JPH031196A - Automatic accompanying device - Google Patents

Abstract

PURPOSE:To suppress the disagreeableness to hear of an ensemble musical performance to the minimum even if there is an error in a melody input position by reading in melody data of a melody store part and performing a melody, unless the melody is inputted in a position where the melody goes in. CONSTITUTION:A musical performance control part 5 reads in successively accompaniment data from an accompaniment data store part 4, and executes an accompaniment musical performance. On the other hand, in a melody input means 1, an acoustic signal is always inputted from the start of the accompaniment musical performance and sent to a melody detecting part 2. In this case, at every one bar, detecting information related to whether a melody is inputted in its bar or not is derived. The musical performance control part 5 collates the melody detecting information with melody position information of the accompaniment data store part 4, while performing the accompaniment, and in the case the melody is not detected, the corresponding melody data is read in from a melody store part 3, and the musical performance of the melody is started together with the accompaniment. In such a way, even in the case a melody input person forgets a position of a melody input, unnaturalness of an ensemble can be reduced.

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention relates to an automatic accompaniment device, and more particularly to an automatic accompaniment device that performs f′i accompaniment performance in response to melody singing by a singer or melody performance by a musical instrument. .

(Prior art) Traditionally, there has been a method of actually playing the 1'F for vocal melodies and musical instrument melodies (Method 1).
Also, a method called karaoke (Method 2) of playing back performances recorded on tapes or compact discs is used. In addition, in recent years, computer-assisted automatic performance (method 3)% type% In these conventional techniques, - the position where the melody should exist in the performance of the song is assumed in advance, and the person inputting the melody inputs the position where the melody will be placed. It is common to play the melody after knowing the melody.

In methods 1 to 3 listed in 1., the accompaniment methods are as follows:
The accompanist performs (Method 1), the recorded material is played back (Method 2), and the performance is performed automatically according to the performance data (Method 3).
), the content of the accompaniment is fixed and does not change depending on the content of the melody performance.

Note that method 3 is a method that has come into use due to the recent development and price reduction of electronic equipment, and is a method in which a computer performs automatic performance according to performance data. Automatic performance under computer control can be performed using a product dedicated to automatic performance such as Yamaha's QX3, a general-purpose computer such as NEC's PC9801, and commercially available automatic performance software. As an automatic performance control program, for example, a PC from "Come on Music Co., Ltd."
Commercial software called X-PCQ8 can be used.

(Problem to be solved by the invention) By the way, since the person inputting the melody is a human, it is natural that the person inputting the melody may make a mistake in the input position.In other words, it is possible that the person inputting the melody may not input the melody when it should exist, or vice versa. Another common mistake is inputting a melody when it shouldn't exist.

In such cases, in Conventional Method 1, if the accompaniment performer notices the error of the person inputting the melody, the accompaniment player performs the accompaniment according to the error, thereby minimizing the unnatural sounding of the unsampled melody and accompaniment. can be held to 0
Accompaniment performance that corresponds to the mistake is the place where the melody should be included.
For example, an accompanist may play a melody when there is no melody input, or when a melody input is received at a position where the melody cannot be entered, the accompaniment may be switched to the accompaniment at the position where the melody should be. However, in method 1, the Eerr player is always required, and the accompanist also needs skills to provide accompaniment in response to mistakes made by the person inputting the melody. Therefore,
It was not possible to perform accompaniment easily.

Method 2 only plays back the recorded performance and does not have a way to change the accompaniment content in response to errors made by the person inputting the melody. Therefore, unless the person inputting the melody notices and corrects it, the melody and accompaniment are There is no way to improve the difficulty of listening to an unsampled performance.Also, even if you notice a mistake, the accompaniment will continue without interruption, so if you want to correct it, start with the correct position from the middle of the melody. ! , which is often difficult for beginners to singing or playing.

In method 3, accompaniment performance is performed according to accompaniment data created in advance. With this method, the accompaniment data can be changed, but the changed data must be created before the accompaniment performance, and the accompaniment data must be changed at the same time as the melody is input while performing the accompaniment performance. In this way, in Methods 2 and 3, it was not possible to realize accompaniment performance in response to mistakes, as shown in the example of Method 1 above.

The purpose of the present invention is to solve the above-mentioned problems in conventional methods 1, 2.3, and to minimize the difficulty of listening to an unsampled performance even if the melody inputter makes a mistake in the melody input position. An object of the present invention is to provide an automatic accompaniment device that is capable of performing automatic accompaniment and does not require an accompanist.

(Means for Solving the Problems) An automatic accompaniment device according to a first aspect of the present invention includes a melody input means for inputting a melody as an acoustic signal, and an analysis of the acoustic signal input from the melody input means,
A melody detection section that obtains melody detection information indicating whether or not a melody has been input; a melody storage section that stores the pitch and length of the melody as melody information; and a melody storage section that stores music information necessary for playing the accompaniment as accompaniment data. and an accompaniment data storage section that describes the position in the song where the melody is to be input as the melody input position;
From the melody detection information obtained by the melody detection section and the melody input position of the accompaniment data storage section, it is determined whether the melody is input as specified by the melody input position, and the melody information is retrieved from the melody storage section. and a performance control section that reads the melody information and performs a performance in combination with the accompaniment data in the accompaniment data storage section.

The automatic f'P playing device according to the second invention of the present invention is as follows:
a melody input means for inputting a melody as a sound Ut=; a melody detection section that analyzes the 1g signal of the human inputted sound from the melody manual means to obtain melody detection information indicating whether or not a melody has been input; A melody storage section that stores the pitch and length of the melody as melody information, and accompaniment data that stores the musical information necessary for playing the accompaniment as accompaniment data, and describes the position in the song where the melody should be input as the melody input position. a melody position detection section that detects the position of the input melody from the information obtained by the melody detection section and the melody information in the melody storage section; and a performance control section that reads accompaniment data from the accompaniment data storage section and performs the performance.

(Function) The first aspect of the present invention is to provide a melody detection unit that detects whether a melody is input, and if a melody is not input at a position where a melody should be input, the melody data in the melody storage unit is read and the melody is detected. The melody detection section then detects whether a melody has been input, and ends the melody performance when the melody input is detected. . As a result, even if you forget the position where the melody is input, the person inputting the melody can quickly notice the mistake, and during that time, the melody is also played in addition to the IT performance. The unnaturalness of the unsample is reduced. Furthermore, since the melody performance serves as a guide, it is possible to easily start playing the original melody even in the middle.

Melody detection is performed by analyzing the audio signal input from the melody input means, as described below. First, the input melody audio signal is sequentially A,
/D converted and stored as waveform data.

The waveform data is analyzed and converted into time-series power and pitch data. To obtain the power data and pitch data, for example, a generally well-known method using an autocorrelation function can be used. The pitch extraction method is F-Digimol Audio Processing J (Sadahiro Furui) published by Tokai University Press.
A variety of schemes are known, and any of these schemes can be used, as described on pages 57-59 of .

In order to obtain detection information indicating whether or not a melody has been input, there are two possible methods: using only the determined power data, or using both power data and pitch data. For those that only use power data, there is a method of determining whether there is a melody input based on whether a predetermined power threshold is exceeded, or a method of determining whether a melody input is present based on whether the power data exceeds a threshold of cumulative value over a certain period of time. There are ways to do this. Furthermore, when power and pitch are used together, a method can be considered to increase the accuracy of determining melody input by comparing the melody data to be input with the determined pitch.

The performance data necessary for the automatic accompaniment of the present invention includes the above-mentioned melody performance data for prompting the melody input person to input a melody, and performance data for If performance, which is the original purpose. In addition, information indicating the position in the song where the melody should be input is stored together with the tV performance data as information necessary to determine whether the melody is input as expected. The position of means a set of time points on the time axis, that is, an interval.

During performance, while playing the accompaniment data, it is determined at regular intervals whether a melody input is expected at that time, and it is determined whether the actual melody input matches the determination. At that time, if there is no melody input even though it should be input, the melody data is read and a performance with melody is performed along with accompaniment data.
Furthermore, if the assumed melody input and the actual input match, only the accompaniment data is played. At this time, if the melody data is already being played, the reading and playing of the melody data is stopped.

The second aspect of the present invention is to deal with the case where a melody is input at a position where the melody should not be input. If the accompaniment position is different from the one in the middle, an accompaniment start position matching the detected melody position is found, the accompaniment performance up to that point is stopped, and the accompaniment performance is started from the found accompaniment start position. As a result, even if the person inputting the melody makes a mistake in the starting position of the melody, it is possible to switch the accompaniment to an appropriate accompaniment for the melody.

In the present invention, similarly to the first invention, power data and pitch data are used as information for determining the position of an input melody. First, use the above method to detect whether or not a melody has been input using power data and pitch data, and then check the correspondence between the pitch and data of the input melody and the pitch of the stored melody information. and find the input melody position. How to request a response:
A method of converting pitch data into interval data by the method described in Japanese Patent Application No. 63-46118, and then comparing it with the interval series of the retained melody data to find the position with the most similar melody, or There is a method of converting the pitch series of the retained melody into pitch data and performing a similar comparison to find the position.

During a performance, while playing the accompaniment data, it is determined whether a melody input is expected at that time, and it is determined whether the actual melody input matches the determination. At that time,
If there is no melody input even though there should be no melody input, find the position of the input melody in the song using the above method, stop the previous rt-accompaniment performance, and start the accompaniment performance at the found position. Switch. The accompaniment data is written in a format that allows the data to be referenced depending on the desired position to start the accompaniment performance, so that the accompaniment performance can be started from any position.

As described above, by analyzing the input melody and controlling the way the melody and accompaniment are played depending on whether the melody is input at the expected position, it is possible to eliminate unsamples that may be difficult to hear due to the error of the person inputting the melody. Therefore, it is possible to realize an automatic accompaniment device that can reduce noise.

(Example) Examples of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of an automatic accompaniment device according to a first aspect of the present invention.

The melody storage unit 3 stores melody information to be input in advance. The melody information represents the pitch and length of each note of the melody in chronological order.

If the melody to be input is shown, for example, on the five-line staff in FIG. 3, the melody information can be expressed as shown in FIG. Here, the length of a note is expressed as a numerical value when the length of an eighth note is set as 1.

Further, the accompaniment data storage section 4 stores accompaniment data for the melody in the melody storage section 3. Like the melody information, the accompaniment data is expressed as information such as the serial number of the note, pitch, note length, etc., and is converted into a format that can be automatically played under the control of the performance control section 5. Further, in the C1 performance data storage section 4, a position where a melody is to be input with respect to accompaniment data is described.

FIG. 5 shows an example of a description of a melody position instruction in accompaniment data. In FIG. 5, information as to whether a melody is included in each measure of the accompaniment data and, if a melody is included, the melody position is shown by the measure number of the melody. A value of -1 indicates that no melody is included in that bar, and any other value indicates that the melody of that bar number in the melody data in the melody storage section should be input.

An example of operation will be described below. First, when an external performance start command is input to the performance control section 5, the performance control section 5 sequentially reads accompaniment data from the accompaniment data storage section 4.
A melody input person who performs an accompaniment performance according to the data inputs a melody at a position where melody input is expected in accordance with the accompaniment performance. On the other hand, the melody input means always receives the acoustic signal from the start of the accompaniment performance.

The input acoustic signal is sent to the melody detection section 2, and the A/
The waveform data is sequentially converted into waveform data by D conversion, and then the waveform data is converted into time-series power data and vibration data by analysis using an autocorrelation method.

Next, it is possible to determine whether a melody has been input, for example, by determining whether the accumulated value of power data over a certain period of time exceeds a threshold value. In other words, the power data is P" P L + 921...
・pt,...2 guns, constant time is S, threshold is E
If th is a certain interval Mt (from time t-S to t)
If the cumulative power value Et=ΣPi is larger than Eth (Et>Eth), it is determined that a melody exists in section mMt. The melody detection section 2 sends this determination result as melody detection information to the performance control section 5 at regular intervals. Here, S is the time of one bar, and the interval at which Et is calculated is also one bar. That is, for each bar, detection information is obtained as to whether or not a melody has been input within that bar, and the result is sent to the performance control section 5.

The performance control unit 5 obtains melody detection information for the previous bar while playing the accompaniment, compares it with melody position information in the accompaniment data storage unit 4, and determines whether the presence or absence of the melody matches. Considering the example of FIG. 5, the presence or absence of the assumed melody can be known by the value (-1 if there is no melody, otherwise melody present), so the reference result and the output of the melody detection unit 2 can be known. All you have to do is to compare them, and if they match, only the accompaniment performance will continue. However, if a melody is not detected even though a melody input is expected, the melody data corresponding to the bar number of the accompaniment data that is currently being read is read from the melody storage section 3, and f! Start playing the melody with the r sound.

Additionally, if the melody detection information matches the expected melody, but the melody is being played,
To stop playing a melody even though melody input has started.

Next, the second invention of the present invention will be explained.
FIG. 2 is a block diagram showing an embodiment of the automatic accompaniment device according to the second invention. Similar to the first invention, the melody storage section 3 describes the pitch and length information of the melody to be input, and the accompaniment data storage section 4 describes performance data necessary for accompaniment to the melody and the corresponding melody position. There is.

When a performance start command is input to the performance control section 5 from the outside,
The performance control section 5 sequentially reads accompaniment data from the accompaniment data storage section 4 and performs an accompaniment performance.

Similar to the first invention, the melody detection unit 2 sequentially captures input audio signals, performs A/D conversion and analysis to obtain power data and bit data, determines whether a melody has been input, and determines whether or not a melody has been input. The result is sent to the performance control section 5 as melody detection information.

While playing the accompaniment, the performance control unit 5 obtains melody detection information, compares it with the melody position information in the accompaniment data storage unit 4, and determines whether the presence or absence of the melody matches in the same manner as in the first invention. Make a judgment. As a result, if a melody is input at a position where a melody should not be input, a request signal is sent to the melody position detection section 6, and the position of the input melody in the song is determined.
The result is sent to the performance control section 5.

When the melody position detection section 6 receives the request signal from the performance control section 5, the melody position detection section 6 calculates the position of the input melody from the power data and pitch data obtained by the melody detection section and the melody data in the melody storage section 3. To find the 0 position, please refer to the above mentioned patent application No. 63-46118.
After converting the pitch data into interval data using the method described in , the melody sequence stored in the melody storage unit 3 is compared with the melody sequence to find a position with a similar melody.Here, melody detection An example will be explained in which information is obtained in units of one measure.

In response to a request from the performance control section 5, the pitch data for the previous measure is sent from the melody detection section 2 to the melody position detection section 6, and the pitch data is converted into pitch data l-.

Next, the degree of similarity between melodies is calculated between the human-generated melody (1)-9 data for one measure and the melody data for each measure in the melody storage section, and the position of the melody for one measure that is similar to the input melody is calculated. seek. Since the method of finding similar melodies based on the degree of similarity between melodies is known, the explanation thereof will be omitted here. Here, it is not necessary to calculate the similarity for all measures of the melody data, and the processing can be made more efficient by setting some measures that are likely to be input incorrectly in advance.
As described above, when the position of the input melody in the song is determined, it is sent to the performance control section 5 as melody position information.

The performance 11m section 5 determines the position of accompaniment data corresponding to the determined melody position, and switches to accompaniment performance from that position. Considering Fig. 5, for example, if the measure number of the accompaniment currently being played is 3 and the melody position information is melody position [1], then the person inputting the melody starts singing the melody at position 1 two measures earlier. , and switches to playing from the accompaniment data two measures ahead.

(Effects of the Invention) As explained above, in the automatic accompaniment device according to the first aspect of the present invention, even if the person inputting the melody forgets the position of inputting the melody, the person inputting the melody can notice the error at an early stage. Also, during that time, not only the accompaniment but also the melody is played, which reduces the unnaturalness of the ensample. Furthermore, since the melody performance serves as a guide, it is possible to easily start playing the original melody even in the middle.

Furthermore, in the second aspect of the present invention, even if the person inputting the melody starts the melody at a position where the melody does not fit, it is possible to switch to fP performance that matches the melody. It is possible to reduce hearing difficulty.

FIG. 3 is a diagram showing an example of a melody to be inputted using a 5-line staff; FIG. 4 is a diagram showing an example of expressing a part of the melody in FIG. 3 as melody information; FIG. 5 2 is a diagram showing an example of describing melody positions in the accompaniment data storage section 4. FIG.

1... Melody input means, 2... Melody detection section,
3... Melody storage section, 4... Accompaniment data storage section,
5... Performance control section, 6... Melody position detection section.

Claims (2)

[Claims] (1) A melody input means for inputting a melody as an acoustic signal; a melody detection unit that analyzes the acoustic signal input by the melody input means to obtain melody detection information indicating whether a melody is input; A melody storage section that stores the pitch and length of the melody as melody information, and accompaniment data that stores the musical information necessary for playing the accompaniment as accompaniment data, and describes the position in the song where the melody should be input as the melody input position. the melody storage section while determining whether or not the melody is input as specified by the melody input position from the melody detection information obtained by the storage section and the melody detection section and the melody input position of the accompaniment data storage section. 1. An automatic accompaniment device comprising: a performance control unit that reads melody information from a music player and performs a performance in combination with the melody information and accompaniment data in the accompaniment data storage unit. (2) a melody input means for inputting a melody as an acoustic signal; a melody detection unit that analyzes the acoustic signal input by the melody input means and obtains melody detection information indicating whether a melody is input; A melody storage section that stores the pitch and length of the melody as melody information, and accompaniment data that stores the musical information necessary for playing the accompaniment as accompaniment data, and describes the position in the song where the melody should be input as the melody input position. a melody position detection section that detects the position of the input melody from the information obtained by the melody detection section and the melody information in the melody storage section; 1. An automatic accompaniment apparatus comprising: a performance control section that reads accompaniment data from the accompaniment data storage section and performs a performance.