JP4736046B2 - Waveform data production method, waveform data production apparatus, program, and waveform memory production method - Google Patents

Description

  The present invention relates to a waveform data production method, a waveform data production apparatus, a program, and a waveform memory production method suitable for use in a device that applies vibrato to a musical sound signal.

  Various techniques for adding vibrato to a musical sound signal have been proposed. As the simplest one, a modulation waveform of about several Hz is generated by a low frequency oscillator (LFO), and the pitch of a musical tone signal is simply modulated by this modulation waveform. Patent Document 1 discloses a technique for sampling a musical tone played with vibrato and cutting out waveform data of an attack part and a loop part from the sampling result. When playing a musical sound, the attack part is first played only once, and then the loop part is played repeatedly. Patent Document 2 discloses a technique of providing two systems of waveform memory in the loop portion. According to this technique, the pitch is modulated by a modulation waveform by a low frequency oscillator (LFO), and waveform data of both waveform memories are read out in the same phase. Both waveform data are mixed, but the mixing ratio of both waveform data is increased or decreased by the modulation waveform.

Japanese Examined Patent Publication No. 02-010440 JP 60-090391 A

  However, among the above-described technologies, the technology that simply modulates the pitch of the tone signal with the modulation waveform cannot give the tone signal a timbre change according to vibrato, and the tone signal becomes unnatural. there were. Also, according to the technique disclosed in Patent Document 1, the vibrato period and depth are determined by the waveform data itself, and thus the vibrato period and depth cannot be changed. Therefore, if there are various variations in the vibrato period, depth, etc., there is a problem that the required number of waveform memories has to be sampled, and the required capacity of the waveform memory increases.

  On the other hand, in the technique disclosed in Patent Document 2, the cycle and depth of the vibrato can be freely set according to the setting state of the modulation waveform and the mixing ratio. However, Patent Document 2 does not disclose a method for creating waveform data to be stored in two types of waveform memories. The present invention has been made in view of the above-described circumstances, and provides a waveform data production method, a waveform data production apparatus, a program, and a waveform memory production method capable of generating a plurality of types of waveform data when vibrato is imparted to a musical sound signal. The purpose is that.

In order to solve the above problems, the present invention is characterized by having the following configuration. The parentheses are examples.
In the waveform data production method according to claim 1, the frequency component of the original waveform data (30) whose pitch varies is analyzed, and the frequency (fkq) of the pitch component (k = 1) and the harmonic component (k> 2) is analyzed. ) (80, 82, 84) and the analysis process for obtaining the original analysis data consisting of the amplitude level trajectory, and a part of the time range in the original analysis data, the head part of the original analysis data being A flattening step (38) for obtaining flattening analysis data by correcting the original analysis data so that a frequency of the pitch component becomes a constant flattening pitch frequency (fp) within a flattening range not including; Within the flattening range, the frequency of the pitch component of the corresponding point in the original analysis data becomes the average of the frequency of the pitch component in the original analysis data by the user's operation. When one or a plurality of first type points (upper points P0 to P4) which are points on the time axis higher than the value are designated, the flattening analysis of the pitch component and the overtone component for each first type point. A first amplitude level acquisition process (24) for acquiring a first amplitude level (LUkq), which is an amplitude level in the data, and a corresponding operation in the original analysis data by a user operation within the flattening range. One or a plurality of second type points (lower points R0 to R5) that are points on the time axis at which the frequency of the pitch component of the point is lower than the average value of the frequency of the pitch component in the original analysis data are designated. Then, for each second type point, a second amplitude level (LDkq), which is an amplitude level in the flattening analysis data of the pitch component and the harmonic component, is obtained. A second amplitude level acquisition process (26) to be obtained and an amplitude level corresponding to the first amplitude level is applied to points other than the first type point of the pitch component and the harmonic component in the flattening analysis data. A first amplitude correction process (40) for generating first amplitude correction analysis data (upper deterministic analysis data) and a pitch component and a harmonic component other than the second type point in the flattening analysis data A second amplitude correction process (42) for generating second amplitude correction analysis data (lower deterministic analysis data) obtained by applying an amplitude level corresponding to the second amplitude level to a point; A synthesis step (44) of synthesizing first waveform data based on the first amplitude correction analysis data and synthesizing second waveform data based on the second amplitude correction analysis data. It characterized by having a 46,58,60,62,64) and.
Furthermore, in the configuration according to claim 2, in the waveform data production method according to claim 1, the first amplitude correction process includes the last point (P4) on the time axis among the first type points. ) Thereafter, the first flattening analysis data is changed so that the amplitude levels of the pitch component and the harmonic component of the flattening analysis data become the first amplitude level related to the last point (P4). The amplitude correction analysis data (upper deterministic analysis data) is generated, and the second amplitude correction process is performed after the last point (R5) on the time axis among the second type points. By changing the flattening analysis data so that the amplitude level of the pitch component and the overtone component of the flattening analysis data becomes the second amplitude level related to the last point (R5), the second vibration level is changed. A process of generating width correction analysis data (lower deterministic analysis data), in the time range after the first point of the first type point and after the first point of the second type point To a loop range designating process for designating a time range that is an integral multiple of a period related to the flattened pitch frequency (fp) as a loop range, and the first waveform data is the first amplitude correction analysis data. A loop portion based on the data of the loop range, and an attack portion based on data of a time range from the beginning of the original waveform data to immediately before the flattening range, and the second waveform The data includes a loop portion based on the data of the loop range in the second amplitude correction analysis data, and a time range from the beginning of the original waveform data to immediately before the flattening range. It has an attack part based on data, It is characterized by the above-mentioned.
Furthermore, in the configuration according to claim 3, in the waveform data production method according to claim 1, a plurality of the first type points (P0 to P4) and the second type points (R0 to R5) are respectively designated. In the first amplitude correction process, in the first amplitude correction analysis data, the amplitude levels of the pitch component and the harmonic component at the plurality of first type points respectively correspond to the first amplitude correction process. The amplitude level of the pitch component and the harmonic component at the intermediate point between the first type points is set to the amplitude level of the pitch component and the harmonic component at the first type point before and after the intermediate point. Is set to an amplitude level obtained by interpolation, and the second amplitude correction process includes the plurality of second types in the second amplitude correction analysis data. The amplitude levels of the pitch component and the harmonic component at the point are respectively set to the corresponding second amplitude levels, and the amplitude levels of the pitch component and the harmonic component at the intermediate point between the respective second type points are The amplitude level of the pitch component and the harmonic component at the second type point before and after the intermediate point is set to an amplitude level obtained by interpolation, and the synthesis process (44, 46, 58, 60, 62, 64) Based on the amplitude level of the pitch component and the harmonic component at the first type point and the corrected amplitude level which is an amplitude level set by interpolating between the pitch component and the harmonic component at the first type point. While synthesizing the first waveform data, the amplitude of the pitch component and the harmonic component at the second type point A process of synthesizing the second waveform data based on a corrected amplitude level that is an amplitude level set by interpolating between a bell and a pitch component and a harmonic component at the second type point. It is characterized by.
Furthermore, in the configuration according to claim 4, in the waveform data production method according to claim 3, the first amplitude correction process further includes a last point on the time axis among the first type points. (P4) After that, the flattening analysis data is changed so that the amplitude levels of the pitch component and the harmonic component of the flattening analysis data become the first amplitude level related to the last point (P4). The first amplitude correction analysis data (upper deterministic analysis data) is generated, and the second amplitude correction process further includes the last point (R5) on the time axis among the second type points. ) Thereafter, the flattening analysis data is changed so that the amplitude levels of the pitch component and the harmonic component of the flattening analysis data become the second amplitude level related to the last point (R5). Thus, the second amplitude correction analysis data (lower deterministic analysis data) is generated.
Furthermore, in the configuration according to claim 5, in the waveform data production method according to claim 3 or 4, the first point after the first point among the first type points and the first point among the second type points. A loop range specifying step of specifying a time range that is an integral multiple of a period related to the flattening pitch frequency (fp) as a loop range from the subsequent time ranges, and the first waveform data includes the first waveform data A loop portion based on the data of the loop range among the amplitude correction analysis data of one, and an attack portion based on the data of the time range from the beginning of the original waveform data to immediately before the flattening range . The second waveform data includes a loop portion based on the loop range data of the second amplitude correction analysis data and the flattening range from the beginning of the original waveform data. And an attack unit based on data in a time range until immediately before the enclosure .
According to a sixth aspect of the present invention, there is provided the waveform data production apparatus according to any one of the first to fifth aspects.
The program according to claim 7 is characterized by causing a processing device to execute the waveform data production method according to any one of claims 1 to 5.
The waveform memory production method according to claim 8 is a waveform memory production method for producing a waveform memory used for generation of a musical sound signal, wherein the waveform memory according to any one of claims 1 to 5 is used. The method includes a step of executing a data production method and a step of writing the produced first and second waveform data to a recording medium.

  Thus, according to the present invention, the flattened analysis data is generated by flattening the pitch of the original analysis data, and the amplitude level corresponding to the first and second amplitude levels is set for each harmonic component. Since the first and second waveform data are combined by applying each, the first and second waveform data are generated so that the amplitude levels of these frequency components are different while substantially matching the phase and frequency of each frequency component. can do. As a result, when the musical tone signal is generated by mixing the first and second waveform data with different weights, the timbre can be smoothly changed in accordance with the weighting variation.

1． Hardware Configuration of Embodiment Next, the configuration of a waveform data generation apparatus according to an embodiment of the present invention will be described with reference to FIG. In the figure, reference numeral 1 denotes a CPU which controls other components based on a control program and control information stored in the ROM 2. A RAM 3 is used as a work area, a buffer area, or an area for storing various programs. Reference numeral 4 denotes a timer, which performs a time measuring operation and a timer interrupt to the CPU 1. An operation unit 5 includes a panel switch equipped with various operation switches and a pointing device such as a mouse. Reference numeral 6 denotes a panel display for performing various displays such as an original waveform to be processed. 7 is a MIDI interface for exchanging MIDI events with external MIDI devices, and 8 is a recording medium such as CD-ROM (Compact Disk-Read Only Memory), HD (hard magnetic disk), FD (flexible magnetic disk), etc. 9 is a driving device for accessing 9.

  Reference numeral 10 denotes a waveform memory which stores waveform data of an original waveform and waveform data created in the present apparatus, and a plurality of waveform data can be written and read out. Reference numeral 11 denotes an access management unit that manages access time slots of the waveform memory 10 so that accesses from the writing circuit 13, the sound source unit 15, or the buffer 14 to the waveform memory 10 do not collide with each other. An external waveform input terminal 12 and a writing circuit 13 sample the original waveform signal input from the external waveform input terminal 12 and write it in the waveform memory 10. Reference numeral 14 denotes a buffer, which transfers waveform data written to the waveform memory 10 from the recording medium 9 or RAM 3 or waveform data read from the waveform memory 10 to the CPU 1 or RAM 3. A sound source unit 15 generates a musical sound signal using the waveform data read from the waveform memory 10. Reference numeral 16 denotes a sound system which outputs a musical sound signal output from the sound source unit 15. Reference numeral 17 denotes a bus line, which is used for exchanging information between the above-described elements.

  Here, in ROM 2 or RAM 3, in addition to a program for performing general control, a waveform analysis program for analyzing original waveform data input from the outside, and processing of waveform data synthesized based on the analysis result, A sound source waveform creation program for editing and generating sound source waveform data, a performance processing program for causing the sound source waveform data to function as a waveform memory sound source, and the like are stored. The CPU 1 performs various controls according to inputs from the panel switch 5 and the MIDI interface 7 according to various control programs stored in the ROM 2 or the RAM 3. Further, at the time of waveform data analysis processing and sound source waveform data creation processing, the data in the waveform memory 10 is read and written via the buffer 14, the waveform data is read out, analyzed, processed, edited, and stored in the waveform memory 10 again. Writing is performed, waveform data supplied from a recording medium 9 or a communication path (not shown) is written to the waveform memory 10, and conversely, the waveform memory 10 is supplied to the recording medium 9 or the communication path.

  Further, the CPU 1 controls the tone generation state of the tone generation channel of the tone generator unit 15 according to the performance information supplied from the MIDI interface 7, the recording medium 9, the RAM 3, or the like when performing the performance process. For example, when a note-on signal indicating the start of sound generation is input from the MIDI interface 7, it is necessary to assign the generation of the musical sound to one of the sound generation channels of the sound source unit 15 and generate the music sound to the assigned sound generation channel. Various musical tone parameters (pitch information, vibrato control information, waveform selection information, volume envelope control information, effect information, etc.) are supplied and a sounding start instruction is given. In response to this, the tone generator 15 uses the assigned tone generation channel to generate a tone corresponding to the tone parameter described above using the waveform data read from the waveform memory 10 according to the waveform selection information. .

2． Operation of the embodiment
2.1. Outline of Operation First, the frequency component of the musical sound waveform of a natural musical instrument will be described with reference to FIGS. When the original waveform data in which the musical sound waveform is recorded is subjected to FFT (Fast Fourier Transform) analysis, the frequency component of the original waveform data can be separated into a frequency component that is continuous on the time axis and a frequency component that is intermittent on the time axis. When the waveform data is synthesized based on the former frequency component, the waveform data of the “period component” of the original waveform data is obtained, and when the waveform data is synthesized based on the latter frequency component, the “noise component” of the original waveform data is obtained. can get. FIG. 2A shows a saxophone musical sound waveform (original waveform data), FIG. 2B shows its periodic component, and FIG. 2C shows its noise component.

Next, an outline of processing in the present embodiment will be described with reference to FIG. FIG. 2 is a functional block diagram showing the processing (program) executed in the CPU 1 and the contents of various data.
In the figure, reference numeral 32 denotes a waveform recording unit, which stores the original waveform data 30 input via the writing circuit 13 in the RAM 3. The original waveform data 30 is played and recorded with vibrato added. Reference numeral 34 denotes an FFT analysis processing unit, which performs FFT analysis processing on the original waveform data 30. Here, first, a window function is applied to the original waveform data 30 with respect to a frame having a length of “8” times the approximate value of the pitch period, and a frequency component within the range of the frame is analyzed. Next, the position of the frame is shifted backward by “1/64” (1/8 of the approximate value of the pitch period) of the frame length on the time axis, and the frequency component is similarly analyzed. When this process is repeated for the entire original waveform data 30, changes in frequency components on the time axis are obtained.

  Of course, the pitch changes when vibrato is generated, but it is not necessary to set the frame length precisely to “8 times the pitch”. However, the frame length may be constant. According to such processing, the frequency of the frequency component, its amplitude level, and its phase are obtained for each frame. Data in such a format is referred to as “analysis data” in this specification, and the analysis data of the original waveform data 30 is referred to as “original analysis data”. An example of the distribution of frequency components in the original analysis data is shown in FIG. In FIG. 4, reference numeral 80 is a locus of pitch components, 82 is a locus of secondary harmonics, and 84 is a locus of tertiary harmonics, which are continuously generated on the time axis. As described above, the frequency component is classified into a component having a continuous locus on the time axis (hereinafter referred to as “deterministic frequency component”) and a component other than that. A continuous component separation unit 36 performs such classification according to the continuity of each analyzed frequency component. Reference numeral 50 denotes a waveform synthesizer that synthesizes deterministic waveform data (that is, periodic component waveform data in FIG. 2B) based only on analysis data of deterministic frequency components among these frequency components. Next, reference numeral 52 denotes a subtraction unit that subtracts deterministic waveform data from original waveform data. The residual waveform as the subtraction result corresponds to the noise component in FIG.

  By the way, when the original waveform data 30 is observed with a reduced time axis, as shown in a waveform shape display section 110 and a pitch change display section 112 in the setting screen 100 (FIG. 5) described later, the pitch and amplitude The fluctuation caused by the vibrato performance occurs. Reference numeral 38 denotes a flattening processing unit, which is deterministic so that the pitch of the time range (flattening region) specified by the flattening region specifying unit 22 in the time range of the deterministic frequency component is constant. The frequency in the original analysis data data relating to the frequency component is changed. The analysis data after such change is referred to as “flattening analysis data”. Here, an arbitrary point in the original waveform data where the pitch is near the maximum value (at least higher in frequency than the average value of the pitch) is called an upper point, and the pitch is near the minimum value (at least than the average value of the pitch) An arbitrary point (with a low frequency) is called a lower point. Reference numeral 24 denotes an upper point designating section, and 26 denotes a lower point designating section, each designating a plurality of upper and lower points.

  Here, the specified upper and lower points are directly specified for the original waveform data or the original analysis data, but the time axis of the original waveform data is the time axis of the flattening analysis data. Furthermore, it is common to the time axis of other various waveform data and analysis data described later. Therefore, the designated upper and lower points are also applied to the flattening analysis data and the like. Reference numeral 40 denotes an upper waveform generation unit that generates “upper deterministic component” analysis data based on amplitude levels of deterministic frequency components at a plurality of upper points in the flattening analysis data. That is, in the flattening analysis data, since the upper point is intermittently generated on the time axis, an amplitude level that is preferable as the upper point is also intermittently generated on the time axis. Therefore, by applying the amplitude level of this upper point to a certain range of analysis data, it is possible to obtain analysis data of the upper deterministic component which is preferable by using any point in this range as the upper point. is there. The analysis data obtained in this way is called “upper deterministic analysis data”. Reference numeral 42 denotes a lower waveform creation unit, which is preferably used by applying the amplitude level at a plurality of lower points to the above range so that any point in this range is used as the lower point. The analysis data of the component, that is, the “lower deterministic analysis data” is obtained.

  By the way, although the residual waveform is composed of frequency components that are cut off in time, as shown in FIG. 4, it is understood that the frequency components are not generated at all, but have a certain tendency. In the example of FIG. 4, frequency components are concentrated at about 0.5 times, 0.6 times, 1.5 times, 2.5 times the pitch frequency. A formant is formed by these frequency components. Such formants due to the residual waveform are generated at the upper point and the lower point, respectively. Reference numeral 54 denotes an upper formant continuation processing unit that creates analysis data of the upper residual waveform based on the frequency components of the residual waveform at the upper point. That is, since the upper point is intermittently generated on the time axis, the frequency component dispersion tendency of the residual waveform at the upper point is also intermittently generated on the entire time axis. Therefore, the analysis data of the upper residual waveform is created by complementing the frequency components having the same dispersion tendency in the section other than the upper point. Reference numeral 56 denotes a lower formant continuation processing unit that creates lower residual waveform analysis data in the same manner as the upper residual waveform analysis data.

  Reference numerals 44, 46, 58, and 60 denote waveform data creation units that generate corresponding waveform data based on the analysis data of the upper and lower deterministic components and the upper and lower residual waveform data, respectively. Each waveform data is composed of an attack portion that is reproduced only once at the start of sound generation, and a loop portion that is repeatedly reproduced thereafter. Reference numeral 28 denotes a loop designating unit that designates a time range (loop range) corresponding to the waveform data of the loop unit. Here, the designated loop range is common to all of the waveform data creation units 44, 46, 58 and 60. 62 and 64 are adders for the final sound source having a deterministic component and a noise component by synthesizing the deterministic waveform data and the waveform data of the residual waveform for the upper side and the lower side, respectively. Create waveform data.

  That is, in the addition unit 62, the attack parts of the upper deterministic waveform data and the upper residual waveform data are added to generate the upper attack part waveform data, and the upper deterministic waveform data and the upper residual waveform data Each loop part is added, and upper loop part waveform data is generated. Then, the upper waveform data 66 is completed by making both waveform data continuous. Similarly, in the adder 64, the attack parts of the lower deterministic waveform data and the lower residual waveform data are added to generate lower attack part waveform data, and lower deterministic waveform data and The loop portions of the lower residual waveform data are added to generate the lower loop portion waveform data. Then, the lower waveform data 68 is completed by making both waveform data continuous.

2.2. Specific operation
2.2.1. Display of Setting Screen 100 When a predetermined operation is executed in the operation unit 5, the setting screen 100 shown in FIG. In the setting screen 100, reference numeral 102 denotes a mouse cursor that moves in the setting screen 100 according to the operation state of the mouse in the operation unit 5. A waveform shape display unit 110 displays the instantaneous value level of waveform data to be processed, that is, the waveform shape, with the horizontal axis as a time axis. A pitch change display unit 112 displays the pitch change of the waveform data with the horizontal axis as a time axis. The “pitch change” is a difference from the initial value of the pitch of the waveform data. 114 is a flattening instruction button for instructing a flattening process. Reference numeral 116 denotes a trial listening button for auditioning the flattened waveform data. Reference numeral 118 denotes an upper creation button for instructing creation of the upper waveform data 66. A preview button 120 is used to audition the created upper waveform data 66.

  Reference numeral 122 denotes a lower creation button for instructing creation of the lower waveform data 68. Reference numeral 124 denotes a test listening button for testing the created lower waveform data 68. Reference numeral 126 denotes an original waveform audition button for auditioning the original waveform data 30. Reference numeral 128 denotes a loop setting button for setting loop addresses in the upper and lower waveform data 66 and 68. Reference numeral 136 denotes a read button for reading new original waveform data. An execution button 134 is provided for instructing execution of flattening processing, upper / lower waveform generation processing, and the like. A save button 132 is provided to save the created upper / lower waveform data 66 and 68 in the waveform memory 10. Reference numeral 130 denotes an end button. When this button is clicked with the mouse, the setting screen 100 is closed.

2.2.2. Reading the original waveform data When the setting screen 100 is initially displayed on the panel display 6, nothing is displayed on the waveform shape display unit 110 and the pitch change display unit 112. Here, when the read button 136 is clicked with the mouse, a selection screen for selecting whether the source of the original waveform data is the external waveform input terminal 12 or the recording medium 9 is displayed. Here, when the external waveform input terminal 12 is selected as the supply source, the audio signal received from the external waveform input terminal 12 is sampled, and the result is written as the original waveform data 30 in the waveform memory 10. On the other hand, when the recording medium 9 is selected as the supply source, a list of waveform data files stored in the recording medium 9 is displayed on the panel display 6. When the user selects an arbitrary waveform data file in this list, the contents of the waveform data file are read out and written into the waveform memory 10 as original waveform data 30.

  When the original waveform data 30 is stored in the waveform memory 10, the CPU 1 executes a program related to the FFT analysis processing unit 34. As a result, as a result of the FFT analysis processing on the original waveform data 30, for example, a change state of the frequency component of the original waveform data 30 on the time axis as shown in FIG. 4 is obtained. The FFT analysis processing result is recorded in the RAM 3 as original analysis data. Next, in the CPU 1, a program related to the waveform synthesis unit 50 is executed, and deterministic waveform data is synthesized based on the FFT analysis processing result. The synthesized deterministic waveform data is stored in the waveform memory 10. Next, in the CPU 1, a program related to the subtraction unit 52 is executed, deterministic waveform data is subtracted from the original waveform data 30, and the result is stored in the waveform memory 10 as waveform data of the residual waveform. Next, the instantaneous value at each point of the original waveform data 30 is displayed on the waveform shape display unit 110, and the lowest frequency component among the deterministic frequency components is regarded as “pitch”, and the initial value of the pitch In contrast, a pitch change that is a difference in pitch detected in each frame is displayed on the pitch change display unit 112.

2.2.3. Here, when the flattening instruction button 114 is clicked with the mouse, it becomes possible to designate a flattened area in the pitch change display unit 112. That is, as shown in FIG. 6A, in the pitch change display unit 112, it is possible to select an arbitrary time range as a flattened region by dragging with the mouse. As described above, when the flattened area is designated, the timings of the head and the end of the flattened area are displayed by the cursors 150 and 152. When the execution button 134 is clicked with the mouse in such a state, the program related to the flattening processing unit 38 is executed. As a result, the pitch in the flattened region is flattened, and the pitch change after the flattening process is displayed on the pitch change display unit 112 as shown in FIG.

  Here, the specific content of the flattening process in the flattening process part 38 is demonstrated. First, the width of the flattened area is set to “n frames”, and a frame number k (where k = 1 to n) is assigned to each frame included in the flattened area. In addition, the number of trajectories of deterministic frequency components is set to “m”, and a trajectory number q is assigned to each trajectory in a range of “1 to m” in order of decreasing frequency. In the k-th frame, the frequency related to the q-th trajectory is referred to as a frequency fkq. Here, the frequency fk1 is, for example, a pitch frequency. Next, the pitch of the top portion of the flattened area, that is, the portion corresponding to the cursor 150 is referred to as a flattened pitch frequency fp. Next, with respect to the frequency fkq (where k = 1 to n, q = 1 to m) of each trajectory of each frame in the original waveform data 30, a flat “f′kq = fkq · (fp / fk1)” is obtained. The frequency f′kq is calculated. Of the flattening frequency f'kq, the pitch frequency f'k1 corresponding to q = 1 always coincides with the flattening pitch frequency fp. The frequency corresponding to the trajectory number q of “2” or more is determined based on the ratio between the flattening pitch frequency fp and the pitch frequency fk1 of each frame in the original waveform data 30. The result of changing the frequency fkq to the flattening frequency f'kq in the analysis data of the original waveform data is stored in the RAM 3 as flattening analysis data.

2.3. Upper / Lower Analysis Data Creation Processing After the flattening analysis data is created by the flattening processing unit 38, when the upper creation button 118 is clicked with the mouse, the upper point can be designated on the pitch change display unit 112. It becomes possible. That is, as shown in FIG. 7 (a), the pitch change display unit 112 displays the pitch change of the original waveform data and the pitch change of the flattening analysis data while being superimposed on each other. When an arbitrary location is clicked with the mouse, that location is selected as the upper point. The selected upper point is displayed on the pitch change display unit 112 by the cursors 160 to 166. As described above, when the execution button 134 is clicked with the mouse in a state where a plurality of upper points are designated, the programs related to the upper waveform creation unit 40 and the upper formant continuation processing unit 54 are executed. As a result, the upper deterministic analysis data is obtained by the upper waveform creation unit 40 and the analysis data of the upper residual waveform is created by the upper formant continuation processing unit 54.

  Here, the details of the processing in the upper waveform creation unit 40 will be described with reference to FIGS. 7B and 7C. In the k-th frame in the flattening analysis data, the amplitude level of the frequency component related to the flattening frequency f'kq is referred to as Lkq. FIG. 7B shows an example of the amplitude level Lkq (the example of the figure is the level of the pitch component of q = 1) and the upper points P0 to P4 designated by the cursors 160 to 166. The upper waveform creation unit 40 converts the level Lkq at each upper point into an amplitude level LUkq shown in FIG. 7C, and outputs the result as upper deterministic analysis data. Here, the amplitude level LUkq is determined as follows. First, until reaching the first upper point P0, the amplitude level LUkq of the upper deterministic analysis data is equal to the amplitude level Lkq of the flattening analysis data, and the amplitude level LUkq is also set at each of the designated upper points P0 to P4. , Equal to the amplitude level Lkq. Next, between each of the upper points P0 to P4, the amplitude level LUkq is set to be equal to a value obtained by linearly interpolating the amplitude level LUkq of the flattening analysis data at each of the upper points P0 to P4. After the last upper point P4, the amplitude level LUkq is set to be equal to the amplitude level LUkq at the upper point P4.

  Further, when the lower creation button 122 is clicked with the mouse on the setting screen 100, the same processing is executed for the lower side. That is, in such a case, as shown in FIG. 8A, the pitch change display unit 112 displays the pitch change of the original waveform data and the pitch change of the flattening analysis data while being superimposed. . When an arbitrary place in the flattened area is clicked with the mouse, the place is selected as a lower point, and these are displayed by the cursors 170 to 178. When the execution button 134 is clicked with the mouse in such a state, the program related to the lower waveform creation unit 42 and the lower formant continuation processing unit 56 is executed, and the lower deterministic analysis data and the lower residual waveform are displayed. Analysis data is created. Here, the details of the processing in the lower waveform creation unit 42 will be described with reference to FIGS. 8B and 8C. The lower waveform creation unit 42 converts the amplitude level Lkq related to the qth trajectory of the flattened deterministic frequency component into the amplitude level LDkq. The lower points of the cursors 170 to 178 are R0 to R5, and the amplitude level LDkq of the lower deterministic analysis data is equal to the amplitude level Lkq until reaching the first lower point R0, and each lower point R0 to R5. The amplitude level LDkq is equal to the amplitude level Lkq. Next, between the lower points R0 to R5, the amplitude level LDkq is set to be equal to a value obtained by linear interpolation of the amplitude levels LDkq at the lower points R0 to R5. After the last lower point R5, the amplitude level LDkq is set to be equal to the amplitude level LDkq at the lower point R5.

2.4. Loop Setting As described above, after the upper and lower deterministic analysis data having the amplitude level LUkq and the amplitude level LDkq are obtained, when the loop setting button 128 is clicked with the mouse, the loop is displayed in the pitch change display unit 112. A range can be specified. That is, as shown in FIG. 9, the pitch change display unit 112 displays the pitch change before and after the flattening process while being superimposed, and further, the cursors 160 to 166 indicating the upper points P0 to P4 and / or Cursors 170 to 178 indicating the lower points R0 to R5 are displayed. Here, in the pitch change display section 112, an arbitrary time range after the top upper point P0 and after the top lower point R0 (however, a time that is an integral multiple of the pitch period after flattening) The range) can be dragged with the mouse to select the loop range.

  As described above, when the loop range is designated, the timings of the beginning and end of the loop range are displayed by the cursors 180 and 182. The loop range set here is a common loop range for the waveform data creation units 44, 46, 58 and 60. When the execution button 134 is clicked with the mouse while the loop range is set, the program related to the waveform data creation units 44, 46, 58, and 60 is executed. Further, in the waveform data creation units 44, 46, 58, 60, the time range from the beginning of the waveform data to immediately before the flattened area is set to a common “attack range”. Accordingly, the pitch frequency at the end portion of the attack range matches the pitch frequency at the loop portion (that is, the flattened pitch frequency fp).

  First, in the waveform data creation unit 44, the flattening frequency f'kq of each harmonic component of the upper deterministic analysis data is corrected. The details of the correction will be described. As described above, the loop range is an integral multiple of the pitch period. If the multiple is A, the phase of the pitch component advances by “2Aπ” within the loop range. Therefore, the phase of the n-th overtone should advance by “2nAπ”, but actually, the frequency fkq is corrected to the flattening frequency f′kq during the flattening process, so that the upper deterministic analysis data The advance of the phase of the nth harmonic within the loop range does not exactly match “2nAπ”. Here, in accordance with the flattening frequency f′kq, the phase in which the n-th overtone actually travels is represented by a solid line in FIG. The result of multiplying the ratio of the actually advanced phase and “2nAπ” by the flattening frequency f′kq of each frame is referred to as a corrected flattening frequency f ″ kq. The phase within the loop range by the corrected flattening frequency f ″ kq Is exactly “2nAπ” as indicated by the broken line in FIG.

  Further, the amplitude level LUkq of the pitch component and each harmonic component related to the upper deterministic analysis data is corrected so as to coincide at the loop start point and the loop end point. The corrected amplitude level is referred to as a corrected amplitude level LU'kq. The contents of the correction method will be described with reference to FIG. It is assumed that the original amplitude level LUkq within the loop range has changed as shown by the solid line in FIG. Then, assuming that the difference in amplitude level between the loop start point and the loop end point is B, the length of the loop range is TL, and the elapsed time from the loop start point is T, the corrected amplitude level LU′kq in each frame is “ LU′kq = LUkq + B (T / TL) ”is set. When the corrected flattening frequency f ″ kq and the corrected amplitude level LU′kq are obtained in this way and then the waveform data is synthesized based on these, the loop portion waveform data corresponding to the upper deterministic analysis data is obtained. On the other hand, in the waveform data creation unit 58 related to the residual waveform, the analysis data of the loop range described above is cut out from the analysis data of the upper residual waveform generated by the upper formant continuation processing unit 54 and cut out. The converted analysis data is converted into waveform data, and the converted waveform data is added to the loop data of the upper deterministic component in the adder 62, thereby completing the loop of the upper waveform data 66. .

  In the waveform data creation unit 44, the following correction is performed on the upper deterministic analysis data of the attack range. That is, first, the frequency of each harmonic component of the attack part is changed so as to gradually approach the corrected flattening frequency f ″ kq of each harmonic component of the loop part at the end of the attack part and finally match. The harmonic component levels of these attack parts are changed so that they gradually approach the corrected amplitude level LU′kq of each harmonic component of the loop part at the end of the attack part, and finally coincide with each other. When the waveform data is synthesized based on the upper deterministic analysis data of the attack range modified as described above, the attack portion waveform data of the upper deterministic component is generated, while the waveform data creation unit 58 related to the residual waveform is generated. Is extracted from the analysis data of the upper residual waveform generated by the upper formant continuation processing unit 54. The analysis data of the attack range described above is extracted. The analyzed data is converted into waveform data, and the converted waveform data is added to the attack part waveform data of the upper deterministic component in the adder 62, thereby completing the attack part of the upper waveform data 66.

  Although the method for creating the upper waveform data 66 has been described in detail above, the lower waveform data 68 is created in exactly the same manner. That is, in the waveform data creation unit 46, the flattening frequency f′kq of each harmonic component of the lower deterministic analysis data is corrected to obtain a corrected flattening frequency f ″ kq. Also, the amplitude of each harmonic component The level LDkq is modified to the modified amplitude level LD′kq so as to match at the loop start point and the loop end point. Note that the modified flattening frequency f ″ kq is relative to the upper and lower deterministic components. The corrected amplitude levels LU′kq and LD′kq are different values on the upper side and the lower side. Next, in the waveform data creation unit 46, loop portion waveform data of the lower deterministic component is obtained based on the modified flattening frequency f ″ kq and the modified amplitude level LD′kq. In the creation unit 60, the analysis data of the loop range is extracted from the analysis data of the lower residual waveform generated by the lower formant continuation processing unit 56, and is converted into waveform data. The adder 64 adds the lower deterministic component loop data to the loop waveform data, thereby completing the lower waveform data 68 loop.

  Further, in the waveform data creation unit 46, the frequency of each harmonic component of the lower deterministic component of the attack unit gradually increases to the corrected flattening frequency f ″ kq of each harmonic component of the loop unit at the end of the attack unit. The levels of the harmonic components of these attack parts gradually approach the corrected amplitude level LD′kq of each harmonic component of the loop part at the end of the attack part. When the waveform data is synthesized based on the lower deterministic analysis data of the attack range thus modified, the attack part waveform data of the lower deterministic component is generated. On the other hand, in the waveform data creation unit 60, among the analysis data of the lower residual waveform generated by the lower formant continuation processing unit 56, the analysis data of the attack range described above. Is cut out, and the extracted analysis data is converted into waveform data, and the converted waveform data is added to the attack part waveform data of the lower deterministic component in the adder 64, thereby lower waveform data. 68 attack parts are completed.

2.5. Audition When the audition button 116 corresponding to the flattening analysis data is clicked with the mouse, waveform data is synthesized based on the flattening analysis data, and the synthesized waveform data is generated via the sound source unit 15 and the sound system 16. The Similarly, when the audition button 120 or 124 related to the upper and lower deterministic analysis data is clicked with the mouse, the waveform data is synthesized based on the corresponding analysis data, and the synthesized waveform data is the sound source unit 15. The sound is generated via the sound system 16. When the original waveform audition button 126 is clicked with the mouse, the original waveform data 30 is sounded. Thereby, the user can easily confirm whether or not analysis data is generated in a desired state in each process.

2.6. Music signal synthesis After the upper and lower waveform data 66 and 68 are stored in the waveform memory 10, for example, the MIDI interface 7 instructs the waveform data generating device to generate a sound corresponding to these waveform data. When the signal is supplied, the CPU 1 instructs the sound source unit 15 to sound the waveform data. In response to this, the tone generator 15 synthesizes a tone signal having vibrato based on the algorithm shown in FIG. In FIG. 11, in the waveform memory 10, the area where the upper waveform data 66 is stored is referred to as an upper area 10a, and the area where the lower waveform data 68 is stored is referred to as a lower area 10b.

  The sound generation instruction from the CPU 1 includes waveform address information WA indicating an attack portion and a loop portion of waveform data used for sound generation, pitch shift information PS0 (cent) of the waveform data, and a volume envelope to be added to the waveform data. Waveform envelope information WE indicating the shape of the waveform. Here, the pitch shift information PS0 will be described. First, each waveform data stored in the waveform memory 10 is obtained by sampling a musical tone signal having a certain pitch at a predetermined sampling period. If the sampling period at the time of recording is equal to the sampling period at the time of reproduction, if the waveform data is read by “1” samples at each sampling period, the read tone signal has the same pitch as that at the time of recording. . However, when actually reading the waveform data, the pitch to be generated differs from the pitch at the time of sampling. The pitch shift information PS0 represents a difference in cents between a pitch when the waveform data is read out by “1” samples for each sampling period and a pitch to be actually generated. For example, if the former pitch is “C3” and the latter pitch is “C4”, the pitch shift information PS0 is “1200” (cents).

  Reference numeral 208 denotes a low-frequency oscillator that outputs a triangular wave having the vibrato frequency fv in accordance with the vibrato frequency fv. This triangular wave is called a vibrato control signal Sv, and its maximum value is “1” and its minimum value is “−1”. Reference numeral 206 denotes a multiplier. When a maximum pitch shift fluctuation amount Δfpsmax, which is the maximum fluctuation amount to be added to the pitch shift information PS0 (cent), is designated, the vibrato control signal Sv is multiplied by the maximum pitch shift fluctuation amount Δfpsmax. To do. The pitch shift fluctuation amount Δfps (cent) as the multiplication result is added to the pitch shift information PS0 in the adder 200, and the adder 200 outputs the pitch shift information PS1 with vibrato added. A linear conversion unit 201 converts pitch shift information PS1 that is frequency ratio information of a cent scale into an f number that is frequency ratio information of a linear scale. Reference numeral 202 denotes a phase generator, which generates and outputs phase information PH by accumulating the f number from the linear conversion unit 201 every sampling period. Reference numeral 204 denotes an address generator, which outputs a common address signal AD for reading out the upper and lower regions 10a and 10b based on the phase signal PH and the waveform address information WA. As described above, the upper and lower waveform data 66 and 68 stored in the upper and lower regions 10a and 10b are each composed of an attack portion and a loop portion. In the address generator 204, an address signal AD is generated so that the attack portion is read only once at first, and then the loop portion is repeatedly read according to the phase related to the phase signal PH and the waveform address information WA. The

As a result, the upper and lower musical tone signals having the same phase and frequency-modulated by the pitch shift fluctuation amount Δfps are read out from the upper and lower regions 10a and 10b of the waveform memory 10 at each sampling period. become. The read upper and lower tone signals are weighted by multipliers 214 and 218 and synthesized by adder 222. The synthesized tone signal is subjected to amplitude modulation in the multiplier 224, and the result is output to the sound system 16 as a tone signal Sout having vibrato. A delay circuit 210 delays the vibrato control signal Sv by a time corresponding to the designated phase difference φ1. A multiplier 212 multiplies the delayed vibrato control signal Sv by the maximum weighting coefficient ΔG1max and outputs the result as a weighting coefficient ΔG1. The maximum weighting coefficient ΔG1max is a coefficient set in the range of “0 to 0.5”. The weighting coefficient ΔG1 is a triangular wave signal that varies within a range of ± ΔG1max.

  Reference numeral 216 denotes an adder, which adds the weighting coefficient ΔG1 and the basic weight value “0.5” and supplies the addition result to the multiplier 214 as the upper weight value GU. An adder 220 subtracts the weighting coefficient ΔG1 from the basic weight value “0.5” and supplies the subtraction result to the multiplier 218 as the lower weight value GD. As a result, it can be seen that the total weight (GU + GD) given to the upper and lower musical tone signals is always “1”. When the phase difference φ1 is set to “0”, the delay time in the delay circuit 210 is also set to “0”, the timing at which the traveling speed of the address signal AD is maximized (the timing at which the peak of the pitch of the tone signal appears), and multiplication The timing at which the weighting of the upper musical tone signal is maximized in the unit 214 is completely coincident. Further, by specifying a value other than “0” as the phase difference φ1, it is also possible to shift the timing of both.

  Next, a delay circuit 226 delays the vibrato control signal Sv by a time corresponding to the designated phase difference φ2. Here, the phase difference φ2 is the phase difference between the pitch vibrato and the amplitude vibrato applied to the musical tone signal. A multiplier 228 multiplies the delayed vibrato control signal Sv by the maximum amplitude variation coefficient ΔG2max and outputs the result as an amplitude variation coefficient ΔG2. The maximum amplitude variation coefficient ΔG2max is a coefficient set in the range of “0 to 1.0”. The amplitude variation coefficient ΔG2 is a triangular wave signal that varies within a range of ± ΔG2max. An adder 230 adds the amplitude variation coefficient ΔG 2 and the basic amplitude coefficient “1”, and supplies the gain Gout as the addition result to the envelope generation unit 232. Based on the waveform envelope information WE, the envelope generator 232 generates a volume envelope waveform for controlling the time change of the volume from the rising edge to the falling edge of the musical sound signal, and combines the volume envelope waveform with the gain Gout. This is supplied to a multiplier 224 as a volume envelope waveform ED subjected to amplitude modulation. In the multiplier 224, the tone signal output from the adder 222 is subjected to amplitude control by the volume envelope ED, and the result is supplied to the sound system 16 as a tone signal Sout for each sampling period.

  Here, FIGS. 12A and 12B show examples of the pitch shift fluctuation amount Δfps and the weighting coefficient ΔG1 when shallow vibrato is applied and when deep vibrato is applied. In this embodiment, both vibratos can be realized by simply switching the parameters of the maximum pitch shift fluctuation amount Δfpsmax and the maximum weighting coefficient ΔG1max. These parameters do not need to be fixed during sound generation, and can be changed variously during sound generation. Therefore, for example, a so-called delay vibrato that gradually transitions from the shallow vibrato state of FIG. 12A to the deep vibrato state of FIG. 12B can be easily realized.

2.7. Mounting on the final product The waveform data generator of this example can be used as it is as a sound source or an electronic musical instrument, but an inexpensive final product generally has a function for creating new waveform data for a sound source. Not done. Therefore, the sound source waveform data (upper / lower waveform data 66, 68) obtained as described above is written in a waveform memory (semiconductor memory, CD-ROM, etc.), and the sound source, electronic musical instrument, synthesizer, etc. Implemented in the final product. In these final products, an algorithm equivalent to that described with reference to FIG. 11 is executed, whereby a tone signal can be synthesized while freely setting parameters such as the vibrato depth and period.

3． Modifications The present invention is not limited to the above-described embodiments, and various modifications can be made as follows, for example.
(1) A communication interface circuit for connecting to a communication network such as a LAN or the Internet is provided for the waveform data generation device of the above embodiment, and waveform data and various programs are downloaded from the server via the communication network, or The generated sound source waveform data may be supplied to another device. Furthermore, a keyboard operator or the like can be connected to the waveform data generation device, and a performance can be performed using this keyboard operator. In the above description, a program for causing the CPU 1 to execute waveform analysis, sound source waveform creation, performance processing, and the like is stored in the ROM 2 or the RAM 3. Alternatively, the program may be executed by receiving an external supply from a CD-ROM (recording medium 9) and installing it on a hard magnetic disk (recording medium 9). Alternatively, the program may be executed by being downloaded from a server on a network to a hard magnetic disk (recording medium 9) via a communication line (not shown).

(2) Further, in the above embodiment, the attack range is a section immediately before the flattened region, but the attack range may be set independently for the flattened region. In this case, it is preferable that the frequency of each frequency component in the attack range and the amplitude level thereof are corrected so as to be continuous with these values at the head portion of the loop portion, and the correction result is used as the attack portion of the waveform data.

(3) Although a plurality of upper points P0 to P4 and a plurality of lower points R0 to R5 are specified in the above embodiment, it is not always necessary to specify a plurality of these points. That is, the upper point and the lower point are designated one by one for the flattening analysis data, and the amplitude level Lkq at each one point is set as the amplitude level LUkq of the entire upper point, or the amplitude level of the entire lower point. You may set to LDkq.

(4) Further, the upper points P0 to P4 and the lower points R0 to R5 do not need to be designated by the user, and the maximum and minimum values of the pitch of the original waveform data are automatically detected to detect one or a plurality of points. The upper and lower points may be determined automatically.

(5) In the above embodiment, the sound source waveform data has the attack part and the loop part. However, the loop part is not necessarily provided, and the sound source waveform data is constituted only by the attack part. Also good. For example, in a percussion instrument such as a vibraphone, an approximate maximum time from the start of sound generation to the end of sound generation can be predicted. Therefore, it is preferable to secure an attack portion for the length of this maximum time.

(6) In the above embodiment, two types of sound source waveform data are generated. However, the number of sound source waveform data is not limited to “2”, and three or more types of waveform data are generated and reproduced. Sometimes these may be mixed.

1 is a block diagram of a waveform data generation device according to an embodiment of the present invention. It is a figure which shows the example of the frequency component of the musical sound waveform of a natural musical instrument. It is a block diagram of processing contents and data according to an embodiment. It is a figure which shows the example of distribution of the frequency component in original analysis data. It is a figure which shows the setting screen 100 displayed on the panel indicator. It is a figure which shows the example of a display of the pitch change display part 112 in the planarization process. FIG. 6 is an operation explanatory diagram of an upper waveform creation unit 40. FIG. 11 is an operation explanatory diagram of the lower waveform creation unit. 6 is an operation explanatory diagram of the loop designating unit 28. FIG. FIG. 6 is an operation explanatory diagram of waveform data creation units 44 and 46; It is a block diagram of the algorithm applied to the sound source part 15 at the time of reproduction | regeneration of a musical tone signal. 4 is a waveform diagram of each part of an algorithm applied to a sound source unit 15. FIG.

Explanation of symbols

  1: CPU, 2: ROM, 3: RAM, 4: Timer, 5: Operation unit, 6: Panel display, 7: MIDI interface, 8: Drive device, 9: Recording medium, 10: Waveform memory, 10a: Upper side Area: 10b: Lower area, 11: Access management section, 12: External waveform input terminal, 13: Write circuit, 14: Buffer, 15: Sound source section, 16: Sound system, 17: Bus line, 22: Flattening Area designation unit, 24: upper point designation unit, 26: lower point designation unit, 28: loop designation unit, 30: original waveform data, 32: waveform recording unit, 34: FFT analysis processing unit, 36: continuous component separation unit , 38: flattening processing unit, 40: upper waveform creation unit, 42: lower waveform creation unit, 44, 46, 58, 60: waveform data creation unit, 50: waveform synthesis unit, 52: subtraction unit, 54: upper side Formant Processing unit, 56: lower formant continuation processing unit, 62, 64: addition unit, 66: upper waveform data, 68: lower waveform data, 80, 82, 84: locus, 100: setting screen, 102: mouse Cursor: 110: Waveform shape display section, 112: Pitch change display section, 114: Flattening instruction button, 116, 120, 124: Audition button, 118: Upper creation button, 122: Lower creation button, 126: Original waveform audition Button: 128: Loop setting button, 130: End button, 132: Save button, 134: Execution button, 136: Read button, 150, 152, 160-166, 170-178: Cursor, 180, 182: Cursor, 200: Adder, 201: linear converter, 202: phase generator, 204: address generator, 206: multiplier 08: low-frequency oscillator, 210: delay circuit 214, 218: Multiplier, 220, 222: adder, 224, 228: Multiplier, 226: delay circuit 232: the envelope generator.

Claims (8)

Analyzing the frequency component of the original waveform data with varying pitch and analyzing the frequency of the pitch component and harmonic component frequency, and obtaining the original analysis data consisting of the amplitude level trajectory,
The original analysis data so that the frequency of the pitch component is a constant flattening pitch frequency within a flattening range that is a partial time range in the original analysis data and does not include the head portion of the original analysis data. A flattening process to obtain flattening analysis data by correcting
Within the flattening range, a point on the time axis at which the frequency of the pitch component of the corresponding point in the original analysis data becomes higher than the average value of the frequency of the pitch component in the original analysis data by the user operation When one or more first type points are specified, a first amplitude level that is an amplitude level in the flattening analysis data of the pitch component and the harmonic component is acquired for each first type point. Amplitude level acquisition process of
Within the flattening range, a point on the time axis at which the frequency of the pitch component of the corresponding point in the original analysis data is lower than the average value of the frequency of the pitch component in the original analysis data by the user operation When one or more second type points are designated, a second amplitude level that is an amplitude level in the flattening analysis data of the pitch component and the harmonic component is acquired for each second type point. Amplitude level acquisition process of
First amplitude correction analysis data is generated by applying an amplitude level corresponding to the first amplitude level to points other than the first type point of the pitch component and the harmonic component in the flattening analysis data. 1 amplitude correction process,
Second amplitude correction analysis data is generated by applying an amplitude level corresponding to the second amplitude level to points other than the second type point of the pitch component and the harmonic component in the flattening analysis data. 2 amplitude correction process;
And combining the first waveform data based on the first amplitude correction analysis data and combining the second waveform data based on the second amplitude correction analysis data. Waveform data production method. In the first amplitude correction process, after the last point on the time axis among the first type points, the amplitude levels of the pitch component and the harmonic component of the flattening analysis data are related to the first point related to the last point. Generating the first amplitude correction analysis data by changing the flattening analysis data so that the amplitude level becomes
In the second amplitude correction process, the amplitude level of the pitch component and the harmonic component of the flattening analysis data after the last point on the time axis among the second type points is the second point related to the last point. Generating the second amplitude correction analysis data by changing the flattening analysis data so as to have an amplitude level of
The time range that is an integral multiple of the period related to the flattening pitch frequency is selected as a loop range from the time point after the first point of the first type point and from the time point after the first point of the second type point. A loop range specifying process for specifying
The first waveform data includes a loop portion based on the data of the loop range in the first amplitude correction analysis data, and data of a time range from the beginning of the original waveform data to immediately before the flattening range. The second waveform data includes a loop portion based on the data of the loop range of the second amplitude correction analysis data, and the flatness from the top of the original waveform data. The waveform data production method according to claim 1, further comprising an attack unit based on data in a time range immediately before the conversion range. A plurality of the first type points and the second type points are designated respectively.
In the first amplitude correction process, in the first amplitude correction analysis data, the amplitude levels of the pitch component and the harmonic component at the plurality of first type points are respectively set to the corresponding first amplitude levels. And an amplitude level obtained by interpolating the amplitude level of the pitch component and the harmonic component at the intermediate point between the first type points and the amplitude level of the pitch component and the harmonic component at the first type point before and after the intermediate point. To set the level,
In the second amplitude correction process, in the second amplitude correction analysis data, the amplitude levels of the pitch component and the harmonic component at the plurality of second type points are respectively set to the corresponding second amplitude levels. Amplitude obtained by interpolating the amplitude level of the pitch component and harmonic component at the intermediate point between the second type points and the amplitude level of the pitch component and harmonic component at the second type point before and after the intermediate point. To set the level,
The synthesis process includes a modified amplitude that is an amplitude level set by interpolating between the amplitude level of the pitch component and the harmonic component at the first type point and the pitch component and the harmonic component at the first type point. The first waveform data is synthesized based on the level and interpolated between the amplitude level of the pitch component and harmonic component at the second type point and the pitch component and harmonic component at the second type point. The waveform data production method according to claim 1, wherein the second waveform data is synthesized based on a corrected amplitude level that is a set amplitude level. In the first amplitude correction process, after the last point on the time axis among the first type points, the amplitude levels of the pitch component and the harmonic component of the flattening analysis data are related to the last point. Generating the first amplitude correction analysis data by changing the flattening analysis data to a first amplitude level;
In the second amplitude correction process, the amplitude level of the pitch component and the harmonic component of the flattening analysis data is related to the last point after the last point on the time axis among the second type points. The waveform data production method according to claim 3, wherein the second amplitude correction analysis data is generated by changing the flattening analysis data so as to become a second amplitude level. The time range that is an integral multiple of the period related to the flattening pitch frequency is selected as a loop range from the time point after the first point of the first type point and from the time point after the first point of the second type point. A loop range specifying process for specifying
The first waveform data includes a loop portion based on the data of the loop range in the first amplitude correction analysis data, and data of a time range from the beginning of the original waveform data to immediately before the flattening range. The second waveform data includes a loop portion based on the data of the loop range of the second amplitude correction analysis data, and the flatness from the top of the original waveform data. 5. The waveform data production method according to claim 3, further comprising: an attack unit based on data in a time range up to immediately before the conversion range .   6. A waveform data production apparatus for executing the waveform data production method according to claim 1.   A program for causing a processing device to execute the waveform data production method according to any one of claims 1 to 5. A waveform memory production method for producing a waveform memory used to generate a musical sound signal,
A process of executing the waveform data production method according to any one of claims 1 to 5,
And a step of writing the produced first and second waveform data to a recording medium.

Images