JP7280055B2 - Song Pronunciation Character String Automatic Correction Program and Song Pronunciation Character String Automatic Correction Apparatus - Google Patents

Description

Article 30, Paragraph 2 of the Patent Law applies https://cmusic. kawai. On the website of Kawai Musical Instruments Co., Ltd. published at the address of jp/a/Download/smz, Kawai Musical Instruments Co., Ltd. about the automatic correction program for singing pronunciation character strings and the automatic correction device for singing pronunciation character strings invented by Naoyuki Tanaka Seisakusho released it on December 18, 2018.

The present invention relates to a function of singing lyrics displayed in a musical score in music creation software such as a DAW (digital audio workstation), and more specifically, automatic correction of singing pronunciation character strings that automatically corrects pronunciation character strings of lyrics. The present invention relates to a program and a singing pronunciation character string automatic correction device having the function.

A DAW is software that creates music by recording musical instrument performances and inputting musical notes. The created music can be output as audio files such as WAV and mp3, and can be edited on the DAW. can. In addition, input notes can be converted into MIDI data and saved as SMF (Standard MIDI File).

When displaying DAW music data, there are cases where lyrics are assigned to a musical score, as shown in FIG. 7, for example. When the lyrics are pronounced using the function to sing the score, the ``ha'' in the lyrics ``ki'', ``no'', ``u'' and ``ha'' is pronounced as ``ha''. On the other hand, the phonetic character string "wa" for singing was set as needed.

Further, as a technique for correcting the pronunciation of words in a sentence, a speech synthesizer described in Patent Document 1 has been proposed. This speech synthesizer performs language processing on sentences containing kanji and kana, and synthesizes the results into speech, correcting the pronunciation so that it is read out in the manner intended by the sentence creator. be.

JP-A-9-258763

The above-mentioned setting of phonetic strings for singing has the problem that it takes a lot of time and effort because it is necessary to search for the lyrics one by one and change the settings of the phonetic strings for the entire displayed score. there were. In the musical score example shown in Fig. 7, it is necessary to set the phonetic character string "wa" by separately specifying "ha" in the upper row and "ha" in the lower row. there were.
Further, the technique of performing pronunciation correction by a speech synthesizer described in Patent Document 1 does not automatically perform correction by matching pronunciation character strings to lyrics in a musical score.

The present invention has been proposed in view of the above-mentioned circumstances. Automatic correction of singing pronunciation character strings that automatically changes lyrics of the same kind in musical scores when setting pronunciation character strings for singing with respect to lyrics of musical scores. An object of the present invention is to provide a program and an automatic correction device for singing pronunciation character strings having the function.

In order to achieve the above object, the singing pronunciation character string automatic correction display program of the present invention, when setting pronunciation character strings for singing, for lyrics in a musical score, finds lyrics that require setting of similar pronunciation character strings. Estimate and automatically set phonetic character strings for singing.
That is, claim 1 is an automatic editing program in which a computer determines whether or not the lyrics of the entire musical score are to be corrected. The lyric information of the lyrics for which the pronunciation for singing has been set is analyzed, and other lyrics in the musical score with similar analysis results are changed in the same manner to set the pronunciation for singing, and the analysis results of the lyric information are similar. "Suru" is characterized by having the same pronunciation timing and the same lyrics as the lyrics for which the pronunciation for singing is set .

Claim 2 is a singing pronunciation character string automatic correction program,
The analysis result of the lyric information is similar when the lyric is the same as the lyric for which the singing pronunciation is set, and the preceding two lyric match.

Claim 3 is a singing pronunciation character string automatic correction program,
If (A) is the lyric object to be edited and (B) is the lyric object to be determined whether or not to be corrected,
When the analysis result of the lyric information is similar, the lyric character strings of the lyric object (A) and the lyric object (B) match, and the preceding rests of the lyric object (A) and the lyric object (B) are It is characterized by the fact that the number of notes up to is the same.

Claim 4 is a singing pronunciation character string automatic correction program,
If (A) is the lyric object to be edited and (B) is the lyric object to be determined whether or not to be corrected,
If the analysis result of the lyric information is similar, the lyric character string of the lyric object (A) is "u", the pronunciation character string of the lyric object (A) is set to "o", and the lyric object (A) is set to "o". The last vowel of the lyric before A) is "o", the lyric character string of the lyric object (B) is "u", and the last vowel of the lyric before the lyric object (B) is "o". It is characterized by

The apparatus for automatically correcting a singing pronunciation character string according to claim 5 is an apparatus for automatically correcting a singing pronunciation character string for setting the pronunciation for singing to the lyrics in the musical score displayed on the display unit (21),
an editing input unit (22) for specifying a lyric object for which the user sets the pronunciation for singing;
a lyric object search unit (23) for retrieving another lyric object in the musical score;
a changed lyric estimation unit (24) for analyzing lyric information of a lyric object for which singing pronunciation is set and estimating whether or not to change the lyric of the another lyric object;
a singing pronunciation setting unit (25) that sets singing pronunciation for lyrics that need to be changed based on the result of the changed lyrics estimating unit ;
When estimating whether or not to change the lyrics of the separate lyric object, the changed lyric estimating unit (24) estimates the lyric having the same timing and the same characters as the setting of the singing pronunciation as the lyric to be changed. It is characterized by

Claim 6 is a singing pronunciation character string automatic correction device,
When estimating whether or not to change the lyrics of the another lyric object, the changed lyric estimation unit (24) detects that the lyrics that are the same as the lyrics for which the pronunciation for singing has been set and the preceding two lyrics match. It is characterized by estimating that the lyrics need to be changed in the case.

According to the singing pronunciation character string automatic correction program and the singing pronunciation character string automatic correction device of the present invention, when the user sets the pronunciation character strings for singing with respect to the lyrics in the musical score, It is possible to guess the lyrics that need to set phonetic character strings and automatically set the phonetic character strings for singing. Therefore, it is not necessary to set a singing pronunciation character string for each lyric character string.

1 is a block diagram showing the hardware configuration of a music data display device equipped with a singing pronunciation character string automatic correction device of the present invention; FIG. FIG. 2 is a block diagram for explaining the function of the automatic pronunciation character string correcting device for singing; FIG. 10 is a diagram showing an example of musical score for explaining a lyric object; FIG. 10 is a flow chart diagram showing a processing procedure for automatic correction of a singing pronunciation character string; FIG. 4 is a diagram showing an example of musical scores displayed on the display unit; FIG. 4 is a diagram showing an example of musical scores displayed on the display unit; FIG. 10 is a diagram showing an example of musical score with lyrics displayed;

A music data display device equipped with the device for automatically correcting a singing pronunciation character string of the present invention is constructed on a general personal computer (PC) by installing a music data display program including a program for automatically correcting a singing pronunciation character string. be.
As shown in FIG. 1, the personal computer includes a central processing unit (CPU) 11, a ROM 12, a RAM 13, a large-capacity storage device HDD 14, a CD/DVD drive 15, an operation unit 16 for inputting music data, a USB An interface 17, a sound device 18 having a sound source and a speaker unit for reproducing specified sounds corresponding to music data, and a display device 19 having a display unit such as an LCD for displaying music data. 20 is connected.
In this embodiment, an example in which a music data display device is constructed with a general PC will be described, but it may be constructed with a smartphone, a tablet, or dedicated hardware.

The CPU 11 processes or calculates data and controls various components connected via the bus 20 .
A control procedure (computer program) for the CPU 11 is stored in the ROM 12 in advance, and the CPU 11 starts by executing the program.
The RAM 13 is used as a work memory for data input/output and transmission/reception, and as a temporary storage for controlling each component. The CPU 11 executes a music data display program (singing pronunciation character string automatic correction program) in the RAM 13 to display music data including musical scores and lyrics on the display device 19 .
A desired program such as a music data display program is stored in the HDD 14, copied to the RAM 13 and executed.
The CD/DVD drive 15 loads programs and data from a CD or the like to the HDD 14, writes data and programs in the PC to a CD-R or the like, and outputs them to the outside.

The operation unit 16 is composed of a keyboard, a mouse, etc. By operating the keyboard, operating the cursor position (cursor movement) using the mouse, and performing a click operation, music data such as musical notes and lyrics can be input and displayed on the display unit of the display device 19 . perform editing work.
The USB interface 17 is for connecting a printer or other external devices via USB.
The sound device 18 includes an audio driver, a line terminal, a headphone terminal, and the like, and reproduces designated sounds corresponding to music data.
The display device 19 includes a video card and a display, and displays music information such as musical notes and lyric information that can be displayed on a musical score, and operation means for inputting and editing such information.

FIG. 2 is a block diagram for explaining the function of the singing pronunciation character string automatic correcting device for setting the singing pronunciation for the lyrics in the musical score displayed on the display unit 21 of the music data display device.
The singing pronunciation character string automatic correction device includes an editing input unit 22 for specifying a lyric object for which the user sets the pronunciation for singing, a lyric object searching unit 23 for searching for another lyric object in the musical score, and the pronunciation for singing. a changed lyric estimation unit 24 for analyzing the lyric information of the received lyric object and estimating whether or not to change the lyrics of the other lyric object; and a storage unit 26 for storing music information.

When the user sets the pronunciation for singing (character string for singing) for the lyrics (character string for singing) of the musical score drawn on the display unit 21, the editing input unit 22 converts the characters for singing of the corresponding lyric object. Edit columns. The lyric object in this case means character information corresponding to each note in the musical score, and includes lyric information of lyric character strings and singing character strings. For example, in the musical score shown in FIG. handle.

The lyric object search unit 23 searches for lyric objects in the musical score.
The changed lyric estimation unit 24 analyzes the lyric information of the lyric object for which the pronunciation for singing is set by the editing input unit 22, and retrieves the lyric object searched by the lyric object searching unit 23 (what is the lyric object for which the pronunciation for singing is set? It is estimated whether or not the lyrics of another lyrics object) are to be changed (whether or not the lyrics need to be changed). This estimation is determined by whether or not the analysis results of the lyric information of the lyric objects are similar, and the estimation procedure will be described later.

A singing pronunciation setting unit 25 automatically sets singing pronunciations for lyrics that need to be changed based on the estimation result of the changed lyrics estimating unit 24 . When the pronunciation for singing is set, the pronunciation for singing is set by enclosing the lyric character string for which the pronunciation character string is set with a rectangle 50, for example, as shown in FIG. characters are displayed so that they can be distinguished from characters that are not Instead of enclosing the lyric string with a rectangle, the color of the text of the lyric may be changed.

The storage unit 26 stores musical score information for displaying musical scores and lyric objects as lyric information. Musical score information includes note information, time information (Tick), bar line position information, time signature information (numerator, denominator), key signature information (Number of # and ♭ (# is positive, ♭ is negative)). . The lyric object of each note includes the lyric information of the phonetic character string when the lyric character string corresponding to the note and characters for singing are set.

Next, a procedure for estimating lyrics that need to be changed based on whether or not the analysis results of the lyric information of the lyric objects are similar to automatically correct the singing pronunciation character strings will be described with reference to the flowchart of FIG. Let (A) be the lyric object manually edited by the user, and (B) be the lyric object whose correction target or not is determined automatically by estimating the lyric to be changed.

First, the edit input section 22 detects that the pronunciation character string of the lyric object A has been edited (step 31). For example, the user edits the pronunciation character string "wa" for "ha" (lyric object A) of "ki", "no", "u", and "ha" in the musical scores shown in FIGS. In this case, the edit input unit 22 detects that the lyric object "ha" has been edited.

Subsequently, in response to the fact that the lyric object "ha" has been edited, another lyric object B in the musical score is detected (step 32).
If another lyric object B exists (step 33), it is determined whether or not the lyric character strings of the lyric object A and the lyric object B match (step 34). In the above example, if the lyric object A is "ha", it is determined whether the lyric object B is "ha".

If the lyric character strings "ha" and "ha" of the lyric objects match in step 34, the pronounced character string "ha" of the lyric object A before editing and the pronounced character string "ha" of the lyric object B match. (step 35).
For example, if there are multiple ``ha'' lyrics in a musical score, and the user intentionally sets the pronunciation character strings for only some of the ``wa'' to characters other than ``wa'', the ``wa'' in the musical score This is because even if the pronunciation character string is set to "wa", the lyrics that are intentionally set to other than "wa" are not automatically corrected to "wa".

If the pronunciation character strings of the lyric objects match in step 35, it is determined whether or not the pronunciation timings of the lyric object A and the lyric object B match (step 36).
When the sounding timings match, the time information (Tick) of the notes held by the lyric object A and the lyric object B are compared, and if the sounding times match, it is determined that the sounding timings match.

If the pronunciation timings match in step 36, the pronunciation character string of lyrics object B is set to the same pronunciation character string as that of lyrics object A (step 38).
In the above example, as shown in the musical score of FIG. 6, when the user sets the pronunciation character string of "wa" (lyric object A) in the upper part of the lyric to "wa", the lower part of the lyric having the same pronunciation timing: "Ha" (lyrics object B) automatically has its phonetic character string set to "wa". When the pronunciation character string is changed, the lyric character string is surrounded by a square 60 so as to be displayed so as to be distinguished from the characters for which the pronunciation for singing is not set.
At this time, using the lyrics object B as a reference, processing for correcting other lyrics is performed.

If the phonetic strings of the lyric objects do not match in step 35, the two lyric strings before and after lyric object A and the two lyric strings before and after lyric object B match. (step 37), and if they match, the phonetic character string of the lyric object B is set to the same phonetic character string as that of the lyric object A (step 38).
In the above example, as shown in the musical score of FIG. 5, when the user sets the phonetic character string of "wa" (lyrics object A) at the top of the lyrics to "wa", the same character "ha" is For "wa" (lyrics object B) in the lower row, since both the previous "u" and the two previous "no" match, the phonetic character string is automatically set to "wa". When the pronunciation character string is changed, the lyric character string is surrounded by a square 50 so as to be displayed so as to be distinguished from the characters for which the pronunciation for singing is not set.

If the lyric strings do not match at step 34, if the phonetic strings do not match at step 35, and if the preceding two lyric strings do not match at step 37, step 38 sets the phonetic strings for lyric object B. If so, the process returns to step 32 and the search for another lyric object in the score is repeated.
Further, if there is no lyric object B in step 33, it is determined that all the lyric character strings in the musical score are set to singing character strings, so the automatic correction of the singing pronunciation character strings in the musical score ends. .

According to the above example, as an example where the analysis result of the lyrics information is similar, the same character ("wa" in the above example) as the lyrics for which the pronunciation for singing is set and the pronunciation timing are the same (step 34 , steps 35 and 36), or when the lyrics are the same as the lyrics for which the pronunciation for singing is set (step 35) and the two previous lyrics match (step 37), but the following cases may be taken as an example in which the analysis results of the lyric information are similar.
As a first example, the lyric character strings of the lyric object (A) and the lyric object (B) match, and the number of notes up to the preceding rest of each of the lyric object (A) and the lyric object (B) is if it matches. The number of notes is determined from the note information stored in the storage unit 26 .
As a second example, the lyric character string of the lyric object (A) is "U", the pronunciation character string of the lyric object (A) is set to "O", and When the lyric last vowel is "o", the lyric character string of the lyric object (B) is "u", and the lyric last vowel before the lyric object (B) is "o". Vowel information is determined from the lyric information of the lyric object stored in the storage unit 26 .
As a third example, when morphological analysis is performed and the lyric character string of the lyric object (A) is the particle "wa" or "he", the lyric characters of the lyric object (A) and the lyric object (B) If the columns match and the lyric string in the lyric object (B) is also a particle. The information as to whether it is a particle is determined from the lyric information of the lyric object stored in the storage unit 26 .

By combining the above-described examples with similar analysis results of lyric information into the automatic correction flowchart of FIG. can.

In addition, in the above-described singing pronunciation character string automatic correction device, the process of automatically setting the singing pronunciation for the lyrics that need to be changed can be performed by using an artificial device operated by a computer program according to the procedure of the flowchart shown in FIG. It may also be done using the brain (AI).

21...Display section 22...Edit input section 23...Lyric object search section 24...Changed lyric estimating section 25...Singing pronunciation setting section 26...Storage section.

Claims (6)

An automatic editing program in which a computer determines whether or not the lyrics of the entire score are to be corrected,
When the user sets the singing pronunciation for the lyrics in the score,
Analyzing the lyric information of the lyrics for which the singing pronunciation has been set, and making similar changes to other lyrics in the musical score for which the analysis results are similar to set the singing pronunciation ,
If the analysis result of the lyrics information is similar, it means that the lyrics have the same pronunciation timing and are the same as the lyrics for which the pronunciation for singing is set.
A singing pronunciation character string automatic correction program characterized by: An automatic editing program in which a computer determines whether or not the lyrics of the entire score are to be corrected,
When the user sets the singing pronunciation for the lyrics in the score,
Analyzing the lyric information of the lyrics for which the singing pronunciation has been set, and making similar changes to other lyrics in the musical score for which the analysis results are similar to set the singing pronunciation,
The analysis results of the lyric information are similar when the lyric is the same as the lyric for which the singing pronunciation is set, and the preceding two lyric match.
A singing pronunciation character string automatic correction program characterized by : An automatic editing program in which a computer determines whether or not the lyrics of the entire score are to be corrected,
When the user sets the singing pronunciation for the lyrics in the score,
Analyzing the lyric information of the lyrics for which the singing pronunciation has been set, and making similar changes to other lyrics in the musical score for which the analysis results are similar to set the singing pronunciation,
If (A) is the lyric object to be edited and (B) is the lyric object to be determined whether or not to be corrected,
When the analysis result of the lyric information is similar, the lyric character strings of the lyric object (A) and the lyric object (B) match, and the preceding rests of the lyric object (A) and the lyric object (B) are If the number of notes up to
A singing pronunciation character string automatic correction program characterized by : An automatic editing program in which a computer determines whether or not the lyrics of the entire score are to be corrected,
When the user sets the singing pronunciation for the lyrics in the score,
Analyzing the lyric information of the lyrics for which the singing pronunciation has been set, and making similar changes to other lyrics in the musical score for which the analysis results are similar to set the singing pronunciation,
If (A) is the lyric object to be edited and (B) is the lyric object to be determined whether or not to be corrected,
If the analysis result of the lyric information is similar, the lyric character string of the lyric object (A) is "u", the pronunciation character string of the lyric object (A) is set to "o", and the lyric object (A) is set to "o". A) the last vowel of the lyric before A) is "o", the lyric character string of the lyric object (B) is "u", and the last vowel of the lyric before the lyric object (B) is "o" is
A singing pronunciation character string automatic correction program characterized by : A singing pronunciation character string automatic correction device for setting singing pronunciation for lyrics in a musical score displayed on a display unit,
an editing input for specifying a lyric object for which the user sets the singing pronunciation;
a lyric object searcher for searching for another lyric object in the score;
a changed lyrics estimation unit that analyzes the lyrics information of the lyrics object for which the pronunciation for singing is set and estimates whether or not to change the lyrics of the another lyrics object;
a singing pronunciation setting unit that sets singing pronunciation for lyrics that need to be changed based on the result of the changed lyrics estimating unit ;
When estimating whether or not to change the lyrics of the other lyric object, the changed lyric estimating unit estimates that the lyric having the same pronunciation timing and the same characters as those set for the singing pronunciation is the lyric that needs to be changed.
A singing pronunciation character string automatic correction device characterized by: A singing pronunciation character string automatic correction device for setting singing pronunciation for lyrics in a musical score displayed on a display unit,
an editing input for specifying a lyric object for which the user sets the singing pronunciation;
a lyric object searcher for searching for another lyric object in the score;
a changed lyrics estimation unit that analyzes the lyrics information of the lyrics object for which the pronunciation for singing is set and estimates whether or not to change the lyrics of the another lyrics object;
a singing pronunciation setting unit that sets singing pronunciation for lyrics that need to be changed based on the result of the changed lyrics estimating unit ;
When estimating whether or not to change the lyrics of the other lyric object, the changed lyric estimating unit uses the same lyrics as the lyrics for which the singing pronunciation is set, and furthermore, the preceding two lyrics in the pronunciation character string are used. Predict lyrics that need to be changed if matched
A singing pronunciation character string automatic correction device characterized by:

Images