JP2007047293A - Musical sound generating device and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a musical sound generating device for generating musical sound using wave form data provided for each pitch, while suppressing an occupancy percentage of a semiconductor memory. <P>SOLUTION: The wave form data for indicating a time change of a wave peak value of the musical sound is stored in a ROM 13 for each pitch. The wave form data is divided into three parts of partial wave form data of a first wave form data, a middle wave form data and a last wave form data. A CPU 11 stores the first wave form data for each pitch in a RAM 12 beforehand and limits the other partial wave form data in data for generating the musical sound in which generation start is indicated by MIDI data input from an electronic musical instrument 20. Thereby, while suppressing the occupancy percentage of the wave form in the RAM 12, the musical sound is generated by using the wave form provided for each pitch. The sound is generated by outputting audio data (wave peak value) generated from the wave form data to a sound source system. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a musical sound generating apparatus that sequentially generates a peak value of a musical sound to be generated using waveform data representing a temporal change in a peak value of a musical sound and generates the musical sound.

  Musical sound is usually generated by generating a peak value of a musical sound at a predetermined time interval. The generated peak value is generated by D / A converting the generated peak value and converting an analog signal (audio signal) obtained thereby into sound.

  As a method for generating such a peak value, there is known a PCM sound source method that uses waveform data representing a temporal change in the peak value of a musical sound (Patent Documents 1 to 4). In the PCM sound source system, generally, waveform data obtained by sampling a musical sound (waveform) at a predetermined sampling period is used. For musical sounds that are artificially created or that can be created, the waveform data can be directly generated (Patent Document 1).

  Most musical sounds have inherent time changes during pronunciation. For this reason, as the waveform data, it is common to prepare data that can reproduce the time change during the tone generation. Such waveform data is composed of a lot of sampling data (crest value). As a result, even in the case of waveform data (Patent Document 3) assuming that the sound generation time is controlled by repeatedly reading out a part of sampling data, the actual amount of data is large.

  The waveform data usually changes depending on the pitch of the musical sound. However, the conventional musical tone generator employing the PCM tone generator method has prepared only a plurality of waveform data and assigned a different pitch range for each waveform data because of the size of the waveform data. As a result, the musical tones of each pitch within the pitch range are pronounced using the same waveform data. In this case, the pitch difference is realized by changing the speed (step width) of reading the sampling data from the waveform data.

It can be said that waveform data should be prepared for each pitch in order to be able to produce a higher-quality musical tone. Since the waveform data used for sound generation needs to be processed at high speed, it must be prepared in a semiconductor memory (storage device) having a high access speed. For this reason, if the waveform data prepared for each pitch is simply copied (developed) to the semiconductor memory, the ratio (exclusive ratio) of the waveform data in the semiconductor memory may increase. When the ratio increases, the portion that can be used in the semiconductor memory for other processing decreases, and thus a problem that the processing cannot be performed at high speed is likely to occur. For this reason, a large-capacity semiconductor memory is often mounted. However, when waveform data is prepared for each pitch, it is considered important to suppress the occupation ratio of the semiconductor memory.
Japanese Patent No. 2571559 Japanese Patent No. 2558519 JP 58-178395 A Japanese Patent Laid-Open No. 11-296174

  An object of the present invention is to provide a musical sound generating apparatus capable of generating musical sounds using waveform data prepared for each pitch while suppressing the occupation rate of semiconductor memory.

  The musical tone generator according to the present invention can acquire waveform data on the premise that the musical tone is generated by sequentially generating the peak values of the musical tone to be generated using the waveform data representing the temporal change of the peak value of the musical tone. When using waveform data different for each pitch of the tone for generating the peak value of the tone to be sounded, the data acquisition unit, the storage unit for storing the waveform data acquired by the data acquisition unit, and dividing the waveform data The first partial waveform data located at the head of the plurality of partial waveform data obtained in this manner is acquired by the data acquisition means for each waveform data and stored in advance in the storage means, and the first partial waveform data is continued. The second partial waveform data, which is partial waveform data, includes data management means for acquiring the data acquisition means and storing it in the storage means as needed.

  It is desirable that the data management means causes the data acquisition means to acquire the second partial waveform data necessary for generating the peak value of the musical tone to be generated and store it in the storage means as needed. Further, it is desirable that the second partial waveform data of the waveform data corresponding to the pitch in the vicinity of the pitch is acquired in advance by the data acquisition unit and stored in the storage unit based on the pitch of the musical sound to be generated. Further, when a plurality of waveform data different for each pitch is prepared in accordance with a velocity value indicating the intensity of generating a musical sound, the data management means obtains the first partial waveform data for each waveform data. The second partial waveform data is acquired by the data acquisition unit as needed according to the velocity value specified for the tone to be generated and stored in the storage unit. Is desirable.

  The program of the present invention is based on the premise that a musical tone generator for generating a musical tone by sequentially generating a peak value of a musical tone to be generated using waveform data representing a temporal change in a peak value of a musical tone is executed. A function that can acquire data, a function that can store waveform data acquired by the function that can be acquired in a storage device, and a waveform data that is different for each pitch of the musical sound to generate a peak value of the musical sound to be generated. The first partial waveform data located at the head of the plurality of partial waveform data obtained by dividing the data is acquired by a function that can be acquired for each waveform data and stored in a storage device by a function that can be stored in advance, Second partial waveform data, which is partial waveform data following the first partial waveform data, can be acquired by the function that can be acquired and stored in the storage device at any time. A function of storing the ability to realize.

  The present invention provides a plurality of partial waveforms obtained by dividing the waveform data when generating the peak value of the tone to be generated using the waveform data representing the temporal change of the peak value of the tone prepared for each pitch. First partial waveform data positioned at the head of the data is stored in a storage device such as a semiconductor memory for each waveform data (pitch), and a second portion which is partial waveform data following the first partial waveform data The waveform data is dynamically stored in the storage device at any time according to the situation. Therefore, compared with the case where all the waveform data is stored in the storage device, the occupation ratio of the waveform data in the storage device can be kept low. Thereby, the storage device can be shared in a more appropriate form for executing many processes. Since the musical sound is generated using the waveform data prepared for each pitch, a higher-quality musical sound can be generated.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
<First Embodiment>
FIG. 1 is a diagram for explaining the configuration of a musical sound generator according to the first embodiment.

  As shown in FIG. 1, the musical sound generator 10 stores a CPU 11 that controls the entire apparatus 10, a RAM 12 that the CPU 11 uses for work, a program executed by the CPU 11, various control data, waveform data, and the like. ROM 103, for example, a keyboard or pointing device (such as a mouse), a medium driving device for accessing a recording medium such as a CD-ROM or DVD, an input unit 14 including those interfaces, and a display for displaying an image on the display device A MIDI interface (I / F) 16 for transmitting and receiving MIDI data between the unit 15 and an external device, and an audio interface (I / F) for outputting digital data (peak value) of a musical tone to be generated to the external device F) 17.

  Connected to the MIDI I / F 16 is an electronic musical instrument 20 having a keyboard 21 and a MIDI interface (I / F) 22 for generating and outputting MIDI data indicating the operation contents for the keyboard 21, and the audio I / F 17. Is connected to a sound source system 30 that can input a peak value and convert it into sound. Thereby, the musical sound generating apparatus 10 according to the present embodiment inputs MIDI data from the MIDI I / F 22 of the electronic musical instrument 20, generates a peak value of a musical sound whose pronunciation is instructed by the MIDI data, and uses it as a sound source. By outputting to the system 30, the musical sound is generated. The sound source system 30 includes at least a D / A converter, an amplifier, and a speaker, for example.

  The musical sound generator 10 is realized by, for example, mounting an application program (hereinafter referred to as “application”) for operating as the musical sound generator 10 on a personal computer (PC). For example, the application is recorded on a recording medium accessible by the medium driving device constituting the input unit 14 and provided. The application may be distributed via a network such as the Internet.

  A writable memory (for example, a flash memory) is employed as the ROM 13 so that an application can be installed (installed). On the recording medium on which the application is recorded, a large number of various waveform data for generating the peak value of the musical sound are recorded. Some of the waveform data are prepared for each pitch. The PC that implements the musical sound generation device 10 may be one in which an auxiliary storage device such as a hard disk device is mounted in addition to or instead of the ROM 13. The application is hereinafter referred to as “musical sound generation application”.

  Waveform data prepared for each pitch can be stored in the ROM 13. Further, it is possible to copy from a recording medium as required. Here, for the sake of explanation, it is assumed that the data is stored in the ROM 13 for convenience. Similarly, only waveform data prepared for each pitch is assumed as waveform data used for the sound generation.

  The waveform data stored in the ROM 13 is copied to the RAM 12 and used. This is because the RAM 12 generally has a higher access speed than the ROM 13. As a result, it is possible to generate a peak value of a musical tone to be generated at a higher speed.

  By preparing waveform data for each pitch, sampling data (crest value) constituting the waveform data is simply read out sequentially. Thus, it is not necessary to perform interpolation according to the speed of reading data (step width). Since the necessity of performing the interpolation operation is avoided, the peak value of the musical sound to be generated can be generated more easily.

  In the present embodiment, each waveform data is divided into three, that is, the waveform data of the portion used at the start of sound generation (first partial waveform data; hereinafter referred to as “first waveform data”), and the portion used for subsequent sound generation Waveform data (hereinafter referred to as “intermediate waveform data”), and further divided into waveform data used for subsequent sound generation (hereinafter referred to as “final waveform data”). Equivalent to partial pronunciation data). The intermediate waveform data and the final waveform data are prepared so that sound generation by one of them is performed after sound generation by one of them is completed. As a result, after the end of the sound generation by the first waveform data, until the end of the tone sound itself, the parts for sound generation are alternately shown as intermediate waveform data → final waveform data, final waveform data → intermediate waveform data. Waveform data is switched.

  In the present embodiment, the top waveform data is loaded (copied) in advance into the RAM 12 for each waveform data (pitch). After that, the intermediate waveform data corresponding to the pitch of the musical tone instructed to start sounding by the MIDI data input by the user's performance operation on the electronic musical instrument 20 is loaded. Thereafter, the final waveform data and the intermediate waveform data are alternately loaded until the end of the tone generation, and one of them is erased when it becomes unnecessary. For this reason, partial waveform data other than the top waveform data stored in the RAM 12 is muted by inputting MIDI data for instructing the start of tone generation, the time elapsed from the start of tone generation, inputting MIDI data for instructing the muting, and the like. It will change depending on the situation.

  In this way, only the necessary partial waveform data other than the top waveform data is stored in the RAM 12 as long as necessary. For this reason, the ratio (exclusive ratio) of the waveform data stored in the RAM 12 can be significantly reduced as compared with the case where all the partial waveform data is stored for each waveform data. As a result, a decrease in speed in processing that is not related to tone generation (processing by executing a program other than the tone generation application) can be further avoided or suppressed. Since the musical sound is generated using the waveform data prepared for each pitch, it is possible to generate a higher quality musical sound.

  In this embodiment, various data shown in FIGS. 2 to 5 are managed in order to realize the storage of the partial waveform data other than the head waveform data as described above in the RAM 12. The data will be specifically described with reference to FIGS.

  FIG. 2 is a diagram for explaining the configuration of timbre data. The timbre data is data prepared for each timbre in order to manage the waveform data in timbre units. As shown in FIG. 2, the timbre data includes data iID indicating the corresponding channel number, data iNoteOnCnt indicating the number of tones being generated, and the storage location of the pitch data of the first pitch (for example, the lowest pitch) among the corresponding pitch data. Data pTD, data pPrev indicating the storage location of the previous tone color data, and data pNext indicating the storage location of the next tone color data.

  FIG. 3 is a diagram for explaining the configuration of pitch data. The pitch data is data prepared for each tone color and for each pitch. As the data pTD, data indicating the storage location of the pitch data of the leading pitch in the corresponding tone color is stored in each tone color data. As shown in FIG. 3, the pitch data includes data iPitch (pitch number: note number) indicating the corresponding pitch, and data iStatus indicating the state of the musical tone of the pitch (0 is muted, 1 is sounding) 2 represents data being silenced), data pSD indicating the storage location of the corresponding source data, data pPrev indicating the storage location of the previous pitch data, and data pNext indicating the storage location of the next pitch data It is equipped with.

  FIG. 4 is a diagram for explaining the configuration of source log data. The source data is data prepared for each waveform data. As shown in FIG. 4, the source data includes data pDAta [0] indicating the start address of the start waveform data copied to the RAM 12 of the corresponding waveform data, and data pDAta indicating the start address of the intermediate waveform data copied to the RAM 12. [1], data pDAta [2] indicating the start address of the final waveform data copied to the RAM 12, data fileName indicating the file name (original storage location) of the waveform data, and data fLength indicating the length of the entire waveform data , Data iIndex indicating partial waveform data being referred to for generating a crest value (indicated as “buffer” in the figure), data minVel indicating the lower limit value of the velocity range corresponding to the waveform data, data maxVel indicating the upper limit value, PPre indicating the storage location of the pitch data of , And a data pNext, indicating the storage location of the next pitch data.

  FIG. 5 is a diagram for explaining the structure of the pronunciation data. The sounding data is data prepared for each waveform data being sounded. As shown in FIG. 5, the sound generation data includes data iID as identification information, data iTone indicating the corresponding timbre data (channel) (for example, data iID in the corresponding timbre data), data iPitch indicating the pitch of the musical tone, Data iOnVel indicating the velocity, data iStatus indicating the state of the musical tone of the pitch (value of 0 indicates mute, value of 1 is sounding, value of 2 indicates mute processing), storage of corresponding source data The data pSD indicating the location, the data lOnStart indicating the sound generation start time, the data lOffStart indicating the mute start time, the data fRelease indicating the attenuation rate during the mute processing, the data pPrev indicating the storage location of the previous sound generation data, and the next Data pNext indicating the storage location of the pitch data is provided. The data iOnVel is a velocity value extracted from the MIDI data input from the electronic musical instrument 20. As is well known, the velocity value represents the speed at which a key is pressed, that is, the strength with which a musical sound is produced.

  Data other than the sound generation data, that is, timbre data, pitch data, and source data, for example, is data that is not deleted during execution of the musical tone generating application. For this reason, loading (copying) of partial waveform data other than the top waveform data to the RAM 12 is performed as necessary, while generating or erasing the sound generation data as required, and various data including existing sound generation data. It is designed to be done by updating. Hereinafter, with reference to various flowcharts shown in FIGS. 6 to 10, the operation of the musical tone generator 10 that stores necessary partial waveform data in the RAM 12 while managing those data will be described in detail.

  FIG. 6 is a flowchart of the entire process. This shows the flow of excerpts of the processing executed from when the musical tone generating application is started to when it is terminated. Accordingly, the CPU 11 is realized by reading the musical tone generating application from the ROM 13 to the RAM 12 and executing it. First, the entire process will be described in detail with reference to FIG.

  First, in step 101, a driver for the MIDI I / F 16 is loaded from the ROM 13 to the RAM 12 and activated. In the subsequent step 102, a driver for the audio I / F 17 is similarly loaded and activated. Thereafter, an audio data output thread for outputting the peak value (audio data) of the tone to be generated from the audio I / F 17 is activated (step 103), and stored in the ROM 13 as a tone definition file. 4 is read and expanded in the RAM 12 (step 104), and for each waveform data, the top waveform data (indicated as “tone data” in the figure) is read from the ROM 13 and loaded into the RAM 12 (step 105). The initialization is completed by the load, and the process proceeds to step 106 to wait for data input by the input unit 14 or the MIDI I / F 16. Here, for the data input by operating the input unit 14, only an instruction to end the musical sound generating application is assumed.

  When a user presses or releases a key constituting the keyboard 21 of the electronic musical instrument 20, the MIDI I / F 22 generates the contents of the performance operation and MIDI data corresponding to the key on which the operation has been performed. Then, it is output to the MIDI I / F 16 of the musical tone generator 10. When a keyboard or a pointing device constituting the input unit 14 is operated, the operation content is notified from the input unit 14 to the CPU 11. Thereby, the process proceeds from step 106 to step 107, and it is determined whether or not the input is an end command of the musical tone generating application. If the user performs an operation for terminating the application on the input unit 14, the determination is yes and the process proceeds to step 110. Otherwise, the determination is no and the process moves to step 108.

  In step 108, it is determined whether or not the input is MIDI data. If the MIDI I / F 16 has input MIDI data, the determination is YES, the process proceeds to step 109, the MIDI IN process corresponding to the input is executed, and then the process returns to step 106. On the other hand, if not, the determination is no and the process returns to step 106.

  In step 110 where the determination in step 107 is YES and the process proceeds, the audio data output thread is terminated. The termination is performed, for example, by substituting a value for instructing termination into the thread termination management variable. Thereafter, various data shown in FIGS. 2 to 5 stored in the RAM 12 are deleted (step 111), the driver of the audio I / F 17 is terminated (released) (step 112), and the driver of the MIDI I / F 16 is turned on. End (open) (step 113). After erasing various drivers and data from the RAM 12 in this manner, a series of processing is terminated, that is, the musical tone generating application is terminated.

  FIG. 7 is a flowchart of the MIDI IN process executed as step 109 described above. Next, the IN process will be described in detail with reference to the flowchart shown in FIG.

  This IN process is a process for dealing with MIDI data input via the MIDI I / F 17. There are many types of MIDI data, but here, only data related to the present invention, that is, only two types of MIDI data of note-on message and note-off message will be described.

  First, in step 201, it is determined whether or not the input MIDI data is a note-on message instructing the start of tone generation. If the MIDI data is not a note-on message, the determination is yes and the process moves to step 206. Otherwise, the determination is yes and the process moves to step 202.

  In step 202, a channel number, a note number (pitch), and a velocity value are extracted from the input MIDI data, and substituted into variables ich, iPit, and iVel, respectively. In the next step 203, the pitch data designated by the values of the variables ich and iPit, that is, the channel that is newly instructed to start the tone generation, the corresponding pitch data is identified from the pitch, and the index value is set in the variable td. Substitute and update the pitch data specified by the value of the variable td. Specifically, the value of the data iStatus is changed to 1. Data iNoteOnCnt in the corresponding timbre data is updated by incrementing the value. Thereafter, the process proceeds to step 204.

  In step 204, a location (position) where the pronunciation data is not stored is specified from the area secured in the RAM 12 for storing the pronunciation data, and an index value indicating the location is substituted into the variable nd. In the next step 205, the pronunciation data is stored in a location specified by the value of the variable nd. Specifically, the value of variable ich is represented as data iTone (indicated as “nd.iTone” in the figure), the value of variable iPit as data iPitch (indicated as “nd.iPitch” in the figure), and data iOnVel (“nd. .IOnVel "), the value of the variable iVel, the data iStatus (indicated as" nd.iStatus "in the figure), 1 and the pitch data data pSD (indicated as" nd.pSD "in the figure) A value indicating the current time is stored as “td.pSD” in the figure) and data lOnStart (indicated as “nd.lOnStart” in the figure). Thereafter, the series of processing is terminated.

  In this way, by inputting the MIDI data of the note-on message, the tone color data and pitch data specified from the MIDI data are updated, and new sound generation data is generated.

  In step 206 where the determination in step 201 is NO and the process proceeds to step 206, it is determined whether or not the input MIDI data is a note-off message instructing the end of tone generation. If the MIDI data is a note-off message, the determination is yes and the process moves to step 207. Otherwise, the determination is no and the series of processing ends here.

  In step 207, a channel number and a note number (pitch) are extracted from the input MIDI data, and are substituted into variables ich and iPit, respectively. In the next step 208, sounding data specified by the values of the variables ich and iPit (phonetic data storing these values as data iTone and iPitch) is specified, and the index value is substituted into the variable td. Further, the pitch data specified by each value of the variables ich and iPit is specified, the index value is substituted into the variable td, and the value of the pitch data data iStatus specified by the value of the variable td is changed to 2. . After the change, the process proceeds to step 209, where the data iStatus in the pronunciation data specified by the value of the variable nd is 2, the data fRelease (indicated as “nd.fRelease” in the figure) 1.0, the data lOffStart ( A value indicating the current time is stored as “nd.lOffStart” in the figure. In this way, after updating the pitch data and the sound generation data specified by the MIDI data of the note-off message, the series of processes is terminated.

  FIG. 8 is a flowchart of processing realized by executing the audio data output thread activated in step 103 in the overall processing shown in FIG. Next, the processing will be described in detail with reference to FIG. This output thread is used to generate a peak value (audio data) of a musical tone to be generated and output via the audio I / F 17 at a sampling period.

  First, in step 301, initialization is performed such that initial values are substituted into various variables, and an area (output buffer) for temporarily storing audio data to be output via the audio I / F 17 is secured in the RAM 12. In the next step 302, it is determined whether or not an end is instructed. As described above, in step 110 of the overall process shown in FIG. 6, a value for instructing termination is substituted into the termination management variable of the audio data output thread. For this reason, if the value is assigned to the variable, the determination is YES, and the series of processing ends here. Otherwise, the determination is no and the process moves to step 303.

  In step 303, output data creation processing for creating (generating) the audio data of the musical sound to be generated and storing it in the output buffer is executed. In the next step 304, a process for sequentially outputting the audio data stored in the output buffer at the sampling period is executed. After the execution, the process returns to step 302. Thereby, while the musical tone generating application is activated, the processing loop formed in steps 302 to 304 is repeatedly executed.

  Next, the output data creation process executed as step 303 will be described in detail with reference to the flowchart shown in FIG. In the creation process, audio data for a plurality of sampling periods is created at a time and stored in an output buffer. “LSamples” in the figure is a constant indicating the number of audio data created at one time.

  First, in step 401, 0 is substituted for variable i. In the subsequent step 402, it is determined whether or not the value of the variable i is less than a constant lSamples. The value of the variable i is incremented every time audio data for one sampling period is created. For this reason, when all the audio data to be created at one time has not been created, the value of the variable i satisfies the relationship of less than the constant lSamples, so the determination is YES, and the routine goes to Step 402. If not, that is, if all the audio data to be created has been created, the determination is no, and the series of processing ends here.

  In step 403, a value (= lTime + i / lSR) obtained by adding the value obtained by dividing the value of the variable i by the constant lSR to the value of the variable lTime is substituted into the variable lSTime. The value assigned to the variable lTime is a value indicating the time when the next audio data is to be output, and the constant lSR is a value indicating the sampling rate of the audio data stored in the output buffer. Accordingly, the value obtained by dividing the value of the variable i by the constant lSR is a value (value indicating time) obtained by multiplying the value of the variable i by the sampling period. For this reason, a value indicating the time at which target audio data is to be output is substituted into the variable lSTTime. After the substitution, the process proceeds to step 404.

  In step 404, 0 is substituted into a variable iValue for creating audio data for one sampling period. In the next step 405, an index value for designating the storage location of the pronunciation data located at the head of the pronunciation data storage area is substituted for the variable nd. In the next step 406, it is determined whether or not pronunciation data exists in the storage location specified by the value of the variable nd. If the sound data does not exist, the determination is no and the process moves to step 416. Otherwise, the determination is yes and the process moves to step 407.

  In step 407, it is determined whether or not the value of the data iStatus at the storage location specified by the value of the variable nd is not 0 indicating that the sound is muted. If the value is 0, the determination is no and the process moves to step 410. Otherwise, the determination is yes and the process moves to step 408.

  A determination of YES in step 407 means that the musical sound corresponding to the pronunciation data currently being watched should be pronounced. Therefore, in step 409, sampling data reading processing for reading sampling data for generating the musical sound from the RAM 12 is executed. As a result, the sampling data is substituted into the variable iVtmp.

  In step 409 following step 408, it is determined whether or not the value of the data iStatus is 2. If the musical sound corresponding to the current pronunciation data is currently being muted, the determination is yes and the process moves to step 412. Otherwise, the determination is no, and a value obtained by adding the value of the variable iVtmp to the previous value (= iValue + iVtmp) is substituted for the variable iValue (step 410), and the storage location of the next pronunciation data is stored in the variable nd. Is substituted (step 411). After the substitution, the process returns to step 406.

  In this way, for musical sounds that are not being muted, the sampling data assigned to the variable iVtmp by executing the sampling data reading process in step 408 is added to the value of the variable iValue so far. Thus, the sound is generated according to the sampling data (waveform data) until the mute process is started.

  In steps 412 to 415, where the determination of step 409 is YES and the process proceeds, processing for gradually decreasing the volume of the musical sound to be generated is performed until it can be considered that the sound has been muted. Since the value of the sampling data constituting the waveform data changes greatly, whether or not it can be regarded as silenced is determined based on the value of the data fRelease that is updated during the silencing process. “RED_OFF” and “RELEASE” in the figure are a constant set for the determination and a constant set for the data fRelease update, respectively.

  First, in step 412, a value (= iVtmp * fRelease) obtained by multiplying the value of the data fRelease in the pronunciation data of interest by the previous value is substituted for the variable iVtmp. In the following step 413, it is determined whether or not the value of the data fRelease is less than a constant RED_OFF. If the value is equal to or greater than the constant RED_OFF, the determination is no, the data fRelease is updated to a value obtained by multiplying the previous value by the constant RELEASE in step 414, and then the process proceeds to step 410. Otherwise, the determination is yes and the process proceeds to step 415 where the value of the data iStatus is set to 0 and the other sound data is deleted by performing an operation of resetting other data. The partial waveform data loaded in the RAM 12 is deleted (released) for updating various data and for sound generation. As a result of the release, the partial waveform data for tone generation of the musical tone from which the tone generation data has been deleted is only the top waveform data. Various data other than the pronunciation data are updated by decrementing the value of the data iNoteOnCnt for the timbre data, updating the value of the data iStatus for the pitch data, and changing the data iIndex to a value indicating the first waveform data for the source data. Update each. After performing such an update, the process proceeds to step 410 described above.

  In step 415, along with the deletion of the sound data, the sound data having an index value indicating the storage location of the sound data as the data pPrev or pNext is updated with the data pPrev or pNext. If the phonetic data to be erased is located at the end, that is, if the index value indicating the storage location is not the maximum, another phonetic data is moved to the storage location where the phonetic data has been deleted, and the phonetic data continues from the beginning. Let them do. As a result, if there is no sound data in the storage location specified by the value of the variable nd and the determination in step 406 is NO, no other sound data to be noted is present. The absence of other noteworthy pronunciation data means that the creation of audio data for one sampling period has been completed.

  In step 416 where the determination in step 406 is NO and the process proceeds to step 416, the value of the variable iValue is set as audio data for one sampling period, and stored in the storage location specified by the value of the variable i in the output buffer together with the value of the variable lTime. Store. After the storage, the value of the variable i is incremented in step 417, and the process returns to step 402.

  FIG. 10 is a flowchart of the sampling data reading process executed as step 408 described above. Finally, the data reading process will be described in detail with reference to FIG.

  As described above, since the waveform data is prepared for each pitch, the sampling data extracted (read) from the waveform data can be basically used as it is for sound generation. Thereby, in the sampling data reading process, the position (address) of the sampling data from which the sampling data in the target partial pronunciation data is to be read is specified, and the sampling data at that position is read. Further, as necessary, the partial waveform data to be referred to next is loaded into the RAM 12 and the loaded partial waveform data is erased (released). Here, for convenience of explanation, it is assumed that all partial waveform data have the same size (the same number of sampling data). “SIZE” shown in the figure is a constant indicating the size.

  First, in step 501, a value obtained by subtracting the value of the data lOnStart in the sound generation data stored in the storage location specified by the value of the variable nd from the value of the variable lStim, that is, the sound of the musical sound being sounded by the sound data. A value indicating the time elapsed since the start of the process is substituted into a variable lT. In the next step 502, a remainder (= (lT * 1SR)% SIZE) obtained by dividing the multiplication result obtained by multiplying the value of the variable lT by the constant lSR by the constant SIZE is substituted for the variable lPos. Thereafter, the data pSD in the pronunciation data of interest is substituted into the variable sd in step 503, and the process proceeds to step 504.

  In step 504, it is determined whether or not the value of the variable lPos is 0, that is, whether or not the reading of the sampling data from the partial waveform data referred to so far has been completed. If the reading is not completed, the determination is no and the process moves to step 509. Otherwise, the determination is yes and the process moves to step 505.

  In steps 505 to 508, processing for switching the partial waveform data to be read next to another data is performed when the sampling data is completely read.

  First, in step 505, it is determined whether or not the value of the data iIndex in the source data specified by the value of the variable sd is not 0, that is, whether or not the partial waveform data referred to so far is not the top waveform data. If the partial waveform data is the top waveform data, the determination is no, the value of the data iIndex is updated to 1 in step 506, and the process proceeds to step 509. Otherwise, the determination is yes and the process moves to step 507.

  In step 507, the partial waveform data indicated by the value of the data iIndex is released (erased) from the RAM 12. In the next step 508, the value of the data iIndex is updated to a value obtained by adding 1 to the value obtained by dividing the previous value by 2 (= (sd.iIndex)% 2 + 1). The process proceeds to step 509 after the update.

  The value of the data iIndex is 0, 1, or 2. When the process proceeds to step 507, the value is 1 or 2. Therefore, if the value is 1, it is updated to 2, and if the value is 2, it is updated to 1. Thus, after the head waveform data is referred to, the intermediate waveform data and the final waveform data are alternately referred to until the musical sound is muted.

  In step 509, partial waveform data corresponding to the value of the data iIndex is set as a reference target. In the next step 510, sampling data designated by the value of the variable lPos in the target partial waveform data (data of the value of the variable lPos from the head) is read and substituted into the variable iVtmp. In the next step 511, in order to reflect the value of the variable iVtmp, that is, the value of the data iOnVel (velocity value) in the pronunciation data in the read sampling data, the value of the data iOnVel is associated with the value of the sampling data. A value obtained by multiplying the value divided by the value of the data maxVel in the source data to be substituted is substituted into the variable iVtmp. As a result, the sampling data assigned to the variable iVtmp is updated in accordance with the ratio between the velocity value when the user actually presses the key and the upper limit value of the velocity range of the waveform data, and the process proceeds to step 512.

  In step 512, it is determined whether or not the value of the variable lPos is equal to or greater than a value obtained by multiplying the constant SIZE by 0.8. If the magnitude relationship is satisfied, the determination is yes, the process proceeds to step 513, and partial waveform data corresponding to a value obtained by adding 1 to the value obtained by dividing the value of the data iIndex by 2 is read from the ROM 13. After loading into the RAM 12, the series of processing ends. Otherwise, the determination is no and the series of processing ends here.

  The partial waveform data loaded in step 513 is intermediate waveform data when the value of the data iIndex is 0 or 2, and is final waveform data when the value is 1. When such partial waveform data is in a situation where 80% or more of the target partial waveform data is currently sampled, the sampling data to be read in the partial waveform data is less than 20% in advance. Are loaded into the RAM 12. Thereby, the partial waveform data to be referred to next is surely prepared in the RAM 12 at the timing of switching the partial waveform data to be referred to.

In the present embodiment, the musical sound to be generated is designated by the user operating the electronic musical instrument 20, but the designation may be automatically made by reproduction data such as a standard MIDI file. The waveform data is divided into three partial waveform data, but the number of divisions is at least better than that. The sampling data may be subjected to compression processing.
<Second Embodiment>
In the first embodiment, partial waveform data other than the top waveform data is loaded into the RAM 12 only with partial waveform data used for creating audio data of musical sounds to be generated. On the other hand, in the second embodiment, partial waveform data used for sound generation of a musical sound having a pitch close to that of the musical sound to be generated is loaded in the RAM 12 in advance. This is because, in music, it is relatively common that after a musical tone having a certain pitch is generated, a musical tone having a pitch close to that is generated. In the second embodiment, paying attention to that, and by speculatively loading partial waveform data that is considered to be likely to be used next, the musical sound to be generated can always be sounded smoothly and reliably. I'm letting Since the partial waveform data to be speculatively loaded is limited to the musical sound to be generated and the musical sound having a similar pitch, the increase in the occupation rate of the RAM 12 due to the waveform data can be suppressed to a slight level.

  The configuration of the musical sound generating device in the second embodiment is basically the same as that in the first embodiment. The operation is largely the same or basically the same. For this reason, only the parts different from the first embodiment will be described using the reference numerals given in the first embodiment as they are.

  In the second embodiment, the MIDI IN process executed as step 109 in the overall process shown in FIG. 6 is different from that of the first embodiment. Therefore, the different parts will be described in detail with reference to the flowcharts shown in FIG. 11 and FIG.

  FIG. 11 is a flowchart of the MIDI IN process in the second embodiment. As shown in FIG. 11, in the second embodiment, after the processing of step 205 is executed, the processing shifts to step 601 and the peripheral pitch whose pitch is close to the musical sound indicated by the sound generation data generated and stored in step 205. The peripheral waveform data reading process for reading the partial waveform data for tone generation from the ROM 13 and loading it into the RAM 12 is executed. A series of processing ends after the execution.

  FIG. 12 is a flowchart of the peripheral pitch data reading process executed as step 601. Next, the data reading process will be described in detail with reference to FIG.

  A performance operation on the keyboard 21 may be performed using both hands. For this reason, in this embodiment, the keyboard 21 is divided into a high frequency range that is assumed to be performed with the right hand and a portion other than the high frequency range is divided into low frequencies that are assumed to be performed with the left hand. In addition, speculative loading of partial waveform data is performed. “DIVPIT” shown in the figure is a constant set as the pitch (note number) of the key that becomes the boundary of these sound ranges. “RNGHIGH” and “RNGLOW” are constants set in the high sound range and the low sound range as the peripheral pitch range in which the partial waveform data is speculatively loaded.

  First, in step 701, the constant RNHIGH is substituted for the variable iRng. In the next step 702, intermediate waveform data is read from the ROM 13 and loaded into the RAM 12 as partial waveform data for musical tone generation indicated by the sound generation data newly stored in the storage location specified by the value of the variable nd. Thereby, the intermediate waveform data for tone generation instructed to start the sound generation by the input MIDI data is loaded at an earlier stage than in the first embodiment. After the loading, the process proceeds to step 703.

  In step 703, whether or not the value of the data iPit indicating the pitch (note number) extracted from the MIDI data is less than the constant DVPIT, that is, whether the musical sound instructing the start of sounding of the MIDI data belongs to the low pitch range. Judge whether or not. If the musical sound has a pitch belonging to the low frequency range, the determination is yes, and after the constant RNGLOW is substituted for the variable iRng in step 704, the process proceeds to step 705. Otherwise, the determination is no, and the process of step 0705 is then executed.

  In step 705 and subsequent steps, intermediate waveform data for tone generation in which the input MIDI data is higher and lower by the value of the variable i than the tone instructing the start of sound generation is necessary while the value of the variable i is sequentially incremented. In response to this, processing for loading into the RAM 12 is performed. As a result, in the present embodiment, among the target musical sounds (musical sounds of peripheral pitches), the high and low-pitched musical sounds are separated from the musical sound instructed to start sounding by the value of the variable iRng. Loads intermediate waveform data for musical sound generation that is not sounding.

  First, in step 705, 0 is substituted for variable i. In the subsequent step 706, the data pNext and pPrev in the pitch data corresponding to the newly stored pronunciation data are substituted for the variables th and tl, respectively. In the next step 707, it is determined whether or not the value of the variable i is less than the value of the variable iRng. If the value of the variable i is equal to or greater than the value of the variable iRng, the determination is NO, and all the intermediate waveform data to be speculatively loaded has been loaded, and the series of processing ends here. Otherwise, the determination is yes and the process moves to step 708.

  In step 708, it is determined whether the value of the data iStatus in the pitch data designated by the value of the variable th is 0, that is, whether the musical sound specified by the pitch data is muted. If the tone is being produced, the determination is no and the process moves to step 710. Otherwise, the determination is yes, and in step 709, the intermediate waveform data for tone generation is read from the ROM 13 and loaded into the RAM 12, and then the process proceeds to step 710.

  Similarly, in step 710, it is determined whether or not the value of data iStatus in the pitch data designated by the value of variable tl is zero. If the musical sound specified by the pitch data is being sounded, the determination is no and the process moves to step 712. Otherwise, the determination is yes, and in step 711, the intermediate waveform data for tone generation is read from the ROM 13 and loaded into the RAM 12, and then the process proceeds to step 712.

  In step 712, the data pNect and pPrev in the pitch data designated by the values so far are substituted for the variables th and tl, respectively. As a result, the pitch data of interest is changed to pitch data that is larger by 1 in the note number on the high pitch side, and is changed to pitch data that is smaller in the note number by 1 on the lower pitch side. In the next step 713, the value of the variable i is incremented. Thereafter, the process returns to step 707.

  In this way, the intermediate waveform data for generating a musical sound that should not be generated is loaded in advance, that is, at a stage before it should be generated. Therefore, although detailed description is omitted, partial waveform data corresponding to a value obtained by adding 1 to the value obtained by dividing the value of the data iIndex by 2 in step 512 in the sampling data reading process shown in FIG. Whether it has already been loaded or not is determined together. As a result, when the value of the variable lPos is equal to or greater than the value obtained by multiplying the constant SIZE by 0.8 and the partial waveform data to be loaded next is not loaded, the determination is made to be YES. . In this way, unnecessary loading of partial waveform data is avoided.

  In the present embodiment (first and second embodiments), one type of waveform data (consisting of three partial waveform data) is prepared for each pitch (here, each tone color). Since the waveform data differs depending on the velocity, a plurality of waveform data having different target velocity ranges for each pitch may be prepared. When multiple waveform data with different velocity ranges are prepared in this way, for example, the first waveform data with different velocity ranges is loaded in advance, and the partial waveform data is the velocity value of the tone to be generated. You can load the one that meets your needs as needed. By doing so, it becomes possible to produce a musical tone using waveform data corresponding to the velocity value for each pitch.

It is a figure explaining the structure of the musical tone generator by 1st Embodiment. It is a figure explaining the structure of timbre data. It is a figure explaining the structure of pitch data. It is a figure explaining the structure of source data. It is a figure explaining the structure of pronunciation data. It is a flowchart of the whole process. It is a flowchart of a MIDI IN process. It is a flowchart of the process implement | achieved by execution of an audio data output thread | sled. It is a flowchart of an output data creation process. It is a flowchart of a sampling data reading process. It is a flowchart of a MIDI IN process (2nd Embodiment). It is a flowchart of a peripheral pitch data reading process.

Explanation of symbols

10 Musical sound generator 11 CPU
12 RAM
13 ROM
14 Input unit 15 Display unit 16, 22 MIDI interface (I / F)
17 Audio increment (I / F)
20 Electronic musical instrument 21 Keyboard 30 Sound source system

Claims (5)

In a musical sound generator for generating the musical sound by sequentially generating the peak value of the musical sound to be generated using the waveform data representing the temporal change of the peak value of the musical sound,
Data acquisition means capable of acquiring the waveform data;
Storage means for storing the waveform data acquired by the data acquisition means;
In the case where waveform data different for each pitch of the musical sound is used to generate the peak value of the musical sound to be generated, the first portion located at the head of the plurality of partial waveform data obtained by dividing the waveform data Waveform data is acquired by the data acquisition means for each waveform data and stored in the storage means in advance, and second partial waveform data that is partial waveform data following the first partial waveform data is obtained at any time. Data management means to be acquired by the means and stored in the storage means;
A musical sound generating device comprising: The data management means causes the data acquisition means to acquire the second partial waveform data necessary for generating the peak value of the musical tone to be sounded, and store it in the storage means.
The musical tone generator according to claim 1. The data management means causes the data acquisition means to acquire in advance the second partial waveform data of the waveform data corresponding to the pitch in the vicinity of the pitch based on the pitch of the musical sound to be generated and store it in the storage means. Let
The musical sound generator according to claim 2, wherein In the case where a plurality of waveform data different for each pitch is prepared in accordance with a velocity value indicating the strength of sounding the musical sound, the data management unit is configured to store the first partial waveform data for each waveform data. The second partial waveform data is acquired by the data acquisition means at any time according to the velocity value designated for the musical tone to be generated, and is acquired by the data acquisition means and stored in advance in the storage means. Storing in the storage means;
The musical tone generator according to claim 1. A program for causing a musical sound generator to sequentially generate a peak value of a musical sound to be generated using waveform data representing a temporal change in a peak value of a musical sound and to generate the musical sound,
A function capable of acquiring the waveform data;
A function of storing waveform data acquired by the function that can be acquired in a storage device;
In the case where waveform data different for each pitch of the musical sound is used to generate the peak value of the musical sound to be generated, the first portion located at the head of the plurality of partial waveform data obtained by dividing the waveform data The waveform data is acquired for each waveform data by the function that can be acquired and stored in the storage device by the function that can be stored in advance, and the second partial waveform data that is the partial waveform data following the first partial waveform data is A function to be acquired by the function that can be acquired at any time, and to be stored by the function that can be stored in the storage device;
A program to realize