JP2010217476A - Waveform data generating method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To separate and obtain low-speed attenuation waveform data and high-speed attenuation waveform data from original waveform data representing waveform of musical instrument sound. <P>SOLUTION: A plurality of peaks corresponding to harmonic components are detected from a spectral distribution obtained by performing Fourier transform of the whole original waveform data collectively, and a plurality of spectrum components corresponding to the plurality of peaks are specified. The plurality of specified spectrum components and spectrum components adjacent thereto from the spectrum components obtained by performing the collective Fourier transform are extracted and subjected to inverse Fourier transform to obtain the low-speed attenuation waveform data. The high-speed attenuation waveform data is obtained by subtracting the low-speed attenuation waveform data from the original waveform data. Also, the low-speed attenuation component waveform data may be obtained by detecting a peak locus with slow attenuation from a plurality of peak loci of the spectrum components over the whole original waveform and using the detected peak locus. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a waveform data generation method for generating waveform data used for generating a musical tone signal in a musical tone signal generating apparatus applied to an electronic musical instrument, a sequencer, and the like.

  Conventionally, for example, as shown in Patent Document 1 below, waveform data representing a decaying musical tone waveform is stored in a memory for each key pitch, and in response to a key depression, the corresponding key is depressed. 2. Description of the Related Art Music signal generators that read waveform data over time and generate music signals are known. In this musical tone signal generator, the waveform data read position is varied according to the key touch. That is, when the key touch is strong, the waveform data is read from the head, and the reading position is shifted backward as the key touch becomes weak, so that a tone signal having an amplitude corresponding to the key touch is generated. Further, in this Patent Document 1, also for percussion instrument sounds, waveform data representing the musical sound waveform is stored in a memory, and the reading position of the waveform data is made different according to the impact strength in the same manner as described above. It is also shown that a musical sound signal corresponding to the hitting intensity is generated.

Japanese Patent Publication No. 3-5758

  In general, attenuation-type instrument sounds and percussion instrument sounds having various pitches have subtle differences in musical sound waveforms that change with time if the performance sound intensity (key touch strength or impact strength) differs. However, in the above-described prior art, although the reading position differs with respect to the intensity of the generated musical sound, the same waveform data is read out, so that a realistic musical sound signal can be generated according to the intensity of the generated musical sound. Can not. In order to solve this problem, a plurality of waveform data respectively representing a plurality of musical sound waveforms having different strengths may be stored in a memory, and the waveform data to be read out may be switched according to the strength of the generated musical sound. . However, when a plurality of waveform data having different strengths are prepared, there is a problem that the amount of waveform data increases and a large capacity memory is required. In particular, in the case of an instrument sound having a large number of pitches constituting a scale, it is necessary to prepare waveform data for each pitch or for each predetermined pitch range, and the amount of waveform data is extremely large.

  In view of such circumstances, the present inventors have discovered the following as a result of conducting a number of experiments. As shown in FIG. 6A, the decaying musical tone waveform immediately after the start of sound generation includes a frequency component and a noise component (anharmonic component) other than the fundamental component and the harmonic component in addition to the fundamental component and the harmonic component. The waveform itself fluctuates with time and is not stable. On the other hand, when time elapses from the start of sound generation, that is, after the stable point, as shown in FIG. 6B, the musical tone waveform of the attenuation system is almost composed of fundamental wave components and harmonic components and becomes stable. Based on this, the present inventors, as shown in FIG. 7, divide the musical sound waveform of the attenuation system into a component waveform (fast decay component waveform) with fast decay and a component waveform (slow decay component waveform) with slow decay, We found that the high-speed decay component waveform changes according to the intensity of the generated musical sound, but the low-speed attenuation component waveform hardly changes even if the intensity of the generated musical sound is different. FIG. 7A schematically shows the entire original tone waveform, FIG. 7B schematically shows the fast decay component waveform, and FIG. 7C schematically shows the slow decay component waveform. Yes. In our experiments, we have found that this tendency is strong for musical sound waveforms such as electric pianos, toms, and timpani.

  Then, the inventors of the present invention provide fast decay waveform data composed of a plurality of waveform data extracted from a performance sound played at a plurality of levels of intensity and one waveform data extracted from a performance sound played at a high intensity. Is stored in a waveform memory, and at least one fast decay waveform data is selected from a plurality of fast decay waveform data according to the strength of the generated musical sound and read out. A musical tone signal generator has been developed that reads out low-speed decay waveform data and mixes and outputs the read out high-speed decay waveform data and low-speed decay waveform data. Further, in reading out the low-speed decay waveform data in this musical tone signal generator, the generated musical tone is read by reading the address farther from the top address of the waveform data as the reading start address with the passage of time as the strength of the generated musical tone becomes weaker. The time from the start to the end of the musical sound becomes shorter as the intensity of the instrument becomes weaker. According to this, a plurality of fast decay waveform data that represents a fast decay component waveform that ends in a short time from the start of generation, and one slow decay waveform data that represents a slow decay component waveform that ends after a long time has elapsed since the start of generation, Only by storing a small amount of waveform data consisting of the above in the waveform memory, a real tone signal can be generated according to the strength of the tone to be generated.

  An object of the present invention is to provide a waveform data generation method suitable for generating waveform data used in the musical tone signal generator as described above.

  In order to achieve the above object, a waveform data generation method according to a first aspect of the present invention includes a step of collectively Fourier-transforming the entire original waveform data representing an instrument sound waveform to generate a spectrum component of the entire instrument sound waveform. (Step S104) and detecting a plurality of peaks corresponding to a fundamental wave component and a harmonic component of the instrument sound waveform from a spectrum distribution represented by a spectrum component of the entire generated instrument sound waveform, A step of identifying a plurality of spectral components corresponding to a peak (step S106), and a spectrum relating to the plurality of identified spectral components and spectral components in the vicinity thereof among the spectral components of the generated instrument sound waveform as a whole. A component extracting step (steps S108 and S110) and an inverse Fourier transform of the extracted spectral component; A step of generating low-speed attenuation waveform data representing a low-speed attenuation component waveform (step S112), and subtracting the generated low-speed attenuation waveform data from the original waveform data to obtain high-speed attenuation waveform data representing a high-speed attenuation component waveform. Generating (step SS116).

  According to the first invention, from the original waveform data representing the original waveform, it is possible to obtain the slow decay waveform data representing the slow decay component waveform corresponding to the harmonic component composed of the fundamental wave component and the harmonic component, and It is possible to acquire fast decay waveform data representing a fast decay component waveform corresponding to an anharmonic component not included in the harmonic component. Also, in acquiring the slow decay waveform data, the fundamental wave component and the overtone component of the instrument sound waveform are obtained from the spectrum distribution of the entire instrument sound waveform obtained by performing Fourier transform on the entire original waveform data representing the instrument sound waveform. A plurality of spectral components corresponding to a plurality of peaks were identified by detecting a plurality of peaks corresponding to Since attenuation waveform data representing a harmonic component is acquired using the plurality of specified spectral components, low-speed attenuation waveform data can be generated in a relatively simple and short time.

  The waveform data generation method according to the second invention sequentially Fourier transforms the original waveform data for each frame defined by the time window while moving the time window with respect to the original waveform data representing the instrument sound waveform. A plurality of peak trajectories connecting the plurality of peaks of the acquired spectral components over the plurality of frames, respectively, and obtaining the spectral components of the plurality of frames over the whole of the original waveform data (steps S204 to S212). (S214), a step (S216) of extracting a slow-decaying peak trajectory from the detected plurality of peak trajectories, and an inverse Fourier transform of the spectral components related to the extracted peak trajectory. A step of generating slow decay waveform data representing a slow decay component waveform (step S218); By subtracting the slow decay waveform data said generated from the shape data, and a step (S222) for generating a fast decay waveform data representative of the fast decay component waveform.

  According to the second aspect of the invention, the slow decay peak locus corresponds to the harmonic component composed of the fundamental wave component and the harmonic component of the original waveform, so that the low-speed decay component corresponding to the harmonic component is obtained from the original waveform data representing the original waveform. Low-speed decay waveform data representing a waveform can be acquired, and fast decay waveform data representing a fast decay component waveform corresponding to an anharmonic component not included in the harmonic component can be obtained. Further, in this case, by detecting a plurality of peak trajectories respectively connecting a plurality of peaks of spectral components acquired over a plurality of frames, by extracting a peak trajectory with slow decay from the detected plurality of peak trajectories, Since the spectral component corresponding to the harmonic component is acquired, the processing is complicated and the processing time is relatively long as compared with the first invention, but the slow decay waveform data is generated with high accuracy. Can do.

  Furthermore, if the waveform data generation method as in the first and second inventions is applied to a plurality of original waveforms having different strengths, which are related to one type of musical instrument, the strength of one type of musical instrument is reduced. A plurality of sets of high-speed attenuation waveform data and low-speed attenuation waveform data corresponding to a plurality of different instrument sound waveforms can be prepared. Then, by selecting a plurality of high-speed attenuation waveform data and one low-speed attenuation waveform data from a plurality of sets of prepared high-speed attenuation waveform data and low-speed attenuation waveform data as required, It can be used to generate musical sound signals.

It is a whole block diagram of a waveform data generation device. It is a flowchart which shows the waveform data generation program for the 1st waveform data generation method. (A) is a figure which shows roughly the result of carrying out FFT processing of the original waveform data collectively, (B) is a figure which shows an extraction window notionally, (C) is a figure showing the component extracted. FIG. It is a flowchart which shows the waveform data generation program for the 2nd waveform data generation method. (A) is a figure which shows roughly the peak locus | trajectory of an original waveform, (B) is a figure which shows roughly the peak locus | trajectory of a slow decay component waveform. (A) is a diagram schematically showing a decaying tone waveform immediately after the start of sound generation, and (B) is a diagram schematically showing a decaying tone waveform in a stable state over time from the start of tone generation. (A) is a diagram schematically showing the whole attenuation-type musical sound waveform, (B) is a diagram schematically showing a high-speed attenuation component waveform in the musical sound waveform, and (C) is the musical sound waveform. It is a figure which shows the inside low-speed decay component waveform in simulation. It is the whole block diagram of the electronic musical instrument to which the musical tone signal generator which generates a musical tone signal using the generated waveform data was applied. FIG. 9 is a block diagram showing a specific configuration of the tone generator circuit of FIG. 8. It is a memory map which shows the structure of a waveform memory. It is a memory map which shows the structure of a timbre parameter memory. (A) is the schematic showing the waveform of the several high-speed decay waveform data memorize | stored in the waveform memory, (B) is the schematic showing the waveform of one slow decay waveform data memorize | stored in the waveform memory. is there. It is a flowchart which shows a note-on event processing program. It is a flowchart which shows another note-on event processing program.

  Hereinafter, the present invention will be described. First, a waveform data generation method according to the present invention will be described, and then a tone signal generator for generating a tone signal using the waveform data generated by the waveform data generation method. Will be described.

a1. Waveform Data Generation Device First, a waveform data generation device used for the waveform data generation method of the present invention will be described. As shown in FIG. 1, the waveform data generation apparatus includes a plurality of panel switches 11, a display 12, a waveform memory 13, a writing circuit 14, a buffer circuit 15, a sound source circuit 16, and an access management circuit 17.

  The plurality of panel switches 11 are provided on the operation panel, and are operated by an operator to instruct the operation of the waveform data generation device. The display device 12 is composed of a liquid crystal display provided on the operation panel, and displays characters, numbers, figures, particularly musical sound waveform and waveform analysis results. The panel switch 11 and the display device 12 are connected to the bus 20.

  The waveform memory 13 is composed of a writable and readable memory, and stores waveform data representing an original waveform (instrument sound waveform) and waveform data generated in the waveform data generating apparatus. The writing circuit 14 controls writing of waveform data to the waveform memory 13. The writing circuit 14 is connected to an input terminal 14a for inputting waveform data representing the instrument sound waveform obtained by sampling various instrument sound waveforms at a predetermined sampling rate and A / D converting. The buffer circuit 15 controls the transfer of waveform data from the waveform memory 13 to other circuits and the transfer of waveform data from other circuits to the waveform memory 13. The tone generator circuit 16 generates a digital tone signal using the waveform data read from the waveform memory 13 and outputs the generated digital tone signal to the sound system 18. The sound system 18 includes a D / A converter, an analog amplifier, and a speaker, and emits a musical sound corresponding to the digital musical sound signal supplied from the sound source circuit 16. These writing circuit 14, buffer circuit 15 and tone generator circuit 16 are also connected to the bus 20. The access management circuit 17 is connected between the waveform memory 13, the write circuit 14, the buffer circuit 15, and the sound source circuit 16, writes waveform data into the waveform memory 13 by the write circuit 14, and the waveform memory 13 by the buffer circuit 15. The access time slot for the waveform memory 13 is managed so that the transfer of the waveform data with respect to and the readout of the waveform data from the waveform memory 13 by the sound source circuit 16 do not collide.

  The waveform data generation apparatus also includes a CPU 31, a ROM 32, a RAM 33, a timer 34, a drive circuit 35, and an external interface circuit 36 connected to the bus 20. The CPU 31, the ROM 32, the RAM 33, and the timer 34 constitute a computer main body. In particular, the CPU 31 executes a waveform data generation program to be described later. The drive circuit 35 controls the storage and reading of various data and programs with respect to the external recording device 37 such as the hard disk HD, the flash memory, and the compact disk CD. The external interface circuit 36 can be connected to an external MIDI device such as an electronic musical instrument or a sequencer, and can be connected to a server via a communication network. The waveform data generation program described above is stored in the external recording device 37 or taken into the RAM 33 or the external recording device 37 via the external interface circuit 36. Hereinafter, the first and second waveform data generation methods using this waveform data generation apparatus will be described.

a2. First Waveform Data Generation Method A first waveform data generation method will be described. First, an operator prepares digital waveform data (hereinafter, this digital waveform data is referred to as original waveform data) representing an instrumental sound waveform of an attenuation system having a desired instrument type, a desired pitch, and a desired intensity. However, in the case of percussion instrument sounds, original waveform data having a desired instrument type and a desired strength is prepared. This original waveform data represents an instrument sound waveform from the start of sound generation to the end of sound generation of one instrument sound. In preparing the original waveform data, a device that stores the original waveform data in advance is connected to the input terminal 14a. Further, the original waveform data may be recorded in advance in the external recording device 37, or the original waveform data may be taken in from the external interface circuit 36.

  After preparing the original waveform data, the operator operates the panel switch 11 to instruct the start of execution of the waveform data generation program of FIG. The execution of this waveform data generation program is started in step S100, and the CPU 31 controls the writing circuit 14 to write the desired original waveform data in the waveform memory 13 in step S102. In this writing, the original waveform data input through the input terminal 14 a and the external interface circuit 36 as described above or the original waveform data recorded in the external recording device 37 is written into the waveform memory 13 by the writing circuit 14. It is.

  After the process of step S102, in step S104, the CPU 31 collectively performs fast Fourier transform processing (hereinafter simply referred to as FFT processing) on the original waveform data over the entire range from the start of sound generation to the end of sound generation. As a result of the collective FFT processing, information (spectral component) according to the time lapse of frequency, amplitude and phase, that is, information (spectral component) including time change of frequency, amplitude and phase is extracted from the original waveform data over the entire range. Is done. Next, in step S106, the CPU 31 uses a plurality of spectral components corresponding to a plurality of peaks whose amplitude values exceed a predetermined value using spectral components over the entire range of the original waveform data obtained by the FFT process (step S106). Frequency component) is detected. In this case, the frequency and amplitude information over the entire range of the original waveform data is used for all the information (spectral components) obtained from the FFT result, and the original waveform data over the entire range is used regardless of the passage of time. A spectral distribution is derived. Thereafter, a plurality of spectrum components (frequency components) respectively corresponding to a plurality of peaks whose amplitude values exceed a predetermined value are detected using the derived spectrum distribution. FIG. 3A shows the derived spectral distribution. The peak part A in the figure corresponds to the fundamental frequency and harmonic frequency (harmonic component) of the musical instrument sound, and the lower part B of the peak part A corresponds to an anharmonic component including noise whose frequency is not stable. As is also considered from the spectrum distribution of FIG. 3A, the predetermined value used for detecting the plurality of peaks (plural spectrum components) has a magnitude exceeding the B portion of FIG. It is preferable that the function value gradually decreases as the frequency increases.

  Next, in step S108, the CPU 31 generates a window function representing a plurality of extraction windows having a predetermined small width, each centered on a spectrum component (frequency component) corresponding to the detected plurality of peaks. This window function is for extracting the plurality of spectral components and spectral components in the vicinity thereof, and defines a plurality of frequency regions having a predetermined width on the frequency axis. FIG. 3B shows this window function.

  After the process of step S108, the CPU 31 generates the window generated by the process of step S108 from all the information (spectral components) regarding the original waveform data obtained by the FFT process of step S104 in step S110. Extract the spectral components defined by the function. That is, spectral components related to frequencies belonging to a plurality of frequency regions defined by the window function are extracted from all spectral components related to the original waveform data. Next, in step S112, the CPU 31 performs inverse fast Fourier transform processing (hereinafter simply referred to as inverse FFT processing) on the extracted spectral component to add and synthesize waveform data representing a slow decay component waveform. In step S114, the CPU 31 controls the writing circuit 54 to write the added combined waveform data into the waveform memory 13 as low-speed decay waveform data.

  After the process of step S114, the CPU 31 changes the slow decay waveform data newly stored in the waveform memory 13 from the original waveform data stored in the waveform memory 13 from the beginning to the end of both waveform data in step S116. Subtract sequentially. This subtraction result corresponds to the fast decay waveform data of the present invention obtained by removing the slow decay waveform data from the original waveform data. Next, in step S118, the CPU 31 writes the waveform data as the subtraction result into the waveform memory 13 as fast decay waveform data, and in step S120, the execution of this waveform data generation program is terminated. As a result, a set of slow decay waveform data and fast decay waveform data is stored in the waveform memory 13.

  After a pair of low-speed decay waveform data and high-speed decay waveform data is stored in the waveform memory 13 in this way, the same instrument type and pitch (in the case of percussion instruments, the same instrument type as described above) 2 is prepared, and the waveform data generation program of FIG. 2 described above is executed again. As a result, a set of low-speed attenuation waveform data and high-speed attenuation that represent the same instrument type and pitch (in the case of percussion instruments, the same instrument type as described above), and represent an instrument sound waveform having a strength different from that described above. Waveform data is stored in the waveform memory 13. By repeating such a process, a set of slow decays representing a musical instrument sound waveform of one instrument type and pitch (one instrument type in the case of percussion instruments) with a strength different from that described above. Waveform data and fast decay waveform data are sequentially stored in the waveform memory 13. For example, when used in a musical sound signal generator (electronic musical instrument) described later, the key touch strength VEL and the operator touch strength VEL are “127”, “80”, “48”, “16” (or Low-speed attenuation waveform data and high-speed attenuation waveform data corresponding to “127”, “85”, “43”, “1”) are sequentially stored in the waveform memory 13. Note that waveform data necessary for tone synthesis in a tone signal generator to be described later includes a plurality of key touch strengths VEL or operator touch strengths VEL (for example, “127”, “80”, “48”, “16”). Or a plurality of fast decay waveform data corresponding to “127”, “85”, “43”, “1”) and the largest key touch strength VEL or operator touch strength VEL (for example, “127”). Since it is one low-speed decay waveform data, other waveform data may be deleted from the waveform memory 13 or may not be stored in the waveform memory 13 during the storage processing in steps S122 and S126.

  Furthermore, after storing in the waveform memory 13 a waveform set consisting of a plurality of fast decay waveform data and at least one slow decay waveform data related to the one instrument type and pitch, the same instrument type and different pitches are used. The waveform set is stored in the waveform memory 13 by the same processing as described above. Then, after storing the waveform set of the waveform set for all pitches or all ranges, the waveform set for different instrument types is stored in the waveform memory 13 by the same processing as described above. As a result, by such waveform data generation processing, the waveform memory 13 prepares a waveform set related to a desired plurality of instrument types, necessary pitches, and a plurality of attenuation-type instrument sounds in the waveform memory 13. Will be.

  According to this first waveform data generation method, it is possible to generate a plurality of fast decay waveform data and one slow decay waveform data from the original waveform data in a relatively simple and short time. This first waveform data generation method can be used to generate a plurality of high-speed attenuation waveform data and one low-speed attenuation waveform data related to various instrument sounds of attenuation systems. It is most suitable for generating one fast decay waveform data and one slow decay waveform data.

a3. Second Waveform Data Generation Method Next, a second waveform data generation method will be described. In this case as well, the operator, in the same way as in the first waveform data generation method, is the original waveform data of the attenuation system having the desired instrument type, desired pitch, and desired intensity (in the case of percussion instrument sounds). For the desired instrument type and original waveform data of desired strength). Then, the operator instructs the start of execution of the waveform data generation program of FIG. 4 by operating the panel switch 11. The execution of the waveform data generation program is started in step S200, and the CPU 31 controls the write circuit 14 in step S202 to control the writing circuit 14 to store the original waveform data in the waveform memory, as in the case of the first waveform data generation method. Write to 13.

  After the process of step S202, in step S204, the CPU 31 sets one time window from the beginning of the original waveform data, and extracts the original waveform data for the first frame included in this time window. The width of this time window is, for example, about 8 times the period of the fundamental frequency component. Next, in step S206, the CPU 31 performs FFT processing on the extracted original waveform data for one frame to acquire spectrum information for one frame. This spectrum information includes three pieces of information on frequency, amplitude, and phase regarding the original waveform data for one frame. Next, in step S208, the CPU 31 performs spectrum analysis on the spectrum information, detects a peak of the spectrum distribution, and detects three pieces of information on the frequency, amplitude, and phase of each detected peak. Next, in step S210, the CPU 31 determines whether the end of the original waveform data, that is, the extraction of the last one frame of waveform data has been completed. If the extraction of the waveform data for the last frame has not been completed, the CPU 31 determines “No” in step S210, moves the time window in step S212, and acquires the waveform data for the next frame. It takes out and returns to step S206. The moving time of this time window is about 1/8 of the period of the fundamental frequency component, for example. Then, the CPU 31 executes the above-described processing of steps S206 and S208 to detect a plurality of peaks related to the original waveform data for the next one frame, and to obtain three pieces of information on the frequency, amplitude and phase related to the plurality of peaks. Detect each. By repeatedly executing the cyclic processing composed of steps S206 to S212, the frequency, amplitude, and phase relating to a plurality of peaks of the original waveform data are obtained for each frame over a plurality of frames from the beginning to the end of the original waveform data. Three pieces of information to be represented are respectively acquired.

  When the extraction of the last one frame of waveform data is completed, the CPU 31 determines “Yes” in step S210 and detects a plurality of peak loci in step S214. In the detection of the peak trajectory, only the peaks in which the frequency, amplitude and phase relating to the peaks of adjacent frames are smoothly connected are extracted. That is, the frequency, amplitude, and phase related to the peak of one adjacent frame are compared with the frequency, amplitude, and phase related to the peak of the other frame, and only the peak having the smallest change is extracted. The peaks in the previous frames are connected for each frequency. Thereby, noise and anharmonic components are removed. FIG. 5A schematically shows the peak locus with respect to the original waveform.

  After the process of step S214, the CPU 31 extracts a slow-decayed peak locus from the detected plurality of peak loci in step S216. Specifically, only the peak trajectory remaining beyond a predetermined time from the beginning of the waveform data and lasting for a predetermined time is extracted. FIG. 5B shows a state where only some of the plurality of peak loci are extracted from the plurality of peak loci shown in FIG. The peak locus shown in FIG. 5B is a peak locus of a slow decay component (harmonic component) that is slowly attenuated. In the extraction of the slowly decaying peak locus, instead of automatically detecting the peak locus by computer processing as described above, the peak locus shown in FIG. The peak locus with slow decay may be extracted by a selection operation by an operator using the panel switch 11.

  Next, in step S218, the CPU 31 performs inverse FFT processing on the spectrum information corresponding to the plurality of extracted peak trajectories for all frames extending from the beginning to the end of the original waveform data to generate a slow decay component waveform. Add and synthesize the waveform data to represent. In step S220, the CPU 31 controls the writing circuit 14 to write the combined waveform data into the waveform memory 13 as the slow decay waveform data. The above-described FFT processing for each frame, detection processing for a plurality of peaks, detection processing for a plurality of peak trajectories, and inverse FFT processing (sinusoidal wave synthesis processing) are disclosed in, for example, Japanese Patent Laid-Open Nos. 2000-10565 and 2000-. Since this is a well-known technique described in Japanese Patent Laid-Open No. 056774, Japanese Patent Laid-Open No. 2001-1000076, Japanese Patent Laid-Open No. 2003-263170, and the like, detailed description thereof is omitted.

  Next, in step S222, the CPU 31 sequentially subtracts the slow decay waveform data newly stored in the waveform memory 13 from the original waveform data stored in the waveform memory 13 from the beginning to the end of both waveform data. This subtraction result corresponds to the fast decay waveform data of the present invention obtained by removing the slow decay waveform data from the original waveform data. In step S224, the CPU 31 writes the waveform data as the subtraction result into the waveform memory 13 as the fast decay waveform data, and in step S226, the execution of the waveform data generation program ends. As a result, a set of slow decay waveform data and fast decay waveform data is stored in the waveform memory 13.

  After a set of low-speed decay waveform data and high-speed decay waveform data is stored in the waveform memory 13 in this way, as in the case of the first waveform data generation method, different instrument sounds and different pitches are used. The waveform data generation process described above is executed for instrument sounds and different instrument types, and a set of low-speed attenuations for a desired plurality of instrument types, required pitches, and a plurality of intensity attenuation instrument sounds. A waveform set including waveform data and fast decay waveform data is stored in the waveform memory 13. It should be noted that a plurality of fast decay waveform data corresponding to a plurality of key touch strengths VEL or operation element touch strengths VEL and one slow decay waveform data of the largest key touch strength VEL or operation element touch strengths VEL are waveformd. The remaining waveform data in the memory 13 may be deleted from the waveform memory 13 or may not be stored in the waveform memory 13 as in the case of the first waveform data generation method.

  In the second waveform data generation method, compared to the first waveform data generation method, the generation of the fast decay waveform data and the slow decay waveform data is complicated because of the peak locus detection process, It takes a long time to generate both waveform data. However, according to the second waveform data generation method, the high-speed decay waveform data and the low-speed decay waveform data can be generated with higher accuracy than the first waveform data generation method. This second waveform data generation method can also be used to generate a plurality of high-speed attenuation waveform data related to various attenuation-type instrument sounds and a single low-speed attenuation waveform data. It is most suitable for generating one fast decay waveform data and one slow decay waveform data.

b. A musical tone signal generator to which the generated waveform data is applied Hereinafter, a musical tone signal is generated using waveform data composed of the fast decay waveform data and the slow decay waveform data generated by the first and second waveform data generation methods. A musical tone signal generator that generates the sound signal will be described. FIG. 8 is an overall block diagram of an electronic musical instrument to which the musical tone signal generator is applied. This electronic musical instrument includes a keyboard 51, a plurality of performance operators 52, a plurality of panel operators 53, a display 54, and a tone generator circuit 55.

  The keyboard 51 is composed of a plurality of white keys and black keys which are operated by the performer and each designates the pitch of the generated musical sound and instructs the generation of the musical sound. The keyboard 51 also includes a key touch detection mechanism for detecting a key touch strength VEL such as a key pressing speed and a key pressing pressure. The key pressing operation and key touch strength VEL of each key on the keyboard 51 are detected by a detection circuit 51 a connected to the bus 56. The detection circuit 51a provides a key touch signal representing a key-on signal KON and a key-off signal KOF representing a key pressing / release operation of each key, a key number NN representing a key that has been depressed / released, and a key touch signal representing a key touch strength VEL. Output. The plurality of performance operators 52 are a plurality of switches respectively corresponding to a plurality of types of percussion instruments that are operated by the performer to instruct the generation of percussion instrument sounds. The plurality of performance operators 52 are also provided with an operator touch detection mechanism for detecting an operator touch strength VEL such as a pressing speed and a pressing pressure thereof. The pressing operation of the plurality of performance operators 52 and the operator touch strength VEL are detected by a detection circuit 52 a connected to the bus 56. The detection circuit 52a is a switch-on signal SWON indicating the pressing operation of each performance operator 52, instrument type information IN indicating the percussion instrument type corresponding to the pressed performance operator 52, and an operation indicating the operator touch strength VEL. The child touch signal is output to the bus 56.

  The plurality of panel operators 53 are composed of a plurality of switches, volumes, and the like provided on the operation panel of the electronic musical instrument, and are operated by the performer to instruct various operations of the electronic musical instrument including a musical sound signal generation mode. The operation of the plurality of panel operators 53 is detected by a detection circuit 53 a connected to the bus 56. The detection circuit 53 a outputs a detection signal indicating the operation of the plurality of panel operators 53 to the bus 56. The display unit 54 is composed of a liquid crystal display, CRT, or the like provided on the operation panel, and displays characters, numbers, figures, and the like. The display 54 is connected to a display circuit 54 a connected to the bus 56. The display circuit 54 a controls display on the display device 54 in accordance with image data supplied via the bus 56.

  The tone generator 55 is connected to the bus 56, generates a digital musical tone signal according to various control signals supplied via the bus 56, and outputs the digital musical tone signal to the effect circuit 57. As shown in FIG. 9, the tone generator circuit 55 includes a plurality of tone generation channels ch1, ch2,... Chn. The plurality of tone generation channels ch1, ch2,... Chn are configured in the same manner, and each channel performs processing for generating a digital musical tone signal for each predetermined sampling period, a readout circuit 55a, a digital control filter 55b and a digital control amplifier 55c. The read circuit 55a reads waveform data stored in a waveform memory WM, which will be described later, in accordance with control parameters (read start address, read end address, pitch shift amount, etc.) and a sound generation start instruction from the computer main body, which will be described later. The digital musical tone signal is generated, and the generated digital musical tone signal is output to the digital control filter 55b. Further, the read circuit 55a performs an interpolation operation on the waveform data read from the waveform memory WM as necessary, or decompresses the waveform data stored in the waveform memory WM if it is compressed.

  The digital control filter 55b is a digital signal composed of waveform data output from the readout circuit 55a in response to control parameters (filter control parameter group, etc.), sound generation start instruction, release start instruction, and rapid attenuation instruction from the computer main body, which will be described later. The frequency characteristic of the tone signal is controlled and output to the digital control amplifier 55c. The digital control amplifier 55c is composed of waveform data output from the digital control filter 55b in response to control parameters (amplitude control parameter group, etc.), a sounding start instruction, a release start instruction, and a rapid attenuation instruction from a computer main body to be described later. The amplitude characteristic (amplitude envelope) of the digital musical tone signal is controlled and output to the channel accumulation circuit 55d. The channel accumulating circuit 55d accumulates, ie, mixes and outputs digital musical tone signals from a plurality of tone generation channels ch1, ch2,... Chn.

  The effect circuit 57 gives effects such as chorus and reverb according to control parameters from a computer main body, which will be described later, to the digital musical tone signal output from the tone generator circuit 55 and outputs the result to the sound system 58. Even in the application of this effect, the effect circuit 57 processes the digital musical tone signal every predetermined sampling period. The sound system 58 includes a D / A converter, an analog amplifier, and a speaker, and emits a musical sound corresponding to the digital musical sound signal supplied from the effect circuit 57.

  The electronic musical instrument also includes a CPU 61, a ROM 62, a RAM 63, a timer 64, an external storage device 65, a MIDI interface circuit 66, and an external interface circuit 67 connected to the bus 56, respectively. The CPU 61, the ROM 62, the RAM 63, and the timer 64 constitute a computer main body. In particular, the CPU 61 executes a note-on event processing program shown in FIG. The external storage device 65 includes various recording media such as a hard disk HD and a flash memory that are pre-installed in the electronic musical instrument, a compact disc CD and a flexible disc FD that can be mounted on the electronic musical instrument, and a drive unit for each recording medium. It is possible to store and read various data and programs.

  The ROM 62 or the external storage device 65 is provided with a waveform memory WM storing waveform data, which will be described in detail later, and a timbre parameter memory PM storing timbre control parameters, and various data including automatic performance data and automatic rhythm data. And various programs including the note-on event processing program. Automatic performance data refers to key performance event data (key-on and key-off events including key touch strength VEL) and percussion instrument performance event data (operator touch strength VEL including operator touch strength VEL) related to percussion sound. Event) is stored in time series as the music progresses. Further, the automatic rhythm data is data in which percussion instrument performance event data (percussion instrument operation event including operator touch strength VEL) related to percussion instrument sounds is stored over a plurality of measures over time for each rhythm pattern type such as march and waltz. . These data and programs may be stored in advance in the ROM 62 or the external storage device 65, or may be taken in from the outside via the MIDI interface circuit 66 or the external interface circuit 67. An external MIDI device 71 such as another electronic musical instrument or a sequencer can be connected to the MIDI interface circuit 66. The external interface circuit 67 can be connected to the server 73 via the communication network 72. Then, the various data and programs described above are taken into the electronic musical instrument from the external MIDI device 71 or the server 73.

  Next, the waveform memory WM and the timbre parameter memory PM will be described. As shown in FIG. 10A, the waveform memory WM is divided into a plurality of storage areas respectively corresponding to a plurality of timbres. The plurality of timbres include a continuous tone, but since the continuous tone is not directly related to the present invention, only the decaying tone will be described in the following description. In each of the storage areas corresponding to timbres, a plurality of waveform sets made up of waveform data groups respectively corresponding to a plurality of keys (a plurality of key pitches) are stored. Each waveform set includes a plurality of fast decay waveform data and one slow decay waveform data.

  The high-speed attenuation waveform data is waveform data representing a high-speed attenuation component waveform over about 1 second obtained by extracting only the fast decay component waveform from the attenuation-type musical sound waveform generated by the first and second waveform data generation methods described above. . The plurality of high-speed decay waveform data respectively correspond to different strengths (key touch strength VEL) of the generated musical sound, and in this case, correspond to four different strengths. For example, if the key touch strength VEL is represented by “1” to “127”, the key touch strength VEL is “127” and “80” as shown in order from the top on the left side of FIG. , “48” and “16”, four pieces of waveform data representing the fast decay component waveforms extracted from the performance sounds are stored as a plurality of fast decay waveform data. The value “127” is the maximum value of the entire range “1” to “127” of the key touch strength VEL, and the values “80”, “48”, and “16” are the entire range of the key touch strength VEL, respectively. Is approximately the median value of the three lower regions divided into four. The reason why the fast decay waveform data corresponding to the key touch strength “127” is adopted is that the same key touch strength as that of the next slow decay waveform data is adopted, thereby simplifying the processing for preparing the waveform data. Because. Instead of the value “127”, the median value “112” of the uppermost area divided into four may be adopted. Furthermore, since the key touch strength VEL is not an absolute scale but a relative scale, other values may be used for the values “80”, “48”, and “16”. It is not limited to the four stages “127”, “80”, “48”, and “16”, and may be three stages or five or more stages. One slow decay waveform data is waveform data representing a slow decay component waveform over about 10 seconds extracted from a performance sound played with the strongest intensity. For example, the slow decay waveform data is waveform data representing a slow decay component waveform extracted from a performance sound having a key touch strength VEL of “127”.

  Waveform data related to percussion instrument sounds is also stored in the waveform memory WM, but percussion instrument sounds need not be prepared for each of a plurality of keys (a plurality of key pitches), and only one is required. Only one waveform set is stored in the storage area corresponding to. However, also in this case, during one waveform set, from the performance sound (percussion instrument sound) whose operation element touch strength VEL is “127”, “80”, “48”, and “16”, as described above. Four waveform data representing the extracted fast decay component waveform are included as a plurality of fast decay waveform data. The one waveform set also includes one waveform data representing a low-speed decay component waveform extracted from a performance sound (percussion instrument sound) whose operation element touch VEL is “127”. Also in this case, as described above, the value of the operator touch strength VEL is not limited to “127”, “80”, “48”, and “16”, and at the same time, the number can be changed as appropriate. is there. The method for generating the high-speed attenuation waveform data and the low-speed attenuation waveform data of the attenuation sound and percussion instrument sound corresponding to a plurality of key pitches is as described above. These fast decay waveform data and slow decay waveform data may be stored in the waveform memory WM as they are, or may be compressed and stored. For a plurality of high-speed decay waveform data corresponding to a plurality of different key touch strengths VEL (or operation element touch strength VEL) stored in the waveform memory WM, the amplitude (volume level) thereof is set to the key touch strength. However, in this embodiment, even if the key touch strength VEL (or the manipulator touch strength VEL) is different, a plurality of high-speed decay waveforms are possible. The amplitude of the data is almost the same. This is because the waveform data can be stored with high accuracy by increasing the amplitude of the fast decay waveform data corresponding to the small key touch strength VEL (or the operator touch strength VEL).

  The timbre parameter memory PM is also divided into a plurality of storage areas respectively corresponding to a plurality of timbres, as shown in FIG. 10B. Also in this case, since the continuous tone is not directly related to the present invention, only the decaying tone will be described in the following description. Each storage area corresponding to a timbre includes a head including a timbre name information representing a timbre name, a waveform set control parameter group, a filter control parameter group, an amplitude control parameter group, other control parameter groups, and a plurality of Selection information relating to the waveform set is stored. The waveform set control parameter group, the filter control parameter group, and the amplitude control parameter group are supplied to the readout circuit 55a, the digital control filter 55b, and the digital control amplifier 55c in the sound source circuit 55, respectively, and the readout circuit 55a, the digital control filter 55b, and This is a parameter group relating to one tone color used for generating a digital musical tone signal in the digital control amplifier 55c. The other control parameter groups are other parameters that are supplied to the tone generator circuit 55 and the effect circuit 57 and used to generate a digital tone signal of one tone color.

  The selection information related to the plurality of waveform sets corresponds to the plurality of waveform sets in FIG. 10A, respectively. The selection information related to each waveform set includes original pitch, fast decay waveform selection information, and slow decay waveform selection information. The original pitch is a parameter for pitch shifting when generating a digital musical sound signal using waveform data. From the information indicating the pitch of the performance sound recorded when generating the waveform data, the sampling frequency when recording the performance sound, etc. Become. In this embodiment, waveform data for each key is prepared to simplify the explanation, and the sampling frequency for generating waveform data and the sampling frequency for generating a digital musical tone signal in the tone generator circuit 55 are described. Is the same as the original pitch, the information representing only the pitch (key pitch) of the performance sound when the waveform data is generated can be used. Note that the pitch shift amount supplied from the computer main unit to the readout circuit 55a is the difference between the pitch of the key pressed and the pitch of the performance sound at the time of recording employed to generate the waveform data. is there. The fast decay waveform selection information is a plurality of sets of start address information and tail address information representing the start and end addresses of each storage area of a plurality of fast decay waveform data in each waveform set. The plurality of sets of head address information and tail address information are arranged in order of strong key touch. The slow decay waveform selection information is a set of head address information and tail address information representing the start and end addresses of one slow decay waveform data storage area in each waveform set, respectively.

  In the timbre parameter memory PM, as to the percussion instrument sound, as in the case described above, the selection related to the header, waveform set control parameter group, filter control parameter group, amplitude control parameter group, other control parameter group, and waveform set is selected. Information is stored. However, basically, for percussion instrument sounds, one waveform set is prepared for each instrument type, so selection information related to the waveform set includes the sampling frequency at the time of waveform data generation and the digital in the tone generator circuit 55. Only when the sampling frequency for generating a musical sound signal is the same, only the fast decay waveform selection information and the slow decay waveform selection information for one waveform set are required. However, when generating percussion instrument sound signals having a plurality of different pitches for one type of percussion instrument, such as timpani, a plurality of waveform data corresponding to a plurality of pitches is prepared, or only one waveform data is prepared. Prepare and change the playback pitch. When preparing a plurality of waveform data, as described above, only when the sampling frequency when generating the waveform data and the sampling frequency for generating the digital musical tone signal in the tone generator circuit 55 are the same, the original pitch is used. Is no longer necessary. However, when only one waveform data is prepared, information on the pitch of the performance sound recorded for generating the waveform data is necessary. The fast decay waveform selection information and the slow decay waveform selection information are the same as in the case of the decay sound generated by the key depression described above.

  Next, the operation of the electronic musical instrument configured as described above will be described. When the performer operates one of the plurality of panel operators 53 and selects a tone of a decaying musical tone (for example, an electric piano) generated by the performance operation of the keyboard 51, the CPU 61 executes a program (not shown). Thus, the waveform set control parameter group, the filter control parameter group, the amplitude control parameter group, and other control parameter groups are read from the storage area corresponding to the selected tone color in the tone color parameter memory PM, and temporarily stored in the RAM 63. Keep it.

  In this state, when the performer presses any key on the keyboard 51, the CPU 61 starts executing the note-on event processing program of FIG. 12 in step S300. After starting execution of the note-on event processing program, the CPU 61 inputs a note number NN representing the depressed key detected by the detection circuit 51a in step S302 and a key touch signal representing the key touch strength VEL. Next, in step S304, the CPU 61 generates two musical tone signals related to the depressed key, and two free sound channels 1ch to nch of the tone generator circuit 55 are generated for the key. Assign sound generation channels as first and second sound generation channels.

  After the processing in step S304, in step S306, the CPU 61 stores a plurality of selection information related to the plurality of waveform sets 1, 2,... Stored in the storage area corresponding to the selected timbre in the timbre parameter memory. Compares the original pitch in the selection information with the inputted note number NN, and specifies selection information corresponding to the key pressed, that is, specifies selection information related to the waveform set NN corresponding to the note number NN To do. Next, in step S308, the CPU 61 selects a key touch strength from among a plurality of sets of address information respectively representing the beginning and end storage addresses of each storage area of the plurality of fast decay waveform data included in the selection information. A set of address information respectively representing the storage addresses at the beginning and end of one storage area of one fast decay waveform data corresponding to the length VEL is selected. In selecting this set of address information, a plurality of key touch strengths (in this case, “127”, “80”, “48”, “16”) respectively corresponding to a plurality of fast decay waveform data. Of these, fast decay waveform selection information (a set of head address information and tail address information) relating to one fast decay waveform data corresponding to one key touch strength closest to the key touch strength VEL is selected.

  Next, in step S310, the CPU 61 performs a preparation process for generating a fast decay component waveform for the assigned first sound generation channel. In this high-speed attenuation component waveform preparation process, the CPU 61 outputs the selected set of head address information and tail address information to the first tone generation channel of the tone generator 55 as a read start address and a read end address, respectively. The waveform set control parameter group, the filter control parameter group, the amplitude control parameter group, and other control parameter groups temporarily stored in the RAM 63 by the program processing (not shown) are output to the first tone generation channel. In the preparation process for generating the fast decay component waveform, the CPU 61 calculates the volume level according to the key touch strength VEL, that is, the volume level of the digital musical tone signal generated in the second tone generation channel. And a balanced volume level is calculated, and a control parameter representing the calculated volume level is output to the first tone generation channel. This is because each maximum amplitude of a plurality of fast decay waveform data corresponding to a plurality of different key touch strengths stored in the waveform memory WM is substantially the same in this embodiment, and corresponds to the key touch strength VEL. Because it is not. The first tone generation channel temporarily stores various control parameters including the output readout start address, readout end address, and control parameter indicating the volume level, and prepares for generation of a digital musical tone signal. Do.

After the processing in step S310, the CPU 61 uses the key touch strength VEL in step S312, and selects the storage address represented by the selected head address information as the key touch strength VEL decreases. The address approaching the storage address represented by the tail address information is determined as the reading start address of the slow decay waveform data. Specifically, for example, as shown in FIG. 11B, if the top and end storage addresses are ADtop and ADend, respectively, the read start address ADstart is calculated by the following equation 1.
ADstart = ADtop + (ADend−ADtop−INTmin) · {1− (VEL−1) / 126} Equation 1
Note that INTmin in Equation 1 is a fixed value representing a value equal to the address interval from the read start address ADstart to the end storage address ADend when the key touch strength VEL is the minimum “1”. However, when the calculation result of Equation 1 includes a decimal part, the calculation result is converted to an integer by rounding off.

  According to Equation 1, when the key touch strength VEL is the maximum “127”, the read start address ADstart is set to the top storage address ADtop. As the key touch strength VEL decreases, the read start address ADstart changes from the top storage address ADtop to the end storage address ADend. When the key touch strength VEL is the minimum “1”, the read start address ADstart is set to the address on the top storage address ADtop side by the address interval INTmin from the end storage address ADend.

  However, the formula 1 is based on a model simplified for easy understanding of the present embodiment, and is not a formula that can be generalized. Here, attention should not be paid to the calculated numerical value, but rather the tendency of the calculated numerical value. Further, the top address ADtop in Equation 1 is not necessarily the top storage address. That is, the read start address ADstart is set to the address closest to the top when the key touch strength VEL is the maximum value “127”, and changes backward from the address closest to the top as the key touch strength VEL decreases. It may be set to the address to be. In calculating the read start address ADstart, various functions and tables can be used depending on the type of musical instrument. In particular, the read start address ADstart is defined by a linear scale, and the volume level (key touch strength VEL) of the digital musical tone signal is defined by a decibel scale. Therefore, when the key touch strength VEL decreases on the decibel scale. The read start address ADstart linearly changes backward from the top storage address ADtop. By determining the read start address ADstart, the volume level of the generated digital tone signal can be controlled by the key touch strength VEL without controlling the volume level of the digital tone signal generated by reading the waveform data. It becomes a thing corresponding to.

  Next, in step S314, the CPU 61 performs a preparation process for generating a slow decay component waveform for the assigned second tone generation channel. In the preparation process of the slow decay component waveform, the CPU 61 outputs the tail address information of the slow decay waveform data together with the determined read start address ADstart to the second tone generation channel of the tone generator circuit 55 as a read end address. The waveform set control parameter group, the filter control parameter group, the amplitude control parameter group, and other control parameter groups temporarily stored in the RAM 63 by the program process (not shown) are output to the second tone generation channel. The second tone generation channel temporarily stores the output read start address, read end address, and various control parameters, and prepares for generation of a digital musical tone signal. For the second tone generation channel, the volume level must be controlled so that the volume balance is achieved with the digital tone signal generated by the first tone generation channel, but the volume level corresponding to the key touch strength VEL is required. No control is necessary.

  After the processing in step S314, the CPU 61 instructs the first and second sound generation channels to start sounding in step S316, and ends the execution of the note-on event processing program in step S318. The first and second sound generation channels start to generate a digital musical tone signal in response to the sound generation start instruction. In the first tone generation channel, the readout circuit 55a sequentially reads out the high-speed decay waveform data stored in the waveform memory WM while stepping up the address as time elapses from the read start address. Generates and outputs a digital musical tone signal composed of a high-speed decay component waveform represented. In this case, in this embodiment, the sampling frequency of the performance sound when generating the waveform data is the same as the sampling frequency for generating the digital musical tone signal by the tone generator circuit 55, and a waveform set is prepared for each key. Therefore, pitch shift is not required, and the address increment is “1” for each sampling period. However, in other cases, it is necessary to perform pitch shift control. The reading of the high-speed decay waveform data ends when the address step reaches the read end address. When the reading of the high-speed decay waveform data is completed, the first tone generation channel is released to become an empty channel. As a result, a fast decay component waveform based on the fast decay waveform data specified by the note number NN and the key touch VEL is generated.

    The digital musical tone signal output from the readout circuit 55a is frequency-controlled by the digital control filter 55b, and the amplitude characteristic is controlled by the digital control amplifier 55c, and is output to the channel accumulation circuit 55d. In particular, regarding the control of the amplitude characteristic in the first tone generation channel, the digital control type amplifier 55c uses the control parameter indicating the volume level supplied to the tone generator circuit 55 by the processing of step S20, so that the digital control type amplifier 55c is in a steady state of the digital tone signal. The volume level is controlled according to the key touch strength VEL. In this case, the digital control amplifier 55c basically changes the amplitude characteristic of the digital musical tone signal based on the read high-speed decay waveform data unless a release start instruction and a rapid decay instruction are received as will be described later. There is no time change.

  On the other hand, also in the second tone generation channel, the readout circuit 55a sequentially reads out the low-speed decay waveform data stored in the waveform memory WM while stepping up the address as time elapses from the read-out start address. A digital musical tone signal composed of a slow decay component waveform represented by data is generated and output. Also in this case, in this embodiment, the sampling frequency of the performance sound when generating the waveform data is the same as the sampling frequency for generating the digital musical tone signal by the tone generator circuit 55, and a waveform set is prepared for each key. Therefore, no pitch shift is required, and the address increment is “1” for each sampling period. However, in other cases, it is necessary to perform pitch shift control. Then, the reading of the low-speed decay waveform data also ends when the address step reaches the read end address. Also in this case, when the reading of the slow decay waveform data is completed, the second sound generation channel is released and becomes an empty channel. As a result, the slow decay waveform data designated by the note number NN starts to be read from the readout start address designated by the key touch strength VEL, and the slow decay component waveform is generated based on the read slow decay waveform data. .

  The frequency characteristic of the digital musical tone signal output from the readout circuit 55a is also controlled by the digital control filter 55b, the amplitude characteristic is controlled by the digital control amplifier 55c, and is output to the channel accumulation circuit 55d. Regarding the control of the amplitude characteristic in the second tone generation channel, unlike the first tone generation channel, the steady volume level of the digital musical tone signal corresponding to the key touch strength VEL is read out from the low-speed decay waveform data described above. Since it depends on the start address ADstart, the volume level is not controlled in accordance with the key touch strength VEL, and other volume levels are controlled as necessary. Also in this case, the digital control amplifier 55c basically has an amplitude characteristic of the digital musical tone signal based on the read high-speed decay waveform data unless receiving a release start instruction and a rapid decay instruction as will be described later. The time is not changed.

  The channel accumulation circuit 55d adds the digital musical tone signal composed of the fast decay component waveform output from the first tone generation channel and the digital musical tone signal composed of the slow decay component waveform output from the second tone generation channel. The mixed digital musical tone signal is output to the effect circuit 57. The effect circuit 57 gives an effect based on an effect control parameter supplied by program processing (not shown) and outputs the effect to the sound system 58. The sound system 58 converts the output digital musical sound signal into an analog musical sound signal and emits the sound through a speaker.

  In the generation of the tone signal as described above, the key pressing time by the performer is sufficiently long, and the reading circuit 55a in the tone generation channel of the tone generator circuit 55 reads the address for reading the slow decay waveform data up to the reading end address. When it reaches, all the high-speed decay waveform data and the low-speed decay waveform data are output as digital musical tone signals. However, if the key that has been pressed is released while the digital musical tone signal is being generated, the CPU 61 executes a program (not shown) to generate a tone generator to which generation of musical tone has been assigned to the released key. An instruction signal for instructing the start of release is output to the two sound generation channels in the circuit 55. The digital control amplifier 55c in the two sound generation channels attenuates the amplitude of the digital musical tone signal from the digital control filter 55b relatively rapidly. That is, the digital control amplifier 55c releases the digital musical tone signal from the digital control filter 55b. The digital control filter 55b also slightly changes the frequency characteristic of the digital musical tone signal as required in synchronization with the release process. Then, the tone generation channel in which the digital musical tone signal has been attenuated by the release process is released to be an empty channel.

  Further, when a large number of keys are pressed and released on the keyboard 51 within a short period of time, and the number of sound generation channels is insufficient, the CPU 61 executes the program (not shown) to generate the most sound generation channels related to the released keys. An instruction signal for instructing the start of rapid attenuation is output to a sound generation channel generating a digital musical tone signal having a small amplitude level. The digital control amplifier 55c in the tone generation channel that has received this instruction signal rapidly attenuates the amplitude of the digital musical tone signal from the digital control filter 55b. Also in this case, the digital control filter 55b slightly changes the frequency characteristic of the digital musical tone signal as needed in synchronization with the rapid attenuation processing of the digital control amplifier 55c. Also in this case, the tone generation channel after the digital musical tone signal has been attenuated is released to be an empty channel. As a result, even when a large number of keys are pressed and released within a short time, a tone generation channel for generating a tone of a newly pressed key is secured.

  Next, a case where any one of the plurality of performance operators 52 is operated to generate a percussion instrument sound will be described. The CPU 61 executes a modified program (not shown) obtained by modifying the above-described note-on event processing program shown in FIG. 12, and controls the pronunciation of percussion instrument sounds. The execution of the deformation program is started in response to the input of the switch-on signal SWON from the detection circuit 52a. Then, instead of the input of the note number NN and the key touch strength VEL in step S302 of FIG. 12, the instrument type information IN and the operator touch strength VEL are input from the detection circuit 52a. Further, instead of specifying the selection information regarding the waveform set NN corresponding to the note number NN in step S306, the percussion instrument timbre represented by the input instrument type information IN among the plurality of storage areas in the timbre parameter memory PM. Specify the storage area corresponding to. Then, the waveform set control parameter group, filter control parameter group, amplitude control parameter group and other control parameter groups in the designated storage area are read out and temporarily stored in the RAM 63. In this case, only one selection information (high-speed attenuation waveform selection information and low-speed attenuation waveform selection information) stored in the storage area is designated. The other processes in steps S304 and S308 to S316 in FIG. 12 are the same as the control process for generating a tone signal by pressing the key.

  As a result, even when any one of the plurality of performance operators 52 is turned on, one fast decay waveform data among the plurality of fast decay waveform data is designated according to the operator touch strength VEL. . Further, the read start address of one low-speed decay waveform data is determined according to the operator touch strength VEL. Accordingly, in the generation of a percussion instrument sound by turning on the performance operator 52, as in the case of the key depression, a digital musical tone signal based on one fast decay waveform data designated by the operator touch strength VEL, and the operation A digital musical tone signal based on the low-speed decay waveform data in which the reading start address is designated by the child touch strength VEL is mixed and output. Also, in the case of the digital musical tone signal related to the percussion instrument sound, the volume level of the mixed high-speed decay waveform data is determined by the operator touch strength as in the case of the digital musical tone signal having a pitch corresponding to the key pitch described above. The volume level is controlled according to VEL.

  Further, even when the key performance event data and the percussion instrument performance event data are read out by the reproduction of the automatic performance data by the execution of the automatic performance program (not shown), the note-on event processing program of FIG. Attenuation-type musical sounds and percussion instrument sounds are generated by the key performance described above. That is, the automatic performance data includes percussion instrument performances including key performance event data including key-on and key-off events including key touch strength VEL and note number NN, and percussion instrument operation events including operator touch strength VEL and instrument type information IN. Event data is stored over time. The automatic rhythm data is obtained by storing percussion instrument performance event data including a percussion instrument operation event including an operator touch strength VEL and instrument type information IN as time passes. These key performance event data and percussion instrument performance event data are the key-on signal KON, key-off signal KOF, note number NN, key touch strength VEL, and switch-on signal SWON output from the detection circuits 51a and 52a of the above embodiment. This is the same as the operator touch strength VEL, instrument type information IN, and the like. Therefore, if the generation of musical sounds is controlled as described above using the key performance event data and the percussion instrument performance event data output during the reproduction of the automatic performance data and the automatic rhythm data, the musical sounds generated by the automatic performance and the automatic rhythm are not generated. The musical tone generated by the keyboard 51 and the plurality of performance operators 52 is exactly the same.

  According to the above musical sound signal generator, the waveform memory WM includes fast decay waveform data composed of a plurality of waveform data extracted from performance sounds played at a plurality of levels of intensity and performance sounds played at a high intensity. Low-speed decay waveform data including one extracted waveform data is stored. The first tone generation channel to which the tone generation in the tone generator circuit 55 is assigned is 1 according to strength information (key touch strength VEL and operator touch strength VEL) from among a plurality of waveform data as fast decay waveform data. One waveform data is selected and read, and a digital musical tone signal that is a high-speed attenuation component waveform is generated based on the read high-speed attenuation waveform data. At this time, the volume level of the digital musical tone signal, which is a fast decay component waveform, is controlled according to the key touch strength VEL and the operator touch strength VEL. The second tone generation channel to which the tone generation in the tone generator circuit 55 is assigned reads one waveform data as the slow decay waveform data, and based on the read slow decay waveform data, the digital tone signal that is the slow decay component waveform is read out. Generate. Then, the channel accumulation circuit 55d mixes and outputs the digital musical tone signal generated by the first tone generation channel and the digital musical tone signal generated by the second tone generation channel.

  The fast decay component waveform changes according to the strength of the generated musical tone, but the slow decay component waveform hardly changes depending on the strength of the generated musical tone, so the mixed digital musical tone signal increases the generated musical tone waveform that changes according to the strength. It will be expressed with accuracy. Therefore, according to the above musical tone signal generator, a realistic musical tone signal can be generated. The waveform memory WM also includes a plurality of fast decay waveform data representing a fast decay component waveform that ends in a short time from the start of generation, and a slow decay that represents one slow decay waveform that ends after a long time has passed since the start of generation. Since only waveform data is stored, the amount of waveform data for generating a musical sound signal can be reduced, and the waveform memory WM does not require a large capacity. Further, the time for using the first tone generation channel to generate the digital musical tone signal that is the fast decay component waveform is limited to a short time immediately after the start of the generation of the musical tone signal. Therefore, even if two sound generation channels are used for generating one musical sound, the occupied time of the sound generation channel is not so long as compared with a method using only one sound generation channel for generating one musical sound.

  Further, the second tone generation channel becomes farther from the top address of the waveform data as the intensity of the musical sound signal represented by the strength information becomes weaker as the waveform data as the slow decay waveform data stored in the waveform memory WM. The address is read out as time elapses with the reading start address. Therefore, as the intensity of the generated tone signal becomes weaker, the time from the start to the end of waveform data reading becomes shorter and the time from the start to the end of generation of the tone signal becomes shorter. This coincides with the characteristics of the attenuation-type musical instrument sound in which the time from the start to the end of the musical sound becomes shorter as the strength of the generated musical sound becomes weaker. Therefore, it is possible to better mimic the decaying tone by a simple method of changing the read start address according to the strength information.

  In the first tone generation channel, the volume level of the digital tone signal is controlled according to the key touch strength VEL and the operator touch strength VEL when generating the digital tone signal which is a high-speed decay component waveform. . Therefore, even if there is high-speed decay waveform data with weak and weak musical tone signals, the amplitude level can be increased when storing it in the waveform memory WM, and high-precision high-speed decay waveform data can be reproduced. It becomes possible.

  In the above musical tone signal generator, when the fast decay waveform data in the waveform memory WM is designated, one fast decay waveform data corresponding to the key touch strength closest to the key touch strength VEL is selected. The digital musical tone signal is generated using the selected one high-speed decay waveform data. However, instead of this, two high-speed attenuation waveform data corresponding to two key touch strengths sandwiching the key touch strength VEL among a plurality of key touch strengths respectively corresponding to the plurality of high-speed attenuation waveform data. It is also possible to generate a digital tone signal by selecting and interpolating (cross-fading) the two selected fast decay waveform data.

  In this case, the note-on event processing program of FIG. 13 modified from the note-on event processing program of FIG. 12 is stored in the ROM 62 or the external storage device 65, and the CPU 61 executes the note-on event processing program of FIG. . In the note-on event processing program of FIG. 13, the same parts as those of the note-on event processing program of FIG. 12 are denoted by the same reference numerals, and description thereof is omitted.

  In the note-on event processing program of FIG. 13, in step S304a instead of step S304 of FIG. 12, the CPU 61 generates a plurality of tone generator circuits 55 for the key in order to generate a tone signal related to the key that has been pressed. Of these sound generation channels, three free sound generation channels are assigned as first to third sound generation channels. In step S308a instead of step S308 in FIG. 12, the CPU 61 includes a plurality of sets of address information representing the storage addresses at the beginning and end of each storage area of the plurality of fast decay waveform data included in the selection information. 2 sets of address information (start address information and end address information) respectively representing the storage addresses at the beginning and end of the storage areas of the two fast decay waveform data corresponding to the two key touch strengths sandwiching the key touch strength VEL. ) Is selected.

In step S320, the CPU 61 calculates the mixing ratio of the two high-speed attenuation component waveforms represented by the selected two high-speed attenuation waveform data. In this case, the two key touch strengths sandwiching the key touch strength VEL are VEL1 and VEL2, respectively, and the mixing ratios of the fast decay component waveforms corresponding to the key touch strengths VEL1 and VEL2 are MIX1 and MIX2, respectively. Then, the mixing ratios MIX1 and MIX2 are determined by the calculations of the following formulas 2 and 3.
MIX1 = | VEL2-VEL | / | VEL1-VEL2 |
MIX2 = | VEL1-VEL | / | VEL1-VEL2 |
According to this, the mixing ratio of the two high-speed attenuation component waveforms is determined as MIX1 and MIX2 by the proportional distribution, and the two high-speed attenuation component waveforms are interpolated according to the key touch strength VEL. A synthesized waveform can be obtained. The determined mixing ratios MIX1 and MIX2 are output to the assigned first and second sound generation channels in the tone generator circuit 55, respectively.

  As in the case of Equation 1, Equations 2 and 3 are also equations based on a simplified model for easy understanding of the present embodiment, and are not equations that can be generalized. Here too, attention should not be paid to the calculated numerical value, but rather the tendency of the calculated numerical value. In particular, the mixing ratios MIX1 and MIX2 are actually expressed in decibel scales instead of linear scales. In addition, the mixing ratios MIX1 and MIX2 actually require an element for controlling the volume level in accordance with the key touch strength VEL in addition to the crossfade element shown in Equations 2 and 3. The decibel values of the ratios MIX1 and MIX2 need to be values adjusted so that the volume of the fast decay component waveform after mixing has a volume level corresponding to the key touch strength VEL. Specifically, the volume level (decibel value) of the digital musical tone signal that is the low-speed decay component waveform described above may be added to the right side of Equations 2 and 3.

  In step S310a instead of step S310 in FIG. 12, the CPU 61 performs a preparation process for generating the two selected fast decay component waveforms for the assigned first and second sound generation channels. . In the preparation processing of the high-speed attenuation component waveform, one set of address information of the selected two sets of address information is output to the first tone generation channel of the tone generator circuit 55 as a read start address and a read end address. The other set of address information of the selected two sets of address information is output to the second tone generation channel of the tone generator circuit 55 as a read start address and a read end address. The point that the waveform set control parameter group, the filter control parameter group, the amplitude control parameter group, and other control parameter groups are output to the first and second sound generation channels is the same as in the above embodiment. The first and second tone generation channels temporarily store the output read start address, read end address, and various control parameters, and prepare for generation of a digital musical tone signal.

  In step S314a instead of step S314 in FIG. 12, the CPU 61 performs a preparation process similar to that in the above-described embodiment for generating a slow decay component waveform for the assigned third sounding channel. In step S316a instead of step S316 in FIG. 12, the CPU 61 instructs the first to third sound generation channels to start sound generation. Other processes are the same as those in FIG.

  By executing the note-on event processing program according to the modified example, the first to third sound generation channels in the tone generator circuit 55 start to generate a digital musical tone signal in response to the sound generation start instruction. In this case, the first and second sound generation channels generate and output digital musical tone signals respectively representing two high-speed attenuation component waveforms according to two different high-speed attenuation waveform data. In the first tone generation channel, the digital control amplifier 55c multiplies the digital musical tone signal output from the digital control filter 55b by the mixing ratio MIX1 output in the process of step S320, thereby obtaining the digital musical tone signal. The amplitude is controlled in proportion to the mixing ratio MIX1. In the second tone generation channel, the digital control amplifier 55c multiplies the digital musical tone signal output from the digital control filter 55b by the mixing ratio MIX2 output by the process of step S320, thereby obtaining the digital musical tone. The amplitude of the signal is controlled in proportion to the mixing ratio MIX2.

  The third tone generation channel generates and outputs a digital musical tone signal representing a slow decay component waveform according to the slow decay waveform data. The digital musical tone signal generation process in the first to third sound generation channels is the same as that described above. The digital musical tone signals output from the first to third sound generation channels are output to the channel accumulation circuit 55d. The channel accumulation circuit 55d adds the digital musical tone signals output from these first to third tone generation channels and outputs the result. As a result, the two high-speed attenuation component waveforms output from the first and second sound generation channels are mixed at the mixing ratios MIX1 and MIX2, and are mixed with the low-speed attenuation component waveform output from the third sound generation channel. The

  With respect to two high-speed decay waveform components according to this modification, percussion instrument sounds are generated by a plurality of performance operators 52, automatic performance sounds are generated based on automatic performance data, and automatic rhythm sounds (percussion instrument sounds) are generated based on automatic rhythm data. It also applies to outbreaks.

  As a result, according to the musical tone signal generating apparatus according to this modification, the fast-decay waveform data to be finally mixed is a musical tone signal represented by strength information (key touch strength VEL and operator touch strength VEL). And the waveform data of the two intensity steps sandwiching the strength are interpolated (cross-fade) in accordance with the strength of the two and the two strength steps sandwiching the strength. Therefore, even if the number of waveform data corresponding to the intensity level as the high-speed attenuation waveform data stored in the waveform memory WM is reduced, the high-speed attenuation waveform data corresponding to the strength of the musical sound signal represented by the strength information can be obtained with high accuracy. Obtainable.

  In the musical tone signal generating apparatus, the key touch strength VEL (or the operator touch strength VEL) is set to the minimum value “VEL” as the read start address of the low-speed decay waveform data by the processing in step S312 of FIGS. The address value linearly changing from the address ADend-INTmin to the address ADtop as the value changes from “1” to the maximum value “127” is calculated. However, instead of this, a function is provided that defines an address value that changes nonlinearly from the address ADend-INTmin to the address ADstart as the key touch strength VEL changes from the minimum value “1” to the maximum value “127”. In this case, an address value that changes nonlinearly with respect to changes in the key touch strength VEL and the operator touch strength VEL may be calculated using the function.

  Also, a plurality of read start positions are stored in correspondence with a plurality of key touch strengths VEL (or operation element touch strengths VEL), and are stored by the processing of step S312 in FIGS. The read start address of the slow decay waveform data may be determined using a plurality of read start positions. In this case, as the slow decay waveform selection information for each waveform set, in addition to the start and end addresses of the storage area of the slow decay waveform data, it is an address between these start and end storage addresses, and the generated musical tone signal It is sufficient to store a plurality of pieces of intermediate address information representing a plurality of read start addresses corresponding to the intensity of the. As shown in FIG. 11B, the intermediate address information represents a read start address corresponding to, for example, key touch strengths VEL “96”, “64”, “32”, “1”. When the reading start position where the key touch strength VEL is the maximum value “127” is set to “0 seconds” and converted into time related to reading waveform data, the four intermediate addresses are “2 seconds” from the head side, The positions correspond to “4 seconds”, “6 seconds”, and “8 seconds”, respectively.

In the process of step S312 of FIGS. 12 and 13, if the key touch strength VEL input by the process of step S302 is the maximum key touch (ie, “127”), the process is specified by the process of step S306. The start address of the slow decay waveform data included in the slow decay waveform selection information in the selected information is determined as the read start address. Further, if the key touch strength VEL matches any one of the plurality of key touch strengths “96”, “64”, “32”, “1”, an intermediate corresponding to the matched key touch strength VEL. Address information is determined as a read start address. If the key touch strength VEL does not match any of the plurality of key touch strengths “127”, “96”, “64”, “32”, and “1”, among the plurality of key touch strengths, Two key touch strengths on both sides between which the key touch strength VEL input by the process of step S302 is sandwiched are selected. Assuming that the two selected key touch strengths are VEL1, VEL2 (VEL1> VEL2), and two intermediate addresses AD1, AD2 (AD2> AD1) corresponding to the key touch strengths VEL1, VEL2, read start address ADstart Is calculated by the interpolation calculation of the following equation 4. However, Formula 4 is also a formula based on a model simplified for easy understanding of the present embodiment, and is not a formula that can be generalized. Here, the calculated numerical value should not be noted, but rather the tendency of the calculated numerical value should be noted.
ADstart = AD1 + (AD2−AD1) ・ (VEL−VEL1) / (VEL2−VEL1)… Equation 4

  As described above, when a plurality of read start addresses corresponding to the key touch strengths “96”, “64”, “32”, and “1” are set as a plurality of intermediate address information, a plurality of key touches are provided. Since the interval of the strength VEL is substantially uniform as “31” or “32”, the read start address changes substantially linearly according to the key touch strength VEL (or the operator strength VEL). However, if the intervals of the addresses represented by the plurality of intermediate addresses are made different, for example, if the addresses are gradually increased, the read start address changes nonlinearly with respect to the change in the key touch strength VEL. .

  Further, in the musical tone signal generating apparatus, a waveform set for each tone color in the waveform memory WM is prepared for each key pitch (note number NN) with respect to a decaying musical tone waveform whose frequency changes according to the key pitch. I did it. However, instead of this, a waveform set is prepared for each key tone range (for example, a half octave) including a plurality of key pitches as a waveform set for each tone color in the waveform memory WM for the attenuation-type musical sound waveform. It may be. In this case, a plurality of waveform sets respectively corresponding to a plurality of key ranges are stored in the waveform memory WM. In addition to the header and various control parameters, selection information regarding a plurality of waveform sets corresponding to the plurality of waveform sets is stored in the storage area corresponding to each tone color of the tone color parameter memory PM. The selection information regarding each waveform set is configured in the same manner as in the above embodiment, and the original pitch in the selection information includes information indicating the pitch of the original performance sound when the waveform data related to the waveform set is generated. In this modification, the key number (pitch) specified by the note number NN of the pressed key and the note number NN according to the difference between the pitch of the performance sound included in the original pitch. A digital musical tone signal (a digital musical tone signal having a fast decay component waveform and a slow decay component waveform) having a specified key pitch is generated.

  Specifically, first, the CPU 61 selects, from the timbre parameter memory PM, selection information related to the waveform set of the key range to which the pressed key belongs, and the fast decay waveform data and the slow decay in the selected selection information. Address information relating to the waveform data is output to the first and second tone generation channels (or the first to third tone generation channels) to which the keys have been pressed and assigned keys, respectively. However, regarding the address information related to the fast decay waveform data, the address information of one (or two) fast decay waveform data corresponding to the key touch strength VEL is selected as in the case of the above embodiment. At the same time as the output of the address information, the CPU 61 calculates the difference between the key pitch (pitch) corresponding to the pressed key and the pitch represented by the original pitch in the selected selection information. The pitch shift amount (cent) is output to the first and second tone generation channels (or the first to third tone generation channels).

  The readout circuit 55a in the first and second tone generation channels (or the first to third tone generation channels) determines the waveform data read rate according to the pitch shift amount and is designated by the address information. Read high-speed decay waveform data and low-speed decay waveform data. As a result, the reading circuit 55a generates a digital musical tone signal having a key pitch corresponding to the note number NN and outputs the digital musical tone signal to the digital control filter 55b. However, since the reading rate determined according to the pitch shift amount normally includes a decimal part, the reading address of the waveform data also includes an integer part and a decimal part. Therefore, in reading out the waveform data, a plurality of sample values of the waveform data are read out using the integer part, and a final digital musical tone signal is generated by interpolation calculation using the decimal part. The processing after the generation of the digital tone signal is the same as in the above embodiment. In this modification as well, it is assumed that the sampling frequency at the time of waveform data generation and the sampling frequency at the time of generation of the digital tone signal of the tone generator circuit 55 are the same. When these two sampling frequencies are different, it is necessary to determine the pitch shift amount in consideration of the difference between the sampling frequencies.

DESCRIPTION OF SYMBOLS 11 ... Panel switch, 12 ... Display, 13 ... Waveform memory, 14 ... Write circuit, 16 ... Sound source circuit, 18 ... Sound system, 31 ... CPU, 51 ... Keyboard, 52 ... Performance operator, 55 ... Sound source circuit

Claims (2)

Fourier transforming the whole of the original waveform data representing the instrument sound waveform, and generating a spectral component of the entire instrument sound waveform;
Detecting a plurality of peaks corresponding to a fundamental wave component and a harmonic overtone component of the instrument sound waveform from a spectrum distribution represented by a spectrum component of the entire generated instrument sound waveform, and a plurality of peaks corresponding to the plurality of peaks. Identifying a spectral component;
A step of extracting spectral components related to the identified plurality of spectral components and spectral components in the vicinity thereof among spectral components of the generated instrument sound waveform; and
A step of generating a slow decay waveform data representing a slow decay component waveform by performing an inverse Fourier transform on the extracted spectral component;
Subtracting the generated slow decay waveform data from the original waveform data to generate fast decay waveform data representing a fast decay component waveform. While moving the time window with respect to the original waveform data representing the musical instrument sound waveform, the original waveform data is sequentially Fourier transformed for each frame defined by the time window, and the spectrum of a plurality of frames over the entire original waveform data is obtained. Obtaining each of the components;
Detecting a plurality of peak loci respectively connecting a plurality of peaks of the acquired spectral components over the plurality of frames;
Extracting a slowly decaying peak trajectory from the detected plurality of peak trajectories;
A process of generating a slow decay waveform data representing a slow decay component waveform by performing an inverse Fourier transform on a spectrum component relating to the extracted peak locus;
Subtracting the generated slow decay waveform data from the original waveform data to generate fast decay waveform data representing a fast decay component waveform.

Images