JP2000242285A - Karaoke system capable of measuring compass by computer and reflecting the measured result - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a user to enjoy KARAOKE with an easy-to-sing key matching with the compass of user's voice by measuring the compass before the user sings and calculating the best key by an arithmetic unit for every singing. SOLUTION: The user measures the compass interactively over a look at a CRT 18 on which the instruction from the arithmetic unit 11 is displayed. The measured result of the compass is stored in a storage device 12 in the form of a compass function. The arithmetic unit 1 extracts a histogram of the pitch distribution of the channel of MIDI data that a vocal sings. The arithmetic unit 11 calculates a cross correction function from the compass function and a histogram of the pitch of a sung song. In the calculation of the cross correction function, the compass function is fixed and the pitch of the sung song is shifted to calculate a correlation value. The arithmetic unit 11 transposes the key of the sung song to the pitch having the largest correlation value. The arithmetic unit 11 outputs the MIDI data of the sung song transposed to the key which is thus calculated to a music synthesizing device 13.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a karaoke system for outputting accompaniment suited to a vocal range.

[0002]

2. Description of the Related Art In a conventional karaoke system, accompaniment is output with a predetermined key. Therefore, if the key does not match the user's voice range, the user cannot sing. Therefore, singing forcibly often damages the throat such as vocal cord polyps. Therefore, a remote control or a main body provided with a key changing function, and a key that can be freely changed in real time during a performance is widely used. There is also a system in which the pitch of a user's voice is extracted in real time during a performance and an accompaniment is matched to the pitch.

[0003]

However, according to the prior art described above, the method of changing keys in real time during a performance has the following problems. (B) It is difficult to listen to the changed key and match the singing voice. (B) Even if you can sing temporarily, you may not be able to sing if the key is raised in the middle of a song such as rust. (C) Gives an unpleasant feeling to the listener. Therefore, the present invention solves the above problem by calculating an optimum key for the user before playing.

[0004]

According to a first aspect of the present invention, there is provided a karaoke system which calculates an optimum key for a user before performing. The present invention calculates the key before playing using the following means. (1) Voice range is measured for each user, and the result is stored in a storage device. (2) When a song to be sung is selected, the pitch distribution of the melody of the song is extracted. (3) From the vocal range information and the pitch distribution, a key that the user can easily sing is calculated by a cross-correlation function. (4) The accompaniment is modulated and output to the key. Further, the invention of claim 2 realizes the above-mentioned vocal range measurement by causing a user to sing a certain song and inputting its impression. The invention according to claim 3 is to determine whether or not the vocal range measurement according to the first aspect is made to be heard by a user at a sound of a certain pitch, and the sound is made to be uttered by a microphone at a loud and small volume, and the pitch of the sound is extracted to be able to utter. Has been realized.

[0005]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention is shown in FIG. Each device constituting the present invention will be described. 1
1 is an arithmetic unit for extracting a pitch from an audio signal, extracting a pitch histogram from MIDI data of a singing song, storing a result of vocal range measurement in a storage device and reading it when necessary, and selecting an optimal key for a user. Calculation processing such as calculation is performed. Reference numeral 12 denotes a storage device for storing a result of the vocal range measurement. Reference numeral 13 denotes a music synthesis device which calculates a karaoke accompaniment sound signal from the MIDI signal sent from the arithmetic device. A mixer 14 combines the singing voice of the user input from the microphone with the karaoke accompaniment input from the music composition device 13 and outputs the synthesized voice to the speaker 2. 15 is a microphone,
Input a voice signal from the user. 16 is an input device,
Enter singing songs and key changes. Reference numeral 17 denotes a music data input device, such as a modem, DVD, or C connected to the host station.
D, etc., to input MIDI data of the song. 18
Is a CRT, on which lyric information during singing and instructions for voice range measurement are displayed. Reference numeral 2 denotes a speaker, which outputs a singing voice and accompaniment.

The operation of the present invention is as follows. First, the user operates the CRT 18 on which an instruction from the arithmetic unit 11 is displayed.
And measure the vocal range interactively. The specific measurement procedure will be described later. The result of the vocal range measurement is stored in the storage device 12. The result of the vocal range measurement is stored in the form of a vocal range function uniquely defined in the present invention. The vocal range function is a function indicating the ease of singing at each pitch. A point where a certain pitch is most easily sung is set to 100 points, and a point where no singing is possible is set to 0 points. The MIDI data of the song selected by the input device 16 is read into the arithmetic unit 11 from the music data input device. Arithmetic unit 1
1 extracts a histogram of the pitch distribution of the channel sung by the vocal of the MIDI data. Specifically, MI
Extract NoteOn and NoteOFF events of DI. FIG. 2 shows an outline of calculating a key suitable for a user.
The arithmetic device 11 calculates a cross-correlation function (3) from the vocal range function (1) and the pitch histogram (2) of the song.
In the calculation of the cross-correlation function, the vocal range function is fixed, and the correlation value is calculated by shifting the pitch of the song. The arithmetic unit 11 transposes the key of the song to the pitch having the maximum correlation value. This is because a key that is most likely to sing is calculated when many notes go to a pitch having a large vocal range function, that is, a pitch that is easy to sing. The arithmetic unit 11 outputs the MIDI data of the song tuned to the key calculated in this way to the music synthesizing unit 13. The music synthesizer 13 creates a sound signal of karaoke accompaniment from the MIDI data,
Send to The mixer 14 synthesizes the accompaniment sent from the music synthesizer 13 and the audio signal from the microphone 15, and
Output to

As an embodiment of the vocal range measurement, a vocal range measurement method in which a user sings a certain piece of music and inputs his / her impression is described. The arithmetic unit 11 reads the sample music for measurement from the music data input device 17. The arithmetic unit 11 is a CRT 18
Shows the range of the sample song. The user sings a sample song in that range. The user inputs an impression of whether or not the throat is tired by singing from the input device 16 to the arithmetic device 11. The impression is selected from the following in three stages. (1) I can't sing because it doesn't fit my vocal range. (2) I can sing but some throats get tired. (3) I can sing without difficulty. Arithmetic unit 11
Adds a vocal range function from the impression information and the range of the sample music. The score to be added is the highest score in the center of the sound range in the range of the pitch within the sound range of the sample music. If the impression is 1, no points are added. When the impression is 2, the upper limit and the lower limit of the range of the sample music are set to 0 points, and the center pitch of the range is set to 50 points. The vocal range function is defined by the following equation as shown in FIG.

(Equation 1) Here, p is the pitch, S (p) is the vocal range function at the pitch p, H and L are the upper and lower limits of the range of the sample music, and M is the center pitch of the range. When the impression is 3, the upper limit and the lower limit of the range of the sample music are set to 50 points, and the center pitch of the range is set to 100 points. The vocal range function is defined by the following equation as shown in FIG.

(Equation 2) Here, p is the pitch, S (p) is the vocal range function at the pitch p, H and L are the upper and lower limits of the range of the sample music, and M is the center pitch of the range. Add S (p) to the vocal function. If the previously added score overlaps, the maximum value of the previous score and S (p) is taken (see FIG. 5). After adding S (p) to the vocal range function, the arithmetic unit 11 or the user shifts the range of the sample music and measures again. With the above repetition, the user and the arithmetic device 11 interactively create a vocal range function.

As an embodiment of vocal range measurement, a vocal range measuring method is described in which a user hears a sound of a certain pitch, the sound is uttered by a microphone at each loudness and loudness, and the pitch of the voice is extracted to determine whether or not the utterance can be made. . The arithmetic unit 11 allows the user to hear a certain pitch from the speaker 2. The user speaks the pitch from the microphone 15. The arithmetic device 11 extracts the pitch from the audio signal input from the microphone 15. The arithmetic unit 11 compares the pitch heard by the user with the pitch of the audio signal, and determines whether the utterance has been correctly made. When the utterance can be made, the arithmetic unit 11 raises the pitch,
Expand the range of sounds that can be uttered. The vocal range is defined as the point at which speech cannot be made. At the same pitch, the user utters both a low volume and a high volume, and when both can be uttered, the pitch can be uttered. This is because, as a vocal problem for many people, the pitch can not be obtained with a small voice, and singing at a pitch that cannot be pronounced with a small voice can hurt the throat. The determination as to whether the pitch has been uttered is performed as follows. The arithmetic unit 11 extracts the utterance section of the audio signal from the volume of the audio signal. Separation is done by measuring the volume when not speaking in advance, and setting the volume to about three times that as the threshold,
Separate audio signals. The arithmetic unit 11 divides the voice signal in the utterance section into short-section windows, and extracts the pitch in each window. Pitch extraction is performed using, for example, an autocorrelation function. The arithmetic unit 11 calculates the number of windows corresponding to the designated pitch,
If the number of windows equal to or more than 50% of the number of windows in the utterance section matches,
The utterance is a success. Here, the reason why the value is set to 50% is that the utterance may be vibrato, and a pitch shifted by a semitone from the designated pitch may be extracted. The vocal range is defined from the lowest sound that can be uttered to the highest sound. The vocal range functions include, for example, the following. FIG. 6 shows the vocal range function having a high score at the center of the vocal range. The formula is as follows:

(Equation 3) Here, p is the pitch, and V (p) is the value of the vocal range function at the pitch p. H and L are the upper and lower limits of the vocal range, and M is the center pitch of the vocal range. Also, measure the range of the spoken voice,
FIG. 7 shows the range with a high score. The formula is as follows:

(Equation 4) Here, p is the pitch, and V (p) is the value of the vocal range function at the pitch p. H and L are the upper and lower limits of the vocal range. B and A are the upper and lower limits of the range of the speaking voice, and C is the pitch at the center of the range of the speaking voice.

[0009]

As described above, the user measures the vocal range before singing, and the arithmetic unit 11 calculates the optimum key for each singing song. Can be enjoyed. The present invention alleviates the burden on the throat caused by singing with an unreasonable key, which is a problem of the conventional karaoke.

[Brief description of the drawings]

FIG. 1 shows an embodiment of the present invention.

FIG. 2 shows a method of calculating a key suitable for a voice range.

FIG. 3 shows a vocal range function of an impression 2 of vocal range measurement using a sample music piece.

FIG. 4 shows a vocal range function of an impression 3 of vocal range measurement using a sample song.

FIG. 5 shows a vocal range function of vocal range measurement using a sample song.

FIG. 6 shows a function weighted at the center among voice range functions of voice range measurement by utterance of one sound.

FIG. 7 shows a function in which a voice range is weighted among voice range functions of voice range measurement based on utterance of one sound.

[Explanation of symbols]

 Reference Signs List 11 arithmetic unit 12 storage device 13 music composition device 14 mixer 15 microphone 16 input device 17 music data input device 18 CRT 2 speaker

Claims (3)

[Claims] 1. A karaoke system wherein a karaoke system is characterized in that the result of vocal range measurement and the pitch distribution of a singing song are transposed to an optimal accompaniment suited to the vocal range by a cross-correlation function and output. 2. A vocal range measuring method in which a user sings a song and inputs his / her impression. 3. A vocal range measuring method in which a user hears a sound of a certain pitch, utters the sound at a volume of each of large and small, and extracts a pitch of the voice to determine whether or not the utterance was possible.