JP6677041B2 - Performance analyzer and program - Google Patents

Description

  The present invention relates to a performance analyzer for analyzing a bowing state in the performance of a bowed instrument from a sound signal including a performance sound of a bowed instrument such as a violin, viola, cello, and contrabass, and further, a computer-executable apparatus therefor. About the program.

  In playing a bowed instrument such as a violin, it is required that the bowing order (up / down) be performed accurately. For example, in a violin playing lesson, evaluating whether or not the bowing order was performed correctly is one important factor. Such bowing evaluation can be performed, for example, by visually confirming the performance operation performed by the player, but it is preferable that the bowing evaluation can be automatically performed without direct visual recognition by a human. Therefore, the present invention intends to provide a technique capable of directly analyzing a bowing state in playing a bowed instrument from an acoustic signal of a playing sound of the bowed instrument.

  By the way, various techniques related to bowing of a bowed musical instrument such as a violin are known. For example, in Patent Literature 1 below, in accordance with a real-time performance of a violin, a stroke speed, a pressure, and a sounding point (a position on a string where the bow touches) of a bow operated by a player are detected using a sensor. It has been shown to generate an electrical signal indicative of the result. However, this cannot analyze the bowing state of the bowed instrument from the acoustic signal of the performance sound. In Patent Document 2 below, a moving speed of a bow is calculated from a volume value (velocity value) in performance data of the MIDI standard, and a return position of the bow is detected using the value (specifically, the calculated moving speed). , The trajectory of the bow is calculated, and the trajectory is compared with a parameter indicating the length of the bow to detect the turn-back position of the bow and obtain this time). This is a technique intended to add control data according to the bowing state to performance data of the MIDI standard, and also cannot analyze the bowing state of a bowed instrument from an acoustic signal of a performance sound. Absent. Further, since the position of the bow turn is detected based on the calculation of the moving speed of the bow from the volume value, it is not suitable for quickly detecting the turn of the bow.

  Further, in Patent Document 3 below, a fixed point, which is a bowing point for preventing a bowing (bowing) from changing in the bowing pattern extraction, is input to a note number of the electronic musical score data, and is fixed. It is shown that a bowing pattern is determined according to points, extracted as a bowing pattern, and a bowing score with the extracted bowing pattern is output. However, this also cannot analyze the bowing state of the bowed instrument from the acoustic signal of the performance sound. Further, in Patent Document 4 described below, a plurality of strings are produced by converting the performance sound of a plucked musical instrument such as a guitar into an electric signal using a microphone and comparing the electric signal of the performance sound with a predetermined evaluation criterion. Is detected and evaluated that a stroke performance of stroking is performed for one stroke. This also cannot analyze the bowing state of the bowed instrument from the acoustic signal of the performance sound.

  Patent Literature 5 below discloses a musical tone control device for an electronic musical instrument suitable for simulating a bowed musical instrument performing using a movable playing member such as a bow. Thus, a fricative sound generated in real time is filtered to generate a tone control signal for bow pressure and bow speed, and a simulated sound of a bowed instrument is controlled by the tone control signal. This is a technique for simulating the sound of a bowed instrument like a silent violin, and cannot analyze the bowing state of a bowed instrument from an acoustic signal of a performance sound.

JP 05-127589 A JP 2001-159892 A JP 2009-265516 A JP 2011-0699900 A JP 05-165465 A

  The present invention has been made in view of the above points, and has as its object to provide a performance analyzer and a program for analyzing a bowing state in playing of a bowed instrument from an acoustic signal of the bowed instrument. The object of the present invention is to make it possible to efficiently detect the return operation of.

  A performance analysis device according to the present invention includes: an acquisition unit that acquires an acoustic signal; an extraction unit that extracts a signal component outside the range of a target bowed instrument from the acoustic signal; and a bow operation timing based on the level of the extracted signal component. And a detection unit that detects

  In a bowed musical instrument, when an operation of returning a bow is performed, pulse noise is generated. Such noise includes a component outside the range of the bowed instrument (for example, a component in a lower range than the original range of the instrument). Therefore, a signal component outside the range of the bowed instrument is extracted from the acoustic signal of the performance sound of the bowed instrument, and the bow is returned based on the level of this signal component (for example, it is determined that the level is equal to or greater than a predetermined threshold). It is possible to detect the occurrence of pulse noise when the operation is performed. Thereby, the bow operation timing (bow return operation timing) can be detected.

  The present invention may be embodied not only as an invention of a device, but also as an invention of a program for causing a computer to execute the steps for realizing the functions of the performance analysis device.

  According to the present invention, since the bow operation timing is detected based on the specific sound range component in the acquired sound signal, complicated processing is unnecessary, and the bow return operation can be efficiently detected. It has the effect.

1 is a conceptual block diagram illustrating a configuration example of a performance analysis device according to an embodiment of the present invention. FIG. 3A is a diagram illustrating an example of an acquired acoustic signal by a volume envelope, and FIG. 3B is a diagram illustrating an example of a low-frequency noise signal outside a sound range of a target bowed instrument. FIG. 1 is a block diagram showing an example of an electrical hardware configuration when a performance analysis device is configured using a computer. 9 is a flowchart illustrating an example of bowing direction estimation processing. 9 is a flowchart illustrating an example of evaluation and display processing. FIG. 6 is a diagram showing a display example according to the evaluation and display processing of FIG. 5.

  FIG. 1 is a conceptual block diagram illustrating a configuration example of a performance analysis device according to an embodiment. The performance analysis device 100 functions to analyze the bowing state (bow operation timing, ie, bow return timing and bowing direction, etc.) from an acoustic signal generated in accordance with the performance of the bowed instrument. It can be used for any variety of purposes, such as for learning the playing method of the ensemble or for evaluating the performance of the bowed instrument. The performance analysis device 100 roughly includes an acquisition unit 10, an extraction unit 11, a detection unit 12, an estimation unit 13, an evaluation unit 14, a range setting unit 15, and a display unit 16. The performance analyzer 100 may be constituted by any computer device such as a general-purpose personal computer capable of executing the operation of each unit (various processes for analyzing an audio signal described later), or may be capable of executing the operation of each unit. And a dedicated hardware device (such as an integrated circuit) configured as described above.

  The acquisition unit 10 acquires, as an analysis target, an acoustic signal AS generated according to a performance of a bowed musical instrument. The acquisition unit 10 may be configured to acquire the acoustic signal AS by acquiring the musical performance of the user's instrument in real time via a sound collecting means such as a microphone and converting the acquired sound to digital, or The recorded data may be obtained from a recording medium on which data of the musical performance sound of the user is recorded, or the sound signal AS may be obtained from a remote place via a communication network. In this case, the encoding format of the audio signal AS acquired from the outside is not limited to the PCM format, and may be data compressed in an appropriate compression format. The acquisition unit 10 may decode the audio compression data as necessary. It may have a chemical function. The acquired acoustic signal AS is buffer-stored as appropriate in a FIFO or a RAM, and its volume envelope is detected in a time series, and is temporarily stored for use in bowing estimation processing described later. FIG. 2A illustrates an example of the acoustic signal AS acquired by the acquisition unit 10 using a volume envelope.

  The extracting unit 11 extracts, from the acquired acoustic signal, a signal component outside the range of the bowed instrument to be analyzed. The reason for extracting the signal component outside the range of the bowed instrument in this way is to exclude the signal component of the normal performance sound based on the vibration of the string caused by the bow operation (that is, the signal component within the range of the target bowed instrument), It is intended to extract friction noise during bow turning operation. For example, assuming that the bowed instrument to be analyzed is a violin, the violin has a range of pitch names G3 to D # 7, and therefore belongs to any band other than the band including the frequency component of the range including overtone components. Extract signal components. Since a stringed instrument sound generally includes harmonic components to a considerably high frequency range, it is preferable to extract a low-frequency side signal component as a signal component outside the range of the target bowed instrument. As an example, when the bowed musical instrument to be analyzed is a violin, the extraction unit 11 uses a low-pass band-pass filter having a band of about 10 Hz to 100 Hz (or a low-pass filter having a high-frequency cutoff frequency of about 100 Hz) to generate a low-frequency band. The signal component may be extracted. Of course, the low-frequency band extracted by this low-pass bandpass filter (low-pass filter) is variably set according to the range of the target bowed instrument. For example, when the bowed instrument to be analyzed is a viola, the pass band of the low-pass band-pass filter (low-pass filter) in the extraction unit 11 extracts a signal component on the lower frequency side than a high-frequency cutoff frequency lower than the violin. It is set as follows. Further, when the bowed instrument to be analyzed is a cello or a contrabass, a lower high-frequency cutoff frequency corresponding to each is set, and a signal component on a lower frequency side than each high-frequency cutoff frequency is set. You.

  FIG. 2B shows an example of a low-frequency noise signal outside the range of the target bowed string instrument extracted by the extraction unit 11 corresponding to the acoustic signal of FIG. The portion where the low-frequency noise signal shows a high level in a spike shape indicates a portion where significant friction noise has occurred during the bowing operation. Based on the level of the extracted signal component (low-frequency noise signal) outside the musical range of the instrument, the detecting unit 12 determines when the level exceeds a predetermined threshold at the timing of the bow returning operation (bow operation timing). Detected as t1, t2, t3,... The bow operation timing detection information BD indicating the bow operation timings t1, t2, t3,... Detected from the acoustic signal in this manner can be used in various applications. For example, as described below, it is used in the estimating unit 13 to estimate the bowing direction. Alternatively, it can be appropriately used in the other use device 17 as a signal indicating the rising of each performance sound (note) included in the acoustic signal AS.

  The estimating unit 13 estimates the bowing direction based on the bowing operation timing detection information BD indicating the bowing operation timing detected by the detecting unit 12. Since the standard bowing operation of a bowed instrument consists of alternately performing the raising bow operation and the lowering bow operation, if the initial bowing direction is specified for at least an appropriate partial performance section, the bowing operation is thereafter performed. By estimating that the bowing direction switches at each timing, it is possible to estimate the bowing direction based only on the bow operation timing detection information BD.

  A more accurate bowing direction estimation method is to take into account the volume of the acoustic signal AS of the performance sound. It is known that, when operated with the same level of force, a loud bow sounds more loudly than an up bow sound, so in one embodiment, the estimating unit 13 determines the adjacent bow operation timing May be configured to estimate the bowing direction at each of the detected bow operation timings based on the correlation between the sound volumes of the acoustic signals AS in the sections sandwiched by (i.e., the correlation between the volumes in the adjacent sections). . More specifically, a section starting from a certain bow operation timing (tentatively referred to as a first bow operation timing; for example, t2 in FIG. 2B) detected by the detection unit 12 (tentatively defining this as a first section) ), The volume of the acoustic signal AS in the section starting from the immediately preceding bow operation timing (this is temporarily referred to as a second bow operation timing; for example, t1 in FIG. 2B) (this is temporarily referred to as a second section). ), The sound volume of the sound signal AS at the first bow operation timing is estimated when the sound volume difference is larger than a predetermined first threshold value when the sound volume difference is larger than a predetermined first threshold value. Is smaller than the volume of the second bow operation timing sound signal AS, and if the volume difference is smaller than a predetermined second threshold value, it is estimated to be the second bowing direction (raising bow). Good.

In FIG. 2, reference numerals U and D indicate examples of bowing directions estimated by a combination of the bow operation timings t1, t2, t3,... Indicated by the bow operation timing detection information BD and the volume of the acoustic signal AS. ing. D indicates a down bow, and U indicates an up bow. For example, the volume of the sound signal AS ₂ in the section that starts a certain arch operation timing t2 that has been detected, a second predetermined than the volume of the sound signal AS ₁ in a section that starts the bow operation timing t1 immediately before Since it is smaller than the threshold value, the bowing direction is estimated to be the upward bow U. Conversely, the volume of the sound signal AS ₁ in a section that starts a certain arch operation timing t1 is than a predetermined first threshold value than the volume of the sound signal AS ₂ in the section that starts the bow operation timing t2 immediately thereafter Since it is large, the bowing direction can be estimated to be the downbow D. Hereinafter, similarly, some bow volume of the sound signal AS ₃ the operation timing t3 in the section to the start point, the first predetermined than the volume of the sound signal AS ₂ in the section that starts the bow operation timing t2 immediately preceding Since it is larger than the threshold value, the bowing direction is estimated to be the downward bow D. The magnitude of the volume can be determined by comparing the maximum value of each volume envelope, or may be determined by an average value instead of the maximum value. Note that these thresholds may be determined as appropriate.

  The estimating unit 13 is not limited to estimating the bow raising / lowering for each sound, and may be configured to estimate the alternate raising / lowering of a bow in a tremolo performance during one sound, or may be configured to estimate a bow in a slur performance. It may be configured to estimate the repetition of the bow returning operation in the same direction. In this case, for example, the values of the first threshold value and / or the second threshold value are appropriately set according to the estimation target performance method such as tremolo or slur.

  The display unit 14 is a device that visually presents information indicating the bowing direction estimated by the estimation unit 13. When visually indicating the estimated bowing direction, any display form may be adopted. For example, appropriate icons (for example, U and D) indicating the bowing direction may be simply arranged and displayed in chronological order, or each note on the musical score display of the music played by the acquired acoustic signal AS may be displayed. An icon indicating the estimated bowing direction may be displayed in association with. This display is not limited to a display, and may be a printout by a printer. The presentation of the information indicating the estimated bowing direction is not limited to the visual method, and may be presented audibly by a notification sound or the like. Alternatively, information indicating the estimated bowing direction may be recorded on a recording medium, and later read out from the recording medium and presented visually or audibly. As described above, visually or audibly presenting the information indicating the estimated bowing direction is based on, for example, whether the bowing is appropriate based on the sound signal AS played by the student at the place of training the bowed instrument. It is useful not only for teachers to evaluate, but also for students or performers themselves to confirm that their bowing was appropriate.

  The evaluation unit 15 compares the information indicating the exemplary bowing direction related to the music played by the acquired sound signal AS with the information indicating the bowing direction estimated by the estimation unit 13, thereby obtaining the acquired music. It is configured to evaluate bowing in the acoustic signal AS. This evaluation result can be presented to a user (a teacher, a student, a performer, or the like) via the display unit 14. An arbitrary display form of the evaluation result can be adopted. For example, the musical score of the musical piece played by the acquired acoustic signal AS is displayed on a display screen, and an icon indicating an evaluation result corresponding to each note on the musical score (an icon indicating that the bow is correct, An icon indicating that the bowing was wrong) may be displayed. Alternatively, a score obtained by scoring a bow based on the evaluation result may be presented. The evaluation result by the evaluation unit 15 may be accumulated and stored in a recording medium for each user (student or performer). By accumulating and storing the evaluation results for each individual user (student or player), it is possible to statistically elucidate the habit of bowing for each student or player or a place on a musical score where the bowing is likely to be mistaken. In addition, it is possible to statistically elucidate, for the entire student or performer group, a part on the score where the bowing is likely to be mistaken or a part on the score that is not likely to be mistaken.

  In addition, the evaluation part 15 can utilize the music database 20 for such evaluation of bowing. The music database 20 stores musical score data of many music. The evaluation unit 15 accesses the music database 20 to acquire music score data of a music piece played by the acquired audio signal AS, and based on the acquired music score data, information indicating an exemplary bowing direction for the music piece. Get. If the musical score data stored in the music database 20 includes bowing direction information (bowing information), it may be used as it is as information indicating the exemplary bowing direction. If the musical score data stored in the music database 20 does not include bowing information, a normal bowing rule (for example, alternately lowering bow and raising bow, return bow at sound breaks, measure lines) The following sound is basically played with a downward bow, and the bow is not returned in the same note and in the case of slurs and ties, etc.), and the default bow operation information is automatically associated with each note in the song. What is necessary is just to generate | generate and to use it as the information which shows the bowing direction of the said model. Even if the score data includes bowing information, it does not present bowing information for all notes in the score, but only for major notes (notes for which you want to specify the bowing direction in particular). Many present bowing information. In such a case, in order to specify bowing information for all notes in the score, it is preferable to automatically generate the default bowing information for notes with insufficient bowing information. . In other words, for musical notes for which specific bowing information is specified in the musical score data, the bowing information is adopted, and for other notes, default bowing information is automatically generated according to the above-mentioned normal bowing rule. The combination is used as information indicating the bowing direction of the model for the music.

  The automatically generated default bowing information (information indicating the exemplary bowing direction) may be stored in a memory dedicated to the user, or the user may rewrite (update) the music database 20. ) May be stored in the music database 20 if authorized. Also, the default bow information created by each user may be shared or rewritten between a plurality of users.For example, such shared bow information may be stored and managed on a cloud. You may.

  The range setting unit 16 identifies the type of the target bowed instrument with respect to the acoustic signal AS acquired by the acquisition unit 10 and sets the range of the signal component to be extracted by the extraction unit 11 according to the identified instrument type. It is configured as follows. The identification of the target bowed instrument type in the range setting unit 16 may be performed in response to a user's input operation (selection operation) of the target bowed instrument, or the sound signal acquired by the acquisition unit 10. The type of the target bowed instrument may be automatically identified based on the analysis of the tone color and the tone range of the AS. In the extraction unit 11, the pass band of the low-pass band-pass filter (low-pass filter) is variably set to a predetermined band lower than the lowest sound of the target bowed instrument in accordance with the setting by the sound range setting unit 16.

  FIG. 3 is a block diagram showing an example of an electrical configuration when a computer is used as hardware of the performance analysis device 100 of FIG. The performance analyzer 100 includes a microprocessor unit (CPU) 101, a memory 102, an operation device 103, a display device 104, an audio interface (“audio I / F”) 105, a communication interface (“communication I / F”) 106, and the like. including. The CPU 101 executes various programs stored in the memory 102 to control the overall operation of the performance analyzer 100. The memory 102 includes a ROM, a RAM, and an external storage device. The external storage device is composed of an appropriate recording medium such as a magnetic disk, an optical disk, and a flash memory.

  The operation device 103 includes, for example, an operation input device such as a mouse and a keyboard, and a mechanism for detecting the operation input. The display device 104 includes a display and a mechanism for controlling display on the display.

  The audio I / F 105 includes an input terminal for an audio signal (audio signal), an output terminal for an audio signal, an analog-digital converter, and a digital-analog converter. The performance analyzer 100 can acquire an acoustic signal AS generated by a user (student or performer) playing a bowed musical instrument in real time via a microphone (not shown) and an audio I / F 105. In that case, the microphone and the audio I / F 105 constitute the acquisition unit 10.

  The communication I / F 106 includes an Ethernet (registered trademark) interface, a USB interface (detachable storage medium interface), and the like. The performance analysis device 100 can connect to a communication network such as the Internet via the communication I / F 106 and communicate with a remote place. The performance analyzer 100 can acquire, via the communication I / F 106, an acoustic signal AS generated by a real-time performance of a musical instrument of a remote user (student or performer) or an already recorded acoustic signal AS. In that case, the communication I / F 106 constitutes the acquisition unit 10.

  FIG. 4 is a flowchart schematically illustrating an example of the bowing direction estimating process including a software program non-transiently stored in the memory 102 and executable by the CPU 101.

  In step S1, the CPU 101 acquires an acoustic signal AS to be analyzed (operation of the acquiring unit 10 in FIG. 1). As an example, the CPU 101 acquires the generated acoustic signal in substantially real time via the audio I / F 104, and stores it in a buffer.

  In step S2, the CPU 101 selects an audio signal of a predetermined time length from the buffered audio signal AS as a processing target in chronological order.

  In step S3, the CPU 101 extracts a signal component (low-frequency noise signal component) outside the range of the target bowed instrument from the acoustic signal of a predetermined time length selected as the processing target. This corresponds to the operation of the extraction unit 11 in FIG. 1, and consists of applying a low-pass bandpass filter (low-pass filter) to the audio signal as described above.

  In step S4, the CPU 101 detects a pulse having a level equal to or higher than a predetermined threshold from the extracted noise signal component, detects this timing as a bow operation timing, and outputs information indicating the detected bow operation timing (bow operation timing detection information). BD) is stored. This corresponds to the operation of the detection unit 12 in FIG.

  In step S5, the CPU 101 divides the audio signal of the predetermined time length selected as the processing target into a plurality of sections at the detected bow operation timing. In step S6, the CPU 101 compares the sound volume with the immediately preceding section for each section divided by the detected bow operation timing. If the sound volume is higher than the immediately preceding section (the sound volume difference is greater than the predetermined value). If the volume is lower than the first section (if the volume difference is smaller than the predetermined second threshold), it is estimated that the volume is lower than the immediately preceding section (if the volume difference is smaller than the second threshold). The estimation result is stored as information indicating the bowing direction (bowing estimation information U / D) in association with the bow operation timing detection information BD. This corresponds to the operation of the estimation unit 13 in FIG.

  In step S7, the CPU 101 determines whether the estimation in step S6 has been completed for all sections relating to the audio signal of the predetermined time length selected as a processing target, and if NO, repeats the processing in step S6. If the estimation in step S6 has been completed for all sections, the process proceeds to step S8, and it is determined whether the processing has been completed for all of the acquired audio signals AS. If NO, the process returns to step S2. In step S2, the earliest audio signal having a predetermined time length among the unprocessed audio signals AS is selected as a processing target. Hereinafter, the processes of steps S3 to S8 are repeated for the audio signal of a predetermined time length newly selected as the processing target. When the processing has been completed for all of the acquired audio signals AS, the processing in FIG. 4 is completed.

  FIG. 5 is a flowchart schematically illustrating an example of an evaluation and display process including a software program non-transiently stored in the memory 102 and executable by the CPU 101. This evaluation and display processing may be performed simultaneously and in parallel with the processing of FIG. 4, or may be performed at another time. Note that the evaluation and display processing shown in FIG. 5 corresponds to the operation of the display unit 14 and the evaluation unit 15 in FIG.

  In step S11, the CPU 101 obtains music score data of a music piece played by the obtained sound signal AS (for example, obtains music score data from the music database 20 in FIG. 1), and displays a screen of the display device 104 based on the music score data. Display the score at the top. For reference, an example of a musical score to be displayed is partially shown in FIG.

  In step S12, the CPU 101 determines whether bowing information is included in the acquired score data. If the bowing information is not included or if only a part of the bowing information is included, the process proceeds to step S13, and as described above, the normal bowing rule (for example, the sound of alternately performing the downward bow and the upward bow, the sound Returns a bow at a break, the sound immediately after the bar line is basically played with a down bow, and within the same note and does not return the bow in the case of slurs and ties, etc.), it corresponds to each note in the song In addition, bowing information is automatically generated and stored. For example, in FIG. 6 (a), a bowing symbol indicating a raising bow is written at a note location indicated by a reference symbol B1, and a bowing symbol indicating a lowering bow is written at a note location denoted by a reference symbol B2 or B3. . In step S13, exemplary bowing information is generated for other note locations in accordance with normal bowing rules. FIG. 6D shows an example of exemplary bow movement information generated corresponding to FIG. 6A. In FIG. 6D, a raised bow and a lowered bow are shown using known music symbols.

  In step S14, the CPU 101 acquires the bow operation timing detection information BD and the bowing estimation information U / D generated (stored) in the processing of FIG.

  In step S15, the CPU 101 displays an appropriate icon indicating the detected bow operation timing on the screen of the display device 104 based on the obtained bow operation timing detection information BD, and furthermore, the acquired bow operation estimation information U An appropriate icon indicating the estimated bowing direction is displayed on the screen of the display device 104 based on / D. FIG. 6B shows an example of icon display of the detected bow operation timing. In this example, the icon is displayed as a pulse figure. FIG. 6C shows an example of icon display indicating the estimated bowing direction (that is, the bowing state in the sound signal AS). In this example, “U” indicating an upward bow and “downward” are illustrated. An icon consisting of the letter "D" is displayed.

  In step S16, based on the acquired bow operation timing detection information BD and the bowing estimation information U / D, the CPU 101 sets exemplary bowing information at the performance (sounding) timing indicated by the information BD and bowing estimation corresponding thereto. The information is compared with the information U / D, and an evaluation result is generated, displayed, and stored. FIG. 6C shows an example of an icon indicating the evaluation result. That is, comparing the bowing state in the acoustic signal AS shown in FIG. 6C and the exemplary bowing information shown in FIG. 6D, the second and third beats of the second bar in the acoustic signal AS are compared. Bowing is not correct. Therefore, in FIG. 6C, a predetermined icon W indicating that the bowing is incorrect is displayed in combination with an icon (D or U) indicating the bowing direction.

  In step S17, the CPU 101 determines whether to terminate the evaluation and display processing. If NO, the CPU 101 returns to step S14 and repeats the processing in steps S14 to S17.

  In the above-described embodiment, the case where the signal component outside the range of the target bowed string instrument to be extracted from the acoustic signal is a low-frequency noise signal component has been described. You may make it extract a signal component.

  As described above, one embodiment of the present invention has been described, but the present invention is not limited to the above-described embodiment, and various modifications may be made within the scope of the claims and the technical idea described in the specification and the drawings. Deformation is possible. For example, the performance analysis device 100 according to the present invention can be applied to a training system in which student terminals and teacher terminals are connected via the Internet. In this case, the teacher's terminal (performance analysis device 100) acquires the practice sound (acoustic signal AS) of the bowed instrument performance from the student's terminal on the Internet via the communication I / F 106, and The performance analysis based on bowing direction estimation and evaluation is displayed. Teachers can provide guidance to students, such as making comments based on them. In that case, the teacher's terminal (the performance analysis device 100) may be configured to simultaneously acquire a plurality of acoustic signals AS and analyze them simultaneously by time-division processing or parallel processing. Alternatively, on the contrary, the student terminal (performance analysis device 100) acquires the musical instrument performance sound (model) from the teacher terminal on the Internet via the communication I / F 106, and acquires the acquired performance sound (model). ) May be displayed. The student can practice the practice effectively by learning the characteristics of the performance based on the display and comparing it with the performance analysis results of the student. Of course, the terminal (performance analysis device 100) of each student may be configured to simultaneously acquire a plurality of audio signals AS and analyze them simultaneously by time-division processing or parallel processing.

  As another example, the terminals (performance analysis devices 100) of a plurality of users are connected by wire or wirelessly to exchange information (including bowing information as an analysis result) with each other, and each user's bowing The information may be able to be confirmed by other users. For example, when a plurality of users are playing their own bowed instruments in an orchestra or the like, bowing information of their own performance is transmitted from their own terminal (performance analysis device 100) to another user's terminal (performance analysis device 100). , It is possible to use a method of making it possible to perform an appropriate notification when the bowing directions are different among the users. Further, as described above, the bowing information of a plurality of users transmitted and received by wire or wireless may be centrally managed in a device for an administrator (instructor or teacher), whereby the bowing may be mistaken. It is possible to statistically elucidate a place on a score that is easy to make, a place on a score that is hard to be mistaken, and a place where another specific user performs earlier or later than the other users. Such an effect can be expected to be reported to the user or the instructor (teacher) in real time during the practice performance. In an orchestra or the like, it is important that the bowing timing and bowing direction of all the players of each part are the same for each performance part. By applying the present invention, appropriate assistance can be performed for that purpose.

  Further, the performance analysis apparatus 100 of the present invention may be configured so that at least a part of the functions of the units 10 to 16 is performed in a distributed manner by one or more computer devices on a network. The present invention also provides a computer with a step of acquiring an acoustic signal, a step of extracting a signal component outside the range of the target bowed string instrument from the acoustic signal, and detecting a bow operation timing based on the level of the extracted signal component. And a step of executing the program. The present invention also includes a step of acquiring an acoustic signal, a step of extracting a signal component outside the range of the target bowed instrument from the acoustic signal, and a step of detecting a bow operation timing based on the level of the extracted signal component. And may be configured as a computer or processor implemented method invention.

REFERENCE SIGNS LIST 100 Performance analyzer, 10 acquisition unit, 11 extraction unit, 12 detection unit, 13 estimation unit, 14 display unit, 15 evaluation unit, 16 range setting unit, 101 microprocessor unit (CPU), 102 memory, 103 operating device, 104 Display device, 105 audio interface, 106 communication interface.

Claims (10)

An acquisition unit that acquires an acoustic signal;
An extraction unit that extracts a signal component outside the range of the target bowed instrument from the acoustic signal,
A performance analysis device comprising: a detection unit that detects a bow operation timing based on the level of the extracted signal component. The performance according to claim 1, further comprising: an estimating unit configured to estimate a bowing direction at each of the detected bow operation timings based on a correlation between sound volumes of the acoustic signals in each section sandwiched by adjacent bow operation timings. Analysis equipment.   The estimating unit may be configured such that the volume of the audio signal in the first section starting from the first bow operation timing is larger than the volume of the audio signal in the second section immediately before the first section, and the volume difference is a predetermined volume. If it is greater than one threshold, the bowing direction at the first bowing operation timing is estimated as the first bowing direction, while the volume of the sound signal in the first section is the volume of the sound signal in the second section. 3. The performance analysis device according to claim 2, wherein when the volume difference is smaller than a predetermined second threshold value, the bowing direction at the first bowing operation timing is estimated as a second bowing direction. 4.   The performance analysis device according to any one of claims 1 to 3, wherein the extraction unit includes a filter that extracts a signal component in a lower frequency range than the range of the target bowed instrument.   The performance analysis device according to claim 1, further comprising a device that presents information indicating the detected bow operation timing or the estimated bowing direction.   3. The apparatus according to claim 2, further comprising: an evaluator configured to evaluate bowing in the acquired acoustic signal by comparing information indicating an exemplary bowing direction with respect to a certain song with information indicating the estimated bowing direction. A performance analyzer according to any one of claims 1 to 4.   The sound signal acquired by the acquisition unit, further includes a range setting unit that identifies a type of the target bowed instrument and sets a range of the signal component to be extracted by the extraction unit according to the identified instrument type. Item 7. The performance analyzer according to any one of Items 1 to 6. The acquisition unit is configured to acquire a plurality of the acoustic signals in a wired or wireless manner,
The extraction unit extracts the signal component for each of the acquired acoustic signals,
The performance analyzer according to any one of claims 1 to 7, wherein the detection unit detects the bow operation timing for each of the acquired acoustic signals. A plurality of the performance analyzers according to any one of claims 1 to 8,
A system configured to connect a plurality of the performance analyzers by wire or wirelessly to communicate information. On the computer,
Obtaining an acoustic signal;
Extracting a signal component outside the range of the target bowed instrument from the acoustic signal;
Detecting the bow operation timing based on the level of the extracted signal component.

Images