JPH04331990A - Voice electronic musical instrument - Google Patents

Abstract

PURPOSE:To accurately reflect the musical representation of a score and the correspondence relation between notes and a text on a voice even when the notes indicated on the score and the text corresponding to the notes are stored as data as they are. CONSTITUTION:When voice data on one voice are stored corresponding to plural successive notes (YES in step 118), the current pitch is bent (121) according to following pitch data while the pitch of the note corresponding to voice data following the current voice data is read out (step 120) to perform voice generation up to a current loop part. Namely, when the current pitch is E4 and following data is C4, the pitch is bent from the E4 to C4, and consequently a voice is generated in voicing mode wherein slurs are given to E4 and C4.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio electronic musical instrument that allows lyrics to be uttered using the pitch and duration of the notes that make up a piece of music.

0002

BACKGROUND OF THE INVENTION Conventionally, audio electronic musical instruments have been proposed that not only generate musical instrument sounds but also synthesize voices (for example, Japanese Patent Laid-Open No. 59-131939 (see official bulletin). This audio electronic musical instrument can selectively set a storage mode and a playback mode, and in the storage mode, by key operation, the pitch and length of the notes that make up the music are recorded. Lyrics consisting of voice data of 50 sounds such as "i", "u", "e", "o", etc. can be stored in the RAM. Then, after storing the pitch and duration along with the audio data in the RAM, when the playback mode is selected and set, the pitch and duration corresponding to the audio data are read out from the RAM, and the audio data is read out from the RAM. An audio signal is generated based on the pitch and duration. Therefore, by reproducing and emitting the audio signal through a speaker, the lyrics of a specific song can be generated from the speaker in accordance with the pitch and duration of the song.

0003

[Problem to be Solved by the Invention] However, in a song, for example, as in the lyrics "Gee" in the 6th measure of the song title "Sakura Sakura" shown in Figure 2, the pitch is E4.
In some cases, a slur is added over an eighth note of pitch C4, and a single sound, "Gee", is pronounced over two notes of different pitches. In such a case, if you simply generate sounds corresponding to the pitch and length of the notes, the sound of "gi" will be generated with a pitch of E4 and an eighth note length, and then the vowel "i" will be produced with a pitch of E4 and an eighth note length.
will occur with a pitch of C4 and an eighth note length. Therefore, a slur exists between pitches E4 and C4, and pitch E4
Even though the pitch should be lowered continuously from C4 to "gi", the sound "gi" at pitch E4 and the sound "i" at pitch C4 are uttered discontinuously. Therefore, the slur is pronounced in a manner that ignores its existence.

[0004] Also, depending on the song, for example, the lyrics ``Kakkou'' in the first measure of the song title ``Kakko'' shown in Figure 4 may be
As in, if the pitch C5 of a quarter note is assigned the two sounds "ka", and the pitch A4 of a quarter note is assigned the two sounds "kou", in other words, there are multiple sounds for one note. may be assigned. Therefore, in such a case, it is not possible to reproduce the lyrics using simple audio processing that generates a single voice corresponding to the pitch and duration of the notes as in the past. Of course, I've written the lyrics "kakkou" in advance, with eighth note length and pitch C5, "ka" and "tsu".
If you divide and store the audio data into 4 pieces of ``ko'' and ``u'' that are eighth note length and pitch A4, you can create a single voice corresponding to the pitch and length of the note. Even if it occurs, it's "cool"
This makes it possible to generate a sound. However, dividing the musical score into note lengths shorter than the notes shown on the musical score and storing them in correspondence with the lyrics has the disadvantage that the operation at the time of storage becomes complicated.

The present invention has been made in view of such conventional problems, and even when the notes shown on a musical score and the lyrics corresponding to the notes are stored as data, The object of the present invention is to provide an audio electronic musical instrument that allows the musical expression of a musical score and the correspondence between notes and lyrics to be accurately reflected in audio.

[0006]

[Means for Solving the Problems] In order to solve the above problems, the invention set forth in claim 1 provides information on the pitch and length of a plurality of notes constituting a music piece and the lyrics corresponding to each note. a storage means for storing audio data; a reading means for reading out the pitch and length of each note and the audio data from the storage means; 1. An audio electronic musical instrument comprising: an audio signal generating means for generating an audio signal based on the
When the audio data of audio is stored corresponding to a plurality of consecutive notes, the pitch of the note corresponding to the next audio data following the current audio data is read out, and the pitch of the note corresponding to the audio data at the current time is read out. Pitch variable means is provided for continuously changing the pitch of the audio signal from the pitch of the note corresponding to the next audio data.

The invention according to claim 2 also includes a storage means for storing the pitches and lengths of a plurality of notes constituting a music piece and voice data of lyrics corresponding to each note; reading means for reading the pitch and length of each note and the audio data; and audio signal generating means for generating an audio signal based on the pitch, length and audio data read by the reading means. In the audio electronic musical instrument, when audio data consisting of a plurality of sounds is stored corresponding to the pitch and duration of one note, a calculation means for calculating the duration of each sound is provided, The signal generating means is configured to generate the audio signal based on the tone length, the pitch, and the audio data calculated by the calculating means.

[0008]

[Operation] In the invention according to claim 1, when the audio data of one voice is stored corresponding to a plurality of consecutive notes, for example, the audio data "Gee" is pitch E4 and is an eighth note. If the sound data "gi" and the continuous pitch C4 and the long eighth note "-" are stored, when generating the sound signal, only the pitch E4 of the current sound data is stored. First, the pitch C4 of the audio data that follows is read out. Then, the audio signal based on the audio data is continuously changed from E4 to C4 by the pitch variable means. Therefore, when this audio signal is reproduced, the sound "Gee" is uttered in a slurred sound form that continuously changes pitch from E4 to C4.

Further, in the invention according to claim 2, when a plurality of audio data are stored corresponding to the pitch and duration of one note, for example, the audio data of the lyric "Kakkou""
If "Ko" has a pitch of a quarter note and is stored with a pitch of A4, the calculating means calculates the pitch of each voice. Then, the audio signal generating means generates an audio signal based on the tone length calculated by the calculating means, the pitch, and the audio data. Therefore, when this audio signal is reproduced, the sounds "ko" and "u" are pronounced with the calculated tone length.

0010

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. That is, FIG. 1 is a block diagram showing the overall configuration of this embodiment, and the CPU 1 executes processes such as reading data input on the keyboard 2 and storing it in the sequence memory 3 in a sequence data creation mode to be described later. At the same time, in the sequence execution mode, the sound generator 4 is controlled in accordance with the data stored in the sequence memory 3. The keyboard 2 includes a mode setting key for selectively setting the sequence data creation mode and sequence execution mode, a tempo input key for inputting music tempo data, and further includes 50 syllables, long sounds, and consonant sounds. A voice code input key for inputting the code, a pitch input key consisting of multiple alphabet keys and numeric keys, a pitch input key consisting of multiple numeric keys and a rest key (*), and an operation when input is completed. An input end key and the like are provided.

The sequence memory 3 is provided with a storage area having the format shown in FIGS. 3 and 5. The storage area is composed of a header 12, a lyric data storage area 6 and a melody data storage area 9 provided for each address, and the melody data storage area 9 is composed of a pitch data storage area 7 and a pitch data storage area 7. It consists of a data storage area 8. In the header 12, a message indicating the number of beats per minute is displayed according to the operation of the tempo input key.
Tempo data such as ``72'' and ``138'' are stored in the audio data storage area 6, and audio codes consisting of a single voice such as ``sa'', ``ku'', ``ra'', or ``ka'' are stored in the audio data storage area 6. ,
A voice code consisting of a plurality of voices, such as ``ko'', is stored for each address. Further, the pitch data storage area 7 stores data indicating the pitch of notes such as A4 and B4, and the note length data storage area 8 stores the note length with the sixteenth note length as "1". In other words, "4" is stored for a quarter note, "8" is stored for a half note, and "2" is stored for an eighth note.

On the other hand, the audio data memory 10 stores PCM data such as 50 sounds, consonants, long sounds, etc. of generated sounds generated at a constant reference pitch, and the PCM data is stored in the attack section, loop section, and release section. It is divided so that the parts necessary for controlling the sound generator 4 can be read out as appropriate. Further, a plurality of different PCM data are stored for different sound ranges for the same voice (multi-sampling), so as to eliminate formant deviation.

[0013] The sound generator 4 is connected to the CPU 1
reads out the PCM data of the voice specified by from the voice data memory 10, performs signal processing for pitch conversion so that it becomes the specified pitch, and also performs the total generation time of the attack section, loop section, and release section. executes tone length processing, etc., which repeats generation of the loop portion so as to match the value of the tone length data stored in the tone length data storage area 8. The generation time is measured based on the clock pulse from the tempo clock 11, and the pulse output operation of the tempo clock 11 is controlled by the CPU 1 according to the value of the tempo data stored in the header 12.

[0014] Furthermore, the audio signal which has been pitch-converted and tone length-processed by the sound generator 4 is inputted to the speaker 14 via the sound generation section 13, and is
4 and output the sound to the outside.

Next, the operation of this embodiment according to the above configuration will be explained with reference to a series of flowcharts shown in FIGS. 6, 7, and 8. This flowchart is executed by turning on the power (not shown), and first, based on the operating state of the mode setting key provided on the keyboard 2,
Mode determination is performed (step 101). In this mode determination, if the sequence data creation mode is set, the tempo data input by the user of this audio electronic musical instrument by operating the tempo input key is stored in the header of the sequence memory 3. 1
2 (step 102). That is, when storing the song "Sakura Sakura" shown in FIG. 2, the tempo "72" of the song is stored in the header 12 as shown in FIG. 3, and the song "Cuckoo" shown in FIG. When storing the music, the tempo "138" of the song is stored in the header 12 as shown in FIG.

Next, the voice code for each note is input by operating the voice code input key provided on the keyboard 2 (step 103). That is, Figure 2
The sounds "sa", "ku", "ra", etc. per note of the lyrics of the song shown in FIG. 1,
Like the 1st and 2nd beats of the 2nd measure, each note has a ``ka'' sound.
, if 2 voices of "Kou" are assigned, 2
Voice is input by operating the voice code input key.

Subsequently, it is determined whether the lyrics = "*" (step 104). That is, when the melody is a rest and there are no lyrics on the beats, as in the third beat of the first and second bars of the song shown in FIG. It is necessary to operate the rest key (*) provided at 2, and it is determined in step 104 whether or not this rest key (*) has been operated. If the rest key (*) is operated, the determination at step 104 becomes YES, and the rest length is input as a 16th note length (step 105). Therefore, like the 3rd beat of the 1st and 2nd measures of the song shown in Figure 4,
If it is a quarter rest, "4" is input as the rest length by operating the numerical keys.

On the other hand, if the determination in step 104 is NO and there are notes constituting the melody, the pitch and length of the notes constituting the melody are determined by a plurality of alphabets provided on the keyboard 2. Input is made by operating a pitch input key consisting of a key and a numeric key, and a pitch input key consisting of a plurality of numeric keys (step 105).
. Then, in step 107 following step 105 or step 106, the data input by each key operation is stored in the sequence memory 3, and until the input end key is operated and the determination in step 108 becomes YES, step 103 The loop from step 108 is repeated. Therefore, during this period, instrument users ““
Enter the audio code, pitch, and length of each note of the song "Sakura Sakura" for the entire song, such as "Sa""A4""4" and "ku""A4""4". Once this is done, the lyrics and pitch of the melody of the entire song are stored in the sequence memory 3.

[0019] At this time, if one voice spans two consecutive eighth notes with a slur, for example, in the 6th measure and the part corresponding to the second beat of the lyrics "Gee", ,
By key operation, “G”, “E4”, “2”, “-”, “
By inputting "C4""2"", the voice code, pitch and length are stored in correspondence with each note as shown in FIG. In addition, when two voices correspond to one quadrupty note, as in the part corresponding to "ka" and "kou" on the first and second beats in the 1st and 2nd bars of the song shown in Figure 4. ``Ka''``C4''``4'',``Kou''``A''
By inputting ``4'', the voice code, pitch, and length are stored in association with each note as shown in FIG.

[0020] When the input end key is operated after inputting all the audio codes, pitches, and durations of one of the songs, the determination in step 108 becomes YES, and the process returns to step 101. Therefore, in this state, if the sequence execution key provided on the keyboard 2 is operated to set the sequence execution mode, the process proceeds from step 101 to step 109 (FIG. 7), and the data is stored in the header 12 of the sequence memory 3. The current tempo data is set in the tempo clock 11. continuation,
At the tempo set in the tempo clock 11, the sequence data consisting of lyrics data and melody data (pitch, length) stored in the sequence memory 4 is
They are read out sequentially from the first address (step 110)
. In other words, if the sequence data shown in FIG. 3 is stored in the sequence memory 3,
4", "ku A4 4", ... are sequentially changed to the tempo "
It is read out at a speed of 72".

Next, it is determined whether or not the read lyrics code is a rest code (*) (step 111).
, this determination is YES, and the third and sixth shown in FIG.
If the sequence data at the th address is a rest code such as "* * 4", the rest length "4" of this sequence data is extracted and rest processing is executed (step 112). Here, the rest process is simply a process that does not instruct the sound generator 4 to produce sound, and therefore, the sound from the speaker 14 is stopped during this time.

On the other hand, if the determination in step 111 is NO and the read sequence data is a voice code other than a rest code, then the read sequence data contains a plurality of lyrics data. It is determined whether it is a voice or not (step 113). If this determination is NO and the lyrics data is a singular sound such as "sa", "ku", or "ra", an attack part is sounded according to the pitch data (step 114). As a result, speaker 1
From 4 onwards, the sound corresponding to the lyric data at the time, such as "sa" or "ku", starts to be emitted at a pitch according to the pitch data.

Subsequently, after the loop time is calculated based on the note length data (step 115), the loop part is sounded for the loop time calculated in the previous step according to the note pitch data (step 116). difference"
The vowel continues to be pronounced continuously for the loop time. Furthermore, in the following step 117, the next audio data is prefetched. In other words, if the current voice data is "sa" stored at the first address of the sequence data shown in FIG. 3, the voice data "ku" at the next address is read in advance. Subsequently, it is determined whether or not this prefetched audio data is the extension code "-" (step 118). If the audio data is "ku", this determination is NO, so step 119 is performed. Then, the release section is made to emit a sound in accordance with the pitch data, thereby attenuating the sound being emitted from the speaker 14. After the process of step 119 is executed, the discrimination process from step 110 is repeated, and as a result, for example, the sounds that make up the lyrics of the song "Sakura Sakura" are sequentially divided into pitches and lengths of the corresponding notes. It is uttered with.

[0024] At this time, when the lyric "gi" is uttered at the 6th measure and the 2nd beat of the song,
When the next audio data is read out in step 117, the next audio data becomes the extended code "-", so the determination in step 118 becomes YES. Therefore, the process proceeds from step 118 to step 120, and the next note length and pitch data are extracted. In other words, for "-C42" marked with * in FIG. 3, the pitch data "C4" and the pitch data "2" are extracted.

Furthermore, the current pitch is pitch bent in accordance with the next pitch data while the loop section up to the present is being sounded (step 121). In other words, while sounding the loop part of "gi" which is the loop part up to now, according to C4 which is the pitch data of the next "-", E4 which is the pitch of said "gi"
is pitch bent to C4, and then the processing from step 115 described above is executed. This will result in "
In the pronunciation process of "Gee", the pitch gradually decreases from E4 to C4, creating a pronunciation form in which E4 and C4 are slurred. Therefore, in the song shown in Figure 2, the lyrics in the 6th measure are E4, E4, C4, D
It is pronounced as "Kagi-ri" with a pitch of 3 and a slur on E4 and C4, and it is possible to generate a sound that matches the musical score without ignoring the presence of the slur. .

On the other hand, if the lyrics data includes multiple voices, such as the lyrics data "ka" and "kou" stored at the first address and the next address of the sequence data shown in FIG. The determination is NO, and the process proceeds from step 113 to step 122 (FIG. 8), where the sound length of each voice is calculated from the sound length and the number of generated voices. This calculation of the tone length is performed by dividing the tone length by the number of generated pronunciations.
”, and the tone length is “4” as shown in FIG. 5, so the tone lengths of the sounds “ko” and “u” are each “2”. Therefore, when sequence data is generated for one voice in the next step 123, sequence data consisting of "ko A4 2" is generated for the voice "ko".

[0027] In the next step 124,
First, the loop time is calculated based on the note length "2" in the sequence data of "Ko A4 2", and then the attack, loop, and release parts are sounded according to the pitch data and the loop time (step 125). Furthermore, it is determined whether or not the generated sound data has ended, that is, whether or not the sound generation of all of the plurality of generated sound data has been completed (step 126). In this example, since the pronunciation of "ko" has not been completed yet, this determination is NO, and the process from step 123 is executed again, and in step 123, the sequence data for the remaining one sound "ko" is ""A42" is generated. Subsequently, the processes of steps 124 and 125 described above are executed, whereby the voice "ko" is uttered with pitch A4 and pitch length "2". Of course, the same goes for "Ka", "Ka C5 2" and "っ C5".
2'' sequence data is generated, and each pronunciation is pronounced with pitch C5 and tone length "2".

Therefore, in the above-described sequence data creation mode, the lyrics ``Kakko'' are changed to ``ka'' and ``tsu'', which are eighth note length and pitch C5, and ``ka'' and ``tsu'', which are eighth note length and pitch A4. There is no need to divide and store the audio data into four pieces, ``ko'' and ``u,'' but instead store the lyrics for ``ka'' and ``kou,'' each corresponding to a quarter note according to the score. It is now possible to accurately generate a "cuckoo" sound according to the musical score.

[0029] When the utterance of each utterance pronunciation data is completed, the determination in step 126 becomes YES.
By the process determination from step 110 again, the next sequence data is read out, and a voice corresponding to the pitch and length of the lyrics and musical score is uttered.

[0030]

As explained above, when the audio data of one voice is stored in correspondence with a plurality of consecutive notes, the present invention reads out the pitch of the next audio data following the current audio data. The pitch of the audio signal is continuously changed from the pitch of the current audio data to the pitch of the next audio data. Therefore, it is possible to make one sound correspond to multiple notes of different pitches and produce the sound in a pronunciation form with slurs, without ignoring the presence of slurs attached to the multiple notes. , it becomes possible to generate audio that matches the musical score.

[0031] Furthermore, when audio data consisting of a plurality of sounds is stored corresponding to the pitch and duration of one note, the duration of each sound is calculated, and the calculated duration, pitch, and The audio signal is generated based on audio data. Therefore, when multiple voices are assigned to one note, if multiple voices are stored in correspondence with each note according to the musical score, it becomes possible to generate the voice accurately according to the musical score. . Therefore,
Both inventions allow the musical expression of the score and the correspondence between the notes and the lyrics to be accurately reflected in the audio by simply storing the notes shown on the score and the lyrics corresponding to the notes as data. It becomes possible.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the overall structure of an embodiment of the present invention.

FIG. 2 is a musical score diagram showing an example of a song.

FIG. 3 is an explanatory diagram showing the contents of sequence data of the music shown in FIG. 3;

FIG. 4 is a musical score diagram showing an example of a song.

FIG. 5 is an explanatory diagram showing the contents of sequence data of the music shown in FIG. 4;

FIG. 6 is a flowchart showing the operation of this embodiment.

FIG. 7 is a flowchart following FIG. 6;

FIG. 8 is a flowchart following FIG. 7;

[Explanation of symbols]

1 CPU 2 Keyboard 3 Sequence memory 4 Sound generator 6 Lyrics data storage area 7 Audio data storage area 8. Tone length data storage area 9 Melody data storage area

Claims (2)

[Claims] 1. Storage means for storing the pitches and lengths of a plurality of notes constituting a music piece and voice data of lyrics corresponding to each note; An audio electronic musical instrument comprising a reading means for reading audio data, and an audio signal generating means for generating an audio signal based on the pitch, duration, and audio data read by the reading means. When data is stored corresponding to a plurality of consecutive notes, the pitch of the note corresponding to the next audio data following the current audio data is read out, and the pitch of the note corresponding to the audio data at the current time is read out. An audio electronic musical instrument characterized in that a pitch variable means is provided for continuously changing the pitch of the audio signal at the pitch of a note corresponding to the next audio data. 2. Storage means for storing the pitches and lengths of a plurality of notes constituting a music piece and audio data of lyrics corresponding to each note; An audio electronic musical instrument comprising a reading means for reading out audio data, and an audio signal generating means for generating an audio signal based on the pitch, duration and audio data read by the reading means. When audio data consisting of a plurality of voices corresponding to pitch and note length is stored, calculation means for calculating the note length per one voice is provided, and the audio signal generating means is configured to calculate the note length per voice. An audio electronic musical instrument, characterized in that the audio signal is generated based on the tone length, the pitch, and the audio data.