the next note Ratio (PWFL
R) and information on the pitch ratio of the preceding sound (PWBPR), the intensity ratio of the preceding sound (PWBIR), and the pitch ratio of the preceding sound (PWBLR), which indicate the relationship between the two sounds of the preceding sound and the preceding sound. In. In addition,
The information of the ratio among these pieces of information is not limited to the ratio, and may be difference information. By the way, each piece of information in these pieces of record information corresponds to each piece of information in the attribute information added in the attribute assignment information processing. The correspondence is as follows. Les code information attribute information pitch (PWPIT) pitch information strength (PWINT) intensity information tone length (PWLEN) pitch ratio of the sound length information Tsugion (PWFPR) following tone information Tsugion intensity ratio ( PWFIR) Back sound information Next sound pitch ratio (PWFLR) Back sound information Previous sound pitch ratio (PWBPR) Previous sound information Previous sound intensity ratio (PWBIR) Previous sound information Previous sound pitch ratio (PWBLR) Previous Sound information As described above, the values of these information are, after clipping of each partial waveform data, the event of the note corresponding to the cut position of the performance data corresponding to the phrase waveform data from which the partial waveform data is cut, and the before and after events. The note is automatically determined based on the event. However, a configuration may be adopted in which the user can individually input the value of each piece of information or change the value of each piece of information automatically determined.

Further, a start address (P) in the storage area of the partial waveform pool where the partial waveform is stored.
WSA), end address (PWEA), beat address (PWBA), and loop address (PWLA). In this case, the start address and the end address are essential information, but the beat address and the loop address may not be provided. The beat address is
A beat position (a position at which a beat is felt) in the partial waveform is shown. When the partial waveform data is used, the beat position is set to the timing of the musical note in the performance data. The beat address is, for example, performance data played by software, and is suitable information when the start position of each played note is not clear. Further, the loop address is the address of the loop waveform when a steady portion having little change in the partial waveform is replaced with the repetition of the loop waveform. Even when the partial waveform is shortened by the loop, the sound duration (PWLE) of the partial waveform is
N) Information does not change. Thus, the sound duration (PWLE)
N) The information is not the length of the waveform data but information indicating the length of the note when the sound is played.

Further, the degree of slur (PWSLR) and the vibrato depth (PW
VIB) and other information. here,
The value of the degree of the slur is such that the note-on of the previous sound and the note-on of the current sound overlap, and the overlap is small and the intensity difference between the previous sound and the current sound is small. The vibrato depth is the ratio of the fluctuation width of the pitch of the partial waveform data to the basic pitch of the same partial waveform data. The above-described partial waveform management database is divided into two databases: a database that stores address information such as a start address and an end address of the partial waveform, and a database that stores the remaining information of the partial waveform for searching for the partial waveform. May be. In this case, only the latter search database is required to select the partial waveform, and if the former address information database is incorporated in the sound source 16,
The processing to be performed by the CPU 10 can be reduced.

Here, in the tone color set creation tool having the data configuration as shown in FIG. 5, the procedure is summarized in accordance with the flowchart of FIG. 9 showing the procedure for creating the tone color set for performance shown in FIGS. explain. This procedure is performed by a user who creates a tone color set. First, a user who creates a tone color set prepares performance data of a plurality of phrases including at least a note having a pitch required for a recording part in a performance data storage area of FIG. 5 (step S1). In this case, notes having characteristic information necessary for recording include pitch, velocity, velocity,
Performance data of a phrase including a note having characteristic information such as preceding sound information and a pitch difference from the preceding sound may be used. Next, while the musical tone automatically played by the waveform data recording / reproducing apparatus 1 is supplied to the player based on the prepared performance data of a plurality of phrases, the phrase waveform played or sung by the player is displayed on the waveform data recording / reproducing apparatus 1. 5 is recorded in any one of the phrase waveform storage areas (step S2).

The recorded phrase waveform data is divided by the waveform data recording / reproducing apparatus 1 into partial waveform data based on the performance data in the recorded data, so that the phrase waveform data is divided into partial waveform data.
Attribute information related to a note corresponding to the divided partial waveform data is added. The start position and end position of each of the divided partial waveform data and the attribute information given to each of the partial waveforms are used as partial waveform management information of each of the partial waveforms in FIG.
Is stored in the partial waveform management information storage area (step S
3). Next, the partial waveform data is selectively combined according to the operation of the user, and tone parameters such as an envelope parameter are prepared, and the selection information indicating the selected partial waveform data and the prepared tone parameters are stored in the tone management. The information is stored in the tone color management information storage area of FIG. 5 (step S4).

Based on the timbre data for one timbre completed by the above steps S1 to S4, timbre data for one timbre of the timbre set of FIG. 6 is prepared in the storage area as follows. First, among the plurality of partial waveform data stored in the phrase waveform storage area of FIG. 5, the partial waveform data selected by the selection information in the timbre management information is copied to the partial waveform pool of the timbre of FIG. Start position and end position of each partial waveform data after copying, and I of each copied partial waveform data
Attribute information including D, name, creator, etc. is stored in the partial waveform management database of FIG. Next, the timbre parameters in the timbre management information of FIG. 5 are copied to the timbre parameter storage area of the timbre of FIG. Finally, header information indicating the type of the timbre, the timbre name, the timbre data capacity, and the like is shown in FIG.
And the tone color data for performance in the tone color set is completed. In step S3, after automatically determining the division position of the phrase waveform data into the partial waveform data and the attribute information to be added, the user may be able to arbitrarily modify the division position and the attribute information.

Next, to apply the performance timbre set shown in FIG. 6 to the automatic performance in the waveform data recording / reproducing apparatus 1 according to the present invention, the performance data is processed in advance before the automatic performance. The performance data is arbitrary performance data desired by the user for automatic performance. Accordingly, a flowchart of performance data processing when a tone color set is applied to an automatic performance is shown in FIG. 10, and the description thereof will be given below. When the performance data processing shown in FIG. 10 is started,
In step S10, performance data to be processed is specified. This designation is made by the user in the ROM 12 or RAM 1
3 designates the performance data to be automatically played out of the performance data stored in the memory 3. Here, for the sake of simplicity, the performance data is a single-melody performance data having a one-part configuration. Next, in step S11, the pointer is set to the first note of the specified performance data. Then, step S12
At the note indicated by the pointer (in this case, the first note of the performance data), pitch, intensity (velocity), pitch, previous / next note pitch ratio, previous / next note intensity ratio, previous note sound/
Based on the characteristic information such as the sound length ratio of the next sound, information for search in the partial waveform management database of the specified timbre in the timbre set is searched, and optimal partial waveform data is detected.

Next, in step S13, designating information for designating the optimum partial waveform data detected corresponding to the note indicated by the pointer is embedded in the performance data. Specifically, immediately before the performance event data corresponding to the note, the designation information of the partial waveform data which is a meta event or a system exclusive message is inserted. Subsequently, in step S14, the pointer is moved to the next note. In step S15, it is determined whether or not there is a note at the movement destination. If there is a note, the process returns to step S12.
If there is a note in this way, the processes of steps S12 to S15 are repeatedly performed, and the designation information of the optimum partial waveform data corresponding to one note is embedded in the performance data for each process. Become. When the embedding of the specification information of the partial waveform data into the last note of the performance data is completed, it is determined in step S15 that there is no note, and the process proceeds to step S16. In step S16, a process of storing the performance data, in which the designation information of the partial waveform data is embedded, under another name is performed. When the above processing ends, the performance data processing processing ends. When the performance data after processing is designated and the waveform data recording / reproducing apparatus 1 automatically performs, the start timing (note-on timing) of each note in the performance data
Each time, the partial waveform data specified by the specification information embedded immediately before the note-on event is read from the HDD 23 and sent to the reproduction circuit 24 together with the tone color parameter corresponding to the note-on. In this way, a musical tone corresponding to each note in the performance data is generated based on the specified partial waveform data and automatically played.

In step S13, the designation information of the partial waveform data is embedded as a meta event or a system exclusive. However, the present invention is not limited to this, and the designation information may be embedded as other events. If the performance data to be automatically played is performance data of a plurality of parts, the processing of FIG. 10 may be performed only on a desired one of the parts, or an arbitrary plurality of the parts may be performed. It may be performed for the part of. Further, the processing in FIG. 10 may be not only performance data of a single melody but also performance data in which a plurality of notes are simultaneously overlapped. However, if it is desired to select partial waveform data based on characteristic information such as whether the preceding sound or the succeeding sound is upper or lower, or the pitch difference between the preceding sound and the succeeding sound, the performance data to be processed is all monophonic. Must be performance data.

FIG. 11 is a flowchart showing the details of the partial waveform selection process performed in step S12 in the performance data processing process shown in FIG. In the partial waveform selection processing shown in FIG. 11, in step S30, current sound data of the current sound pitch (PPIT), the current sound intensity (PINT), and the current sound pitch (PLEN) are obtained. Next, in step S31, next sound data of the next sound pitch (FPIT), the next sound intensity (FINT), and the next sound pitch (FLEN) are obtained. Further, in step S32, preceding sound data of the pitch of the preceding sound (BPIT), the intensity of the preceding sound (BINT), and the length of the preceding sound (BLEN) are obtained. Next, in step S33, the pitch ratio of the next sound (FPR) and the pitch ratio of the previous sound (BPR) are calculated using the acquired data by the following equations (1) and (2). . FPR = (FPIT-PPIT) / PPIT (1) BPR = (BPIT-PPIT) / PPIT (2)

Further, the intensity ratio of the next sound (FIR) and the intensity ratio of the preceding sound (BIR) are calculated by the following equations (3) and (4) at step S34 using the acquired data. You. FIR = (FINT-PINT) / PINT (3) BIR = (BINT-PINT) / PINT (4) Furthermore, using the acquired data, step S35 is performed.
At the pitch ratio of the next sound (FLR) and the pitch ratio of the previous sound (BLR)
Is calculated by the following equations (5) and (6). FLR = (FLEN-PLEN) / PLEN (5) BLR = (BLEN-PLEN) / PLEN (6) Further, other detections are performed in step S36. The degree of slur is detected on the basis of the variation in the intensity of a plurality of preceding and succeeding sounds, and the swing width of the pitch bend data is detected as the vibrato depth.

The pitch of the current sound (PPIT), the intensity of the current sound (PINT), the length of the current sound (PLEN) obtained in step S30, and steps S33 to S3
5, the pitch ratio of the next sound (FPR), the pitch ratio of the previous sound (BPR), the intensity ratio of the next sound (FIR), the intensity ratio of the previous sound (BIR), the pitch ratio of the next sound (FLR), the information of the ratio of the length ratio of the preceding sound (BLR), and the information detected in step S36 become characteristic information of the sound of the note indicated by the pointer. Then, in step S37, a pattern comparison between the characteristic information and the characteristic information of each partial waveform is performed to select a partial waveform that has the best matching characteristic information of both. The ID of the selected partial waveform is stored in the SPWID register. Thus, the partial waveform selection processing in step S12 ends.

The pattern comparison process executed in step S37 is as shown in the flowchart of FIG. In this pattern comparison process, step S
The pitch (PPIT), intensity (PINT), and duration (PLEN) of the current sound corresponding to the note indicated by the pointer at 40 are represented by the pitch (P) of the sound in each partial waveform.
WPIT), intensity (PWINT), duration (PWLEN)
Are compared with each other, and based on the comparison result, candidate partial waveforms are narrowed down in order to reduce the amount of calculation. As an example of a narrowing-down method, first, the pitch (P
WPIT) and the pitch of the current sound (PPIT) are in the range ΔP
The difference between the sound intensity of the partial waveform (PWINT) and the intensity of the current sound (PINT) is in the range ΔI, and the difference between the sound length of the partial waveform sound (PWLEN) and the sound length of the current sound (PWLEN) is in the range ΔL. Then, the partial waveforms respectively included are selected. In this case, if the number of selected partial waveforms is too small, the range Δ
P, ΔI, and ΔL are expanded (the conditions are relaxed), and the selection of the partial waveform is executed again according to the range. However, if a sufficient number of partial waveforms has been obtained as the number of selected partial waveforms, the narrowing-down process ends. Further, the following narrowing method may be used. First, for all partial waveforms, a distance (PND), which will be described later, indicating the degree of separation between each partial waveform and the current sound is calculated in order to check the similarity between each partial waveform regarding pitch, intensity, and duration. Next, n partial waveforms are selected in ascending order of distance (PND), which is a similar order.

After narrowing down the partial waveform in this way,
The distance (PND) regarding the pitch, intensity, and duration of each of the partial waveforms narrowed down in step S41 and the note indicated by the pointer is calculated by the following equation (7). However, if the distance (PND) is calculated using the equation (7) in step S40, this step is skipped. PND ² = ap (PPIT-PWPIT) ² + bp (PINT-PWINT) ² + cp (PLEN-PWLEN) ² (7) where ap, bp and cp are PND
This is a coefficient for calculation. Also, the pitch ratio, intensity ratio, and pitch ratio of the next note in each partial waveform narrowed down in step S42, and the pitch ratio, intensity ratio, and pitch ratio of the next note in the note of the note indicated by the pointer Distance to (FND)
Is calculated by the following equation (8). FND ² = af (FPR-PWFPR) ² + bf (FIR-PWFIR) ² + cf (FLR-PWFLR) ² (8) where af, bf, cf are FND
This is a coefficient for calculation.

Further, the pitch ratio, intensity ratio, length ratio of the preceding sound in each partial waveform narrowed down in step S43, and
The distance (BND) with respect to the pitch ratio, intensity ratio, and length ratio of the preceding sound in the sound of the note indicated by the pointer is calculated by the following equation (9). BND ² = ab (BPR-PWBPR) ² + bb (BIR-PWBIR) ² + cb (BLR-PWBLR) ² (9) where ab, bb and cb are BND
This is a coefficient for calculation. Next, the distance (PND) and the distance (F) calculated in steps S41 to S43 are calculated.
Using ND) and distance (BND), the total distance (TOTALD) indicating the overall similarity between the partial waveform and the note pointed to by the pointer is calculated in step S44 by the following equation (10). TOTALD = at · PND + bt · FND + ct · BND (10) where at, bt, ct are TO
These are coefficients for calculating TALD, and the tendency of these coefficients is in the order of at>ct> bt. As described above, the coefficients for calculating the PND, the coefficients for calculating the FND, and the BN
By setting any one or more of the coefficients for calculating D and the coefficient for calculating TOTALD to zero, the corresponding element can be prevented from being included in the distance calculation.

When the total distance (TOTALD) is calculated for each of the narrowed partial waveforms in step S44,
The total distance (TOTAL calculated in step S45)
D) Select the partial waveform with the smallest value and set its ID to SP
Save in WID register. Thus, the pattern comparison processing ends. In step S12 in the performance data processing, one of the partial waveforms is selected regardless of the state of the current data of the note indicated by the pointer in the performance data. However, a minimum condition is set for the matching degree between the current data and the attribute data, and when the attribute data of any of the partial waveforms are too far apart and are not similar, it can be determined that “there is no corresponding partial waveform”. Is also good. And
When it is determined that there is no partial waveform, it is preferable to warn the user or to generate a sound with a normal waveform memory sound source instead of the partial waveform. Further, instead of selecting only one partial waveform, a second candidate, a third candidate,...
May be automatically selected, and one partial waveform can be selected from the candidates in accordance with a user's instruction.

FIG. 1 is a flowchart of a note-on event process when a timbre set is applied to a real-time performance in the waveform data recording / reproducing apparatus 1 according to the present invention.
3 and the description is given below. When a note-on event is generated by pressing a key included in the operation unit 15 or a note-on event is supplied via the MIDI interface 25, a note-on event process is started and generated in step S20. The attribute information of the partial waveform management database of the designated timbre in the timbre set is searched based on the information such as the pitch and the intensity in the note-on event, and the optimal partial waveform data is selected. The selection criterion for the partial waveform data in this case is (1)
(2) the pitch and intensity of the note-on, (3) the pitch of the note-on, and information on whether the pitch of the previous note is up or down, (4) the note-on (5) Other combinations of information such as the note-on and the intensity of the previous sound, and the like.

When the optimum partial waveform data corresponding to the note-on event is selected, a channel to be sounded by the reproducing circuit 24 is assigned in step S21, and the selected partial waveform is assigned to the sound channel assigned in step S22. Data information and tone parameters are set. In this case, the pitch shift amount is also set based on the pitch difference between the pitch at the note-on event and the pitch of the selected partial waveform data. Next, step S23
The note-on is transmitted to the tone generation channel assigned by, and the reproduction circuit 24 reproduces the tone waveform based on the set partial waveform data, the tone color parameter, and the pitch shift amount. When the musical tone waveform corresponding to the note-on event is reproduced, the note-on event process ends, and when a note-on event based on the real-time performance occurs again, the note-on event process is activated again and the tone corresponding to the note-on event described above is activated. The process of reproducing the waveform is repeatedly performed.

In the real-time performance, the length of the sound to be played is not known until the corresponding note-off, so that it is not possible to control the time axis of the partial waveform data according to the sound length. Therefore, in the partial waveform data for real-time performance, a loop portion for performing loop playback is set in the steady portion after the attack of the musical tone waveform as in the case of a normal waveform memory sound source, and an attenuation envelope is provided in accordance with the loop playback and note-off. The control of the musical tone length is performed according to. Also, as a modified example of applying the tone set to real-time performance, if a sound delay of several seconds to several tens of seconds is allowed from the occurrence of a note-on event to the sound generation, the performance event during that time should be stored in a buffer. Accordingly, when selecting the partial waveform data, information on the note length of the note-on, the subsequent sound, and the like can be used.

FIG. 14 shows a flowchart of the partial waveform selection processing executed in step S20 in the note-on event processing. In the partial waveform selection process of step S20, the pitch (PPIT) and the intensity of the current sound (PINT) of which the note-on is performed in step S51 are acquired. However, the pitch of the current sound (PLEN) cannot be acquired because the note is on. Next, step S5
At 2 the pitch of the previous note (BPI
T), previous sound intensity (BINT), previous sound duration (BLE)
N) is obtained. Next, step S5
The pitch ratio (BPR) between the pitch of the current sound (PPIT) and the pitch of the previous sound (BPIT) obtained in step (3) is as described in (2) above.
It is calculated by the formula. Further, the intensity of the current sound (PINT) and the intensity of the previous sound (BIN) acquired in step S54.
The intensity ratio (BIR) with respect to T) is calculated by the above equation (4). Further, other detections are performed in step S55. In this detection, the degree of slur is detected based on the overlap time with the previous sound, the variation of the intensity, and the like, and the fluctuation width of the pitch bend data is determined by the vibrato. Detect as depth.

Next, from step S51 to step S51
The pitch (PPI) of the current note-on note acquired at 55
T), pattern comparison is performed between the characteristic data of the current sound intensity (PINT), the calculated pitch ratio of the previous sound (BPR), and the intensity ratio of the previous sound (BIR), and the characteristic data of each partial waveform. Thus, a partial waveform that best matches these feature data is selected in step S56. The ID of the selected partial waveform is stored in the SPWID register. Thus, the partial waveform selection processing in step S20 ends. Note that the pattern comparison processing can be performed in the same manner as the above-described pattern comparison processing shown in FIG. However, since there is no data of the next sound, a pattern comparison is performed based on data on the current sound and the previous sound. Specifically, the distance (FND) for the next sound is not calculated, and the distance (PND) for the current sound is calculated with the coefficient cp set to zero.
In the calculation of the distance (BND) for the preceding sound,
In the equation (9), the pitch of the preceding sound (BLEN) is used in place of the pitch ratio (BLR) of the third term, and the pitch of the preceding sound (= PWBLR × PWLEN) is used instead of the pitch ratio (PWBLR). I do. Then, the calculated distance (PND) and the distance (B
ND), the total distance (TOTALD) is calculated, the partial waveform having the smallest total distance (TOTALD) is selected, and its ID is stored in the SPWID register.

Incidentally, in the waveform data recording / reproducing apparatus 1 according to the present invention, the timbre set can be applied to a combination of the automatic performance and the real-time performance. In the automatic performance at this time, an automatic performance based on performance data of, for example, 16 parts is performed, and any of the parts is designated as a real-time performance part. If the first part is a real-time performance part, musical tones of the second part to the sixteenth part are generated by automatic performance. In a real-time performance performed by a player in accordance with the automatically played musical tone, a musical tone waveform is generated by executing the note-on event processing shown in FIG. When performing in real time in accordance with any automatic performance such as such automatic rhythm, automatic accompaniment, and accompaniment part, a tempo clock can be obtained from the automatic performance. Therefore, in this case, if the performance data of the real-time performance part performed by the player is prepared, the performer generally performs the performance according to the real-time performance part, so that the note-on event and the note-off event in the real-time performance are performed. Etc., the future situation of the event can be detected. Therefore, in the process of selecting a partial waveform in step S20 of the note-on event process shown in FIG. 13, a partial waveform can be selected using the note length information of the current note-on and the information of the sound to be played next. Become like Note that the performance data processed in the performance data processing process of FIG. 10 may be used as the performance data for automatic performance.

As described above, when performing an automatic performance, a tempo clock can be obtained from the automatic performance.
That is, the tempo timer generates the tempo interrupt only at the time of the automatic performance at the time interval corresponding to the tempo.
Each time this tempo interrupt occurs, the tempo counter counts up by a predetermined value. The current count value POS of the tempo counter indicates the progress position of the automatic performance in the performance data. However, this position is not an address position on the memory of the performance data, but a position based on a tempo clock such as what bar, what beat, what clock, and so on. The performance data also includes a real-time performance part played by the player, and the current position of the real-time performance part is determined in synchronization with the automatic performance of the other parts. .

FIG. 15 is a flowchart of an automatic performance start process in which the tempo counter is initialized. When the automatic performance start process is started, step S
At 60, performance data to be automatically played is specified. Next, in step S61, the value POS of the tempo counter is set to an initial value. Next, the first event position in the performance data specified in step S62 is determined, and the tempo timer is started in step S63. As a result, a tempo interrupt is generated at regular intervals by the tempo timer, and in the tempo interrupt processing, the events of the performance data are sequentially reproduced from the determined event.

FIG. 16 is a flowchart of a tempo interrupt process for updating the tempo counter. When a tempo interrupt occurs, a tempo interrupt process is started and a step S7 is executed.
When the tempo is interrupted at 0, the value POS of the tempo counter is counted up and updated by a predetermined value. Then
In step S71, it is determined whether the current time is the event position. Here, if it is determined that the event is at the event position, the process branches to step S72, where the event at that position is reproduced. When the reproduction process ends, the position of the next event is determined at step S73, and the process proceeds to step S74. in this case,
If the event is a real-time performance part event, generation and control of musical tones in accordance with the event are not performed. If the event is an event of another performance part, generation and control of musical sounds in accordance with the event are performed. It is.
If it is determined in step S71 that the current position is not the event position, the process directly proceeds to step S74 to determine whether the current time is the end position. If the end position is determined, the process branches to step S75 to stop the tempo timer and end the tempo interrupt process. In this case, the automatic performance ends. If it is determined in step S74 that the current position is not the end position, the tempo interrupt processing ends.

Next, FIG. 17 shows a flowchart of a note-on process of a real-time performance performed in combination with an automatic performance. When the note-on of the real-time performance is detected, the note-on process of the real-time performance is started, and the note number of the note-on detected in step S80 is stored in the PPIT register, and the velocity is stored in the PINT register. . Next, in step S81, the pitch of the previous sound (BPI
T), previous sound intensity (BINT), previous sound duration (BLE)
N) is obtained. Furthermore, in a range around the value POS of the tempo counter indicating the current performance position, an event in the performance data of the real-time performance part in the performance data corresponding to the detected note number (PPIT) and velocity (PINT) is determined. It is detected in step S82. Subsequently, step S83
The pitch length (PLEN) data of the note-on sound is acquired based on the sound of the event detected at. In this case, note length information is obtained by detecting a note-off event corresponding to the event (note-on event) from the performance data of the real-time performance part.

Further, based on the next sound in the performance data of the event detected in step S84, the next sound of the next sound pitch (FPIT), the next sound intensity (FINT), and the next sound pitch (FLEN) Data is obtained. Next, a partial waveform selection process is performed based on the note-on data obtained in step S85 and data obtained from the performance data based on the note-on data. In this partial waveform selection process, steps S33 to S3 in the partial waveform selection process shown in FIG.
7 is performed. As a result, a partial waveform having feature data most matching the feature data of the note-on sound is selected. When the optimum partial waveform corresponding to the note-on sound is selected, a channel to be sounded by the reproduction circuit 24 is assigned in step S86, and the selected partial waveform data of the selected sound channel is assigned to the tone generation channel assigned in step S87. Information and tone parameters are set. In this case, the pitch shift amount is also set based on the pitch difference between the detected note-on note number (PPIT) and the pitch of the selected partial waveform.

Next, note-on is sent to the tone generation channel assigned in step S88, and the reproduction circuit 2
Reference numeral 4 reproduces a musical tone waveform based on the set partial waveform, tone color parameter, and pitch shift amount. When the musical sound waveform corresponding to the note-on is reproduced, the note-on processing of the real-time performance ends, and when the note-on of the real-time performance occurs again, the note-on processing is restarted to generate the musical sound waveform corresponding to the above-described note-on event. Playback processing is repeatedly performed. In the note-on processing of the real-time performance shown in FIG. 17, the pitch (PPIT), the intensity (PINT), the pitch of the previous sound (BPIT), the intensity of the previous sound (BINT), the pitch of the previous sound are calculated from the real-time performance. Although (BLEN) is acquired, data obtained from performance data of a real-time performance part may be used as these data used for selecting a partial waveform. In this way, a partial waveform to be selected at each note-on of the real-time performance can be selected in advance before the real-time performance. However, the data of the real-time performance is used as it is for controlling the parameters for sound generation.

Note that the key range is divided into predetermined ranges, and a note having a pitch required for multi-sampling in which a sampling waveform is allocated to each divided range, is used to record the performance data of the part to be recorded in the phrase waveform recording shown in FIG. If all of them are included in the phrase, all the necessary partial waveform data of the pitch will be recorded in the recorded data.
Therefore, in this case, the performance based on the multi-sampled partial waveform data can be performed immediately after the performance of the phrase is completed. At this time, all the notes of the required pitch and performance intensity may be included in the phrase of the performance data. Further, if the partial waveform data obtained by dividing the phrase waveform data is set so as to be used as a tone color set, the procedure for creating the selected tone color set can be omitted and the performance can be immediately performed.

Further, each information may be used for a purpose other than its original purpose. For example, when singing voice is sampled, “strength information” may be used as “phoneme information” for determining a phoneme of a voice. Specifically, for example, strength = 60 is assigned as information indicating the phoneme “Ah-ah-ah-”, and strength = 61 is used as the information indicating the phoneme “La-La-La-”.
Is assigned, and strength = 62 is assigned as the information indicating the phoneme “d ぅ -d ぅ -d ぅ-”. As described above, when “strength information” is used as “phoneme information”, “strength information”
, It becomes possible to collect partial waveform data corresponding to the same phoneme. Furthermore, the performance data to be processed in the performance data processing may be used as it is as the corresponding performance data that is the performance data used for recording. In this way, after the recorded waveform data is divided into partial waveform data, the obtained divided waveform data can be pasted directly to the performance data, simplifying the process of detecting the optimal partial waveform data. Can be Furthermore, by editing each note of the performance data to which the partial waveform data is pasted, it is possible to indirectly edit the recorded phrase waveform data. Furthermore,
After recording the phrase waveform, a person may input performance data corresponding to the phrase waveform and create attribute information based on the performance data. Furthermore, the recorded phrase waveform may be frequency-analyzed, the corresponding performance data may be automatically generated, and attribute information may be created based on the performance data.

[0059]

Since the present invention is configured as described above, according to the waveform generating method of the present invention, it is possible to generate a waveform by using an optimum partial waveform corresponding to the tone generation control information. It is possible to generate a musical tone with a rich expression. Further, according to the performance data processing method of the present invention, it is possible to select in advance the partial waveform data most suitable for the characteristics of each note in the performance data, and to add the partial waveform data to each note of the performance data. Therefore, by using the performance data, it is possible to select the optimal partial waveform data in the waveform generation in the automatic performance without increasing the load on the CPU 10 when performing the automatic performance. Further, according to the waveform generation method of the present invention, even when a musical tone is generated by a real-time performance, the musical tone generated by the current performance according to a performance to be performed in the future by using performance information corresponding to the real-time performance. Characteristics can be controlled. Furthermore, if one or a plurality of partial waveforms having attribute data close to the characteristic data of the sound to be played and the sound to be played before or the sound to be played next are selected, the sound to be played is An optimal partial waveform can be selected as a partial waveform for generating a musical tone.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a hardware configuration of a waveform data recording / reproducing apparatus including a waveform generation method and a performance data processing method according to the present invention.

FIG. 2 is a diagram showing a relationship between synchronized performance data and phrase waveform data in a waveform data recording / reproducing apparatus according to the present invention.

FIG. 3 is a diagram showing an aspect when recording waveform data in the waveform data recording / reproducing apparatus according to the present invention.

FIG. 4 is a diagram showing an outline of a configuration for performing a process of dividing phrase waveform data and a process of providing attribute information in the waveform data recording / reproducing apparatus according to the present invention.

FIG. 5 is a diagram showing a configuration of recorded data in the waveform data recording / reproducing apparatus according to the present invention.

FIG. 6 is a diagram showing a configuration of a performance timbre set in the waveform data recording / reproducing apparatus according to the present invention.

FIG. 7 is a diagram showing a detailed configuration of a partial waveform record in a performance timbre set in the waveform data recording / reproducing apparatus according to the present invention.

FIG. 8 is a diagram showing a configuration of a partial waveform ID of a partial waveform record in a performance timbre set in the waveform data recording / reproducing apparatus according to the present invention.

FIG. 9 is a flowchart showing a procedure for creating a timbre set in the waveform data recording / reproducing apparatus according to the present invention.

FIG. 10 is a flowchart of performance data processing performed by the waveform data recording / reproducing apparatus according to the present invention.

FIG. 11 is a flowchart showing the processing of step S12 in the flowchart of the performance data processing executed by the waveform data recording / reproducing apparatus according to the present invention.

FIG. 12 is a flowchart of a pattern comparison process in the flowchart shown in FIG. 11;

FIG. 13 is a flowchart of a note-on event process executed by the waveform data recording / reproducing apparatus according to the present invention.

FIG. 14 shows a step S in a note-on event process executed by the waveform data recording / reproducing apparatus according to the present invention.
20 is a flowchart showing a process 20.

FIG. 15 is a flowchart of an automatic performance start process executed by the waveform data recording / reproducing apparatus according to the present invention.

FIG. 16 is a flowchart of a tempo interruption process executed during automatic performance of the waveform data recording / reproducing apparatus according to the present invention.

FIG. 17 is a flowchart of note-on processing in a real-time performance performed in combination with an automatic performance of the waveform data recording / reproducing apparatus according to the present invention.

[Explanation of symbols]

1 waveform data recording / reproducing device, 10 CPU, 11 timer, 12 ROM, 13 RAM, 14 display, 1
5 controls, 16 sound sources, 17 mixers, 18 DACs,
19 sound system, 20 microphone, 21
ADC, 22 recording circuit, 23 HDD, 24 reproduction circuit, 25 MIDI interface, 26 other interfaces, 30 performance data, 31 automatic performance means,
32 player, 33 waveform recording means, 34 phrase waveform data, 35 clock synchronization line, 36 headphones, 37 speakers, 38 speaker sound removal processing means,
40 phrase waveform data, 41 division processing means, 42
Attribute information adding processing means, 43 partial waveform data, 44
Performance data

Claims (6)

[Claims] 1. A plurality of partial waveforms are stored in a partial waveform memory, and attribute information of each partial waveform stored in the partial waveform memory is stored in an attribute information memory. Referring to a memory, reading a partial waveform having attribute information corresponding to the sound control information from the partial waveform memory, and processing the read partial waveform based on the attribute information and the sound control information. A waveform corresponding to the sounding control information. 2. The attribute information of each partial waveform is stored in an attribute information memory, and a characteristic of each note included in performance data to be processed is compared with the attribute information stored in the attribute information memory. By this, the optimal partial waveform for the feature of the note concerned is detected,
A performance data processing method, wherein specified data for designating a detected partial waveform to the note is assigned to a note in the performance data, and the performance data to which the specified data is assigned is stored. 3. A waveform generating method for reproducing performance data provided with designated data stored by the performance data processing method according to claim 2, wherein a plurality of partial waveforms are stored in a partial waveform memory. When playing a note to which designated data is assigned based on the performance data, by reading a partial waveform from the partial waveform memory based on the designated data assigned to the note to be played back, A waveform generating method, wherein a waveform corresponding to a note is generated. 4. Performance information corresponding to a real-time performance is stored, an automatic performance or an automatic accompaniment is performed in accordance with the generated tempo clock to generate an accompaniment sound, and the performance information is controlled in accordance with the tempo clock. A waveform corresponding to the performance event is generated based on the performance event played in real time in accordance with the accompaniment sound and the played performance information. 5. A waveform selecting apparatus for selecting one or a plurality of partial waveforms from a plurality of partial waveforms stored in a partial waveform memory, wherein for each partial waveform, a sound corresponding to the partial waveform and a sound thereof A database storing attribute data indicating characteristics of two tones of the preceding sound, and a characteristic data of two sounds to be played and a preceding sound to be played are referred to by referring to the database. A search means for extracting one or a plurality of partial waveforms having attribute data close to characteristic data of two sounds as a partial waveform for generating the sound to be played, characterized by comprising: . 6. A waveform selecting apparatus for selecting one or more partial waveforms from a plurality of partial waveforms stored in a partial waveform memory, and for each partial waveform, a sound corresponding to the partial waveform and a sound corresponding to the sound A database that stores attribute data indicating the characteristics of the two tones of the next sound, and the characteristic data of the two tones to be played and the sound to be played before that are referred to the database. A search means for extracting one or a plurality of partial waveforms having attribute data close to characteristic data of two sounds as a partial waveform for generating the sound to be played, characterized by comprising: .