Detection device, mute device for wind instrument, and method for computing radiated sound waveform

Abstract

A detection device 20 comprises: a first actuator 21 that generates sound waves to be transmitted through a pipe 3 of a clarinet 1; a first sensor unit 31 that has a plurality of sensors 31a-31h for detecting the sound waves; and a first attachment part 41 with which the first actuator 21 and the first sensor unit 31 are disposed in the pipe 3. The plurality of sensors 31-31h are arranged at intervals in the longitudinal direction of the pipe 3.

Description

Detection device, mute device for wind instruments, and method for calculating radiated sound waveform

[0001] The present disclosure relates to a detection device, a mute device for a wind instrument, and a method for calculating a radiated sound waveform.

[0002] The '691 patent describes a system for representing the sound of a reed instrument that includes a speaker configured to deliver sound to an air chamber of the reed instrument, a microphone that receives sound from the air chamber, and a processing unit that receives a measurement signal from the microphone. The processing unit generates an output signal from the measurement signal that indicates which musical note is being played by the reed instrument. The system also includes a pressure sensor that sends a signal to the processing unit that indicates that a user of the reed instrument is blowing into the reed instrument.

[0003] Special table 2018-521367 publication

[0004] The system described in Patent Document 1 can identify pitch by detecting the resonant frequency within the tube of a reed instrument, but cannot detect the open / closed state of each of the reed instrument's tone holes, and therefore cannot fully obtain parameters used to calculate the radiated sound waveform. Therefore, with conventional systems, it is difficult to accurately grasp the playing state of a reed instrument. In consideration of the above, one aspect of the present disclosure aims to more accurately detect the playing state of a wind instrument.

[0005] In order to solve the above problems, a detection device according to one aspect of the present disclosure comprises a generating unit that generates waves that propagate within a tube of a wind instrument, and a first sensor unit having a plurality of sensors that detect the waves, wherein the generating unit and the first sensor unit are disposed within the tube, and each of the plurality of sensors is disposed at intervals from one another in the longitudinal direction of the tube.

[0006] A detection device according to one aspect of the present disclosure comprises a first actuator that generates sound waves by vibrating the air inside a tube of a wind instrument, a first sensor unit having a plurality of sensors that detect the sound waves, and a first mounting unit that positions the first actuator and the first sensor unit inside the tube.

[0007] A mute device for a wind instrument according to one aspect of the present disclosure comprises the detection device and a blocking device arranged between the mouthpiece of the wind instrument and the tube, blocking air flowing from the mouthpiece into the tube.

[0008] A method for calculating a radiated sound waveform according to one aspect of the present disclosure estimates a tube shape model representing the tube shape of a wind instrument based on standing wave information relating to the distribution of standing waves generated within the tube of the wind instrument, and calculates a waveform relating to the radiated sound from the wind instrument based on the estimated tube shape model and breath pressure information relating to the breath pressure of a performer playing the wind instrument.

[0009] A method for calculating a radiated sound waveform according to one aspect of the present disclosure estimates a tube shape model representing the tube shape of a wind instrument based on traveling wave information relating to the distribution of traveling waves generated within the tube of the wind instrument, and calculates a waveform relating to the radiated sound from the wind instrument based on the estimated tube shape model and breath pressure information relating to the breath pressure of a performer playing the wind instrument.

[0010] A method for calculating a radiated sound waveform according to one aspect of the present disclosure estimates the shape of the oral cavity of a player playing a wind instrument having a manipulator based on vibration information related to the vibration of air in the oral cavity, and calculates a waveform related to the radiated sound from the wind instrument based on operation information indicating the player's operation of the manipulator, the estimated shape of the oral cavity, and breath pressure information related to the player's breath pressure.

[0011] FIG. 1 is a cross-sectional view along the central axis of a wind instrument mute device and a clarinet according to the first embodiment. FIG. 2 is a top view of a wind instrument mute device and a clarinet according to the first embodiment. FIG. 3 is a front view of a wind instrument mute device and a clarinet according to the first embodiment. FIG. 4 is a block diagram showing an example of the configuration of a detection device for a wind instrument mute device according to the first embodiment. FIG. 5 is a diagram showing an example of a tone hole opening and closing pattern of the clarinet of FIG. 1. FIG. 6 is a cross-sectional view showing the mouthpiece and reed of FIG. 1. FIG. 7 is a diagram showing the shape of a vibrating part of a reed. FIG. 8 is an explanatory diagram of incident waves and reflected waves in an oral cavity with a simple shape. FIG. 9 is an explanatory diagram of incident waves and reflected waves in an oral cavity with a more complex shape. FIG. 11 is a schematic explanatory diagram of incident waves, reflected waves, and transmitted waves in an oral cavity with a more complex shape. FIG. 12 is a block diagram showing the configuration of a physical model sound source unit. FIG. 13 is a configuration diagram of a reed dynamic characteristics block. FIG. 14 is a block diagram showing an example of the configuration of a detection device according to the second embodiment. FIG. 15 is a cross-sectional view showing a simple physical model according to the second embodiment.

[0012] A: First embodiment A1: Configuration of mute device for wind instruments

[0013] Fig. 1 is a cross-sectional view along the central axis of a wind instrument mute device and a clarinet according to the first embodiment. Fig. 2 is a top view of the wind instrument mute device and a clarinet according to the first embodiment. Fig. 3 is a front view of the wind instrument mute device and a clarinet according to the first embodiment. While Fig. 1 is a view of the wind instrument mute device and a clarinet viewed from the horizontal direction, Fig. 2 is a view of the wind instrument mute device and a clarinet viewed from above.

[0014] The wind instrument mute device 10 is attached between the mouthpiece 2 and the tube 3 of a clarinet 1, which is a wind instrument. In addition to the mouthpiece 2 and the tube 3, the clarinet 1 also includes a reed 4. In FIG. 1, the mouthpiece 2 is oriented so that the bottom of the page faces downward.

[0015] The tube 3 has a plurality of tone holes 5a-5e opening in the longitudinal direction of the tube 3. A plurality of controls 6a-6e are provided corresponding to the tone holes 5a-5e. The performer can control the open / closed state of each of the tone holes 5a-5e by operating each of the controls 6a-6e. Note that while five tone holes and five controls are shown in Fig. 1 as a schematic diagram, the number of tone holes and controls actually differs from the number of tone holes and controls that an actual clarinet has.

[0016] The mute device 10 for a wind instrument includes a detection device 20 and a cut-off device 50 .

[0017] The detection device 20 includes a first actuator 21 as a generating unit, a first sensor unit 31, a first mounting unit 41, a second actuator 22, a second sensor unit 32, a second mounting unit 42, a third actuator 23, a third sensor unit 33, a third mounting unit 43, a fourth sensor unit 34, and a fourth mounting unit 44.

[0018] The first mounting portion 41 disposes the first actuator 21 and the first sensor portion 31 inside the pipe 3. The first mounting portion 41 is a cylindrical member having a flange-shaped end, and extends in the longitudinal direction of the pipe 3. One end of the first mounting portion 41 is disposed near the inlet of the pipe 3. The other end of the first mounting portion 41 protrudes from the outlet of the pipe 3 to the outside of the pipe 3. The other end of the first mounting portion 41 is fixed to the pipe 3 via a fixing member 7. In FIG. 1 , the first mounting portion 41 and the shutoff device 50 are formed as separate bodies, but the first mounting portion 41 and the shutoff device 50 may also be formed as a single body.

[0019] As shown in Fig. 3, the fixing member 7 has a center portion 71, three arms 72a, 72b, and 72c, and three hook portions 73a, 73b, and 73c. ​​The center portion 71 is attached to the other end of the first attachment portion 41. For example, as shown in Fig. 3, the center portion 71 is annular, and the other end of the first attachment portion 41 is fitted into the center portion 71.

[0020] Three arms 72a, 72b, and 72c extend radially from a central portion 71. The three arms 72a, 72b, and 72c are flexible. Three hooks 73a, 73b, and 73c are fixed to the ends of the three arms 72a, 72b, and 72c, respectively. Each of the three hooks 73a, 73b, and 73c is designed to hook onto an end of the pipe 3. This allows the other end of the first mounting portion 41 to be fixed to the pipe 3 without wobbling. Therefore, the position of the first mounting portion 41 relative to the pipe 3 is fixed. Note that the other end of the first mounting portion 41 does not necessarily have to protrude from the pipe 3 as long as it is fixed to the pipe 3 without wobbling.

[0021] The first actuator 21 generates waves in the audible sound band that propagate within the tube 3. Hereinafter, waves in the audible sound band will be referred to as sound waves. The first actuator 21 is provided on the flange portion 41a of the first mounting portion 41. The first actuator 21 is a speaker, a horn driver, a piezoelectric element, or the like. Therefore, the first actuator 21 generates sound waves and vibrates the air within the tube 3. Note that the first actuator 21 may also be provided on the first mounting portion 41 in a portion near the outlet of the tube 3.

[0022] The first sensor unit 31 is provided on the cylindrical portion 41b of the first mounting portion 41. The first sensor unit 31 has eight sensors 31a, 31b, 31c, 31d, 31e, 31f, 31g, and 31h. Each of the eight sensors 31a to 31h is a microphone, a pressure transducer, or the like. Note that, although the first sensor unit 31 has eight sensors in FIG. 1, the number of sensors is not limited to this.

[0023] The eight sensors 31a to 31h are arranged at intervals from one another in the longitudinal direction of the pipe 3. In particular, in this example, the eight sensors 31a to 31h are arranged at equal intervals parallel to the central axis C1 of the pipe 3. The eight sensors 31a to 31h are preferably arranged at intervals of 5 to 15 cm, but are not particularly limited to these values. Furthermore, the sensors 31a to 31h are preferably arranged along the central axis C1.

[0024] The second mounting portion 42 positions the second actuator 22 and the second sensor portion 32 on the lead 4. The second mounting portion 42 has a first cantilever 42a and a second cantilever 42b. One end of the first cantilever 42a is fixed to the isolating device 50. The second actuator 22 is provided on the other end of the first cantilever 42a. The second actuator 22 is pressed against the lead 4 by the first cantilever 42a.

[0025] One end of the second cantilever 42b is fixed to the isolating device 50. The other end of the second cantilever 42b is provided with the second sensor portion 32. The second sensor portion 32 is pressed against the lead 4 by the second cantilever 42b.

[0026] The second actuator 22 vibrates the reed 4. The second actuator 22 is an actuator that uses a piezoelectric element. The second actuator 22 can displace the reed 4 by utilizing the inverse piezoelectric effect of the piezoelectric element. For example, when a sine wave voltage is applied to the piezoelectric element, the second actuator 22 can vibrate the reed 4 at the frequency of the sine wave.

[0027] The second actuator 22 generates vibrations having frequency components equal to or higher than the lowest pitch of the clarinet 1. This vibration may be generated, for example, by applying to the piezoelectric element a voltage waveform that has been generated by passing a random noise waveform through a high-pass filter that reduces frequency components lower than the lowest pitch of the clarinet 1. Alternatively, this vibration may be generated by applying to the piezoelectric element a voltage waveform that is a sweep of a sine wave equal to or higher than the lowest pitch.

[0028] The second sensor unit 32 has one or more sensors and detects vibration of the lead 4. The second sensor unit 32 is a contact-type vibration sensor. The second sensor unit 32 detects the displacement, displacement speed, and displacement acceleration of the lead 4. The vibration sensor is a sensor using a piezoelectric element. The vibration sensor converts pressure changes caused by vibration of the lead 4 into an electrical signal by utilizing the piezoelectric effect of the piezoelectric element. The second sensor unit 32 may be configured as an acceleration sensor using MEMS (Micro Electro Mechanical Systems), an acceleration sensor using optical fiber, or the like. The vibration of the lead 4 is not limited to being detected by a contact-type vibration sensor, and may also be detected by a non-contact sensor. For example, the vibration of the lead 4 may be detected by a photocoupler.

[0029] The third mounting portion 43 positions the third actuator 23 and the third sensor portion 33 within the mouthpiece 2 .

[0030] The third actuator 23 vibrates the air in the mouth of the player and is a speaker, a horn driver, a piezoelectric element, or the like.

[0031] The third sensor unit 33 detects vibrations of the air in the mouth of the performer. The third sensor unit 33 is a microphone, a pressure transducer, or the like. The pressure transducer may utilize the resistance change of an isotropic conductor or the piezoelectric effect of a piezoelectric element. The third sensor unit 33 may also be an acceleration sensor.

[0032] The fourth mounting portion 44 positions the fourth sensor portion 34 within the mouthpiece 2. The fourth sensor portion 34 has one or more sensors. The fourth sensor portion 34 detects the breath pressure of the player and converts the detected breath pressure into an electrical signal. The converted electrical signal will be referred to hereinafter as a breath pressure signal. The fourth sensor portion 34 may be a gas pressure sensor, a wind speed sensor, a flow speed sensor, or the like.

[0033] The cutoff device 50 is disposed between the mouthpiece 2 and the tube 3 of the clarinet 1. The cutoff device 50 is a cylindrical member. It has a main body 50a, a small-diameter portion 50b, and an annular portion 50c. The outer diameter of the small-diameter portion 50b is the same as or smaller than the inner diameter of the tube 3. The small-diameter portion 50b fits into the tube 3. The annular portion 50c is annular, and its inner diameter is the same as or smaller than the outer diameter of the mouthpiece 2. The annular portion 50c fits into the mouthpiece 2. Note that while the cutoff device 50 is composed of the main body 50a, the small-diameter portion 50b, and the annular portion 50c, this configuration is suitable for the shape of the clarinet 1 and may not be suitable for all other reed instruments. The shape of the cutoff device 50 can be modified as appropriate depending on the shape of the reed instrument.

[0034] The shutoff device 50 and the first mounting portion 41 are joined at the surface of the shutoff device 50 facing the tube 3, i.e., between the bottom surface of the small diameter portion 50b and the end surface of the flange portion 41a of the first mounting portion 41. The inner wall surface 50d of the shutoff device 50 and the inner wall surface 41c of the first mounting portion 41 are continuous without any steps. The shutoff device 50 blocks air from flowing into the tube 3 from the mouthpiece 2. In other words, the space inside the mouthpiece 2 and the space inside the tube 3 are not connected.

[0035] Meanwhile, the space inside the mouthpiece 2 and the space inside the first mounting portion 41 are connected. As described above, the space inside the first mounting portion 41 is formed by the inner wall surface 41c. Therefore, air flowing in from the mouthpiece 2 passes through the inside of the first mounting portion 41 and is released to the outside from the end of the first mounting portion 41. Therefore, even when a player of the clarinet 1 blows into the mouthpiece 2, the blown breath does not flow into the tube 3, and therefore the player's breath does not contribute to the generation of air vibrations within the tube 3. Therefore, the wind instrument mute device 10 functions as a mute device for the clarinet 1. Note that the main body 50a may have an air vent hole. If the main body 50a has an air vent hole, even when the player blows into the mouthpiece 2, the blown breath will flow out through the air vent hole and will not flow into the tube 3. Furthermore, if the main body 50a has an air vent hole, the space inside the first mounting portion 41 does not need to be provided.

[0036] A2: Configuration of the Detection Device Hereinafter, the configuration of the detection device of the wind instrument mute device 10 according to the first embodiment will be described with reference to FIG.

[0037] 4 is a block diagram showing an example of the configuration of the detection device of the wind instrument mute device 10 according to the first embodiment. The detection device 20 includes a first actuator 21, a second actuator 22, a third actuator 23, a first sensor unit 31, a second sensor unit 32, a third sensor unit 33, a fourth sensor unit 34, a storage device 80, and a processing device 90.

[0038] The storage device 80 is a computer-readable recording medium (e.g., a non-transitive recording medium readable by a computer). The storage device 80 includes a non-volatile memory and a volatile memory. The non-volatile memory is, for example, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), and an electrically erasable programmable read-only memory (EEPROM). The volatile memory is, for example, a random access memory (RAM).

[0039] The storage device 80 stores a program p1 and various information. The program p1 defines the operation of the detection device 20. The storage device 80 may store the program p1 read from a storage device in a server (not shown). In this case, the storage device in the server is an example of a computer-readable recording medium.

[0040] The processing device 90 includes one or more central processing units (CPUs). The one or more CPUs are examples of one or more processors. Each of the processing device, processor, and CPU is an example of a computer.

[0041] The processing device 90 reads a program p1 from the storage device 80. By executing the program p1, the processing device 90 functions as a control unit 91, a first signal processing unit 92, a pattern estimation unit 92a, a tubular body shape model estimation unit 92b, a second signal processing unit 93, a third signal processing unit 94, a fourth signal processing unit 95, and a physical model sound source unit 96. At least one of the control unit 91, the first signal processing unit 92, the pattern estimation unit 92a, the tubular body shape model estimation unit 92b, the second signal processing unit 93, the third signal processing unit 94, the fourth signal processing unit 95, and the physical model sound source unit 96 may be realized by a circuit such as a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA).

[0042] The control unit 91 controls the first actuator 21 , the second actuator 22 , and the third actuator 23 .

[0043] The first signal processing unit 92 is composed of an amplifier circuit and a sample-and-hold circuit. The first signal processing unit 92 acquires detection signals output from the eight sensors 31a to 31h of the first sensor unit 31 while the first actuator 21 is in operation. Based on the acquired detection signals, the first signal processing unit 92 extracts standing wave information, which is information regarding the distribution of standing waves generated within the tube 3, and outputs the extracted standing wave information. During the operation of the first actuator 21, a sinusoidal sweep waveform in the audible sound band may be input to the first actuator 21. The sweep waveform may also be applied to the first actuator 21 in semitone increments.

[0044] The standing wave information is extracted as follows. First, the control unit 91 vibrates the first actuator 21 at a frequency in the audible range. When the vibration of the first actuator 21 vibrates the air inside the tube 3, standing waves are generated inside the tube 3 according to the open / closed states of the sound holes 5a to 5e. The standing wave is a composite wave of a progressive wave and a backward-reaching wave. The progressive wave is a wave that travels from the entrance of the tube 3 toward the exit of the tube 3, and the backward-reaching wave is a wave that travels from the exit of the tube 3 toward the entrance of the tube 3 as a result of the progressive wave being reflected at the exit of the tube 3.

[0045] Each of the sensors 31a to 31h detects the sound pressure of a standing wave at the point where the sensor 31a to 31h is located and outputs the detected sound pressure. Each of the sensors 31a to 31h detects the superposition of a forward wave and a backward wave as a standing wave. When each of the sensors 31a to 31h is a microphone, the sensor 31a to 31h outputs a sound pressure level corresponding to the detected amplitude strength. The output sound pressure level from each of the sensors 31a to 31h varies in a sinusoidal manner, but the change pattern of each output sound pressure level differs depending on the position of each of the sensors 31a to 31h.

[0046] The amplitude of the sound pressure detected by a sensor located at an antinode of the standing wave is maximum, and the amplitude of the sound pressure detected by a sensor located at a node of the standing wave is minimum. The distribution of the amplitude of the change in each output sound pressure level within the pipe 3 corresponds to the sound pressure distribution of the standing wave generated within the pipe 3.

[0047] Therefore, the first signal processing unit 92 acquires each output sound pressure level that varies in a sinusoidal waveform from each of the sensors 31 a to 31 h. The first signal processing unit 92 amplifies each acquired output sound pressure level in an amplifier circuit. The first signal processing unit 92 detects the peak value of each amplified output sound pressure level from each amplified output sound pressure level using a sample-and-hold circuit.

[0048] Each detected peak value indicates the sound pressure distribution of the standing wave at each sensor position. Therefore, the first signal processing unit 92 can estimate the sound pressure distribution of the standing wave in the pipe 3 based on the amplitude of the sound pressure level detected by each sensor. The first signal processing unit 92 outputs amplitude information of the sound pressure level at each sensor 31a to 31h to the pattern estimation unit 92a. In this way, the standing wave information is based on information output from the multiple sensors 31a to 31h.

[0049] The pattern estimation unit 92a estimates the sound hole opening and closing pattern based on the amplitude information of the sound pressure levels of the sensors 31a to 31h.

[0050] FIG. 5 is a diagram showing an example of a tone hole opening / closing pattern for the clarinet 1 of FIG. 1 . In FIG. 5 , a specific pitch i (i=1, 2, ..., N) of the clarinet 1 is associated with the open / closed state of each tone hole of the clarinet 1. Tone hole numbers are assigned in order according to the distance from the outlet of the tube 3, with the tone hole closest to the outlet of the tube 3 being numbered 1 and the tone hole farthest from the outlet of the tube 3 being numbered 24. The open / closed state of each tone hole is represented by 0 or 1. Here, "0" indicates a closed state of the tone hole, and "1" indicates an open state of the tone hole. The pattern in which the open / closed states of each tone hole are arranged in the order of the tone hole numbers is the tone hole opening / closing pattern of the clarinet 1, and the tone hole opening / closing pattern for a specific pitch i is called OCP(i). The tone hole opening / closing pattern OCP(i) is a discrete value.

[0051] According to Figure 5, the tone hole opening and closing pattern OCP(1) is a value for the pitch "E3." When the pitch "E3" is produced, all tone holes are closed, so the tone hole opening and closing pattern OCP(1) has values ​​of "0" corresponding to each tone hole number. The tone hole opening and closing pattern OCP(i) is a value for the pitch "F#3." When the pitch "F#3" is produced, the tone hole opening and closing pattern OCP(i) has values ​​of "1" corresponding to tone hole numbers 1 and 2, and values ​​of "0" corresponding to tone hole numbers 3 to 24.

[0052] The tone hole opening and closing pattern OCP(i) is information relating to the opening and closing status of each of the multiple tone holes of the clarinet 1. Therefore, the tone hole opening and closing pattern OCP(i) can also be said to be information relating to the fingering status of the player of the clarinet 1. Below, a machine learning method using a classification technique for the tone hole opening and closing pattern OCP(i) will be described.

[0053] In this embodiment, learning is performed using a support vector machine as supervised learning using a classification method. Input data to the support vector machine is acquired as follows: While the clarinet 1 is being fingered in a manner corresponding to each tone hole opening and closing pattern OCP(i), white noise is input to the first sensor unit 31 for a fixed period of time. The white noise input to the first sensor unit 31 may be generated, for example, from the first actuator 21, or from another actuator, speaker, or the like. If the multiple sensors of the first sensor unit 31 are composed of m microphones, the output sound pressure levels of the m microphones are time-averaged. The time-average of each output sound pressure level is attempted a total of n times for each tone hole opening and closing pattern.

[0054] The output sound pressure level of the first microphone, the output sound pressure level of the second microphone, ..., and the output sound pressure level of the mth microphone in the first trial are represented as L1(1) decibels, L2(1) decibels, ..., and Lm(1) decibels, respectively. The output sound pressure level of the first microphone, the output sound pressure level of the second microphone, ..., and the output sound pressure level of the mth microphone in the nth trial are represented as L1(n) decibels, L2(n) decibels, ..., and Lm(n) decibels, respectively.

[0055] The first trial result, I1 (L1(1), L2(1), ..., Lm(1)), the second trial result, I2 (L1(2), L2(2), ..., Lm(2)), ..., the nth trial result, In (L1(n), L2(n), ..., Lm(n)), are given as input data to the support vector machine. Hereinafter, I1, I2, ..., In are referred to as standing wave sound pressure distribution information. In addition, the sound hole opening and closing pattern OCP(i) is given as output data of the support vector machine.

[0056] The machine learning device for the support vector machine model inputs standing wave sound pressure distribution information I1 to In as input data into the support vector machine model, and trains the support vector machine model to establish a correlation between the standing wave sound pressure distribution information I1 to In as input data and the sound hole opening / closing pattern OCP(i) as output data.

[0057] When the machine learning device determines that a predetermined training termination condition is satisfied, it terminates the training and stores the support vector machine model at that point as a trained model in the storage device 80. The predetermined training termination condition is, for example, that the number of iterations of the series of training processes reaches a predetermined number.

[0058] In this way, the relationship between the combination of the output sound pressure levels of the m microphones and the sound hole opening / closing pattern OCP(i) is learned.

[0059] The pattern estimation unit 92a outputs the estimated sound hole opening and closing pattern to the tubular body shape model estimation unit 92b. The pattern estimation unit 92a has a trained support vector machine. Therefore, when amplitude information of the sine wave signals from each sensor of the first sensor unit 31, i.e., a combination of output sound pressure levels from each sensor, is input, the pattern estimation unit 92a predicts the corresponding sound hole opening and closing pattern OCP(i).

[0060] The tube shape model estimating unit 92b receives the estimated sound hole opening and closing pattern from the pattern estimating unit 92a. The tube shape model estimating unit 92b references the sound hole opening and closing pattern database 81 and estimates a tube shape model based on the input sound hole opening and closing pattern and the sound hole opening and closing pattern database 81. The tube shape model estimating unit 92b outputs the estimated tube shape model to the physical model sound source unit 96.

[0061] The second signal processing unit 93 calculates reed resonance characteristic information based on the vibration information of the reed 4 input from the second sensor unit 32, and outputs the information to the physical model sound source unit 96. More specifically, the second signal processing unit 93 acquires, as vibration information of the reed 4, a waveform related to the displacement of the reed 4 from the second sensor unit 32, as follows: the control unit 91 activates the second actuator 22, the second actuator 22 vibrates the reed 4, and the second signal processing unit 93 acquires, as vibration information of the reed 4, a waveform related to the displacement of the reed 4 from the second sensor unit 32. In this way, the second signal processing unit 93 can also be said to be a reed state estimation unit that estimates the state of change in the shape of the reed 4 based on the detection result by the second sensor unit 32.

[0062] FIG. 6 is a cross-sectional view showing the mouthpiece 2 and the reed 4. FIG. 7 is a diagram showing the shape of the vibrating portion of the reed 4. As shown in FIG. 6, the x-axis is an axis passing through the origin O, when the tip of the mouthpiece 2 is the origin O, and is parallel to the axis of the mouthpiece 2. That is, when the mouthpiece 2 and the tube 3 are connected, the x-axis is an axis parallel to the central axis C1 of the tube 3. The y-axis is an axis passing through the origin O and aligned along the direction of the tip opening. That is, the x-axis and the y-axis are perpendicular to each other. The positive direction of the y-axis is downward in FIG. 6, i.e., the direction in which the reed 4 moves away from the mouthpiece 2 when the reed 4 vibrates.

[0063] In FIG. ｒ is the effective vibration length of the reed 4, that is, the length of the portion of the reed 4 that substantially contributes to vibration. b is the width of the tip of the reed 4. In FIG. ｅｘｔ is the external force in the y direction that the player applies to the reed 4 through the player's lips. H is the external force F ｅｘｔ is the y-coordinate value of the tip of the lead 4 when y is 0. ０ is the external force F ｅｘｔis the y coordinate value of the tip of the lead 4 when the force is applied to the lead 4. In other words, y ０ indicates the static opening of the reed 4. １ is the area of ​​the gap between the reed 4 and the mouthpiece 2. １ is called the lead gap area.

[0064] m ｒ is the effective vibration mass of the reed 4, that is, the mass of the part of the reed 4 that vibrates. ｒ is called the reed effective seismic mass. ｒ is the effective vibration area of ​​the reed 4. ｒ is called the effective reed vibration area. μ ｒ is the effective mass of the lead 4 per unit area. ｒ is called the lead unit effective mass. H is the external force F ｅｘｔ is the y-coordinate value of the tip of the reed 4 when is 0. In other words, H is the initial value of the static opening degree of the reed 4.

[0065] The second signal processing unit 93 calculates the reed resonance characteristic based on the acquired waveform relating to the displacement of the reed 4 along the y-axis, and outputs the result as reed resonance characteristic information. The reed resonance characteristic information includes the resonant frequency f ｒ and the opening y of the lead 4 along the y axis ０ is.

[0066] The second signal processing unit 93 also calculates the static displacement of the lead 4 at the contact position between the lead 4 and the second sensor unit 32 by passing the acquired waveform relating to the displacement of the lead 4 along the y axis through an LPF. Furthermore, the second signal processing unit 93 calculates the opening degree y of the tip portion of the lead 4 based on the static displacement at the contact position between the lead 4 and the second sensor unit 32. ０ Calculate the following.

[0067] The displacement of the tip of the lead 4 may be detected using a photocoupler. In this case, the static opening of the tip of the lead may be obtained by applying a low-pass filter to the detection signal of the photocoupler. This allows the displacement of the tip of the lead 4 to be detected more accurately.

[0068] The third signal processing unit 94 estimates the shape of the performer's oral cavity in real time based on the waves reflected in the performer's oral cavity, and outputs information about the estimated oral cavity shape as oral cavity shape information to the physical model sound source unit 96. In this way, the third signal processing unit 94 can also be said to be an oral cavity shape estimation unit that estimates the shape of the oral cavity based on the detection results by the third sensor unit 33.

[0069] More specifically, when the player holds the mouthpiece 2 in his / her mouth and the control unit 91 vibrates the third actuator 23, the third signal processing unit 94 acquires from the third sensor unit 33 a sound pressure signal detected by the third sensor unit 33. At this time, the sound pressure signal acquired by the third signal processing unit 94 includes signals due to reflected waves reflected in the player's oral cavity and airway.

[0070] For example, the Ware and Aki algorithm is employed as a method for estimating the oral cavity shape in real time in the third signal processing unit 94. Because the oral cavity and the airway are continuous, the oral cavity shape estimated by the algorithm also includes the airway shape.

[0071] 8 is an explanatory diagram of incident and reflected waves in a simple oral cavity. In the oral cavity shown in FIG. 8, the cross-sectional area of ​​the oral cavity is x ０ At the position, the inlet cross-sectional area A ０ From the first cross-sectional area A １ Under the condition that the pressure and volumetric flow velocity are conserved even at the point of change in cross-sectional area, that is, at the point of discontinuity, the sound reflection coefficient r is ｉ , the reflected wave is P ｒ is expressed as follows:

[0072] r = (P ｒ / P ｉ ) × (A ０ -A １ ) / (A ０ +A １ ) ... (1)

[0073] Here, the inlet cross-sectional area A ０ Since is known, the first cross-sectional area A１ is calculated as in the following equation (2).

[0074] A １ = [(P ｒ -r x P ｉ ) / (P ｒ +r × P ｉ ) ]×A ０ ... (2)

[0075] 9 is an explanatory diagram of incident and reflected waves in a more complicatedly shaped oral cavity. ０ , 2x ０ , 3x ０ , and 4x ０ The cross-sectional area of ​​the oral cavity at the position of ０ , A １ , A ２ , A ３ 10 is a schematic explanatory diagram of incident waves, reflected waves, and transmitted waves in a more complicatedly shaped oral cavity. ｉ When incident, at time T, the position x ０ The reflected wave P(T) reflected by the cross section at time 2T and position 2x ０ The reflected wave P(2T) reflected by the cross section at is observed.

[0076] Here, using equation (2), the first cross-sectional area A １ and the second cross-sectional area A ２ At time 3T, the position 3x ０ Reflected wave Pr from the cross section １ and the multiple reflected wave Pr shown by the dashed line ２ The mixed wave of the multiple reflected wave Pr ２ is the inlet cross-sectional area A ０ , first cross-sectional area A １ , and the second cross-sectional area A ２ Therefore, the mixed wave is multiply reflected by Pr ２ Therefore, the reflected wave Pr can be removed from the mixed wave. １ Since it is separated, the third cross-sectional area A ３The Ware and Aki algorithm estimates the oral cavity shape taking into account the effect of such multiple reflections.

[0077] The fourth signal processing unit 95 acquires from the fourth sensor unit 34 a breath pressure signal detected by the fourth sensor unit 34. The fourth signal processing unit 95 extracts information related to breath pressure based on the acquired breath pressure signal and outputs the extracted information related to breath pressure as breath pressure information to the physical model sound source unit 96. In this way, the fourth signal processing unit 95 can also be said to be a breath pressure information generating unit that generates breath pressure information based on the detection result by the fourth sensor unit 34.

[0078] More specifically, the signal detected by the fourth sensor unit 34 is a superimposed signal generated by the second actuator 22, the signal generated by the third actuator 23, and the signal generated from the mouth of the performer. The signals generated by the second actuator 22 and the third actuator 23 are mainly signals with frequencies in the audible range, but the signal related to the breath pressure generated from the mouth of the performer is a signal with frequency components lower than the audible range.

[0079] To separate the signal detected by the second actuator 22 from the signal detected by the fourth sensor 34, the second actuator 22 can vibrate the reed 4 with a small amplitude to prevent interference with air vibrations. To separate the signal detected by the third actuator 23 from the signal detected by the fourth sensor 34, a low-pass filter can be used. Therefore, the signal related to breath pressure is separated from the signals generated by the second actuator 22 and the third actuator 23 by vibrating the reed 4 with a small amplitude using the second actuator 22 and by using a low-pass filter. This makes it possible to extract a breath pressure signal that is not affected by resonance in the oral cavity.

[0080] The fourth signal processing unit 95 can extract only the breath pressure information by passing the acquired breath pressure signal through a low-pass filter.

[0081] The physical model sound source unit 96 calculates a radiated sound waveform based on the output signal from the tubular body shape model estimator 92b, the output signal from the second signal processor 93, the output signal from the third signal processor 94, and the output signal from the fourth signal processor 95. The radiated sound waveform output from the physical model sound source unit 96 is output to the amplifier AM and amplified by the amplifier AM. The radiated sound waveform amplified by the amplifier AM is output to the speaker SP and converted into a sound signal by the speaker SP. Note that headphones, earphones, etc. may be used instead of the speaker SP. The configuration and operation of the physical model sound source unit 96 will be described below with reference to FIGS. 11 and 12 .

[0082] Fig. 11 is a block diagram showing the configuration of the physical model sound source unit 96. As shown in Fig. 11, the physical model sound source unit 96 has a reed dynamic characteristics block 961, a tube propagation realization block 962, an oral cavity propagation realization block 963, and a radial propagation realization block 964.

[0083] 12 is a block diagram of the read dynamic characteristics block 961. The read dynamic characteristics block 961 includes a model related to the dynamic characteristics of a single read. As shown in FIG. 12, the read dynamic characteristics block 961 includes a subtraction unit 961a, an inversion unit 961b, a read dynamic characteristics filter unit 961c, a first calculation unit 961d, a constant multiplication unit 961e, a second calculation unit 961f, and a multiplication unit 961g.

[0084] The reed dynamics block 961 includes the breath pressure signal p ０ The subtractor 961a receives the breath pressure signal p(t) and the pressure signal p(t) directly below the reed 4. ０ The pressure signal p(t) directly below the reed 4 is subtracted from the breath pressure signal p(t) to obtain the breath pressure signal p ０ The fourth signal processor 95 outputs Δp(t), which is the difference between the pressure signal p(t) and the pressure signal p(t) just below the reed 4. The difference value Δp(t) is input to the inverter 961b and the second calculator 961f. The output signal from the inverter 961b is input to the reed dynamic characteristic filter 961c. The breath pressure signal acquired by the fourth signal processor 95 is input to the reed dynamic characteristic filter p(t) input to the oral cavity side waveguide model. ０In the case where the oral cavity propagation realization block is not provided, the breath pressure signal acquired by the fourth signal processing unit 95 is related to the breath pressure signal p ０ (t).

[0085] The reed dynamic characteristic filter section 961c further includes a resonant frequency f ｒ , Q factor Q ｒ , lead unit effective mass μ ｒ , static opening of lead 4 y ０ , and the initial value H of the static opening of the reed 4. The reed dynamic characteristic filter unit 961c calculates the reed 4 resonance frequency f ｒ , Q factor Q ｒ , lead unit effective mass μ ｒ , static opening of lead 4 y ０ , and an initial value H of the static opening of the lead 4 as parameters. In this example, a second-order IIR (Infinite Impulse Response) filter is used as the digital filter.

[0086] When the difference value Δp(t) is 0, the output y of the read dynamic characteristic filter unit 961c is y ０ and the breath pressure signal p ０ It decreases monotonically as (t) increases.

[0087] The output y of the reed dynamic characteristic filter unit 961c is input to the first calculation unit 961d. The first calculation unit 961d outputs an output G(y) as the calculation result. The output G(y) is a function that limits the displacement of the reed 4. The output G(y) is an output in which the displacement along the y-axis in the region where the difference value Δp(t) indicates a negative value is limited. Furthermore, the output G(y) clips values ​​in the region where the y coordinate value is negative to 0. In other words, the output G(y) limits the y value to a value equal to or greater than 0.

[0088] The output G(y) is input to the constant multiplication unit 961e. The constant multiplication unit 961e multiplies the output G(y) by b. This calculation is performed by multiplying the area S of the gap between the reed 4 and the mouthpiece 2 by b. １ The coefficient b corresponds to the width of the tip of the lead 4.

[0089] The second calculation unit 961f calculates and outputs the air flow velocity u in the gap between the reed 4 and the mouthpiece 2 according to the following equation (3).

[0090] u=(2×Δp(t) / ρ) １／２ ...(3)

[0091] The multiplication unit 961g calculates the area S of the gap between the reed 4 and the mouthpiece 2. １ is multiplied by the flow velocity u, and the multiplication result is output as the volumetric flow velocity U(t) in the gap between the reed 4 and the mouthpiece 2.

[0092] Referring again to FIG. 11 , the tubular propagation realization block 962 has a tubular side waveguide model 962a. Tube shape information estimated by the tubular shape model estimator 92b is input to the tubular side waveguide model 962a. This forms a tubular side waveguide model. The volume flow velocity U(t) in the gap between the reed 4 and the mouthpiece 2 is input to the tubular propagation realization block 962 from the reed dynamics block 961. The tubular propagation realization block 962 outputs a pressure signal p(t) directly below the reed 4 to the reed dynamics block 961. The transfer function Z of the tubular propagation realization block 962 is ｉｎ (f) is expressed as p(f) / U(f).

[0093] Furthermore, the tubular body side waveguide model 962a calculates the sound pressure signal p ｏｕｔ (t) to the radial propagation realization block 964.

[0094] The oral cavity propagation realization block 963 has an oral cavity side waveguide model 963a. The oral cavity shape information estimated by the third signal processing unit 94 is input to the oral cavity side waveguide model 963a. In this way, an oral cavity side waveguide model is formed. The oral cavity propagation realization block 963 receives the volume flow velocity U(t) from the reed dynamics block 961. The oral cavity propagation realization block 963 calculates the breath pressure signal p ０ (t) is output to the read dynamics block 961.

[0095] The radiation propagation realization block 964 includes the sound pressure signal p ｏｕｔThe radiation propagation realization block 964 receives the sound pressure signal p ｏｕｔ (t), the radiated sound pressure signal p ｒａｄ The radial propagation realization block 964 is realized by, for example, an IIR filter.

[0096] In this way, the processing device 90 estimates a tube shape model based on the tone hole opening and closing pattern estimated by the pattern estimation unit 92a, and calculates a waveform related to the sound radiated from the clarinet 1 based on the estimated tube shape model, the estimated changes in the shape and characteristics of the reed 4, the estimated shape of the oral cavity, and the generated breath pressure information. With this configuration, the physical model sound source unit 96 can accommodate changes in timbre due to embouchure techniques, including bending and flaxenoid techniques. The processing device 90 estimates whether a bending technique is being used based on the distortion of the reed 4, and whether flaxenoid techniques are being used based on the shape of the performer's oral cavity.

[0097] A3: Summary of the First Embodiment As explained above, the detection device 20 according to the first embodiment comprises a first actuator 21 and a first sensor unit 31. The first actuator 21 generates sound waves that propagate within the tube 3 of the clarinet 1. The first sensor unit 31 has multiple sensors 31a-31h that detect the sound waves. The first actuator 21 and first sensor unit 31 are disposed within the tube 3. The multiple sensors 31a-31h are disposed at intervals from one another in the longitudinal direction of the tube 3.

[0098] The detection device 20 according to the first embodiment includes multiple sensors 31a-31h, which detect the distribution of standing waves related to sound pressure generated within the tube 3. From the distribution of standing waves, the open / closed states of the multiple tone holes, i.e., the tone hole opening / closing pattern, can be estimated. Therefore, the detection device 20 according to the first embodiment can more accurately detect the playing state of the clarinet 1. More specifically, by estimating the tone hole opening / closing pattern, parameters necessary for accurately calculating the radiated sound waveform of the clarinet 1 can be obtained. For example, even if the frequency of the emitted sound is the same, it is possible to identify playing techniques that result in different tone hole opening / closing patterns. Furthermore, because the multiple sensors 31a-31h are spaced apart along the length of the tube 3, the standing waves generated within the tube 3 can be detected more accurately. This allows for more accurate estimation of the tone hole opening / closing pattern.

[0099] In the first embodiment, the clarinet 1 is an example of a "wind instrument," and the first actuator 21 is an example of a "generator."

[0100] The detection device 20 further includes a second actuator 22 and a second sensor unit 32. The second actuator 22 vibrates the reed 4 of the clarinet 1. The second sensor unit 32 has a sensor that detects the vibration of the reed 4. The second actuator 22 and the second sensor unit 32 are disposed on the reed 4.

[0101] This allows the resonance characteristics and static opening of the reed 4 to be detected, so that even if the tone hole opening and closing patterns are the same, it is possible to accommodate playing styles that produce different pitches or timbres depending on the resonance characteristics and static opening of the reed 4. This allows for more accurate detection of the playing state of the clarinet 1.

[0102] The detection device 20 also includes a third actuator 23 and a third sensor unit 33. The third actuator 23 vibrates the air in the mouth of the player playing the clarinet 1. The third sensor unit 33 has a sensor that detects the vibration of the air in the mouth. The third actuator 23 and the third sensor unit 33 are disposed in the mouthpiece 2 of the clarinet 1.

[0103] This allows the shape of the player's mouth to be detected, making it possible to more accurately detect the playing state of the clarinet 1.

[0104] The detection device 20 further includes a fourth sensor unit 34. The fourth sensor unit 34 has a sensor that detects the breath pressure of the player playing the clarinet 1. The fourth sensor unit 34 is disposed within the mouthpiece 2 of the clarinet 1.

[0105] This allows the breath pressure of the player to be detected, so that the playing state of the clarinet 1 can be detected more accurately.

[0106] The generator generates sound waves by vibrating the air inside the tube 3 of the clarinet 1. The first actuator 21 generates sound waves by vibrating the air inside the tube 3 of the clarinet 1.

[0107] According to this, sound waves propagating within the pipe 3 are detected by a plurality of sensors 31a to 31h spaced apart from one another in the longitudinal direction of the pipe 3. Because sound waves propagating within the pipe 3 appear as standing waves, each of the sensors 31a to 31h can detect a sound pressure level corresponding to the sound pressure distribution of the standing waves within the pipe 3. The sound pressure distribution of the standing waves can be estimated from the detection results of the sound pressure levels by each of the sensors 31a to 31h.

[0108] The detection device 20 according to the first embodiment includes a first actuator 21 and a first sensor unit 31. The first actuator 21 and first sensor unit 31 are disposed inside the tube 3. The first actuator 21 generates sound waves by vibrating the air inside the tube 3 of the clarinet 1. The first sensor unit 31 includes multiple sensors 31a to 31h that detect sound waves.

[0109] Sound waves transmitted within the pipe appear as standing waves, and in order to detect the sound pressure distribution of the standing waves, it is preferable to detect the sound waves at a position close to the central axis C1 of the pipe 3. In this way, since multiple sensors 31a to 31h are disposed within the pipe 3, even if the position of the first attachment part is shifted, it is possible to selectively use the sensors close to the central axis C1 of the pipe, and it is possible to more accurately detect the sound pressure distribution of the standing waves.

[0110] The sensors 31a to 31h are arranged at intervals in the longitudinal direction of the pipe 3.

[0111] In this way, the sensors 31a to 31h are arranged at intervals from one another in the longitudinal direction of the pipe 3, so that the sound pressure distribution of standing waves occurring within the pipe 3 can be detected more accurately.

[0112] The mute device 10 for a wind instrument according to the first embodiment includes a detection device 20 and a blocking device 50. The blocking device 50 is provided between the mouthpiece 2 and the tube 3 of the clarinet 1. The blocking device 50 blocks air from flowing from the mouthpiece 2 into the tube 3.

[0113] In this way, the space inside the tube 3 is not in communication with the space inside the blocking device 50, which blocks air from flowing into the tube 3 from the mouthpiece 2. Therefore, the breath blown by the player into the mouthpiece 2 does not flow into the tube 3. As a result, the clarinet 1 is muted.

[0114] The method for calculating a radiated sound waveform according to the first embodiment estimates a tube shape model based on standing wave information, and calculates a waveform related to the radiated sound from the clarinet 1 based on the estimated tube shape model and breath pressure information. The standing wave information is information related to the distribution of standing waves generated within the tube 3 of the clarinet 1. The tube shape model is a model that indicates the tube shape of the clarinet 1. The breath pressure information is information related to the breath pressure of the performer playing the clarinet 1.

[0115] According to this, a tube shape model is determined from the detected standing waves. The tube shape model is a model that estimates pitch information and timbre information from the tube shape corresponding to the player's fingering on the clarinet 1, and therefore a sound waveform is calculated that more accurately reproduces the performance state of the clarinet 1. Furthermore, according to the above aspect, breath pressure information is generated based on the detected breath pressure of the player, and therefore a sound waveform can be calculated that matches the player's performance timing, breath pressure strength, etc., based on the breath pressure information.

[0116] Furthermore, estimating the tube shape model includes inputting standing wave information and outputting tone hole opening and closing patterns for the clarinet 1, and determining the tube shape model based on the output tone hole opening and closing patterns. The tone hole opening and closing patterns are patterns that indicate combinations of the open and closed states of each of the multiple tone holes.

[0117] According to this method, a tone hole opening and closing pattern is estimated from the detected standing waves, and a tube shape model is determined based on the estimated tone hole opening and closing pattern. The tube shape model is a model that estimates pitch information and timbre information from the tube shape that corresponds to the player's fingering on the clarinet 1, so a waveform that more accurately reproduces the playing state of the clarinet 1 can be calculated.

[0118] Furthermore, the frequency of the standing wave is included in the audible sound band.

[0119] According to this, waves in the audible sound band, i.e., sound waves, are less likely to attenuate within the tube 3 and are more likely to appear as standing waves. Therefore, the tone hole opening and closing pattern can be estimated by generating sound waves within the tube 3 and detecting the standing waves with multiple sensors. Furthermore, because the audible sound band is the same as the sound generation band of the clarinet 1, the tone hole opening and closing pattern can be easily estimated.

[0120] The method for calculating a radiated sound waveform according to the first embodiment estimates the state of change in the shape of the reed 4 based on reed vibration information, and calculates a waveform related to the radiated sound based on the estimated tubular body shape model, the estimated state of change in the shape of the reed 4, and the generated breath pressure information. The reed vibration information is information related to the vibration of the reed 4 of the clarinet 1.

[0121] This allows the resonance characteristics of the reed 4 and the static opening of the reed 4 to be estimated from the detected vibration information. Therefore, whether or not the performer is performing a bend technique can be reflected, and a sound waveform that more accurately reproduces the playing state of the clarinet 1 can be calculated.

[0122] In addition, the calculation method for a radiated sound waveform according to the first embodiment estimates a tubular body shape model using a trained model in which the correlation between standing wave sound pressure distribution information and the sound hole opening / closing pattern OCP(i) has been trained.

[0123] According to this, when standing wave sound pressure distribution information is input to a trained model, a sound hole opening and closing pattern OCP(i) corresponding to the standing wave sound pressure distribution information is output, and a tubular body shape model is estimated from the sound hole opening and closing pattern OCP(i) output from the trained model. Therefore, there is no need to develop a complex program for estimating the sound hole opening and closing pattern OCP(i) from the standing wave sound pressure distribution information, so the tubular body shape model can be easily estimated. Note that standing wave sound pressure distribution information is an example of "standing wave information."

[0124] Furthermore, the method for calculating a radiated sound waveform according to the first embodiment estimates the shape of the inside of the oral cavity based on vibration information, and calculates a waveform related to the radiated sound from the clarinet 1 based on operation information, the estimated shape of the inside of the oral cavity, and breath pressure information. The vibration information is information related to the vibration of air in the oral cavity of a player playing the clarinet 1 having the controls 6a to 6e. The operation information is information indicating the player's operation of the controls. The breath pressure information is information related to the player's breath pressure.

[0125] This allows the shape of the player's mouth to be estimated based on the detected vibrations of the air in the mouth. Therefore, the state of the player's embouchure and whether or not the player is using the flageo technique can be reflected, allowing for the calculation of a sound waveform that more accurately reproduces the playing state of the clarinet 1. The flageo technique is a playing style that can produce different pitches even if the tone hole opening and closing patterns are the same, due to the resonance characteristics of the oral cavity resulting from different oral cavity shapes.

[0126] Moreover, the method for calculating a radiated sound waveform according to the first embodiment outputs a waveform relating to the radiated sound as an electrical signal.

[0127] According to this, a sound waveform that reproduces the performance state of the clarinet 1 is output as an electrical signal, so that the performer can listen to the reproduced performance using a speaker SP, headphones, etc.

[0128] In this embodiment, the sound pressure distribution of the standing wave is observed and the sound hole opening / closing pattern is estimated from the sound pressure distribution of the standing wave, but the density distribution of the standing wave of ultrasonic waves or electromagnetic waves, not limited to sound waves, may be observed and the sound hole opening / closing pattern may be estimated from this. Standing waves of ultrasonic waves or electromagnetic waves are difficult to generate inside the pipe 3. However, it is possible to observe the density distribution of the standing wave by installing a reflector at the bell outlet, for example, and generating a reflected wave by the reflector. Therefore, in the case of ultrasonic waves or electromagnetic waves, a reflector may be installed.

[0129] Furthermore, there may be multiple reflectors inside the tube 3. In the case of electromagnetic waves, each of the multiple reflectors is preferably made of metal. In the case of ultrasonic waves, each of the multiple reflectors is preferably circular, and the material of each reflector is not particularly limited. The installation position of each of the multiple reflectors is preferably immediately behind each sound hole. In this case, the installation positions of the multiple sensors are preferably immediately before and immediately behind each sound hole.

[0130] Here, "immediately before each sound hole" refers to the position between that sound hole and the sound hole immediately before it. The sound hole immediately before each sound hole refers to the sound hole adjacent to that sound hole, which is closer to the entrance of the tube 3 than that sound hole. Also, "immediately after each sound hole" refers to the position between that sound hole and the sound hole immediately beyond that sound hole. The sound hole immediately beyond that sound hole refers to the sound hole adjacent to that sound hole, which is closer to the exit of the tube 3 than that sound hole. In other words, the sensor installed immediately before each sound hole and the sensor installed immediately after each sound hole are positioned in the longitudinal direction of the tube 3, with the sound holes sandwiched between them. Also, the diameter of each of the multiple reflectors is preferably large enough to leave a gap between them and the inner diameter of the tube 3, so that ultrasonic waves or electromagnetic waves can pass through the entire tube 3.

[0131] B: Second Embodiment A mute device for a wind instrument, a detection device, and a method for calculating a radiated sound waveform according to a second embodiment will be described below with reference to Figures 13 and 14. Note that, in the following description, for the sake of simplicity, the same reference numerals are used for the same components as in the first embodiment, and descriptions of their functions may be omitted. Also, in the following description, for the sake of simplicity, differences between the second embodiment and the first embodiment will be mainly described.

[0132] B1: Configuration of the Detection Device The configuration of the detection device according to the second embodiment will be described below with reference to Fig. 13. Fig. 13 is a block diagram showing an example configuration of the detection device according to the second embodiment. The detection device 20A includes a first actuator 21, a second actuator 22, a third actuator 23, a first sensor unit 31, a second sensor unit 32, a third sensor unit 33, a fourth sensor unit 34, a storage device 80A, and a processing device 90A.

[0133] The storage device 80A is a computer-readable recording medium, similar to the storage device 80. The storage device 80A includes a nonvolatile memory and a volatile memory. The nonvolatile memory is, for example, a ROM, an EPROM, and an EEPROM. The volatile memory is, for example, a RAM.

[0134] The storage device 80A stores a program p2 and various information. The program p2 defines the operation of the detection device 20A. The storage device 80A may store the program p2 read from a storage device in a server (not shown). In this case, the storage device in the server is an example of a computer-readable recording medium. The storage device 80A differs from the storage device 80 in that it stores the program p2 instead of the program p1 and does not have the tone hole opening and closing pattern database 81, but is the same in other respects.

[0135] The processing device 90A includes one or more CPUs. The one or more CPUs are examples of one or more processors. Each of the processing device, processor, and CPU is an example of a computer.

[0136] The processing device 90A reads the program p2 from the storage device 80A. By executing the program p2, the processing device 90A functions as a control unit 91, a first signal processing unit 92, a tubular body shape estimation unit 92c, a second signal processing unit 93, a third signal processing unit 94, a fourth signal processing unit 95, and a physical model sound source unit 96. At least one of the control unit 91, the first signal processing unit 92, the tubular body shape estimation unit 92c, the second signal processing unit 93, the third signal processing unit 94, the fourth signal processing unit 95, and the physical model sound source unit 96 may be realized by a circuit such as a DSP, an ASIC, a PLD, or an FPGA. The processing device 90A differs from the processing device 90 in that the processing device 90A functions as a tubular body shape estimation unit 92c instead of the pattern estimation unit 92a and the tubular body shape model estimation unit 92b, but is the same in other respects.

[0137] The tubular shape estimating unit 92c acquires amplitude information of the sound pressure levels at the sensors 31a to 31h from the first signal processing unit 92. The tubular shape estimating unit 92c estimates the tubular shape based on the amplitude information of the sound pressure levels at the sensors 31a to 31h.

[0138] The estimation model in the tubular body shape estimation unit 92c is learned by a machine learning method using a regression technique, which will be described below.

[0139] Fig. 14 is a cross-sectional view showing a simplified physical model according to the second embodiment. As shown in Fig. 1, the clarinet 1 originally has a tube shape with multiple tone holes at fixed positions, but the simplified physical model shown in Fig. 14 assumes a tube shape with a single tone hole 5 whose position in the longitudinal direction is variable, and the distance between the single tone hole 5 and the bell outlet 3a is variable.

[0140] 14 shows a control 6 for opening and closing a single tone hole 5. In the simple physical model, a state in which a single tone hole 5 is closed represents a state in which all tone holes are closed. Note that in the simple physical model, a state in which all tone holes are closed may also be represented as a state in which there are no tone holes.

[0141] For example, in the case of the pitch C4, as shown in Figure 5, of the open tone holes in an actual clarinet 1, the closest tone hole to the mouthpiece 2 is the 9th tone hole. To express a shape that can produce the pitch C4 using a simple physical model, it is sufficient to open only one hole at the position of the 9th tone hole and then adjust the position of the bell outlet.

[0142] The positions of the tone holes and the bell outlet section that can produce the pitch C4 are given as continuous-value parameters. In the second embodiment, combinations of the tone hole positions, the bell outlet positions, and the corresponding pitch C4 are prepared in advance as output data of a trained model using a neural network. The positions of the tone holes and the bell outlet section indicate characteristic shape parameters of the simple physical model. Hereinafter, the positions of the tone holes and the bell outlet section will be referred to as model shape parameters. Similarly, for other pitches, combinations of model shape parameters and corresponding pitches are prepared in advance as output data of a trained model using a neural network.

[0143] The neural network model machine learning device inputs standing wave sound pressure distribution information I1 to In as input data to the neural network model, and trains the neural network model to determine the correlation between the standing wave sound pressure distribution information I1 to In as input data and model shape parameters as output data.

[0144] More specifically, the machine learning device inputs the standing wave sound pressure distribution information I1 (L1(1), L2(1), ..., Lm(1)), ..., and In (L1(n), L2(n), ..., Lm(n)) constituting the training data to the input layer of the neural network as input data.

[0145] The machine learning device uses an evaluation function that compares the output data output as an inference result from the output layer of the neural network, i.e., the pitch model shape parameters corresponding to the standing wave sound pressure distribution information I1 to In, with the output data that constitutes the training data, i.e., the model shape parameters, and iteratively adjusts the weights associated with each synapse so that the value of the evaluation function becomes smaller. Here, adjusting the weights associated with each synapse is called backpropagation.

[0146] When the machine learning device determines that a predetermined training termination condition is satisfied, it terminates the machine learning and stores the neural network model at that point as a trained model in the storage device 80 A. The predetermined training termination condition is, for example, that the number of iterations of the series of training processes reaches a predetermined number, and that the value of the evaluation function becomes smaller than an allowable value.

[0147] C: Modifications The present invention is not limited to the above-described embodiment, and various modifications can be adopted within the scope of the present invention. Specific modification examples are exemplified below. Two or more embodiments arbitrarily selected from the following examples may be combined as appropriate within a range that does not contradict each other. In the modifications exemplified below, for elements whose actions and functions are equivalent to those of the above-described embodiment, the symbols used in the above explanation will be used and detailed explanations of each will be omitted as appropriate.

[0148] C1: First Modification In the above embodiment, the multiple sensors of the first sensor unit 31 were arranged at intervals from each other in the longitudinal direction of the pipe 3, but the multiple sensors may also be arranged at intervals from each other in the circumferential direction of the pipe 3, either in addition to or instead of this.

[0149] Ideally, the sound pressure distribution of standing waves generated inside the pipe 3 should be detected on the central axis C1 of the pipe 3. In a configuration in which multiple sensors are arranged circumferentially, even if the fixed positions of the multiple sensors are shifted for some reason, good detection of standing waves is possible by selectively using sensors closer to the central axis C1. Furthermore, even if there is an obstacle such as a protrusion inside the pipe 3, good detection of standing waves is possible by selectively using sensors that are less affected by the obstacle.

[0150] C2: Second Modification In the above embodiment, the first actuator 21 is an actuator that generates waves in the audible frequency band of 20 Hz to 20 kHz. However, it may be an actuator that generates ultrasonic waves, which are in a frequency band higher than the audible frequency band. For example, a piezoelectric element may be used as the actuator that generates ultrasonic waves. In this case, the sensors of the first sensor unit 31 may also be piezoelectric elements.

[0151] Ultrasonic waves are more likely to attenuate within the tube 3 than sound waves. Therefore, the ultrasonic waves attenuate from the position of the first actuator 21 toward the bell outlet 3a, and the component reflected at the bell outlet 3a decreases. Therefore, ultrasonic waves are easily observed as traveling waves within the tube 3. Ultrasonic waves are also easily attenuated at open sound holes. For this reason, multiple sensors are placed at intervals along the length of the tube 3 within the tube 3, and the intensity of the ultrasonic waves is detected by each sensor, thereby measuring the ultrasonic attenuation pattern.

[0152] Because the ultrasonic attenuation pattern is affected by the open / closed state of each tone hole, the tone hole open / close pattern can be estimated from the ultrasonic attenuation pattern. In this way, by using multiple sensors spaced apart along the length of the tube 3 within the tube 3, the ultrasonic attenuation pattern can be measured, thereby further improving the accuracy of tone hole open / close pattern estimation. Furthermore, because no audible sound is generated according to this modification, when the detection device 20 of this modification is applied to a wind instrument mute device 10, the performer can play the clarinet 1 with reduced sound generated around the clarinet 1.

[0153] Each sensor is preferably located as close as possible to each sound hole, and at least one of the sensors is preferably located at bell outlet 3a.

[0154] As described above, it is possible to observe the density distribution of ultrasonic standing waves by installing one or more reflectors at the outlet of the pipe 3 or inside the pipe 3. Therefore, the detection device according to this modified example may detect the density distribution of ultrasonic standing waves.

[0155] C3: Third Modification In the above embodiment, the first actuator 21 that vibrates the air inside the pipe 3 is used as the generator of the wave that propagates inside the pipe 3. However, an oscillator that generates electromagnetic waves and an antenna may also be used. In this case, a plurality of antennas may be used as the plurality of sensors of the first sensor unit 31.

[0156] Like ultrasonic waves, electromagnetic waves tend to attenuate within the pipe 3. Therefore, electromagnetic waves are easily observed as traveling waves within the pipe 3. Electromagnetic waves also tend to attenuate at open sound holes. For this reason, multiple sensors are arranged at intervals along the longitudinal direction of the pipe 3 within the pipe 3, and the electromagnetic wave intensity of the electromagnetic waves is detected by each sensor, thereby measuring the attenuation pattern of the electromagnetic waves.

[0157] Since the attenuation pattern of electromagnetic waves is affected by the open / closed state of each tone hole, the tone hole open / close pattern can be estimated from the attenuation pattern of electromagnetic waves. In this way, by using multiple sensors arranged at intervals along the longitudinal direction of the tube 3 inside the tube 3, the attenuation pattern of electromagnetic waves can be measured, thereby further improving the accuracy of estimating the tone hole open / close pattern. Furthermore, according to this modified example, no audible sound is generated, so when the detection device 20 of this modified example is applied to a wind instrument mute device 10, the performer can play the clarinet 1 with reduced sound generated around the clarinet 1. In this case, taking into consideration the effects of electromagnetic waves on the human body, the output of the generator is set to 100 mW / cm ２ The following is preferable (https: / / acoustics.jp / qanda / answer / 78.html).

[0158] As described above, it is possible to observe the density distribution of standing waves of electromagnetic waves by installing one or more reflectors at the outlet of or inside the pipe 3. Therefore, the detection device according to this modified example may detect the density distribution of standing waves of electromagnetic waves.

[0159] C4: Fourth Modification The amount of attenuation of traveling waves in the audible sound band may be observed, and the sound hole open / close pattern may be estimated based on the traveling wave information. The traveling wave information is information about the distribution of traveling waves. For example, traveling waves in the audible sound band are relatively resistant to attenuation, but tend to be more susceptible to attenuation after the open sound holes. Therefore, by installing sensors of the first sensor unit 31 immediately before each sound hole and immediately after each sound hole, and comparing the outputs of the installed sensors, the open / closed state of each sound hole can be determined.

[0160] For example, the output of the immediately preceding sensor may be compared with the output of the immediately following sensor, and when the output of the immediately following sensor is greater than the output of the immediately preceding sensor and the difference between the output of the immediately following sensor and the output of the immediately preceding sensor is equal to or greater than a certain value, it may be determined that the sound hole to be distinguished is open. The immediately preceding sensor is a sensor provided in a position immediately preceding the sound hole to be distinguished. The immediately following sensor is a sensor provided in a position immediately following the sound hole to be distinguished. The immediately preceding sensor and the immediately following sensor are arranged on either side of the sound hole to be distinguished.

[0161] With this configuration, the open / closed state of each sound hole is determined by comparing the output of the immediately preceding sensor with the output of the immediately following sensor, so there is no need to use a machine learning device to estimate the sound hole open / close pattern, making it possible to estimate the sound hole open / close pattern with a simpler configuration.

[0162] C5: Fifth Modification In the above embodiment, the second actuator 22 is used to vibrate the reed 4, and the second sensor unit 32 is used to detect the vibration of the reed 4. The second sensor unit 32 may be a contact vibration sensor such as a piezoelectric element or an acceleration sensor, or a non-contact vibration sensor such as a photocoupler, but is not limited to these. A strain sensor may also be used to detect the displacement of the reed 4. When a strain sensor is used, there is no need to vibrate the reed 4, and the second actuator 22 is not required.

[0163] C6: Sixth Modification In the first embodiment, a tube shape model is estimated based on the sound hole opening / closing pattern estimated by the pattern estimation unit 92a, and a radiated sound waveform is calculated by the physical model sound source unit 96 based on the estimated tube shape model. In the second embodiment, a radiated sound waveform is calculated by the physical model sound source unit 96 based on the tube shape estimated by the tube shape estimation unit 92c. In contrast, in the fifth modification, a pitch may be determined based on the estimated tube shape model or tube shape, and a recorded waveform corresponding to the pitch may be reproduced as the radiated sound waveform by a PCM (Pulse Code Modulation) sound source unit instead of the physical model sound source unit 96.

[0164] The wind instrument mute device according to this modification may include both the physical model sound source section 96 and the PCM sound source section. For example, the wind instrument mute device may be configured so that the performer can decide whether to calculate the radiated sound waveform using the physical model sound source section 96 or the PCM sound source section.

[0165] C7: Seventh Modification In the above embodiment, the clarinet 1 was used as an example of a wind instrument, but the wind instrument may also be another reed instrument such as an oboe, saxophone, bassoon, etc. In the case of a double reed instrument such as a bassoon, the model relating to the dynamic characteristics of a single reed used in the reed dynamic characteristics block 961 may be changed to a model relating to the dynamic characteristics of a double reed.

[0166] D: Supplementary Notes From the above-described exemplary embodiments, the following configurations can be understood, for example.

[0167] A detection device according to one aspect (aspect 1) of the present disclosure comprises a generating unit that generates waves that propagate within a tube of a wind instrument, and a first sensor unit having a plurality of sensors that detect the waves, wherein the generating unit and the first sensor unit are disposed within the tube, and each of the plurality of sensors is disposed at intervals from one another in the longitudinal direction of the tube.

[0168] According to this aspect, by using multiple sensors, the distribution of standing waves occurring within the tube can be detected, and the open / closed state of each of the multiple tone holes can be inferred from the distribution of standing waves. Because pitch is detected from the inferred open / closed states of the multiple tone holes, the detection result is less susceptible to the influence of resonance characteristics, such as the strength of resonance for a fingered pitch being weaker than the strength of resonance for other pitches. This reduces erroneous pitch detection. Furthermore, because each of the multiple sensors is spaced apart along the length of the tube, standing waves occurring within the tube can be detected more accurately. This allows for more accurate pitch detection.

[0169] A detection device according to one aspect (aspect 2) of the present disclosure further includes a second actuator that vibrates the reed of the wind instrument, and a second sensor unit having one or more sensors that detect the vibration of the reed, the second actuator and the second sensor unit being arranged on the reed.

[0170] According to this aspect, the resonance characteristics of the reed and the static opening of the reed can be detected, so that the playing state of the wind instrument can be detected more accurately.

[0171] A detection device according to one aspect (aspect 3) of the present disclosure further includes a third actuator that vibrates the air in the mouth of a player playing the wind instrument, and a third sensor unit having one or more sensors that detect the vibration of the air in the mouth, wherein the third actuator and the third sensor unit are disposed within the mouthpiece of the wind instrument.

[0172] According to this aspect, the shape of the inside of the mouth of the player can be detected, so that the playing state of the wind instrument can be detected more accurately.

[0173] A detection device according to one aspect (aspect 4) of the present disclosure further includes a fourth sensor unit having one or more sensors for detecting the breath pressure of a player playing the wind instrument, the fourth sensor unit being positioned within the mouthpiece of the wind instrument.

[0174] According to this aspect, the breath pressure of the player can be detected, so the playing state of the wind instrument can be detected more accurately.

[0175] In a detection device according to one aspect (aspect 5) of the present disclosure, the generating unit is a first actuator that generates sound waves by vibrating the air inside the tube of the wind instrument, and each of the plurality of sensors detects the sound waves.

[0176] According to this aspect, sound waves propagating within the pipe are detected by a plurality of sensors spaced apart along the length of the pipe. Because sound waves propagating within the pipe appear as standing waves, each sensor can detect a sound pressure level corresponding to the sound pressure distribution of the standing waves within the pipe. The sound pressure distribution of the standing waves can be estimated from the detection results of the sound pressure levels by each sensor.

[0177] A detection device according to one aspect (aspect 6) of the present disclosure comprises a first actuator that generates sound waves by vibrating the air inside a tube of a wind instrument, a first sensor unit having a plurality of sensors that detect the sound waves, and a first mounting unit that positions the first actuator and the first sensor unit inside the tube.

[0178] The sound waves propagating within the pipe appear as standing waves, and in order to detect the sound pressure distribution of the standing waves, it is preferable to detect the sound waves at a position close to the central axis of the pipe. According to this aspect, since multiple sensors are disposed within the pipe, even if the position of the first attachment part is shifted, it is possible to selectively use the sensors close to the central axis of the pipe, and it is possible to more accurately detect the sound pressure distribution of the standing waves.

[0179] In a detection device according to one aspect (aspect 7) of the present disclosure, each of the plurality of sensors is arranged at intervals from one another in the longitudinal direction of the pipe.

[0180] According to this aspect, the plurality of sensors are arranged at intervals from one another in the longitudinal direction of the pipe, so that the sound pressure distribution of standing waves occurring inside the pipe can be detected more accurately.

[0181] In a detection device according to one aspect (aspect 8) of the present disclosure, the wind instrument has one or more tone holes, and the multiple sensors are arranged in the longitudinal direction, sandwiching the tone holes.

[0182] According to this aspect, each of the multiple sensors is disposed immediately before and immediately after the sound hole, and the open / closed state of each sound hole is determined by comparing the output of the immediately before sensor with the output of the immediately after sensor, so there is no need to use a machine learning device to estimate the sound hole open / close pattern. This makes it possible to estimate the sound hole open / close pattern with a simpler configuration.

[0183] A mute device for a wind instrument according to one aspect (aspect 9) of the present disclosure comprises the detection device and a blocking device that is arranged between the mouthpiece of the wind instrument and the tube and blocks air from flowing into the tube from the mouthpiece.

[0184] In this embodiment, the space inside the tube is not connected to the space inside the blocking device that blocks air from flowing into the tube from the mouthpiece. Therefore, the breath blown into the mouthpiece by the player does not flow into the tube, muting the wind instrument.

[0185] A method for calculating a radiated sound waveform according to one aspect (aspect 10) of the present disclosure estimates a tube shape model representing the tube shape of a wind instrument based on standing wave information relating to the distribution of standing waves generated within the tube of the wind instrument, and calculates a waveform relating to the radiated sound from the wind instrument based on the estimated tube shape model and breath pressure information relating to the breath pressure of the performer playing the wind instrument.

[0186] According to this aspect, a tube shape model is determined from the detected standing waves. The tube shape model is a model that estimates pitch information and timbre information from the tube shape corresponding to the player's fingering on the wind instrument, so a sound waveform that more accurately reproduces the playing state of the wind instrument is calculated. Furthermore, according to this aspect, breath pressure information is generated based on the detected breath pressure of the player, so a sound waveform that matches the player's playing timing, breath pressure strength, etc. is calculated based on the breath pressure information.

[0187] In a method for calculating a radiated sound waveform according to one aspect (aspect 11) of the present disclosure, estimating the tube shape model includes inputting the standing wave information and outputting a sound hole opening / closing pattern indicating a combination of the open / closed states of each of a plurality of sound holes in the wind instrument, and determining the tube shape model based on the output sound hole opening / closing pattern.

[0188] According to this aspect, a tone hole opening and closing pattern is estimated from the detected standing waves, and a tube shape model is determined based on the estimated tone hole opening and closing pattern. The tube shape model is a model that estimates pitch information and timbre information from the tube shape corresponding to the fingering of the player on the wind instrument, so that a sound waveform that more accurately reproduces the playing state of the wind instrument can be calculated.

[0189] In the method for calculating a radiated sound waveform according to one aspect (aspect 12) of the present disclosure, the frequency of the standing wave is included in the audible sound band.

[0190] According to this aspect, waves in the audible sound band, i.e., sound waves, are less likely to attenuate within the tube and tend to appear as standing waves. Therefore, by generating sound waves within the tube and detecting the standing waves with multiple sensors, the sound hole opening and closing pattern can be estimated.

[0191] A method for calculating a radiated sound waveform according to one aspect (aspect 13) of the present disclosure estimates the tube shape model using a trained model in which the correlation between the standing wave information and a sound hole opening / closing pattern indicating the combination of the opening / closing states of each of the multiple sound holes in the wind instrument has been trained.

[0192] According to this aspect, when standing wave information is input to the trained model, a sound hole opening and closing pattern corresponding to the standing wave information is output, and a tube shape model is estimated from the sound hole opening and closing pattern output from the trained model. Therefore, there is no need to develop a complex program for estimating the sound hole opening and closing pattern from the standing wave information, and the tube shape model can be easily estimated.

[0193] A method for calculating a radiated sound waveform according to one aspect (aspect 14) of the present disclosure estimates a tube shape model representing the tube shape of a wind instrument based on traveling wave information regarding the distribution of traveling waves generated within the tube of the wind instrument, and calculates a waveform related to the radiated sound from the wind instrument based on the estimated tube shape model and breath pressure information related to the breath pressure of the performer playing the wind instrument.

[0194] According to this aspect, a tube shape model is determined from the detected traveling wave. The tube shape model is a model that estimates pitch information and timbre information from the tube shape corresponding to the player's fingering on the wind instrument, so a sound waveform that more accurately reproduces the performance state of the wind instrument is calculated. Furthermore, according to this aspect, breath pressure information is generated based on the detected breath pressure of the player, so a sound waveform that matches the player's performance timing, breath pressure strength, etc. is calculated based on the breath pressure information.

[0195] In a method for calculating a radiated sound waveform according to one aspect (aspect 15) of the present disclosure, estimating the tube shape model includes inputting the traveling wave information and outputting a sound hole opening / closing pattern indicating a combination of the open / closed states of each of a plurality of sound holes in the wind instrument, and determining the tube shape model based on the output sound hole opening / closing pattern.

[0196] According to this aspect, a tone hole opening and closing pattern is estimated from the detected traveling waves, and a tube shape model is determined based on the estimated tone hole opening and closing pattern. The tube shape model is a model that estimates pitch information and timbre information from the tube shape corresponding to the fingering of the player on the wind instrument, so that a sound waveform that more accurately reproduces the playing state of the wind instrument can be calculated.

[0197] In a method for calculating a radiated sound waveform according to one aspect (aspect 16) of the present disclosure, the frequency of the traveling wave is a frequency equal to or higher than the ultrasonic band.

[0198] According to this aspect, waves above the ultrasonic band are likely to attenuate within the pipe and appear as traveling waves. Therefore, by generating ultrasonic waves or electromagnetic waves within the pipe and detecting the traveling waves with multiple sensors, the sound hole opening and closing pattern can be estimated.

[0199] A method for calculating a radiated sound waveform according to one aspect (aspect 17) of the present disclosure estimates a change in the shape of the reed based on reed vibration information relating to the vibration of the reed of the wind instrument, and calculates a waveform relating to the radiated sound based on the estimated tube shape model, the estimated change in the shape of the reed, and the generated breath pressure information.

[0200] According to this aspect, the resonance characteristics and static opening of the reed are estimated from the detected vibration information, and therefore, whether or not the performer is performing a bend technique can be estimated, allowing a sound waveform to be calculated that more accurately reproduces the performance state of the wind instrument.

[0201] A method for calculating a radiated sound waveform according to one aspect (aspect 18) of the present disclosure estimates the shape of the oral cavity of a player playing a wind instrument having a manipulator based on vibration information relating to the vibration of air in the oral cavity, and calculates a waveform relating to the radiated sound from the wind instrument based on operation information indicating the player's operation of the manipulator, the estimated shape of the oral cavity, and breath pressure information relating to the player's breath pressure.

[0202] According to this aspect, the shape of the player's mouth is estimated based on the detected vibrations of the air in the mouth, and therefore the state of the player's embouchure and whether or not the player is using the flageo technique can be estimated, allowing for the calculation of a sound waveform that more accurately reproduces the playing state of the wind instrument.

[0203] A method for calculating a radiated sound waveform according to one aspect (aspect 19) of the present disclosure determines a pitch based on the operation information, the shape of the oral cavity, and the breath pressure information, and calculates a recorded waveform corresponding to the determined pitch as a waveform related to the radiated sound.

[0204] According to this aspect, the pitch is determined based on information about the player's operation of the wind instrument, the shape of the player's mouth, and information about the player's breath pressure. If a PCM sound source is prepared that records waveforms corresponding to each pitch, the radiated sound waveform can also be calculated based on the waveforms corresponding to each pitch recorded in the PCM sound source.

[0205] A method for calculating a radiated sound waveform according to one aspect (aspect 20) of the present disclosure outputs a waveform related to the radiated sound as an electrical signal.

[0206] According to this aspect, a sound waveform that reproduces the performance state of a wind instrument is output as an electrical signal, so that the performer can listen to the reproduced performance using a speaker or headphones.

[0207] According to the above aspect, a mute device for a wind instrument according to one aspect (aspect 21) of the present disclosure comprises a processing device having the detection device; a blocking device disposed between the mouthpiece and the tube of the wind instrument and blocking air flowing into the tube from the mouthpiece; a pattern estimation unit that estimates a sound hole opening / closing pattern indicating a combination of the open / closed states of each of the wind instrument's sound holes based on the detection results of the first sensor unit; a reed state estimation unit that estimates a change in the shape of the reed based on the detection results of the second sensor unit; an oral cavity shape estimation unit that estimates the shape of the oral cavity based on the detection results of the third sensor unit; and a breath pressure information generation unit that generates breath pressure information, which is information related to breath pressure, based on the detection results of the fourth sensor unit. The processing device estimates a tube shape model based on the sound hole opening / closing pattern estimated by the pattern estimation unit, and calculates a waveform related to the sound radiated from the wind instrument based on the estimated tube shape model, the estimated change in the shape of the reed, the estimated shape of the oral cavity, and the generated breath pressure information.

[0208] According to this aspect, a tone hole opening and closing pattern is estimated, and a tube shape model is determined based on the estimated tone hole opening and closing pattern. Also, according to this aspect, the resonance characteristics of the reed, the static opening of the reed, and the shape of the performer's mouth are estimated. Also, according to this aspect, a sound waveform is calculated based on breath pressure information, matching the performer's performance timing, breath pressure strength, and the like. Therefore, it is possible to estimate the state of the performer's embouchure and whether the performer is using a bend technique or a flageo technique, and therefore a sound waveform that more accurately reproduces the performance state of the wind instrument is calculated.

[0209] 1...clarinet, 2...mouthpiece, 3...tube, 4...reed, 5a, 5b, 5c, 5d, 5e...tone holes, 6a, 6b, 6c, 6d, 6e...operators, 10...mute device for wind instrument, 20, 20A...detection device, 21...first actuator, 22...second actuator, 23...third actuator, 31...first sensor unit, 32...second sensor unit, 33...third sensor unit, 34...fourth sensor unit, 41...first mounting unit, 42...second mounting unit, 43...third mounting unit, 44...fourth mounting unit, 50...isolating device, OCP...tone hole opening / closing pattern.

Claims

1. A detection device comprising: a generating unit that generates waves that propagate within a tube of a wind instrument; and a first sensor unit having a plurality of sensors that detect the waves, wherein the generating unit and the first sensor unit are disposed within the tube, and each of the plurality of sensors is disposed at intervals from one another in the longitudinal direction of the tube.

2. The detection device according to claim 1, further comprising: a second actuator that vibrates a reed of the wind instrument; and a second sensor unit having one or more sensors that detect vibration of the reed, wherein the second actuator and the second sensor unit are disposed on the reed.

3. The detection device described in claim 1, further comprising: a third actuator that vibrates the air in the mouth of a player playing the wind instrument; and a third sensor unit having one or more sensors that detect the vibration of the air in the mouth, wherein the third actuator and the third sensor unit are positioned within the mouthpiece of the wind instrument.

4. The detection device according to claim 1, further comprising a fourth sensor unit having one or more sensors for detecting the breath pressure of a player playing the wind instrument, the fourth sensor unit being disposed within a mouthpiece of the wind instrument.

5. The detection device according to claim 1, wherein the generating unit is a first actuator that generates sound waves by vibrating the air inside a tube of the wind instrument, and each of the plurality of sensors detects the sound waves.

6. A detection device comprising: a first actuator that generates sound waves by vibrating the air inside a tube of a wind instrument; a first sensor unit having a plurality of sensors that detect the sound waves; and a first mounting unit that positions the first actuator and the first sensor unit inside the tube.

7. The detection device of claim 6, wherein each of said plurality of sensors are spaced apart from one another along the length of said tube.

8. The detection device according to claim 1, wherein the wind instrument has one or more tone holes, and the plurality of sensors are arranged in the longitudinal direction on either side of the tone holes.

9. A muting device for a wind instrument, comprising: a detection device as claimed in any one of claims 1 to 8; and a blocking device provided between the mouthpiece and the tube of the wind instrument, for blocking air flowing from the mouthpiece into the tube.

10. A method for calculating a radiated sound waveform, comprising: estimating a tube shape model representing the tube shape of a wind instrument based on standing wave information relating to the distribution of standing waves generated inside the tube of the wind instrument; and calculating a waveform relating to the sound radiated from the wind instrument based on the estimated tube shape model and breath pressure information relating to the breath pressure of a player playing the wind instrument.

11. A method for calculating a radiated sound waveform as described in claim 10, wherein estimating the tube shape model includes inputting the standing wave information and outputting a sound hole opening / closing pattern indicating a combination of the open / closed states of each of a plurality of sound holes in the wind instrument, and determining the tube shape model based on the output sound hole opening / closing pattern.

12. A method for calculating a radiated sound waveform according to claim 10 or 11, wherein the frequency of the standing wave is included in the audible sound band.

13. A method for calculating a radiated sound waveform as described in claim 10, in which the tube shape model is estimated using a trained model in which the correlation between the standing wave information and a sound hole opening / closing pattern indicating a combination of the open / closed states of each of a plurality of sound holes in the wind instrument has been trained.

14. A method for calculating a radiated sound waveform, comprising: estimating a tube shape model representing the tube shape of a wind instrument based on traveling wave information relating to the distribution of traveling waves generated inside the tube of the wind instrument; and calculating a waveform relating to the sound radiated from the wind instrument based on the estimated tube shape model and breath pressure information relating to the breath pressure of a player playing the wind instrument.

15. A method for calculating a radiated sound waveform as described in claim 14, wherein estimating the tube shape model includes inputting the traveling wave information and outputting a sound hole opening / closing pattern indicating a combination of the open / closed states of each of a plurality of sound holes in the wind instrument, and determining the tube shape model based on the output sound hole opening / closing pattern.

16. A method for calculating a radiated sound waveform as described in claim 13 or 14, wherein the frequency of the traveling wave is a frequency equal to or higher than the ultrasonic band.

17. A method for calculating a radiated sound waveform as described in claim 10 or claim 14, comprising estimating a state of change in the shape of the reed based on reed vibration information relating to the vibration of the reed of the wind instrument, and calculating a waveform relating to the radiated sound based on the estimated tube shape model, the estimated state of change in the shape of the reed, and the generated breath pressure information.

18. A method for calculating a radiated sound waveform, comprising: estimating the shape of the oral cavity of a player playing a wind instrument having a control device based on vibration information related to vibrations of air in the oral cavity; and calculating a waveform related to sound radiated from the wind instrument based on operation information indicating the operation of the player on the control device, the estimated shape of the oral cavity, and breath pressure information related to the breath pressure of the player.

19. A method for calculating a radiated sound waveform as described in claim 18, further comprising the steps of: determining a pitch based on said operation information, the shape of the oral cavity, and said breath pressure information; and calculating a recorded waveform corresponding to said determined pitch as a waveform related to said radiated sound.

20. The method for calculating a radiated sound waveform according to claim 10 or 14, further comprising outputting a waveform relating to the radiated sound as an electrical signal.

Images