JP2004138634A - "karaoke" singing equipment, program, and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform rating in which karaoke singing ability is suitably evaluated corresponding even to vibrato and figuration which are generally considered to be a skillful technique. <P>SOLUTION: When a special singing method like vibrato or figuration is used, the pitch data D1 of song voice changes roughly periodically in a pitch direction around a reference pitch. As a result, although the average value D31 of the absolute value D3 of a value obtained by subtracting rating reference data D2 from the pitch data D1 becomes a large value, the absolute value D42 of the average value D41 of D1-D2 is close to the rating reference data D2. On the other hand, when the same verification is performed to the case where a singer is a so-called poor singer who sings a song by deviating from the reference pitch without using the special singing method, D31, D42 become almost equal. Since the difference of them, D31-D42, shows different values clearly for both cases, presence of absence of use of the special singing method can be judged when a determination value for differentiating both is used in karaoke singing equipment. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to singing scoring in a karaoke apparatus, and more specifically, to a technique for obtaining an appropriate scoring result even when using a technique such as vibrato or fist.
[0002]
[Prior art]
Conventionally, a scoring function is well known as an additional function of a karaoke apparatus. This scoring function generates singing data indicating a singing state such as a pitch, a voice volume, or a tempo of a singer by sampling a singer's voice signal input from a microphone. The singing data is compared with scoring reference data such as main melody data in the karaoke data, and a predetermined score is given based on the comparison result to generate scoring data. When the singing part is completed, the scores in the scoring data are totaled to calculate a total score. The total score is displayed as it is on a scoreboard or a display, or a video reflecting the total score such as a video including a predetermined message or a predetermined expression content is output to a display (for example, see Patent Document 1).
[0003]
[Patent Document 1]
JP-A-11-305786
[0004]
[Problems to be solved by the invention]
However, in the conventional singing scoring method, superiority of singing ability is determined based on the degree of coincidence of the singing state with a uniquely determined scoring standard. For example, when looking at the degree of coincidence of pitches, it is determined that the singing ability is excellent when the difference between the two pitches is as close to zero as possible. Therefore, using techniques such as vibrato and fist, for example, the difference between the pitch data obtained from the voice signal of the singer and the pitch data of the singing melody as the scoring standard increases, There was a problem that even if you felt like you were singing well, your score would be bad.
[0005]
Therefore, an object of the present invention is to provide a karaoke apparatus capable of scoring with appropriate evaluation of singing ability in response to vibrato and fist which are generally good techniques when singing karaoke.
[0006]
Means for Solving the Problems and Effects of the Invention
(1) In a karaoke apparatus according to claim 1, which has been made to solve the above-mentioned problems, the music data storage means stores karaoke performance data for playing a karaoke music and pitch data of the singing melody of the karaoke music. The karaoke performance means reads out the karaoke performance data of the designated karaoke music from the music data storage means and performs the karaoke performance. When a karaoke singing voice signal is input via the voice signal input means, the pitch extracting means samples the voice signal and extracts the karaoke singing pitch data. Then, the scoring means reads out, from the music data storage means, scoring reference data corresponding to the performance music in synchronization with the karaoke performance by the karaoke playing means, and reads out the read scoring reference data and the pitch data extracted by the pitch extracting means. The scoring is based on the following, but in more detail, the following scoring is performed.
[0007]
In other words, the first pitch difference calculating means calculates the difference between the plurality of pitch data extracted by the pitch extracting means during the predetermined period and the corresponding pitch data in the scoring reference data read from the music data storage means. The absolute values of the differences are calculated, and the average of the absolute values is calculated. Also, the second pitch difference calculating means may calculate the plurality of pitch data extracted by the pitch extracting means during a predetermined period and the corresponding pitch data in the scoring reference data read from the music data storage means. Calculate the absolute value of the average of the differences. Then, the special singing method determining means determines whether the difference between the pitch difference calculated by the first pitch difference calculating means and the pitch difference calculated by the second pitch difference calculating means is a special pitch using vibrato or fist. If the singing method other than the singing method is larger than the determination value that cannot be obtained, it is determined that the method is a special singing method using vibrato or fist. If the special singing method determining means determines that the method is a special singing method using vibrato or fist, the score is calculated using the pitch difference calculated by the second pitch difference calculating means.
[0008]
Here, the special singing method determining means determines whether or not the special singing method is a special singing method using vibrato or fist based on the difference between the pitch differences calculated by the first and second pitch difference calculating means. The reason why the determination can be made will be described.
[0009]
First, considering the characteristics of the singing voice signal when using techniques such as vibrato and fist, when focusing on the pitch, almost the same difference in the pitch direction around the reference pitch A "sinusoidal" waveform signal that changes almost periodically between pitches having a. In general, the sampling interval for scoring is shorter than this change period and does not coincide with the change period. A situation occurs that is significantly deviated from the height. However, when looking at the entire vibrato portion, the state deviates from the reference pitch to a higher side at one sampling timing, the state deviates from the reference pitch to a lower side at another sampling timing, and the reference tone at another sampling timing. A state in which the pitch matches or is very close to the pitch is obtained. When the pitches of the entire vibrato portion are averaged, they are close to the reference pitch.
[0010]
On the other hand, if you do not use techniques such as vibrato or fist, but sing without the reference pitch, so-called `` poor '', at almost all sampling timings, Will be off the same side. In other words, if it deviates to the higher side from the reference pitch, it deviates to the higher side by almost the same amount at almost all sampling timings, and if it deviates to the lower side from the reference pitch, almost all sampling timings At the same time, the sign of the pitch difference is the same, and the absolute value of the pitch difference is also substantially equal.
[0011]
Based on these analyses, it was thought that the difference between the two could be reflected by the following physical quantities. In other words, when using a technique such as vibrato or fist, and singing at the correct pitch, the waveform signal becomes a "sine wave" as described above, and thus, at a predetermined sampling interval from the singing voice signal. For the plurality of pitch data obtained during the predetermined period, the average value of the absolute value of the difference from the pitch data serving as the grading reference is relatively large, but the average of the difference from the pitch data serving as the grading reference is averaged. The absolute value of the value becomes relatively small. On the other hand, when singing with the reference pitch removed without using techniques such as vibrato or fist, the average value of the absolute value of the difference from the pitch data serving as the scoring standard and the scoring standard are compared. Both absolute values of the average value of the difference from the pitch data become relatively large values. Therefore, it is considered that the difference can be determined by focusing on these differences, and the above-described determination method of the special singing method determination unit is employed. The “judgment value that cannot be obtained by a singing method other than the special singing method using vibrato or fist” used for this judgment may be obtained by, for example, an experiment. For example, using a technique such as vibrato or fist and singing at the correct pitch, the average value of the absolute value of the difference from the pitch data serving as the scoring reference is calculated. If it is, for example, about two tones (four semitones), the decision value is set to three semitones.
[0012]
If the singing voice signal is analyzed by frequency analysis or the like to obtain signal waveform information, it is possible to determine the presence or absence of vibrato from the waveform. However, in this case, it is necessary to perform frequency analysis such as FFT, and the amount of calculation becomes relatively large. On the other hand, in the case of the present invention, since it can be dealt with by the simple four arithmetic operations, it is very advantageous in that it is possible to perform appropriate scoring by virtue of the presence or absence of vibrato by relatively simple calculation.
[0013]
Then, in the karaoke apparatus of the present invention, when it is determined that the special singing method using vibrato or fist, scoring is performed using the pitch difference calculated by the second pitch difference calculating means. This is because the pitch difference calculated by the second pitch difference calculating means is viewed as a whole, as can be seen from the characteristic analysis of the voice signal of the singing when the technique such as vibrato or fist described above is used. This is because it is considered that this reflects the degree of coincidence with the reference pitch in the case where the pitch is changed. By scoring in this manner, when singing a karaoke song, scoring that appropriately evaluates the singing ability can be performed in response to vibrato and fist, which are generally good techniques.
[0014]
(2) As described in claim 2, when the special singing method determining unit determines that the special singing method uses vibrato or fist, scoring is performed by adding a predetermined score. Good. As described above, scoring using the pitch difference calculated by the second pitch difference calculating means is an unfair disadvantageous scoring in the case of a special singing method using vibrato or fist. It is a reluctant response to not be done. However, techniques such as vibrato and fist are generally considered to be good techniques, so scoring more favorably than singing without such techniques or saying `` It is effective from the viewpoint of "scoring evaluated". Therefore, techniques such as vibrato or fist are set as points to be added, and when such a technique is used, a predetermined score is added.
[0015]
In addition, such a score as a scoring result may be displayed on a display or the like as in the past, but in this case, if a technique such as vibrato or fist is recognized and added, for example, the fact is added. May be displayed together. In this way, the singer knows that his technique has been recognized and added points, and the singer is more satisfied.
[0016]
(3) By the way, in the scoring, with respect to the predetermined period during which the pitch extracting means extracts a plurality of pieces of pitch data as calculation targets of the first pitch difference calculating means and the second pitch difference calculating means, as follows. Some ideas can be considered.
For example, a period in which the pitch data of the same pitch to be calculated in the scoring reference data read from the music data storage means continues is adopted as the “predetermined period”. When the same pitch continues, there is a possibility that singing using vibrato or fist is performed in part or all of the period. Therefore, by targeting all parts during this period, scoring that appropriately evaluates the singing ability can be performed regardless of where vibrato or fist is used.
[0017]
Of course, it is also possible to adopt not a whole period of the pitch data of the same pitch to be calculated in the scoring reference data but a part of the period as the “predetermined period”. However, in order to properly grasp the property that the absolute value of the average value of the difference from the pitch data serving as the scoring reference becomes relatively small, among the “sinusoidal” waveform signals as described above, This is preferable because it is considered in the range of an integral multiple of one cycle. However, in order to detect one cycle, if the singing voice signal is analyzed by frequency analysis or the like to obtain signal waveform information as described above, it is necessary to perform frequency analysis such as FFT. The amount increases.
[0018]
Thus, the difference between the pitch data extracted by the pitch extracting means and the corresponding pitch data in the scoring reference data read from the music data storage means is converted to positive or negative. It is conceivable to estimate one cycle of a pitch waveform due to vibrato or fist based on the timing of the execution. For example, the period from the timing of inversion from positive to negative to the next timing of inversion from positive to negative may be calculated as one cycle.
[0019]
The predetermined period set in this way is set as one comparison target range, and during a period in which pitch data of the same pitch continues, the average value of the pitch difference is calculated while sequentially shifting the comparison target range. It is also conceivable to employ a so-called “moving average” technique.
[0020]
(4) As described in claim 5, when the scoring means in the karaoke apparatus according to any one of claims 1 to 4 is realized by a computer, it can be provided as, for example, a program executed by the computer. Such a program is recorded on a computer-readable recording medium such as a flexible disk, a magneto-optical disk, a CD-ROM, a hard disk, a ROM, and a RAM, and loaded into the computer as necessary. By executing the program, or loading and executing the program via a network, a function as a scoring unit can be realized.
[0021]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments to which the present invention is applied will be described with reference to the drawings. It is needless to say that the embodiments of the present invention are not limited to the following examples, and can take various forms within the technical scope of the present invention.
[0022]
FIG. 1 is a block diagram illustrating the configuration of the karaoke apparatus according to the present embodiment. In this embodiment, a plurality of karaoke apparatuses 1a, 1b, 1c... Are connected to a local area network (LAN) 30 to form a karaoke system. Although the basic configuration is the same for all karaoke apparatuses 1, only the karaoke apparatus 1a can be connected to the host computer 2 via the communication network 3. Then, the karaoke apparatus 1a can acquire music information and image information related to karaoke from the host computer 2 connected via the communication network 3. The music information and the image information relating to the karaoke acquired by the karaoke apparatus 1a can be acquired by the other karaoke apparatuses 1b, 1c,.
[0023]
The host computer 2 can access the karaoke apparatus 1a via the communication network 3 and sends the karaoke apparatus 1a the music data such as the latest popular music or the number of times of what music. Related information including log data indicating whether or not a performance has been performed can be received from the karaoke apparatus 1a and managed. In this case, the “log data indicating what music has been played and how many times” means not the log data of the karaoke apparatus 1a alone but the log data of the entire karaoke system. That is, the log data of the other karaoke apparatuses 1b, 1c,... Is also collected in the karaoke apparatus 1a, and the log data of the entire karaoke system is transmitted from the karaoke apparatus 1a to the host computer 2.
[0024]
Further, the host computer 2 has a database, which stores music information as content data used for music performance, image information such as background images or lyrics, and the like. In addition, the host computer 2 stores the upgraded system program and the like in addition to the content data in a database, reads out the database from the database as needed, and transmits the system program to the karaoke apparatus 1.
[0025]
Next, the configuration of the karaoke apparatus 1a will be described.
As shown in FIG. 1, the karaoke apparatus 1a is connected to a host computer 2 via a communication network 3 to transmit and receive various kinds of information, an operation panel 10 for performing music reservation and the like, a karaoke apparatus 1a. CPU 14 for overall control, RAM 15 for temporarily storing various information, sound source reproducing device 18 for reproducing performances, amplifier 20 for amplifying electric signals related to music information, and inputting electric signals from amplifier 20 A speaker 22 that outputs accompaniment music and the singing voice of the user, a microphone (hereinafter simply referred to as a microphone) 23 for inputting the singing voice of the user and the like to the amplifier 20, and a singing voice input from the microphone 23 is sampled. Scoring unit 28 that extracts singing pitches by analysis and compares it with scoring reference data to score, music information and image information for karaoke, etc. EEPROM 12 and hard disk 13 storing seed data, video reproducing device 24 for visualizing image information and the like, display device 26 for displaying background information and lyrics as image information, and network for connecting to LAN 30 An interface 17 is provided. The operation panel 10, the EEPROM 12, the hard disk 13, the RAM 15, the network interface 17, the sound source playback device 18, the video playback device 24, and the scoring unit 28 are connected to the CPU 14.
[0026]
Since the other karaoke devices 1b, 1c,... Do not include the communication device 19 but include other components, only the karaoke device 1a will be described here, and the configuration of the other karaoke devices 1b, 1c,. The description of is omitted.
The communication device 19 is a modulation / demodulation device that modulates and demodulates a signal, and is configured to be accessible to the host computer 2 through the communication network 3 under the control of the CPU 14. Thus, the communication device 19 can receive music data and the like transmitted from the host computer 2 via the communication network 3 and transmit the related information to the host computer 2.
[0027]
The operation panel 10 is operated by the user to select an arbitrary song, adjust keys for performance sounds, adjust a volume balance between performance and song, and perform various adjustments such as echo, volume, and tone. An operation unit and a display unit (not shown) for displaying, for example, a selected music number are provided. By operating the operation unit, the user can make a reservation registration of the music to be reproduced in the karaoke apparatus 1.
[0028]
The EEPROM 12 stores a system program and setting data necessary for various settings.
The hard disk 13 stores content data such as music information and image information and log data such as performance records. When a song is selected via the operation unit of the operation panel 10, the CPU 14 calls up the music information including the lyrics data, the image data including the video data, and the music information including the performance data from the hard disk 13, and The data is output in synchronization with the sound source reproducing device 18. The hard disk 13 stores performance data, lyrics data, and the like corresponding to a song number as identification information.
[0029]
Thereafter, the performance data output from the CPU 14 is converted into an analog performance sound signal in the sound source reproducing device 18 and then sent to the amplifier 20 to be electrically amplified. The amplifier 20 mixes the singing sound signal of the user (singer) input via the microphone 23 at an appropriate ratio, and the mixed singing sound signal and performance sound signal are transmitted from the amplifier 20 to the speaker 22. And output as voice and performance sound from the speaker 22 to the outside.
[0030]
On the other hand, the video reproduction device 24 reproduces an image based on the image information read from the hard disk 13 under the control of the CPU 14. Thereby, the lyrics data output by the CPU 14 is combined with the video data in the video playback device 24, and the lyrics telop is displayed on the screen of the display device 26 together with the background video.
[0031]
With such a configuration, the user can sing using the microphone 23 in accordance with the karaoke performance played from the speaker 22 while referring to the lyrics telop displayed on the display device 26.
Further, the karaoke apparatus 1a of the present embodiment has a singing scoring function for scoring the singing state of the person singing along with the karaoke performance and outputting the scoring result. When the user wants to use the singing score function with the karaoke apparatus 1a, he or she presses a predetermined operation key following a series of music numbers for performance reservation from an operation panel 10 such as a panel or a remote controller. Thereby, the music number to which the code indicating the operation command of the singing scoring function is added is transferred to the CPU 14. When performing the music performance processing, the CPU 14 performs the performance reservation processing on the assumption that the singing scoring function is activated. Of course, the song grading function may be activated by a user's billing procedure or the like. For example, a separate charging device is connected to the karaoke device 1a, and a singing scoring function is activated by inputting a predetermined fee to the charging device at an appropriate time, such as immediately before the music to be scored is played or during a prelude. Such a configuration may be adopted.
[0032]
The scoring unit 28 performs singing scoring. The scoring unit 28 is configured to input a singing voice from the microphone 23. The scoring unit 28 extracts the singing pitch and rhythm as a scoring element by sampling and analyzing the input singing voice. On the other hand, the accompaniment music generation data stored in the hard disk 13 is received via the data bus and the CPU 14, and the vocal data (singing melody data) in the data is acquired as scoring reference data. Then, the scoring element based on the acquired scoring reference data is compared with the scoring element extracted from the above-mentioned singing voice, and the singing voice is scored based on how close it is to the vocal data. The vocal data is also used to realize a function of outputting the accompaniment music sound from the speaker 22 with a relatively small volume as a so-called guide melody function.
[0033]
The data of the scoring result is transferred to the CPU 14, and the CPU 14 receives the transferred scoring data and stores it in the RAM 15. When detecting the end point of the scoring section in the karaoke music, the CPU 14 causes the scoring unit 28 to stop generating scoring data, accumulates scoring data accumulated in the RAM 15 during the scoring section, and displays the scoring result on the display device 26. Display.
[0034]
This is the description of the configuration and schematic operation for scoring, but in the karaoke apparatus 1a of the present embodiment, it is possible to judge vibrato / fist and perform more appropriate scoring. This is in view of the following problems. In other words, the conventional method of judging the degree of coincidence between the pitch data extracted from the singing voice and the pitch data of the vocal data in the accompaniment music generation data requires a technique such as vibrato or fist. The difference from the high data increases. Therefore, even if the person listening on the side feels that they are singing well, the score will be bad. Therefore, in the karaoke apparatus 1a of the present embodiment, it is possible to make a determination in consideration of the presence of vibrato and fist in a part for determining the degree of coincidence of pitch data.
[0035]
This will be described with reference to FIGS. FIG. 2 is a flowchart illustrating an algorithm for judging the degree of coincidence of pitch data in the scoring process performed by the scoring unit 28 in the embodiment. FIG. 3 is a diagram illustrating pitch data D1 extracted from singing voice. It is explanatory drawing of the scoring method based on scoring reference data D2. A pitch difference between the pitch data D1 and the scoring reference data D2 is calculated for each note, and a predetermined score is given according to the obtained pitch difference. In the example shown in FIG. 3, a vibrato / fist exists in the latter half of the last note. FIG. 4 is an explanatory diagram of a determination method according to the present embodiment, and FIG. 5 is an explanatory diagram of a conventional determination method.
[0036]
As shown in FIG. 2, the scoring unit 28 first takes in a singing voice signal input from the microphone 23 (S102). Then, the captured singing voice signal is sampled and analyzed at a predetermined timing (for example, a constant cycle) to extract pitch data D1 (S104). Further, the accompaniment music generation data is received via the data bus, and the pitch data in the vocal data in the data is acquired as the scoring reference data D2 (S106).
[0037]
In the next S108, the absolute value D3 of the value obtained by subtracting the scoring reference data D2 corresponding to the singing timing at which the pitch data D1 was obtained from the pitch data D1 extracted from the singing voice is calculated. , An average value D31 of the absolute value D3 is calculated. For the “predetermined period”, for example, it is conceivable to adopt a period in which the pitch data of the same pitch continues in the scoring reference data D2 to be calculated. When the same pitch continues, there is a possibility that singing using vibrato or fist is performed in part or all of the period. Therefore, by targeting all parts during this period, scoring that appropriately evaluates the singing ability can be performed regardless of where vibrato or fist is used.
[0038]
The calculation of the average value D31 in S108 will be specifically described with reference to FIG.
The vertical axis in FIG. 4A indicates the magnitude of the frequency, the horizontal axis indicates the time axis, and corresponds to the frequency acquired in S106 of the scoring reference data D2. The black circles are the pitch data D1 of the singing voice extracted in S104, and the curve connecting these black circles continuously indicates the frequency of the singing voice when singing using vibrato or fist. . Then, the absolute value D3 of the value obtained by subtracting the scoring reference data D2 from the pitch data D1 is indicated by a black circle in FIG. 4B. The vertical axis in FIG. Is the time axis, and the average value D31 of the absolute value D3 is indicated by a dashed line, and is shown in a bar graph by drawing a line from the time axis to a black circle in FIG. The average length value corresponds to D31.
[0039]
Returning to the description of the flowchart of FIG. 2, in the next S110, a value D4 is calculated by subtracting the scoring reference data D2 corresponding to the singing timing at which the pitch data D1 was obtained from the pitch data D1 extracted from the singing voice. Then, the absolute value D42 of the average value D41 of the subtraction value D4 during the predetermined section is calculated. This "predetermined period" is the same as the predetermined period in S108.
[0040]
The subtraction value D4 in S110 is indicated by a black circle in FIG. Note that the vertical axis in FIG. 4B is the frequency, and the horizontal axis is the time axis. Further, in FIG. 4C, the average value D41 of the subtraction value D4 is indicated by a dashed line, and the absolute value D42 of the average value D41 is indicated by a two-dot chain line.
[0041]
In the next S112, it is determined whether or not a value obtained by subtracting the absolute value D42 (of the average value D41) calculated in S110 from the average value D31 (of the absolute value D3) calculated in S108 is equal to or more than a predetermined value α. As the predetermined value α, for example, a value such as 3 degrees (3 semitones) can be considered. This predetermined value α will be supplemented later.
[0042]
If (D31−D42) ≧ α (S112: YES), since this one sound is considered to be a special singing method using vibrato or fist, it is calculated in S110 (of the average value D41). The value D42 is set as the difference from the pitch used as the scoring reference (S114). Then, a predetermined point is added by using the special singing method using vibrato or fist (S116). This additional point gives a certain score.
[0043]
In this case, the scoring unit 28 performs scoring based on the pitch difference obtained in S114 of FIG. 2 and adds a score to be added obtained in S116 in the main scoring process executed separately, and performs scoring. Do. In addition, the score as a result of the scoring as described above is displayed on the display device 26. If the score is added through the process of S116, the fact is added (for example, “O points added by vibrato or fist”). Etc.) may be displayed. In this way, the singer can know that a technique such as vibrato or fist has been recognized and added points, and the singer's satisfaction increases.
[0044]
On the other hand, if (D31-D42) <α (S112: NO), it is considered that this one sound has not been subjected to the special singing method using vibrato or fist, and thus is calculated in S108 (the absolute value D3). ) The average value D31 is set as the difference from the pitch used as the scoring reference (S118). The scoring unit 28 in this case performs scoring based on the pitch difference obtained in S118 in FIG. 2 in a main scoring process that is separately executed.
[0045]
Here, as described in the description of S112, if (D31−D42) ≧ α, it is considered that the special singing method uses vibrato or fist, and if (D31−D42) <α, vibrato or fist. The reason why it is considered that the special singing method using is not performed will be described.
[0046]
Focusing on the pitch of the singing voice when using techniques such as vibrato or fist, the pitch between the pitches with almost the same difference in the pitch direction around the reference pitch changes almost periodically The result is a "sinusoidal" waveform signal. That is, it is a waveform signal as shown in FIG. Since the sampling interval for scoring is shorter than this change period, it does not coincide with the change period. Therefore, as can be seen from FIG. 4 (a), there is a state in which, at a certain sampling timing, the reference pitch is greatly deviated, although the singing is performed at the correct pitch as a whole. However, when looking at the entire vibrato portion, the state deviates from the reference pitch to a higher side at one sampling timing, the state deviates from the reference pitch to a lower side at another sampling timing, and the reference tone at another sampling timing. High or very close states are obtained. Accordingly, when the pitches of the entire vibrato portion are averaged, the pitch is a reference pitch, such as D41 (= D4 (= D1-D2) average value) indicated by a dashed line in FIG. 4C. It becomes close to the scoring reference data D2.
[0047]
For comparison, a similar test will be conducted for a so-called “bad” case in which the user does not use techniques such as vibrato and fist, but sings with the reference pitch removed. FIG. 5A shows pitch data D1 and scoring reference data D2 when singing with a reference pitch removed without using a technique such as vibrato or fist. The characteristic of the pitch data D1 extracted from the singing voice in this case is that the pitch data D1 deviates to the same side with respect to the reference pitch at almost all sampling timings. That is, when the reference pitch is deviated to a higher side (a side having a higher frequency with respect to the scoring reference data D2), the deviates to the higher side by almost the same amount at almost all sampling timings. In the case where the pitch deviates to the lower side (the side where the frequency is smaller than the scoring reference data D2), it deviates to the lower side by almost the same amount at almost all sampling timings. The absolute values of the differences are also substantially equal.
[0048]
When the processing of S108 and S110 in FIG. 2 is performed on such data D1 and D2, the “absolute value D3 of the value obtained by subtracting the scoring reference data D2 from the pitch data D1” in S108 is shown in FIG. Since this is indicated by a black circle in b), the average value D31 of the absolute value D3 is indicated by a dashed line in FIG. On the other hand, since the “value D4 obtained by subtracting the scoring reference data D2 from the pitch data D1” in S110 is indicated by a black circle in FIG. 5C, the absolute value D42 of the average value D41 of the subtraction value D4 is shown in FIG. Is shown by a two-dot chain line. That is, D31 and D42 have the same value.
[0049]
Due to the difference between D31 and D42 shown in FIGS. 4 and 5, singing is performed without the reference pitch without using a technique such as vibrato or fist as shown in FIG. In this case, the value of (D31-D42) becomes almost zero, while singing well in a state close to the reference pitch using a technique such as vibrato or fist as shown in FIG. In this case, the value of (D31-D42) becomes a certain size. For example, if the width of the vibrato is about two to two and a half (4 to 5 semitones) with respect to the reference pitch, the value of (D31-D42) will be one and a half (3 semitones) depending on the sampling timing. ). Such a value is a value that cannot be obtained when singing out of the reference pitch without using a technique such as vibrato or fist as shown in FIG. Therefore, for example, the predetermined value α for the determination in S112 of FIG. 2 may be set to about three semitones. Of course, the width of the vibrato may vary depending on the song or singer, and may be smaller than the above example. Therefore, the predetermined value α for determination is a value such as two semitones or one semitone smaller than the above example. May be adopted.
[0050]
When determining the determination value α, it may be determined by, for example, an experiment. Using a technique such as vibrato or fist and singing at the correct pitch, the average value of the absolute value of the difference from the pitch data serving as the scoring reference is calculated. For example, if it is about two tones (four semitones), the judgment value is set to three semitones. If it is about one sound (two semitones), the judgment value is made one semitone. . For example, the analysis may be based on the degree of vibrato or fist of the singer singing the original song. The singing voice portion (vocal portion) is extracted from the music information containing the singing voice of the singer, and the value of (D31-D42) is calculated in the vibrato and fist portions. It is also conceivable to calculate such a sample value based on the vocal portions of a plurality of singers and take an average of them to determine the predetermined value α for determination.
[0051]
As described above, according to the karaoke apparatus 1a of the present embodiment (the same applies to the other karaoke apparatuses 1b, 1c, etc.), the following effects can be obtained.
(1) When singing karaoke, scoring that appropriately evaluates singing ability can be performed in response to vibrato and fist, which are generally good techniques.
[0052]
Note that if the singing voice signal is analyzed by frequency analysis or the like to obtain signal waveform information, the presence or absence of vibrato can be determined from the waveform, but in this case, it is necessary to perform frequency analysis such as FFT. Yes, the amount of calculation is relatively large. On the other hand, in the case of the present embodiment, as can be understood from the description of S108, S110, and S112 in FIG. It is very advantageous in that it can perform scoring.
[0053]
(2) In addition, as shown in S116 of FIG. 2, when a special singing method using vibrato or fist is performed, points are added. Techniques such as vibrato and fist are generally considered to be good techniques, so scoring more favorably than singing without such techniques means that `` the singing ability was properly assessed. "Scoring" can be realized.
[0054]
In the present embodiment, the hard disk 13 corresponds to “song data storage means”, and the CPU 14 and the sound source reproducing device 18 correspond to “karaoke performance means”. The microphone 23 corresponds to an audio signal input unit, and the scoring unit 28 corresponds to a pitch extracting unit and a scoring unit. In addition, among the processing in FIG. 2, S108, S110, and S112 respectively execute processing as “first pitch difference calculating means”, “second pitch difference calculating means”, and “special singing method determining means”. Is equivalent to
[0055]
Although the embodiments have been described above, the present invention is not limited to the above-described embodiments, and can be implemented in various modes. I will explain some of them.
(1) In the above embodiment, the pitches having the same pitch in the scoring reference data D2 are defined as “predetermined sections” of the data to be calculated for the average value D31 in S108 and the absolute value D42 in S110 in FIG. A period during which the data continued was adopted. However, instead of the entire period in which the pitch data of the same pitch continues, a part of the period may be adopted as the “predetermined period”. However, since it is determined whether or not the special singing method using vibrato or fist is used based on the magnitude of the value of (D31-D42) in S112 in FIG. 2, the following point is devised. It is desirable to apply.
[0056]
In other words, the vibrato or fist portion is a “sinusoidal” waveform signal that changes almost periodically between pitches having substantially the same difference in the pitch direction around the reference pitch, so that the whole 4C, the absolute value D42 of the average value D41 of the difference D4 between the pitch data D1 based on the singing voice and the scoring reference data D2 is relatively small, as indicated by the two-dot chain line in FIG. It has the property of becoming. Since this characteristic is used to determine whether or not a special singing method is used, in order to properly grasp the characteristics of vibrato or fisting, the “sine wave” waveform signal must be an integral multiple of one cycle. It is preferable because it is considered in the range. For example, FIG. 4 shows data for three cycles. Actually, even if the vibrato or fist portion has a longer period, the calculation may be performed focusing on, for example, only three periods.
[0057]
However, if a singing voice signal is analyzed by frequency analysis or the like to obtain signal waveform information in order to detect one cycle, it is necessary to perform frequency analysis such as FFT, which requires a relatively large amount of calculation. . For this reason, for example, after extracting the pitch data D2 in S104 of FIG. 2 and obtaining the scoring reference data D2 in S106, one cycle of the pitch waveform is changed at the timing when the difference (D1-D2) changes between positive and negative. It can be estimated. For example, the period from the timing of inversion from positive to negative to the next timing of inversion from positive to negative may be calculated as one cycle. Of course, conversely, the period from the timing of inversion from negative to positive to the next timing of inversion from negative to positive may be calculated as one cycle. Further, for example, a period from the timing when the signal is inverted from positive to negative to the next timing when the signal is inverted from negative to positive may be set as a half period, and the period may be doubled to calculate one period. Then, the above-mentioned predetermined period is set to an integral multiple of one cycle estimated by the cycle estimating means.
[0058]
Of course, the change period is not always constant even though it is a "sine wave" waveform signal, but considering the actual singing situation, the change period when singing using vibrato or fist is almost constant Since it is assumed that there are many cases, there is no particular problem even with the above-described method.
[0059]
Then, the processing of S108 to S112 in FIG. 2 is performed in the “predetermined period” of an integral multiple of one cycle set as described above. In addition, when only a part of the period in which the pitch data of the same pitch continues was determined to be the determination section, only the determination section happened to use a special singing method in another section instead of a special singing method such as vibrato. It is also possible. Therefore, for example, in the predetermined determination section appropriately extracted from the first half and the second half of the period in which the pitch data of the same pitch continues, the processing of S108 to S112 in FIG. If the method is acceptable, measures to add points may be taken.
[0060]
In addition, even when the entirety of the period in which the pitch data of the same pitch continues is used as the determination section, or when a part of the data is used as the determination section, the following “moving average” technique is used. It is also possible. For example, in the case where the entirety of the period in which the pitch data of the same pitch continues is the determination section, the predetermined period of two cycles is set as one comparison target range, and the period in which the pitch data of the same pitch continues. In the meantime, the average value of the pitch difference is calculated while sequentially shifting the comparison target range.
[0061]
(2) In the above embodiment, the tune number and the like are input by the operation unit of the operation panel 10 provided on the main body of the karaoke apparatus 1. However, for example, a remote controller or the like connected by infrared communication or wireless communication based on the Bluetooth standard is used. The above-mentioned operation buttons may be prepared, and music selection data based on the operation may be transmitted to the karaoke apparatus main body side.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a schematic configuration of a karaoke apparatus according to an embodiment.
FIG. 2 is a flowchart illustrating an algorithm for determining the degree of coincidence of pitch data in the scoring process performed in the karaoke apparatus of the embodiment.
FIG. 3 is an explanatory diagram of a scoring method based on pitch data D1 extracted from singing voice and scoring reference data D2.
FIG. 4 is an explanatory diagram of a determination method according to the present embodiment.
FIG. 5 is an explanatory diagram of a conventional determination method.
[Explanation of symbols]
1a, 1b, 1c, karaoke device, 2 host computer, 3 communication network, 10 input device, 12 EEPROM, 13 hard disk, 14 CPU, 15 RAM, 17 network interface, 18 sound source reproduction Device, 20 amplifier, 22 speaker, 23 microphone, 24 video playback device, 26 display device, 28 scoring unit.

Claims (6)

Song data storage means for storing karaoke performance data for playing a karaoke song and scoring reference data including pitch data of singing melody of the karaoke song;
Karaoke performance means for performing karaoke performance by reading karaoke performance data of the designated karaoke music from the music data storage means;
Voice signal input means for inputting a voice signal of karaoke singing;
Pitch extraction means for sampling the audio signal input via the audio signal input means and extracting the pitch data of the karaoke song,
In synchronization with the karaoke performance by the karaoke performance means, the scoring reference data corresponding to the performance music is read from the music data storage means, and the read scoring reference data and the pitch data extracted by the pitch extraction means are read out. Scoring means for scoring based on
A karaoke device comprising:
The scoring means,
Calculating the absolute value of the difference between the plurality of pitch data extracted by the pitch extraction means during a predetermined period and the corresponding pitch data in the scoring reference data read from the music data storage means; First pitch difference calculating means for calculating an average value of the absolute values,
Calculating an absolute value of an average value of a difference between a plurality of pitch data extracted by the pitch extracting means during a predetermined period and corresponding pitch data in the scoring reference data read from the music data storage means; Second pitch difference calculating means;
In a singing method other than the special singing method using vibrato or fist, a difference between the pitch difference calculated by the first pitch difference calculating means and the pitch difference calculated by the second pitch difference calculating means is different. Special singing method determining means to determine that it is a special singing method using the vibrato or fist, if the determination value is larger than an impossible value,
If the special singing method determining means is determined to be a special singing method using vibrato or fist, the scoring using the pitch difference calculated by the second pitch difference calculating means A karaoke apparatus characterized by performing. The karaoke apparatus according to claim 1,
The karaoke apparatus, wherein the scoring means performs scoring by adding a predetermined score when the special singing method judging means judges that the special singing method uses vibrato or fist. The karaoke device according to claim 1 or 2,
A predetermined period during which the pitch extracting means extracts a plurality of pitch data as a calculation target of the first pitch difference calculating means and the second pitch difference calculating means is a scoring read out from the music data storage means. A karaoke apparatus characterized by a period in which pitch data of the same pitch to be calculated in the reference data continues. The karaoke device according to claim 1 or 2,
The scoring means,
Further, based on the timing at which the difference between the pitch data extracted by the pitch extracting means and the corresponding pitch data in the scoring reference data read from the music data storage means changes between positive and negative, the sound generated by vibrato or fisting is used. A period estimating means for estimating one period of the high waveform,
The predetermined period during which the pitch extracting means extracts a plurality of pitch data as a calculation target of the first pitch difference calculating means and the second pitch difference calculating means is one cycle estimated by the cycle estimating means. A karaoke apparatus characterized by being an integral multiple of. A program for causing a computer to function as the scoring unit in the karaoke apparatus according to claim 1. A recording medium in which a program for causing a computer to function as the scoring unit in the karaoke apparatus according to claim 1 is recorded.

Images