JP3518253B2 - Data editing device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data editing device, and more particularly to a data editing device capable of editing audio data.

[0002]

2. Description of the Related Art As a conventional technique, there is known a sound producing apparatus provided with a speech synthesis circuit such as a formant sound source. The formant sound source is formed by synthesizing a formant formed by frequency analysis of a sound. It is a sound source that generates an audio signal.

Here, the voice will be briefly described.
The voice is, for example, a so-called Japanese syllabary in which various pronunciations are pronounced, in particular, a voice when the lyrics of a song are pronounced.
Sounds are called syllables. Japanese "ka" etc. correspond to syllables.
In addition, syllables are composed of phoneme strings on the time axis,
For example, the syllable “ka” is the phoneme string “CL (7.5 ms) +
kha (4 × 7.5 ms) + aj (infinite length) ”(the value in the parentheses represents each sounding time). Then, the minimum unit element in the phoneme string, for example, “CL” or “kh”
"a" and "aj" are called phonemes. The formant sound source is a sound source that inputs information on such a phoneme string to synthesize formants corresponding to each phoneme described in the phoneme string information to generate a voice signal.

[0004]

It is not necessary to edit the voice data in a voice pronunciation device that produces only a predetermined voice, but in order for the voice production device to produce a wide variety of voices, It is essential to edit voice data, and the present inventor devises a voice data editing device.

The voice data described here is the MIDI for creating the musical tone data which has been conventionally known.
It is based on the standard. By creating the audio data based on the MIDI standard, there is an advantage that the audio data can be created in association with the conventional musical sound data.
Then, the voice data editing apparatus also uses the musical tone data (MI
Although it is created on the basis of a DI data) editing device, since the voice data is constructed according to a specific rule, it is difficult for a beginner who is not familiar with the rule to edit the voice data.

An object of the present invention is to provide a data editing device or a recording medium of a computer program capable of easily editing audio data.

[0007]

According to one aspect of the present invention, audio data including syllable data composed of a plurality of phoneme data whose occurrence time is defined by duration data that is independent of reproduction tempo, and the audio data. A data editing device capable of editing the syllable data in song data including timing data depending on a reproduction tempo that defines a sounding timing of the data is one of the plurality of phoneme data constituting the syllable data. And a changing unit for changing the duration data corresponding to the phoneme data designated by the phoneme designating unit to an occurrence time designated independently of the designation of the reproduction tempo.

The operator can designate any one of the phoneme data that constitutes the syllable data and change the occurrence time data corresponding to the phoneme data to a desired value.

According to another aspect of the present invention, audio data including syllable data composed of a plurality of phoneme data whose generation time is defined by duration data independent of reproduction tempo and sounding timing of the audio data are defined. A data editing apparatus capable of editing song data including timing data depending on a specified reproduction tempo is a syllable data and each syllable data determined according to characteristics of a plurality of phoneme data forming the syllable data. Table for storing the correspondence relationship with the sounding timing change value of No. 1, the designating unit for designating the voice data, and the pronunciation of the syllable data located at the head among the syllable data constituting the voice data designated by the designating unit. Detecting means for detecting the timing change value from the table, and voice specified by the specifying means Changing means for changing the timing data for the data by the sounding timing change value detected by the detecting means.

By designating the voice data, the operator can change the tone generation timing data for the voice data by a desired value according to the table.

According to still another aspect of the present invention, voice data including syllable data composed of a plurality of phoneme data whose generation time is defined by duration data that is independent of reproduction tempo and sounding timing of the voice data. The data editing device capable of editing the syllable data in the song data including the timing data depending on the reproduction tempo that defines the syllabic data includes a designating means for designating the syllable data, and a syllable data designated by the designating means. It has a detecting means for detecting phoneme data belonging to a consonant among a plurality of included phoneme data, and a shortening means for shortening and changing the duration data of the phoneme data detected by the detecting means.

By specifying the syllable data, the operator can shorten and change the duration data of the phoneme data included in the syllable data.

[0013]

FIG. 4 is a diagram showing an example of a voice data editing screen of a data editing apparatus according to an embodiment of the present invention. The voice data includes note data and phoneme string data. The note data is a note number (for example, E
3) and the pronunciation start time. The phoneme string data includes data related to each phoneme and expiration, that is, data expressing breathing when a person produces a voice (hereinafter, referred to as expiration data). The phoneme string data has a one-to-one correspondence with each character displayed as "lyrics" in FIG. Once the "lyrics" are decided, the phoneme string data is automatically decided.

For example, the data on the second line will be described. The lyric “ka” is the phoneme string data “CL + kha + a
j ”, and the pitch of F # 3 (note)
Will be pronounced for a predetermined duration. As shown in FIG.
By enumerating the above data, it is possible to create voice data for one song for causing the voice sounding device to sing a desired song.

FIG. 5 is a diagram showing the connection between the personal computer 1 and the external device 21. The personal computer 1 includes the data editing device according to this embodiment.
The personal computer 1 may be a sequencer. The external device 21 includes a voice pronunciation device.

First, the configuration of the external device 21 will be described. The detection circuit 33 detects the operation of the operator 34 and generates an operator signal. The operator 34 is, for example, a setting operator such as a switch or a performance operator such as a keyboard.

In addition to the detection circuit 33, the bus 22 has a display circuit 28, a sound source circuit 29, an effect circuit 30, and a voice synthesis circuit 3.
2, interface 23, RAM 24, ROM 25,
The CPU 26 is connected.

The ROM 25 stores formant data for synthesizing voice, other various data, and a computer program. The RAM 24 stores flags, buffers, accompaniment data, and the like. The computer program may be stored in the RAM 24 instead of being stored in the ROM 25. CPU 26 is ROM 25 or RAM
According to the computer program stored in 24,
Calculate or control.

The timer 27 is connected to the CPU 26.
The CPU 26 can obtain time information from the timer 27. The CPU 26 reproduces accompaniment data or audio data (song data) according to the time information.

The CPU 26 reads the accompaniment data stored in the RAM 24 to generate musical tone parameters and effect parameters, and supplies them to the tone generator circuit 29 and the effect circuit 30, respectively.

The CPU 26 can also generate a tone parameter and an effect parameter in accordance with a note-on signal generated by the detection circuit 33 and supply them to the tone generator circuit 29 and the effect circuit 30, respectively.

The tone generator circuit 29 generates a tone signal in accordance with the tone parameter supplied (for example, the accompaniment data). The effect circuit 30 adds effects such as delay and reverb to the musical tone signal generated by the tone generator circuit 29 according to the supplied effect parameter, and supplies it to the sound system 31. The sound system 31 includes a D / A converter and a speaker, converts a supplied digital tone signal into an analog signal, and produces a sound.

The tone generator circuit 29 is a waveform memory type,
The FM system, the physical model system, the harmonic synthesis system, the formant synthesis system, the VCO + VCF + VCA analog synthesizer system, or the like may be used.

Interface 23 includes a MIDI interface or other interface for a communication network. For example, the MID of the external device 21
The I interface 23 and the MIDI interface 8 of the personal computer 1 are connected by a MIDI cable. The external device 21 and the personal computer 1 can perform MIDI communication.

The CPU 26 receives voice data (song data) from the personal computer 1 via the interface 23 and stores it in the RAM 24. The voice data includes note data and phoneme string data as described above. Details will be described later with reference to FIG. CP
The U 26 reads out the voice data stored in the RAM 24 and supplies the formant data to the voice synthesis circuit 32 based on the formant data stored in the ROM 25. The formant data is, for example, formant center frequency (center frequency of a mountain forming a formant) data required to form a formant corresponding to each phoneme, formant band width (bandwidth of a mountain forming a formant) data, The formant level (peak level of mountains forming the formant) data and the like.

The voice synthesis circuit 32 generates a voice signal according to the supplied formant data. The audio signal is
It has a predetermined pitch and corresponds to a singing voice. Voice synthesis circuit 32
May be a formant synthesis method (formant sound source) or any other method. The formant synthesis method has, for example, the configuration shown in FIG. 1 of JP-A-3-200299.

The effect circuit 30 adds an effect such as delay to the sound signal generated by the sound synthesizing circuit 32 according to the supplied effect parameter and supplies the sound signal to the sound system 31. The sound system 31 converts the supplied digital audio signal into an analog format and produces sound.

The voice synthesis circuit 32 and the sound source circuit 29
Is not limited to being configured by using dedicated hardware, but may be configured by using a DSP + microprogram, or may be configured by a program of CPU + software.

Further, a plurality of tone generation channels may be formed by using one voice synthesis circuit or tone generator circuit in a time division manner, or one tone generation channel may be formed by using a plurality of voice synthesizer circuits or tone generator circuits. A plurality of sound generation channels may be configured by one voice synthesis circuit or sound source circuit.

Next, the configuration of the personal computer 1 will be described. The detection circuit 9 detects a movement operation or a switch operation of the mouse 10 and generates a mouse signal. Detection circuit 1
Reference numeral 1 detects a key (numerical key, character key, etc.) input on the keyboard 12 and generates a key signal. The operator can use the mouse 10 or the keyboard 12 to edit the voice data.

The display circuit 7 is the personal computer 1.
Can be connected to the display 45 which is an external device. The display 45 can display the edit screen shown in FIG. The operator can edit the voice data while referring to the edit screen on the display 45.

To the bus 2, in addition to the detection circuit 9, the detection circuit 11 and the display circuit 7, the interface 8, the external storage device 13, the RAM 3, the ROM 4 and the CPU 5 are connected.

The ROM 4 stores various parameters and computer programs. The RAM 3 stores flags, buffers, voice data (song data), and the like. The RAM 3 can also store a computer program, audio data, or the like externally supplied via the external storage device 13 or the interface 8. CPU5
Performs calculation or control for editing audio data according to a computer program stored in the RAM 3 or the ROM 4.

The timer 6 supplies time information to the CPU 5. The CPU 5 performs interrupt processing at predetermined time intervals according to the time information.

Interface 8 includes a MIDI interface or other interface for a communication network. As described above, the MIDI interface 8 is connected to the MIDI interface 23 of the external device 21 by the MIDI cable. The personal computer 1 can transmit audio data and the like to the external device 21 via the MIDI interface 8.

Further, the communication interface 8 is connected to a communication network 41 such as a local area network (LAN), the Internet or a telephone line. A server computer 42 is connected to the communication network 41. The personal computer 1 can receive voice data or a computer program from the server computer 42 via the communication network 41.

The external storage device 13 includes an interface for the external storage device, and is connected to the bus 2 via the interface. The external storage device 13 is, for example, a floppy disk drive (FDD), a hard disk drive (HDD), a magneto-optical disk (MO) drive, C
It is a D-ROM (compact disc-read only memory) drive or the like. The audio data is stored in the external storage device 1.
3 or RAM3.

The computer program or the like may be stored in the external storage device 13 (for example, hard disk) without being stored in the ROM 4. R from hard disk
By reading the computer program or the like into the AM 3, the CPU 5 can be caused to perform the same operation as when the computer program or the like is stored in the ROM 4. In this way, by adding the computer program or the like from another external storage medium such as a CD-ROM to the hard disk, it is possible to easily add or upgrade the computer program or the like.

The communication interface 8 is connected to a communication network 41 such as a local area network (LAN), the Internet or a telephone line, and is connected to the server computer 42 via the communication network 41. If no computer program or the like is stored in the external storage device 13, the server computer 42
Computer programs can be downloaded from. The personal computer 1 serving as a client includes a communication interface 8 and a communication network 4.
A command requesting download of a computer program or the like is transmitted to the server computer 42 via 1. The server computer 42 receives this command,
Distributing the requested computer program and the like to the personal computer 1 via the communication network 41, and the personal computer 1 receives these computer programs and the like via the communication interface 8 and stores them in the external storage device 13. This completes the download.

FIG. 6A shows an example of the structure of song data 50 including audio data. The song data 50 is stored in the RAM 3 or the like of the personal computer 1. In the song data 50, the initial setting data SD is located at the beginning, the end data ED is located at the end, and the array of data DT is located between them. End data ED is song data 5
It means the end of 0.

The initial setting data SD is setting data necessary for performance such as tempo data. This initial setting data S
According to D, sound source setting for sounding the data DT (voice data) is made.

The data DT includes timing data 51, note numbers 52, velocities 53, gate times 54 and phoneme string data 55. The data DT is data corresponding to the pronunciation of one note.

The timing data 51 is the phoneme string data 5
5 shows the start timing for producing 5, and corresponds to the note-on timing. The timing data 51 is represented by the number of clocks when the time of 1/480 of a quarter note is one clock, for example. If you change the tempo of the song, the real time of one clock also changes. The timing data 51 can be considered as relative time.

The note number 52 indicates the pitch at which the phoneme string data 55 is produced. The velocity 53 indicates the volume when the phoneme string data 55 is sounded. Gate time 54
Indicates the time from note-on to note-off when the phoneme string data 55 is sounded. The gate time 54 is represented by the number of clocks like the timing data 51.

The phoneme string data 55 is an array of data in which phonemes and durations (phoneme pronunciation lengths) are set as one set.
For example, “CL (1 × 7.5 ms) + kha (4 × 7.5
ms) + aj (infinite length) "is stored as one phoneme string data. The last phoneme aj of the phoneme string data 55 has an infinite duration in principle. This is to change the syllable currently being pronounced so that the pronunciation is smoothly connected to the next syllable to be pronounced. The detailed data structure is shown below.

FIG. 6B shows an example of the structure of the phoneme string data 55. The phoneme string data 55 has a part number 61 and a device number 62 at the beginning, and the phoneme data PD or the exhalation data 65 are arranged after that.

The part number 61 is an identification number assigned to the voice synthesis circuit 32 (FIG. 5) for performing sound generation based on the phoneme string data. In FIG. 5, the external device 2
1 shows that there is one voice synthesis circuit 32,
Two or more voice synthesis circuits 32 may be provided. For example, by inserting a sound source board into the external device 21, the number of voice synthesis circuits 32 can be increased. One voice synthesis circuit 32 can generate a voice signal of one voice part (vocal part). At that time, the part number 61 of one voice synthesis circuit 32 can be set to "1", and the part number 61 of the other voice synthesis circuit 32 can be set to "2".

The device number 62 indicates an identification number assigned to the external device 21 equipped with the voice synthesis circuit. Although FIG. 5 shows the case where one external device 21 is connected to the personal computer 1,
Two or more external devices 21 may be connected. At that time, 1
The device number 62 of one external device 21 is "1", and the device number 62 of another external device 21 is "2".
Can be

The phoneme data PD is a set of phoneme number 63 and duration 64. The phoneme number 63 is a number representing the type of phoneme that is predetermined for each phoneme.
The phonemes are, for example, “CL”, “kha”, and “aj”. The duration 64 is a phoneme pronunciation length, and is, for example, “1 × 7.5 ms”, “4 × 7.5 ms”, or “infinite length”. The duration 64 can be represented as 1 for a time of 7.5 ms. For example, "1",
“4” and “0” are data stored as the duration 94. However, "0" represents infinite length. The duration 64 is represented by the actual time that is actually generated with 7.5 ms as one unit.
Gate time 54 and timing data 51 shown in (A)
Different from If the duration 64 is represented by the number of clocks like the gate time 54, there is a possibility that the sound generation time will change depending on the model of the external device 21 (FIG. 5) used. That is, depending on the model of the external device 21, 1
There are adverse effects due to different clock times. In order to prevent such an adverse effect, it is necessary to express the duration in real time (absolute time).

The expiration data 65 is data indicating that breathing is expressed. When the exhalation data 65 is located after the phoneme data PD, after the phoneme data PD is sounded,
We show that the pronunciation is changed to silence for breath expression. The exhalation data 65 is not limited to being located at the end of the phoneme string data 55, and may be located between the phoneme data PD and another phoneme data PD.

Next, the main function of the data editing function is
I will explain one. (1) Syllable Editing Function In the phoneme sequence forming a syllable, the duration 64 at which each phoneme is emitted is predetermined. For example, in the case of the syllable "ka" as described above, CL is 7.5 ms and k
ha is set to 4 × 7.5 ms and aj is set to an infinite length. However, if this time is fixed, only one pattern of syllable "ka" can be pronounced, resulting in monotonous pronunciation.

The syllable editing function is a function capable of editing the duration of each phoneme. For example, the duration of the phoneme “kha” is changed from 4 × 7.5 ms to 3
It can be changed to × 7.5 ms or 5 × 7.5 ms. By using this function, the duration can be freely set, so that even when pronouncing the syllable "ka", it is possible to pronounce various "ka".

Further, the syllable editing function can add a new phoneme to the phoneme string data forming the syllable. For example, by adding a new phoneme to the phoneme sequence “CL + kha + aj”, it is possible to realize the pronunciation of “ka” that seems to be nasal. Furthermore, instead of adding phonemes, syllables may be added.

The processing for realizing the above syllable editing function will be described later with reference to the flowchart of FIG.

(2) Sound Generation Timing Change Function The external device 21 (FIG. 5) can simultaneously generate a plurality of parts of the accompaniment part and vocal (voice) part. 1A shows an example of data on the time axis of an accompaniment part, and FIG. 1B shows an example of data on the time axis of a vocal part (song data).

The accompaniment data is key-on KON in chronological order.
Based on 11, key-off KOFF11, key-on KON12, key-off KOFF12, and key-on KON13, musical tone is generated by conventionally known MIDI data.

The upper part of FIG. 1 (B) shows the key-on / off data stored in the voice data, in the order of time, key-on KON1, key-off KOFF1, key-on KON.
2 occurs. The key-on KON1 is a key-on KON12 at a desired position of the accompaniment data, for example, when creating voice data based on the normal accompaniment data (FIG. 1A).
Is set to occur at the same timing as. The key-on / off data is generated based on the data DT of FIG. 6 (A). The interval between the key-on KON1 and the key-off KOFF1 corresponds to the gate time 54 in FIG. 6 (A).

The lower part of FIG. 1 (B) shows the phoneme string data 55 (for example, in the case of syllable "ka") shown in FIG. 6 (A). For example, phoneme CL (1 × 7.5 ms), phoneme kh
a (4 × 7.5 ms) and phoneme aj (infinite length) are generated.

Here, the sound generation timing will be further described. Phonemes include unvoiced consonants and voiced vowels. In the case of the syllable “ka”, “CL” and “kha” are consonants and “aj” is a vowel. When the consonant “CL” and the consonant “kha” are generated, the sound “ka” is not heard as a sound, and when the vowel “aj” is generated thereafter, it is heard as a sound at that time.

Phoneme C at the timing of key-on KON1
Generation of L starts, but after the time (5 × 7.5 ms) during which the consonant CL + kha is generated has passed, the vowel “a
When the pronunciation of "j" is started, "ka" is actually heard as a sound. On the other hand, in the accompaniment data (FIG. 1 (A)), a musical sound is produced at the timing of key-on KON12, and the voice data in the song data (FIG. 1 (B)) is slightly delayed from the timing of key-on KON1. "Ka"
You will be able to hear the pronunciation. Therefore, the key-on KON1 of the voice data is the accompaniment data (see FIG.
Although the timing is the same as that of the key-on KON12 of (A)), the pronunciation of "ka" becomes audible at a timing slightly later than the pronunciation of the accompaniment data. In particular, the total duration of the consonant sequence included at the beginning of the phoneme sequence data (5 × for “CL + kha”)
When 7.5 ms) is long, the delay time becomes large.
This delay time may make the listener feel uncomfortable. The duration of the consonant sequence differs depending on the syllable. Depending on the syllable, there is one in which a delay is hardly felt and one in which a big delay is felt. However, listeners who have no knowledge of phonemes cannot distinguish it.

In order to eliminate the inconvenience, a tone generation timing changing function is used. When the listener determines that a certain lyrics (syllable) is behind the accompaniment, the data is changed (edited) so as to automatically eliminate the delay only by designating the syllable.

FIG. 1C shows the audio data after editing the audio data of FIG. 1B. As shown in the upper part of FIG. 1C, the timings of key-on KON1, key-off KOFF1, and key-on KON2 are all advanced by 10 clocks. Specifically, the timing data 51 shown in FIG. 6A may be reduced by 10 clocks. This 10
The clock time is determined for each syllable by the conversion table shown in FIG.

FIG. 1D shows the ROM 4 or RAM of FIG.
3 shows an example of the conversion table stored in FIG. The conversion table stores the number of clocks to be advanced for each syllable.
For example, “ka” means 10 clocks. "Ah,
Each of the syllables “i, u, e, o” does not need to be included in the conversion table because it does not include the phoneme of the consonant and thus does not cause a delay in pronunciation. Since the duration of the consonant string is determined by the syllable, it is set in advance how fast the consonant sequence can be eliminated.

The lower part of FIG. 1C shows the vocalization timing of the edited phoneme string data. Phoneme string CL + kha + aj
Occurs at the timing of key-on KON1.
Phoneme string CL (1 x 7.5 ms) + kha (4 x 7.5 m)
The duration of (s) + aj (infinite length) remains unchanged.
The timing of key-on KON1 etc. is advanced as compared with FIG. 1 (B), and accordingly, the generation of phoneme string data is advanced.

As described above, by advancing the timing of key-on KON1 or the like, the delay in the pronunciation of "ka" can be reduced or eliminated. The actual sounding timing of "ka" can be brought close to the timing of the key-on KON12 of the accompaniment data. Also, since the operator only has to specify the lyrics (for example, “ka”) that feel the delay in pronunciation, pronunciation does not cause delay with a simple operation, without knowing how many clocks should be advanced. Can be done.

The processing for realizing the above-mentioned tone generation timing changing function will be described later with reference to the flowchart of FIG.

(3) Phoneme time shortening function FIG. 2A shows voice data for pronouncing "ka" as in FIG. 1B. The upper part of FIG. 2A shows key-on KON1, key-off KOFF1, and key-on KON2 in order of time. The lower part of FIG. 2A shows the phoneme string data C.
L (1 × 7.5 ms) + kha (4 × 7.5 ms) + a
Indicates j (infinite length).

FIG. 2B shows an example of pronunciation of voice data after increasing the tempo of song data. The voice data is
You can change the tempo. For example, in FIG. 6A, the tempo data in the initial setting data SD of the song data 50 may be changed. When the tempo is increased, key-on KON1 and key-off KO shown in the upper part of FIG.
The timing of FF1 and key-on KON2 will be earlier,
The duration of the phoneme string shown in the lower part of FIG. 2 (B) does not change. This is because the timing of key-on or the like is represented by the number of clocks (corresponding to the timing data 51 and the gate time 54 in FIG. 6A), and the duration 6 of the phoneme string data.
4 (FIG. 6 (B)) is represented in real time.
Changing the tempo is changing the generation timing time of one clock.

Phoneme string C at the timing of key-on KON1
Generation of L + kha + aj starts, but next key-on K
The generation of the phoneme string is stopped at the timing of ON2, and the generation of the next phoneme string data is started. As a result, the consonant CL is generated, and then the consonant kha is generated, and the generation of the consonant kha shifts to the pronunciation of the next phoneme string data in the middle, and the vowel aj is not generated at all.

As described above, the phoneme sequence is consonant CL + kh.
Since "a" is not heard as a sound while "a" is generated and as a sound when the vowel aj is generated, "ka" cannot be heard as a sound. In this way, if the tempo is increased, some lyrics (syllables) may not be heard.

To eliminate the inconvenience, a phoneme time shortening function is used. After changing the tempo to be faster, if the operator designates a lyrics (syllable) that is not pronounced, the operator automatically changes (edits) the data so that the lyrics are pronounced.

FIG. 2C shows the song data after editing the song data of FIG. 2B. As shown in the upper part of FIG. 2C, key-on KON1 and key-off KOFF
1. The timing of key-on KON2 is the same as in FIG. 2 (B). As shown in the lower part of FIG. 2 (C), the duration of the consonant kha in the phoneme string data CL + kha + aj is shortened by one unit time (7.5 ms). That is,
Duration of consonant kha from 4x7.5ms to 3x
Change to 7.5 ms. The changed phoneme sequence data is CL
(1 × 7.5 ms) + kha (3 × 7.5 ms) + aj
(Infinite length).

By shortening the duration of the consonant kha, the consonant CL is generated at the timing of the key-on KON1, the consonant kha is then generated, and then the vowel a is generated.
j occurs. The vowel aj is generated by the next key-on KON.
Although the operation shifts to the pronunciation of the next phoneme string data in the middle of 2, the vowel aj is generated, so that the “ka” is heard as a sound.

That is, by reducing the duration of the consonant, it is possible to hear the pronunciation of the voice that cannot be heard due to the change in tempo. However, when the duration is 1 × 7.5 ms, it becomes 0 ms when it is further reduced, and the consonant is not generated at all, and different syllables are generated. Therefore, the duration is 2 x 7.5 m
It is necessary to reduce the duration only for consonants of s or more.

The processing for realizing the above-mentioned phoneme time shortening function will be described later with reference to the flowchart of FIG.

FIGS. 3A to 3D are views for explaining exhalation data (hereinafter represented by “∇”) and long sound symbols (represented by “−”). As shown at the top of FIG. 3, key-on KON1, key-off KOFF1, and key-on KON2 occur in time order. The sound generation timing of each voice data at that time is shown in FIGS.

FIG. 3A shows the sounding timing of the voice data "ka", as in the above case. A consonant CL (1 × 7.5 ms) is generated at the timing of key-on KON1, a consonant kha (4 × 7.5 ms) is subsequently generated, and subsequently,
The vowel aj (infinite length) is generated. After that, key off KO
FF1 occurs, but key-off KOFF1 is ignored,
The vowel aj (infinite length) continues to be generated. After that, when the key-on KON2 occurs, the pronunciation changes from the generation of the vowel aj (infinite length) to the generation of the next phoneme sequence. Since there is no exhalation data after "ka", it is pronounced so that "ka" and the next syllable are smoothly connected.

FIG. 3B shows the tone generation timing of the song data “ka∇”. “∇” indicates exhalation data. Consonant CL (1 x 7.5 m at the timing of key-on KON1
s) is generated, followed by consonant kha (4 × 7.5 ms)
Occurs, followed by the vowel aj (infinite length). After that, when the key-off KOFF1 occurs, the generation of the vowel aj (infinite length) shifts to the silent state. That is, it becomes a state in which exhalation, that is, breathing is expressed. After that, key-on K
When ON2 occurs, the generation of the next syllable (phoneme string) is started.

FIG. 3C shows the sounding timing of the song data "Kai". When the key-on KON1 occurs, the consonant CL (1 × 7.5 m
s) is generated, followed by consonant kha (4 × 7.5 ms)
Occurs, followed by the vowel aj (infinite length). After that, when the key-off KOFF1 is generated, the vowel aj (infinite length) is generated and the sound "i" immediately after that is generated, so that the vowel ij (3 × 7.5 ms) is generated and 3 × 7.
It becomes silent after 5 ms. That is, the pronunciation of "i" ends. The duration of this vowel ij (3 × 7.
5 ms) is a predetermined time. afterwards,
When the key-on KON2 occurs, the generation of the next syllable (phoneme sequence) is started.

FIG. 3D shows the tone generation timing of the song data "kaka". When the key-on KON1 occurs, the consonant CL (1 × 7.5 m
s) is generated, followed by consonant kha (4 × 7.5 ms)
Occurs, followed by the vowel aj (infinite length). After that, when the key-off KOFF1 occurs, the consonant CL (1 × 7.5 ms) is generated because the vowel aj (infinite length) is generated and the sound “ka” immediately after that is generated, and the consonant kha ( 4 × 7.5 ms), and the vowel a
j (2 × 7.5 ms) is generated. After that, when the key-on KON2 is generated, the pronunciation shifts from the generation of the vowel aj (2 × 7.5 ms) to the generation of the next syllable (phoneme sequence).
The duration of this vowel aj (2 x 7.5 ms) is
It is a predetermined time.

As shown in FIG. 3 (C) and FIG. 3 (D), a phoneme other than the last phoneme in the phoneme sequence that has a phoneme of infinite duration includes a long sound symbol ("-"). It can be determined that it is a phoneme string.

FIG. 7 shows the CP of the personal computer 1.
It is a flowchart which shows the process of the main routine which U5 performs. In step SA1, initial setting processing is performed.
The initial setting includes initialization of registers and flags, for example.

At Step SA2, it is checked whether or not the reproduction start of the song data is instructed. For example,
The operator can instruct the reproduction start by clicking the reproduction start switch displayed on the display 45 using the mouse 10. When the reproduction start is instructed, the process proceeds to step SA3, and when the reproduction start is not instructed, the process proceeds to step SA5.

At step SA3, it is checked whether or not song data is selected. If song data is selected, the process proceeds to step SA4, and if song data is not selected, the process proceeds to step SA5.

At Step SA4, the selected song data is reproduced. Details of the reproduction process will be described later with reference to FIG. Then, it progresses to step SA5.

At Step SA5, it is checked whether or not the reproduction stop is instructed. For example, the operator can instruct the reproduction stop by clicking the reproduction stop switch displayed on the display 45 with the mouse 10. When the reproduction stop is instructed, the process proceeds to step SA6, and when the reproduction stop is not instructed, the process proceeds to step SA7.

At step SA6, the flag run is set to 0. When the flag run is 1, it indicates that the song data is currently being reproduced, and when the flag run is 0, it indicates that the song data is not currently being reproduced. Then, it progresses to step SA7.

At Step SA7, it is checked whether or not the tempo change is instructed. For example, the operator can use the mouse 10 or the keyboard 12 to instruct to change the tempo value displayed on the display 45. When the tempo change is instructed, the process proceeds to step SA8, and when the tempo change is not instructed, the step SA is performed.
Proceed to 9.

At step SA8, the generated clock timing is changed based on the changed tempo. That is,
This corresponds to changing the generation time interval of one clock.
Then, it progresses to step SA9.

At Step SA9, it is checked whether or not the start of the data correction mode process is instructed. For example,
The operator can use the mouse 10 to instruct the data correction mode displayed on the display 45.
When the mode is instructed, the process proceeds to step SA10, and when the mode is not instructed, the process proceeds to step SA11.

At step SA10, a data correction mode process is performed. In the data correction mode, the processing of the syllable editing function, the pronunciation timing editing function (FIG. 1), and the phoneme time shortening function (FIG. 2) described above is performed. The details of the process are
It will be described later with reference to FIG. Then, it progresses to step SA11.

At step SA11, other processing is performed. The other process is, for example, a process of reading the song data from the external storage device 13 or the like into the RAM 3. afterwards,
Go to step SA12.

At Step SA12, it is checked whether or not the end is instructed. For example, the operator can instruct the end by clicking the end displayed on the display 45 using the mouse 10. If the end is not instructed, the process returns to step SA2 and the above process is repeated. When the termination is instructed, the processing of the main routine is terminated.

FIG. 8 is a flow chart showing details of the reproducing process shown in step SA4 of FIG.

At step SB1, all data up to the start timing data of the selected song data is transmitted. That is, of the song data 50 of FIG. 6 (A),
The initial setting data SD at the head is transmitted to the external device 21.

At step SB2, the flag run is set to 1. By setting the flag run to 1, it is recorded that the reproduction is currently being performed.

At step SB3, 0 is set in the register time. The register time is a register for storing the elapsed time after the start of reproduction is instructed. The register time is incremented in the interrupt processing that is performed at predetermined time intervals thereafter. The interrupt process will be described later with reference to the flowchart of FIG.

At step SB4, the above-mentioned start timing data is set in the register timing. The register timing is a register for storing timing data.

In step SB5, the data read position (read pointer) is set to the data (note number 52 in FIG. 6A) next to the leading timing data.

Thereafter, in the interrupt processing shown in FIG. 9, the reading and transmitting processing for the note number 52 and below is performed. With the above, the reproduction process ends, and the process returns to the process of the main routine of FIG. 7.

FIG. 9 shows the CP of the personal computer 1.
It is a flowchart which shows the interruption processing which U5 performs.
The interrupt process is performed every one clock representing the timing data. One clock is, for example, 1/48 of a quarter note
It has a cycle of zero. The personal computer 1 operates with a clock having a higher frequency than that.

At step SC1, it is checked whether the flag run is 1 or not. When the flag run is 1, it means that the song data is currently being reproduced, and the process proceeds to step SC2. When the flag run is 0, it means that the song data is not currently being reproduced, and step SC1
Proceed to 7.

At step SC2, it is checked whether or not the value of the register time is the same as the value of the register timing. That is, it is checked whether or not the time register time is the data sounding timing at which the value of the head timing data is reached. When it reaches, it proceeds to step SC3, and when it does not reach, it proceeds to step SC14.

At step SC3, one data is read from the song data. Step SB5 of the reproduction process of FIG.
Since the read pointer is set to the note number 52 in FIG. 6A, the note number 52 is read.

At step SC4, it is checked whether the read data is the timing data 51 (FIG. 6A). Since the read data is the note number, it is determined that it is not the timing data, and the process proceeds to step SC5 according to the NO arrow.

At step SC5, it is checked whether the read data is the phoneme string data 55 (FIG. 6 (A)). Since the read data is the note number, NO
Follow the arrow to move to step SC9.

At step SC9, it is checked whether the read data has the gate time 54 (FIG. 6A). Since the read data is the note number, NO
Follow the arrow to move to step SC11.

At step SC11, it is checked whether the read data is the end data ED (FIG. 6 (A)). Since the read data is the note number, NO
Follow the arrow to move to step SC12.

At step SC12, the read note number is stored in the data holding area in the RAM 3. Then, it returns to step SC3.

At step SC3, the next data in the song data is read. That is, the velocity 53 (see FIG.
(A)) is read. Thereafter, the same step processing as in the case of the note number 52 is performed, and step SC12
And the velocity 53 is stored in the data holding area. A note number and velocity are stored in the data holding area. Then, it returns to step SC3.

At step SC3, the next data is read. That is, the gate time 54 (FIG. 6A) is read. Then, the process proceeds to step SC9 via steps SC4 and SC5. In step SC9, the process proceeds to step SC10 following the YES arrow.

At step SC10, the value of the read gate time 54 is set in the register gate. The register gate is a register for holding the value of the gate time. Then, it returns to step SC3.

At step SC3, the next data is read. That is, the phoneme string data 55 (FIG. 6A) is read. Then, through step SC4, step SC
Go to 5. In step SC5, the process proceeds to step SC6 following the YES arrow.

At step SC6, the read phoneme string data 55 is transmitted to the external device 21 (FIG. 5). Then, in step SC7, the key-on data and the note number 52 and velocity 53 stored in the data holding area are transmitted to the external device 21.

The song data in the RAM (see FIG.
(A) is arranged in the order of note number 52, velocity 53, gate number 54, and phoneme string data 55. However, when transmitting to the external device 21, due to the nature of the external device 21, first the phoneme string data is transmitted. 55 is sent (SC
6) After that, it is necessary to transmit the key-on event including the note number 52 and the velocity 53 (SC7).

Next, in step SC8, the transmitted note number 52 and velocity 53 are erased from the data holding area. Then, it returns to step SC3.

At step SC3, the next data is read. That is, the timing data 51 (FIG. 6A) is read. Then, in step SC4, the process proceeds to step SC13 by following the YES arrow. At step SC13, the read timing data is stored in the register timin.
Set to g. Then, it progresses to step SC17.

At step SC17, the conventionally-known automatic performance process for other parts is performed. For example, the automatic performance data of the accompaniment part can be transmitted to the external device 21. Then, in step SC18, the value of the time register time is incremented. That is,
The time register time will be incremented at predetermined time intervals. After that, interrupt processing is completed,
Return to the process before the interrupt.

When the next interrupt is made, step SC
1 through step SC2, register tim
It is checked whether the value of e has reached the value of the timing register timing. If not reached, the process proceeds to step SC14.

At step SC14, the time register ti
It is checked whether the value of me has reached the value of (gate time register gate + timing register timing). If not reached, the process proceeds to step SC17 following the NO arrow, and the above-described processing is repeated.
When it reaches, it means that the key-off timing has been reached, and the process proceeds to step SC15 following the YES arrow.

At step SC15, the key-off data is transmitted to the external device 21. Then, step SC1
7 and repeats the above processing.

When the end data ED (FIG. 6 (A)) is read in step SC3 by repeating the above interrupt processing, it means the end of the song data, and through step SC4, SC5, SC9, step SC1
Go to 1. In step SC11, the process proceeds to step SC16 following the YES arrow.

At step SC16, the flag run is set to 0.
And record that song data is not currently being played back. Then, it progresses to step SC17 and performs the process shown above.

When the next interrupt is made, step SC
In 1, it is determined that the flag run is 0, and the process immediately proceeds to step SC17 to perform the above-described processing.

FIG. 10 shows the CPU 26 of the external device 21.
6 is a flowchart showing a MIDI receiving process performed by the.
In the interrupt processing of FIG. 9, the personal computer 1
Sends the above data via MIDI. External device 2
1 receives the data, the external device 2
1 performs the following processing.

At step SD1, it is checked whether any one of the phoneme string data, key-on data, note number, velocity and key-off data, that is, voice data is received. When it is received, the process proceeds to step SD2 according to the YES arrow, and when it is not received, the process of step SD2 is bypassed to proceed to step SD3.

At step SD2, the data is transmitted to the voice synthesis circuit 32 based on the received data. When the voice synthesis circuit 32 is a formant sound source, the formant center frequency, the formant band width, etc. are transmitted to the formant sound source based on the received phoneme string data and the like. Then, it progresses to step SD3.

At step SD3, other received data processing is performed. For example, when the data of the accompaniment part is received, the data is transmitted to the tone generator circuit 29 or the effect circuit 30. After that, the MIDI reception process ends.

FIG. 11 is a flow chart showing details of the data correction mode process shown in step SA10 of FIG.

At step SE1, it is checked whether the voice data edit screen (FIG. 4) is displayed. If not displayed, follow the NO arrow to step SE
2. If it is displayed, follow the YES arrow to proceed to step SE4 when displayed.

At step SE2, selection of song data (FIG. 6 (A)) desired to be corrected is accepted.
For example, the operator can select song data by operating the mouse.

Next, in step SE3, an audio data edit screen is displayed based on the selected song data.
Then, it progresses to step SE4.

At step SE4, it is checked whether or not the selection of the editing range is designated. For example, the operator can select a range of one line or a plurality of lines on the edit screen shown in FIG. One line is data of one note and lyrics corresponding to it.

When the edit range is selected, the range selected in step SE5 is read according to the YES arrow.
It is stored in AM3 and the process proceeds to step SE6. When no editing range is selected, step SE5 is bypassed and the process proceeds to step SE6.

At step SE6, it is checked whether or not the tone generation timing editing process (FIG. 1) is instructed.
The operator can give the instruction using the mouse. When instructed, the sounding timing process is performed in step SE7 according to the YES arrow, and then the process proceeds to step SE8. When not instructed, N
By following the arrow of O, step SE7 is bypassed and the process proceeds to step SE8. Details of the pronunciation timing processing will be described later in FIG.
It will be described with reference to the flowchart of FIG.

At step SE8, it is checked whether or not the phoneme time reduction process (FIG. 2) is instructed. The operator can give the instruction using the mouse. If instructed, follow the YES arrow to step S
The phoneme time shortening process is performed at E9, and then the process proceeds to step SE10. If not instructed, follow the NO arrow to bypass step SE9 and skip step SE1.
Go to 0. The details of the phoneme time reduction process will be described later with reference to the flowchart of FIG.

In step SE10, it is checked whether or not syllable editing processing is instructed. The operator can give the instruction using the mouse. If it is instructed, follow the YES arrow to proceed to step SE11, and if not, follow the NO arrow to bypass steps SE11 and SE12 and proceed to step SE1.
Go to 3.

At step SE11, it is checked whether or not the selected range is one musical tone pronunciation (one line of the edit screen of FIG. 4). YES for one tone sound
The syllable editing process is performed in step SE12 according to the arrow, and then the process proceeds to step SE13. If it is not one musical tone pronunciation, immediately follow the NO arrow and proceed to step SE1.
Go to 3. Details of the syllable editing process will be described later with reference to the flowchart of FIG.

At step SE13, other data correction processing is performed. Next, in step SE14, it is checked whether or not the end of the data edit mode has been instructed. The operator can give the instruction using the mouse.
If not instructed, the process returns to step SE4 following the NO arrow and repeats the above processing. When instructed, the YES arrow is followed to end the data correction mode processing and return to the processing of the main routine of FIG.

FIG. 12 is a flow chart showing details of the tone generation timing changing process (FIG. 1) shown in step SE7 of FIG.

In step SF1, the data read position (read pointer) is set at the beginning of the selected range. Specifically, in the song data of FIG. 6A, the read pointer is set at the address position of the timing data 51. One line of the edit screen (FIG. 4) corresponds to one data DT of FIG. 6 (A). Then, it progresses to step SF2.

At step SF2, one data is read. That is, the timing data 51 is read.

Next, in step SF3, since the read data is the timing data, the flow advances to step SF4 according to the YES arrow.

At step SF4, it is checked whether or not there is phoneme string data which is sounded at the timing indicated by the timing data. Specifically, the phoneme string data 55 is inserted between the timing data and the next timing data.
(FIG. 6 (A)) is checked. If there is phoneme string data, follow the YES arrow to step SF
If there is no phoneme string data, the process proceeds to step SF7 immediately following the NO arrow.

At step SF5, the conversion table (see FIG.
(D)), the delay clock of the leading phonetic character in the phoneme string data is extracted. For example, when the lyrics corresponding to the phoneme string data is "kai", the delay clock of the first phonetic character "ka" is extracted. Then, with reference to the conversion table (FIG. 1D), 10 clocks are extracted as a delay clock corresponding to "ka".

At step SF6, the read timing data is subtracted by the delay clock number, and the song data is rewritten using the value as new timing data. Then, it progresses to step SF7.

At step SF7, it is checked whether the data read out above is the final data within the selected range. When only one row is selected, since it is the final data, the YES timing arrow is followed to end the tone generation timing changing processing and return to the data correction mode processing of FIG.

When a plurality of rows are selected, since the data is not the final data, the process returns to step SF2 following the NO arrow in step SF7. In step SF2, the next data is read. For example, the note number 52 located next to the timing data 51 is read. Next, in step SF3, since the read data is the note number and not the timing data, the process returns to step SF2 via step SF7. In step SF2, the next data is read. Thereafter, similar processing is performed, and the processing is repeated until the next timing data is read. When the timing data is read, the process proceeds to step SF4 following the YES arrow in step SF3, and the above-described processing is repeated. This allows the timing data for multiple rows to be modified. When the process of the last line of the selected lines is completed, the sound generation timing changing process is ended according to the YES arrow in step SF7.

As described above, by subtracting the predetermined number of clocks from the timing data, as shown in FIG. 1, even if the duration of the consonant string at the beginning is long, it is possible to eliminate or reduce the delay of the sounding timing. it can.

FIG. 13 is a flow chart showing details of the phoneme time shortening process (FIG. 2) shown in step SE9 of FIG.

In step SG1, the data read position (read pointer) is set at the beginning of the selected range. That is, the read pointer is set at the position of the address of the timing data 51 in the first row. afterwards,
Go to step SG2.

At step SG2, one data is read. That is, the timing data 51 is read according to the read pointer.

Next, in step SG3, it is determined whether the read data is phoneme string data. This time,
Since the read data is timing data, it is determined that the data is not phoneme string data, and the process proceeds to step SG6. In step SG6, it is determined whether the read data is the final data in the selected range. Since the read timing data is not the final data this time, the process returns to step SG2 following the NO arrow. Hereinafter, the above processing is repeated until the phoneme string data is read. When the phoneme string data is read out, the process proceeds to step SG4 following the YES arrow in step SG3.

In step SG4, it is checked whether or not there is a duration of "2" or more among the durations in the phoneme string data, which is for the phoneme of the consonant. When the duration is present, in order to shorten the phoneme time, the process proceeds to step SG5 following the YES arrow. If the duration is not present, the phoneme time cannot be shortened, so the process proceeds to step SG6 following the NO arrow.

At step SG5, the duration is subtracted by 1 × 7.5 ms. That is, the duration "2" or more for the phoneme of the consonant is subtracted by "1". Then, it progresses to step SG6.

In step SG6, it is checked whether the data read out above is the final data within the selected range. If the selected line is only one line, the phoneme time reduction process is ended following the YES arrow and the process returns to the data correction mode process of FIG.

If the selected line is a plurality of lines, the process returns to step SG2 following the NO arrow. Hereinafter, the same process is performed for the next line. When the process of the last line is completed, the phoneme time reduction process is completed according to the YES arrow in step SG6.

As described above, by shortening the duration of the consonant, as shown in FIG. 2, it is possible to correct the inconvenience that the lyrics are not pronounced by increasing the tempo of the song data, and the lyrics are properly pronounced. be able to.

In the above flow chart, a case has been described in which the sound generation time is shortened by "1 × 7.5 ms" in one process, but the present invention is not limited to this. The time may be shortened by 2 × 7.5 ms or more, or the shortening time may be changed according to the syllable. The unit time of duration is 7.5 ms.
Not limited to

FIG. 14 is a flow chart showing details of the syllable editing process shown in step SE12 of FIG.

In step SH1, the syllable editing screen is displayed based on the phoneme string data included in the selected row.
FIG. 15 shows an example of the syllable editing screen. In the syllable editing screen, phoneme string data of the syllable "ka" is shown. The syllable “ka” is composed of phonemes “CL”, “kha”, and “aj”. The phoneme "CL" has a duration of "1 x 7.5 m.
s ”, and the phoneme“ kha ”has a duration of“ 4 ×
The phoneme “aj” has a duration of “infinite length (∞)”.

Next, in step SH2, it is checked whether or not there is a data change input, and the process waits until that input. The operator can specify a phoneme using a mouse or a keyboard and input a duration value. If there is an input, follow the YES arrow to step SH3.
Go to.

At step SH3, the duration value is rewritten based on the change input. Specifically, the value of the corresponding duration 64 in the phoneme string data 55 (FIG. 6 (B)) is rewritten. For example, the duration of the phoneme “kha” is changed from “4 × 7.5 ms” to “3 × 7.5 m”.
s ".

Next, in step SH4, it is checked whether or not an instruction to end the voice editing process has been given. This instruction can be given by the operator using the mouse. If the end is not instructed, follow the NO arrow to step SH2.
Return to and repeat the above process. When the end is instructed,
According to the YES arrow, the syllable editing process ends, and the process shown in FIG.
Return to the data correction mode processing of.

As described above, since the operator can set the duration of each phoneme to an arbitrary value, various sounds can be produced even with the same syllable.

Other syllable editing processing will be described. Figure 15
In, when the phoneme switch 72 is clicked with the mouse, a list of phonemes is displayed below it. Then, when the insertion switch 73 is clicked, the phoneme string “CL + kha +
A desired phoneme can be inserted into "aj". By inserting a phoneme, it is possible to generate a phoneme sequence that expresses, for example, a "ka" that looks like a nose. By clicking the delete switch 74, a desired phoneme can be deleted from the phoneme string.

By clicking the syllable switch 71, syllables can be inserted or deleted.

When the pre-change switch 75 is clicked, the phoneme string before editing can be sounded, and when the post-change switch 76 is clicked, the phoneme string after editing can be sounded. The operator can edit or create various voices while checking the actual voice.

The data editing apparatus according to this embodiment is not limited to the form of a personal computer and application software, but may be of an electronic musical instrument or a sequencer. The application software may be stored in a storage medium such as a magnetic disk, an optical disk, or a semiconductor memory and supplied to a personal computer, or may be supplied via a network.

The format of the song data is "event + relative time" in which the time of occurrence of a performance event like a standard MIDI file is represented by the time from the previous event, and the time of occurrence of a performance event is the song or bar. "Event + Absolute time" expressed in absolute time, "Pitch (rest) + note length" that represents performance data in terms of note pitch and note length or rest and rest length, minimum resolution of performance A "solid method" in which a memory area is secured for each and the performance event is stored in the memory area corresponding to the time when the performance event occurs.
The format such as

The song data may have a format in which the data of a plurality of channels are mixed, or may have a format in which the data of each channel is separated for each track.

The present invention has been described above with reference to the embodiments.
The present invention is not limited to these. For example, it will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

[0173]

As described above, according to the present invention,
The operator can specify any one of the phoneme data that constitutes the syllable data and change the occurrence time data corresponding to the phoneme data to a desired value, so that various syllable data is generated. be able to.

By designating the voice data, the operator can change the tone generation timing data of the voice data by a desired value according to the table, so that the delay of the tone generation timing of the voice data can be reduced or It can be lost.

By designating the syllable data, the operator can shorten and change the duration data of the phoneme data included in the syllable data.
Even if the duration data of the phoneme data belonging to the consonant is long, it is possible to prevent the non-pronouncing.

[Brief description of drawings]

FIG. 1 is a diagram for explaining sound generation timing changing processing performed by a data editing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram for explaining a phoneme time reduction process performed by the data editing apparatus according to the present embodiment.

FIG. 3 is a diagram showing a timing of sound generation.

FIG. 4 is a diagram showing an example of an audio data editing screen displayed on the data editing apparatus according to the present embodiment.

FIG. 5 is a diagram showing a hardware configuration of a system in which a personal computer and an external device are connected.

FIG. 6A is a diagram showing an example of song data, and FIG. 6B is a diagram showing an example of phoneme string data.

FIG. 7 is a flowchart showing a process of a main routine performed by a CPU of a personal computer.

FIG. 8 is a flowchart showing details of a reproduction process shown in step SA4 of FIG.

FIG. 9 is a flowchart showing an interrupt process performed by a CPU of a personal computer.

FIG. 10 is a flowchart showing a reception process performed by a CPU of an external device.

11 is a flowchart showing details of the data correction mode processing shown in step SA10 of FIG.

FIG. 12 is a flow chart showing details of sound generation timing changing processing shown in step SE7 of FIG.

FIG. 13 is a flowchart showing details of the phoneme time reduction process shown in step SE9 of FIG.

FIG. 14 is a flowchart showing details of the syllable editing process shown in step SE12 of FIG.

FIG. 15 is a diagram showing an example of a syllable editing screen displayed on the data editing device according to the present embodiment.

[Explanation of symbols]

1 personal computer, 2 buses, 3
RAM, 4 ROM, 5 CPU, 6 timer, 7 display circuit, 8 interface,
9, 11 Detection circuit, 10 mouse, 12
Keyboard, 13 external storage device, 21 external device, 22 bus, 23 interface,
24 RAM, 25 ROM, 26 CP
U, 27 timer, 28 display circuit, 29
Sound source circuit, 30 effect circuit, 31 sound system, 32 voice synthesis circuit, 33 detection circuit, 34 manipulator, 41 communication network,
42 server computer, 45 display, 50 song data, 51 timing data, 52 note number, 53 velocity, 54 gate time, 55 phoneme sequence data,
ED End Data, 61 Part Number, 6
2 device number, 63 phoneme number, 64
Duration, 65 exhalation data, 71 syllable switch, 72 phoneme switch, 73 insertion switch, 74 deletion switch, 75 before change switch, 76 after change switch

─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-5-341793 (JP, A) JP-A-8-248993 (JP, A) JP-A-4-170600 (JP, A) JP-A-4- 162000 (JP, A) JP 53-116705 (JP, A) JP 55-130596 (JP, A) JP 8-234771 (JP, A) JP 2001-209381 (JP, A) (JP 58) Fields investigated (Int.Cl. ⁷ , DB name) G10L 13/00 G10L 13/06 G10H 1/00 101 G10H 1/00 102

Claims (8)

(57) [Claims] 1. Audio data including syllable data composed of a plurality of phoneme data whose generation time is defined by duration data that is independent of reproduction tempo, and timing dependent on reproduction tempo that defines sounding timing of the audio data. A data editing device capable of editing the syllable data in song data including data, comprising: a phoneme designating unit for designating one of the plurality of phoneme data constituting the syllable data; and the phoneme designating unit. A data editing apparatus having a changing unit that changes duration data corresponding to phoneme data designated by the means to an occurrence time designated independently of the playback tempo designation. 2. Audio data including syllable data composed of a plurality of phoneme data whose generation time is defined by duration data which is independent of reproduction tempo, and timing dependent on reproduction tempo which defines sounding timing of the audio data. A data editing device capable of editing song data including data, each syllabic data determined according to characteristics of a plurality of phoneme data forming the syllable data, and a sounding timing change value of the syllable data. Table for storing the correspondence relation of the sound data, a designating unit for designating the sound data, and a sounding timing change value for the syllable data located at the head among the syllable data forming the sound data designated by the designating unit. The detection means for detecting from the Data editing apparatus and a changing means for changing the serial timing data by tone generation timing change value detected by said detection means. 3. Audio data including syllable data composed of a plurality of phoneme data whose occurrence time is defined by duration data which is independent of reproduction tempo, and timing dependent on reproduction tempo which defines sounding timing of the audio data. A data editing device capable of editing the syllable data in song data including data, the designating means for designating the syllable data, and the plurality of phoneme data included in the syllable data designated by the designating means. A data editing apparatus comprising: a detecting unit that detects phoneme data belonging to a consonant; and a shortening unit that shortens and changes duration data of the phoneme data detected by the detecting unit. 4. The detecting means is means for detecting phoneme data belonging to a consonant and having corresponding duration data equal to or larger than a predetermined value, and the shortening means shortens the duration data by a predetermined value. The data editing device according to claim 3, which is a means for changing the data. 5. A timing dependent on the reproduction tempo that defines sound data including syllable data composed of a plurality of phoneme data whose generation time is defined by duration data that is independent of the reproduction tempo and a sounding timing of the sound data. A medium for recording a computer program capable of editing the syllable data in song data including data, and (a) a step of specifying one of the plurality of phoneme data constituting the syllable data. , (B) a step of specifying the occurrence time after the change of the specified phoneme data independently of the designation of the reproduction tempo, and (c) a step of setting the duration data corresponding to the specified phoneme data in the step ( A medium storing a program for causing a computer to execute the procedure for changing to the occurrence time specified in b). body. 6. Audio data including syllable data composed of a plurality of phoneme data whose generation time is defined by duration data which is independent of reproduction tempo, and timing dependent on the reproduction tempo which defines sounding timing of the audio data. A medium for recording a computer program capable of editing song data including data, comprising: (a) a procedure for specifying the voice data; and (b) characteristics of a plurality of phoneme data constituting the syllable data. From the table storing the correspondence relationship between each syllable data determined according to the syllable data and the sounding timing change value of the syllable data, a predetermined one corresponding to the syllable data located at the head among the syllable data forming the specified voice data is stored. And (c) the voice data identified in step (a) above. A medium for recording a program for causing a computer to execute the procedure of changing the timing data according to the sounding timing change value acquired in step (b). 7. Timing dependent on a reproduction tempo that defines sound data including syllable data composed of a plurality of phoneme data whose generation time is defined by duration data that is independent of the reproduction tempo, and sound generation timing of the sound data. A medium for recording a computer program capable of editing the syllable data in song data including data; (a) a procedure for identifying syllable data; and (b) included in the identified syllable data. A medium recording a program for causing a computer to execute a procedure of detecting phoneme data belonging to a consonant among a plurality of phoneme data and (c) a procedure of shortening and changing duration data of the detected phoneme data. 8. The step (b) is a step of detecting phoneme data which is phoneme data belonging to a consonant and the corresponding duration data of which is a predetermined value or more, and the step (c) is a step of detecting the duration data by a predetermined value. The medium on which the program according to claim 7 is recorded, which is a procedure for performing a shortening conversion by a value.