JP4940408B2 - Electric artificial larynx - Google Patents

Description

  The present invention relates to an electric artificial larynx.

  Healthy individuals generate sounds by vibrating their vocal cords when their exhalation passes through the larynx (glottis), and changing the frequency components of the sound with articulators such as the tongue, chin, and lips, and various utterances It is carried out. The sound generated by the vibration of the vocal cords is called “glottal sound source” because it is the sound that is the basis of the utterance.

  In contrast, a laryngectomy person who has removed the larynx is normally unable to utter by his own intention, but if the articulator remains, the artificially created sound will be generated. It is possible to speak in spite of incompleteness by generating it in the mouth instead of or sending it into the mouth.

  One of the devices that artificially create sound that can replace such glottal sound sources is the electric artificial larynx. The electric artificial larynx generates a sound that substitutes for the glottal sound source mechanically or electromechanically, and guides the sound into the oral cavity through vibration of the neck, etc., thereby supporting the vocalization by the laryngectomy To do. This type of electric artificial larynx is equipped with a diaphragm and a voice coil motor that vibrates the diaphragm to generate sound in the oral cavity or guides sound into the oral cavity. There are some which change the generated sound by changing the period and intensity of vibration.

For example, the electric artificial larynx disclosed in Patent Document 1 stores a sequence of a series of cycles forming a melody of a song in a storage medium, and vibrates the diaphragm at a cycle according to the sequence, thereby causing the larynx to vibrate. Supports singing by extractors. Moreover, the electric artificial larynx disclosed in Patent Document 2 is equipped with a sensor for detecting the pressure of exhaled air exhaled from the trachea of the laryngectomy, and the vibration of the diaphragm is detected according to the detection value of the sensor. The period is changed.
JP 07-433 International Publication No. WO99 / 012501

  By the way, while the voice of a healthy person is uttered with the intention of continuing to utter at the same loudness and pitch, the basic period and sound pressure of the waveform change minutely. On the other hand, in the electric artificial larynx disclosed in Patent Documents 1 and 2, when a sound to be generated is determined, a predetermined intensity and period of hitting set in advance according to the sound continues. There is a problem that it gives a natural and unnatural impression, and no matter who uses it, it can only produce a similar voice.

  On the other hand, in the paper “The role of“ waveform fluctuation ”in the natural nature of vowels” in the 47th issue of the Acoustical Society of Japan, vol. It is published along with several experimental results that the performance is improved.

  However, the electric artificial larynx used in the experiment of this paper is of a type that guides the sound of an arbitrary waveform generated by a speaker into the oral cavity by a pipe. In other words, a mechanism for minutely changing the basic period of the waveform and the sound pressure, which can be realized even with an electric artificial larynx of the type using a diaphragm and a drive unit that has a large restriction on the waveform that can be generated as sound, is still invented. It has not been.

  The present invention has been devised under such a background, and provides a mechanism that can further improve the naturalness of a voice generated by an electric artificial larynx using a diaphragm and a drive unit. With the goal. It is another object of the present invention to provide a mechanism that can add personality to a voice generated by an electric artificial larynx using a diaphragm and a drive unit.

  An electric artificial larynx, which is a preferred embodiment of the present invention, receives a vibration plate and a pulse signal, and generates vibration by applying vibration according to the period or / and duty ratio indicated by the pulse signal to the vibration plate. A drive unit for generating sound, an original sound generation instructing unit for instructing the generation of sound, and the generation of sound by the original sound generation instructing unit while the period or And / or a signal supply unit that supplies a series of pulse signals obtained by giving a minute change in the duty ratio to the drive unit.

  In this aspect, a plurality of period designation data respectively designating a period within a predetermined range on the low period side and the high period side of the waveform of the basic pulse signal, the temporal change of the fundamental period measured in real voice A memory that stores a set of a plurality of period designation data created based on the quantity, and the signal supply unit reads each period designation data stored in the memory, and the period designated by the period designation data These pulse signals may be sequentially supplied to the drive unit as a series of pulse signals obtained by applying the minute change.

  Also, a plurality of duty ratio specifying data respectively specifying duty ratios within a predetermined range on the low duty ratio side and the high duty ratio side of the waveform of the basic pulse signal, and the sound pressure time measured in real voice Further comprising a memory storing a plurality of sets of duty ratio designating data created based on the amount of change, and the signal supply unit reads each duty ratio designating data stored in the memory and designates the duty ratios thereof Each of the pulse signals having the duty ratio specified by the data may be sequentially supplied to the driving unit as a series of pulse signals obtained by giving the minute change.

  A plurality of period designating data for designating periods within a predetermined range on the low period side and the high period side of the waveform of the basic pulse signal, the power spectrum density being inversely proportional to the power of the frequency Further comprising a memory storing a set of a plurality of period specifying data, the signal supply unit reads each period specifying data stored in the memory, each of the pulse signals of the period specified by the period specifying data, You may make it supply sequentially to the said drive part as a series of pulse signals obtained by giving the said micro change.

  A plurality of duty ratio designating data for designating a duty ratio within a predetermined range on the low duty ratio side and the high duty ratio side of the waveform of the basic pulse signal, the power spectrum density being a power of the frequency The memory further stores a set of a plurality of duty ratio designation data that is inversely proportional, and the signal supply unit reads each duty ratio designation data stored in the memory, and the duty ratio designated by the duty ratio designation data You may make it supply each pulse signal used as a ratio to the said drive part sequentially as a series of pulse signals obtained by giving the said minute change.

  According to the present invention, the naturalness of a voice generated by an electric artificial larynx using a diaphragm and a drive unit can be further improved. Moreover, individuality can be imparted to the voice generated by the electric artificial larynx.

(Embodiment of the Invention)
Embodiments of the present invention will be described below with reference to the drawings. The electric artificial larynx according to the present embodiment artificially generates a sound as a voice in the oral cavity of the laryngectomy, or guides the artificially generated sound into the oral cavity. As a result, the frequency component of the sound is changed through the articulators such as the tongue, jaw, and lips, and is emitted from the user's mouth.

  FIG. 1 is a left side view of an electric artificial larynx 10 according to the present embodiment, and FIG. 2 is a top view thereof. The direction from left to right in FIGS. 1 and 2 is the front-rear direction, and the direction from top to bottom in FIG. 1 is the vertical direction. Further, the inside of the chain line in FIG. 1 shows the internal structure of the electric artificial larynx 10. As shown in both drawings, the electric artificial larynx 10 is connected to a taper-shaped voice coil motor holder 30 (corresponding to “driving unit” in the claims) in front of the housing 20, and the inside of the voice coil motor holder 30. The diaphragm 50 driven by the voice coil motor 31 is exposed from the forefront.

  The housing | casing 20 has comprised the substantially cylindrical shape of the size which can be accommodated in a user's palm. A hole (not shown) is passed through the upper portion of the housing 20 from the outer periphery toward the inner periphery, and the push button 21 is exposed from the hole. The push button 21 (corresponding to the “original sound generation instruction section” in the claims) is an operator that instructs the generation of sound, and is linked to an ON / OFF switch 22 (see FIG. 3) including a contact and a spring. . Each circuit element including the ON / OFF switch 22 is mounted inside the housing 20. Details of these circuit elements will be described later.

  Next, the voice coil motor 31 inside the voice coil motor holder 30 and the diaphragm 50 driven thereby will be described in detail. As shown in FIG. 1, a voice coil motor 31 includes a cylindrical magnet 32 having both ends magnetized to N and S poles, a cylindrical bobbin 33 surrounding the magnet 32 from the front and sides, and a bobbin. A coil 34 wound around the outer periphery of 33, and a yoke 35 that surrounds them from the rear and sides. The yoke 35 is open at the front, and a circular rubber 36 is stretched so as to close the opening. In the center of the circular rubber 36, a shaft 37 is penetrated in the front-rear direction and fixed. One end of the shaft 37 is fixed to the bobbin 33. The diaphragm 50 has a thin columnar shape and is made of urethane resin. The diaphragm 50 is fixed to the inside of the front edge of the voice coil motor holder 30 via a rubber material 38.

  The vibration coil 50 is driven by the voice coil motor 31 using Fleming's law. That is, when a pulse signal is supplied to the coil 34, the operation is repeated in which the other end of the shaft 37 pushed forward by the magnetic force applied to the bobbin 33 sliding in the yoke 35 strikes and separates the diaphragm 50. . As a result, the diaphragm 50 vibrates. When the supply of the pulse signal to the coil 34 is stopped, the shaft 37 is retracted by the action of the restoring force of the circular rubber 36.

  FIG. 3 is a block diagram showing a circuit configuration of the electric artificial larynx 10. In addition to the ON / OFF switch 22, the electric artificial larynx 10 according to the present embodiment includes a power source 23, a control unit 24 (corresponding to a part of “signal supply unit” in the claims), and a pulse signal output unit 25 (claims). And a storage unit 26 (corresponding to “memory” in the claims). The power supply 23, the pulse signal output unit 25, the control unit 24, and the storage unit 26 are built in the housing 20. Further, as described above, the push button 21 forming a part of the ON / OFF switch 22 is exposed from the hole in the upper part of the housing 20.

  The power source 23 is a battery, for example, and supplies power to the control unit 24 and the like. The pulse signal output unit 25 outputs a pulse signal for driving it to the voice coil motor 31 under the control of the control unit 24.

  The control unit 24 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and the like. Then, the control unit 24 controls the operation of the pulse signal output unit 25 in accordance with the stored contents of the storage unit 26, thereby making a minute change in the cycle and duty ratio of a series of pulse signals output from the pulse signal output unit 25. give.

  The storage content of the storage unit 26 and the control content of the pulse signal output unit 25 by the control unit 24 will be further described in detail. The storage unit 26 is a memory such as an EEPROM (Electrically Erasable and Programmable Read Only Memory). The storage unit 26 stores pulse signal generation data (corresponding to “period designation data” and “duty ratio designation data” in the claims) arranged in the time series of reading. This pulse signal generation data is a pair of high data that determines the time during which the pulse signal is in a high state and low data that determines the time during which the pulse signal is in a low state. Note that the pulse signal generated based on the pulse signal generation data is set such that the cycle and the duty ratio change minutely with the passage of time.

  When an ON signal is supplied from the ON / OFF switch 22 as the push button 21 is pressed, the control unit 24 reads the pulse signal generation data stored in the storage unit 26 one by one. Then, the control unit 24 instructs the pulse signal output unit 25 to set the output signal to a high state, and counts down the high data according to a predetermined clock signal. When the count value becomes zero, the pulse signal output unit 25 is instructed to set the output signal to the low state, and the low data is counted down according to a predetermined clock signal. When the count value becomes zero, the next pair of pulse signal generation data is read and the same operation is repeated. When all the pulse signal generation data stored in the storage unit 26 has been read, the same operation is repeated after returning to the top of the data. Thereby, even when the push button 21 is pressed for a time longer than the reproducible time by the pulse signal generation data, the pulse signal can be continuously generated.

  A series of pulse signal generation data to be stored in the storage unit 26 is generated, for example, according to the following procedure. First, vowels (eg, “A”, “I”, “U”, “E”, “O”) spoken naturally by healthy subjects or patients before laryngectomy are sampled at a predetermined cycle. Obtain sampling data. Note that the vowel to be sampled preferably includes, for example, 32 or more pitch waveforms (described later).

  FIG. 4A is a diagram showing a waveform of the vowel “A” of the sampling data obtained in this way. As shown in this figure, the vowel waveform is configured by repeating a basic waveform (pitch waveform) having substantially the same shape at a predetermined cycle (basic cycle). In order to generate the pulse signal generation data, first, information on each basic period of the basic waveform is obtained. Specifically, as shown in FIG. 4A, the basic periods a0, a1, a2, a3, a4, a5... Are obtained using one of the zero-cross points of the sound waveform as a guide. Alternatively, as shown in FIG. 4B, the basic periods b0, b1, b2, b3, b4, b5.

  Next, information about each sound pressure of the sampling data is obtained. Specifically, as shown in FIG. 5A, for each pitch waveform, values connecting the positive and negative peaks (peak-to-peak values) k0, k1, k2, k3, k4, k5,.・ Sequentially. Alternatively, as shown in FIG. 5B, the effective values R1, R2, R3, R4, R5,...

  And high data is calculated | required based on the information regarding the sound pressure calculated | required as mentioned above. More specifically, a value obtained by multiplying the peak-to-peak values k0, k1, k2, k3, k4, k5. Alternatively, a value obtained by multiplying the effective values R1, R2, R3, R4, R5,.

  Next, low data is obtained according to the basic period and high data. More specifically, a value obtained by multiplying the basic periods a0, a1, a2, a3, a4, a5... (Or the basic periods b0, b1, b2, b3, b4, b5...) By a predetermined coefficient. A value obtained by subtracting high data from a value obtained by dividing each of them by the period τ of the clock signal of the control unit 24 (the clock signal used when counting down as described above) is defined as low data.

  The pulse signal generation data obtained by the above processing has the following characteristics. That is, the time corresponding to the value obtained by adding the high data and the low data corresponds to the basic period of each pitch waveform of the sampling data described above. Each high data corresponds to the sound pressure of each pitch waveform. The pulse signal generation data generated in this way is stored in the storage area of the storage unit 26 in order. The control unit 24 reads the pulse signal generation data stored in the storage unit 26 in the order in which they are stored, sets the output of the pulse signal output unit 25 to the high state for the time corresponding to the high data, and corresponds to the low data. The output of the pulse signal output unit 25 is set to the low state for the time. As a result, the pulse signal output from the pulse signal output unit 25 has the basic period corresponding to the sampled waveform data and the sound pressure, and the minute change (the minute change of the basic period and the sound pressure) of the waveform data. A minute change).

  The shaft 37 of the voice coil motor 31 strikes the diaphragm 50 with a period and intensity according to the waveform of the pulse signal supplied from the pulse signal output unit 25. Therefore, when the user holds the casing 20 of the electric artificial larynx 10 in his / her palm, presses the diaphragm 50 against his / her neck and presses the push button 21, the vibration of the diaphragm 50 is transmitted from the neck to the oral cavity. Sound. The user can utter a desired phoneme from the mouth by changing the frequency component of the sound by means of articulators such as the tongue, jaw and lips. The basic period of sound generated in the oral cavity depends on the period at which the shaft 37 of the voice coil motor 31 strikes the diaphragm 50, and the sound pressure is the strength at which the shaft 37 strikes the diaphragm 50. Depends on.

  In two round frames in FIG. 3, an example of a waveform of a pulse signal output from the pulse signal output unit 25 to the voice coil motor 31 under the control of the control unit 24 and a voice coil motor 31 corresponding to the waveform. The waveform of the sound generated in the oral cavity by driving is described. The rising interval (cycle) of the waveform indicated by T1 of the pulse waveform corresponds to the time interval at which the shaft 37 of the voice coil motor 31 strikes the diaphragm 50. Therefore, the narrower the interval, the higher the frequency of vibration of the diaphragm 50 and the higher the fundamental frequency of sound generated in the oral cavity. Further, the time when the waveform is at the high level indicated by T2 of the waveform, that is, the duty ratio, corresponds to the strength with which the shaft 37 of the voice coil motor 31 strikes the diaphragm 50. Therefore, the higher the duty ratio, the greater the vibration intensity of the diaphragm 50 and the greater the sound pressure generated in the oral cavity.

  As described in the background section, the basic period and sound pressure of a healthy person change slightly while the voice of a healthy person is intended to continue speaking at the same loudness and pitch. ing.

  Therefore, the fundamental frequency of the pulse signal generated according to the series of pulse signal generation data stored in the storage unit 26 is the fundamental frequency of the voice (for example, about 125 Hz for a male voice and about 250 Hz for a female voice. ) In a predetermined range on the high frequency side and low frequency side. For this reason, the diaphragm 50 is hit according to the pulse signal generation data, and the fundamental frequency of the sound uttered in the oral cavity by the vibration also changes so as to have a minute change that is physiologically acceptable to the listener.

  Similarly, each duty ratio indicated by the series of pulse signal generation data stored in the storage unit 26 falls within a predetermined range on the high duty ratio side and the low duty ratio side of the basic duty ratio (for example, 50/100). scatter. For this reason, the sound pressure of the sound uttered in the oral cavity by the vibration of the diaphragm 50 according to the pulse signal generation data also changes so as to have a minute change that is physiologically acceptable to the listener.

  In the electric artificial larynx 10 according to the present embodiment described above, a pulse signal is supplied from the pulse signal output unit 25 to the voice coil motor 31 when the user gives an instruction to generate sound by pressing the push button 21. Is done. The shaft 37 of the voice coil motor 31 strikes the diaphragm 50 in accordance with this pulse signal, and the vibration generated by the impact is transmitted from the user's neck to the oral cavity, thereby generating sound in the oral cavity. Then, the period and duty ratio of the pulse signal that determines the basic period and sound pressure of the sound are not constant, but slightly change.

  Furthermore, the cycle and duty ratio of the pulse signal output from the pulse signal output unit 25 are changed according to a series of pulse signal generation data in the storage unit 26. The pulse signal generation data is acquired based on the sound waveform of the sound recorded from a healthy person or a patient before laryngectomy. Therefore, the basic period and sound pressure of the sound can be changed to have a minute change that is physiologically acceptable to the listener, not just a change, and a sound closer to that generated from the human vocal cords can be artificially generated. Can be produced. As a result, the voice is closer to the human voice.

(Other embodiments)
The present invention can be modified in various ways.

  Although the electric artificial larynx 10 according to the above embodiment can give a minute change within a predetermined range to the period and duty ratio of the pulse signal that determines the vibration of the diaphragm 50, the basic period and duty ratio itself are Only one kind is prepared. On the other hand, by storing a plurality of sets of pulse signal generation data in the storage unit 26, the basic period of sound generated in the oral cavity and the sound pressure itself may be switched. The electric artificial larynx 10 according to this modification individually records voices of healthy persons or the like having different voice fundamental frequencies and loudness, and pulse signal generation data obtained from each of the recorded voices. Each set (for example, a set for men and a set for women) is associated with the basic period and sound pressure index data and stored in the storage unit 26.

  In the electric artificial larynx 10 according to the above-described embodiment, a series of pulse signal generation data designating a basic period and a duty ratio is stored in the storage unit 26, and pulses having a basic period and a duty ratio designated by those data are stored. Signals are sequentially supplied from the pulse signal output unit 25 to the voice coil motor 31. On the other hand, an algorithm for generating a series of pulse signals having minute changes in the basic period and the duty ratio may be mounted, and the control unit 24 may be operated according to the algorithm. By adopting such a configuration, it is not necessary to previously store a set of pulse signal generation data in the storage unit 26.

  The electric artificial larynx 10 according to the above embodiment gives both the basic period and the sound pressure of the sound generated in the oral cavity by giving a minute change to the period and duty ratio of the pulse signal output from the pulse signal output unit 25. It is designed to change subtly. On the other hand, it is also possible to give a minute change to the period of the pulse signal while keeping the duty ratio constant so as to slightly change only the basic period of the sound generated in the oral cavity. Further, only the sound pressure of the sound generated in the oral cavity may be slightly changed by giving a slight change to the duty ratio with a constant period of the pulse signal.

The electric artificial larynx 10 according to the embodiment described above is based on the waveform of the pulse signal output from the pulse signal output unit 25 and the duty ratio, and the waveform of the sound recorded from a healthy person or a patient before laryngectomy. Although it is changed minutely by the data based on it, it may be changed minutely by using data created by calculation. For example, as shown in FIG. 6, the pulse signal output unit 25 is used by using basic cycle designation data and duty ratio designation data whose power spectral density has a characteristic of 1 / f ⁿ (n = about 0.5 to 1.5). Even if the period and the duty ratio of the pulse signal output from are changed, a sound close to that generated from the human vocal cords can be produced.

  The pulse signal generation data stored in the storage unit 26 of the electric artificial larynx 10 according to the above embodiment includes the high data for determining the time when the pulse signal is in the high state and the time when the pulse signal is in the low state. It is a pair of raw data to be determined. Then, the control unit 24 alternately repeats the high data countdown and the low data countdown indicated by the pulse signal generation data, and outputs the pulse signal so that the output signal is in the high state while counting down the high data. While instructing the unit 25, the pulse signal output unit 25 is instructed so that the output signal becomes low while the low data is being counted down. On the other hand, data indicating a reference value (for example, an average value) of high data and low data and a series of data indicating a deviation (small change width) from the reference value may be stored in the storage unit 26. Good. In general, since the width of the minute change is smaller than the absolute value of the reference value itself (approximately 1 to several percent), the amount of data stored in the storage unit 26 can be reduced by doing so. In addition, the amount of data can be reduced by storing data indicating a difference from the immediately preceding data instead of a deviation from the reference value.

  In the above embodiment, a series of pulse signal generation data generated based on sampling data of vowels sounded by a healthy person or the like is stored in the storage unit 26. On the other hand, when recording data of the user's own voice before the larynx is extracted remains, pulse signal generation data may be acquired from the recording data and stored in the storage unit 26. When the voice of the person can be used directly, a microphone, a sampling circuit, and an audio processing unit may be provided, and pulse signal generation data may be directly generated based on the sampled audio by driving them.

  The electric artificial larynx 10 according to the above embodiment is equipped with a voice coil motor 31 as a drive unit that vibrates the diaphragm 23. On the other hand, other elements such as a piezoelectric element and an electrostrictive element may be mounted instead of the voice coil motor 31.

  When the ON signal is supplied from the ON / OFF switch 22 as the push button 21 is pressed, the control unit 24 of the electric artificial larynx 10 according to the above embodiment receives the pulse signal generation data stored in the storage unit 26. Although reading is performed one by one from the beginning, this reading position may be changed each time the push button 21 is pressed. According to this modified example, each time the push button 21 is pressed, it is possible to generate a sound having a minute pattern with a different pattern.

  The electric artificial larynx 10 according to the above embodiment is equipped with only the push button 21 which is an operator for instructing the generation of sound, but another operator for instructing a change in sound pressure or period of the sound itself is provided. It may be installed.

It is a left view of the electric artificial larynx concerning embodiment. It is a top view of the electric artificial larynx according to the embodiment. It is a circuit block diagram of the electric artificial larynx concerning embodiment. It is a figure which shows the specific method of the fundamental frequency concerning embodiment. It is a figure which shows the specific technique of the level concerning embodiment. It is a figure which shows the frequency characteristic of the minute change given to the period and duty ratio of a pulse signal concerning embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Electric artificial larynx, 20 ... Housing | casing, 30 ... Voice coil motor holder, 31 ... Voice coil motor (equivalent to the "drive part" of a claim), 50 ... Diaphragm, 21 ... Push button ("Original sound generation instruction 22 ... ON / OFF switch, 32 ... magnet, 33 ... bobbin, 34 ... coil, 35 ... yoke, 36 ... circular rubber, 37 ... shaft, 23 ... power supply, 24 ... control unit ("signal supply") ... Pulse signal output unit (corresponding to part of “signal supply unit”), 26... Storage unit (corresponding to “memory”)

Claims (6)

In an electric artificial larynx that is used while the exposed diaphragm is pressed against the user's skin ,
A holding part for holding the diaphragm so as to be exposed;
A drive unit that receives supply of a pulse signal and generates a sound by applying vibration according to a period or / and a duty ratio indicated by the pulse signal to the diaphragm;
An original sound generation instructing unit for instructing sound generation;
While the generation of sound is instructed by the original sound generation instructing unit, a series of pulse signals obtained by giving a minute change in the period or / and the duty ratio to the waveform of the pulse signal that is the basis for generating the sound A signal supply unit for supplying to the drive unit;
An electric artificial larynx characterized by comprising A plurality of period designation data respectively designating a period within a predetermined range on the low period side and the high period side of the waveform of the basic pulse signal, and based on a temporal change amount of the fundamental period measured in the real voice Memory that stores a set of created multiple cycle specification data,
Further comprising
The signal supply unit is
Read each cycle designation data stored in the memory, and sequentially supply each pulse signal of the cycle designated by those cycle designation data to the drive unit as a series of pulse signals obtained by giving the minute change,
The electric artificial larynx according to claim 1. A plurality of duty ratio designating data for designating a duty ratio within a predetermined range on the low duty ratio side and the high duty ratio side of the waveform of the basic pulse signal, and the temporal change in sound pressure measured in the real voice Memory that stores multiple sets of duty ratio specification data created based on the quantity,
Further comprising
The signal supply unit is
Each duty ratio designation data stored in the memory is read, and each pulse signal having a duty ratio designated by the duty ratio designation data is sent to the drive unit as a series of pulse signals obtained by giving the minute change. Supply sequentially,
The electric artificial larynx according to claim 1. The signal supply unit changes the reading position from the memory each time generation of sound is instructed by the sound source generation instruction unit.
The electric artificial larynx according to claim 2 or 3, characterized in that. A diaphragm,
A drive unit that receives supply of a pulse signal and generates vibration by applying vibration according to a period or / and a duty ratio indicated by the pulse signal to the diaphragm;
An original sound generation instructing unit for instructing sound generation;
While the generation of sound is instructed by the original sound generation instructing unit, a series of pulse signals obtained by giving a minute change in the period or / and duty ratio to the waveform of the pulse signal that is the basis for generating the sound A signal supply unit for supplying to the drive unit;
A plurality of period specifying data respectively specifying periods within a predetermined range on the low period side and the high period side of the waveform of the basic pulse signal, and a plurality of such data whose power spectral density is inversely proportional to the power of the frequency A set of cycle specification data or a plurality of duty ratio specification data for specifying duty ratios within a predetermined range on the low duty ratio side and high duty ratio side of the basic pulse signal waveform, and the power spectral density A memory storing a plurality of sets of duty ratio designation data such that is inversely proportional to the power of the frequency,
With
The signal supply unit is
Each period designation data or each duty ratio designation data stored in the memory is read out, and each pulse signal having a period designated by the period designation data or a pulse signal having a duty ratio designated by the duty ratio designation data Are sequentially supplied to the drive unit as a series of pulse signals obtained by giving the minute change,
An electric artificial larynx characterized by this. The signal supply unit changes the reading position from the memory each time generation of sound is instructed by the sound source generation instruction unit.
The electric artificial larynx according to claim 5.

Images