JP2001209390A - System representing candidate for choice and receiving reservation of selected melody in karaoke device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system capable of identifying individuals according to voiceprint data of more users than ever, representing candidates for choice, and receiving reservation of selected melodies in a KARAOKE device having only a limited storage capacity. SOLUTION: This system collects voiceprint data of singers when playing KARAOKE and classifies the collected voiceprint data into several patterns by a voiceprint classifying means; makes the IDs of the KARAOKE melodies, which are being performed when the voiceprints are collected, correspond to the classification to which the voiceprint belong, and records the IDs in a voiceprint/melody music recording book; a recorded music ID group, made to correspond to the classification to which the voiceprint data obtained in a melody choice mode belong, from the voiceprint/melody music recording book; and displays a list of the names of music corresponding thereto for letting the users choose melodies from them for reservation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for presenting song selection candidates in a karaoke apparatus and accepting a reservation for song selection. More particularly, the present invention relates to karaoke songs recorded in association with customer's voiceprint data obtained from a microphone. And a system for receiving a music selection reservation.

[0002]

2. Description of the Related Art For example, as disclosed in Japanese Patent Application Laid-Open No. 9-146563, a personal ID of a customer is specified by a voiceprint of the customer, and personal information such as a singing history and a scoring result is centrally managed. A karaoke system which is connected to a database via a communication line, searches the database for the personal information of the customer using the specified personal ID as a key, and provides various services such as displaying the music selection history and the scoring result of the customer. Devices are known.

[0003] On the other hand, the present inventors have also provided customers (users) with a view to more efficiently providing such services to customers.
The voiceprint data and singing history information of the customer specified by the personal ID obtained from the host device are downloaded from the host device, and the voiceprint data of the customer input to the microphone is collated with the previously downloaded voiceprint data, and the encoded voiceprint data is encoded. Japanese Patent Application No. 10-187443 has already disclosed a communication karaoke performance terminal which displays a music selection history associated with a music karaoke and allows the customer to make a music selection reservation.

[0004]

By the way, it goes without saying that there is an upper limit to the capacity of the storage area that can be provided in the karaoke apparatus, and the number of the above-mentioned voiceprint data and singing history information that can be stored in the karaoke apparatus at one time is limited. . for that reason,
In the case where a large number of users use the service at once, there is a risk that the above-described uploading and downloading will be performed many times during the use of karaoke, causing the place to disappear during karaoke, The cost of these communications is not ridiculous.

[0005] Of course, these problems may be solved by changing the hardware of the karaoke apparatus to increase the storage capacity. However, karaoke operators usually operate and manage a large number of karaoke devices installed in various places, and changing all of these hardware requires a great deal of labor and cost, so it is not always a reality. This is not a simple measure.

The present invention has been made in view of such circumstances, and in a karaoke apparatus having only a limited storage capacity, a karaoke apparatus capable of performing individual identification based on voiceprint data of more users than before. The present invention aims to provide a system that presents music selection candidates and receives a music selection reservation.

[0007]

According to a first aspect of the present invention, there is provided a karaoke apparatus for presenting a tune selection candidate and accepting a tune selection. First voiceprint detection means for collecting voiceprint data of a singer based on a voice signal from a singing microphone, and voiceprint classification for classifying the voiceprint data obtained by the first voiceprint detection means into several types. Singing history recording means for recording a song ID of a karaoke song being played when the voiceprint data was collected in a voiceprint / singing song record book in association with the classification to which the voiceprint data belongs, and a song selection mode A second voiceprint detecting means for sometimes collecting voiceprint data based on a voice signal from a microphone, and the voiceprint data obtained by the second voiceprint detecting means is classified by the voiceprint classification. A music ID group is read from the voiceprint / singing music record book, which is recorded in association with the classification to which the voiceprint data belongs, and based on the music ID group, the names of the music are listed. And a music selection reservation receiving means for receiving an input for specifying one of the music title groups displayed as the music selection candidates and subscribing the corresponding music ID to the music selection reservation list by this means. I will prepare it.

According to a second aspect of the present invention, in the system according to the first aspect, the voiceprint classification means classifies the voiceprint data into a maximum of N types,
If the singing history recording means matches the voiceprint data with the voiceprint data belonging to one of the already existing classifications, the singing history recording means determines the song ID of the karaoke music that was being performed when the voiceprint data was collected by the classification. Is recorded in the voiceprint / singing song record book, and the collected voiceprint data is not the same as any voiceprint data belonging to the already existing classification, and the number of already existing classifications is less than N. In this case, a classification to which the collected voiceprint data belongs is provided, and the classification is associated with the music ID of the karaoke tune played when the voiceprint data was collected, and recorded in the voiceprint / singing music record book, If the collected voiceprint data is not the same as any of the voiceprint data belonging to the already existing classification, and if N classifications already exist, the predetermined similarity judgment is performed. The algorithm determines the similarity between the voiceprint data belonging to each category and the voiceprint data belonging to another category, and the two categories to which the two voiceprint data determined to be most similar belong belong to one category. After integrating and reducing the number of already existing classifications, the classification to which the collected voiceprint data belongs is provided, and the classification is associated with the music ID of the karaoke music that was being performed when the voiceprint data was collected. It is a means for recording in a voiceprint / singing music record book.

According to a third aspect of the present invention, there is provided the system according to the second aspect, wherein each song ID recorded in the voiceprint / singing song record book has a song ID. A counter indicating the number of times the karaoke song corresponding to the singing is sung is attached, and the singing history recording unit deletes the voiceprint data when the collected voiceprint data matches the voiceprint data belonging to any of the existing classifications. When the song ID of the karaoke song played at the time of collection is recorded in the voiceprint / singing song record book in association with the classification, the same song ID as the song ID has already been recorded in association with the classification. In this case, the unit is configured to update the value of the counter attached to the music ID without recording the music ID.

According to a fourth aspect of the present invention, in the system according to any one of the first to third aspects, the singing history recording means associates the music ID. If M or more song IDs have already been associated with the classification of, some of the already recorded M song IDs are deleted by a predetermined deletion algorithm to reduce the number of song IDs. It is a means for recording the music ID in the voiceprint / singing music record book after the number is reduced.

According to a fifth aspect of the present invention, in the system according to the third or fourth aspect, the deletion algorithm gives priority to a music ID having a small counter value. It shall be deleted.

According to a sixth aspect of the present invention, there is provided the system according to any one of the third to fifth aspects, wherein each music recorded in the voiceprint / singing music record book. The ID is accompanied by the last date on which each song ID was recorded or the counter was updated, and the deletion algorithm preferentially deletes the song ID with the oldest date and time. .

Further, the invention according to a seventh aspect of the present invention is the system according to any one of the first to sixth aspects, wherein the voiceprint / singing music record book is one-to-one in the classification.
N having a storage area for a predetermined number of voiceprint data corresponding to
Singing history recording means, if there is a file in which the same voiceprint data as the collected voiceprint data is recorded, the singing history recording means was performed when the voiceprint data was collected in the file. If the song ID of the karaoke song is recorded and there is no file in which the same voiceprint data as the collected voiceprint data is recorded, and a file in which the voiceprint data is not recorded remains,
The voiceprint data is recorded in the file, and the song ID of the karaoke song that was being played when the voiceprint data was collected is recorded, and there is no file in which the same voiceprint data as the collected voiceprint data is recorded. ,And,
If there is no remaining file in which no voiceprint data is recorded, the similarity between two files recorded in each file is determined by a similarity determination algorithm between the files, and the files are determined to be most similar. The two files in which the two voiceprint data are recorded are specified, the recorded contents of one file are transcribed to the other file, and the voiceprint data and music ID recorded in the transcribed file are also recorded. Is deleted to generate an unregistered file in which voiceprint data is not recorded, and the file ID of the karaoke song that was being played when the voiceprint data was collected is recorded in this file.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS === Karaoke Apparatus === FIG. 1 shows a schematic configuration of a karaoke apparatus according to an embodiment of the present invention, which presents music selection candidates and receives a music selection reservation. The karaoke apparatus is mainly composed of a main computer 3 comprising a CPU, ROM, RAM, etc., which controls the whole apparatus. The karaoke apparatus is provided with a voice control device 7, a video control device 9, a hard disk 11, user interface 13,
Various peripheral devices such as a voiceprint extraction device 15 are connected.

The voice control device 7 outputs voice signals input to a singing microphone, karaoke accompaniment music, voice signals from the video control device 9 and the like from a speaker. The image control device 9 outputs a lyrics image, a background image, and the like to a display. The hard disk 11 stores accompaniment music data and lyrics video data in a data format such as a MIDI code, and karaoke data 19 including attribute data such as song names and singer names in association with song IDs. The user interface 13 has a function of transmitting an operation input signal transmitted from various operation input devices such as an operation panel and a remote control device to the main computer 3. The voiceprint extraction device 15 is connected to the voice control device 7 and performs processing such as vector quantization and HMM (Hidden Markov Model) on an input voice signal sent from the voice control device 7 and includes data components to be described later. It has a function of generating voiceprint data and transmitting the voiceprint data to the main computer 3.

=== Operation of Karaoke Device === When a user performs a reservation operation using an operation panel or a remote control device, the user interface 13 sends the song ID of the karaoke song specified by the user to the main computer 3. introduce. The main computer 3 records the music ID transmitted from the user interface 13 in the reservation table 21 secured on the RAM.

The main computer 3 has a reservation table 2
The karaoke data 19 corresponding to each music ID recorded in 1 is sequentially read out from the hard disk 11, and the voice control device 7 is driven in accordance with the accompaniment music data included therein to reproduce and output the karaoke accompaniment music to the speaker. A lyric image synchronized with karaoke accompaniment music is output to a display based on the lyric image data included in the karaoke data 19. Further, the video control device 9 superimposes the background video sent from the connected video playback device 17 such as a DVD device on the lyrics video and outputs it to the display.

=== Music Selection Reservation Function from Voiceprint === In addition to the normal music selection reservation function described above, this karaoke apparatus acquires voiceprint data from a voice input signal input to a microphone, and acquires the voiceprint data. Is displayed on the display, and a function is provided for operating in a music selection reservation mode using a voiceprint for allowing the user to make a music selection from the list. Here, the voiceprint data is data composed of three voiceprint parameters of a formant level L, a formant center frequency F, and a formant bandwidth W, which are characteristic amounts related to the resonance frequency of the vocal tract.

On the other hand, the association between the voiceprint data and the karaoke music is stored and managed in N files on the hard disk 11. FIG. 2 shows the data structure of this file. One or two voiceprint data and a plurality of music IDs are recorded in each file, and the music ID further includes a counter indicating the number of times a karaoke music corresponding to the music ID has been selected in the past, Data indicating the last date on which the karaoke song corresponding to the ID was sung (the date on which the song ID was recorded in the file or the last update date of the counter) is attached. Up to M music IDs can be recorded in each file. Also,
The value M and the value of the number of files N described above are set to appropriate values in consideration of the storage capacity of the hard disk 11.

In the music selection reservation mode, the karaoke apparatus operates as follows. When using the song selection reservation mode,
The user performs a predetermined operation on the user interface 13 to start use in this mode. When recognizing that the use start operation has been performed, the main computer 3 displays a predetermined message on a display and urges the microphone to speak a predetermined word. When the user responds and speaks into the microphone,
The main computer 3 drives the voice print extracting device 15 to obtain voice print data from the voice signal input to the microphone.

Next, the main computer 3 stores the voiceprint data corresponding to the acquired voiceprint data in the hard disk 1.
1 is checked by a comparison algorithm described later. As a result, if there is a matching voiceprint data, the song name and the singer name corresponding to the music ID associated with the voiceprint data are entered in the karaoke data 19.
And displays a list listing these on the display, and prompts the user to perform a music selection reservation operation from the list.
When the user operates the operation panel or the remote controller to select a music ID from the list, the main computer 3 records the music ID in the reservation table 21.

=== Comparison Algorithm === The comparison algorithm described above will be described in detail. The comparison algorithm is as follows, when the formant levels of certain voiceprint data a and b are La and Lb, the center frequency of the formant is Fa and Fb, and the bandwidth of the formant is Wa and Wb, R = | La-Lb | + | Fa-Fb | + | Wa-Wb | (where || is an absolute value) is calculated, and if this value is smaller than a certain value Rc, the two voiceprint data a and b are the same. It is determined that there is. On the other hand, when the value of R is equal to or greater than Rc, it is determined that the voiceprint data a and b are not the same voiceprint data. Further, the comparison algorithm determines the degree of similarity based on the value of the similarity R, and determines that the smaller the similarity R is, the more similar the two data are.

=== File Management === Next, the manner in which voiceprint data and music IDs are stored and managed by the files will be described with reference to the flowchart of FIG. When the performance of the karaoke music recorded in the reservation table 21 is started, the main computer 3 acquires the voiceprint data of the singer from the microphone (101 to 1).
02) The voiceprint data acquired here is compared with the voiceprint data recorded in each file by the above-described comparison algorithm. Then, it is checked whether the same voiceprint data as the obtained voiceprint data has already been recorded in any file (103). If the same voiceprint data is found as a result of the survey, the music I of the karaoke music being played is stored in the file in which the voiceprint data is recorded.
D is recorded (104 → 106 → 108).

At this time, if the music ID is already recorded in the file, the music ID being played is not recorded in the file, but the value of the counter attached to the music ID already recorded is recorded. Is incremented by 1 (105). When M music IDs are already recorded in the file, each music ID recorded in the file is recorded.
By checking the counter of D, the one with the smallest counter value and the oldest last date is specified, and its music ID is deleted (if there are a plurality of corresponding music IDs, all of them are deleted). After that, the music ID is newly recorded in the file (106 → 107 → 108).

On the other hand, the comparison (10
As a result of 3), if no voiceprint data corresponding to the obtained voiceprint data exists in any file, it is further determined whether there is any unregistered file on the hard disk 11 to which no voiceprint data is associated. Is checked (110). If there is an unregistered file, the voiceprint data and the song ID of the karaoke song being played are stored in the unregistered file (if there are a plurality of unregistered files, any one of them). Is recorded (112). On the other hand, when there is no unregistered file, the main computer 3 executes a file merging process described later to create an unregistered file,
The voiceprint data and the music ID of the karaoke music being played are recorded in this (111 → 112).

=== File Merging Process === Next, the above-described file merging process will be described with reference to FIG. In this process, the main computer 3
Calculates the similarity R for all combinations of voiceprint data respectively recorded in two different files. If two voiceprint data are recorded in one file, the processing is performed for each of them (20
1). Next, two files having the smallest calculated similarity R, that is, two files in which the two most similar voiceprint data are recorded are specified (20).
2). Then, the music ID recorded in these two specified files is transcribed to one file, and the voiceprint data recorded in the transcription source file is transcribed to the transcription destination file. At this time, if two voiceprint data have already been recorded in the transcription destination file, the voiceprint data of the two files and the voiceprint data of the transcription source voiceprint data in total are respectively three voiceprint data of the voiceprint data. The similarity R is calculated for the combination and the most similar 2
One voiceprint data is specified, and the two specified voiceprint data are recorded in the destination file (203). Which file is set as the transcription source or the transcription destination is determined based on an appropriate algorithm, for example, setting the one with the smaller number of recorded music IDs as the transcription source. On the other hand, after the above processing, the main computer 3 reads the voiceprint data and the music ID recorded in the transcription source file.
Are changed to an unregistered file (an unregistered file is generated) (204).

According to the file merging process described above, it is possible to manage the association between the voiceprint data and the singing history while keeping the number of files at a fixed number N. Each time a file is merged, the number of music IDs recorded in the file increases, but the number of music IDs recorded in one file is limited to the maximum value M. Therefore, the occupation amount of the hard disk 11 by these files is kept below a certain value. Therefore, efficient use of the hard disk 11 is achieved, and in a karaoke apparatus having only a limited storage capacity, individual identification based on voiceprint data corresponding to a large number of users becomes possible.

Further, the singing histories of a plurality of persons are mixed in one file by the file merging process.
At the time of operation in the song selection reservation mode based on the voiceprint data described above, karaoke songs that have not been sung by the user at the time of song selection reservation will be displayed, but in general, people with similar voice qualities will be similar Since music is often selected, there is no particular inconvenience in karaoke operation by the file merging process. Also, by displaying a karaoke song sung by another person having a voice quality similar to that of the user, it is possible to expect an effect such that a new karaoke song matching the user's voice quality can be found.

=== Other Applications === By the way, when there are many music IDs with the minimum value of the counter, the music IDs may be deleted in order from the one with the oldest selection date.

The comparison algorithm is not limited to the one described here, but may be another algorithm as long as it can determine the same or similar voiceprint data.

[0031]

As described above, according to the system of the present invention, in a karaoke apparatus having only a limited storage capacity, it is possible to identify individuals with voiceprint data of a larger number of users than before. A system for presenting music selection candidates and accepting music selection reservations in the device can be configured.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a schematic configuration of a system for presenting music selection candidates and receiving a music selection reservation in a karaoke apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a data structure of a file for storing and managing voiceprint data and music ID in one embodiment of the present invention.

FIG. 3 is a flowchart illustrating a file merging process according to an embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 3 main computer 5 bus 7 audio control device 9 video control device 11 hard disk 13 user interface 15 voiceprint extraction device 17 video playback device

Claims (7)

[Claims] 1. A first voiceprint detecting means for collecting voiceprint data of a singer based on a voice signal from a singing microphone during a karaoke performance, and several voiceprint data obtained by the first voiceprint detecting means. Voiceprint classification means for classifying the voiceprint data, and singing for recording the music ID of the karaoke tune played at the time of collecting the voiceprint data in the voiceprint / singing music record book in association with the classification to which the voiceprint data belongs. History recording means, second voiceprint detection means for collecting voiceprint data based on a voice signal from a microphone in a music selection mode, and classifying the voiceprint data obtained by the second voiceprint detection means by the voiceprint classification means. Read out a song ID group recorded in association with the type to which the voiceprint data belongs from the voiceprint / singing song record book, based on the song ID group. Candidate selection display means for displaying the names of the songs in a list, and receiving an input for specifying one of the group of song names displayed as the candidate for song selection by this means, and assigning the corresponding song ID to the song selection reservation list. A system for presenting a song selection candidate and accepting a song selection reservation in a karaoke apparatus comprising a song selection reservation accepting means for joining. 2. The system according to claim 1, wherein the voiceprint classifying means classifies the voiceprint data into a maximum of N types, and wherein the singing history recording means includes the collected voiceprint data. If the voiceprint data belongs to any of the categories, the music I of the karaoke music that was being performed when the voiceprint data was collected.
D is recorded in the voiceprint / singing song record book in association with the classification, and the collected voiceprint data is not the same as any voiceprint data belonging to the already existing classification, and the number of the already existing classification is N. If the number is less than the number, the classification to which the collected voiceprint data belongs is provided, and the music ID of the karaoke music that was played when the voiceprint data was collected in this classification.
Are recorded in the voiceprint / singing music record book in association with each other. If the collected voiceprint data is not the same as any voiceprint data belonging to the already existing classification, and N classifications already exist, A predetermined similarity determination algorithm is used to determine the similarity between the voiceprint data belonging to each class and the voiceprint data belonging to another class,
After integrating the two classifications to which the two voiceprint data determined to be most similar belong into one classification to reduce the number of already existing classifications, a classification to which the collected voiceprint data belongs is provided, and this classification is assigned to this classification. A system for recording the voiceprint / singing music record book in association with the music ID of the karaoke music being played when the voiceprint data was collected. 3. The system according to claim 2, wherein each music ID recorded in the voiceprint / singing music record book.
A counter indicating the number of times the karaoke song corresponding to the song ID has been sung, and the singing history recording means matches the voiceprint data belonging to any of the existing classifications with the collected voiceprint data. When the song ID of the karaoke song played when the voiceprint data was collected is recorded in the voiceprint / singing song record book in association with the classification, the song ID
If the same music ID has already been recorded in association with the classification, the means is a means for updating the value of the counter attached to the music ID without recording the music ID. . 4. The system according to claim 1, wherein the singing history recording unit has already assigned M or more song IDs to the category to which the song ID is to be associated. If the number of music IDs has already been recorded, some of the M music IDs already recorded are deleted by a predetermined deletion algorithm to reduce the number of music IDs, and then the voiceprint / singing music record book of the music IDs is deleted. It is a means for performing recording on a. 5. The system according to claim 3, wherein the deletion algorithm preferentially deletes a music ID having a small counter value. 6. The system according to claim 3, wherein each song ID is recorded in the voiceprint / singing song record book, or the counter is updated. Is attached, and the deletion algorithm deletes the music ID with the oldest date and time in priority. 7. The system according to claim 1, wherein the voiceprint / singing music record has a storage area for a predetermined number of voiceprint data corresponding to the classification on a one-to-one basis. Singing history recording means, when there is a file in which the same voiceprint data as the collected voiceprint data is recorded, the song was recorded when the voiceprint data was collected in the file. The song ID of the karaoke song is recorded. If there is no file in which the same voiceprint data as the collected voiceprint data is recorded and a file in which no voiceprint data remains remains, the file The voiceprint data is recorded, and the music ID of the karaoke tune played at the time of collecting the voiceprint data is recorded, and the same voiceprint as the collected voiceprint data is recorded. If there is no file in which the data is recorded and there is no remaining file in which no voiceprint data is recorded, a similarity determination algorithm is performed between the two files for the two voiceprint data recorded in each file. To determine the two files in which the two voiceprint data determined to be most similar are recorded, and the recorded contents of one file are transcribed to the other file and become the transcription source. All contents such as voiceprint data and music ID recorded in the file are deleted to generate an unregistered file in which no voiceprint data is recorded, and the karaoke performed when the voiceprint data was collected in this file. It is a means for recording the music ID of the music.