JP2011075619A - Musical score tracing device - Google Patents

Abstract

A music score tracking technique having high followability and a small amount of calculation is provided.
An evaluation of each position of a score is performed by adding 1 if the pitch data of the score and the played pitch data match, and subtracting 1 if they do not match. In a score tracking apparatus that calculates a value and estimates the score position that maximizes the evaluation value as the current score position, an upper limit is set for the evaluation value. For example, the upper limit is set to 5. By setting an upper limit for the evaluation value, if the performer changes the performance position in the middle, the time until the evaluation value for the correct performance position after the change becomes smaller than the evaluation value for the correct performance position after the change. It becomes shorter and the time to follow the correct performance position after the change is shortened.
[Selection] Figure 7

Description

  The present invention relates to a technique for estimating where a performer is performing on a score.

  A musical score tracking apparatus described in Patent Document 1 is known as a conventional technique. In the score tracking device described in Patent Document 1, 1 is added when the pitch data of the score and the played pitch data match, and 1 is subtracted when they do not match. Thus, the evaluation value at each position of the score is calculated, and the position of the score that maximizes the evaluation value is estimated as the current score position.

  For example, as shown in FIG. 18, when the pitch data of the score is “Domicoside Red ...” and the played pitch data is also “Domisoside Red ...”, the position of the number surrounded by the line in FIG. Estimated as position. The numbers in FIG. 18 represent the evaluation values at each position of the score.

  Hereinafter, the time when the played pitch data “SO” was input will be described as an example. When the third played pitch data “SO” is input, the evaluation value at each position of the score is calculated. If the yth position of the score in the xth played pitch data is expressed as (x, y), the evaluation value of each position (3, y) (y = 1,..., Y) of the score is calculated. The Y is the total number of musical pitch data.

  The evaluation value at the position (3, 1) is 0 because the pitch data “SO” played at the position (3, 1) does not match the pitch data “DO” of the score. The evaluation value at the position (3, 2) does not match the pitch data “mi” played at the position (2, 1) and the pitch data “do” of the score, and the performance at the position (3, 2). The pitch data “So” thus set and the pitch data “Mi” of the score do not coincide with each other, and thus become zero. The evaluation value at the position (3, 3) indicates that the pitch data “do” played at the position (1, 1) matches the pitch data “do” of the musical score, and the evaluation is performed at the position (2, 2). The pitch data “Mi” matches the pitch data “Mi” of the score, and the played pitch data “So” at the position (3, 3) matches the pitch data “So” of the score. Therefore, 3 = 1 + 1 + 1. In this way, the evaluation value of each position (3, y) (y = 1,..., Y) of the score is calculated.

  In this example, the evaluation value 3 at the position (3, 3) is the maximum among the evaluation values at each position (3, y) (y = 1,..., Y) of the score. Therefore, the third position of the score is estimated as the position of the score at the time when the pitch data “So” is input.

U.S. Pat. No. 4,745,836

  However, in the score tracking device described in Patent Document 1, the evaluation value increases without limit when the correct performance continues, and it is difficult to follow when the performer changes the performance position on the way. There was a problem.

  In order to solve the above problem, an upper limit value is provided for the evaluation value.

  By setting an upper limit for the evaluation value, if the performer changes the performance position in the middle, the time until the evaluation value for the correct performance position after the change becomes smaller than the evaluation value for the correct performance position after the change. It becomes shorter and the time to follow the correct performance position after the change is shortened.

The functional block diagram of the example of a score tracking device. The functional block diagram of the example of the accompaniment reproducing | regenerating apparatus using a score tracking device. The functional block diagram of the example of the score display apparatus using a score tracking apparatus. The flowchart of the example of a score tracking method. The flowchart which illustrates the detail of step S1. The flowchart which illustrates the detail of step S2. The figure which illustrates the fluctuation | variation of the evaluation value in case an upper limit is provided, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of performance mistake, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of performance mistake, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value in case the upper limit is not provided, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value in case an upper limit is provided, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of setting an upper limit to 3, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of making an upper limit into 4, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of making an upper limit into 5, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of setting an upper limit to 6, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of setting an upper limit to 7, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value at the time of making an upper limit into 8, and the position of the estimated score. The figure which illustrates the fluctuation | variation of the evaluation value by background art, and the position of the estimated score.

  Hereinafter, embodiments of the present invention will be described in detail.

  The score tracking device 1 includes, for example, an evaluation value calculation unit 11 and an evaluation value maximum value search unit 12 as shown in FIG. FIG. 4 shows a flowchart of an example of a score tracking method.

  The evaluation value calculation unit 11 calculates the evaluation value at each position of the score every time the played pitch data is input (step S1). The calculated evaluation value is sent to the evaluation value maximum value search unit 12.

  The score evaluation value represents the degree of coincidence between the pitch data of the score up to that position and the played pitch data, and the pitch data of the score up to that position and the played pitch The higher the value is, the higher the value is, and the predetermined upper limit value is set as an index.

  The pitch data played and the pitch data of the score are expressed by, for example, MIDI (registered trademark) Note Number, and the pitch data of the score is input to the evaluation value calculation unit 11 in the format of SMF (Standard MIDI File). . In this case, the evaluation value calculation unit 11 calculates an evaluation value at each position of the score for each MIDI Note On input. In this example, only the pitch data played and the pitch data of the score are used, and other information such as the length of the sound is ignored.

  Specifically, the evaluation value calculation unit 11 calculates the evaluation value of the position of the score, each pitch data of the score up to that position, and played pitch data corresponding to each pitch data of the score. Are determined, and if they match, 1 is added, and if they do not match, 1 is subtracted.

  Each time the played pitch data p is input, the evaluation value calculation unit 11 calculates an evaluation value at each position of the score by performing, for example, the processing of FIG.

  Y is substituted for i (step S11). Y is the total number of musical pitch data.

  It is determined whether the played pitch data p matches the pitch data s (i) at the i-th position of the score (step S12).

  If they match in step S12, a value obtained by adding 1 to the (i-1) -th evaluation value W (i-1) of the score calculated when the pitch data played immediately before is input. The temporary evaluation value W ′ (i) at the i-th position is set (step S13). The smaller value of the temporary evaluation value W ′ (i) and the upper limit value (5 in this example) of the evaluation value is set as the evaluation value W (i) at the i-th position of the score (step S14).

  If they do not match in step S12, the value obtained by subtracting 1 from the (i-1) th evaluation value W (i-1) of the score calculated when the pitch data played immediately before is input. The temporary evaluation value W ′ (i) at the i-th position is set (step S15). The larger value of the temporary evaluation value W ′ (i) and the lower limit value (0 in this example) of the evaluation value is set as the evaluation value W (i) at the i-th position of the score (step S16).

  i-1 is substituted for i (step S17).

  It is determined whether i is greater than 1 (step S18). If i> 1, the process returns to step S12.

  If i> 1 is not satisfied in step S18, it is determined whether the played pitch data p matches the first pitch data s (1) of the score (step S19).

  If p = s (1) in step S19, the evaluation value W (1) at the first position of the score is set to 1 (step S110).

  If p = s (1) is not satisfied in step S19, the evaluation value W (1) at the first position of the score is set to 1 (step S111).

  The maximum evaluation value search unit 12 searches for the position of the score where the evaluation value is highest (step S2, FIG. 4). The position of the score where the searched evaluation value is the highest is estimated as the current score position.

  The evaluation value maximum value search unit 12 estimates the position of the score by performing, for example, the process of FIG.

  First, each variable is initialized (step S21). That is, 1 is substituted for i, 0 is substituted for max, and 1 is substituted for Rh. max is a variable representing the maximum value of the evaluation value, and Rh is a variable representing the position closest to the position (Rh0 + 1) next to the position Rh0 of the previously estimated score. In order to estimate the position of the score for the first time this time, if there is no previously estimated score position Rh0, for example, Rh0 = 1.

  It is determined whether max and the i-th evaluation value W (i) of the score match (step S22).

  If max = W (i), it is determined whether or not | (Rh0 + 1) −Rh |> | (Rh0 + 1) −i | (step S23). | (Rh0 + 1) −Rh | is the distance between the position (Rh0 + 1) next to the previously estimated score Rh0 and the position Rh tentatively determined to be close to the position (Rh0 + 1), and | (Rh0 + 1) −i | is the distance between the position (Rh0 + 1) next to the previously estimated score position Rh0 and the current position i.

  If | (Rh0 + 1) −Rh |> | (Rh0 + 1) −i |, i is substituted for Rh (step S24), and i is incremented by 1 (step S25).

  If | (Rh0 + 1) −Rh |> | (Rh0 + 1) −i | is not satisfied, i is incremented by 1 (step S25).

  If there are a plurality of score positions at which the evaluation value is maximized in steps S22 to S24, the position closest to the position (Rh0 + 1) next to the previously estimated score position Rh0 among the plurality of positions. Is selected.

  When it is determined in step S22 that max = W (i) is not satisfied, it is determined whether max <W (i) is satisfied (step S26).

  If max <W (i), W (i) is substituted for max (step S27), and i is substituted for Rh (step S28). Thereafter, i is incremented by 1 (step S25).

  If max <W (i) is not satisfied, i is incremented by 1 (step S25).

  After step S25, it is determined whether i <Y + 1 (step S29). If i <Y + 1, the process returns to step S22.

  If i <Y + 1, the current position Rh is the position closest to the position (Rh0 + 1) next to the previously estimated score position Rh0 with the maximum evaluation value. In this case, Rh is substituted for Rh0 (step S210), and the process of step S2 is completed. Next, when the played pitch data comes in, the processing of step S1 and step S2 is performed in the same manner as described above. At this time, Rh0 is used.

[Specific Example 1]
As in the example given in the background art section, it is assumed that the pitch data of the musical score is “Domicoside Red ...” and the pitch data that is played is also “Domisoside Red ...” as shown in FIG. At this time, the position of the number surrounded by the line in FIG. 7 is estimated as the current score position. The position of the number surrounded by the broken line in FIG. 7 gives the maximum evaluation value, but is not estimated as the current score position because it is far from the position next to the previously estimated score position. Indicates. Further, 5 after the pitch indicates that the pitch is 1 octave higher, and 3 after the pitch indicates that the pitch is 1 octave lower. For example, “do 5” is 1 octave higher than “do”. “F” means “Fa” of the pitch name. The same applies to the other drawings.

  Hereinafter, the time when the third played pitch data “SO” is input will be described as an example. When the third played pitch data “SO” is input, the evaluation value at each position of the score is calculated. If the yth position of the score in the xth played pitch data is expressed as (x, y), the evaluation value of each position (3, y) (y = 1,..., Y) of the score is calculated. The Y is the total number of musical pitch data.

  The evaluation value at the position (3, 1) is 0 because the pitch data “So” played at the position (3, 1) does not match the pitch data “Do” of the score. The evaluation value at the position (3, 2) does not match the pitch data “mi” played at the position (2, 1) and the pitch data “do” of the score, and the performance at the position (3, 2). The pitch data “So” thus set and the pitch data “Mi” of the score do not coincide with each other, and thus become zero. The evaluation value at the position (3, 3) indicates that the pitch data “do” played at the position (1, 1) matches the pitch data “do” of the musical score, and the evaluation is performed at the position (2, 2). The pitch data “Mi” matches the pitch data “Mi” of the score, and the played pitch data “So” at the position (3, 3) matches the pitch data “So” of the score. Therefore, 3 = 1 + 1 + 1. In this way, the evaluation value of each position (3, y) (y = 1,..., Y) of the score is calculated. The numbers in FIG. 7 represent the evaluation values at each position of the score.

  In this example, the evaluation value 3 at the position (3, 3) is the maximum among the evaluation values at each position (3, y) (y = 1,..., Y) of the score. Therefore, the third position of the score is estimated as the position of the score at the time when the pitch data “So” is input.

  One of the differences between the present invention and the background art is that an upper limit value is provided for the evaluation value. In FIG. 7, the upper limit value of the evaluation value is 5. Since the pitch data “Re” played at the position (6, 6) matches the pitch data “Le” of the score, the evaluation value of the position (6, 6) is 6 = in the background art of FIG. In the present invention, the evaluation value at the position (6, 6) is 5 in order to prevent the evaluation value from exceeding the upper limit value.

[Specific Example 2]
Next, description will be made by taking as an example the case of FIG. 8 in which the pitch data of the score is “Domisido Red ...” as in FIG. 7, but there is an error in the performance and the played pitch data is “Domira Sid Red ...”. do.

  The evaluation value at the position (3, 3) indicates that the pitch data “do” played at the position (1, 1) matches the pitch data “do” of the musical score, and the evaluation is performed at the position (2, 2). The pitch data “Mi” matches the pitch data “Mi” of the score, but the played pitch data “La” at the position (3, 3) matches the pitch data “So” of the score. Therefore, 1 = 1 + 1-1.

  The evaluation value at the position (3, 8) indicates that the pitch data “do” played at the position (1, 6) does not match the pitch data “le” of the score, and the performance at the position (2, 7). The pitch data “Mi” and the pitch data “Do” of the score do not match, but the played pitch data “La” and the pitch data “La” of the score at the position (3, 8) are different. Since they match, 1 = 0 + 0 + 1.

  In the third played pitch data, there are two positions where the evaluation value is maximized: position (3, 3) and position (3, 8). In this case, the evaluation value maximum value search unit 12 estimates a position closest to the position next to the previously estimated score position as a new score position. In this example, at position (3, 3) and position (3, 8), position (3, 3) is the next position (3, 3) after the previously estimated score position (2, 2). ), The position (3, 3) is selected as a new score position. As described above, when there are a plurality of score positions at which the evaluation value has the highest evaluation value, the evaluation value maximum value search unit 12 is a position next to the previously estimated score position among the plurality of score positions. By making the position closest to the position of the new score, the estimation accuracy of the position of the score is increased.

[Specific Example 3]
Next, description will be made by taking as an example the case of FIG. 9 in which the pitch data of the score is “Domiso Sidred ...” as in FIG. 7, but the performance data is “Domifu Sidred ...” due to an error in performance. do.

  The evaluation value at the position (3, 3) indicates that the pitch data “do” played at the position (1, 1) matches the pitch data “do” of the musical score, and the evaluation is performed at the position (2, 2). The pitch data “Mi” matches the pitch data “Mi” of the score, but the played pitch data “F” at the position (3, 3) matches the pitch data “So” of the score. Therefore, 1 = 1 + 1-1.

  The evaluation value of the position (3, A) indicates that the pitch data “do” played at the position (1, A-2) matches the pitch data “do” of the score, and the position (2, A-1 The pitch data “mi” played in “” matches the pitch data “mi” in score, and the pitch data “fu” played in position (3, A) and the pitch data “fu” in score Therefore, 3 = 1 + 1 + 1.

  In this case, since the evaluation value 3 of the position (3, A) is the maximum value, the estimated score position is the position (3, A).

  The evaluation value at the position (4, 4) is that the pitch data “do” played at the position (1, 1) matches the pitch data “do” of the score, and the evaluation is performed at the position (2, 2). The pitch data “Mi” matches the pitch data “Mi” of the score, and the played pitch data “F” at the position (3, 3) does not match the pitch data “So” of the score. However, since the pitch data “Shi” played at the position (4, 4) matches the pitch data “Shi” of the score, 2 = 1 + 1−1 + 1.

  The evaluation value at the position (4, A + 1) indicates that the pitch data “do” played at the position (1, A-2) matches the pitch data “do” of the score, and the position (2, A + 1) The played pitch data “mi” in -1) matches the pitch data “mi” of the score, and the played pitch data “fu” in the position (3, A) and the pitch data of the score “ 2 ”, but the pitch data“ shi ”played at the position (4, A + 1) and the pitch data“ re ”of the score do not match, so 2 = 1 + 1 + 1−1.

  In this case, there are two positions of the score (4, 4) and (4, A + 1) that maximize the evaluation value, but the position (4) next to the previously estimated position (3, A) of the score. , A + 1) is selected as the position of the new score (4, A + 1).

  In the fifth to seventh played pitch data, position (5, 5), position (6, 6), and position (7, 7) give the maximum evaluation value, respectively. Each is the estimated score position.

  In this way, if the performance is wrong, the score estimated at the position (position (3, A), position (4, A + 1)) where the pitch data of the same score is the same as the pitch data of the wrong performance. However, if the correct performance is continued, the evaluation amount at the correct position increases and the evaluation amount at the wrong position decreases, so that the estimated score position returns to the correct position.

[Specific Examples 4 and 5]
Concrete example 4 in FIG. 10 shows the fluctuation of the evaluation value when the upper limit value is not defined, and corresponds to the background art. Specific example 5 in FIG. 11 shows the fluctuation of the evaluation value when 5 is set as the upper limit value, and corresponds to an example of the present invention.

  The pitch data of the scores in FIGS. 10 and 11 is Mozart's “Kirakira Star”, and the performer replays from the middle (15th played pitch data).

  When the upper limit value is not set for the evaluation value as shown in FIG. 10, it follows the correct position in the 22nd played pitch data, and the time from replaying to following the correct performance position. Is long. This is because the evaluation value goes up indefinitely, so that it takes a longer time until the evaluation value of the performance position before replaying becomes smaller than the evaluation value of the correct performance position after replaying.

  On the other hand, when the upper limit value is defined in the evaluation value as shown in FIG. 11, the correct position is followed in the 17th played pitch data, compared with the case where the upper limit value is not defined. It takes a short time to follow the correct performance position after redrawing. By providing an upper limit value for the evaluation value, the time until the evaluation value of the performance position before replaying becomes smaller than the evaluation value of the correct performance position after replaying is shortened.

  In this way, by setting the upper limit value for the evaluation value, when the performer changes the performance position in the middle, the evaluation value for the performance position before the change becomes smaller than the evaluation value for the correct performance position after the change. The time required to follow the correct performance position after the change is shortened.

[Optimum upper limit range]
If the upper limit value is set small, there is a high possibility of leaving the correct performance position if a performance is missed, but the followability when the performance position is changed is improved. On the other hand, if the upper limit value is set high, the possibility of leaving the correct performance position in the event of a performance mistake is reduced, but the followability when the performance position is changed is reduced.

  FIG. 12 to FIG. 17 show the evaluation value and the estimated variation in the position of the score when the upper limit values are 3, 4, 5, 6, 7, and 8, respectively. In this example, the performance was missed at the position of the sixth played pitch data “Mi” and the position of the tenth played pitch data “Le 5”, and the 18th played pitch data. Replaying from the position of the pitch data “do”.

  If the upper limit value is 3 or less as shown in FIG. 12, an erroneous performance position is likely to be estimated when the performance is missed or the performance position is changed.

  As shown in FIGS. 13 to 17, as the upper limit value is increased, the followability when the performance position is changed decreases.

  The upper limit is preferably 4 or more and 7 or less for practical use. Even within that range, it is most preferable from experience to set the upper limit to 5. However, the upper limit value is appropriately set according to the required performance and specifications.

[Modifications, etc.]
As illustrated in FIG. 2, the score tracking device 1 can be used as part of an accompaniment playback system. The performance position information generated by the score tracking device 1, the tempo information estimated by the tempo estimation device 2, and the accompaniment score information are input to the accompaniment playback device 3. The accompaniment reproducing device 3 generates an accompaniment sound using these pieces of information.

  With this accompaniment playback system, the left hand part can be played automatically when the player plays the right hand part in a piano-only song, the orchestra part can be played automatically when the two-hand piano part is played, When a player plays a part, the second part can be automatically played.

  Further, as illustrated in FIG. 3, the score tracking device 1 can be used as a part of an automatic turning device. The performance position information generated by the score tracking device 1 and the pitch data of the score are input to the score display device 4. For example, the score display device 4 constituted by an LCD displays the position of the current score using these pieces of information.

  The present invention is not limited to the above-described embodiment, and can be modified as appropriate without departing from the spirit of the present invention.

1 Music Score Tracking Device 11 Evaluation Value Calculation Unit 12 Evaluation Value Maximum Value Search Unit

Claims (5)

The score evaluation value represents the degree of coincidence between the pitch data of the score up to that position and the played pitch data, and the pitch data of the score up to that position and the played pitch An evaluation value calculation that takes a higher value as the data matches and calculates an evaluation value at each position of the score from the pitch data of the score and the played pitch data as an index with a predetermined upper limit. And
An evaluation value maximum value search unit for searching for the position of the score with the highest evaluation value;
A music score tracking device. The music score tracking device according to claim 1,
The evaluation value calculation unit determines whether the evaluation value of the position of the score matches the pitch data of the score up to that position and the played pitch data corresponding to the pitch data of the score. Determine if it matches, add 1 if it matches, and subtract 1 if it doesn't match,
A music score tracking device characterized by that. The music score tracking device according to claim 2,
The upper limit is 3 to 7.
A music score tracking device characterized by that. The music score tracking device according to claim 2,
The upper limit is 5.
A music score tracking device characterized by that. The musical score tracking device according to any one of claims 1 to 4,
If there are a plurality of score positions at which the evaluation value is highest, the evaluation value maximum value search unit is a position closest to the position next to the previously estimated score position among the plurality of score positions. Is estimated as the position of the new score,
A music score tracking device characterized by that.

Images