JP3832147B2 - Song data processing method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention specifies the pitch and length of a plurality of musical sounds recorded on a recording medium.pluralSong data that changes the tone characteristics such as pitch, amplitude, and tone of each tone that is played based on the song data, and changes the tone characteristics such as pitch, amplitude, tone, etc. How to processTo the lawRelated.
[0002]
[Prior art]
Conventionally, it is possible to generate a series of musical sounds by playing music data composed of a series of note data that are recorded on a recording medium and respectively specify the pitches and lengths of a plurality of musical sounds. well known. In this case, in order to give a real feeling to the generated musical sound, the musical sound characteristics such as the pitch, amplitude, and tone color of the musical sound are changed with time in accordance with control signals from an envelope generator, a low frequency transmitter, and the like.
[0003]
[Problems to be solved by the invention]
However, the time change of the musical sound characteristics imparted by the above conventional method is simple, and it is difficult to generate a realistic musical sound like a natural musical instrument. Especially, the musical sound having various performance expressions is reproduced. I couldn't. On the other hand, it is possible in principle to perform replay method simulation by inputting continuous data such as pitch bend data and control change data into the music data using a MIDI (Musical Instrument Digital Interface) sequencer. However, a normal MIDI sequencer has only the functions of continuous data input, such as typing one point at a time, drawing a straight line, drawing a certain form of curve, and freehand input. It is difficult to input control data for musical tones that have a feeling and rich performance expression.
[0004]
SUMMARY OF THE INVENTION
  The present invention has been made to address the above-described problems, and its purpose is to specify the pitch and length of a plurality of musical sounds recorded on a recording medium.pluralBy processing music data consisting of note data in a simple way, the musical sound characteristics such as pitch, amplitude, and timbre of the musical sound to be played are changed over time to create a musical sound that has a realistic feel and rich performance expression. An object of the present invention is to provide a music data processing method that enables reproduction.
[0008]
  In order to achieve the above object, a first structural feature of the present invention is that a plurality of musical tone modification data each representing a time change of at least one musical tone characteristic among pitch, amplitude, and timbre relating to one musical tone are represented by time. Divided into multiple parts on the axis and dividedThe end values of each part are all the same valueEach part is processed and prepared in advance so that the at least one musical sound characteristic of the musical sound represented by the note data in the music data is changed as a musical sound modification data for changing the time. Select multiple parts for different positions on the time axis from the parts, and connect the selected multiple parts to form musical sound modification data for changing the at least one musical sound characteristic of one musical sound over time The musical tone modification data thus formed is pasted into the music data.
[0009]
  According to the first structural feature, in order to form musical tone modification data for one musical tone, a plurality of parts at different positions on the time axis are selected and the selected plurality of parts are connected. As described above, various musical tone modification data can be formed, and a musical tone rich in realism, rich in performance expression, and varied can be generated. In addition, each part prepared in advance is divided into multiple partsThe end values of each part are all the same valueTherefore, it is easy to connect each part no matter which part is selected.
[0010]
  In this case, the length of any one of the plurality of connected parts may be adjusted according to the length of the note represented by the note data in the music data. According to this, it is possible to easily form musical tone modification data for notes having various note lengths without preparing musical tone modification data corresponding to various note lengths.
  Further, in the music data processing method according to the invention of the above-described configuration and other inventions to be described later, the musical sound modification data is obtained by analyzing the pitch, amplitude and tone color of a natural musical instrument sound. According to this, it becomes possible to easily generate a sound quality equivalent to a performance sound of a natural musical instrument.
[0011]
  The second structural feature of the present invention is that musical tone modification data representing a time change of at least one musical tone characteristic among pitch, amplitude, and timbre related to one musical tone is prepared in advance. In order to change the musical tone characteristics of the musical tone represented by the note data over time, the prepared musical tone modification data is pasted into the song data,When a plurality of musical sounds are generated at the same time by a plurality of note data stored in one track in the song data, the portion where the plurality of musical sounds are generated at the same time,The musical tone modification data relating to one musical tone among the plurality of musical sounds generated at the same time is selected, and the selected musical tone modification data is shared by a plurality of note data respectively corresponding to the plurality of musical sounds generated simultaneously. It is to be pasted on.
[0012]
  The third structural feature of the present invention is that musical tone modification data representing a time change of at least one musical tone characteristic among pitch, amplitude, and timbre related to one musical tone is prepared in advance. In order to change the musical tone characteristics of the musical tone represented by the note data over time, the prepared musical tone modification data is pasted into the song data,When a plurality of musical sounds are generated at the same time by a plurality of note data stored in one track in the song data, the portion where the plurality of musical sounds are generated at the same time,The musical tone modification data relating to the plurality of musical sounds generated at the same time are synthesized, and the synthesized musical tone modification data is pasted in common to the plurality of note data respectively corresponding to the plurality of musical sounds generated at the same time. It is to have done.
[0013]
  According to a fourth structural feature of the present invention, musical tone modification data representing a time change of at least one musical tone characteristic among pitch, amplitude, and tone color relating to one musical tone is prepared in advance.In order to temporally change the at least one musical tone characteristic of the musical tone represented by the note data in the musical piece data, the prepared musical tone modification data is pasted into the musical piece data, and at the time of pasting, the musical piece data When multiple musical sounds are generated simultaneously by multiple note data stored in one track,A plurality of note data corresponding to the plurality of musical sounds generated simultaneously.SeparateMultiple tracks eachRemarksRememberBeforeMultiple tracksRespectivelyRememberedpluralThe prepared musical tone modification data is pasted on the note data.
[0014]
  these2nd to 4thAccording to the structural features of the song dataOccur on one track at a timeMultiple musical soundsMultiple corresponding toEven if note data is included, the tone characteristic control data by multiple tone modification dataOne trackIt no longer exists on the time axis, and it is possible to easily add to the tone a time change in the tone characteristic based on the tone modification data.
[0015]
  Furthermore, the present invention5thThe structural features of each specify the pitch and length of multiple musical sounds.MultipleNote data andMultipleIn a music data processing method for processing music data consisting of tempo data representing the playback tempo of the note data of a plurality of music data, a plurality of music sound modifications representing a time change of at least one music characteristic among pitch, amplitude and tone color relating to one music sound Data is prepared in advance, and in order to change the musical tone characteristic of the musical tone represented by the note data in the song data over time, the prepared musical tone modification data is compressed or expanded on the time axis according to the tempo data. This is because it is pasted into the song data.
[0016]
  this5thAccording to the above structural feature, even if the tempo of the musical sound reproduction changes and the absolute sounding time of one musical sound changes, the musical sound modification data can be appropriately applied to the musical sound.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a musical tone generator to which a music data processing method according to the present invention is applied.
[0018]
The musical tone generator includes a CPU 11, a ROM 12, and a RAM 13 that are connected to the bus 10 and constitute a computer main body. The CPU 11 inputs from the outside by executing a program stored in the ROM 12 and, if necessary, a program recorded in the hard disk 14 or an external recording device 15 such as a CD or MO and transferred to the RAM 13 when used. To create musical tone modification data for controlling the musical tone characteristics such as pitch, amplitude, tone color, etc. of the generated musical tone by analyzing the musical instrument sound, or to the music data consisting of a series of note data representing the pitch and length The modification data is pasted or the music data with the same musical tone modification data pasted is reproduced to generate a musical sound.
[0019]
The hard disk 14 and the external recording device 15 record various data described above and later, and the hard disk 14 is incorporated in a drive device 14 a connected to the bus 10. The external recording device 15 is selectively assembled to the drive device 15 a connected to the bus 10.
[0020]
The bus 10 is also connected with a capturing circuit 21, a tone generator circuit 22, a MIDI (Musical Instrument Digital Interface) interface 23, a key input device 24, and a display 25. The capture circuit 21 includes an A / D converter 21a, and converts an analog signal supplied to the external signal input terminal 26 into digital waveform data at a predetermined sampling rate in accordance with an instruction from the CPU 11. The converted waveform data is written to the RAM 13, the hard disk 14, the external recording device 15 and the like as appropriate.
[0021]
The tone generator circuit 22 is provided with a plurality of time division channels for forming musical tone signals, respectively, and the digital musical tone signals respectively formed by these time division channels are converted into analog signals by a built-in D / A converter. The circuit 22 is connected to a waveform memory 27 and a sound system 28. The waveform memory 27 and the sound system 28 are connected to the circuit 22. The waveform memory 27 stores a plurality of types of musical sound waveform data for use in forming the musical sound signal. Each musical sound waveform data is composed of a plurality of waveform sample values to which amplitude envelopes from the attack start to the release end of the musical sound are added. Also, only the waveforms of the attack and sustain sections (loop sections) with amplitude envelopes are prepared as musical sound waveform data, and when the musical sound is generated, the waveform of the sustain section can be used as a release waveform, or as musical sound waveform data Only the waveform of the sustain portion (loop portion) may be prepared, and the waveform of the sustain portion may be used as an attack waveform and a release waveform when a musical sound is generated. The sound system 28 is composed of an amplifier, a speaker, and the like, and amplifies an analog musical tone signal from the sound source circuit 22 to generate a musical tone.
[0022]
The MIDI interface 23 is connected to other musical tone generation control devices such as a musical instrument such as a keyboard, other musical instruments, a personal computer, an automatic performance device (sequencer), and the like. High information) Various for controlling the generation of musical tone such as KC, key-on signal, key-off signal, note length information, key-on time, velocity information (key touch information), tone color control information, performance tempo information, effect control information Input musical tone control information. The key input device 24 includes a keyboard, a mouse, and the like, and performs various instructions to the CPU 11 or directly inputs various data according to the display on the display unit 25 or independently. The display device 25 displays characters and figures in accordance with instructions from the CPU 11.
[0023]
Next, using the musical tone generator configured as described above, (a) a musical tone modification data creation mode for creating musical tone modification data for time-varying musical tone characteristics comprising the pitch, amplitude and tone of the musical tone; (b) A music data processing mode for processing the music data by pasting the created music sound modification data to music data consisting of a series of note data that respectively specify the pitch and length of a plurality of music sounds; and (c) Each will be described in the order of the tone generation mode for generating a tone based on the processed music data. Each of these modes is set by the user operating the key input device 24 alone or in accordance with an instruction on the display 25.
[0024]
a. Music modification data creation mode
FIG. 2 is a functional block diagram showing the operation of the musical tone generator of FIG. 1 in the musical tone modification data creation mode. First, in this mode, the microphone 30 is connected to the external signal input terminal 26, and external sound (for example, natural musical instrument sound) is input to the capture circuit 21 through the microphone 30. Instead of the microphone 30, a recording device such as a tape recorder that records external sound in advance may be connected to the external signal input terminal 26 to input the external sound signal to the capturing circuit 21.
[0025]
The external sound signal is sampled at a predetermined sampling rate, A / D converted by the A / D converter 21a, and stored in the waveform data storage area 32 by the recording control means 31. In this case, the A / D converter 21a is provided in the capturing circuit 21, and the recording control means 31 responds to an instruction using the key input device 24, and the drive device 14a (or the drive device 15a) and This is realized by program processing (not shown) by the CPU 11 cooperating with the fetch circuit 21. The waveform data storage area 32 is an area provided in the hard disk 14 or the external recording device 15 for storing sampling data of an external sound signal.
[0026]
In this case, various natural musical instruments are played in various modes, and waveform data relating to the performance sound is accumulated and recorded in the waveform data storage area 32. For example, a sound with a strong trumpet, a sound with a medium strength, a sound with a weak blow, a sound with a stutter, a sound with a slur, a sound with a fast rise, a blow with a slow rise Record waveform data related to sound.
[0027]
After recording the waveform data, in response to the operation of the key input device 24, the CPU 11 analyzes the recorded waveform data by a program process (not shown). This analysis process is shown in FIG. 2 as a function block diagram as the analysis means 33. The analysis means 33 analyzes each recorded waveform data, and the tone characteristics such as pitch, amplitude, spectrum (tone color), etc. Extract the time change of. Each time change is composed of, for example, characteristic value data representing an instantaneous value of each musical sound characteristic and time data representing the continuous time of the instantaneous value by the number of steps of the minimum time resolution in the MIDI standard, that is, event data. And MIDI sequence data consisting of duration data.
[0028]
The time change of the pitch includes fluctuation of the frequency in the attack part and release part of the musical tone, vibrato, pitch bend, pitch change due to slur, etc. This time change is expressed as, for example, a MIDI standard pitch bend sequence. . The time change of the amplitude includes the change of the amplitude envelope from the attack part to the release of the musical sound, the amplitude change due to tremolo, accent, slur, etc. This time change is, for example, as a volume or expression sequence of the MIDI standard. Expressed.
[0029]
The time change of the spectrum is a change in the spectrum characteristic from the attack part to the release part of the musical sound. There are various elements in spectral characteristics, and one specific example is brightness data. The brightness data indicates the amplitude ratio between the fundamental tone of the recorded waveform and each overtone, and is expressed as a MIDI standard brightness sequence. Other examples include data indicating the formant shape, cut-off frequency, data indicating the time variation of the spectrum spread, and the like. Instead of the MIDI standard brightness data, it may be represented by sequence data representing a time change such as a filter cutoff frequency or resonance.
[0030]
FIG. 3 is an analysis data value showing the time variation of the pitch, amplitude (volume) and spectrum of the trumpet sound separately from the sound of the trumpet strongly blown, the sound of the medium strength and the sound of weakly blown in order from the left. Is shown. FIG. 4 is an enlarged view of each analysis data value related to the sound of strongly blowing the trumpet.
[0031]
Next, in response to the operation of the key input device 24, the CPU 11 divides each analysis data of pitch, amplitude and spectrum from the attack to the release into a plurality of pieces on the time axis by program processing (not shown). Analyzing the analysis data, standardizing each of the divided analysis data, and combining the analysis data related to the pitch, amplitude and spectrum for each of the divided units into one set and indexing the part data A set is created and recorded in the part data storage area 37. In FIG. 2, the function for creating the part data set is shown as a part creation unit 34, and the function for generating auxiliary data for indexing is shown as an auxiliary data generation unit 35. The writing means 36 having the function of writing the part data set corresponds to the drive device 14a (or the drive device 15a) and the program processing for writing, and the part data storage area 37 is stored in the hard disk 14 or the external recording device 15. This is an area provided for storing part data.
[0032]
First, the decomposition process will be described. The time change of the musical sound characteristics such as pitch, amplitude, spectrum and the like is decomposed into an attack part, a sustain part, a sustain part with vibrato, a release part, a connection part, etc. in the time axis direction. In this case, as shown in FIG. 4, the display 25 displays the time change of the pitch, amplitude, and spectrum, and the user uses the key input device 24 including a mouse on the time axis of the pitch, amplitude, and spectrum. By specifying a common point, the pitch, amplitude, and spectrum are divided into a plurality of parts (attack, sustain, release, etc.) having a common time axis. In this case, the division point may be determined by searching for a point at which the manner of time change changes based on at least two time changes of pitch, amplitude, and spectrum. Specifically, for example, before, during and after changes in the boundary between non-stationary parts (attack, release, etc.) and stationary parts (sustains, etc.) and changes in characteristics (slur, vibrato, etc.) It is recommended to specify the boundary as a dividing point.
[0033]
As described above, the disassembly process may be performed with a manual operation by the user, but the disassembly process may be automatically performed by a program process. In this case, the division point determination criteria as described above may be programmed, and a plurality of parts may be created by determining the division points when the program is executed.
[0034]
Next, the normalization process will be described. For each musical tone characteristic (pitch, amplitude, and spectrum) related to each analyzed instrument sound, the analysis data for each of the parts (the attack part, the sustain part, and the release part) is normalized. In addition, the normalization as used in this specification is defined as a common definition for each musical sound characteristic (pitch, amplitude, and spectrum), and a common connection point value for each part (attack, sustain, and release). It means that it is almost matched with a predetermined value set in advance. Specifically, when an attack part, a sustain part, and a release part are adopted as each part, the last of the attack part, the first of the sustain part, the last of the sustain part, and the first of the release part are respectively set to predetermined values. Equivalent to equalization.
[0035]
First, the pitch normalization process will be described. In this process, in addition to the process of adjusting the connection point value according to the common definition of the standardization to a predetermined value, the frequency change of the input instrument sound is changed. A process expressed as a frequency change (pitch change) with respect to the reference frequency is also performed. Specifically, the frequency corresponding to the pitch of the input musical instrument sound (scale frequency such as A4 and E4) is set as the reference frequency, and the analysis data representing the time variation of the pitch for each part is used as the pitch deviation amount with respect to the reference frequency. The data is converted into data representing the time change of each, and thereafter, the pitch deviation amount of the connection point is set to substantially “0”. In other words, if the normalization processing for the attack portion, the converted data is processed so that the last pitch deviation amount is almost “0”, and if the normalization processing is for the sustain portion, the conversion data is converted to the first and last pitches. Both are processed so that the deviation amounts are substantially “0”, and if the normalization processing is related to the release portion, the converted data is processed so that the initial pitch deviation amounts are both substantially “0”. In this case, the reference frequency may be input using the key input device 24, or a frequency close to the scale frequency may be automatically set as the reference frequency based on the analysis data.
[0036]
Next, a normalization process related to amplitude will be described. In this process, a process of making the connection point value according to the common definition of the normalization substantially coincide with a predetermined value is performed. Specifically, even if the recorded instrument sounds are different in intensity (overall volume), the analysis data for each part is shifted and the gain is adjusted so that each part (attack part, sustain part) And the release point) are set to a predetermined value. That is, the last amplitude level of the attack portion, the first and last amplitude levels of the sustain portion, and the first amplitude level of the release portion are corrected to substantially the same predetermined value. This leads to approximating the amplitude level of the sustain portion to a predetermined value. As for the spectrum, as in the case of the amplitude, the analysis data for each part is shifted or the gain is adjusted, so that the value of the connection point of each part (attack part, sustain part and release part) is almost predetermined. Adjust to the value. 5 to 7 are diagrams illustrating examples of data values representing temporal changes in the pitch, amplitude, and spectrum of the normalized attack portion, sustain portion, and release portion.
[0037]
Since the change width of the value of each part is standardized and reduced by the normalization process as described above, the time change of each pitch, amplitude and spectrum can be expressed with a small number of bits, and the capacity of the part data storage area 37 can be expressed. Can be reduced. In particular, in the part data representing the pitch variation characteristic, the pitch variation with respect to the reference frequency is represented, so that the capacity of the part data storage area 37 can be further reduced. In the MIDI standard, the number of bits of data representing the pitch, amplitude, and timbre changes of the musical tone signal is limited, and the data of each part by the standardization process can be easily expressed in a format according to the MIDI standard. In addition, a limited number of bits can be effectively used according to the MIDI standard.
[0038]
In addition, regarding the musical tone characteristics of the sustain portion, particularly the amplitude and spectrum, a fine change is important, not a rough change such as monotonous decrease, so that the analysis data regarding the amplitude and spectrum is subjected to high-pass filter processing. Good. In this case, by the high-pass filter processing, the amplitude and spectrum of the sustain portion are superposed on the data that is maintained at a substantially constant value as a whole. The data values of are almost the same. A rough change such as a monotonous decrease in the sustain portion can be appropriately added when connecting part data to be described later.
[0039]
However, for instrument sounds with a tremolo effect and instrument sounds with a vibrato effect, temporal changes in amplitude and spectrum should not be removed by the high-pass filter process, but should remain as temporal changes. Therefore, the cutoff frequency of the high pass filter needs to be set to a very small value. FIG. 8 shows an example of temporal changes in the pitch and amplitude of the instrument sound that is not decomposed into an attack part, a sustain part, and a release part, but is a musical instrument sound to which a tremolo effect has been applied. .
[0040]
Next, an operation of forming a part data set and storing it in the part data storage area 37 for each unit (attack part, sustain part and release part) obtained by dividing the standardized parts data will be described. As described above, standardized pitch change data, amplitude change data, and spectrum change data representing time changes in pitch, amplitude, and spectrum for each divided unit (attack part, sustain part, and release part) are grouped. Thereafter, the user inputs auxiliary data for indexing using the key input device 24, thereby generating the auxiliary data. Then, the generated auxiliary data is added to the grouped part data to create a set of part data sets, and each part data set is stored in the part data storage area 37. In this case, the auxiliary data for indexing is data for identifying instrument sounds that are input based on the instrument name, sound intensity, staggered or slur sound, fast or slow rising sound, presence or absence of tremolo and vibrato, etc. This data includes data representing parts such as an attack part, a sustain part, and a release part, and is used for searching for each part data. The auxiliary data is automatically generated by the instrument sound input, analysis, decomposition, normalization processing, or other processing in place of or in addition to the user input. Data can also be used.
[0041]
FIG. 9A shows index data, pitch change data, amplitude change data, and spectrum change data, and each part data set set for each divided unit (attack part, sustain part, and release part). Shows the data format. FIG. 9B shows an example of pitch change data, where ΔPitch is data representing the degree of pitch change with respect to the reference frequency (data representing the change in pitch with respect to the fundamental frequency in accordance with the MIDI standard), and ST is The time for which each degree of change continues is represented by the number of steps.
[0042]
In the above description, each part data is formed based on a musical instrument sound signal inputted from the outside. However, each part data may be created from changes in output of various sensors attached to a natural musical instrument. For example, each part data is created from the output of the bow speed sensor composed of the pressure sensor attached to the bow of the violin, each part data is created from the output of the performance operator of the electronic musical instrument, the breath sensor of the electronic wind instrument, Each part data may be created from the lip sensor.
[0043]
In the above description, the part data (musical tone modification data) representing the musical tone characteristics is converted into data conforming to the MIDI standard. However, in a system that can use a data format other than the MIDI standard, the part data is converted into a format other than the above. It may be converted into the data. For example, the detected time change of the musical sound characteristic may be expressed as a time function, or may be expressed by a broken line approximation based on the target value and the rate.
[0044]
b. Song data processing mode
Next, a music data processing mode for processing the music data by pasting the created part data (musical sound modification data) into music data composed of a series of note data will be described. FIG. 10 is a functional block diagram showing the operation of the musical tone generator of FIG. 1 in the music data processing mode.
[0045]
First, the user designates one piece of song data by inputting a song name or the like from a plurality of pieces of song data recorded in advance on the hard disk 14 or the external recording device 15 using the key input device 24. The CPU 11 reads out the designated music data from the hard disk 14 or the external recording device 15 according to the designation of the music data, and writes it in the RAM 13. The functions of the key input device 24, the CPU 11, and the RAM 13 are described as the music data input means 41 in FIG.
[0046]
The part selection data input means 42a displays information for selecting a part data set in response to the user's operation of the key input device 24, for example, information indicating an instrument name, a fast rising sound, an attack part, etc. for each note. To create. Specifically, the user inputs information necessary for designating a part data set in accordance with the display on the display 25, and the CPU 11 stores the same information for each note in the RAM 13 by program processing (not shown). deep. Therefore, the part selection data input means 42a corresponds to a function realized by the CPU 11, the RAM 13, the key input device 24, the display 25, and the like.
[0047]
The feature analysis unit 42b automatically creates information for selecting a part data set for each note by analyzing the selected music data. For example, you can specify an appropriate instrument name based on pitch, note length, tempo, etc., or specify strong, medium, or weak based on velocity data (representing the strength of the sound) contained in the song data, The slur is specified based on the fact that the rising part of the sound overlaps the sound generation period of the previous sound, or the staccato is specified because the sound generation period is shorter than the note length. Specifically, the CPU 11 automatically analyzes the music data written in the RAM 13 by program processing (not shown), and creates information for selecting each part data set based on the analysis result for each note. And stored in the RAM 13. Therefore, the feature analysis unit 42b corresponds to a function realized by the CPU 11, the RAM 13, and the like.
[0048]
In the above description, information for selecting a part data set is formed manually or automatically for each note. A unit for forming information for selecting these part data sets may be a phrase unit composed of a plurality of notes or a part of one note. That is, when a single and related effect, for example, a slur effect is given to a phrase composed of a plurality of notes, information for selecting a part data set for a plurality of notes belonging to the phrase is simultaneously provided. Create it. In addition, when an effect such as vibrato is given only to a part of one note, information for selecting a part data set or a part of the part data set may be formed for a part of one note. .
[0049]
The part designating unit 42c forms selection data for selecting a part data set for each note based on the information created by the part selection data input unit 42a and / or the feature analysis unit 42b. In this case, the selection data is formed only by the information created by the part selection data input means 42a (manual selection mode) automatically or in accordance with a user instruction from the key input device 24, or created by the feature analysis means 42b. The selection data is formed only by the information (automatic selection mode), or the selection data is formed by using the information respectively created by both the part selection data input means 42a and the feature analysis means 42b (semi-automatic selection mode). Specifically, these operations are executed by a program process (not shown) by the CPU 11 mixed with the operation of the key input device 24, and the parts specifying means 42c is realized by the CPU 11, the RAM 13, the key input device 24, and the like. Corresponding to the function to be performed.
[0050]
The part selection means 42d selects a part data set corresponding to the selection data from a plurality of part data sets stored in the part data storage area 37 using the formed selection data. Specifically, the CPU 11 performs a plurality of sets stored in the part data storage area 37 provided in the hard disk 14 or the external storage device 15 in accordance with an instruction input using the key input device 24 by program processing (not shown). A plurality of part data sets corresponding to the selected data are sequentially read out (part data sets of the attack part, the sustain part, and the release part), and a plurality of one note data that are read out are searched. The part data sets are collectively stored in the RAM. Such an operation is sequentially performed one note at a time, and a set of parts data sets for one note or a plurality of notes belonging to a predetermined unit is stored in the RAM 13.
[0051]
In the above description, in order to form musical tone modification data for one note, a plurality of part data sets (attack part, sustain part and release part) at different positions on the time axis stored in the part data storage area 37 are used. Part data sets) were selected independently. However, a plurality of part data sets at different positions on the time axis may be simultaneously selected in relation to each other.
[0052]
The processing and connection means 43 is a part of the piece data in the music data including the instruction by the user, which is divided into the attack part, the sustain part, the release part, etc. of the part data set selected by the part selection means 42d. Two notes are connected while being processed from an attack part to a release part, and tone modification data is created for one note for each tone characteristic of pitch, amplitude and spectrum. Specifically, the CPU 11 uses a song process and a plurality of parts data stored in the RAM 13 by a program process (not shown) to generate a musical sound from the attack unit to the release unit in response to an instruction from the key input device 24. The modification data is formed for each tone characteristic. Therefore, the processing and connection means 43 corresponds to a function realized by the CPU 11, RAM 13, key input device 24, and the like.
[0053]
In the processing and connecting operation, as shown in FIGS. 11 and 12 (musical tone modification data related to amplitude), the musical data for one note is obtained by sequentially connecting the selected part data such as the attack part, the sustain part and the release part. The modification data is formed for each tone characteristic of pitch, amplitude and spectrum. At this time, when the part data relating to the prepared sustain part is long and the note length represented by the tone modification data connecting the parts data is longer than the corresponding note length, as shown in FIG. Then, only a part of the part data of the sustain part is cut out, and a part of the cut part data is connected between the attack part and the release part. On the contrary, when the part data relating to the prepared sustain portion is short and the note length represented by the musical tone modification data connecting the parts data is shorter than the corresponding note length, as shown in FIG. The part data of the sustain part is repeatedly used.
[0054]
In this case, the values of the end point of the attack part data, the start and end points of the sustain part data, and the start point of the release part data are standardized and set to almost the same value. The connection is easily made. Even in this connection, if necessary, particularly when using part data of the sustain part, the level of either the part data before connection or the part data after connection is shifted, By adjusting the gain, the data value may be slightly modified to smoothly connect the connection portions. Further, as shown in FIG. 13, the rear part of the front part data and the front part of the rear part data in the connection part may be connected by crossfading.
[0055]
In addition, when a slow change in the tone characteristic is added, such as a gentle decrease in the sustain portion as shown in FIG. 11, the user can use the previously stored data or the key input device 24 and the display 25 to It is preferable to modify a part of the part data before or after the connection of each part data based on the data inputted.
[0056]
Further, when pasting the musical sound modification data into the music data, it is necessary to compress or expand the musical sound modification data (part data) on the time axis according to the reproduction tempo of the music data. That is, the musical sound modification data represents a time change of the musical sound characteristic such as a time change of pitch, volume or spectrum (tone color) every 10 milliseconds. Therefore, in order to paste the musical sound modification data into the song data while maintaining its time change characteristics, the length of the time axis of the musical sound modification data according to the tempo of the position where the musical sound modification data is to be pasted Must be adjusted appropriately.
[0057]
For example, it is assumed that musical tone modification data is recorded at a time interval of 10 milliseconds. This 10 milliseconds corresponds to one clock when the tempo of automatic performance is 125 beats / minute and the resolution of a quarter note is 48 clocks. Assuming that the tempo of 125 beats / minute is the standard tempo, when the tempo at the position where the musical sound modification data is to be pasted in the song data is slower than the standard tempo, as shown in FIG. The part data of the part, the sustain part and the release part are compressed on the time axis in accordance with the ratio of the standard tempo and the tempo in the music data before or after connecting the part data. This means that if the tempo of the automatic performance is slowed down, the readout clock of the musical tone modification data is slowed down, and the musical tone modification data read out due to the slowdown of the readout clock is expanded on the time axis. This is to correct in advance. On the contrary, when the tempo at the position where the musical tone modification data is to be pasted is faster than the standard tempo, the part data is connected to the standard tempo and the tempo in the song data before or after the connection of the part data. It expands on the time axis according to the ratio. Specifically, the step number data ST in FIG. 9B may be changed in accordance with the ratio between the two tempos.
[0058]
The pasting means 44 processes the music data for one song by sequentially pasting the musical tone modification data relating to the pitch, amplitude and spectrum created by the processing and connecting means 43 to the music data one note at a time. is there. The song data output means 45 sequentially stores the song data to which the musical tone modification data is pasted in the RAM 13 and also in the hard disk 14 or the external recording device 15. Specifically, the CPU 11 pastes the musical tone modification data for each note stored in the RAM 13 to the music data stored in the RAM 13 according to an instruction by the operation of the key input device 24 by program processing (not shown). The music data that has been pasted and pasted is stored in the RAM 13 and also stored in the hard disk 14 or the external recording device 15. Accordingly, the pasting means 44 and the music data output means 45 correspond to functions realized by the CPU 11, RAM 13, key input device 24, drive devices 14a, 15a, and the like.
[0059]
The operation for pasting the musical tone modification data will be described in detail with reference to FIG. In FIG. 14, (A) shows a part of a musical score of music, (B) shows music data corresponding to the musical score before pasting musical tone modification data, and (C) shows the music data. In FIG. 10, pitch change data (music data pasted with FIG. 10 (B) is shown as musical tone modification data. In FIGS. 10 (B) and (C), “NOTE” represents a note name, and “K #” Indicates the key code of the pitch name, “ST” is the time to the next event (ie, the number of steps corresponding to the note length in (B), the step to the next pitch name or pitch change in (C) "GT" indicates the number of steps corresponding to the gate time, and "VEL" indicates velocity (sound strength) in the case of data relating to the pitch name, and in the case relating to pitch change. Indicates the degree of pitch change. This represents the time until the next note or rest, and the gate time represents the time from the start of the attack portion to the end of the sustain portion (key-on time), and in (B), the number of steps of the gate time GT Is larger than the number of steps of the note length ST, it means that the notes relating to these data are played with a slur, and the numerical value “192” indicates the bar position.
[0060]
14A to 14C, the note length step number ST of this note is “32”. The pitch change data changes every step number “17”, “2”, “1”, “8”, “2”, “2” as shown in FIG. 9B. Therefore, in this case, the note length step number “32” in (B) is divided into “17”, “2”, “1”, “8”, “2”, “2”, and (1) As in C), pitch change data ΔPitch is added to the data relating to the note names NOTE, #K, note length ST, gate time GT, and velocity VEL. Such work is performed one note at a time from the beginning to the end of the song data.
[0061]
As for the amplitude change data and the spectrum change data, similarly to the case of the pitch change data, the amplitude change data and the spectrum change data can be pasted on the music data. When a plurality of types of musical tone modification data among the pitch variation data, the amplitude variation data, and the spectrum variation data are pasted to the music data, the plurality of types of musical tone modification data are pasted in a state of being synchronized with each other. Here, “synchronization” means maintaining the time relationship between pitch change, amplitude change, and spectrum change when extracted by analysis.
[0062]
When the pitch change data is to be pasted into the music data, pitch bend data for changing the pitch may already be recorded in the music data. In this case, pitch change data (musical tone modification data) formed from the part data set may be added to the pitch bend data and pasted into the music data. In addition, regarding amplitude (volume) and spectrum (timbre), amplitude change data formed from the parts data set is also recorded when data representing amplitude and spectrum temporal changes for one musical tone is already recorded in the music data. And spectral change data (musical sound modification data) may be added and pasted into the music data. In some cases, the data representing the temporal change in amplitude and spectrum for one musical tone already recorded is removed, and the pitch variation data, amplitude variation data and spectral variation data (musical tone modification data) formed from the part data set are removed. ) May be pasted into the song data, or the already recorded data may be left as it is, and the pasting of the musical tone modification data may be stopped.
[0063]
If the song data is composed of MIDI format sequence data, only one pitch bend, volume data and brightness data can be specified for each storage track. In the method, it is suitable to paste the musical sound modification data on the music data of the track on which the solo part is recorded.
[0064]
When pasting musical tone modification data to the song data of a track in which parts that generate two or more musical sounds at the same time are recorded in the MIDI format, the musical tone modification data for a plurality of musical sounds that are simultaneously generated by the above-described method is used. Since the pasting is not performed accurately, the following method should be adopted.
[0065]
The first method is to select musical tone modification data relating to any one of the plurality of musical sounds for a portion where a plurality of musical sounds are generated simultaneously in various musical tone characteristics of pitch, amplitude and spectrum, The musical tone modification data is pasted into the song data by pasting the selected musical tone modification data in common to the plurality of note data respectively corresponding to the plurality of musical sounds generated simultaneously. In this case, among the plurality of musical sounds, it is preferable that the musical sound modification data to be pasted is the musical sound modification data related to the musical sound having the largest amplitude value at each time point. Alternatively, in the various musical tone characteristics, the musical tone modification data relating to the musical tone having the greatest degree of change from the reference value of the musical tone characteristic at each time point may be used as the musical tone modification data to be pasted.
[0066]
In the second method, a plurality of tone modification data (a plurality of pitch change data, a plurality of amplitude change data, or a plurality of spectrum change data) are synthesized, and the synthesized tone modification data is generated at the same time. Are pasted in common to a plurality of note data. In addition, when the sound generation start timing or sound generation end timing of a plurality of musical sounds that are simultaneously generated is shifted and a transition is made from a state in which one musical sound is generated to a state in which a plurality of musical sounds are simultaneously generated, or a plurality of musical sounds When shifting from a state in which sounds are generated simultaneously to a state in which only one tone is generated, it is preferable to crossfade a plurality of tone modification data at the transition portion.
[0067]
As a third method, a plurality of musical note data relating to a plurality of musical sounds that are stored in one track and pronounced at the same time are separated, stored in a plurality of tracks, and stored in a plurality of tracks. Paste the musical tone modification data for each note data. For example, when a plurality of musical sounds are generated at the same time as shown in the score of FIG. 15A, the song data is as shown in FIG. In (B), the note length step number ST being “0” means that it is generated simultaneously with the next note data. Such music data is stored separately in a plurality of tracks 1 to 3 as shown in FIGS. 15 (C) to 15 (E). That is, with respect to the note data in which the note length step number ST is “0” in (B), the note length step number ST of the note that is simultaneously pronounced and simultaneously stops sounding is adopted, and new song data for different tracks is adopted. Is remembered.
[0068]
In addition, regarding separation of note data in this way, in the case of a track storing song data that produces guitar-tone musical tones, each string that is simultaneously generated by a fingering method's automatic analysis technology is recorded to which string. It is preferable that note data is distributed to a plurality of tracks corresponding to each string. In this way, the musical tone modification data is added to the musical note data separately stored in a plurality of tracks, and the above-described pasting method for each musical note may be used.
[0069]
According to the first to third methods, even if the music data includes note data that simultaneously generates a plurality of musical sounds, the musical sound characteristic control data based on the plurality of musical sound modification data is superimposed on the time axis. Therefore, it is possible to easily add a time change of the tone characteristic based on the tone modification data to the tone.
[0070]
In addition, in the processing of music data as described above, part data prepared in advance for each part such as an attack part, a sustain part, and a release part is selectively connected, and pitch change data for one note. Various musical tone modification data composed of amplitude change data and spectrum change data are formed, and the musical tone modification data thus formed is pasted to each note data in the music data. As a result, it is possible to easily perform processing for adding musical tone modification data to music data and to add various musical tone modification data to musical note data.
[0071]
c. Music generation mode
Next, a tone generation mode for generating a tone based on the music data pasted with the tone modification data will be described. In this case, the user designates the music data using the key input device 24 while using the display unit 25. When the song data is stored in the hard disk 14 or the external recording device 15, the stored song data is transferred and stored in the RAM 13. Then, when instructing to start the reproduction of the song data, the CPU 11 sequentially reads the note data and the musical tone modification data in the song data stored in the RAM 13 by a program process (not shown). In this case, the CPU 11 measures the number of steps ST stored together with the note data and the tone modification data by the built-in timer means that performs a counting operation according to the tempo data TEMPO in the music data, and the same note data and the tone modification data. Data is read sequentially according to the measurement. Note that music data stored in a plurality of tracks are read simultaneously.
[0072]
The read note data and tone modification data are output to the tone generator circuit 22 via the bus 10 by program processing (not shown) by the CPU 11. When the note data is read, the CPU 11 assigns the note data to an empty channel in the tone generator circuit 22, outputs the note data to the empty channel, and instructs generation of a musical tone. The sound source circuit 22 forms a musical sound signal in accordance with the supplied note data in the empty channel in cooperation with the waveform memory 27 and outputs it to the sound system 28. That is, the tone generator circuit 22 generates a musical tone signal having a pitch corresponding to the key code K # and a volume corresponding to the velocity data VEL. In this case, the gate time data GT defines the time from the attack start of the musical sound to the end of the sustain, and the total time of the number of steps ST for one note defines the note length of one note.
[0073]
On the other hand, when the tone modification data composed of pitch change data, amplitude change data and spectrum data is read, the CPU 11 reads the tone modification data read out for the channel to which the note data in the tone generator circuit 22 is assigned. Output data. As a result, the tone generator circuit 22 changes and controls the pitch, amplitude (volume), and spectrum (tone color) of the tone signal according to the tone modification data for a time corresponding to the step number ST. In this case, particularly with respect to the amplitude envelope, the basic amplitude (volume) of the tone signal is controlled by the velocity data VEL described above, and the tone amplitude (volume) of the tone is further changed and controlled by the tone modification data relating to the amplitude. In addition, regarding the spectrum (tone color), control by tone modification data may be added to control by velocity data VEL.
[0074]
Also, other data included in the music data that controls the musical tone signal and various data not included in the data are output to the channel. Thus, the pitch, amplitude (volume) and spectrum (tone color) of the generated tone signal are controlled by the tone modification data and other various data.
[0075]
Thus, the generated tone signal is controlled by time-varying tone modification data, and this tone modification data is generated based on the pitch, amplitude (volume) and spectrum (tone) of natural instrument sounds, etc. The tone quality of the musical tone to be played is improved, and it is possible to generate a musical tone having a sense of realism and a rich performance expression.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an example of a musical sound generator to which a music data processing method according to the present invention is applied.
FIG. 2 is a functional block diagram showing an operation of the musical sound generating device of FIG. 1 in a musical sound modification data creation mode.
FIG. 3 is a diagram illustrating temporal changes in pitch, amplitude, and spectrum of a trumpet sound separately for a sound that is strongly blown by a trumpet, a sound that is blown at a medium intensity, and a sound that is weakly blown.
FIG. 4 is a diagram illustrating a time change in pitch, amplitude, and spectrum with respect to a sound strongly blown by the trumpet in FIG. 3;
FIG. 5 is a diagram illustrating temporal changes in pitch, amplitude, and spectrum of a standardized attack part.
FIG. 6 is a diagram illustrating temporal changes in pitch, amplitude, and spectrum of a normalized sustain portion.
FIG. 7 is a diagram illustrating temporal changes in each pitch, amplitude, and spectrum of a standardized release unit.
FIG. 8 is a diagram illustrating an example of temporal changes in pitch and amplitude of a standardized instrument sound that is a musical instrument sound to which a tremolo effect is applied.
9A is a format diagram of a part data set, and FIG. 9B is a detailed format diagram of pitch change data in the part data set.
FIG. 10 is a functional block diagram showing the operation of the musical tone generator of FIG. 1 in the song data processing mode.
FIG. 11 is an explanatory diagram for explaining an example of connection of each part data of an attack part, a sustain part, and a release part.
FIG. 12 is an explanatory diagram for explaining another example of connection of each part data of an attack part, a sustain part, and a release part.
FIG. 13 is an explanatory diagram for explaining another example of connection of each part data of the sustain portion and the release portion.
14A is a partial diagram showing an example of a score, FIG. 14B is a data format diagram showing a part of music data before processing corresponding to the score, and FIG. It is a data format figure which shows a part of corresponding music data after a process.
15A is a partial diagram showing an example of a score, FIG. 15B is a data format diagram showing a part of music data before processing corresponding to the score, and FIGS. 15C to 15E. FIG. 4 is a data format diagram of song data obtained by separating each piece of note data corresponding to a plurality of musical sounds that are simultaneously generated in the song data.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Bus, 11 ... CPU, 12 ... ROM, 13 ... RAM, 14 ... Hard disk, 15 ... External recording device, 21 ... Capture circuit, 22 ... Sound source circuit, 23 ... MIDI interface, 24 ... Key input device, 25 ... Display 26 ... external signal input terminal, 30 ... microphone, 32 ... waveform data storage area, 33 ... analyzing means, 34 ... part creating means, 35 ... auxiliary data generating means, 36 ... writing means, 37 ... part data storage area, 41 ... Music data input means, 42a ... Parts selection data input means, 42b ... Feature analysis means, 42c ... Parts designation means, 42d ... Parts selection means, 43 ... Processing and connection means, 44 ... Pasting means, 45 ... Music data Output means.

Claims (7)

In a song data processing method for processing song data consisting of a plurality of note data that are recorded on a recording medium and respectively specify the pitch and length of a plurality of musical sounds,
Dividing a plurality of musical tone modification data respectively representing a time change of at least one musical tone characteristic among pitch, amplitude and tone color related to one musical tone into a plurality of parts on the time axis, and end portions of the divided parts Each of the above parts is processed and prepared in advance so that all of the values are the same ,
A plurality of parts related to different positions on the time axis are selected from the plurality of prepared parts as musical sound modification data for temporally changing the at least one musical sound characteristic of the musical sound represented by the note data in the song data. And
Connecting the plurality of selected parts to form musical tone modification data for temporally changing the at least one musical tone characteristic of one musical tone;
A music data processing method characterized in that the formed music modification data is pasted into the music data. In the music data processing method according to claim 1,
A song data processing method characterized in that the length of any one of the plurality of connected parts is adjusted in accordance with the length of a note represented by the note data in the song data. . In a song data processing method for processing song data consisting of a plurality of note data that are recorded on a recording medium and respectively specify the pitch and length of a plurality of musical sounds,
Musical tone modification data representing a time change of at least one musical tone characteristic among pitch, amplitude and tone color relating to one musical tone is prepared in advance;
In order to temporally change the at least one musical tone characteristic of the musical tone represented by the note data in the musical piece data, the prepared musical tone modification data is pasted into the musical piece data, and at the time of pasting, the musical piece data When a plurality of musical sounds are simultaneously generated by a plurality of musical note data stored in one track, a portion of the plurality of musical sounds generated at the same time for a portion where the plurality of musical sounds are generated simultaneously. The musical tone modification data relating to one musical tone is selected, and the selected musical tone modification data is commonly pasted to a plurality of note data respectively corresponding to the plurality of musical sounds generated simultaneously. Song data processing method. In a song data processing method for processing song data consisting of a plurality of note data that are recorded on a recording medium and respectively specify the pitch and length of a plurality of musical sounds,
Musical tone modification data representing a time change of at least one musical tone characteristic among pitch, amplitude and tone color relating to one musical tone is prepared in advance;
In order to temporally change the at least one musical tone characteristic of the musical tone represented by the note data in the musical piece data, the prepared musical tone modification data is pasted into the musical piece data, and at the time of pasting, the musical piece data When a plurality of musical sounds are generated simultaneously by a plurality of musical note data stored in one track, each musical sound related to the plurality of musical sounds generated at the same time for a portion where the plurality of musical sounds are generated simultaneously. Song data characterized by combining modification data and pasting the combined music modification data to a plurality of note data respectively corresponding to the plurality of musical sounds generated simultaneously. Processing method. In a song data processing method for processing song data consisting of a plurality of note data that are recorded on a recording medium and respectively specify the pitch and length of a plurality of musical sounds,
Musical tone modification data representing a time change of at least one musical tone characteristic among pitch, amplitude and tone color relating to one musical tone is prepared in advance;
To the time change at least one of the musical characteristics of a musical tone represented by note data in the music data, pasting the prepared musical tone-modified data in said song data, the time of the paste, songs When a plurality of musical sounds are generated simultaneously by a plurality of note data stored in one track in the data, a plurality of tracks are separated by separating the plurality of note data corresponding to the plurality of musical sounds generated simultaneously. music data processing method which respectively is memorize, over the previous SL plurality of note data respectively are stored in a plurality of tracks, characterized in that the said prepared musical tone modified data to the pasted respectively. Song data processing for processing song data that is recorded on a recording medium and that is composed of a plurality of note data that respectively specify the pitch and length of a plurality of musical sounds, and tempo data that represents the playback tempo of the plurality of note data In the method
Preparing in advance a plurality of musical tone modification data representing temporal changes in at least one musical tone characteristic among pitch, amplitude and tone color relating to one musical tone;
In order to temporally change the at least one musical tone characteristic of the musical tone represented by the note data in the music data, the prepared musical tone modification data is compressed or expanded on the time axis according to the tempo data, and A song data processing method characterized by being pasted into song data.   7. The music data processing method according to claim 1, wherein the musical tone modification data is obtained by analyzing a pitch, amplitude, and tone color of a natural musical instrument sound.

Images