JP2012037562A - Musical sound generating device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To restrict an increase of a burden applied to a musical sound generating processing part even if the number of sound generation channels increases.SOLUTION: A sound source part 20 generates musical sound waveforms with a sound generation unit consisting of two sound generation channels. A sound source register 23 stores tone control data of each the sound generation unit. A waveform memory 21 stores a pair of waveform data; stereo data. A phase generation part 111 generates a phase that is common to both of Lch and Rch of the sound generation unit by accumulating F numbers that are common to Lch and Rch. A waveform reading part 112 reads the waveform data of Lch and Rch from the waveform memory 21 based on respective waveform designating information of Lch and Rch. Musical sound characteristics of the waveform data of Lch and Rch having been read by the waveform reading part 112 are controlled in a characteristic control part 113 based on a characteristic control parameter that is common to Lch and Rch.

Description

  The present invention relates to a musical sound generating apparatus capable of efficiently generating musical sounds even when the number of sound generation channels is increased.

  A conventional musical sound generator includes a performance information input unit that receives performance information from a MIDI (Musical Instrument Digital Interface), keyboard, sequencer, or the like, a control unit (CPU) that generates sound source parameters used for musical sound generation, And a sound source unit that forms a musical sound based on the parameters. The control unit (CPU) executes control processing such as channel assignment and generation of sound source parameters in accordance with the input performance information, and supplies musical sound parameters generated in the sound source unit and a sound generation start instruction. The tone generator section includes a tone generator register for storing a tone parameter used for generating a tone, and a tone generation processing section for performing a tone generation process. Musical tone parameters supplied from the control unit (CPU) are stored in a tone generator register, and a musical tone generation processing unit performs musical tone generation processing based on the musical tone parameters. In the tone generator unit, a tone generation process instructed by the control unit (CPU) is performed by executing a tone generation process based on the tone parameter stored in the tone generator register. As a sound source unit, a waveform memory sound source including a waveform memory in which an actual musical sound waveform is recorded as waveform data is conventionally known. In this waveform memory sound source, waveform data for each tone color and tone range is stored in the waveform memory, and the tone generation processing unit reads waveform data corresponding to the tone color and pitch of the tone to be generated from this waveform memory, The characteristics of the read waveform data are controlled by a control waveform such as a volume control envelope generated based on the sound source parameter. In order to simultaneously generate a plurality of musical tones, a plurality of tone generation channels are formed in the tone generation processing section, and tone generation processing in each tone generation channel is performed in a time-sharing manner. As a result, musical tone waveform samples are calculated and generated for a plurality of tone generation channels for each sampling period, and tone equivalent to the number of tone generation channels is generated.

When a new sound generation instruction (note-on event) occurs, an empty sound channel is detected and assigned, and a musical tone is generated in the assigned sound channel. In this case, if tone generation processing has been performed for all the sound generation channels, that is, if sound generation is assigned to all sound generation channels, the sound generation channel having the least influence is selected even if the sound is muted. Then, a so-called truncation process is performed in which the volume of the musical sound being sounded in the sound generation channel is rapidly attenuated (dumped), and the sound generation channel is released for generating a new music sound. As a channel to which truncation processing is to be performed (truncated channel), a method of selecting a channel with the smallest volume of the currently sounding tone among all the sounding channels is common.
The output unit of such a musical tone generator generally has a stereo configuration, and in order to generate a high-quality musical tone, stereo waveform data is stored in the waveform memory, and the musical tone to be generated is supported from this waveform memory. The stereo sound is also generated by reading the stereo waveform data.

Japanese Patent No. 2671690 Japanese Patent No. 3666346 Japanese Patent No. 2915452

  In the case of stereophonic sound generation by a conventional musical sound generator, in order to generate a single stereo sound, a left channel (hereinafter referred to as “Lch”) and a right channel (hereinafter referred to as “Rch”) 2 Two sounding channels (hereinafter referred to as “ch”) were used. That is, the control unit (CPU) secures two sounding channels by performing sounding assignment processing for assigning sounding channels to Lch and Rch, respectively. Also, since the sound source parameters for each sounding channel are stored in the sound source register, the sound source parameters for Lch and Rch are stored. The musical sound generation processing unit reads waveform data at a speed corresponding to the designated pitch from the waveform memory based on the sound source parameter for each sound generation channel, and the musical sound characteristics of the read waveform data are converted to a volume control envelope or the like. Since the sound is assigned to the Lch, the Lch waveform data is read from the waveform memory based on the Lch sound source parameters, the tone characteristics are controlled, and the Lch tone waveform data is generated. For the sounding channels assigned to Rch, Rch waveform data is read from the waveform memory on the basis of the Rch sound source parameters, and the musical tone characteristics are controlled to generate Rch musical tone waveform data.

As described above, in the conventional tone generator, since two tone generation channels are required to generate a stereo tone waveform sample, there is a desire to increase the number of tone generation channels, and an integrated circuit for tone generation (sound source LSI). ) Tends to increase the number of pronunciation channels that can be handled. However, in order to control the sound channel characteristics during stereo sound generation in the musical sound generation processing unit, it is necessary to perform processing for two channels of Lch and Rch, which increases the burden on the musical sound generation processing unit. was there. In this case, for monophonic sounding, it is only necessary to control the characteristics of only one channel. In stereophonic sounding, if the characteristics of two channels are not controlled at the same time within the same sampling period, musical sounds having different characteristics depending on the timing are produced for Lch and Rch. Therefore, there is a problem that it is necessary to simultaneously perform processing for controlling the characteristics within the same sampling period, which further increases the burden on the musical sound generation processing processing unit.
Accordingly, an object of the present invention is to provide a musical sound generation apparatus that can suppress as much as possible the burden on the musical sound generation processing unit even when the number of sound generation channels is increased.

  In order to achieve the above object, the musical sound generating apparatus of the present invention stores a plurality of waveform data pairs each including stereo-sampled Lch waveform data and Rch waveform data, and mono-sampled mono waveform data. A sound generation unit composed of two sound generation channels operating in a stereo sound generation mode or a monaural sound generation mode according to a mode flag, and N (N is an integer of 1 or more) sound generation units. Each sound generation unit operating in the stereo sound generation mode reads 2ch of waveform data from the waveform memory to generate a 2ch musical sound waveform, and each sound generation unit operating in the monaural sound generation mode is 1ch from the waveform memory. Minute waveform data is read out, and a sound source unit that generates a 1-ch musical sound waveform, and N sound generation units corresponding to N sound generation units Color control data is stored, and each tone color control data is sent to the tone generator register including the mode flag, the first waveform designation information, the second waveform designation information, the F number, and the characteristic parameter. A controller that controls the operation of the sound source unit by writing parameters, and the control unit is one of N sounding units according to a sounding instruction command (note-on) for instructing the sounding of a new musical sound. One is assigned to the generation of the musical sound instructed to be generated, and it is determined whether the musical sound instructed to be generated is stereo or monaural. If the musical sound instructed to be generated is stereo, a mode flag indicating a stereo sound generation mode, Forming tone color control data including first waveform designation information, second waveform designation information, F number and characteristic parameters used for generation, and assigning the tone generator register When the musical sound instructed to be generated is monaural, the mode flag indicating the monaural sounding mode, the first waveform designation information used for generating the musical sound, the F number, and Forming timbre control data including a characteristic parameter, and setting the storage area corresponding to the sounding unit assigned in the sound source register to perform a process of instructing the start of sounding to the assigned sounding unit; In response to a sounding start instruction for one sounding unit from the control unit, the sound source unit is set in the sound source register if the mode flag of the sounding unit stored in the sound source register indicates a stereo sounding mode. Lch (or Rch) waveform data indicated by the first waveform designation information of the sound generation unit and Rch ( Or Lch) waveform data is read from the waveform memory at the speed indicated by the F number of the sound generation unit, and the respective tone characteristics of the read Lch waveform data and Rch waveform data are set in the tone generator register. If control is performed based on the characteristic control parameter of the sound generation unit to form an Lch tone signal and an Rch tone signal, and the mode flag of the sound generation unit stored in the tone generator register indicates the monaural sound mode The mono waveform data indicated by the first waveform designation information of the sounding unit stored in the tone generator register is read from the waveform memory at the speed indicated by the F number of the sounding unit, and the tone characteristics of the read mono waveform data are obtained. Then, a monaural tone signal is formed by controlling based on the characteristic control parameter of the sounding unit set in the tone generator register. It is the most important feature to perform a physical.

According to the present invention, when a sound generation instruction is generated, one sounding unit may be assigned regardless of whether the sound is monaural or stereo. In addition, since there is data commonly used in one sounding unit, in the case of stereo sounding, the amount of data to be set in the sound source register is less than double that in the case of monaural sounding. Furthermore, when starting sound generation, it is only necessary to indicate the start of sound generation for the assigned sound generation unit regardless of monaural sound or stereo sound after setting the sound source register. There is no need to perform control such as “note-on”.
Furthermore, since the F number or control parameter is commonly used by Lch and Rch, the control unit only needs to change the value of the F number or control parameter of the tone generator register by one even for stereophonic sound. It is not necessary to perform control so that the value is changed simultaneously for two channels.
The characteristic control parameter is a parameter for controlling any one of a pitch envelope waveform (pitch EG waveform), a volume envelope waveform (volume EG waveform), a filter envelope waveform (filter EG waveform), and a low frequency waveform (LFO waveform). Any one of the above is applicable, but in addition to this, a release start instruction (note-off) is also applicable.

It is a block diagram which shows the outline | summary of a structure of the musical sound production | generation apparatus concerning the Example of this invention. It is a block diagram which shows the detailed structure of the musical sound production | generation apparatus concerning the Example of this invention. It is a figure which shows the data memorize | stored in the sound source register in the musical sound production | generation apparatus of this invention. It is a figure which shows the data structure of the waveform data memorize | stored in the waveform memory in the musical tone generator of this invention. It is a figure which shows the data structure of the waveform management data and tone color data which are memorize | stored in the flash memory in the musical tone production | generation apparatus of this invention. It is a flowchart of the note-on event process performed in the musical tone generation device of the present invention. It is a flowchart of the note-off event process performed in the musical tone generating apparatus of the present invention. It is a flowchart of the pitch bend event process performed in the tone generation apparatus of the present invention.

FIG. 1 is a block diagram showing an outline of the configuration of a musical tone generation apparatus according to an embodiment of the present invention.
In the musical sound generating apparatus 1 of the present invention shown in FIG. 1, two sound generation channels (hereinafter referred to as “channels”), ie, a left sound channel (hereinafter referred to as “Lch”) and a right channel (hereinafter referred to as “Rch”) are generated. The sound source unit 20 includes N (N is an integer equal to or greater than 1) sounding units. For example, when the number of sounding channels in the sound source unit 20 is 256, the number of sounding units is 128, which is half the number of sounding channels. In order to simultaneously generate a plurality of musical tones in the musical tone generating apparatus 1 according to the present invention, the musical tone generation processing for a plurality of sounding units is performed in a time division manner for each sampling period. Furthermore, since each tone generation unit calculates and generates musical sound waveform samples for two channels in a time-sharing manner, a musical sound waveform sample that is twice the sound generation unit is calculated and generated for each sampling period. Here, in the musical sound generating apparatus 1 according to the present invention, a part of the parameters of the Lch and Rch, which are sound generation units, are shared between the two channels. That is, the type 1 parameters that are individually set for each sound generation channel of Lch and Rch are stored in the sound generation unit area prepared in the sound source register 23 independently for each channel. The type 2 parameter shared between the two channels of Lch and Rch is stored in common in the sound generation unit area prepared in the tone generator register 23. That is, the tone generator register 23 is provided with an area for storing N timbre control data corresponding to N sounding units, and each timbre control data is a type 1 parameter used for the timbre. Lch waveform designation information indicating stereo Lch waveform data, Rch waveform designation information indicating Rch waveform data, and an F number for determining the pitch of a musical tone common to both Lch and Rch channels, which are type 2 parameters And characteristic control parameters.

The waveform memory 21 stores a plurality of waveform data pairs of stereo-sampled Lch waveform data and Rch waveform data. This waveform data pair is stored for each tone color and tone range, and monaural sampled mono waveform data is stored in the waveform memory 21 for waveform data of a tone color that is not suitable for stereo sound generation.
A data structure (memory map) of a large number of waveform data recorded in the waveform memory 21 is shown in FIG. In the data structure shown in FIG. 4, the waveform data WD1m is monaural waveform data, the waveform data WD2sl and WD2sr are stereo waveform data pairs, of which WD2sl is Lch waveform data and WD2sr is Rch waveform data. It is. Similarly, the waveform data WD3sl and WD3sr are other waveform data pairs in stereo, of which WD3sl is Lch waveform data and WD3sr is Rch waveform data. In the waveform memory 21, a large number of such monaural waveform data and stereo waveform data pairs are recorded. Lch waveform data and Rch waveform data are one waveform data when viewed one by one, and are not different from monaural waveform data. Therefore, only one of the Lch waveform data and the Rch waveform data can be read as monaural waveform data.

Further, the phase generation unit 111 accumulates the F number of the sounding unit stored in the sound source register 23 for each of the plurality of sounding units in a time division manner, so that both the Lch and Rch of the sounding unit are obtained. A common phase is generated for ch. The accumulation of each sounding unit is started with a sounding start instruction of the sounding unit received by the receiving unit 117a from the control unit 122 as a trigger. Note that the reception unit 117a receives a sound generation start instruction independently of each other for each of the plurality of sound generation units.
Further, the waveform reading unit 112 time-divisionally, for each of a plurality of sounding units, the phase of the sounding unit supplied from the phase generating unit 111 and the Lch waveform designation information of the sounding unit stored in the sound source register 23. Based on the above, one Lch waveform data is read from the waveform memory 21, and also based on the phase supplied from the phase generator 111 and the Rch waveform designation information of the sound generation unit stored in the sound source register 23, One Rch waveform data is read from the waveform memory 21.

  Furthermore, the control waveform generation unit 116 time-divides each of the plurality of sounding units, the sounding unit whose value changes with time based on the control waveform parameter of the sounding unit stored in the sound source register 23. A common characteristic control waveform is generated for both Lch and Rch. The characteristic control waveform generated by the control waveform generation unit 116 is supplied to the characteristic control unit 113, and the tone characteristics of the Lch waveform data and the Rch waveform data are controlled based on the characteristic control waveform, respectively. Musical tone signals and Rch musical tone signals are formed. The characteristic control waveform generated by the control waveform generator 116 includes a pitch envelope waveform (pitch EG waveform) for controlling the pitch change of the musical tone, a volume envelope waveform (volume EG waveform) for controlling the volume change of the musical tone, and the frequency of the musical tone. There are a filter envelope waveform (filter EG waveform) that controls changes in characteristics, and a low frequency waveform (LFO waveform) that applies low frequency modulation to the pitch and volume of a musical tone. Further, by supplying the LFO waveform generated by the control waveform generator 116 to the phase generator 111, vibrato can be added to the musical sound signal. By supplying this LFO waveform to the characteristic controller 113, the musical sound A wah wah or tremolo can be added to the signal.

  Here, when the control unit 122 receives performance information from MIDI, a keyboard, a sequencer or the like and receives a sound generation instruction command (note-on) for instructing the sound of a new musical sound, the control unit 122 performs stereo sound generation or monaural. Regardless of the sound generation, a process of assigning any one of the N sound generation units to the generation of the musical sound instructed to sound is performed (steps S11 to S15 described later). Next, timbre control data including Lch waveform designation information, Rch waveform designation information, F number, and characteristic control parameters used to generate a musical tone instructed to be generated by the control unit 122 is formed to correspond to the assigned pronunciation unit. Is set in the storage area of the tone generator register 23 (steps S16 and S17). Then, the reception unit 117a is instructed to start sound generation in the sound generation unit assigned (step S18). In the case of stereo sounding, the sounding start is instructed to 2ch of the sounding unit. As a result, using the sound generation start instruction of the sound generation unit received by the reception unit 117a as a trigger, the phase generation unit 111 starts accumulating the F number of the sound generation unit and is common to both the Lch and Rch of the sound generation unit. The phase is generated. The waveform reading unit 112 to which the generated phase is supplied is based on the phase of the sounding unit supplied from the phase generating unit 111 and the Lch waveform designation information of the sounding unit stored in the sound source register 23. One Lch waveform data is read from the memory 21, and one Rch is read from the waveform memory 21 on the basis of the phase supplied from the phase generator 111 and the Rch waveform designation information of the sounding unit stored in the sound source register 23. Read waveform data.

  The Lch waveform data and the Rch waveform data read by the waveform reading unit 112 are respectively controlled based on a characteristic control waveform whose tone characteristics change over time during sound generation generated by the control waveform generation unit 116, and An Lch musical sound waveform sample and an Rch musical sound waveform sample of the sound generation unit are formed. The Lch musical sound waveform sample and the Rch musical sound waveform sample for each sounding unit are output from the characteristic control unit 113 to the Lch side and the Rch side of the stereo configuration accumulation unit (MIX) 114 in a time division manner. The musical sound waveform samples output to the Lch side and the musical sound waveform samples output to the Rch side of the number of sounding units allocated within the sampling period are accumulated. From the MIX 114, an Lch musical sound waveform sample and an Rch musical sound waveform sample accumulated over all sounding units are obtained every 1 DAC period (conversion period of the digital-to-analog converter (DAC) 115), which is one sampling period. , Output to the DAC 115. The stereo DAC 115 converts the Lch musical sound sample and the Rch musical sound sample accumulated for each DAC period into an analog Lch musical signal and an Rch musical signal, and the stereo sound system 22 ( A stereo tone signal is emitted from an amplifier and a speaker.

  When the number of sounding channels is 256 and the number of sounding units is 128 in the musical sound generating device 1 according to the present invention, the phase generator 111, the waveform reading unit 112, the control waveform generator 116, and the characteristic controller 113 at the time of stereo sounding. The number of required processing slots (time slot for time division processing) is indicated in parentheses in each block. That is, since the phase generation unit 111 generates a phase common to 2ch for each sound generation unit, the number of phase processing slots is 128. Since the waveform reading unit 112 reads the waveform data from the waveform memory 21 for each of the two channels for each sound generation unit, the number of processing slots of the waveform data is 256. Further, since the control waveform generation unit 116 generates a control waveform common to 2ch for each sounding unit, the number of processing slots of the control waveform is 128, and the characteristic control unit 113 stores waveform data for 2ch for each sounding unit. Since the tone characteristics are controlled, the number of processing slots for controlling the tone characteristics is 256.

  By the way, when a monaural sound is instructed to be generated for a sound generation unit, the waveform reading unit 112 outputs one mono waveform data based on the Lch waveform designation information (or Rch waveform designation information) and the phase from the phase generation unit 111. Reading from the waveform memory 21 is started. That is, only 1ch waveform data is read out in 2ch sounding units. The mono waveform data read by the waveform readout unit 112 is controlled by the characteristic control unit 113 by the characteristic control waveform that changes in time during the sound generation supplied from the control waveform generation unit 116 in the characteristic control unit 113, so that a monaural musical sound waveform sample is obtained. It is formed and output to the Lch side and the Rch side of the MIX 114. As described above, the MIX 114, the DAC 115, and the sound system 22 have a stereo configuration, and the characteristic control unit 113 performs the same monaural musical sound waveform sample from 1ch of the sound generation unit on the Lch side and Rch side of the MIX 114 at a volume based on the pan parameter. Output to the side. The musical sound waveform samples output to the Lch side and the Rch side are respectively accumulated by the MIX 114 with the musical sound samples output to the Lch side and the Rch side in other sound generation units, and output to the DAC 115. In the stereo DAC 115, the accumulated Lch and Rch musical sound waveform samples are converted into analog Lch and Rch music signals for each DAC period, and the music signal is emitted from the stereo sound system 22. The In this case, a stereo signal of stereo sound and a monaural signal of monaural sound are mixed and emitted from the sound system 22.

  As described above, in the musical sound generating apparatus 1 according to the present invention, two musical sound signals of the Lch musical sound signal and the Rch musical sound signal are formed on the basis of the phase signal common to the Lch and Rch, which are sounding units. Thus, the number of phases generated by the phase generator 111 becomes the number of sounding units that is half the number of musical tones that can be generated, and the processing amount of the phase generator 111 can be reduced and the circuit scale can be reduced. Generally, precise pitch control is required for tone generation, and the phase generation unit generates a phase with a bit length exceeding 20 bits, so the circuit scale tends to increase. Further, since the control waveform generator 116 generates a common control waveform for Lch and Rch based on the characteristic control parameter of each sound generation unit, the number of characteristic control waveforms generated by the control waveform generator 116 can be generated. Therefore, the number of sounding units is reduced to half the number of musical tones, and the processing amount of the control waveform generator 116 can be reduced and the circuit scale can be reduced. Furthermore, since the sound generation assignment process is performed for sound generation units that are half the number of sound generation channels, the processing load on the control unit 122 can be reduced even when the truncation process is performed. In the sound assignment, the amount of processing load for the processing increases exponentially according to the number of sounding channels or sounding units to be assigned. Further, the F number and the characteristic control parameter are common to the two channels that are the sounding unit, and only one setting is required for each sounding unit, so the control unit 122 sets the characteristic control parameter in the sound source register 23. Time can be shortened, and the scale of the tone generator register 23 can be reduced.

  The musical sound generating apparatus 1 according to the present invention can produce stereo and monaural pronunciations, but if a sounding instruction is generated regardless of whether the sounding is monophonic or stereophonic, one sounding unit is assigned. Good. In the musical sound generating device 1, the Lch waveform data and the Rch waveform data are read from the waveform memory 21 by the waveform reading unit 112 at a speed corresponding to the same F number for each sound generation unit in the case of stereo sounding, and the characteristic control is performed. The unit 113 performs characteristic control with the same characteristic control parameter (or characteristic control waveform based on the characteristic control parameter) to form an Lch tone signal and an Rch tone signal in which the pitch and tone characteristics are similarly controlled. In this case, since the two tone signals of the Lch tone signal and the Rch tone signal are formed based on the phase signal common to the Lch and the Rch, the number of phases generated by the phase generator 111 is the number of tones that can be generated. Therefore, the number of sound generation units is reduced to half of the number of units, so that the processing amount of the phase generator 111 can be reduced and the circuit scale can be reduced. Further, since the control waveform generator 116 generates a common control waveform for Lch and Rch based on the characteristic control parameter of each sounding unit, the number of characteristic control waveforms generated by the control waveform generator 116 is generated. The number of sounding units is reduced to half the number of possible musical sounds, and the processing amount of the control waveform generator 116 can be reduced and the circuit scale can be reduced. The characteristic control waveform is any one of a pitch envelope waveform (pitch EG waveform), a volume envelope waveform (volume EG waveform), a filter envelope waveform (filter EG waveform), and a low frequency waveform (LFO waveform).

As described above, since the F number or the characteristic control parameter is commonly used for Lch and Rch, the amount of data to be set in the sound source register 23 is smaller than that in the case of monaural sounding even in stereo sounding. It becomes quantity. For this reason, the set time for the control unit 122 can be shortened. Further, after various parameters are set in the tone generator register 23, the sounding start instruction may be given to the assigned sounding unit regardless of monaural sounding or stereo sounding. That is, it is not necessary to bother to perform control to simultaneously note-on Lch and Rch 2ch even for stereo sounding. Furthermore, since the F number or the characteristic control parameter is used in common for Lch and Rch, the control unit 122 can control the F number or characteristic control parameter for one sounding unit of the tone generator register 23 even in the case of stereo sounding. It is only necessary to change the value in real time, and it is not necessary to change the two parameter values related to the two sounding channels simultaneously in real time, so that the amount of processing applied to the control unit 122 can be reduced.
Although the musical tone generating apparatus 1 of the present invention shown in FIG. 1 is realized by an integrated circuit for generating musical sounds, the processing amount of some blocks can be reduced as described above. Even if it increases, it becomes possible to suppress as much as possible the increase in the scale of the integrated circuit of the sound source unit 20 that generates musical sounds.

Next, FIG. 2 shows a detailed configuration of the musical tone generating apparatus 1 according to the present invention shown in FIG. The musical tone generation device 1 shown in FIG. 2 performs the same musical tone generation operation as that of the musical tone generation device 1 shown in FIG. 1 and has the same operational effects. The blocks assigned the same numbers in FIG. 1 and FIG. Basically perform the same function as each other.
In the musical sound generating device 1 shown in FIG. 2, one sounding unit is configured by two channels of Lch and Rch that generate stereo as described above, and the sound source unit 20 has N (N is an integer of 1 or more) pieces. Has pronunciation units. For example, when the number of sounding channels in the sound source unit 20 is 256, the number of sounding units is 128, which is half the number of sounding channels. In order to simultaneously generate a plurality of musical sounds in the musical sound generation apparatus 1 according to the present invention, the musical sound generation processing for a plurality of pronunciation units is performed in a time division manner. In each sound generation unit, musical sound waveform samples are generated in each of 2ch, so that a musical sound waveform sample that is twice the sound generation unit is calculated and generated in a time division manner for each sampling period.

  In the musical sound generating device 1 shown in FIG. 2, the CPU 10 executes various programs related to musical sound generation to control operations of various functions including musical sound generation in the musical sound generating device 1 (Central Processing Unit). ). A flash ROM (Read Only Memory) 11 is a rewritable nonvolatile memory in which various programs such as a tone generation process program executed by the CPU 10 and timbre data are stored. A RAM (Random Access Memory) 12 is a volatile main memory in the musical tone generating apparatus 1 and a work area of the CPU 10 is set. The operation element 13 is a performance operation element such as a keyboard and various switches, and various instructions can be given to the tone generator 1 by operating the various switches. Furthermore, the display device 14 is a display device for displaying various types of information when generating musical sounds. The communication I / O 15 is an interface that can take in performance information and the like from the outside. For example, the communication I / O 15 is a MIDI interface that transmits and receives MIDI messages with the outside. Based on the control of the CPU 10, the tone generator unit 20 reads out waveform data necessary for musical tone generation from the waveform memory 21, performs processing such as waveform data interpolation, envelope assignment, channel accumulation (mixing), and the like. It is output as a musical sound waveform signal. In this figure, there are two arrows from the accumulation unit (MIX) 20i to the sound system 22, and it is clearly shown that two pieces of data of the stereo configuration pass, but the other one arrow is 1 It does not indicate that one piece of data passes, but a plurality of signals pass in time division. For example, 128 data corresponding to 128 tone generation units are supplied from the F number generator (FG) 20a to the phase generator (PG) 20b in a time division manner. The musical sound waveform signal output from the sound source unit 20 is supplied to the sound system 22 so that a stereo musical sound is emitted. Each unit is connected via a bus line 16.

In the sound source register 23 in the sound source unit 20, 128 timbre control data corresponding to 128 sound units are set. Each tone color control data is formed by the CPU 10 that performs note-on processing, and is set in an area prepared for each tone generation unit of the tone generator register 20. The tone color control data set in the area of the tone generator register 23 includes data for each tone generation unit shown in FIG. 3A and each tone generation channel (256 ch) of Lch and Rch in the tone generation unit shown in FIG. It consists of data.
The data for each sound generation unit shown in FIG. 3A includes pitch shift data (PS) expressed in cent values, LFO frequency, PM depth, FM depth, LFO parameters including AM depth, and a plurality of states. PEG parameters (PEGPs) composed of PEG rates and PEG levels, FEG parameters (FEGPs) composed of FEG rates and FEG levels in a plurality of states, AEG parameters (AEGPs) composed of AEG rates and AEG levels in a plurality of states, Pan data (PANs) composed of Lch pan data (PAN (L)) and Rch pan data (PAN (R)), a mode flag (Mode) indicating a stereo sound mode or a monaural sound mode, and Has been. Here, the pitch shift data (PS) is data for controlling the amount by which the pitch of the waveform data read from the waveform memory is shifted, and is an F number in the cent scale. An F number generation unit (FG) 20a described later generates an F number on a linear scale of frequency ratio, and this PS is a main factor that determines the value of the generated F number. The resolution of PS is 1 cent, and an interpolator is provided to smooth the change of the value so that no noise is generated even if it is changed in real time. In addition, in these parameters, parameters with a lowercase “s” are not composed of a single parameter, but are composed of a plurality of parameters.

Further, the data of the sound generation channels of the Lch and Rch in the sound generation unit shown in FIG. 3B is obtained from the start address (WS), the loop start address (LS), and the end address (WE) which are the waveform designation information of the channel. And other parameters including waveform information parameters (WAPs), compression information for controlling the release when the waveform data is compressed, loop information for controlling loop reading, and the like. In the monaural sound generation mode, monaural waveform data is designated by the Lch waveform designation information, and the Rch waveform designation information is not used.
These three addresses WS, LS, and WE are respectively absolute addresses of the waveform memory, but any one of them may be left as an absolute address and the remaining two may be relative addresses from the absolute address. For example, the loop head address LS may be an absolute address, and the head address WS and the tail address WE may be relative addresses from the loop head address LS. In this case, the two addresses, which are relative addresses, can be shared between the two channels of the sound generation unit, and can be moved to the data for each sound generation unit (FIG. 3A). That is, the waveform designation information of the present invention corresponds to an address that is at least one of the three addresses WS, LS, and WE.
Further, in the tone generator register 23, apart from the region for each sounding unit shown in FIGS. 3A and 3B, a sounding start flag GT (instructing each sounding unit GU to start sounding) GU) area and a release start flag RT (GU) area for instructing the start of release.

  The flash ROM 11 stores a plurality of waveform management data for managing a plurality of waveform data recorded in the waveform memory 21 and a plurality of tone color data that are parameter sets of a plurality of tone colors that can be selected by the tone generator 1. Has been. The tone generator 1 of the present embodiment is a multi-part sound source, and for each part, one tone color can be selected from the plurality of tone colors and set as the tone color of the part. And according to the performance information of a plurality of parts, a musical tone is generated with a tone color set for each part. FIG. 5A shows the data structure of the waveform management data, and FIG. 5B shows the data structure of the timbre data. When reading the waveform data from the waveform memory 21, the waveform data is selected by referring to the tone color data set in the sound source part, and the waveform data is selected from the waveform memory using the waveform management data of the selected waveform data. Read the waveform data.

  The plurality of waveform management data shown in FIG. 5A is data necessary for reading the waveform data from the waveform memory 21, and each waveform management data is each stereo waveform data pair stored in the waveform memory 21. Corresponds to monaural waveform data. WKD1 is waveform management data of monaural waveform data WD1m, WKD2 is waveform management data of a stereo waveform data pair of waveform data WD2sl and WD2sr, and WKD3 is waveform management data of a stereo waveform data pair of waveform data WD3sl and WD3sr. is there. Each waveform management data has the same data structure regardless of whether the corresponding waveform data is a stereo waveform data pair or monaural waveform data. As shown in the expanded view of WKD2, as shown in FIG. 1, waveform address information 2, basic pitch, loop information, and compression information. If the stereo flag is “1”, the corresponding waveform data is a stereo waveform data pair, and if “0”, it is monaural waveform data. When waveform data with a flag set is selected for a new sound, a mode flag (Mode) indicating the stereo sound mode is set in the sound generator register of the sound unit assigned for that sound, and the flag is set. When waveform data that has not been selected is selected, a mode flag (Mode) indicating a monaural sound generation mode is set.

  The waveform address information 1 indicates the storage position of monaural waveform data in the waveform memory 21 when the stereo flag is “0”, and indicates the storage position of Lch waveform data in the waveform memory 21 when the stereo flag is “1”. It consists of information on the start address, loop start address, and end address of monaural waveform data to Lch waveform data. The waveform address information 2 is information (Don't Care) that is not used when the stereo flag is “0”, and is information that indicates the storage position of the Rch waveform data when the stereo flag is “1”. It consists of address, loop head address, and tail address information. The basic pitch is information indicating the pitch of the read waveform data when the waveform data is read without pitch shift (F number = 1), and the loop information indicates whether or not loop reading is performed. Information for correcting the pitch of the loop portion for adjusting the start pitch of the loop waveform to the pitch of the end of the attack waveform. The compression information is compression method information indicating whether or not the waveform data is compressed and a compression method when the waveform data is compressed, and compression coefficient information for releasing the compression of the first frame. At the start of sound generation, the CPU 10 sets these pieces of information in the region of the sound generation unit of the sound source register 23, and the waveform address generation unit 20c, the read & cache unit 20d, the decoder 20e, and the like are controlled based on the set information. The

5B, the flash ROM 11 stores a plurality of timbre data TD1, TD2, TD3,... Of a plurality of timbres, and the timbre data of each timbre is expanded to TD2. As shown, the tone color name, waveform selection information, PEG parameters common to Lch and Rch, FEG parameters, AEG parameters, LFO parameters, and other parameters are included.
Waveform selection information is waveform data (waveform management data for reading out) used to generate a musical sound corresponding to the sound generation instruction according to the pitch (note number) and performance intensity (velocity) related to the sound generation instruction. This is information for selecting. For example, the waveform selection information of the timbre data set in the keyboard part is referred to by the note number or velocity associated with the note-on in response to the note-on when the keyboard as the operation element 13 is played. Waveform management data corresponding to one stereo waveform data pair or monaural waveform data is selected.
The PEG parameter, the FEG parameter, and the AEG parameter are parameters that control the pitch EG waveform, the filter EG waveform, and the volume EG waveform, each having a multi-state polygonal line shape. It also consists of parameters for adjusting some rates or levels according to the note number or velocity. The plurality of states of the volume EG waveform are composed of, for example, five states of attack (or hold), first decay, second decay, sustain and release, and the attack (or hold) level is increased or decreased according to velocity. . The LFO parameter includes an LFO frequency for controlling the frequency of the LFO waveform to be generated, a PM depth for controlling the amplitude (depth of pitch modulation) for modulating the pitch of the musical sound, and a frequency characteristic for modulating the musical sound. It consists of FM depth for controlling the amplitude (depth of filter modulation) of the LFO waveform and AM depth for controlling the amplitude (depth of amplitude modulation) of the LFO waveform for modulating the amplitude of the musical sound.

  If a new note-on instructing the start of musical tone generation is detected in the musical tone generating device 1, the CPU 10 executes a note-on event process shown in FIG. In the note-on event process, the CPU 10 first assigns a new note-on to an empty sounding unit by performing a sounding assignment process, and the assigned sounding unit area in the tone generator register 23 (FIGS. 3A and 3B). ), Tone color control data composed of various new note-on parameters is set. Then, when the CPU 10 sets a sound generation instruction (GT) for the sound generation unit, the sound source unit 20 generates a musical tone corresponding to a new note-on using the set tone color control data. Start in units. Here, when the mode flag (Mode) is set to the stereo sound generation mode, a stereo tone signal is generated from the sound source unit 20 and output to the sound system 22, and the mode flag (Mode) is set to the monaural sound mode. If so, a monaural musical sound signal is generated from the sound source unit 20 and output to the sound system 22.

This note-on event process will be described in more detail. Note-on is a command (MIDI message) for instructing the start of a new musical tone, and as a parameter, a part number indicating which part is the command and a note number indicating the pitch of the musical tone to be generated And a velocity indicating strength. For example, when any key is pressed on the keyboard as the operation element 13, the part number indicating the part controlled by the keyboard, the note number of the pressed key, and the velocity indicating the pressing speed, A note-on containing is generated. In the automatic performance of music data of a plurality of parts, note-on for each part is generated. Further, note-on may be received from an external device via the communication I / O 15.
When any of these note-on is detected, the CPU 10 starts the note-on event processing shown in FIG. 6, and first extracts the part number, note number, and velocity from the note-on generated in step S10. The part number is stored in the area PT secured in the work area of the RAM 12, the note number is stored in the area NN, and the velocity is stored in the area VL. Next, in step S11, a process of detecting an empty pronunciation unit (Unit) that is not assigned a pronunciation from all the pronunciation units is performed, and it is determined in step S12 whether or not an empty pronunciation unit is detected in the detection process. The If it is determined that an empty pronunciation unit has been detected, the process proceeds to step S13, and the Unit number of the empty pronunciation unit is stored in the area AU secured in the work area of the RAM 12. If it is determined in step S12 that an empty pronunciation unit is not detected, the process branches to step S14 to perform a truncated unit determination process, and stores the unit number of the determined truncated pronunciation unit in the area AU.

  In the truncated unit determination process, first, a search target part for searching a truncated sound unit such as a low-priority part for generating a musical sound is determined from all the parts. The RAM 12 is provided with an area in which the CPU 10 records the part number, note number, and the like of the musical tone being sounded in each sounding unit. The CPU 10 refers to the information of the area and determines the part that has been determined. Detects pronunciation units during pronunciation. Next, in a plurality of detected sound generation units, a sound generation unit having a minimum combined value of the amplitude of the Lch waveform data and the sound volume EG waveform within the volume level of the 2ch musical sound being sounded is detected and truncated. Determined as a unit. This is because even in the monaural sound generation mode, a musical sound is generated on the Lch in the sound generation unit. The unit number of the determined truncation pronunciation unit is stored in the area AU, and the truncated unit determination process is terminated.

When the truncated unit determination process in step S14 is completed, a rapid attenuation process (dump process) is performed on the 2ch musical sound of Lch and Rch being generated with the unit number stored in the area AU in step S15. Through the processing so far, one sounding unit (Unit) is assigned to generate a new musical tone, and the assigned sounding unit number is stored in the area AU. When the process of step S13 or step S15 is completed, the process proceeds to step S16, and the waveform data is obtained from the waveform selection information of the tone color data (FIG. 5B) currently set for the part indicated by the part number in the region PT. select. Then, referring to the stereo flag of the waveform management data (FIG. 5A) corresponding to the selected waveform data, if the waveform data is a stereo waveform data pair, the sound generation mode is determined as the stereo sound generation mode, If the waveform data is monaural, the monaural sound generation mode is selected.
Next, in step S17, various parameters used for generating a new musical tone are determined based on the timbre data and the part number, note number, and velocity stored in the areas PT, NN and VL, and the tone generator register The determined various parameters are set in a storage area (FIGS. 3A and 3B) corresponding to the unit number stored in the area AU. Various parameters set by the CPU 10 include a mode flag (Mode) indicating the determined sound generation mode, waveform selection information in the tone color data of the current tone color, waveform data selected in accordance with the note number and the velocity value. LFO in the tone color data of the current tone color common to both the Lch and Rch channels processed according to the Lch waveform designation information, Rch waveform designation information, loop information, compression information, the note number and the velocity value The tone color control data includes parameters, PEG parameters, FEG parameters, AEG parameters, pitch shift data (F number in cent scale) based on the difference between the note number and the basic pitch, and pan data. However, the Rch waveform designation information is set only in the stereo sounding mode, and may not be set in the monaural sounding mode (since there is no information to be set). After setting the timbre control data in the AU area in step S17, the CPU 10 writes “1” in the sound generation start flag GT (AU) of the sound source register 23 in step S18, thereby generating the sound indicated by the unit number of the area AU. Instruct the unit to start pronunciation. As a result, the note-on event process ends, and the sound source unit 20 uses the sound generation start instruction (GT (AU) ← 1) as a trigger, based on the set tone color control data in the assigned sound generation unit (Unit). Generation of Lch and Rch musical sounds (stereo sound generation mode) or generation of monaural chine sounds (monaural sound generation mode) is started.
In the present embodiment, substantially the same processing is performed regardless of whether the sound generation mode is the stereo sound mode or the monaural sound mode in the sound generation assignment processing in steps S12 to S15 marked with “*” and the sound generation start instruction processing in step S18. It is characterized in that it is performed.

Operation of each unit constituting the sound source unit 20 in the time slot of the sound generation unit when generating a musical sound in the sound generation unit (Unit) in which the mode flag (Mode) indicating the stereo sound generation mode of the sound source unit 20 is set The description is given below.
The F number generating unit (FG) 20a that generates the F number includes the modulation data from the LFO 20m (the LFO waveform whose amplitude is adjusted by PM depth) and the pitch shift from the sound source register 23 in the time slot of the sound generation unit. The data (PS) and the pitch EG waveform from the pitch envelope generator (PEG) 20k are supplied, these three data are added, and further converted from a cent scale to a linear scale, and the Lch and Rch An F number common to the pronunciation units is generated. In this case, these three data are data in units of cent values, and the PS directly supplied from the sound source register 23 is the pitch when the waveform data read from the waveform memory 21 is recorded (FIG. 5A ) Basic pitch) and the pitch (note number) of the musical sound to be generated is data expressed as a cent value. The F number generated in the FG 20a is a linear value data in which a rough value is determined by the pitch shift data, and a relatively small adjustment is made to the value by the modulation data and the pitch EG waveform. For example, when a musical tone is generated with the same pitch as the recorded pitch, PS is set to “0”, and the data from the adjustment values LFO20m and PEG20k are ignored. In this case, the FG20a generates the tone. The F number is “1”. The F number from the FG 20a is supplied to the phase generator (PG) 20b, and the PG 20b accumulates the F number supplied for each sampling period in the time slot of the sound generation unit, and the upper bit of the accumulated value is obtained. The lower bits are output to the waveform address generation unit (WAG) 20c as the integer part of the phase common to the sound generation units of Lch and Rch for generating the read address as the fractional part of the same phase. Accumulation of the F number in the PG 20b is started with a sound generation instruction (GT) of the sound generation unit supplied from the reception unit of the sound source register 23 as a trigger, and zero as an initial value. The integer part of the generated phase is data specifying the position of the sample unit of the waveform data, and the decimal part is data specifying the intermediate position between two adjacent samples.

The PEG 20k is supplied with the PEG rate and PEG level PEG parameters (PEGPs) for each state that are common to the Lch and Rch sounding units from the sound source register 23 in the sounding unit time slot. Based on this, a pitch EG waveform consisting of a plurality of states common to the sound generation units of Lch and Rch and having a value that changes every sampling period is generated and supplied to the FG 20a. The generation of the pitch EG is started with a sound generation instruction (GT) of the sound generation unit supplied from the sound source register 23 as a trigger. By supplying the pitch EG waveform to the FG 20a, the reading speed from the waveform memory 21 is appropriately modulated according to the pitch EG, and a pitch modulation effect such as an attack pitch is given.
In the WAG 20c, the Lch and Rch time slots of the sound generation unit, the Lch waveform designation information including the start address, loop start address, and end address of the Lch waveform data from the tone generator register 23, the start address of the Rch waveform data, and the loop start address Rch waveform designation information consisting of the end address and waveform address parameters (WAPs) composed of loop information are supplied, and based on these parameters (WAPs) and the integer part of the phase, the Lch of the sounding unit Waveform address and Rch waveform address are generated. For each channel of the sound generation unit for which loop information that is not loop read is set, the waveform address that proceeds from the sample position indicated by the start address to the sample position indicated by the end address at the same speed as the increase rate of the integer part of the phase for each sampling period Is generated. In each channel of the sound generation unit for which loop information to be read out is set, the sample indicated by the loop head address is moved from the sample position indicated by the head address to the sample position indicated by the tail address at the same speed as the increase speed. A waveform address that repeatedly proceeds from the position to the sample position indicated by the end address is generated. The Lch and Rch waveform addresses (integer part) output from the WAG 20c are supplied to the reading & cashing part 20d, and the fractional part of the Lch and Rch phase is supplied to the intersample interpolation part (INT) 20f. The waveform memory 21 stores compressed waveform data, which is compressed waveform data, and uncompressed linear waveform data, and the read & cash unit 20d uses the Lch and Rch time slots of the sound generation unit. The Lch and Rch waveform data samples are read from the waveform memory 21 in accordance with the supplied Lch and Rch waveform addresses and compression information. If the read Lch and Rch waveform data is compressed waveform data, the decoder (DEC) 20e expands the Lch and Rch time slots of the sound generation unit to obtain the original Lch waveform data sample. And Rch waveform data samples are output. On the other hand, in the case of linear waveform data, it passes through the DEC 20e in the same time slot and is output as an Lch waveform data sample and an Rch waveform data sample.

Here, the compressed waveform data recorded in the waveform memory 21 will be described. Waveform data composed of a plurality of consecutive samples is divided into a plurality of frames, and compression processing is performed to make the waveform data variable length for each frame. . The compressed waveform data for one frame is stored over a predetermined number j of consecutive addresses in the waveform memory 21 capable of storing n-bit data for each address. In the data stored at j addresses, the compressed waveform data having the same number of bits in the frame is sequentially stored in m bits out of n bits. Then, sub information including the compression information is stored in the remaining (nm) bits.
When reading the compressed waveform data from the waveform memory 21 in which the compressed waveform data is stored in this way, the read & cash unit 20d is supplied from the WAG 20c for each sampling period in each of the sound channel Lch and Rch time slots. Each time the waveform address advances by k, a memory address that advances by one is generated, and the waveform memory 21 is accessed by the memory address to read n-bit data, and from the read n-bit data (n−m ) Extract bit data and output sub information including compression information. Also, m-bit data is extracted from the read n-bit data and temporarily stored in the cache memory. Then, the DEC 20e accesses the cache memory by the waveform address in the Lch and Rch time slots of the sound generation unit, sequentially reads the k sample compressed waveform data, and converts the compression information and the compressed waveform data sample to be sequentially read out. Based on the decompression processing in the DEC 20e, the original waveform data sample is restored.
Note that since the compression information used for decompression is the compression information stored in the previous frame read out, the initial stage of the decoder supplied from the sound source register 23 when there is no compression information. Decompressed using the value. For details regarding the compressed waveform data, refer to Japanese Patent No. 3912304.

The latest two samples of the waveform data of the sound units Lch and Rch that are output linearly from the DEC 20e or are expanded and output are INT20f in each time slot of the sound units Lch and Rch. To be supplied. The INT 20f is an Lch and Rch time slot of the sounding unit, and interpolates between the expanded waveform data of each 2 samples in the Lch and Rch based on the fractional part of the sounding unit phase supplied from the WAG 20c. By processing, the interpolated Lch and Rch waveform data are obtained. When four-point interpolation is performed in the INT 20f, the latest four samples of Lch and Rch are supplied from the DEC 20e, and interpolation processing is performed based on the expanded waveform data of the four samples and the fractional part of the phase. The interpolated Lch and Rch waveform data are obtained.
The interpolated Lch and Rch sounding unit waveform data output from the INT 20f is supplied to a digitally controlled filter (DCF) 20g to attenuate the high frequency components. The DCF 20g is supplied with modulation data from the LFO 20m (an LFO waveform whose amplitude is adjusted by FM depth) and a filter EG waveform from the filter envelope generation unit (FEG) 20n in the Lch and Rch time slots of the sound generation unit. Then, the cut-off frequency and Q (selectivity) of the DCF 20g are controlled according to these two data, and the high frequency components of the Lch waveform data and the Rch waveform data are attenuated. The FEG 20n is supplied with FEG parameters (FEGPs) including the FEG rate and FEG level for each state, which are common to the sound generation units of the Lch and Rch, from the sound source register 23 in the time slot of the sound generation unit. Based on the above, a filter EG waveform having a plurality of states common to the sound generation units of Lch and Rch and having a value that changes at each sampling period is generated and supplied to the DCF 20g. The generation of the filter EG waveform is started with a sound generation instruction (GT) supplied from the sound source register 23 as a trigger.

  The signal-processed Lch waveform data and Rch waveform data output from the DCF 20g are supplied to a digitally controlled amplifier (DCA) 20h to control the volume of each of the Lch and Rch. The DCA 20h includes modulation data from the LFO 20m (LFO waveform whose amplitude is adjusted by AM depth) and a volume EG waveform from the volume envelope generator (AEG) 20p in the Lch and Rch time slots of the sound generation unit, and Pan data (PANs) from the sound source register 23 for localizing the sound image is supplied, and the volume of the Lch waveform data is controlled according to the modulation data, the volume EG waveform, and the Lch pan data (PAN (L)) from the LFO 20m. Is output to the Lch side of the accumulation unit (MIX) 20i, and the volume of the Rch waveform data is controlled according to the modulation data from the LFO 20m, the volume EG waveform, and the Rch pan data (PAN (R)), and the MIX 20i is controlled. Is output to the Rch side. In this way, from the DCA 20h, the Lch waveform data (Lch musical tone data) and the Rch waveform data (Rch musical tone data) whose sound image positions are localized by pan data (PANs) are the time slots of the sound generation unit. It is output to the Lch side and Rch side of the MIX 20i. The AEG 20p is supplied with AEG parameters (AEGPs) including AEG rate and AEG level for each state, which are common to the sound source units of the Lch and Rch, from the sound source register 23 in the time slot of the sound generation unit. Based on the above, a volume EG waveform that is common to the sound generation units of Lch and Rch in a plurality of states and that changes in value every sampling cycle is generated and supplied to the DCA 20h. The generation of the volume EG waveform is started with a sound generation instruction (GT) supplied from the sound source register 23 as a trigger.

  Within one DAC period, from the DCA 20h, waveform data samples are output to the Lch side and the Rch side of the accumulation unit (MIX) 20i in time slots of a plurality of sound generation units. In the MIX 20i, the waveform data sample supplied to the Lch side and the waveform data sample supplied to the Rch side are separately accumulated on the Lch side and the Rch side, respectively, in the time slot of each sound generation unit of one DAC period. It is calculated. From the MIX 20i, stereo waveform data composed of Lch and Rch samples, which is an accumulation result for all sound generation units, is output to the DAC 20j for each DAC period. The DAC 20j converts the supplied stereo waveform data into an analog stereo waveform signal for each DAC period, and supplies the converted analog Lch and Rch stereo waveform signals to the stereo sound system 22. The sound system 22 emits the stereo waveform signal.

  The LFO 20m is supplied with LFO parameters (LFOPs) of the LFO frequency, PM depth, FM depth, and AM depth common to the sounding units of Lch and Rch from the sound source register 23 in the time slot of the sounding unit. A common LFO waveform is generated for the Lch and Rch sound units of the frequency indicated by the frequency, and the amplitude is controlled according to PM depth, FM depth, and AM depth, and supplied to the FG 20a, DCF 20g, and DCA 20h. . The generation of the LFO waveform is started with a sound generation instruction (GT) supplied from the sound source register 23 as a trigger. By supplying this LFO waveform to the FG 20a, vibrato can be added to the musical sound signal, and by supplying this LFO waveform to the DCF 20g, wah wah can be added to the musical sound signal, and this LFO waveform is supplied to the DCA 20h. Thus, tremolo can be added to the musical sound signal.

  As described above, the phase generation unit 111 in the configuration shown in FIG. 1 corresponds to the FG 20a and PG 20b, the waveform reading unit 112 corresponds to the WAG 20c, the reading & cashing unit 20d, the DEC 20e, and INT 20f, and the characteristic control unit 113 includes the DCF 20g, The control waveform generator 116 corresponds to DCA 20h, and corresponds to PEG 20k, LFO 20m, FEG 20n, and AEG 20p that generate a control waveform that changes with time during sound generation surrounded by a broken line. The control unit 122 corresponds to the CPU 10.

  The above description of the operation of each part of the sound source unit 20 describes the operation in the case of stereo sounding, paying attention to the time slot of the sounding unit whose mode flag (Mode) is set to the stereo sounding mode. In the waveform memory 21, monaural waveform data and stereo waveform data pairs are recorded for each tone color and each tone range. When a tone is generated in a tone generation unit of the tone generator 1, the waveform memory 21 stores the tone data. When stereo waveform data is selected as the waveform data to be read, a mode flag (Mode) indicating the stereo sound generation mode is set for the sound generation unit, and a stereo tone is generated in the time slot of the sound generation unit. Become so. Further, when a musical tone is generated in a certain tone generation unit of the tone generator 1, if a monaural sound waveform is selected as the waveform data to be read from the waveform memory 21, a mode flag indicating the monaural tone generation mode for that tone generation unit is selected. (Mode) is set, and a monaural musical sound is generated in the time slot of the sounding unit. The waveform data read from the waveform memory 21 is waveform data selected in accordance with the set tone color, note number, velocity, etc., and whether or not the tone color is such that the stereo effect can be understood. In consideration of various requirements such as the sound range, playing frequency of intensity, restriction due to the total capacity of the waveform memory, etc., they are prepared as stereo waveform data pairs or monaural waveform data, respectively.

The operation of each unit of the sound source unit 20 will be described with respect to the sound generation unit (Unit) in which the mode flag (Mode) indicating the monaural sound generation mode is set.
When a new note-on event is detected in the tone generator 1, the CPU 10 assigns a new note-on to an empty sounding unit by performing a sounding assignment process, and assigns it to the assigned sounding unit area in the tone generator register 23. Various new note-on parameters shown in FIGS. 3A and 3B are set. In this case, the mode flag (Mode) set in the tone generator register 23 is set to the monaural sound generation mode, but a 2ch sounding unit is assigned even in the case of monaural sounding note-on. In this embodiment, monaural sound is generated using various Lch resources of the two channels of the sound generation unit 20, but monaural sound may be generated using various Rch resources. .
When various parameters are set in the region of the sounding unit assigned to the new note-on in the sound source register 23 and the sounding start is instructed to the sounding unit, the FG 20a performs the same as in the case of stereo sounding. The F number is generated in the time slot of the sound generation unit, and the PG 20b accumulates the F number for each sampling period and generates the read address in the time slot of the sound generation unit in the same manner as in the case of stereo sounding. A phase composed of an integer part and a decimal part is generated and output to the WAG 20c.

  The WAG 20c is an Lch time slot of the sound generation unit, and an Lch waveform address in the sound generation unit is generated based on the Lch waveform designation information and loop information in the waveform address parameters (WAPs) and the integer part of the phase. The The Lch waveform address (integer part) output from the WAG 20c is supplied to the read & cash unit 20d, and a sample of monaural waveform data is read from the waveform memory 21 using the Lch time slot of the sound generation unit. The read monaural waveform data is expanded in the Lch time slot of the sound generation unit in the DEC 20e, and the original monaural waveform data sample is output. The latest two samples of the monaural waveform data output from the DEC 20e are supplied to the INT 20f, and interpolation processing is performed based on the fractional portion of the phase of the sounding unit supplied from the WAG 20c in the Lch time slot of the sounding unit. By doing so, the interpolated monaural waveform data is obtained. The interpolated monaural sound unit waveform data output from the INT 20f is supplied to the DCF 20g, and in the DCF 20g, according to the modulation data from the LFO 20m and the filter EG waveform from the FEG 20n in the Lch time slot of the sound unit. The high-frequency component is attenuated at the cut-off frequency and Q (selectivity), and monaural waveform data in which the harmonic component is adjusted is output. PEG20k, FEG20n, AEG20p, and LFO20m each generate a pitch EG waveform, a filter EG waveform, an amplitude EG waveform, and an LFO waveform in the time slot of the sounding unit, as in the case of stereo sounding.

  The monaural waveform data output from the DCF 20g is supplied to a digitally controlled amplifier (DCA) 20h, and the volume of the monaural waveform data is controlled in the sound channel Lch and Rch time slots. The DCA 20h is supplied with modulation data from the LFO 20m, a volume EG waveform from the AEG 20p, and pan data (PANs) from the sound source register 23. In this case, the volume of the monaural waveform data is controlled according to the modulation data from the LFO 20m, the volume EG waveform, and the Lch pan data (PAN (L)) and output to the Lch side of the MIX 20i, and the modulation data and volume from the LFO 20m. The volume of the same monaural waveform data is controlled according to the EG waveform and the Rch pan data (PAN (R)) and output to the Rch side of the MIX 20i. In this way, from the DCA 20h, monaural waveform data whose sound image position is localized by pan data (PANs) is converted into Lch musical sound data and Rch musical sound data in the time slot of the corresponding sound generation unit as the Lch musical sound data and the Rch musical sound data. Output to the side.

In the MIX 20i, the monaural waveform data sample supplied to the Lch side in the time slot of the sounding unit is accumulated with the waveform data sample supplied to the Lch side in the time slot of the other sounding unit and supplied to the Rch side. The sampled monaural waveform data samples are accumulated with the waveform data samples supplied to the Rch side in the time slots of the other sounding units, and the Lch and Rch accumulated for all sounding units in one DAC period are accumulated. Waveform data samples are output to the DAC 20j every DAC period. The DAC 20j converts the supplied waveform data sample into an analog stereo waveform signal for each DAC period, and supplies the converted analog Lch and Rch waveform signals to the sound system 22. The sound system 22 emits a stereo waveform signal, which is a mixed sound of a stereo signal of a stereo sounding unit and a monaural sound signal of a monaural sound unit whose sound image is localized by pan data (PANs). Will be.
As described above, in the sound generation unit in which the monaural sound generation mode is set, only the 1ch processing of the 2ch resources of the WAG 20c, the reading & cashing unit 20d and the DEC 20e included in the sound generation unit is used. Will be left behind. Therefore, if all of these resources are diverted to read out monaural waveform data samples in one sampling period of the sounding unit, twice the number of samples can be read out and decoded in one sampling period. In, the pitch can be increased up to twice the normal pitch.

Next, FIG. 7 shows a flowchart of the note-off event process executed by the CPU 10 when the note-off instructing the start of the decay of the musical sound is detected in the musical sound generating apparatus 1 of the present invention. The note-off is a command (MIDI message) for instructing the start of the decay of the musical sound, and as a parameter, a part number indicating which part is the command and a note number indicating the pitch of the musical sound to be attenuated With. For example, when a key that has been pressed so far is released on the keyboard that is the operation element 13, a part number indicating the part that is controlled by the keyboard and a note-off including the note number of the key is generated. . Further, note-off may be received from an external device via the communication I / O 15.
When the note-off is detected, the CPU 10 starts the note-off event process shown in FIG. The note number is stored in the area NN. Next, in step S21, the tone generation unit in which the tone of the pitch indicated by the note number of the region NN is formed by the part indicated by the part number of the region PT from the tone generation units for which the tone generator 20 is generating musical tone data. When the processing for detecting (Unit) is performed and the corresponding sounding unit is detected, the unit number of the sounding unit is stored in the area DU secured in the work area of the RAM 12. In step S22, it is determined whether or not the corresponding sounding unit is detected in step S21. If it is determined that the sounding unit has been detected, the process proceeds to step S23. In step S23, “1” is written in the attenuation start flag RT (DU) of the sound source register 23 to instruct the start of release for the sounding unit indicated by the Unit number of the area DU. As a result, the note-off event process ends, and the PEG 20k, FEG 20n, and AEG 20p of the sound source unit 20 are generating in the time slot of the corresponding sound generation unit using the release start instruction (RT (DU) ← 1) as a trigger. The pitch EG waveform, the filter EG waveform, and the amplitude EG waveform are shifted to the release state. Since the amplitude EG waveform shifted to the release state gradually decreases from the current level to the zero level (−∞ dB) at the AEG rate of the release state supplied from the sound source register 23, the sound generation unit is thereby The volume of the Lch and Rch musical sound data being generated in the Lch and Rch time slots gradually attenuates toward the zero level. If it is determined in step S22 that no detection has been made, the note-off event process is terminated. In addition, when a sounding unit with sufficiently advanced attenuation is detected, the CPU 10 performs an opening process for releasing the sounding unit as the above-mentioned “empty sounding unit”.
The present embodiment is characterized in that substantially the same processing is performed in the attenuation start instruction processing in step S23 marked with “*” regardless of whether the stereo sound generation mode is the monaural sound generation mode.

Next, FIG. 8 shows a flowchart of a pitch bend event process executed by the CPU 10 when a pitch bend command is detected during generation of a musical sound in the musical sound generating apparatus 1 of the present invention. The pitch bend command is a command (MIDI message) for changing in real time the pitch of one or a plurality of musical sounds being generated in one part. As a parameter, a part number indicating which part the command is and a sound With a bend amount indicating a high change amount. For example, when a pitch bend wheel serving as the operator 13 is operated, a pitch bend command including a part number indicating a part controlled by the wheel and a bend amount corresponding to the operation amount is generated. In some cases, a pitch bend command is received from an external device via the communication I / O 15.
When the pitch bend command is detected, the pitch bend event process shown in FIG. 8 is started. In step S30, the part number associated with the pitch bend command is set in the area PT secured in the work area of the RAM 12, and the bend amount is set in the area PV. Store. Next, in step S31, the first sounding unit of the 128 sounding units (Unit) is designated, and the unit number of the designated sounding unit is stored in the area CU secured in the work area of the RAM 12. Then, in step S32, the sound generation unit forming the musical sound data of the part number stored in the region PT is detected backward from the Unit number stored in the region CU. The unit number of the pronunciation unit is stored in the area DU secured in the work area of the RAM 12. Then, it is determined in step S33 whether or not the sounding unit forming the musical sound data is detected. If it is determined that the sounding unit is detected, the process proceeds to step S34.

In step S34, the value of the pitch shift PS (DU) common to the Lch and Rch in the sound generation unit indicated by the Unit number of the area DU in the sound source register 23 is the note number of the musical sound being sounded in the sound generation unit and the area PV. It is changed based on the bend amount. When the process in step S34 is completed, the unit number of the area DU is incremented by 1 in step S35, the unit number is stored in the area CU, and the process returns to step S32. In step S32, a process of detecting a sounding unit that is forming the musical sound data of the part indicated by the part number of the region PT is performed backward from the unit number incremented by 1 stored in the region DU stored in the region CU. Further, if detected, the unit number of the pronunciation unit is stored in the area DU, and then the processing from step S33 is performed. By repeating the processing from step S32 to step S35, all the sounding units that form the musical tone data of the part indicated by the part number in the area PT are detected, and the Lch and the sounding units in those sounding units in the tone generator register 23 are detected. The value of the pitch shift PS (DU) common to Rch is changed based on the bend amount of the region PV. Then, when all the corresponding pronunciation units that form the musical tone of the part number stored in the area PT are detected and it is determined in step S33 that there is no corresponding pronunciation unit, the pitch bend event process ends. As a result of this processing, the pitch of all musical tone data of the part indicated by the part number of the pitch bend command being generated by the sound source 20 changes (bends) according to the bend amount of the pitch bend command.
The present embodiment is characterized in that substantially the same processing is performed regardless of whether the sound generation mode is the stereo sound mode or the monaural sound mode in the pitch change processing in step S34 to which “*” is added.

As described above, note-on event processing, note-off event processing, and pitch bend event processing use the sound generation unit for generating sound data regardless of whether the musical sound data to be controlled is sounded in mono or stereo. On the other hand, processing is performed. In monaural pronunciation, some resources of the pronunciation units are not used and it seems that there is a lot of waste at first glance, but in the current situation where there are many timbres that use high-quality stereo pronunciation, there are too many resources that are wasted in practice. The effect of reducing the load on the CPU 10 is much greater.
Further, in the tone generator unit 20, since the stereo sound mode or the monaural sound mode is set for each sound generation unit, the tone data in the stereo sound mode and the sound unit in the monaural sound mode are generated in parallel. It becomes like this.

In the musical sound generating apparatus of the present invention described above, compressed waveform data is also recorded in the waveform memory. However, the present invention is not limited to this, and only uncompressed waveform data may be recorded. good. In this case, the DEC 20e that performs the decompression process can be omitted. In addition, although the number of sound generation channels of the tone generator according to the present invention is 256 channels, the number of sound generation channels is not limited to this and may be more or less than 256 channels. In this case, the number of sounding units is half the number of sounding channels.
In each block having resources for 2 stereo channels such as WAG 20c, read & cache unit 20d, DEC 20e, INT 20f, DCF 20g, etc., monaural waveform data processing is performed using Lch resources in the time slot of the sound generation unit in the monaural sound generation mode. However, it may be changed so that monaural waveform data is processed using Rch resources.
In the truncation unit determination process, the sound generation unit to be truncated is determined based on the volume level of the Lch of the sound generation unit that generates the musical sound of the search target part. You may make it carry out based on the larger volume level. In addition, the sounding unit to be truncated is determined regardless of whether the sounding unit is before or after the start of the release, but the sounding unit that has already been released is determined as the sounding unit to be preferentially truncated. It may be. First, the search target part was narrowed down, and the truncation unit to be truncated was determined based on the volume level of the sound generation unit generating the musical sound of the narrowed part. The sounding unit to be truncated may be determined based on the volume of the sounding unit that generates the musical sound.
Further, the CPU 11 sets the pitch shift data (PS), which is the cent scale F number, in the tone generator register 23, but instead, may set the linear scale F number. . In this case, the F number generator 20a does not need to convert the F number set by the CPU 11 from a cent scale to a linear scale. However, since the configuration for synthesizing the F number from the CPU 11, the pitch EG from the PEG 20k, and the modulation data from the LFO 20m in the F number generating unit 20a is complicated, a linear scale F number is added to the sound source register 23. It is inappropriate as a design to memorize.
As an example of controlling the value of the control parameter in real time, an example of a pitch bend command has been shown, but real time control is not limited to this. For example, an expression command for controlling the volume in real time, a modulation depth command for controlling the modulation depth such as amplitude by LFO in real time, or a parameter change command for controlling the cutoff frequency in real time may be used.

DESCRIPTION OF SYMBOLS 1 Musical sound production | generation apparatus, 10 CPU, 11 Flash ROM, 12 RAM, 13 Operator, 14 Display, 15 Communication I / O, 16 Bus line, 20 Sound source part, 20a FG, 20b PG, 20c WAG, 20d Reading & Cashe, 20e DEC, 20f INT, 20g DCF, 20h DCA, 20i MIX, 20j DAC, 20k PEG, 20m LFO, 20n FEG, 20p AEG, 21 waveform memory, 22 sound system, 23 tone generator register, 111 phase generator, 112 waveform readout Unit, 113 characteristic control unit, 114 MIX, 115 DAC, 116 control waveform generation unit, 117a reception unit, 122 control unit

Claims (2)

A waveform memory storing a plurality of waveform data pairs each including stereo-sampled Lch waveform data and Rch waveform data, and mono-sampled mono waveform data;
The sound generation unit is composed of two sound generation channels that operate in the stereo sound mode or the monaural sound mode according to the mode flag, and has N (N is an integer of 1 or more) sound units. Each sounding unit that operates operates to read 2ch of waveform data from the waveform memory to generate a 2ch musical sound waveform, and each sounding unit that operates in the monaural sounding mode reads 1ch of waveform data from the waveform memory. A sound source unit for generating a 1ch musical sound waveform,
N tone color control data corresponding to N tone generation units are stored, and each tone color control data includes the mode flag, the first waveform designation information, the second waveform designation information, the F number, and the characteristic parameter. Including sound source register,
A controller that controls the operation of the sound source unit by writing parameters to the sound source register;
In response to a sound generation instruction command (note-on) for instructing the pronunciation of a new musical sound, the control unit
Assign any one of N sounding units to the generation of a musical sound that has been instructed to sound,
Judge whether the musical sound that is pronounced is stereo or monaural,
When the tone sound to be generated is stereo, tone color control data including a mode flag indicating a stereo tone generation mode, first waveform specifying information used for generating the tone, second waveform specifying information, F number, and characteristic parameters Is set in a storage area corresponding to the tone generation unit assigned in the tone generator register, and when the tone instructed for tone generation is monaural, a mode flag indicating the monaural tone generation mode is used to generate the tone. Forming timbre control data including one waveform designation information, F number, and characteristic parameter, and setting it in a storage area corresponding to the tone generation unit assigned in the tone generator register;
For the assigned pronunciation unit, perform the process to instruct the start of pronunciation,
In response to a sounding start instruction for one sounding unit from the control unit,
If the mode flag of the sounding unit stored in the sound source register indicates the stereo sounding mode, Lch (or Rch) waveform data indicated by the first waveform designation information of the sounding unit set in the sound source register; The Rch (or Lch) waveform data indicated by the second waveform designation information of the sounding unit is read from the waveform memory at the speed indicated by the F number of the sounding unit, and the read Lch waveform data and Rch waveform data are read. Are controlled based on the characteristic control parameter of the sound generation unit set in the tone generator register to form an Lch tone signal and an Rch tone signal,
If the mode flag of the sounding unit stored in the sound source register indicates the monaural sounding mode, the mono waveform data indicated by the first waveform designation information of the sounding unit stored in the sound source register is stored in the sounding unit. A monophonic tone signal is formed by reading out the tone characteristics of the mono waveform data read from the waveform memory at the speed indicated by the F number, based on the characteristic control parameters of the sound generation unit set in the tone generator register. A musical sound generating apparatus characterized by performing processing to perform. In response to a control command indicating the F number of a musical tone being generated or the value of a control parameter, the control unit
Identify the sound generation unit that is generating the musical tone to be controlled by the control command,
A process of changing the value of the F number or control parameter of the specified tone generation unit of the sound source register to a value instructed by the control instruction;
The sound source unit can be changed by changing the F number or control parameter value by the control unit.
If the mode flag of the sounding unit stored in the tone generator register indicates the stereo sounding mode, each reading speed when reading the Lch waveform data and the Rch waveform data from the waveform memory is controlled by the changed F number. Alternatively, the control of each tone characteristic of the read Lch waveform data and Rch waveform data is controlled by the changed characteristic parameter,
If the mode flag of the sounding unit stored in the tone generator register indicates the monaural sounding mode, the mono waveform controlled or read by the changed F number at the reading speed when reading mono waveform data from the waveform memory 2. The musical tone generation apparatus according to claim 1, wherein a process for controlling the musical tone characteristic of the data is performed according to the changed characteristic parameter.

Images