WO2015137360A1

WO2015137360A1 - Singing analyzer

Info

Publication number: WO2015137360A1
Application number: PCT/JP2015/057063
Authority: WO
Inventors: 松本　秀一
Original assignee: ヤマハ株式会社
Priority date: 2014-03-10
Filing date: 2015-03-10
Publication date: 2015-09-17
Also published as: JP2015169867A

Abstract

A singing analyzer (100) is provided with an analysis processing unit (22) for specifying a comment (e.g., singing advice (A)) corresponding to a trend of reference voices in a group, from among a plurality of reference voices recorded beforehand, corresponding to the singing voice (V) of a subject singer and a presentation processing unit (24) for presenting the comment specified by the analysis processing unit (22) to the subject singer.

Description

Singing analysis device

The present invention relates to a technique for analyzing a singing voice.

A technology that uses the tendency of singing in the past by singers has been proposed. For example, in Patent Document 1, music that matches the preferences of individual singers by registering music in advance for each of a plurality of groups in which each singer is classified according to the tendency (ie, preference) of past song selection. Is disclosed to a singer.

Japanese Unexamined Patent Publication No. 2012-078387

The technique of Patent Document 1 is a technique that uses the tendency of music selection by each singer to propose music, but a tendency when a large number of singers sing a music (for example, many singers of music fail. Singing advice (pointing out or advice) taking into account the tendency of individual singers to sing a song (for example, pitch errors are likely to occur in the high range), or the above trends If comments such as the evaluation result of the added singing can be presented to the singer, an effective improvement of the singing can be expected.
In view of the above circumstances, an object of the present invention is to present an appropriate comment according to the singing tendency of the singer to the singer.

In order to solve the above problems, the singing analysis device of the present invention specifies a comment corresponding to the tendency of each reference voice of the group corresponding to the singing voice of the target singer among a plurality of reference voices recorded in advance. And a presentation processing unit for presenting the comment specified by the analysis processing unit to the target singer. In the above configuration, since the tendency of each reference voice of the group corresponding to the singing voice of the target singer is specified, a comment appropriate for the singing voice of the target singer can be presented to the target singer. Therefore, there is an advantage that the singing of the target singer can be effectively improved.

In one aspect of the present invention, the analysis processing unit specifies, as the comment, singing advice corresponding to a tendency of each reference voice of a group corresponding to the singing voice of the target singer. In the above aspect, since the singing advice according to the tendency of each reference voice of the group corresponding to the singing voice of the target singer is specified, the singing advice appropriate for the singing voice of the target singer is presented to the target singer. It is possible.

In one aspect of the present invention, the analysis processing unit refers to reference information that specifies singing advice for each of a plurality of groups that classify a plurality of reference sounds that are common to the singing voice of the target singer and the music, and includes a plurality of reference sounds. The singing advice of the group to which the singing voice of the target singer belongs is specified. In the above aspect, the singing voice of the target singer belongs by referring to the reference information that specifies the singing advice for each of a plurality of groups in which a plurality of reference voices in which the singing voice of the target singer and the music are in common are classified. Since the group singing advice is specified, there is an advantage that a suitable singing advice can be presented for each piece of music. Further, according to the configuration in which the analysis processing unit sequentially updates the group of the singing voice while the target singer sings the music, there is an advantage that suitable singing advice can be presented for each section of the target music. In addition, the specific example of each above aspect is later mentioned as 1st Embodiment, for example.

In one aspect of the present invention, the analysis processing unit refers to reference information that specifies singing advice for each music attribute according to the tendency of each reference voice of a group that has collected the voice of the target singer among a plurality of reference voices. Then, the singing advice of the part applicable to a music attribute is specified among the music which an object singer sings. In the above aspect, the singing advice of the location corresponding to the music attribute in the music is specified by referring to the reference information that specifies the singing advice for each music attribute according to the tendency of the plurality of reference sounds of the target singer. Therefore, there is an advantage that suitable singing advice (for example, advice for poor singing for each singer) can be presented for each target singer. For example, the reference information specifies singing advice with specific pitches of successive pitches as music attributes, and the analysis processing unit refers to a location where a specific pitch exists among songs sung by the target singer. According to the configuration that specifies the singing advice specified for the pitch in the information, effective singing advice to overcome the weakness is presented to the singer who is not good at singing that changes the pitch at a specific pitch Is possible. In addition, the specific example of the above aspect is later mentioned, for example as 2nd Embodiment.

In one aspect of the present invention, the analysis processing unit specifies, as the comment, an evaluation result obtained by evaluating the singing voice of the target singer according to the tendency specified by the analysis processing unit. In the above aspect, since the evaluation result of evaluating the singing voice of the target singer according to the tendency of each reference voice of the group corresponding to the singing voice of the target singer is presented, it is appropriate for the singing voice of the target singer. The evaluation result can be presented to the target singer.

By the way, if the singing of a singer is evaluated in consideration of the tendency when a large number of singers sing a song and the tendency of singing by individual singers, an effective improvement of the singing can be expected. In consideration of the above circumstances, the singing analysis apparatus according to another aspect of the present invention responds to the tendency of each reference voice of the group corresponding to the singing voice of the target singer among a plurality of reference voices recorded in advance. And an analysis processing unit for evaluating the singing voice. According to the above configuration, since the singing voice is evaluated (typically scored) according to the singing tendency of the singer, there is an advantage that an evaluation that can effectively contribute to the improvement of the singing can be realized.

It is a lineblock diagram of the song analysis device concerning a 1st embodiment of the present invention. It is explanatory drawing of reference information. It is explanatory drawing at the time of indication. It is a flowchart of a singing advice specific process. It is explanatory drawing of the reference information in 2nd Embodiment. It is a flowchart of the singing advice specific process in 2nd Embodiment.

<First Embodiment>
FIG. 1 is a configuration diagram of a singing analysis apparatus 100 according to the first embodiment of the present invention. The singing analysis device 100 is an information processing device for presenting advice (hereinafter referred to as “singing advice”) regarding the singing of music to a singer of the music (hereinafter referred to as “target singer”), The present invention is realized by a computer system including an arithmetic processing device 12, a storage device 14, a sound collection device 16, and a display device 18. The singing analysis apparatus 100 is suitably used as a karaoke apparatus that reproduces accompaniment sounds of music, for example.

The sound collection device 16 is a device (microphone) that collects ambient sounds. The sound collection device 16 of the first embodiment collects the singing voice V in which the target singer sang a specific music (hereinafter referred to as “target music”). Note that the synthesized voice synthesized by the voice synthesis technique can be used as the singing voice V. The display device 18 (for example, a liquid crystal display panel) displays an image instructed from the arithmetic processing device 12. In the first embodiment, the singing advice A of the target music is displayed on the display device 18. Specifically, at each time point during the song singing by the target singer, singing advice A suitable for the time point is sequentially displayed on the display device 18. In addition, it is also possible to output the singing advice A by sound from a sound emitting device (for example, a speaker).

The storage device 14 stores programs executed by the arithmetic processing device 12 and various data used by the arithmetic processing device 12. A known recording medium such as a semiconductor recording medium or a magnetic recording medium or a combination of a plurality of types of recording media is arbitrarily employed as the storage device 14. The storage device 14 of the first embodiment stores reference information DA for each of a plurality of music pieces. Each reference information DA is used to specify the singing advice A for the music.

FIG. 2 is an explanatory diagram of reference information DA for any one piece of music. As illustrated in FIG. 2, the reference voice group Q is used to generate the reference information DA. The reference voice group Q is a set of a plurality of singing voices (hereinafter referred to as “reference voices”) R recorded in advance. The plurality of reference sounds R included in the reference sound group Q are sounds in which an unspecified number of singers sang arbitrary music. As illustrated in FIG. 2, a plurality of reference voices (a plurality of reference voices with a common song song) R of an arbitrary piece of music is composed of N groups (N is a natural number of 2 or more) groups G [1] ˜ Classified as G [N]. One group G [n] (n = 1 to N) corresponding to an arbitrary musical piece includes a plurality of reference sounds R in which different singers sang the musical piece.

The method of classification (clustering) of each reference sound R is arbitrary, but a method of classifying a plurality of reference sounds R into N groups G [1] to G [N] is preferable from a musical viewpoint. Specifically, each reference voice R is set to N for each range of evaluation index (song scoring result) E which is a difference index between the melody of the singing part of the music and the reference voice R (for example, in increments of 5 points of 100 points). Depending on the method of grouping into groups G [1] to G [N] and the trend of evaluation index E calculated in time series within the music (for example, the tendency of evaluation index E to increase in the second half of the music) Thus, a method of classifying each reference speech R into N groups G [1] to G [N] can be employed. In addition to the above categorization by singing level, classification by gender (male and female), classification by singing level by gender, classification when singing alone, classification when singing with mixed voice, Classification etc. when singing in a group can also be adopted.

2, the reference information DA of the first embodiment includes N unit information U [1] to U [N] corresponding to different groups G [n] of the reference speech R. As representatively shown in FIG. 2 for unit information U [N], any one unit information U [n] is composed of a plurality of time points (hereinafter referred to as “pointed time points”) T (T1, T2,. ...) Singing advice A (A1, A2,...) Is designated for each. The content of the singing advice A is set individually for each indication time T.

The unit information U [n] of an arbitrary piece of music is generated in consideration of the musical tendency of each reference sound R classified into the group G [n] among the plurality of reference sounds R of the music. . Specifically, each time point at which the singing is to be improved is designated as a point in time T by a large number of reference sounds R of the group G [n] in the music, and the content of the improvement of the singing or advice or indication for improvement ( A character string expressing (suggestion) is designated as singing advice A.

FIG. 3 shows a time series of average pitches P [n] over a plurality of reference sounds R of group G [n] and a time series of exemplary pitches P0 of music. The exemplary pitch P0 is a time series of the pitches of each note specified in the musical score of the music, or a time series of the average values of the pitches of the reference sounds R of the group G having the maximum evaluation index E. . As can be understood from FIG. 3, the point in time when the difference (pitch error) between the average pitch P [n] of the group G [n] and the exemplary pitch P0 is maximized is designated as the indication time T. Then, a character string (for example, a message such as “Caution on the pitch!”) For improving the pitch error at the time point is designated as the singing advice A for each indicated time point T.

1 performs overall control of each element of the singing analysis device 100 by executing a program stored in the storage device 14. As illustrated in FIG. 1, the arithmetic processing device 12 according to the first embodiment has a plurality of functions (analysis processing unit 22 and presentation processing unit 24) for presenting singing advice A to a target singer who sings the target music. To realize. A configuration in which each function of the arithmetic processing device 12 is distributed to a plurality of devices, or a configuration in which a dedicated electronic circuit realizes a part of the function of the arithmetic processing device 12 may be employed.

1 specifies the singing advice A to be presented to the target singer. The analysis processing unit 22 of the first embodiment sequentially specifies singing advice A suitable for the singing voice V of the target singer during the singing of the target song. FIG. 4 is a flowchart of a process for the analysis processing unit 22 to specify the singing advice A (hereinafter referred to as “singing advice specifying process”). The singing advice specifying process of FIG. 4 is started with the start of the singing of the target music (reproduction start of the accompaniment sound of the target music).

When the singing advice specifying process is started, the analysis processing unit 22 determines whether or not the target music has been completed (SA1). When the target music has not ended (SA1: NO), the analysis processing unit 22 selects one of the plurality of sections (hereinafter referred to as “fixed length” or “variable length”) on the time axis. The singing voice V of “selected section” is acquired from the sound collection device 16 (SA2). The analysis processing unit 22 selects each section of the music from the beginning to the end in order every time step SA2 is executed, and acquires the singing voice V in the selected section.

The analysis processing unit 22 is a group (hereinafter referred to as “affiliation group”) G to which the singing voice V in the selected section belongs among the N groups G [1] to G [N] into which the plurality of reference voices R of the target music are classified. Is specified (SA3). Specifically, the analysis processing unit 22 calculates the evaluation index E for the singing voice V in the selected section, and among the N groups G [1] to G [N] corresponding to different ranges of the evaluation index E One group G [n] in a range in which the evaluation index E of the singing voice V in the selected section is included is specified as the belonging group G.

The analysis processing unit 22 selects the unit information U corresponding to the group G identified in step SA3 from the N unit information U [1] to U [N] of the reference information DA stored in the storage device 14. (SA4). That is, the analysis processing unit 22 identifies each singing advice A of the group G to which the singing voice V of the target singer belongs among the N groups G [1] to G [N]. When the unit information U of the selected section is specified by the above procedure, the analysis processing unit 22 moves the process to step SA1. Therefore, until the target music ends (SA1: YES), the belonging group G is sequentially updated for each section of the target music, and the unit information U (the time series of the singing advice A) corresponding to the updated belonging group G is obtained. It is specified sequentially. In addition, since the group G corresponding to the singing voice V is not specified at the stage where the first section of the target music is sung (the stage where the singing voice V is not acquired), each unit information U [n] of the reference information DA Singing advice A (for example, a general message such as “Let's sing with emotion”) prepared in advance regardless of the number is presented to the target singer.

1 presents the singing advice A identified by the analysis processing unit 22 in the singing advice identifying process exemplified above to the target singer. Specifically, the presentation processing unit 24 sings the unit information U for the indicated time T at a time that precedes each indicated time T specified by the unit information U specified by the analysis processing unit 22 by a predetermined time. The advice A is displayed on the display device 18. That is, the points to be improved under the respective reference sounds R of the group G to which the singing voice V belongs (that is, the places where the singing voice V of the target singer is supposed to be improved similarly) are sequentially given to the target singer. The target singer can sing with particular attention to a portion of the target music that is likely to fail.

As described above, in the first embodiment, the singing advice A corresponding to the tendency of the plurality of reference sounds R of the group G to which the singing voice V of the target singer belongs is presented to the target singer. That is, singing advice A suitable for the singing voice V of each target singer is presented to the target singer. Therefore, there is an advantage that the singing of the target singer can be effectively improved.

In the first embodiment, in particular, the singing advice A is specified by referring to the reference information DA that specifies the singing advice A for each group G [n] that classifies the plurality of reference sounds R that share the singing voice V and the music. Therefore, the singing advice A suitable for each piece of music is specified. Therefore, the above-described effect that the singing advice A appropriate for the singing voice V of the target music can be presented is particularly remarkable. Moreover, in 1st Embodiment, the affiliation group G of the singing voice V is updated sequentially for every area of a target music during the song of the target music by a target singer. Therefore, there exists an advantage that suitable singing advice A can be shown for every section of object music.

Second Embodiment
A second embodiment of the present invention will be described below. In addition, about the element which an effect | action and function are the same as that of 1st Embodiment in each form illustrated below, the reference | standard referred by description of 1st Embodiment is diverted, and each detailed description is abbreviate | omitted suitably.

The storage device 14 of the second embodiment stores the reference information DB of FIG. 5 instead of the reference information DA of the first embodiment. For the generation of the reference information DB, a set (group) of a plurality of reference voices R of the target singer in the reference voice group Q similar to the first embodiment is used. Specifically, by analyzing a plurality of reference voices R of the target singer in the reference voice group Q, the music attribute X (typically the target singer fails to present singing advice A to the target singer) The music attribute X) that tends to be identified is specified, and the reference information DB that specifies the singing advice A (A1, A2,...) For each of the plurality of music attributes X (X1, X2,. Are stored in the storage device 14. The reference information DB may be generated in advance for each of a plurality of singers, but the reference information DB of the target singer is generated immediately before the singing by the target singer (that is, for each singing). It is also possible.

“Music attribute” means a musical attribute (mode) of a music piece. Specifically, range (high / low), performance mark (A melody, rust, etc.), position within a specific section such as a phrase (protrusion, etc.), sound type (ascending, descending, continuation of same sound, kobushi, modification) Sound), note value (long tone, short passage), rhythm type, legato / staccato, tempo, beat position (second beat back, etc.), chord function (root, non-harmonic sound), etc. Included in the concept of “attribute”.
In the present invention, the music attribute X means a musical attribute (mode) of the song part of the song. For example, in the reference information DB of the target singer analyzed from a plurality of reference sounds R, the singing advice such as “careful of high sounds!” For the music attribute X1 “high range”, which is analyzed from a plurality of reference sounds R. A1 is specified. In the reference information DB of the target singer analyzed from a plurality of reference sounds R, the music attribute X2 of “5 degrees” is a tendency that the singing of each pitch that is in succession at a specific pitch (for example, 5 degrees) is not good. Singing advice A2 such as “Caution for pitch change!” Is specified for (specific pitch). In addition, in the reference information DB of the target singer who has analyzed the tendency that the singing of a specific rhythm is not good, the singing advice A3 such as “note rhythm!” Is specified for the music attribute X3 “specific rhythm”. In the reference information DB of the target singer in which the tendency that the singing of a specific section such as immediately after the start (start of singing) is not good is analyzed, the music attribute X4 “immediately after start” Singing advice A4 such as “!” Is designated.

FIG. 6 is a flowchart of the singing advice specifying process for the analysis processing unit 22 of the second embodiment to specify the singing advice A. Similar to the first embodiment, the singing advice specifying process in FIG. 6 is started when the singing of the target music starts.

When the singing advice specifying process is started, the analysis processing unit 22 refers to the reference information DB of the target singer, so that the portion corresponding to the music attribute X specified by the reference information DB (hereinafter referred to as “pointed section”). (Referred to as “)” (SB1). For example, when a specific sound range (for example, a high sound range) is designated as the music attribute X by the reference information DB, the section of the sound range of the target music is searched as the indicated section, and a specific pitch (for example, 5 degrees) is searched. Is specified in the reference information DB as the music attribute X, a section in which the pitch is around the target musical piece is searched for as the indicated section. When a specific rhythm is specified as the music attribute X in the reference information DB, the rhythm section of the target music is searched as the indicated section, and the specific section (for example, immediately after the start) is referred to as the music attribute X. If the information DB is specified, the section is searched as the indicated section in the target music. Note that it is possible to add a plurality of types of music attributes X to the search for the indicated section. For example, a section of “specific rhythm” and “specific pitch” is searched for as an indicated section.

The analysis processing unit 22 specifies the singing advice A for each indicated section searched from the target music by the above procedure (SB2). Specifically, the analysis processing unit 22 specifies the singing advice A corresponding to the music attribute X of the indicated section from the reference information DB for each of the plurality of indicated sections searched from the target music. The above is a specific example of the singing advice specifying process in the second embodiment.

The presentation processing unit 24 of the second embodiment presents the singing advice A specified by the analysis processing unit 22 in the singing advice specifying process described above to the target singer for each indicated section of the target music. Specifically, the presentation processing unit 24 displays the singing advice A specified by the analysis processing unit 22 for the indicated section on the display device 18 at a time point preceding the starting point of each indicated section in the target music by a predetermined time. Let As understood from the above description, singing advice A for improving the singing of the section is sequentially presented to the target singer in advance of the singing of the indicated section estimated to be weak for the target singer.

As described above, in the second embodiment, since the singing advice A corresponding to the group tendency corresponding to the singing voice of the target singer in the reference voice group Q is presented to the target singer, the first embodiment. Similarly, it is possible to present the singing advice A appropriate for each target singer to the target singer. In the second embodiment, in particular, reference information DB that designates the singing advice A for each music attribute X is referred to according to the tendency of the group of the plurality of reference sounds R uttered by the target singer in the past in the reference sound group Q. Therefore, the effect that the appropriate singing advice A can be presented for each target singer is particularly remarkable. For example, since the reference information DB that designates the singing advice A with a specific pitch as the music attribute X is referred to, for a singer who is not good at singing each pitch at a specific pitch, Effective singing advice A can be presented.

<Third Embodiment>
In 1st Embodiment and 2nd Embodiment, the analysis process part 22 specified the singing advice A according to the tendency of the some reference sound R of the affiliation group G to which the target singer's singing voice V belongs. The analysis processing unit 22 according to the third embodiment specifies comments of evaluation results obtained by evaluating (scoring) the singing voice V according to the tendency of the plurality of reference voices R of the group G to which the singing voice V of the target singer belongs. . Specifically, the evaluation result of evaluating the singing voice V with an emphasis on the evaluation items according to the tendency of each reference voice R of the group G to which the singing voice V belongs is specified.

For example, in consideration of the fact that beginners mainly store the chorus section of music (other than that, they do not store much), the reference voice tends to have a large volume variation and a large pitch error. When the singing voice V belongs to the group G of R (that is, the beginner's group), the analysis processing unit 22 sets the weight value of the evaluation of the chorus section of the music to a larger numerical value than the other sections, and the evaluation result Is calculated. Further, for the group G of the reference speech R that tends to have a higher pitch evaluation result than the inflection evaluation result, the analysis processing unit 22 compares the weight evaluation pitch value with other elements such as intonation. Set the value to a larger value and calculate the evaluation result For the group G of the reference speech R that has a high inflection evaluation result and a high frequency of various singing techniques (Kobushi and Shakuri) compared to the pitch evaluation result, the analysis processing unit 22 uses the inflection and singing techniques. The evaluation result is calculated by setting the evaluation weight to a larger value compared to other elements such as pitch. In addition, for the group G of the reference speech R having a tendency that the sound pressure is large and the pitch fluctuation amount is large (that is, the tendency to sing), the analysis processing unit 22 has a larger weight value for evaluating the inflection than other elements. Set the numerical value and calculate the evaluation result. The presentation processing unit 24 causes the display device 18 to display a comment on the evaluation result specified by the analysis processing unit 22.

In the third embodiment, comments of evaluation results according to the tendency of the plurality of reference sounds R of the group G to which the singing voice V of the target singer belongs are presented to the target singer. That is, a comment appropriate for the singing voice V of each target singer is presented to the target singer. Therefore, there is an advantage that the singing of the target singer can be effectively improved. In the third embodiment, it is possible to omit the presentation of the comment of the evaluation result. That is, the present invention can also be realized as an apparatus (a configuration in which the presentation processing unit 24 is omitted) that evaluates the singing voice V according to the tendency of the plurality of reference voices R of the group G to which the singing voice V of the target singer belongs. .

<Modification>
Each of the above forms can be variously modified. Specific modifications are exemplified below. Two or more aspects arbitrarily selected from the following examples may be appropriately combined.

(1) In the first embodiment, the configuration in which the reference information DA for each piece of music is created in advance has been exemplified. However, the reference information DA can be generated in real time for each song of each piece of music. For example, the group G of each reference sound R similar in musical tendency to the singing sound V of the target singer is extracted from the reference sound group Q, and the analysis processing unit 22 uses each reference sound R of the group G. It is also possible to generate the reference information DA. For example, according to a group G of a plurality of reference sounds R whose evaluation index E is close to the singing voice V (for example, a plurality of reference sounds R within a range of ± 5% with respect to the evaluation index E and the pitch P of the singing voice V). Thus, a configuration for generating the reference information DA is preferable.

(2) In the first embodiment, a plurality of reference sounds R corresponding to a specific music piece in the reference sound group Q are classified into N groups G [1] to G [N]. This method is arbitrary as described in the first embodiment. For example, it is also possible to classify a predetermined number (for example, the top 5%) of reference voices R positioned in descending order of the evaluation index E among a plurality of reference voices R corresponding to a specific music into a group G [n]. is there.

(3) In each of the above-mentioned forms, the point in time when the difference between the average pitch P [n] of the group G [n] and the exemplary pitch P0 is maximized is selected as the indication point T. The method of selecting T is not limited to the above examples. For example, when the evaluation index E of the plurality of reference sounds R included in the group G [n] or the distribution degree of pitches (for example, dispersion or distribution width) increases, the average of the evaluation indices E of the plurality of reference sounds R It is also possible to select the point in time at which the value is minimized as the point-in-time T. It is also possible to select the point in time T when the difference between the average pitch P [n] and the exemplary pitch P0 exceeds a predetermined threshold.

(4) In the first embodiment, the affiliation group G is updated for each section of the target music. For example, the affiliation group G of the selected section is changed according to the evaluation index E of the singing voice V over a plurality of sections including the selected section. It is also possible to specify. Specifically, among the N groups G [1] to G [N] corresponding to different ranges of the evaluation index E, the weighted sum of the evaluation indices E over a plurality of sections with the selected section at the end is included. The group G [n] in the range to be specified is specified as the belonging group G. For example, the weight value applied to the evaluation index E in each section is set to a larger numerical value as it is closer to the selected section.

(5) It is also possible to selectively apply a plurality of reference information D (DA, DB) to specifying the singing advice A in accordance with an instruction from the user. For example, in the first embodiment, a set of reference information DA1 having a high point-in-time T and reference information DA2 having a low point-in-time T is prepared for each piece of music, and the analysis processing unit 22 receives the reference information according to an instruction from the user. A configuration that selectively uses DA1 and reference information DA2 is employed. When the reference information DA1 is applied, the singing advice A is presented at a number of pointed-out times T in the target music (that is, stubborn advice), and when the reference information DA2 is applied, the singing advice A is presented. Time T decreases (ie, sweet eye advice).

(6) In the second embodiment, when the evaluation index E is calculated for each section of the target music and the evaluation index E of the indicated section exceeds a predetermined reference value (that is, when the target singer overcomes weakness) It is also possible to notify the target singer. According to the above configuration, since the target singer can recognize how to overcome weakness, there is an advantage that the singing motivation of the target singer can be maintained.

(7) In the first embodiment and the second embodiment, the presentation of the singing advice A is exemplified, and in the third embodiment, the presentation of the evaluation result is exemplified. However, the content presented to the target singer is limited to the above illustration. Not. As understood from the illustrations of the above-described embodiments, the analysis processing unit 22 identifies comments (singing advice A and evaluation results) according to the tendency of each reference voice R of the group G belonging to the singing voice V of the target singer. It is expressed comprehensively as an element.

(8) The singing analysis device 100 can be realized by a server device (for example, a web server) that communicates with a communication terminal such as a communication karaoke device. For example, in the configuration in which the singing analysis device 100 according to the first embodiment is realized by the server device, the analysis processing unit 22 analyzes the singing advice A corresponding to the group G of the singing voice V received from the communication terminal via the communication network. The presentation processing unit 24 transmits a command for specifying (singing advice specifying process) and causing the target singer to present the singing advice A to the communication terminal.

The singing analysis apparatus according to each aspect described above is realized by hardware (electronic circuit) such as DSP (Digital Signal Processor) dedicated to the presentation of singing advice, and general-purpose arithmetic such as CPU (Central Processing Unit) This is also realized by cooperation between the processing device and the program. The program which concerns on the suitable aspect of this invention is the analysis process part which identifies the comment according to the tendency of each reference audio | voice of the group corresponding to the singing audio | voice of a target singer among the some reference audio | voices recorded beforehand, and The computer is caused to function as a presentation processing unit that presents the comment specified by the analysis processing unit to the target singer. Moreover, the program which concerns on another aspect is an analysis process part which evaluates the said song audio | voice according to the tendency of each reference audio | voice of the group corresponding to the singing audio | voice of a target song person among the some reference audio | voices recorded beforehand. Make the computer work. The program of the present invention can be provided in a form stored in a computer-readable recording medium and installed in the computer. The recording medium is, for example, a non-transitory recording medium, and an optical recording medium (optical disk) such as a CD-ROM is a good example, but a known arbitrary one such as a semiconductor recording medium or a magnetic recording medium This type of recording medium can be included. For example, the program of the present invention can be provided in the form of distribution via a communication network and installed in a computer.

The present invention is also specified as an operation method (singing analysis method) of the song analysis apparatus according to each of the above aspects. The singing analysis method according to a preferred aspect of the present invention is an analysis process for identifying a comment corresponding to a tendency of each reference voice of a group corresponding to the singing voice of the target singer among a plurality of reference voices recorded in advance. And a presentation process for presenting the comment specified in the analysis process to the target singer. In addition, the singing analysis method according to another aspect is an analysis process for evaluating the singing voice according to the tendency of each reference voice of the group corresponding to the singing voice of the target singer among a plurality of reference voices recorded in advance. Includes processes.

This application is based on a Japanese patent application (Japanese Patent Application No. 2014-045957) filed on March 10, 2014, the contents of which are incorporated herein by reference.

According to the present invention, it is possible to present an appropriate comment to the singer according to the singing tendency of the singer.

DESCRIPTION OF SYMBOLS 100 ... Singing analysis apparatus, 12 ... Arithmetic processing apparatus, 14 ... Memory | storage device, 16 ... Sound collection apparatus, 18 ... Display apparatus, 22 ... Analysis processing part, 24 ... Presentation processing part.

Claims

An analysis processing unit for identifying a comment corresponding to a tendency of each reference voice of the group corresponding to the singing voice of the target singer among a plurality of reference voices recorded in advance;
A singing analysis device comprising: a presentation processing unit that presents a comment specified by the analysis processing unit to the target singer.
The singing analysis apparatus according to claim 1, wherein the analysis processing unit specifies, as the comment, singing advice corresponding to a tendency of each reference voice of a group corresponding to the singing voice of the target singer.
The analysis processing unit refers to reference information that specifies singing advice for each of a plurality of groups in which a plurality of reference sounds that are common to the target singer's singing voice and music are classified, and among the plurality of groups, The singing analysis apparatus according to claim 2, wherein the singing advice of the group to which the singing voice of the target singer belongs is specified.
The singing analysis apparatus according to claim 3, wherein the analysis processing unit sequentially updates the group of the singing voice during the singing of the music by the target singer.
The analysis processing unit calculates an evaluation index for the singing voice of the target singer, specifies a group in which the evaluation index is included among the plurality of groups, selects unit information corresponding to the specified group, The singing analysis device according to claim 2, wherein the singing advice specified by the unit information is specified.
The analysis processing unit refers to reference information that specifies singing advice for each music attribute according to the tendency of each reference sound of the group that has collected the sound of the target singer among a plurality of reference sounds, and the target song The singing analysis apparatus according to claim 2, wherein the singing advice is specified at a location corresponding to the music attribute among the songs sung by the person.
The reference information designates a singing advice with a specific pitch of each successive pitch as the music attribute,
The singing analysis apparatus according to claim 6, wherein the analysis processing unit specifies singing advice specified in the pitch by the reference information with respect to a portion where the specific pitch is present in the music sung by the target singer.
The singing analysis apparatus according to claim 1, wherein the analysis processing unit specifies, as the comment, an evaluation result obtained by evaluating the singing voice of the target singer according to the tendency specified by the analysis processing unit.
Identify a comment according to the tendency of each reference voice of the group corresponding to the singing voice of the target singer among a plurality of reference voices recorded in advance,
A singing analysis method for presenting the specified comment to the target singer.