JP5416396B2 - Singing evaluation device and program - Google Patents

Description

  The present invention relates to a technique for evaluating the skill of a singing technique.

  In the karaoke apparatus, an apparatus for scoring a skill of singing or displaying a message indicating the content of singing instruction has been proposed. In such a karaoke apparatus, for example, when scoring or teaching about the vibrato technique, first, the section in which the vibrato technique is used is identified from the singing voice as a model, and the singing voice in the identified section is analyzed. A method of scoring and teaching can be considered. Note that the vibrato technique is a singing technique that gives the sound a rich reverberation by producing a timbre that slightly raises and lowers the pitch while stretching the sound.

As a method for identifying a section in which the vibrato technique is used from a singing voice, Patent Document 1 discloses a method for identifying a vibrato section in which the pitch of input data vibrates. Patent Document 2 discloses a method of detecting a vibrato technique from singing voice and reflecting it in singing evaluation.
JP 2006-30610 A JP 2004-102146 A

A skilled singer uses a technique (hereinafter referred to as a combination technique) in which a technique for extending the same sound for a certain period of time (for example, 500 milliseconds) (hereinafter referred to as a long tone technique) is followed by a direct shift to a vibrato technique. Sometimes.
In the conventional karaoke apparatuses such as the karaoke apparatuses described in Patent Documents 1 and 2, the vibrato technique and the long tone technique are described as being separately evaluated, but a combination technique combining the above techniques is described. It was not possible to perform precise singing evaluations such as evaluating.

  The present invention has been made under the above-described background, and an object thereof is to provide a technique for precisely evaluating the skill of a singing technique.

The singing evaluation apparatus according to the present invention includes a first receiving means for receiving singing voice data representing a singing voice, a first pitch detecting means for detecting a pitch of the singing voice data received by the first receiving means, When it is detected whether or not the state in which the variation in pitch detected by the first pitch detecting means is within a predetermined range has continued for a predetermined time or more, and the continuation for the predetermined time or longer is detected, the detection section First long tone section specifying means for specifying as a long tone section, and detecting whether the pitch detected by the first pitch detecting means periodically fluctuates beyond a predetermined width, and the pitch period If the fluctuation is detected, the long tone interval is detected in the first vibrato section identifying means for identifying the detection section as the vibrato period, by the first long-tone section identifying means If the vibrato period is specified by the first vibrato section identifying means, the difference between the end time and the start time of the vibrato section of the long tone interval is determined whether it is below a predetermined value, the determination result There a first detecting means for detecting a combination of a positive long tone interval and vibrato period, for the long tone interval and vibrato section first detecting means corresponding to the combination detected, the time length of the long tones section Received by the second receiving means, the second receiving means for receiving the model voice data representing the singing voice serving as a model, and the second receiving means for calculating the relative value between the time length of the voice and the vibrato section Second pitch detection means for detecting the pitch of the model voice data, and the variation in pitch detected by the second pitch detection means is within a predetermined range. A second long tone section specifying means for detecting whether or not the state has continued for a predetermined time or more, and specifying the detected section as a long tone section when continuation for the predetermined time or more is detected; Detecting whether or not the pitch detected by the pitch detecting means periodically fluctuates beyond a predetermined width, and when the periodic fluctuation of the pitch is detected, the detected section is specified as a vibrato section. When the long tone section is detected by the second vibrato section specifying means and the second long tone section specifying means, and the vibrato section is specified by the second vibrato section specifying means, the end time of the long tone section And a combination of a long tone period and a vibrato period in which the determination result is affirmative The second detection means for detecting the setting and the long tone interval and the vibrato interval corresponding to the combination detected by the second detection means are the relative values of the time length of the long tone interval and the time length of the vibrato interval. A second relative value calculating means for calculating a relative value of 2, and an evaluation means for calculating an evaluation of the singing voice based on a detection result of the detecting means , wherein the first relative value calculating means calculates Evaluation means for comparing the relative value with the second relative value calculated by the second relative value calculation means and calculating the evaluation of the singing voice based on the comparison result is provided.

The singing evaluation apparatus according to the present invention includes a receiving unit that receives singing voice data representing a singing voice, a pitch detecting unit that detects a pitch of the singing voice data received by the receiving unit, and a pitch detected by the pitch detecting unit. A long tone section specifying means for detecting whether or not the state where the fluctuation is within a predetermined range has continued for a predetermined time or more, and when the continuation for the predetermined time or more is detected, the detection section is specified as a long tone section And detecting whether or not the pitch detected by the pitch detecting means periodically fluctuates beyond a predetermined width, and when a periodic fluctuation of the pitch is detected, the detected section is specified as a vibrato section A long tone section is detected by the vibrato section specifying means and the long tone section specifying means, and the vibrato section is detected by the vibrato section specifying means. If it is specified, it is determined whether the difference between the end time of the long tone period and the start time of the vibrato period is equal to or less than a predetermined value, and the determination result is positive between the long tone period and the vibrato period. Detecting means for detecting a combination; and a relative value calculating means for calculating a relative value between the time length of the long tone section and the time length of the vibrato section for the long tone section and the vibrato section corresponding to the combination detected by the detecting means; An evaluation means for calculating an evaluation of the singing voice based on a detection result of the detection means, wherein the relative value calculated by the relative value calculation means is compared with a predetermined value, and the singing based on the comparison result And an evaluation means for calculating an evaluation of speech.

The program according to the present invention includes a first receiving means for receiving singing voice data representing a singing voice, and a first pitch detecting means for detecting the pitch of the singing voice data received by the first receiving means. Detecting whether or not the state in which the variation in pitch detected by the first pitch detecting means is within a predetermined range has continued for a predetermined time or more, and detecting the continuation for the predetermined time or longer. A first long tone section specifying means for specifying the section as a long tone section, and detecting whether or not the pitch detected by the first pitch detecting means periodically fluctuates beyond a predetermined width; When periodic fluctuation is detected, a first vibrato section specifying means for specifying the detected section as a vibrato section, and a long tone by the first long tone section specifying means. When the interval is detected and the vibrato section is specified by the first vibrato section specifying means, it is determined whether the difference between the end time of the long tone section and the start time of the vibrato section is equal to or less than a predetermined value. First detection means for detecting a combination of a long tone section and a vibrato section for which the determination result is positive, and a long tone section and a vibrato section corresponding to the combination detected by the first detection means A first relative value calculating means for calculating a relative value between the time length of the tone section and the time length of the vibrato section; second receiving means for receiving model voice data representing a singing voice as a model; A second pitch detecting means for detecting the pitch of the model voice data received by the receiving means; and a variation in pitch detected by the second pitch detecting means is predetermined. Second long tone section specifying means for detecting whether or not the state within the range has continued for a predetermined time or more and, when the continuation for the predetermined time or more is detected, specifying the detection section as a long tone section; , Detecting whether or not the pitch detected by the second pitch detecting means periodically fluctuates beyond a predetermined width, and when the periodic fluctuation of the pitch is detected, the detected section is defined as a vibrato section When a long tone section is detected by the second vibrato section specifying means and the second long tone section specifying means, and the vibrato section is specified by the second vibrato section specifying means, the long tone section It is determined whether the difference between the end time of the section and the start time of the vibrato section is equal to or less than a predetermined value, and the long tone section and the vibrato where the determination result is affirmative A second detection means for detecting a combination of sections, and a long tone section and a vibrato section corresponding to the combination detected by the second detection means, with a relative value of the time length of the long tone section and the time length of the vibrato section. A second relative value calculating means for calculating a second relative value; and an evaluation means for calculating an evaluation of the singing voice based on a detection result of the detecting means , wherein the first relative value calculating means The calculated relative value is compared with the second relative value calculated by the second relative value calculating means, and the function is used as an evaluating means for calculating the evaluation of the singing voice based on the comparison result.

The program according to the present invention includes a computer for receiving singing voice data representing singing voice, a pitch detecting means for detecting a pitch of singing voice data received by the receiving means, and a pitch detected by the pitch detecting means. Long tone section specification that detects whether or not the state of fluctuation within a predetermined range continues for a predetermined time or more, and when the continuation for the predetermined time or longer is detected, the detected section is specified as a long tone section And when the pitch detected by the pitch detection means periodically fluctuates beyond a predetermined width, and when a periodic fluctuation of the pitch is detected, the detected section is defined as a vibrato section. A long tone section is detected by the vibrato section specifying means to be specified and the long tone section specifying means, and the vibrato section specifying means is When a brat section is specified, it is determined whether the difference between the end time of the long tone section and the start time of the vibrato section is a predetermined value or less, and a long tone section in which the determination result is affirmative Detection means for detecting a combination of vibrato sections, and a relative value for calculating a relative value of the time length of the long tone section and the time length of the vibrato section for the long tone section and the vibrato section corresponding to the combination detected by the detection means A calculation means and an evaluation means for calculating an evaluation of the singing voice based on a detection result of the detection means , the relative value calculated by the relative value calculation means is compared with a predetermined value, and based on the comparison result And function as evaluation means for calculating the evaluation of the singing voice .

  According to the present invention, the skill of the singing technique can be accurately evaluated.

The karaoke apparatus 1 according to the present invention uses a vibrato technique to sing a singing voice (hereinafter referred to as a model voice) serving as a model and a singing voice (hereinafter referred to as a practitioner voice) of a person who practices singing (hereinafter referred to as a practitioner voice). A section (hereinafter referred to as a vibrato section) or a section sung using a long tone technique (hereinafter referred to as a long tone section). And the karaoke apparatus 1 has a function which evaluates the skill of those singing techniques precisely based on the vibrato area and the long tone area specified in those voices.
Embodiments of the present invention will be described below with reference to the drawings.

(A: Configuration)
FIG. 1 is a block diagram showing an example of a hardware configuration of a karaoke apparatus 1 according to an embodiment of the present invention.
The control unit 11 includes a central processing unit (CPU), a read only memory (ROM), and a random access memory (RAM). The control unit 11 reads and executes the control program stored in the ROM, thereby Control each part.
The display unit 13 includes a display device such as a liquid crystal panel, and displays various images such as karaoke lyrics telop under the control of the control unit 11.
The operation unit 14 includes various operators, and outputs an operation signal indicating the operation content by the operator to the control unit 11.
The microphone 15 outputs an analog audio signal representing the collected audio.
The audio processing unit 16 includes an A / D converter, converts an analog audio signal output from the microphone 15 into digital audio data, and outputs the digital audio data to the control unit 11. The audio processing unit 16 includes a D / A converter, converts digital audio data received from the control unit 11 into an analog audio signal, and outputs the analog audio signal to the speaker 17.
The speaker 17 emits sound based on the audio signal received from the audio processing unit 16.

The storage unit 12 is a storage unit such as a hard disk drive, and includes various storage areas described below.
The accompaniment data storage area 121 stores accompaniment data for each music piece. The accompaniment data is, for example, in a data format such as MIDI (Musical Instrument Digital Interface) format, and information indicating the pitch (pitch) of various musical instruments that accompany each of the above-mentioned music is written along with the progress of the music. The accompaniment data includes melody pitch data indicating the pitch (pitch) for each note (note) of the guide melody of the music.
The lyrics data storage area 122 stores lyrics data indicating the lyrics of each song. The lyrics data is displayed on the display unit 13 as lyrics telop as the karaoke accompaniment progresses.
In the model voice data storage area 123, data representing model voices of each music piece (hereinafter referred to as model voice data) is stored. The model audio data is, for example, audio data in the WAVE format or MP3 (MPEG1 Audio Layer-3) format.
In the trainer voice data storage area 124, for each piece of music, voice data generated by the trainer's singing being picked up by the microphone 15 and converted into digital data by the voice processing unit 16 (hereinafter referred to as “practice voice”). Data) is stored. This trainer voice data is also voice data in the WAVE format or MP3 format.
The vibrato section data storage area 125 stores data (hereinafter referred to as vibrato section data) indicating a vibrato section in the model voice data and the trainer voice data of each music piece. The vibrato section data is data generated by the control unit 11 as will be described later.
The long tone section data storage area 126 stores data indicating the long tone section (hereinafter, long tone section data) in the model voice data and the trainer voice data of each music piece. The long tone section data is data generated by the control unit 11 as will be described later.
The parameter storage area 127 stores a pitch extracted from the model voice data and the trainer voice data in each musical piece, and various parameters extracted from the pitch.

(B: Operation of the embodiment)
Next, the operation of the karaoke apparatus 1 according to the present invention will be described.

(B-1; Karaoke accompaniment processing)
First, karaoke accompaniment processing will be described with reference to the flowchart shown in FIG.
When the practitioner operates the operation unit 14 of the karaoke apparatus 1 and selects a song to sing, an operation signal specifying the selected song is output from the operation unit 14 to the control unit 11.

In step SA100, the control unit 11 causes the karaoke accompaniment to be performed and causes the display unit 13 to display a lyrics telop. That is, the control unit 11 reads the accompaniment data corresponding to the operation signal from the accompaniment data storage area 121 and supplies the accompaniment data to the sound processing unit 16. The audio processing unit 16 converts the accompaniment data into an analog audio signal and outputs it to the speaker 17. The speaker 17 emits sound based on the audio signal received from the audio processing unit 16.
In addition, the control unit 11 reads out lyric data corresponding to the operation signal from the lyric data storage area 122 and supplies the lyric data to the display unit 13. The display unit 13 displays the lyrics telop based on the lyrics data.

The practitioner refers to the lyrics telop displayed on the display unit 13 and sings along with the sound (karaoke accompaniment) emitted from the speaker 17.
Singing voice by the practitioner is picked up by the microphone 15, and practicing voice data is generated based on the picked up singing voice (step SA110). The generated trainer voice data is written in the trainer voice data storage area 124 of the storage unit 12.

  As the karaoke accompaniment progresses, the control unit 11 reads out the model voice data from the model voice data storage area 123 in parallel with the process of generating the trainer voice data as described above.

The control unit 11 extracts a pitch (pitch) in frame units of a predetermined time length (for example, 10 msec) from the trainer voice data generated as the karaoke accompaniment progresses and the read model voice data, “Absolute model pitch data” and “absolute practitioner pitch data” representing the absolute pitch are generated (step SA120). Hereinafter, when the two are not distinguished, they are collectively referred to as “absolute pitch data”.
The generated absolute pitch data is written in the parameter storage area 127.

Next, the control unit 11 calculates a relative value when the pitch represented by the melody pitch data is set as a zero reference value from the generated absolute model pitch data and absolute trainer pitch data. Data "and" relative practitioner pitch data "are generated (step SA130). Hereinafter, when the two are not distinguished, they are collectively referred to as relative pitch data.
The generated relative pitch data is written in the parameter storage area 127 of the storage unit 12.

FIG. 3 shows an example of the pitch represented by the absolute practitioner pitch data generated as described above, together with the melody pitch data. In FIG. 3, the horizontal axis represents time (elapsed time since the music was started). In addition, the vertical axis indicates the pitch of the practitioner voice at each time.
FIG. 4 shows relative practitioner pitch data generated from the absolute practitioner pitch data shown in FIG. In FIG. 4, the vertical axis represents the ratio (pitch deviation) between the pitch of the trainee voice and the pitch of the guide melody. In the figure, “cent” is a unit indicating a frequency ratio between two sounds. For example, when displayed with a deviation of +100 cents from the reference guide melody, the pitch of the trainer voice and the guide melody is a semitone. Indicates that
Here, the absolute pitch data and the relative pitch data regarding the practitioner voice are shown, but similar data is also generated for the model voice.

In step SA140, the control unit 11 determines whether or not one karaoke accompaniment has been completed. If the determination result in step SA140 is “Yes”, the karaoke accompaniment process ends. On the other hand, if the determination result in step SA140 is “No”, the processes in steps SA100 to SA130 are performed on the remaining portion of the music.
As described above, as the karaoke accompaniment progresses, data regarding the pitches of the model voice and the practice voice are sequentially written in the storage unit 12.

(B-2; identification of vibrato section)
When the karaoke accompaniment process ends, the control unit 11 specifies a vibrato section in the model voice and the trainer voice.
In the following, an example of a method for specifying the vibrato section will be briefly described. However, the vibrato section may be specified by another conventionally known method.

  The control unit 11 specifies a vibrato section in the model voice and the trainer voice in light of the following conditions. First, the control unit 11 identifies a point where the pitch represented by the relative pitch data generated in the karaoke accompaniment process changes from negative to positive or from positive to negative (zero crossing) as a zero crossing point. For example, in the example shown in FIG. 4 (relative practitioner pitch data), the time at which the graph is zero-crossed, that is, times t1 to t5, etc. is specified as the zero-cross location.

Next, the control unit 11 measures the time interval at which the zero-cross points appear, (1) the measured time interval is within a predetermined range, and (2) the zero-cross points are continuously detected a predetermined number of times or more, (3) A section in which the pitch amplitude exceeds a predetermined value (for example, 20 cents) is specified as a vibrato section. In short, “a section in which the pitch periodically fluctuates beyond a predetermined width” is specified as a vibrato section.
By this process, in the example shown in FIG. 4, section A is specified as a vibrato section. It should be noted that the vibrato section is specified in the same manner for the other music parts of the trainer voice data and also for the model voice data.
The control unit 11 specifies a start time and an end time for each of the specified vibrato sections, and writes the start time and end time in the vibrato section data storage area 125.

(B-3; identification of long tone section)
The control unit 11 specifies the long tone section in parallel with the above-described specification of the vibrato section. In the following, an example of a method for specifying a long tone section will be briefly described. However, the long tone section may be specified by another conventionally known method.

FIG. 5 shows a pitch indicated by absolute practitioner pitch data by a certain practitioner together with melody pitch data.
First, the control unit 11 specifies a predetermined width so as to include the pitch of each time indicated by the melody pitch data. In the figure, the pitch corresponding to the top and bottom 10 cents of the melody pitch data is a predetermined width. A hatched area (detection area) shows an example of the predetermined width at each time.

  The control unit 11 determines whether or not the pitch at each time indicated by the absolute practitioner pitch data falls within the predetermined range, and exceeds the predetermined time (for example, the time width indicated by k in the figure). The section is referred to as a “long tone section”. In the figure, the absolute practitioner pitch data continues to fall within the predetermined range in the interval from time t1a to t1b and in the interval from time t2a to t2b. Since it is only the section from time t1a to t1b, this section is specified as a long tone section. In short, “a section in which a state in which the pitch variation is within a predetermined range continues for a predetermined time or longer” is specified as a long tone section.

Similarly, the long tone section is specified for the other music parts of the trainer voice data and for the model voice data.
When the control unit 11 specifies the long tone section, the control section 11 specifies the start and end times of the section, and writes the start and end times in the long tone section data storage area 126 of the storage unit 12 as long tone section data.

(B-4; Evaluation of singing)
Next, a method for evaluating a practitioner's voice will be described. In the karaoke apparatus 1, evaluation from a plurality of viewpoints is performed. Below, the viewpoint of each evaluation is each demonstrated.

(B-4-1; Evaluation based on pitch accuracy)
The karaoke apparatus 1 evaluates singing based on pitch accuracy. This evaluation method is substantially the same as the evaluation method used in the conventional karaoke apparatus, and the evaluation method will be briefly described below.
The control unit 11 integrates the absolute value of the difference between the pitch of the practitioner voice and the pitch of the guide melody (the absolute value of the relative practitioner pitch data) with time, Accumulate over the whole. The larger the integrated value, the more the trainer's voice is deviated from the guide melody in the entire music, and the evaluation is performed by the deduction method according to the integrated value.

(B-4-2; Evaluation of vibrato technique)
Next, a method for evaluating the skill of the vibrato technique in the trainer voice will be described with reference to the flowchart shown in FIG. In the present invention, any of the conventionally known evaluation methods may be used as the evaluation method of the vibrato technique. Hereinafter, an example will be briefly described.

In step SB110, the control unit 11 reads out the vibrato section data related to the trainer voice data from the vibrato section data storage area 125.
The control unit 11 extracts the following parameters in each vibrato section in the relative practitioner pitch data (step SB120). In the following description, for example, when the pitch fluctuates periodically as in section A of FIG. 4, the number of vibrations of the pitch per unit time is referred to as “vibrato vibration frequency”.

(1) Standard deviation Df of vibrato frequency
The parameter Df is calculated as a variation (standard deviation) in a plurality of pieces of data of the reciprocal of the time interval at which the relative practitioner pitch data zero-crosses the horizontal axis.
(2) Standard deviation Dp of pitch vibration width
The parameter Dp indicates variation (standard deviation) in the vibration width of each vibration in a series of pitch vibrations found in each vibrato section.

The control unit extracts the two parameters shown above for each vibrato section read in step SB110, and writes the extracted parameters in the parameter storage area 127.
In step SB130, the control unit 11 evaluates the skill of the vibrato technique in the vibrato section based on the parameters written in the parameter storage area 127. Specifically, it is as follows.

The control unit 11 gives higher evaluation (score) as the value of Df is smaller. This is because Df represents a variation in the vibration frequency of vibrato, and is considered to be an excellent vibrato that shows a uniform vibration frequency as the value of Df is closer to zero.
Moreover, the control part 11 gives high evaluation (score), so that the value of Dp is small. This is because Dp means the variation of the pitch fluctuation range in the vibrato section, and the closer the value of Dp is to 0, the better the vibrato that the pitch vibrates with a uniform fluctuation range.
In the following, the score given for the vibrato technique is called the “vibrato score”.

In the vibrato technique, it is known that the higher the periodicity of the pitch variation, that is, the more regular the frequency and vibration width, the richer and more desirable the sound resonates. Therefore, by evaluating based on Df, which is a parameter reflecting the uniformity of the vibration frequency of vibrato, and Dp, a parameter reflecting the uniformity of the vibration width of the pitch, the skill of the vibrato technique is appropriately evaluated. Is possible.
In the vibrato section, the above-described “evaluation based on pitch accuracy” is not performed. This is because the vibrato technique is a technique in which the practitioner sings so as to intentionally deviate from the pitch of the melody pitch data.

(B-4-3; Evaluation of long tone technique)
Next, a method for evaluating the skill of the long tone technique in the trainee voice will be described. In the present invention, any of the conventionally known evaluation methods may be used as the evaluation method of the long tone technique. Hereinafter, an example will be briefly described.

  Since the long tone technique is a technique for producing a sound having the same pitch for a long time with a single breath, the evaluation criteria are that the pitch does not rise or fall during the sound generation or the sound generation is not interrupted during the sound generation.

  The control unit 11 reads out the long tone section data for the trainee voice from the long tone section data storage area 126. Then, the control unit 11 evaluates the skill of the long tone technique in the trainer voice based on the following conditions in each long tone section in the relative trainer pitch data.

The control unit 11 integrates the absolute value of the relative practicer pitch data with respect to each long tone section in the relative practicer pitch data over time, and determines the pitch deviation in the practicer voice as each longtone period of the entire song. It accumulates over.
The integrated value is generally a small value in a long tone section in which an excellent long tone technique is used. Therefore, the control unit 11 adds points according to the integrated value. That is, as the integrated value is smaller, it is considered that the trainer's voice keeps the pitch constant during the long tone period, so a higher score is added.
Also, the duration of the long tone is calculated from the start time and end time of the long tone section, and a higher score is added as the duration is longer. This is because long tone generally has a higher difficulty level as the duration continues, but it becomes a beautiful song.
Hereinafter, the score given for the above long tone technique is referred to as “long tone score”.
Note that the above-described “evaluation based on pitch accuracy” is not performed even in the long tone section. This is because the evaluation of the long tone technique substantially includes evaluation contents based on pitch accuracy.

(B-4-4; Evaluation based on combination of long tone and vibrato)
As described above, the control unit 11 separately evaluates the skill of the vibrato technique and the long tone technique. However, in the karaoke apparatus 1 according to the present invention, a technique (combination technique) further combining the vibrato technique and the long tone technique. ) And the following singing is evaluated based on the combination technique.

Here, an example of the “combination technique” is shown. FIG. 7 shows the absolute model pitch data in a part of the music together with the melody pitch data. In the drawing, the pitch indicated by the absolute model pitch data is continuously within a predetermined width from the pitch of the melody pitch data in the section from the time t1 to the time t2, and is determined to be a long tone section. Further, immediately after the long tone section, in the section from time t3 to t4, the pitch indicated by the absolute model pitch data vibrates with reference to the pitch of the melody pitch data, and is determined to be a vibrato section. Is done.
In the lower part of the figure, the temporal relationship between the long tone section and the vibrato section is schematically shown. A horizontal line band represents a long tone section (LT in the figure), and a wavy line band represents a vibrato section (VT in the figure).

  FIG. 8 shows the vibrato section and the long tone section specified in the trainee voice in the same display manner as the lower part of FIG. As shown in FIGS. 9A to 9C, there are various patterns in the temporal arrangement of the vibrato section and the long tone section.

  As shown in FIG. 5A, when the vibrato period is continuous immediately after the long tone period, that is, when the end time of the long tone period and the start time of the vibrato period coincide with each other, the long tone technique is used as the vibrato technique. Is determined to be used, and a predetermined score (to be referred to as a combination score hereinafter) is added to the singing evaluation.

As shown in FIG. 5B, there is a vibrato section immediately after the long tone section, but there is a time gap between the end time of the long tone section and the start time of the vibrato section (in the figure, 0.2 seconds). When the time gap is less than a predetermined time (0.5 seconds), it is determined that the long tone period and the vibrato period are continuous, and the combination is performed as in (a). A score is added.
Similarly, when the end time of the long tone section is shifted before or after the start time of the vibrato section, similarly, when the time gap is less than a predetermined time, the long tone section It is determined that the section and the vibrato section are continuous.
In the above processing, it is determined that both sections are continuous even though the time gap is detected. The time gap is generated when the long tone section and the vibrato section are specified. This is because it is considered to be derived from the error.
In the present operation example, the predetermined time is set to 0.5 seconds. However, the predetermined time is not limited to this value and may be set as appropriate. For example, it may be changed according to the tempo of the music.

  In FIG. 4C, the time gap between the long tone section and the vibrato section is 1.5 seconds. Thus, the time gap exceeds the predetermined time (0.5 seconds). In this case, it is not determined that the long tone period and the vibrato period are continuous. This is because when the time gap exceeds the predetermined time (0.5 seconds), it is considered that the long tone section and the vibrato section are used independently for separate parts of the music. Because. In this case, a combination score is not added.

Here, a method of calculating the “combination score” described above will be described. In the present embodiment, the combination score is determined based on the ratio of the time lengths of the combined long tone section and vibrato section.
For example, as shown in FIG. 11, when a practitioner sings in the patterns of (a), (b), and (c) in the music portion designated to sing the same pitch for 8 seconds. Will be described. In FIG. 6A, a vibrato section of 2 seconds is detected after a long tone section of 6 seconds. In this case, the vibrato technique is used in the latter half 25% of the section (hereinafter referred to as the combined section) that combines the long tone section and the vibrato section. Further, in FIGS. 5B and 5C, the vibrato technique is used in the second half 50% and 75% of the combination section, respectively.

In this case, the control unit 11 detects a combination section corresponding to each combination section in the trainee voice in the model voice, and calculates a ratio of the vibrato section for the combination section in the model voice. Then, the control unit 11 calculates the ratio of the vibrato section for each combination section in the trainer voice, and adds a higher score as the calculated ratio is closer to the ratio of the vibrato section calculated in the model voice. To do.
For example, for the musical piece portion shown in FIG. 11, when the proportion of the vibrato section is 30% in the model voice, the highest combination score is added when the practice voice is in the form shown in FIG. In the case of (b) and (c), the combination score lower than that in (a) is added.

(B-5; Summary of singing evaluation)
The evaluation contents based on the above-described singing evaluation will be specifically described. 9 and 10 show examples of absolute practitioner pitch data as in FIG. In the following, the evaluation contents for each trainer's voice will be described.

  As shown in FIG. 9, when the practitioner sings the music piece using only the long tone technique, the evaluation contents are as follows. That is, a long tone section is detected in the music part, and a higher long tone score is added as the duration of the long tone section is longer and as the deviation between the absolute trainee pitch data and the guide melody is smaller. . In addition, in the same figure, about the sections other than the long tone section, the above-described “singing evaluation based on pitch accuracy” is performed.

  Moreover, as shown in FIG. 10, when the practitioner sings the music piece using only the vibrato technique, the evaluation contents are as follows. That is, a vibrato section is detected in the music portion, and a higher vibrato score is added as the duration of the vibrato section is longer and the frequency and amplitude of the vibration of the pitch are more uniform. In addition, in the same figure, about the sections other than the vibrato section, the above-described “singing evaluation based on pitch accuracy” is performed.

  Although illustration is omitted, when the practitioner sings the music piece without using either the vibrato technique or the long tone technique, the evaluation is performed only based on the above-described “singing evaluation based on pitch accuracy”. .

  As described above, when the practitioner sings using either the long tone technique or the vibrato technique, each technique is evaluated and a score is added. In addition, for a section in which any technique is not used, evaluation is performed based on pitch accuracy.

  As shown in FIG. 3, when the practitioner sings the same piece of music as described above by combining the long tone technique and the vibrato technique, the evaluation contents are as follows. That is, first, the detected long tone technique is independently evaluated, and a long tone score is added. The detected vibrato technique is also evaluated independently of the long tone technique, and a vibrato score is added. These evaluations are performed for each of the long tone period and the vibrato period.

  Furthermore, a combination score is added to the fact that the long tone technique and the vibrato technique are used in combination. That is, as shown in FIG. 3, when it is detected that a combination technique of singing so as to shift from the long tone technique to the vibrato technique is used, the ratio of the time length between the long tone period and the vibrato period is calculated. Based on this, a combination score is calculated, and the score is added to the singing evaluation.

  Now, the long tone score or vibrato score given when singing with the long tone technique alone or the vibrato technique alone (FIGS. 9 and 10), and the long tone score given when singing with the combination of both techniques (FIG. 3) When the vibrato scores are compared with each other, the scores when the respective techniques are used alone tend to be higher. This is because the duration of the long tone section or vibrato section when singing using the long tone technique or vibrato technique alone can be continued longer than the duration of each technique when singing using both techniques in combination. It is. For example, the durations of the long tone period and the vibrato period in FIG. 3 are shorter than the durations of the long tone period and the vibrato period in FIG.

  However, a sufficiently high value is set for the score of “combination score” added in the case of FIG. 3, and the sum of the scores added in FIG. 3 (long tone score + vibrato score + combination score) is: A high value is obtained when compared with the score in FIG. 9 (long tone score only) or the score in FIG. 10 (vibrato score).

As described above, when a musical tone is sung by combining the long tone technique and the vibrato technique, the evaluation is higher than when the long tone technique and the vibrato technique are singed alone.
In general, when a long tone technique or a vibrato technique is used alone, the singing method tends to feel monotonous or homogeneous. On the other hand, when the long tone technique and the vibrato technique are combined, the emotional excitement in the music is expressed abundantly, and the singing feels excellent. Therefore, according to the karaoke apparatus 1 that detects the “combination technique” that combines the long tone technique and the vibrato technique and highly evaluates the technique, it is possible to perform an evaluation that closely matches the feeling of actually singing a song. it can.

(C: Modification)
As mentioned above, although embodiment of this invention was described, this invention is not limited to embodiment mentioned above, It can implement with another various form. An example is shown below.

(1) In the above-described embodiment, the case where the vibrato section and the long tone section for the model voice data and the trainer voice data are specified after the karaoke accompaniment process has been described. However, the process may be executed in parallel with the karaoke accompaniment process. That is, the trainer voice data is sequentially stored and the model voice data is read out, and the pitch is extracted from the model voice data and the practitioner voice data and the pitch data is generated from each singing voice in the music portion where the performance of the karaoke is finished. May be.
Moreover, in embodiment mentioned above, the case where the evaluation regarding a practitioner audio | voice was performed after completion | finish of a karaoke accompaniment process was demonstrated. However, the singing evaluation process may be sequentially performed as soon as data necessary for the evaluation is prepared as the karaoke accompaniment progresses.

(2) In the above-described embodiment, in the evaluation of each of the vibrato technique and the long tone technique, the case where scoring is performed by the scoring method according to various parameters representing the skill of the technique has been described. However, in each evaluation of the vibrato technique or the long tone technique, evaluation may be performed by a deduction method based on parameters extracted from each technique. For example, with respect to the vibrato technique, the larger the parameters such as the standard deviation Df of the pitch frequency and the standard deviation Dp of the pitch vibration width, the larger the points may be deducted. In the long tone technique, points may be deducted as the pitch deviation from the guide melody increases. Further, with respect to both the above techniques, points may be deducted when there is a section that is not detected in the trainer voice even though both techniques are detected in the model voice. That is, a low evaluation may be performed for the fact that the vibrato technique or the long tone technique is not used at a place where it should be used.

(3) In the above embodiment, the case where the combination score is adjusted based on the ratio of the vibrato section in each “combination section” of the trainer voice and the ratio of the vibrato section in the corresponding sample speech section has been described. . However, the method of adjusting the combination score is not limited to the above method.
For example, the combination score may be adjusted based on the difference between the ratio of the vibrato section in each combination section of the trainer voice and a value (for example, 30%) set in advance in the control program.
Further, the ratio of the vibrato section in each combination section of the trainer voice may be compared with a value based on the attributes of the music such as the genre or tempo of the music. For example, the storage unit of the karaoke apparatus 1 stores a table that associates “music genre” with “ratio of vibrato section”, and if the music data includes genre data, The ratio of the vibrato section corresponding to the genre represented by the data is specified with reference to the above table, and the combination score is based on the difference between the ratio of the vibrato section and the specified ratio in each combination section of the trainer voice. May be adjusted.
In addition, as for the combination score, a predetermined score is added for each combination technique, and the combination score in the whole music may be proportional to the number of combination techniques detected in the music.

(4) In the above-described embodiment, the case where the vibrato technique, the long tone technique, and the combination technique are evaluated for the trainer voice data has been described. However, the data to be processed is not limited to data representing singing voice. For example, it may be audio data representing the performance sound of a musical instrument such as a violin or flute.

(5) In the above-described embodiment, the case where the vibrato section and the long tone section in each musical piece are specified from the model voice and the singing is evaluated based on the both specified sections has been described. However, when data indicating both sections in the model voice is obtained separately, it is not necessary to specify both sections. The case where data indicating both sections is obtained separately is, for example, a case where data indicating both sections is included in advance in the model voice data. In such a case, the data indicating both sections may be read out and used at the time of singing evaluation.

(6) In the above-described embodiment, the case where the parameter types (Df, Dp) are used in the detection of the vibrato section and the evaluation of the technique has been described. However, these parameter types are examples, and in addition to these, the vibrato technique is used. It is possible to evaluate the vibrato technique using various parameters representing the characteristics of

(7) The control program executed by the control unit 11 of the karaoke apparatus 1 in the above-described embodiment is recorded on a recording medium such as a magnetic tape, a magnetic disk, a flexible disk, an optical recording medium, a magneto-optical recording medium, RAM, or ROM. Can be provided in the state. It is also possible to download to the karaoke apparatus 1 via a network such as the Internet.

It is a block diagram which shows the hardware constitutions of the karaoke apparatus. It is a flowchart which shows the flow of a karaoke accompaniment process. It is a figure which shows an example of absolute practitioner pitch data. It is a figure which shows an example of relative practitioner pitch data. It is a figure for demonstrating the identification method of a long tone area. It is a flowchart which shows the flow of a vibrato area specific process. It is a figure which shows an example of absolute model pitch data. It is a figure which shows the relationship between a long tone area and a vibrato area. It is a figure which shows an example of absolute practitioner pitch data. It is a figure which shows an example of absolute practitioner pitch data. It is the figure which showed the aspect of the combination technique.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Karaoke apparatus, 11 ... Control part, 12 ... Memory | storage part, 13 ... Display part, 14 ... Operation part, 15 ... Microphone, 16 ... Sound processing part, 17 ... Speaker, 121 ... Accompaniment data storage area, 122 ... Lyric data Storage area 123: Model voice data storage area 124 ... Trainer voice data storage area 125 ... Vibrato section data storage area 126 ... Long tone section data storage area 127: Parameter storage area

Claims (4)

First receiving means for receiving singing voice data representing the singing voice;
A first pitch detecting means for detecting the pitch of the singing voice data to which the first receiving means is received,
When it is detected whether or not the state in which the variation in pitch detected by the first pitch detecting means is within a predetermined range has continued for a predetermined time or more, and the continuation for the predetermined time or longer is detected, the detection section First long tone section specifying means for specifying as a long tone section;
If the pitch detected by the first pitch detection means periodically changes beyond a predetermined width, and the periodic fluctuation of the pitch is detected, the detected section is defined as a vibrato section. First vibrato section identifying means for identifying;
When a long tone section is detected by the first long tone section specifying means and a vibrato section is specified by the first vibrato section specifying means, the end time of the long tone section and the start time of the vibrato section are First detection means for determining whether or not a difference between the two is less than a predetermined value and detecting a combination of a long tone section and a vibrato section in which the determination result is positive;
First relative value calculating means for calculating a relative value between the time length of the long tone section and the time length of the vibrato section for the long tone section and the vibrato section corresponding to the combination detected by the first detecting means;
A second receiving means for receiving model voice data representing a singing voice as a model;
Second pitch detecting means for detecting the pitch of the model voice data received by the second receiving means;
When it is detected whether or not the state in which the pitch variation detected by the second pitch detection means is within a predetermined range has continued for a predetermined time or more, and the continuation for the predetermined time or longer is detected, the detection section Second long tone section specifying means for specifying as a long tone section;
When the pitch detected by the second pitch detection means periodically changes over a predetermined width, and the periodic fluctuation of the pitch is detected, the detected section is defined as a vibrato section. A second vibrato section identifying means for identifying;
When a long tone section is detected by the second long tone section specifying means and a vibrato section is specified by the second vibrato section specifying means, the end time of the long tone section and the start time of the vibrato section are Second detection means for determining whether the difference between the two is a predetermined value or less, and detecting a combination of a long tone section and a vibrato section in which the determination result is positive;
A second relative value for calculating a second relative value, which is a relative value of the time length of the long tone period and the time length of the vibrato period, for the long tone period and the vibrato period corresponding to the combination detected by the second detection unit. A value calculating means;
Evaluation means for calculating an evaluation of the singing voice based on a detection result of the detection means , wherein the relative value calculated by the first relative value calculation means and the second value calculated by the second relative value calculation means A singing evaluation apparatus comprising: an evaluation unit that compares the relative value of the singing voice and calculates the evaluation of the singing voice based on the comparison result . Receiving means for receiving singing voice data representing the singing voice;
Pitch detecting means for detecting the pitch of the singing voice data received by the receiving means;
It is detected whether or not the state in which the pitch variation detected by the pitch detection means is within a predetermined range has continued for a predetermined time or longer. A long tone section specifying means for specifying a section;
A vibrato that detects whether or not the pitch detected by the pitch detection means periodically fluctuates beyond a predetermined width, and identifies a detected section as a vibrato section when a periodic fluctuation of the pitch is detected Section identification means;
When a long tone section is detected by the long tone section specifying means and a vibrato section is specified by the vibrato section specifying means, a difference between the end time of the long tone section and the start time of the vibrato section is a predetermined value. Detecting means for determining whether or not a combination of a long tone period and a vibrato period in which the determination result is positive;
Relative value calculating means for calculating a relative value between the time length of the long tone section and the time length of the vibrato section for the long tone section and the vibrato section corresponding to the combination detected by the detecting means;
Evaluation means for calculating an evaluation of the singing voice based on a detection result of the detection means , comparing the relative value calculated by the relative value calculation means with a predetermined value, and based on the comparison result, the singing voice A singing evaluation apparatus comprising: an evaluation means for calculating the evaluation of Computer
First receiving means for receiving singing voice data representing the singing voice;
A first pitch detecting means for detecting the pitch of the singing voice data to which the first receiving means is received,
When it is detected whether or not the state in which the variation in pitch detected by the first pitch detecting means is within a predetermined range has continued for a predetermined time or more, and the continuation for the predetermined time or longer is detected, the detection section First long tone section specifying means for specifying as a long tone section;
If the pitch detected by the first pitch detection means periodically changes beyond a predetermined width, and the periodic fluctuation of the pitch is detected, the detected section is defined as a vibrato section. First vibrato section identifying means for identifying;
When a long tone section is detected by the first long tone section specifying means and a vibrato section is specified by the first vibrato section specifying means, the end time of the long tone section and the start time of the vibrato section are First detection means for determining whether or not a difference between the two is less than a predetermined value and detecting a combination of a long tone section and a vibrato section in which the determination result is positive;
First relative value calculating means for calculating a relative value between the time length of the long tone section and the time length of the vibrato section for the long tone section and the vibrato section corresponding to the combination detected by the first detecting means;
A second receiving means for receiving model voice data representing a singing voice as a model;
Second pitch detecting means for detecting the pitch of the model voice data received by the second receiving means;
When it is detected whether or not the state in which the pitch variation detected by the second pitch detection means is within a predetermined range has continued for a predetermined time or more, and the continuation for the predetermined time or longer is detected, the detection section Second long tone section specifying means for specifying as a long tone section;
When the pitch detected by the second pitch detection means periodically changes over a predetermined width, and the periodic fluctuation of the pitch is detected, the detected section is defined as a vibrato section. A second vibrato section identifying means for identifying;
When a long tone section is detected by the second long tone section specifying means and a vibrato section is specified by the second vibrato section specifying means, the end time of the long tone section and the start time of the vibrato section are Second detection means for determining whether the difference between the two is a predetermined value or less, and detecting a combination of a long tone section and a vibrato section in which the determination result is positive;
A second relative value for calculating a second relative value, which is a relative value of the time length of the long tone period and the time length of the vibrato period, for the long tone period and the vibrato period corresponding to the combination detected by the second detection unit. A value calculating means;
Evaluation means for calculating an evaluation of the singing voice based on a detection result of the detection means , wherein the relative value calculated by the first relative value calculation means and the second value calculated by the second relative value calculation means A program for functioning as an evaluation means for comparing the relative values of the singing voices and calculating the evaluation of the singing voice based on the comparison result . Computer
Receiving means for receiving singing voice data representing the singing voice;
Pitch detecting means for detecting the pitch of the singing voice data received by the receiving means;
It is detected whether or not the state in which the pitch variation detected by the pitch detection means is within a predetermined range has continued for a predetermined time or longer. A long tone section specifying means for specifying a section;
A vibrato that detects whether or not the pitch detected by the pitch detection means periodically fluctuates beyond a predetermined width, and identifies a detected section as a vibrato section when a periodic fluctuation of the pitch is detected Section identification means;
When a long tone section is detected by the long tone section specifying means and a vibrato section is specified by the vibrato section specifying means, a difference between the end time of the long tone section and the start time of the vibrato section is a predetermined value. Detecting means for determining whether or not a combination of a long tone period and a vibrato period in which the determination result is positive;
Relative value calculating means for calculating a relative value between the time length of the long tone section and the time length of the vibrato section for the long tone section and the vibrato section corresponding to the combination detected by the detecting means;
Evaluation means for calculating an evaluation of the singing voice based on a detection result of the detection means , comparing the relative value calculated by the relative value calculation means with a predetermined value, and based on the comparison result, the singing voice A program for functioning as an evaluation means for calculating the evaluation .

Images