JP5772485B2 - Song editing apparatus and program - Google Patents

Description

  The present invention relates to a technique for generating performance data for one song by combining a plurality of performance data.

  Various techniques have been proposed in which performance data representing various melodies are collected into a database, and data selected by the user is synthesized to generate music data for one song. There is Patent Document 1 as a document relating to this type of technology. The automatic arrangement device disclosed in this document divides the melody indicated by the data into a plurality of sections when data indicating the melody of the music is supplied, and indicates the degree of undulation of the sound in each divided section. Calculate the undulation frequency. Then, the accompaniment to be assigned to the melody of each section is determined according to the undulation frequency of each section, and song data for one song in which the melody and accompaniment of each section are connected on the time axis is generated.

Japanese Patent No. 2806351 JP 2007-240564 A

By the way, in this type of conventional technology, there is a case where performance data recorded in a database is obtained by dividing performance data for one song into relatively short units such as one measure or two measures. In this case, the music that can be created by combining the performance data in the database becomes similar, and there is a problem that it is not possible to provide an environment for creating music with rich variations.
The present invention has been made in view of such a problem, and an object of the present invention is to make it possible to create a musical composition rich in variations using a plurality of performance data recorded in a database.

  According to the present invention, there is provided a first type of performance data including each event indicating the sound production content of each sound in a sequence of sounds of M (M is an integer of 1 or more) and timing data indicating the execution time of each event; Storage means for storing a second type of performance data including each event indicating the content of pronunciation of each sound in the sequence of sounds following the output and M measures, and timing data indicating the execution time of each event; Editing means for acquiring and arranging a plurality of performance data from the storage means, and arranging the second type of performance data after the first or second type of performance data. For the previous performance data, the end of the sound sequence consisting of the sound indicated by each event of the performance data is the amount corresponding to the approximate output in the sound sequence indicated by the event of the performance data after the arrangement order. Subjected to Kusuru rewriting process, to provide a music editing apparatus comprising an editing means for arranging the performance data corresponding after the performance data over the rewriting process.

  In the present invention, with respect to performance data including each event indicating the sound content of each sound in the arrangement of sounds in the performance section for M measures and timing data indicating the execution time of each event, each event of the performance data Of these, a process for adding a mark indicating that the sound event is generated after a rest longer than the threshold value corresponds to the position where the sound sequence is separated is performed, and the performance data obtained through this process is used as the first type of performance. The performance data is stored in the storage means as data, and for performance data including each event indicating the sound content of each sound in a performance section longer than M measures and timing data indicating the execution time of each event. A process of adding the mark to an event of a sound that is generated through a rest longer than a threshold among the events is performed, and the performance data that has undergone this process is used as the second type. Means for storing in the storage means as the performance data may comprise a.

1 is a block diagram showing a music editing apparatus according to an embodiment of the present invention. It is a figure which shows an example of the data structure of the performance data in the hard disk of the same apparatus. It is a flowchart which shows the data storage process which the apparatus performs. It is a figure which shows the content of the division process which the same apparatus performs. It is a figure which shows the content of the output production | generation process which the apparatus performs. It is a flowchart which shows the edit process which the apparatus performs. It is a figure which shows the content of the evaluation point calculation process which the same apparatus performs. It is a figure which shows the content of the evaluation point calculation process which the same apparatus performs. It is a figure which shows the content of the rewriting process which the apparatus performs. It is a figure which shows the content of the rewriting process which the apparatus performs. It is a figure which shows the content of the rewriting process which the apparatus performs. It is a figure which shows the content of the rewriting process which the apparatus performs. It is a figure which shows the content of the rewriting process which the apparatus performs. It is a flowchart which shows the song synthesis | combination process which the same apparatus performs.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of a music editing device 10 according to an embodiment of the present invention. The music editing apparatus 10 generates a plurality of types of performance data MD as material of a melody that is a sequence of sounds from music performance data SMD created by various creators using a MIDI (Musical Instrument Digital Interface) sequencer. The music performance data EMD for one song is generated by combining the performance data MD stored in the device 10 and selected from the plurality of performance data MD stored in the device 10. The voice signal of the singing voice having the same melody as the melody is output.

As shown in FIG. 1, the music editing device 10 includes a data reading unit 11, an operation unit 12, a display unit 13, a sound emitting unit 14, and a control unit 15. When a storage medium is loaded, the data reading unit 11 reads the music performance data SMD in the medium and supplies it to the control unit 15. FIG. 2 is a diagram showing an example of the data structure of the music performance data SMD. The song performance data SMD includes a note-on event EV _ON (#) (# indicates a note number) and a note-off event EV _OFF (#) indicating the sound generation content, and each event EV _ON (#) (or EV _OFF ). Delta time ΔT _h (h) which is timing data indicating the execution time t of (#)) as the time difference (number of ticks) from the execution time t of each previous event EV _OFF (#) (or EV _ON (#)) = 1, 2,...) (H is an index indicating the arrangement order) and are divided into L tracks TR-j (j = 1 to L). Here, in this embodiment, each creator creates music performance data SMD of one music according to the following rules a1 to f1.
a1. The music to be converted into music performance data shall be in C major.
b1. Event EV ON to indicate the pronunciation and the silencing of each sound from the first sound of the music to be music performance data of up to the end of the sound _(♯) and EV _{OFF (♯)} and tune the delta time ΔT _h indicating its execution time t Described in the first track TR-1 in the performance data SMD.
c1. When there is a performance break position in the music to be converted into music performance data, an event that serves as a mark of the event EV _ON (#) immediately before the break position in the track TR-1 (for example, an event of the note number 126) EV and _ON (126)) and its execution time t second track in the music playing delta time [delta] t _h indicating the (previous event _{EV ON} (♯ delimiter position) same execution time t) from that of data SMD TR -2.
d1. When there is a performance section suitable for the last melody of a song in the music to be converted into music performance data, an event that serves as a mark indicating the performance (for example, an event EV _ON (125) of the note number 125) ) and described in the third track TR-3 for this event _EV oN (125) the execution time t (corresponding delta time ΔT music performance data in the SMD of _h indicating the time) of the start point of the play section.
e1. When there is a performance section suitable for the first melody of the song in the music to be converted into music performance data, an event that serves as a mark indicating this (for example, event EV _ON (124) of note number 124) ) and described in the third track TR-3 of the music performance data in the SMD delta time [delta] t _h indicating the execution time t (the time of the start point of the corresponding play interval) of this event _EV oN (124).
f1. If there is a performance section suitable for the melody of the part (intermediate part) that is neither the first nor the last of the song in the music to be converted into musical performance data, an event that serves as a mark indicating that (for example, a note) the third track number 123 events and _EV ON (123)) of this event _EV ON (123) the execution time t (corresponding delta time ΔT music performance data in the SMD of _h indicating the time) of the start point of the play interval Described in TR-3.

In FIG. 1, the operation unit 12 receives an input operation of various information and supplies a signal indicating the operation content to the control unit 15. The display unit 13 displays various information under the control of the control unit 15. The sound emitting unit 14 outputs the sound signal output from the control unit 15 as a sound. The control unit 15 is a device that serves as a control center of the music editing device 10. The control unit 15 includes a CPU 16, a RAM 17, a ROM 18, and a hard disk 19. The CPU 16 executes programs stored in the ROM 18 and the hard disk 19 while using the RAM 17 as a work area. The hard disk 19 stores a material database MDB, a music database BDB, a speech segment database PDB, and a music editing program PRG. Performance data MD is stored in the material database MDB. Music performance data EMD is stored in the music database BDB. Speech unit data PD is stored in the speech unit database PDB. Here, the speech segment data PD is data defining waveforms of various speech segments that are materials of singing speech, such as a single phoneme and a transition portion between phonemes. The song editing program PRG has the following three functions.
a2. Data accumulation function This is a function for generating a plurality of types of performance data MD from the music performance data SMD and storing it in the material database MDB when the music performance data SMD is supplied from the data reading unit 11.
b2. Editing Function This is a function for acquiring four pieces of performance data MD from the material database MDB, and arranging the acquired performance data MD to generate one piece of music performance data EMD.
c2. This is a function that synthesizes a sound signal of a singing voice having the same melody as the melody of the music performance data EMD and emits the sound signal from the sound emitting unit 14.

  Next, the operation of this embodiment will be described. In the present embodiment, the CPU 16 performs a data accumulation process, an editing process, and a song composition process. Hereinafter, each process is demonstrated in order.

FIG. 3 is a flowchart showing the data accumulation process. A series of processing shown in FIG. 3 is processing executed by the function of the data storage function. A series of processes shown in FIG. 3 is started when the music performance data SMD read by the data reading unit 11 is written in the RAM 17. In executing the series of processes shown in FIG. 3, the user sets the following three items with the operation unit 12 and stores each set value in the RAM 17.
a3. Data processing start position In this item, the first bar position X of the portion used to generate the performance data MD in the music performance data SMD is set.
b3. In this item, if you want to get only performance data MD without the effect, set “No output”. If you want to get both the performance data MD without the effect and the performance data MD with the effect, set “With the effect”. Set.
c3. Transposition In this item, if you want to obtain only the performance data MD in the same key (C major) as the performance data MD, set “No transposition”, and obtain the performance data MD in the same key as the performance data MD and in a different key. If you want, set “With transposition”.

In FIG. 3, the CPU 16 performs a dividing process (S101). In this division processing, as shown in FIG. 4, the CPU 16 sets the performance section after the X bar (X = 2 in the example of FIG. 4) in the music performance data SMD in the RAM 17 to M (M is an integer of 1 or more). : For example, M = 2) divided into performance sections Ri for each measure (i = 1 to Z: Z is the total number of divided performance sections), and the event EV _ON in each performance section Ri _{(♯), EV OFF (♯} ), and the performance data _MD R -i the delta time ΔT _h.

Next, the CPU 16 sets the value of the variable k to the initial value k = 1 (S102), and then the kth performance data among the performance data MD _R-i (i = 1 to Z) divided in step S101. selecting MD _R -k for processing (S103). The CPU 16 performs a first determination process (S104). In the first determination process, CPU 16 forms a melody in the delta-time [Delta] T _h playing interval R-k after each event _{EV ON} the first track TR-1 of the performance data _MD R -k (♯) It is determined whether or not the performance section R-k satisfies the following two conditions a4 and b4 with the length of each sound.
a4. There is at least one sound represented by each event EV _ON (#) in the performance section Rk that has a sound length less than the threshold value TH1. B4. The total length of the sound indicated by each event EV _ON (#) in the performance section Rk is less than the threshold TH2.

If the CPU 16 determines in step S104 that the performance data MD _R -k to be processed satisfies both the conditions a4 and b4 (S104: Yes), the performance data MD _R -k is discarded (S105). . Here, to discard the performance data MD R _-k, melody melody or sound including too short sound contains only little is because hardly utilized as a melody singing voice. CPU16 at step S105, after discarding the performance data _MD R -k, the process proceeds to step S114. The contents of the process in step S114 will be described later.

CPU16 at step S104, if the performance data _MD R -k was processed is determined not to satisfy one or both of the above conditions a4 and b4 (S104: No), performs the second determination process (S106) . In the second determination process, the CPU 16 determines whether “with output” is set in the setting item b3. The CPU 16 performs the output generation process when “with output” is set (S106: Yes) (S107). The output generation process is performed using each event EV _ON (#) and EV _OFF (#) and each event EV indicating each sound of the output and subsequent M (M = 2) melody sounds from the music performance data SMD. is _ON (♯) and _{EV OFF} processing for generating performance data _MD RA -k comprising a delta time [delta] T _h indicating the execution time (♯). In pick-up (upbeat) generating process, CPU 16 is, M from music performance data SMD (M = 2) event _{EV ON} longer playing section than bars (♯) and _{EV OFF} and (♯) excised delta time [Delta] T _h, cut event EV _ON (♯) and _{EV OFF} (♯) and the delta time ΔT _h and performance data _MD RA -k. More specifically, as shown in FIG. 5, the CPU 16 has a measure (first in the example of FIG. 5) immediately before the performance section R- 1 in the second track TR- 2 of the music performance data SMD in the RAM 17. In the event EV _ON (126) in the measure), the execution time t of the event EV _ON (126) closest to the performance section R-1 is set as the start time t _STRT of the after- _effect, and the track TR-j (j = 1 to L) of the music performance data SMD each event _{EV ON} between the time t _STRT to the end of the play interval R-1 in the (♯) and an _{EV OFF} (♯) and play a delta time [delta] t _h of the first data _MD RA -1. CPU16 generates a second performance data _MD RA-2,3-th performance data _MD RA -3 ... Z th performance data _MD RA -Z in the same procedure.

If “no output” is set (S106: No), the CPU 16 proceeds to the next step S108 without executing step S107. In step S108, the CPU 16 performs a third determination process. In the third determination process, the CPU 16 sets the event EV on the second track TR-2 of the performance data MD _R -k (or performance data MD _R -k and MD _RA -k when step S107 is executed). It is determined whether _ON (126) is written. CPU16, when the second track TR-2 to the event _EV ON performance data _MD R -k (126) is not written (S108: No), performs the first marking (S109). In the first marking process, CPU 16 forms a melody in the delta-time [Delta] T _h playing interval R-k after each event _{EV OFF} in the first track TR-1 of the performance data _MD R -k (♯) each It is determined whether there is a rest having a length equal to or greater than the threshold TH3 in the performance section R-k. If there is a rest longer than the threshold value TH3 between the sounds forming the melody in the performance section Rk, the CPU 16 regards the rest as a performance break position. In this case, CPU 16 is configured to append an event _EV ON (126) to the second track TR-2 performance data _MD R -k, selects a previous event _{EV OFF} delimiter position in the track TR-1 (♯) . Then, the CPU 16 sets the second track so that the execution time t of the selected event EV _OFF (#) and the execution time t of the event EV _ON (126) added to the second track TR-2 are the same time. Rewrite the delta time ΔT _h before and after the event EV _ON (126) added in TR-2.

CPU16, when the performance data _MD R -k second track TR-2 to one or more event _EV ON (126) is written (S108: Yes), without executing the step S109, the next step Proceed to S110. In step S110, the CPU 16 performs a fourth determination process. In the fourth determination process, the CPU 16 sets the event EV on the third track TR-3 of the performance data MD _R -k (or performance data MD _R -k and MD _RA -k when step S107 is executed). It is determined whether _ON (125), EV _ON (124), or EV _ON (123) is written. CPU16 at step S110, the event _EV ON ₍₁₂₅₎ to the third track TR-3 of the performance data _{MD R -k, EV ON (124} ), and determines that any unwritten the _EV ON (123) In the case (S110: No), the second marking process is performed (S111). In the second marking process, CPU 16, the sound playing section of the last event _{EV ON} (♯) R-k of the first track TR-1 for each event _{EV ON} performance data _MD R -k (♯) The last sound of the melody is considered. Then, it is determined whether or not this last sound satisfies the following two conditions.
a5. The note name of the note is a C fundamental key (b) b5. The sound length is greater than or equal to the threshold TH4

CPU16, when both of the two conditions a5 and b5, regarded as the performance data _MD R -k contains a melody suited to the last part of the song. In this case, CPU 16 additionally writes the events _EV ON (125) to the beginning of the third track TR-3 of the performance data _MD R -k.

Further, CPU 16, when at least one of the conditions a5 and b5 are not met, including performance data MD R _-k is the first part of the song, the melody suited to the first and last is neither part (middle part) It is considered to be out. In this case, CPU 16 additionally writes the performance data _MD R -k third track TR-3 for the top event _EV ON (124) and event _EV ON (123).

CPU16 determines in step S110, the third track TR-3 in the event _EV ON performance data _{MD R -k (125), EV} ON (124), and either of the _EV ON (123) is written If so (S110: Yes), the process proceeds to the next step S112 without executing step S111. In step S112, a fifth determination process is performed. In the fifth determination process, the CPU 16 determines whether or not “with transposition” is set in the setting item c3. When “with transposition” is set (S112: Yes), the CPU 16 performs transposition processing (S113). The transposition process, (when the step S107 is performed, the performance data _MD R -k and _MD RA -k) the performance data _MD R -k generate performance data MD from the melody was transposed C major different tone To do.

More particularly, CPU 16 may generate performance data _MD R -k first track TR-1 of the performance data _MD R1 -k the note number is rewritten to those minor second higher of each event _{EV ON} (♯) of To do. Also generates performance data _MD R -k first track TR-1 of the performance data _MD R2 -k the note number is rewritten to what length 2 degrees higher for each event _{EV ON} (♯) of. Also generates performance data _MD R -k performance data _MD R3 -k first track TR-1 event _{EV ON} the note number (♯) rewritten to those minor third higher. Also generates performance data _MD R -k performance data _MD R4 -k the note number of the first track TR-1 event _{EV ON} (♯) is rewritten to that major third higher. Also generates performance data _MD R -k first track TR-1 event _{EV ON} (♯) performance data _MD R5 -k Rewritten note number to 4 degrees higher that of the. Also generates performance data _MD R -k performance data _MD R6 -k the note number of the first track TR-1 event _{EV ON} (♯) is rewritten to that increase 4 degrees higher. Also generates performance data _MD R -k first track TR-1 event _{EV ON} (♯) performance data _MD R7 -k the note number is rewritten to 5 degrees higher that of the. Also generates performance data _MD R -k performance data _MD R8 -k the note number of the first track TR-1 event _{EV ON} (♯) is rewritten to that increase 5 degrees higher. Also generates performance data _MD R -k performance data _MD R9 -k first track TR-1 event _{EV ON} the note number (♯) is rewritten to that length 6 degrees higher. Also generates performance data _MD R -k performance data _{MD R10} -k first track TR-1 event _{EV ON} the note number (♯) rewritten to those minor 7 degrees higher. Also generates performance data _MD R -k of the first track TR-1 event _{EV ON} (♯) performance data _{MD R11} -k the note number is rewritten to what length 7 degrees high. Here, in the following, performance data MD _R -k in C major and performance data MD _R1- k, MD _R2- k, MD _R3- k, MD _R4- k, MD _R5- k, MD _R6- k, MD _R7- k, MD _R8- k, MD _R9- k, MD _R10- k, and MD _R11- k are described as performance data MD _R- k.

When “no transposition” is set (S112: No), the CPU 16 proceeds to Step S114 without executing Step S113. In step S114, it is determined whether the variable k and the division number Z of the music performance data SMD are the same. When k <Z is satisfied (S114: No), the CPU 16 proceeds to step S115 to set the variable k to k = k + 1, and then repeats the processing after step S103. When k = Z is satisfied (S114: Yes), the CPU 16 performs the performance data MD _R -k, MD _R1- k, MD _R2- k, MD _R3- k, MD _R4- k, MD _R5- k, MD _R6- k, MD _R7- k, MD _R8- k, MD _R9- k, MD _R10- k, MD _R11- k, MD _RA- k, MD _RA1- k, MD _RA2- k, MD _{RA3- k, MD RA4 -k, MD RA5} -k, MD RA6 -k, MD RA7 -k, MD RA8 -k, MD RA9 -k, MD RA10 -k, MD RA11 material database each of the -k as performance data MD Store in the MDB (S116).

FIG. 6 is a flowchart showing the editing process. A series of processing shown in FIG. 6 is processing executed by the function of the editing function. A series of processing shown in FIG. 6 is started when editing of the performance data MD is instructed by the operation unit 12. In executing the series of processes shown in FIG. 6, the user sets the following items using the operation unit 12 and stores the setting values in the RAM 17.
a6. Chord progression This item sets the codename CN ₁ -CN ₈ of each code of the first to eight measures in the chord progression 8 bars of music.

In FIG. 6, the CPU 16 performs performance data acquisition processing (S201). In the performance data acquisition process, the CPU 16 performs the performance data MD _1-2 as the material of the melody of the first to second measures, the performance data MD _3-4 as the material of the melody of the third to fourth measures, Among the plurality of performance data MD in the material database MDB, the performance data MD _5-6 as the melody material of the fifth to sixth measures and the performance data MD _7-8 as the melody material of the seventh to eighth measures are used. The search result is read out to the RAM 17. More specifically, the CPU 16 performs the following four data search processes.

In the first data search process, the CPU 16 selects candidates for the performance data MD _{1-2 from} all the performance data MD in the material database MDB, in which the event EV _ON (123) is embedded in the third track TR-3. and then calculates the evaluation point P that indicates the goodness of fit between the chords codename CN ₁ and CN ₂ in the setting item a5 each performance data MD that is a candidate of the performance data _{MD 1 to 2.} The procedure for calculating the evaluation point P is as follows. As shown in FIG. 7, performance data MD (performance data MD _R -k, MD _R1- k, MD _R2- k, MD _R3- k, MD _R4- k, MD _R5 -k, MD _R6 -k, MD _R7 -k, MD _R8 -k, MD _R9 -k, MD _R10 -k, MD _R11 -k)), the CPU 16 selects the performance data MD. An event EV _ON (#) for two measures of one track TR-1 is an evaluation target. Then, a total value SUM _TCK of the sound lengths (number of ticks) of each event EV _ON (#) of two measures to be evaluated is obtained. Next, the sound name of each sound indicated by the event EV _ON (#) of the first one measure to be evaluated is the sound of the chord corresponding to the chord name CN ₁ (in the example of FIG. 7, CN ₁ = C. 3 sounds), and each tone of the 5-tone scale (Pentatonic) that shares this chord and root sound (in the example of FIG. 7, CN ₁ = C, so de, re, mi, so, 5 points, and the sound length (number of ticks) that matches each chord sound is 5 points. The sound that does not match each chord sound but matches each sound of the 5 scales The sound length (number of ticks) is multiplied by 3 points, and the sound length (number of ticks) of other sounds is multiplied by 1 point. Next, the pitch name of each sound indicated by the event EV _ON (#) of the second measure 1 to be evaluated is a chord of the chord name CN ₂ (in the example of FIG. 7, CN ₂ = F, so 3 sounds) and each tone of the 5-tone scale (Pentatonic) that shares this chord with the root note (in the example of FIG. 7, CN ₂ = F, so 5 sounds of fa, so, la, de, les) ), 5 points for the sound length (number of ticks) that matches each sound of the chord, and the length of the sound that does not match each sound of the chord but matches each sound of the 5 scales ( The number of ticks) is multiplied by 3 points, and the sound length (number of ticks) of other sounds is multiplied by 1 point. Then, a sum SUM _{PNT of} values obtained by multiplying the sound length (number of ticks) in each event EV _ON (#) of the two measures by the corresponding point is obtained. Then, a value SUM _PNT / SUM _TCK obtained by dividing the total value SUM _TCK by the total value SUM _PNT is set as the evaluation point P of the performance data MD.

Further, as shown in FIG. 8, the performance data MD (MD _RA- k, MD _RA1- k, MD _RA2- k, MD _RA3- k, MD _RA4- k, MD _{RA5) whose} evaluation points P are to be calculated include the _{output. -k, MD RA6 -k, MD RA7} -k, MD RA8 -k, MD RA9 -k, MD RA10 -k, if any) of the _{MD RA11} -k, CPU 16 may first of the performance data MD An event EV _ON (#) for two measures after the end of the output in the track TR-1 is an evaluation target. Then, a total value SUM _TCK of the sound lengths (number of ticks) of each event EV _ON (#) of two measures to be evaluated is obtained. Next, the pitch name of each tone indicated by the event EV _ON (#) of the first one measure to be evaluated is the pitch of the chord corresponding to the chord name CN ₁ (in the example of FIG. 8, CN ₁ = C. 3 sounds), and each tone of the 5-tone scale (Pentatonic) that shares this chord and root sound (in the example of FIG. 8, CN ₁ = C, so de, re, mi, so, 5 points, and the sound length (number of ticks) that matches each chord sound is 5 points. The sound that does not match each chord sound but matches each sound of the 5 scales The sound length (number of ticks) is multiplied by 3 points, and the sound length (number of ticks) of other sounds is multiplied by 1 point. Next, the pitch name of each sound indicated by the event EV _ON (#) of the second measure 1 to be evaluated is a chord of the chord name CN ₂ (in the example of FIG. 8, CN ₂ = F, so 3 sounds) and 5 tones (Pentatonic) sounds that share this chord and root sound (in the example of FIG. 8, CN ₂ = F, so 5 sounds of Fa, So, La, De, and Les) ), 5 points for the sound length (number of ticks) that matches each sound of the chord, and the length of the sound that does not match each sound of the chord but matches each sound of the 5 scales ( The number of ticks) is multiplied by 3 points, and the sound length (number of ticks) of other sounds is multiplied by 1 point. Then, a sum SUM _{PNT of} values obtained by multiplying the sound length (number of ticks) in each event EV _ON (#) of the two measures by the corresponding point is obtained. Then, a value SUM _PNT / SUM _TCK obtained by dividing the total value SUM _TCK by the total value SUM _PNT is set as the evaluation point P of the performance data MD.

In the first data search process, the CPU 16 obtains the evaluation point P for all the performance data MD in which the event EV _ON (123) is embedded, and then determines the performance data MD _1-2 with the highest evaluation point P. Is stored in the RAM 17.

In the second data search process, the CPU 16 selects candidates for the performance data MD _{3 to 4} that have the event EV _ON (124) embedded in the third track TR-3 among all the performance data MD in the material database MDB. and then, converts the fit of the chord codename CN ₃ and CN ₄ in the setting item a5 each performance data MD which is of the performance data MD _{3 to 4} candidate evaluation point P, those evaluation point P has the highest The performance data MD _{3 to 4} are stored in the RAM 17.

In the third data search process, the CPU 16 selects candidates for the performance data MDs _{5 to 6} that have the event EV _ON (124) embedded in the third track TR-3 among all the performance data MD in the material database MDB. The matching score between each performance data MD, which is a candidate for the performance data MD _5-6 , and the chord of the chord names CN ₅ and CN ₆ in the setting item a5 is converted into an evaluation point P, and the one with the highest evaluation point P is converted. stored as performance data _{MD 5~6} to RAM17.

In the fourth data search process, the CPU 16 selects the performance data MD _{7 to 8} that have the event EV _ON (125) embedded in the third track TR-3 among all the performance data MD in the material database MDB. The matching score between each performance data MD that is a candidate for the performance data MD ₇ to _{8 and} the chords of the chord names CN ₇ and CN ₈ in the setting item a5 is converted into an evaluation point P, and the one with the highest evaluation point P is converted. The performance data MD _{7 to 8} is stored in the RAM 17.

In FIG. 6, the CPU 16 performs a first array process (S202). In the first arrangement processing, the CPU 16 arranges the performance data MD ₇ to ₈ after the performance data MD _{5 to 6} , and the events EV _ON (#) and EV _{OFF for} the four measures of the fifth to eighth measures of the song. Performance data MD ₅ to ₈ including (#) are generated. More specifically, in the case where the type of the subsequent performance data MD _{7 to 8} of the two performance data MD _{5 to 6} and MD _{7 to 8} whose order of arrangement is in sequence is the performance data MD without an outcome, The following processing is performed for each of the performance data MD with the output.

a7. In the case where the performance data MD ₇ to ₈ are performance data MD without an output in this case, the CPU 16 performs the events EV _ON (#), EV _OFF (#), and delta of each track TR-j of the performance data MD _{5 to 6} each track TR-j event _{EV oN} performance data _{MD 7 to 8} after the time _{ΔT h (♯), EV OFF} (♯), and the performance data _{MD 5 to 8} those connecting the delta time [delta] T _h.

b7. In the case where the performance data MD ₇ to ₈ are performance data MD with an effect In this case, the CPU 16 performs the events EV _ON (#) and EV _OFF (for the performance data MD _{5 to 6} on the performance data MD _{5 to 6} ). The end of the melody consisting of each sound indicated by #) is the amount of the outfact in the melody consisting of the sound indicated by each event EV _ON (#) and EV _OFF (#) of the performance data MD _{7 to 8} after the arrangement order. Rewrite processing to shorten it. On top of that, each track TR-j event _{EV ON} performance data MD _{'5 to 6} which has passed through the rewriting process _{(♯), EV OFF (♯} ), and each of the performance data _{MD 7 to 8} after the delta time [Delta] T _h event _{EV ON} of the track _{TR-j (♯), EV} OFF (♯), and the performance data _{MD 5~8} those connecting the delta time ΔT _h. The detailed procedure of the rewriting process is as follows. In this rewriting process, first, an after-performance performance section in the melody composed of sounds indicated by the events EV _ON (#) and EV _OFF (#) of the first track TR-1 of the performance data MD _{7 to 8} is performed as an after-performance. and section _{R a,} first track TR-1 event _{EV ON} performance data _{MD 5 to 6} (♯) and _{EV OFF} (♯) pick-up (upbeat) performance than the end of the play interval _{R 5 to 6} of the melody composed of the sound indicated by the The time before the section _RA is set as the reference time t _BASIS (FIG. 9).

Then, the reference time t _BASIS the preceding and the reference time t _BASIS closest time to the execution time t from a series of events _EV ON the second track TR-2 performance data _{MD 5 to 6} (126) Event EV _OFF (126) is selected (FIG. 10). Next, the events EV _ON (#), EV _OFF (#) and the delta time ΔT _h after the selected event EV _OFF (126) in each track TR-j of the performance data MD _5-6 are deleted. (FIG. 11). Next, the last event EV _OFF (#) is selected from among the events EV _ON (#) and EV _OFF (#) that remain in the first track TR-1 of the performance data MD _5-6 , and this is selected. selected event _{EV OFF} (♯) delta indicating the execution time t of the event _{EV OFF} as silencing time coincides with the reference time t _bASIS sound indicated (♯) time [delta] t _h (i.e., the event _{EV OFF} (♯ The data obtained by rewriting the delta time ΔT _h ) before) is set as performance data MD ′ ₅ to ₆ (FIG. 12). When the subsequent performance data MD _7-8 is performance data MD with an effect, each of the performance data MD _7-8 after each track TR-j of the performance data MD ' _5-6 obtained through the above processing. The performance data MD ₅ to ₈ are connected to the tracks TR-j (FIG. 13).

In FIG. 6, the CPU 16 performs a second array process (S203). In the second arrangement process, the CPU 16 arranges the performance data MD ₅ to ₈ after the performance data MD _{3 to 4} , and the events EV _ON (#) and EV _{OFF for} the six bars of the third to eighth bars of the music. Performance data MD ₃ to ₈ including (#) are generated. In the second arrangement process, when the type of musical data _{MD 5 to 8} of the performance of the two that arrangement order is tandem data _{MD 3 to 4} and _{MD 5 to 8} are performance data MD having no pick-up (upbeat) is playing each track TR-j of the event _{EV oN} data _{MD 3~4 (♯), EV OFF} (♯), and each track TR-j of the event _{EV oN} performance data _{MD 5~8} after the delta time ΔT _h (♯ _{), EV OFF (♯),} and the performance data _{MD 3 to 8} those connecting the delta time [delta] T _h. Further, when the type of the performance data MD _{5 to 8} is performance data MD with an effect, the rewriting process described above is performed on the performance data MD ₃ to ₄ . After that, each of the performance data MD _5-8 after the event EV _ON (#), EV _OFF (#) and the delta time ΔT _h of each track TR-j of the performance data MD ′ _3-4 after the rewriting process. event _{EV ON} tracks _{TR-j (♯), EV} OFF (♯), and the performance data _{MD 3 to 8} those connecting the delta time [delta] T _h.

Next, the CPU 16 performs a third array process (S204). In the third arrangement processing, the CPU 16 arranges the performance data MD ₃ to ₈ after the performance data MD _{1 to 2} to generate one piece of music performance data EMD (eight bars of the first to eighth bars). To do. In the third arrangement process, if the type of the performance data MD _3-8 is the performance data MD without the out of the _two performance data MD _1-2 and MD _3-8 , the arrangement order of which is the order, the performance data each track TR-j of the event _{EV oN} of _{MD 1~2 (♯), EV OFF} (♯), and each track TR-j of the event _{EV oN} performance data _{MD 3~8} after the delta time ΔT _h (♯) , _{EV OFF} (♯), and those connecting the delta time ΔT _h and song performance data EMD. Further, when the type of the performance data MD _{3 to 8} is performance data MD with an effect, the rewriting process described above is performed on the performance data MD _{1 and 2} . After that, each of the performance data MD _{3 to 8} after the event EV _ON (#), EV _OFF (#), and the delta time ΔT _h of each track TR-j of the performance data MD ′ _{1 to 2} subjected to the rewriting process. event _{EV ON} of the track _{TR-j (♯), EV} OFF (♯), and those connecting the delta time ΔT _h and song performance data EMD.

  Next, the CPU 16 stores the music performance data EMD for one music obtained by the first to third arrangement processes in the music database BDB (S205).

  FIG. 14 is a flowchart showing the song synthesis process. A series of processes shown in FIG. 14 is a process executed by the function of the song synthesis function. The series of processes shown in FIG. 14 is started when the operation unit 12 instructs the generation of singing voice. In executing the series of processes shown in FIG. 14, the user inputs lyrics to be placed on the melody of the song as a character string of hiragana characters.

  In FIG. 14, the CPU 16 converts a character string input as lyrics into a series of speech segments (S301). The CPU 16 reads the speech unit data PD corresponding to each speech unit obtained in step S301 from the speech unit database PDB to the RAM 17 (S302). The CPU 16 generates a speech signal corresponding to the character string input as lyrics by connecting the speech segment data PD in the RAM 17 (S303). The CPU 16 reads one piece of music performance data EMD stored in the music database BDB into the RAM 17 (S304). The CPU 16 performs a pitch conversion process for matching the pitch of the audio signal generated in step S304 with the melody of the music performance data EMD (S305). The CPU 16 emits the sound signal subjected to the pitch conversion process as sound from the sound emitting unit 14 (S306). Please refer to patent document 2 for the detailed contents of each step of this song composition process.

The above is the details of the present embodiment. According to this embodiment, the following effects can be obtained.
First, in the present embodiment, in the editing process, when the performance data MD having an effect is arranged after the performance data MD, the performance data MD whose arrangement order is before is subjected to a rewrite process, and the performance data that has undergone the rewrite process. The corresponding performance data MD is arranged after the data MD. Therefore, it is possible to create a musical composition rich in variations using a plurality of performance data MD recorded in the database MDB.

  Secondly, in the present embodiment, in the data accumulation processing, performance data MD in a key other than C major is generated by performing transposition processing on performance data MD in C major, and these performance data MD are converted into the original C major. The performance data MD is stored in the material database MDB. Therefore, according to this embodiment, it is possible to create various musical compositions by combining melody of different tones.

Although one embodiment of the present invention has been described above, the present invention may have other embodiments. For example, it is as follows.
(1) In the above embodiment, the music performance data EMD for one piece of music obtained by the editing process may be emitted from the sound emitting unit 14 as a sound signal of the music.

(2) In the above embodiment, four pieces of musical performance data MD are arranged to generate one piece of musical piece performance data EMD. However, 2 to 3 performance data MD and 5 or more performance data MD may be arranged to generate music performance data EMD for one song. Further, the length of the performance section in the performance data MD does not have to be two bars, and may be one bar, or may be three bars or more.

(3) In the first marking process in the above embodiment, after determining the melody break position in the performance section Ri, the event EV _OFF (#) immediately before the break position in the first track TR-j is selected. Then, the delta time ΔT before and after the event EV _ON (126) is rewritten so that the execution time t of the selected event EV _OFF (#) and the event EV _ON (126) in the second track is the same time. However, the event EV _ON (#) immediately after the break position in the first track TR-j (that is, the event EV _ON (#) instructing the sound generation immediately after the rest having a length equal to or greater than the threshold TH3) is selected. Even if the delta time ΔT before and after the event EV _ON (126) is rewritten so that the execution time t of the selected event EV _ON (#) and the event EV _ON (126) in the second track is the same time. Good.

(4) In the above embodiment, the event EV _ON (126), which is the mark of the event EV _OFF (#) immediately before the performance break position in the first track TR-j, is described in the second track TR-j. However, an event that becomes the mark of the event EV _OFF (#) immediately before the performance break position in the first track TR-j may be described in the first track TR-j. For example, the mark of the previous event EV _OFF delimiter position of playing a predetermined type of control change events _(♯), the control change and the immediately preceding event EV _OFF delimiter position of the performance in the first track TR-j _(♯) Events may be described continuously.

(5) In the rewriting process in the above-described embodiment, the reference time t _BASIS is selected from the series of events EV _ON (126) in the second track TR-2 in which the arrangement order is the previous of the two consecutive performance data MD. An event EV _OFF (126) that precedes and is closest to the reference time t _BASIS is selected as the event EV _OFF (126), and is later than the selected event EV _OFF (126) in each track TR-j of the performance data MD. of events _{EV ON (♯), EV OFF} (♯), and to erase the delta time ΔT _h. However, there may be a case where no event EV _ON (126) is written on the second track TR-2. In this case, the CPU 16 _selects the event EV _OFF (after the reference time t _BASIS from the series of events EV _ON (#) and EV _OFF (#) in the first track TR-1 of the performance data MD having the previous arrangement order. ♯) is a an event _{EV OFF} to the time closest to the reference time t _bASIS the execution time t (♯) select, from the selected event _{EV OFF} in each track TR-j of the performance data MD (♯) event _{EV ON} after also _{(♯), EV OFF (♯} ), and may want to erase the delta time ΔT _h.

(6) In the above embodiment, the music performance data SMD and the performance data MD obtained by dividing the music performance data SMD are the note-on event EV _ON (#), the note-off event EV _OFF (#), and the delta time ΔT _h as L The tracks TR-j (j = 1 to L) were divided and arranged. However, the music performance data SMD and the performance data MD obtained by dividing the music performance data SMD may have a different data format. For example, a format in which performance data SMD and MD are arranged in time series in the order of records of track numbers of corresponding tracks, sound source channels of sound sources, timing data indicating event execution times, event types, and event parameters. (List format). Further, the performance data SMD and MD may be in a format in which records consisting only of event types and their parameters are arranged in chronological order. The performance data SMD and MD may be in a piano roll format or a score roll format.

(7) In the above embodiment, the event EV _ON (126) of the note number 126 is used as a mark indicating the performance break position in the music performance data SMD, and the event EV _ON (125) of the note number 125 is the last of the song. The event EV _ON (123) of the note number 123 is set as a mark indicating the performance section appropriate for the melody of the part, the event EV _ON (124) of the note number 124 is set as the mark indicating the performance section appropriate to the melody of the first part of the song. Is a mark indicating a performance section suitable for the melody of the song that is not the first or last part of the song. However, another type of data may be adopted as a mark indicating a performance section suitable for a performance break position or a melody of each part of a song. The data serving as the mark may be data in a format not supported by MIDI.

(8) In the transposition processing of the above embodiment, 11 types of performance data MD different from C major are obtained by replacing the note numbers of each event EV _ON (#) in C major performance data MD _R -k with higher semitones. _R1- k, MD _R2- k, MD _R3- k, MD _R4- k, MD _R5- k, MD _R6- k, MD _R7- k, MD _R8- k, MD _R9- k, MD _R10- k, MD _R11- k was produced. However, the key type may be set by operating the operation unit 12, and only the performance data MD-k of the set key may be generated. Further, after the transposition processing, 11 types of performance data MD _R1 -k, MD _R2 -k, MD _R3 -k, MD _R4 -k, MD _R5 -k, MD _R6 -k, MD _R7 -k, MD _R8 -k , MD _R9 -k, MD _R10 -k, MD _R11 -k may be subjected to a process of removing the articulation symbols in each musical score, and the data content may be corrected to performance data MD indicating C major melody.

(9) In the output generation process of the above embodiment, the CPU 16 may generate the performance data MD _RA -k as follows. First, to produce a copy of the performance data _MD R -k. Then, the play interval R-k in the performance data _MD R -k is the replication origin was the time t0, the time t0 as 1 / h (e.g., h = 2) between the time _{t B} after only bars playing interval event _{EV ON} in _{R B} between _{(♯), EV OFF (♯} ), and replicating the delta time [delta] T. The event _{EV ON} is this replication _{(♯), EV OFF (♯} ), and playing those append before each track TR-j delta time ΔT the original performance data _MD R -k data _{MD RA} - k.

(10) In the performance data acquisition process of the above embodiment, the series of events EV _ON (#) to be evaluated match the chords of chord names CN ₁ and CN ₂ , and they do not match chords but are five tone scales Are classified into three types: those that match the sound of the chord and those that do not match any of the chord and the five-tone scale, and multiply the sound length (number of ticks) of each event EV _ON (#) by the point according to each type did. However, a series of events EV _ON (#) to be evaluated matches those of chord names CN ₁ and CN ₂ , does not match chords but matches chord sounds, chords and five tones Each event EV _ON is classified into four types: one that does not match any note in the scale or one that does not match any other note in the C major scale, one that does not match any sound in the chord, the fifth scale, or the C major scale. The sound length (number of ticks) of (#) may be multiplied by a point corresponding to each type. In this case, if it falls under the third type (one that does not match any of the chord and the fifth note scale, or one that matches the other C major note), multiply by one point, It is better not to multiply the points when it corresponds to the type (one that does not match any of the chord and the five-tone scale, or matches the other C major notes).

(11) In the output generation process of the above embodiment, the event EV _ON () included in each performance data MD _RA -i so that the successive performance data MD _RA -i and MD _RA- (i + 1) have overlapping portions. ♯) and _{EV OFF} (♯) and to determine the cut-out position of the delta time [delta] T _h. However, for example, playing section R-1 to R-3 in the event _{EV ON} in music performance data _{SMD (♯), EV OFF (} ♯), the delta time [Delta] T _h and performance data _MD RA -1, playing interval R- 4~R-6 in the event _{EV ON (♯), EV OFF} (♯), so that the delta-time [delta] T _h and performance data _MD RA -2 ..., the performance data _MD RA -i and _{MD RA} preceding and succeeding - (i + 1) may determine the event _{EV oN} (♯) and _{EV OFF} (♯) and cut-out position of the delta time [delta] T _h including the overlap portion to have no within each performance data _MD RA -i.

(12) In the above embodiment, the music performance data SMD and MD are an array of data indicating the pronunciation content of each sound in the melody, which is a sequence of sounds for a certain length of time. However, the music performance data SMD and MD may be an array of data indicating the pronunciation content of each sound in a sequence of sounds that determine the harmony and rhythm for a certain length of time.

(13) In the first to fourth arrangement processes in the above embodiment, the arrangement order of the two performance data MD (for example, performance data MD _5-6 and MD _7-8 ) to be processed is the previous one. The last event EV _OFF (#) is selected from among the events EV _ON (#) and EV _OFF (#) that remain in the first track TR-1 of the performance data MD _5-6 , and this selected event is selected. EV _OFF (♯) rewrites the delta time [delta] t _h indicating the execution time t of the event _{EV OFF} as silencing time coincides with the reference time t _bASIS sound indicated (♯). However, the difference between the mute time of the sound indicated by the event event EV _OFF (#) remaining without being erased in the first track TR-1 of the performance data MD _5-6 of the previous arrangement order and the reference time t _BASIS is predetermined. may be rewritten delta time [delta] t _h indicating the execution time t of the event EV _OFF to be less than the value _(♯). In short, in the first to fourth arrangement processes, with respect to the performance data MD having the previous arrangement order, the performance data having the arrangement order subsequent to the end of the arrangement of the sounds indicated by the events of the performance data MD. It is only necessary to execute a process of shortening by the amount of approximately the outfact in the sound sequence indicated by each event.

DESCRIPTION OF SYMBOLS 10 ... Song editing apparatus, 11 ... Data reading part, 12 ... Operation part, 13 ... Display part, 14 ... Sound emission part, 15 ... Control part, 16 ... CPU, 17 ... RAM, 18 ... ROM, 19 ... Hard disk.

Claims (5)

M (where M is an integer equal to or greater than 1) bars, a first type of performance data including each event indicating the sound content of each sound and timing data indicating the execution time of each event; Storage means for storing a second type of performance data including each event indicating the sound content of each sound in the subsequent M-bar sound sequence and timing data indicating the execution time of each event;
Editing means for acquiring and arranging a plurality of performance data from the storage means, and arranging the second type of performance data after the first or second type of performance data. For the previous performance data, the end of the sound sequence consisting of the sound indicated by each event of the performance data is the amount corresponding to the approximate output in the sound sequence indicated by the event of the performance data after the arrangement order. A music editing apparatus comprising: editing means for performing a rewriting process for shortening and arranging the performance data after the performance data subjected to the rewriting process. The editing means includes
In the rewriting process,
In the sequence of sounds indicated by each event of the performance data whose arrangement order is later, the performance section of the sound of the after-effects is defined as the after-performance period, and from the end of the performance section consisting of the sound indicated by the event of the previous performance data The time before the above-mentioned outcome performance section is set as the reference time, and the sequence is preceded by the reference time from the events corresponding to the positions of the sound arrangement breaks in the previous performance data, and the reference time is the longest. Select an event whose execution time is the nearest time, erase the event and timing data after this selected event, and the mute time of the sound indicated by the last event among the remaining events without being erased is the reference time. The music editing apparatus according to claim 1, wherein the timing data indicating the execution time of the last event is rewritten so as to substantially coincide with. Of the events of the first and second types of performance data, a mark is added to a sound event that is sounded through a rest longer than a predetermined value,
The editing means includes
In the rewriting process,
Of the events included in the previous performance data in the arrangement order, the event with a mark is set as each event corresponding to the delimiter position, and the event preceding the reference time and the most in the reference time is selected from the events. The music editing apparatus according to claim 2, wherein an event whose execution time is a near time is selected.   Each event indicating each sound in a performance section longer than M bars and the execution time of each event from the music performance data including each event indicating the arrangement of sounds for one song and timing data indicating the execution time of each event 4. The method according to claim 1, further comprising means for cutting out the timing data shown and storing each of the cut-out events and timing data in the storage means as the second type of performance data. The music editing device described in 1. On the computer,
M (where M is an integer equal to or greater than 1) bars, a first type of performance data including each event indicating the sound content of each sound and timing data indicating the execution time of each event; Storage means for storing a second type of performance data including each event indicating the sound content of each sound in the subsequent M-bar sound sequence and timing data indicating the execution time of each event;
Editing means for acquiring and arranging a plurality of performance data from the storage means, and arranging the second type of performance data after the first or second type of performance data. For the previous performance data, the end of the sound sequence consisting of the sound indicated by each event of the performance data is the amount corresponding to the approximate output in the sound sequence indicated by the event of the performance data after the arrangement order. A program that implements a rewriting process for shortening and editing means for arranging the corresponding performance data after the performance data subjected to the rewriting process.

Images