JP2010134207A - Musical performance evaluation system and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To appropriately estimate the level of difficulty of a musical performance piece and an ability of a player. <P>SOLUTION: An evaluation system 10 is provided with: a performance time storage part 22 storing performance times of musical performance for a plurality of musical performance pieces played by each of a plurality of players; a reference player specifying processing part 23 for reading out the performance times from the performance time storage part 22, calculating a total performance time for all musical performance pieces of each player, and specifying the player representing the shortest total performance time as a reference player; a correct/error decision processing part 24 comparing the performance time for the musical performance piece with the performance time for the same musical performance piece played by the reference player, and imparting either answer value representing correct/error to each musical performance piece played by each player; and a variable estimation processing part 25 calculating a value to represent an ability of each player and a value to represent the level of difficulty of each musical performance piece according to an item response theory based on the answer value. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a performance tune which is a task for performance, and a performance evaluation system and performance evaluation program for evaluating performance by a performer.

  The item response theory is a theory for estimating each ability of the subject and the difficulty (difficulty) of the evaluation item based on the response to the evaluation item. As described in Non-Patent Document 1, for example, in order to obtain the correct answer probability using the one-parameter logistic model (The Rasch model), it is necessary to estimate the item difficulty parameter b and the ability parameter θ. ing. In this model, the probability of correctly answering item i is given by the following equation.

p _i (θ) = 1 / (1 + exp ^{− (θ−bi)} )
θ corresponds to the ability of the subject, and bi represents the difficulty level (difficulty level) of item i.

  In addition to the one-parameter logistic model, models with increased parameters (two-parameter logistic model (2PL model) and three-parameter logistic model (3PL model)) are known.

In Patent Document 1, in the two-parameter logistic model,
p (u _j = 1 | θ _i ) = 1 / (1 + ^{exp− (ajθi−bj)} )
A method for estimating the parameter a _j in FIG.

  Further, in Patent Document 2, a test taker having a specific ability is virtually created, and the probability of correct answer to the virtual test taker's problem is calculated from the correct answer rate curve in the item response theory, and the calculated correct answer rate A method has been proposed in which answer determination is performed by comparing a curve and a random number, and an ability value is output based on the answer determination result.

Patent Document 3 proposes a method for estimating a learner's ability using item response theory, and selecting and presenting a teaching material according to the ability and the learner's request.
"Introduction to Item Response Theory, New Analysis Method for Language Test Data", Chapter 4 "Rasch model: PROX Method", pp. 93-117, Kenji Otomo, July 10, 1996, published by Daishukan Shoten JP 2007-279306 A JP 2004-177510 A Japanese Patent Laid-Open No. 2003-271752

  In the item response theory, the answer of the item (question) is given a value of correct answer (value = 1) or incorrect answer (value = 0). Therefore, although it can be applied to the case where the correctness / incorrectness of the problem is clear, there is a problem that it cannot be applied when it is not easy to divide the answer into two, that is, the correct answer and the incorrect answer. For example, in the performance of musical instruments, it has not been easy to estimate the difficulty level of a musical piece and the ability of a player using item response theory.

  It is an object of the present invention to provide a performance evaluation system and a performance evaluation program that can appropriately estimate the difficulty level of a musical piece and the ability of a performer.

An object of the present invention is to provide performance index value storage means for storing an index value for a performance of a plurality of performances performed by each of a plurality of performers,
A reference player who reads out the index value stored in the performance index value storage means and identifies the player who shows the best index value as the reference player based on the index values for all the performance songs for each player Specific means,
For each of a plurality of performances performed by each performer, the index value for the performance music is compared with the index value for the same performance music by the reference player, and the Correctness determination means for assigning any answer value indicating and storing the answer value in the answer value storage means,
Based on the answer value stored in the answer value storage means, according to the item response theory, a value indicating each ability of the performer and a value indicating each difficulty level of the performance music are calculated, and variable storage is performed. This is achieved by a performance evaluation device comprising variable estimation means stored in the means.

In a preferred embodiment, the performance index value storage means stores a performance time of a performance song by each of the performers,
The reference player specifying means determines, as a reference player, a player whose performance total time, which is the sum of the performance times of the performance songs, among the performers who performed all the performance songs.

  In a more preferred embodiment, the correctness determination means determines the correctness based on a ratio of a performance time for the performance music to a performance time for the performance music by the reference player.

In another preferred embodiment, the performance index value storage means stores the number of mistouches that have a pitch different from the pitch that should be originally played in the performance of each of the performers.
The reference player specifying means determines, as a reference player, a player having the smallest total number of mistouched performances among the performers who performed all the performance songs.

  In a more preferred embodiment, the correctness determination means determines the correctness based on a ratio of the number of mistouches for the performance music to the number of mistouches for the performance music by the reference player.

In still another preferred embodiment, the performance index value storage means has a number of mistouches having a pitch different from a pitch that should be originally played in the performance of the performance by each of the performers, and the performance music. Stores all the number of notes that should be played in
The overall mistouch rate, which is the ratio of the total number of mistouches that is the sum of the number of mistouches of the performance songs and the total number of sounds of all the performance songs among the performers who performed all the performance songs, The player with the smallest is identified as the reference player.

  In a more preferred embodiment, the correctness determination means is based on a ratio of a mistouch rate, which is a ratio of the number of mistouches for the performance music to the number of sounds of the performance music, to a mistouch rate for the performance music by the reference player. Then, the correctness is determined.

  Further, in a preferred embodiment, the correctness / incorrectness determination means includes a difference between a mistouch rate, which is a ratio of the number of mistouches for the performance music to the number of sounds of the performance music, and a mistouch rate for the performance music by a reference player. Based on the above, the correctness / incorrectness is determined.

In another preferred embodiment, the performance index value storage means has the number of correct answers having a pitch that matches the pitch that should be originally played in the performance of the performance music performed by each of the performers, and Store all the number of notes that should be played in the performance song,
The reference player specifying means is the ratio of the total number of correct answers of all performance songs to the total number of sounds of all performance songs among the performers who performed all the performance songs. Are identified as reference performers.

  In a more preferred embodiment, the correct / incorrect determination means is based on a ratio of a correct answer rate, which is a ratio of the number of correct answers for the performance music to the number of sounds of the performance music, to a correct answer rate for the performance music by a reference player. Then, the correctness is determined.

  In another preferred embodiment, the correctness / incorrectness determination means includes a difference between a correct answer rate, which is a ratio of the number of correct answers for the performance music to the number of sounds of the performance music, and a correct answer rate for the performance music by the reference player. Based on the above, the correctness / incorrectness is determined.

Another object of the present invention is to provide a computer provided with performance index value storage means for storing an index value for a performance of a plurality of performances by each of a plurality of performers.
A reference player specifying step of reading out the index value stored in the performance index value storage means, and specifying the player showing the best index value as the reference player based on the entire index value for each player;
For each of a plurality of performance songs performed by each performer, the index value for the performance song is compared with the index value for the same performance song by the reference performer, and A correct / incorrect determination step of assigning any value indicating and storing in the answer value storage means;
Based on the value stored in the answer value storage means, according to item response theory, a value indicating each ability of the performer and a value indicating each difficulty level of the performance music are calculated, and variable storage means And a variable estimation step stored in the performance evaluation program.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to provide the performance evaluation system and performance evaluation program which can estimate the difficulty of a performance music and a player's capability appropriately.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a block diagram showing a configuration of an evaluation apparatus according to an embodiment of the present invention. As shown in FIG. 1, an evaluation apparatus 10 according to this embodiment includes a CPU 12, an input / output interface (I / F) 13, an input device 14 such as a keyboard and a mouse, a display device 15, a ROM 16, a RAM 17, and a hard disk device 18. Have The electronic musical instrument 11 is connected to the input / output I / F 13. As shown in FIG. 1, the evaluation apparatus 10 according to the present embodiment can be realized by a normal personal computer.

  The CPU 12 determines whether the performance is calculated based on performance data input from the electronic musical instrument 11, processing for specifying a reference performer based on the performance time data, and correct / incorrect for each performance performed by each performer. Processing, processing for estimating variables based on item response theory from the correctness / incorrectness determination result, and the like are executed.

  The ROM 16 stores a processing program for processing executed by the CPU 12. The RAM 17 temporarily stores variables, parameters, input data, output data, and the like that are necessary in the course of program execution. The hard disk device 18 has a performance time calculated by calculating the performance time, an answer value obtained in the correctness / incorrectness determination process, a variable obtained by an operation according to the item response theory (a value representing ability and a value representing difficulty), and the like. Remember.

  For example, the electronic musical instrument 11 outputs note numbers indicating note on, off, and pitch based on MIDI (Musical Instrument Digital Interface) in response to the player's key operation, and via the input / output I / F 13. Then, information (performance data) based on the MIDI is given to the CPU 12. The CPU 12 stores the performance data in the hard disk device 18.

  FIG. 2 is a block diagram showing functions of the evaluation apparatus 10 according to the present embodiment. As shown in FIG. 2, the evaluation device 10 according to the present embodiment stores a performance time measurement unit 21 and a performance that stores data (performance time data) indicating the performance time of a performance song obtained by the performance time measurement unit 21. By referring to the time data storage unit 22 and the performance time data, a reference player specifying processing unit 23 for specifying a reference player, the performance time data of the reference player, and the performance time data of each player are compared. Then, the correctness / incorrectness determination processing unit 35 for determining correctness / incorrectness for the performance of the musical piece performed by each performer, and the variable estimation processing unit for estimating a variable based on the item response theory based on the answer value obtained by the correctness / incorrectness determination processing unit 35 25, and a variable storage unit 26 for storing the variables obtained by the variable estimation processing unit 25. The functions of the performance time measuring unit 21, the reference player specifying processing unit 23, the correctness determination processing unit 24, and the variable estimation processing unit 25 are mainly realized by the CPU 12. The performance time data storage unit 22 and the result storage unit 26 are provided in the hard disk device 18, for example.

  Processing in the evaluation apparatus 10 configured as described above will be described. In the present embodiment, prior to processing by the reference player specifying processing unit 23, the correct / incorrect determination processing unit 24, and the variable processing unit 25, a plurality of performers who are test subjects each perform a plurality of performance songs that are tasks. Perform using the electronic musical instrument 11. The evaluation device 10 measures the performance time based on the performance data of each performance song. FIG. 3 is a flowchart showing an example of performance time measurement processing executed by the evaluation apparatus 10 according to the present embodiment. Hereinafter, in the present embodiment, the subject is referred to as a performer, and the item indicating the difficulty level is a performance song, and hence is referred to as a performance song without being referred to as an item.

  As shown in FIG. 3, the CPU 12 (performance time measuring unit 21) initializes a timer (step 301) and waits until performance data is input (step 302). When the performance data is input (Yes in step 302), the CPU 12 starts a timer (step 303). After that, if performance data is not input for a predetermined time, it is determined that there is no key depression (Yes in Step 304), and the CPU 12 invalidates the performance tune data to be measured as performance time data by the performer. A value is set and stored in the performance time data storage unit 22 (step 305). In this embodiment, “0” is set as an invalid value, but the present invention is not limited to this, and a value representing a so-called “NULL” value may be used.

  If it is determined that the performance has ended (Yes in step 306), the CPU 12 sets a timer value as performance time data and stores it in the performance time data storage unit 22 (step 307). For example, model data based on MIDI of performance music is stored in the hard disk device 18, and the CPU 12 sequentially compares the input performance data with the model data, and the performance data note and the last note of the model data. Can be determined to be the end of performance. Alternatively, the performance end operation may be performed by the performer, and the CPU 12 may determine that the performance has ended in response to the evaluation device 10 detecting the performance end operation. Until the acquisition of performance time data by all performers is completed (Yes in Step 308), the processing of Step 301 to Step 307 is repeated. Through the performance time measurement process, performance time data iDat [i] [j] is obtained for performance music i (i = 1 to m) and performer j (j = 1 to n).

  Next, variable calculation processing based on performance time data will be described. FIG. 4 is a flowchart showing an example of variable calculation processing according to the present embodiment. As shown in FIG. 4, the CPU 12 (reference player specifying unit 23) reads the performance time data iDat [i] [j] stored in the performance time data storage unit 22 (step 401), and specifies the reference player. Processing is executed (step 402). Next, the CPU 12 (correction determination processing unit 24) determines the correctness / incorrectness for each performance piece by each player based on the performance time data of the reference player (step 403). Thereafter, the CPU 12 (variable estimation processing unit 25) executes a variable estimation process based on the PROX method (step 404). Steps 402 to 404 will be described below.

  5 and 6 are flowcharts showing an example of the reference player specifying process according to the present embodiment. The CPU 12 (reference player specifying unit 23) initializes the minimum performance time parameter iMin (step 501). The minimum performance time parameter iMin is stored in the RAM 17 and has a value “30000”, for example. Next, the CPU 12 initializes a parameter “j” for specifying the performer to “1” (step 502).

  As long as the parameter j is equal to or less than the total number of performers n (Yes in Step 503), the CPU 12 executes the processing from Step 504 onwards. The CPU 12 initializes the performance total time iDat [giline] [j] of the performer to be processed to “0” (step 504), and also initializes the parameter “i” indicating the performance music to “1”. (Step 505).

  The CPU 12 determines whether or not the parameter i is equal to or less than the total number m of musical pieces to be played (step 506). The CPU 12 determines whether or not the performance time data iDat [i] [j] is an invalid value (“0” in the present embodiment) (step 507). If it is determined No in step 507, the CPU 12 adds the performance time iDat [i] [j] to the performance total time iDat [Gilene] [j] for the player j to obtain a new performance total. Time iDat [Giline] [j] is set (step 508). The updated total performance time is stored in the RAM 17. Thereafter, the CPU 12 increments the parameter i for specifying the performance music (step 509), and returns to step 506.

  If it is determined Yes in step 506, that is, if performance time data of all the performance songs is added for a certain player and performance total time data is obtained, the CPU 12 performs the performance for the player. It is determined whether or not the total time data iDat [Gilene] [j] is less than or equal to the minimum performance time iMin (step 601). If it is determined as Yes in step 601, the CPU 12 updates the minimum performance time so that the new minimum performance time iMin is set to iDat [Giline] [j] (step 602). The updated minimum performance time iMin is stored in the RAM 17. Further, the CPU 12 sets a player j that takes the minimum performance time as a reference player (step 603). The reference player information is also stored in the RAM 17.

  Thereafter, the CPU 12 increments the parameter j specifying the performer (step 604) and returns to step 503. Also, if it is determined Yes in step 507, that is, if the performance time data for any performance piece by the player j is an invalid value, or if it is determined No in step 601, Proceeding to step 604, the parameter j is incremented, and the process returns to step 503.

  Through the reference player specifying process shown in FIG. 5 and FIG. 6, a player who has played all performance songs to the end and whose performance total time is minimum is specified as the reference player, and the performance Total time is calculated.

  Next, the correctness determination process according to the present embodiment will be described. In the correctness / incorrectness determination process, based on performance time data iDat [i] [j] of a performance piece by a performer, it is determined whether the performance time data should be handled as a correct answer or an incorrect answer. In the present embodiment, the value given by the determination is referred to as an answer value, and the answer value is “1” for a correct answer and “0” for an incorrect answer. FIG. 7 is a flowchart illustrating an example of correctness determination processing according to the present embodiment.

  As shown in FIG. 7, the CPU 12 (correctness determination processing unit 24) initializes a parameter “j” for specifying the performer to “1” (step 701). As long as the parameter j is equal to or less than the total number of performers n (Yes in Step 702), the CPU 12 executes the processing in Step 702 and the subsequent steps. The CPU 12 initializes the parameter “i” indicating the performance music to “1” (step 703).

  The CPU 12 executes the processing from step 706 onward until the processing is completed for all the performance songs (No in step 704). If it is determined No in step 704, the parameter j indicating the performer is incremented (step 705), and the process returns to step 702.

  The CPU 12 reads out the performance time data iDat [i] [j] specified by the parameters i and j and the performance time data iDat [i] [iMinPerson] for the same performance by the reference player iMinPerson from the RAM 17, and the former. Is calculated for the latter (doRatio = iDat [i] [j] / iDat [i] [iMinPerson]) (step 704). When the performance time data iDat [i] [j] is “0” (Yes in Step 707), the CPU 12 sets “0” as the correct / incorrect answer value doProx [i] [j] (Step 708). ). The answer value doProx [i] [j] is stored in the RAM 17.

  When it is determined No in step 707, the CPU 12 determines whether or not the ratio doRatio calculated in step 706 is smaller than a predetermined threshold do (step 709). When doRatio <threshold do (Yes in Step 709), the CPU 12 sets “1” as the correct / incorrect answer value doProx [i] [j] (Step 710). In the present embodiment, for example, the threshold do is set to “3” or “4”. That is, when a performance piece is played in a time shorter than three times or four times the performance time of the reference player, the answer value is given as a value “1” indicating the correct answer.

  If it is determined No in step 709, the CPU 12 sets “0” as the correct / incorrect answer value doProx [i] [j] (step 711). Next, the CPU 12 increments the parameter i for specifying the performance tune (step 712) and returns to step 704.

  By the correctness / incorrectness determination process, an answer value of “1” or “0” is given to each performance piece performed by each performer. FIG. 8 is a diagram illustrating an example of performance time data iDat [i] [j] of a performance piece performed by each performer. In FIG. 8, the value “0” indicates an invalid value. In the example shown in FIG. 8, the player “7” is the reference player. FIG. 9 is a diagram illustrating an example of the answer value doProx [i] [j] obtained based on the performance time data shown in FIG. In the example of FIG. 9, the threshold do is set to 4.

  Next, variable estimation processing by the CPU 12 (variable estimation processing unit 25) is executed based on the obtained answer value doProx [i] [j]. The variable estimation process is executed by the CPU 12 according to the procedure described on pages 98 to 117 of the non-patent document. Hereinafter, the variable estimation process will be described. FIG. 10 is a flowchart showing the variable estimation processing executed by the variable estimation processing unit.

  As shown in FIG. 10, the CPU 12 refers to the answer value doProx [i] [j] for each performance song, and the performance song whose all answer values are “1” and all the answer values are A performance song such as “0” is excluded from the processing target. Similarly, with reference to the answer value doProx [i] [j] for each performer, the performer in which all the answer values are “1” and all the answer values are “0”. The performer is excluded from the processing target (step 1001).

Next, the CPU 12 calculates a logit collect x for each performance piece and obtains a square x ² for each. Further, the CPU 12 determines the average value x. And its square x. ² is obtained (step 1002). The CPU 12 multiplies the total sum Σx ² of the logit-incorrect squares calculated in step 1002 and the square of the average value by the total number L of performance songs excluding the number of performance songs excluded in step 1001. . ² , the unbiased variance U for the performance music is calculated (step 1003). The unbiased variance U can be obtained according to the following equation:

U = (Σx ² −Lx. ² ) / (L−1)
Similarly, the CPU 12 calculates logit collect y for each performer, and obtains each square y ² . Further, the CPU 12 calculates the average value y. And its squared y. ² is obtained (step 1004). The CPU 12 multiplies the total sum Σy ² of the logit incorrect squares calculated in step 1004 and the square of the average value by the total number N of performers excluding the number of performers excluded in step 1001. . ² , the unbiased variance V for the performance music is calculated (step 1005). The unbiased variance V can be obtained according to the following equation:

V = (Σy ² −Ny. ² ) / (N−1)
Thereafter, the CPU 12 calculates an expansion factor X relating to the difficulty level of the musical composition (step 1007). The expansion element X can be obtained according to the following equation:

X = SQR ((1 + U / 2.89) / (1-UV / 8.35))
Similarly, the CPU 12 calculates an expansion factor Y regarding the player's ability (step 1008). The expansion element Y can be obtained according to the following equation:

Y = SQR ((1 + V / 2.89) / (1-UV / 8.35))
After that, the CPU 12 calculates each difficulty level of the performance piece by each logit collect and the expansion element X of the performance piece, and also performs the player's each logit collect and expansion element Y by the player's respective logit collection and the extension element Y. Each ability is calculated (step 1008). The calculated difficulty level of each performance piece and each ability of the performer are stored in the result storage unit 26.

  The difficulty level of the musical piece and the performance of the performer can be read from the result storage unit 26 in response to the operation of the input device 14 by the operator and displayed on the screen of the display device 15 in the form of a table or the like.

  In the present embodiment, the player who shows the best index value based on the index values for all the performance songs for each performer from the index values for each of the plurality of performance songs performed by each performer. Is identified as the reference performer. Furthermore, in the present embodiment, for each of a plurality of performance songs performed by each performer, the index value for the performance song is compared with the index value for the same performance song by the reference player, A correct / incorrect value, that is, an answer value indicating “1” or “0” is assigned to the musical piece performed by the performer, and each ability of the performer is determined according to the item response theory based on the answer value. A value indicating the degree of difficulty of each performance piece is calculated.

  That is, according to the present embodiment, the player who is considered to have performed the best as a whole is determined as the reference player, and the index value for the performance piece by each player and the same performance piece by the reference player. Based on the comparison of the index values, an answer value indicating correctness is given. Therefore, it is possible to give an appropriate answer value for the performance of a performance song that cannot be determined simply, correct value for each player (subject), and difficulty for each performance song (item). It is possible to appropriately calculate a value indicating the degree.

  In particular, in the present embodiment, the reference player specifying processing unit 23 selects the player having the minimum performance total time, which is the sum of the performance times of the performance songs, among the performers who performed all the performance songs. Decided to be a performer. That is, the performance time is used as an index value, and the total performance time is used as an index value for all performance songs. Thereby, the person who was able to perform the fastest, that is, the total performance time, can be specified as the reference player.

  Further, in the present embodiment, the correctness / incorrectness is determined based on the ratio of the performance time of each player's performance to the performance time of the performance by the reference player. For example, when the performance time of each performer's performance is not an invalid value and the ratio is smaller than a certain threshold value, an answer value indicating a correct answer is given. By using the ratio, it is possible to prevent variation in results for each performance song.

  The present invention is not limited to the above embodiments, and various modifications can be made within the scope of the invention described in the claims, and these are also included in the scope of the present invention. Needless to say.

  For example, in the above-described embodiment, the player who has the minimum performance total time, which is the total performance time of the musical composition, is set as the reference player, and the performance time of the performance by the reference player (reference When the ratio (performance time / reference performance time) to (performance time) is not 0 and smaller than a predetermined threshold value, the correct / incorrect answer value is “1”. However, an index other than the performance total time can also be used.

  For example, it is possible to use the number of miss-touches, that is, the number of keys that are different from the key of the pitch to be originally played, which is shown in the score, based on the performance of each musical piece performed by each performer. In this example, for each performer, the number of missed touches of each performance song is stored in the hard disk device 18. In this example, since the number of miss-touches is considered to be “0”, a value other than “0” is stored as an invalid value when the performance cannot be performed to the end.

  In the reference player specifying process, the CPU 12 calculates, for each player, the total number of mistouches for all the performance songs, determines the player having the smallest total number of mistouches as the reference performer, and determines the reference performer. And the number of miss-touches for each piece of music performed by the reference player are stored in the RAM 17.

  In the correctness / incorrectness determination process, the ratio of the number of mistouches for the performance song to the number of mistouches for the same performance song by the reference player (the number of player mistouches / the number of reference player mistouches) When the ratio is smaller than a predetermined threshold, the correct / incorrect answer value is determined as “1”, and in other cases, the answer value is determined as “0”. The determined answer value doPro [i] [j] is stored in the RAM 17 as in the above embodiment. The variable estimation process using the answer value doPro [i] [j] is the same as in the above embodiment.

  In this example, the person who performed the most accurate performance as a whole can be specified as the reference player by using the viewpoint of accuracy of performance, not the speed of performance, but the number of miss-touches.

  Alternatively, the number of mistouches of each performance song by each performer and the number of keys to be pressed for each performance song are stored in the hard disk device 18, and for each player, the overall mistouch rate for the entire performance song, that is, , (The total number of mistouches for all performance songs / the total number of keys to be pressed for all performance songs) is calculated, and the player with the smallest mistouch rate for the entire performance song is determined as the reference player. You may do it. In this example as well, since it is conceivable that the mistouch rate is “0”, a value other than “0” is stored as an invalid value when the performance cannot be performed to the end.

  In the above example, the reference player information and the mistouch rate (number of mistouches in a certain performance song / number of keys to be played in the performance song) for each of the performance songs of the reference player are stored in the RAM 17. The

  In the above example, in the correctness determination process, the ratio of the player's mistouch rate to the reference player's mistouch rate (performer's mistouch rate / reference player's mistouch rate) is calculated for each piece of music performed by each player. When the ratio is calculated and the ratio is smaller than a predetermined threshold, the correct / incorrect answer value is determined as “1”, and in other cases, the answer value is determined as “0”.

  Also in this example, the person who performed the most accurate performance as a whole can be identified as the reference player by using the viewpoint of accuracy of performance, not the viewpoint of speed, but the mistouch rate.

  In the correctness / incorrectness determination process, correctness / incorrectness may be determined based on a difference between an index value for a performance piece by each player and an index value for the same performance piece by a reference player, instead of a ratio. For example, in the first embodiment, the performance time / reference performance time is calculated, but the correctness / incorrectness may be determined based on the performance time−reference performance time. FIG. 11 is a flowchart illustrating an example of correctness determination processing according to the second embodiment. 11, steps 1101 to 1105 are the same as steps 701 and 705 in FIG. When it is determined Yes in step 1104, the CPU 12 calculates a difference doTime between the performance time iDat [i] [j] and the reference performance time iDat [i] [iMinPerson] (step 1106). Note that doTime does not fall below -iDat [i] [iMinsonson], but can take a negative value.

  When the performance time data iDat [i] [j] is “0” (Yes in Step 1107), the CPU 12 sets “0” as the correct / incorrect answer value doProx [i] [j] (Step 1108). ). The answer value doProx [i] [j] is stored in the RAM 17.

  When it is determined No in step 1107, the CPU 12 determines whether or not the difference value doTime calculated in step 1106 is smaller than a predetermined threshold value b (step 1109). If doTime <b (Yes in Step 1109), the CPU 12 sets “1” as the correct / incorrect answer value doProx [i] [j] (Step 1110). On the other hand, if it is determined No in step 1109, the CPU 12 sets “0” as the correct / incorrect answer value doProx [i] [j] (step 1111). Next, the CPU 12 increments the parameter i for specifying the performance music (step 1112), and returns to step 1104.

  Also, the difference between the number of missed touches of the music piece performed by each player and the number of missed touches (reference number of mistouches) of the same music piece performed by the reference player is calculated, and the difference value is within a predetermined range (d> difference value> f: where d is a predetermined positive number, and f is a negative number of reference mistouches), the CPU 12 sets “1” as the correct / incorrect answer value, otherwise it is the correct / incorrect answer value. “0” may be set, and the set answer value may be stored in the RAM 17.

  Further, the difference between the mistouch rate of the performance music by each player and the mistouch rate (reference mistouch rate) of the same performance music by the reference player is calculated, and the difference value is within a predetermined range (g> difference value> If h: g is a predetermined positive number and h is 0 or a predetermined negative number), the CPU 12 sets “1” as the correct / incorrect answer value, and “0” as the correct / incorrect answer value otherwise. And the set answer value may be stored in the RAM 17. In particular, in this example, since the ratio (mistouch rate) is already used as the index value, it is possible to obtain a good answer value simply by taking the difference value.

  In the above embodiment, the player who has the minimum mistouch rate for the total performance time, the total mistouch, or the entire performance piece is set as the reference player. May be selected. For example, the correct answer rate, that is, the rate at which the key of the pitch indicated in the score is correctly pressed may be used based on the performance of each performance piece of the performer. In this example, the number of correct answers and the number of keys to be pressed are stored in the hard disk device 18 for each performer. The number of correct answers is the number of pitches that correspond to the pitch that should be originally played in the performance of the musical piece by the performer, more specifically, the number of keys that are correctly pressed for the musical piece by the performer. .

  In the reference player specifying process, the CPU 12 calculates, for each performer, the total number of correct answers for all performance songs and the total number of keys to be pressed for all performance songs, and for the entire performance song. The total correct answer rate (total number of correct answers / total number of keys to be pressed) is calculated, the player with the highest correct answer rate for the entire performance is determined as the reference player, and the information about the reference player is obtained. And stored in the RAM 17.

  FIG. 12 is a flowchart showing an example of correctness determination processing according to the third embodiment using the correct answer rate. In FIG. 12, steps 1201 to 1205 are the same as steps 701 and 705 in FIG. When it is determined Yes in step 1104, the CPU 12 calculates the correct answer rate iRatio [i] [j] for the performance music i of the player j to be processed, and the performance music of the reference player iMinPerson. The correct answer rate (reference correct answer rate) iRatio [i] [iMinPerson] for i is calculated (step 1206).

  Next, the CPU 12 calculates a ratio of the correct answer rate to the reference correct answer rate (doRatio = iRatio [i] [j] / iRatio [i] [iMinPerson]) (step 1207). When the correct answer rate iRatio [i] [j] is “0” (Yes in Step 1208), the CPU 12 sets “0” as the correct / incorrect answer value doProx [i] [j] (Step 1209). . The answer value doProx [i] [j] is stored in the RAM 17.

  If it is determined No in step 1208, the CPU 12 determines whether or not the ratio of correct answers calculated in step 1107, doRatio, is greater than a predetermined threshold c (step 1210). When doRatio> c (Yes in Step 1210), the CPU 12 sets “1” as the correct / incorrect answer value doProx [i] [j] (Step 1110). On the other hand, if it is determined No in step 1210, the CPU 12 sets “0” as the correct / incorrect answer value doProx [i] [j] (step 1111). Next, the CPU 12 increments the parameter i for specifying the performance tune (step 1213) and returns to step 1204.

  As described in the third embodiment, as the index value, the correct answer rate representing the better performance can be used as the index value is larger, and the performance with the largest index value (in this example, the overall correct answer rate). Can be the reference performer.

  In addition, the difference between the correct answer rate of the performance song by each performer and the correct answer rate (reference correct answer rate) of the same performance song by the reference performer is calculated, and the difference value is within a predetermined range (i> difference value> j: i is a predetermined positive number and j is 0 or a predetermined negative number), the CPU 12 sets “1” as the correct / incorrect answer value, and “0” as the correct / incorrect answer value otherwise. And the set answer value may be stored in the RAM 17. In particular, in this example, since the ratio (correct answer rate) is already used as the index value, it is possible to obtain a good answer value simply by taking the difference value.

  Further, in the embodiment, in the variable estimation process, the CPU 12 refers to the answer value doProx [i] [j] for each performance song, and the performance song whose answer values are all “1”, and A musical piece whose answer value is “0” is excluded from the processing target. Similarly, with reference to the answer value doProx [i] [j] for each performer, the performer in which all the answer values are “1” and all the answer values are “0”. The performer is excluded from the processing target (see step 1001). However, in this case, it is not possible to estimate the difficulty level and ability of a musical piece and a player who are excluded from the processing target. Therefore, instead of step 1001, the CPU 12 sets only one answer value to “0” by random numbers or “0” for all performance songs having all answer values “1”. For a performance piece such as “”, only one predetermined answer value may be set to “1” by a random number. Similarly, for a player whose answer values are all “1”, the CPU 12 sets only one answer value to “0” based on a random number, or all answer values are “0”. For such a player, variable estimation processing may be performed by setting only one answer value to “1” based on random numbers.

FIG. 1 is a block diagram showing a configuration of an evaluation apparatus according to an embodiment of the present invention. FIG. 2 is a block diagram showing functions of the evaluation apparatus 12 according to the present embodiment. FIG. 3 is a flowchart showing an example of performance time measurement processing executed by the evaluation apparatus 10 according to the present embodiment. FIG. 4 is a flowchart showing an example of variable calculation processing according to the present embodiment. FIG. 5 is a flowchart showing an example of the reference player specifying process according to the present embodiment. FIG. 6 is a flowchart showing an example of the reference player specifying process according to the present embodiment. FIG. 7 is a flowchart illustrating an example of correctness determination processing according to the present embodiment. FIG. 8 is a diagram illustrating an example of performance time data iDat [i] [j] of a performance piece performed by each performer. FIG. 9 is a diagram illustrating an example of the answer value doProx [i] [j] obtained based on the performance time data shown in FIG. FIG. 10 is a flowchart showing the variable estimation processing executed by the variable estimation processing unit. FIG. 11 is a flowchart illustrating an example of correctness determination processing according to the second embodiment of the present invention. FIG. 12 is a flowchart illustrating an example of correctness determination processing according to the third embodiment of the present invention.

Explanation of symbols

10 Evaluation Device 11 Electronic Musical Instrument 12 CPU
13 Input / output I / F
14 Input device 15 Display device 16 ROM
17 RAM
18 Hard Disk Device 21 Performance Time Measurement Unit 22 Performance Time Storage Unit 23 Reference Performer Specific Processing Unit 24 Correctness Determination Processing Unit 25 Variable Estimation Processing Unit 26 Result Storage Unit

Claims (12)

Performance index value storage means for storing an index value for the performance of a plurality of performances performed by each of a plurality of performers,
A reference player who reads out the index value stored in the performance index value storage means and identifies the player who shows the best index value as the reference player based on the index values for all the performance songs for each player Specific means,
For each of a plurality of performance songs performed by each performer, the index value for the performance song is compared with the index value for the same performance song by the reference player, and the Correctness determination means for assigning any answer value indicating and storing the answer value in the answer value storage means,
Based on the answer value stored in the answer value storage means, according to the item response theory, a value indicating each ability of the performer and a value indicating each difficulty level of the performance music are calculated, and variable storage is performed. A performance evaluation apparatus comprising: variable estimation means stored in the means. The performance index value storage means stores a performance time of a performance song by each of the performers,
The reference player specifying means determines, as a reference player, a player having a minimum performance total time, which is a sum of performance times of performance songs, among performers who performed all the performance songs. The performance evaluation apparatus according to claim 1.   3. The performance according to claim 2, wherein the correctness determination means determines the correctness based on a ratio of a performance time for the performance music to a performance time for the performance music by the reference player. Evaluation device. The performance index value storage means stores the number of mistouches having a pitch different from the pitch to be originally played in the performance of the performance music by each of the performers,
2. The player according to claim 1, wherein the reference player specifying unit determines, as a reference player, a player who has the smallest total number of mistouched performances among the performers who performed all the performance songs. Performance evaluation device.   5. The performance according to claim 4, wherein the correctness / incorrectness determination means determines the correctness based on a ratio of the number of mistouches for the performance music to the number of mistouches for the performance music by the reference player. Evaluation device. The performance index value storage means has a number of mistouches that have a pitch different from the pitch that should be originally played in the performance of each of the performers, and all the numbers that should be played in the song. And store
The overall mistouch rate, which is the ratio of the total number of mistouches that is the sum of the number of mistouches of the performance songs and the total number of sounds of all the performance songs among the performers who performed all the performance songs, The performance evaluation device according to claim 1, wherein the player with the smallest is identified as a reference player.   The correctness / incorrectness determination means determines the correctness based on a ratio of a mistouch rate, which is a ratio of the number of mistouches for the performance music to the number of sounds of the performance music, to a mistouch rate for the performance music by a reference player. The performance evaluation apparatus according to claim 6.   The correct / incorrect determination means determines the correct / incorrect based on a difference between a mistouch rate, which is a ratio of the number of mistouches for the performance music to the number of sounds of the performance music, and a mistouch rate for the performance music by a reference player. The performance evaluation apparatus according to claim 6, wherein: In the performance index value storage means, the number of correct answers having a pitch that matches the pitch that should be originally performed in the performance of the performance by each of the performers, and all the sounds that should be originally performed in the performance Store numbers and
The reference player specifying means is the ratio of the total number of correct answers of all performance songs to the total number of sounds of all performance songs among the performers who performed all the performance songs. The performance evaluation apparatus according to claim 1, wherein the performer is identified as a reference performer.   The correctness determination means determines the correctness based on a ratio of a correct answer rate, which is a ratio of the number of correct answers for the performance music to the number of sounds of the performance music, to a correct answer rate for the performance music by a reference player. The performance evaluation apparatus according to claim 9.   The correct / incorrect determination means determines the correct / incorrect based on a difference between a correct answer rate, which is a ratio of the number of correct answers for the performance music to the number of sounds of the performance music, and a correct answer rate for the performance music by a reference player. The performance evaluation apparatus according to claim 9, wherein: For a plurality of performance songs by each of a plurality of performers, a computer having performance index value storage means for storing index values for the performance,
A reference player specifying step of reading out the index value stored in the performance index value storage means, and specifying the player showing the best index value as the reference player based on the entire index value for each player;
For each of a plurality of performance songs performed by each performer, the index value for the performance song is compared with the index value for the same performance song by the reference performer, and A correct / incorrect determination step of assigning any value indicating and storing in the answer value storage means;
Based on the value stored in the answer value storage means, according to item response theory, a value indicating each ability of the performer and a value indicating each difficulty level of the performance music are calculated, and variable storage means And a variable estimation step stored in the performance evaluation program.

Images