JP2008131474A - Voice input device, its manufacturing method and information processing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a close-talking type voice input device capable of being miniaturized and achieving a highly accurate noise eliminating function, its manufacturing method, and an information processing system. <P>SOLUTION: The close-talking type voice input device includes a first microphone 10 provided with a first vibrating membrane 12, a second microphone 20 provided with a second vibrating membrane 22, and a differential signal generation part 30 for generating differential signals indicating a difference between a first voltage signal acquired in the first microphone 10 and a second voltage signal acquired in the second microphone 20. The first and second vibrating members 12 and 22 are disposed such that a noise-strength ratio indicating the ratio of the strength of noise components included in the differential signals to the strength of the noise components included in the first or second voltage signal is lower than a voice-strength ratio indicating the ratio of the strength of input voice components included in the differential signals to the strength of the input voice components included in the first or second voltage signal. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a voice input device, a manufacturing method thereof, and an information processing system.

  In a telephone call, voice recognition, voice recording, etc., it is preferable to pick up only the target voice (user voice). However, in a usage environment of the voice input device, there may be a sound other than the target voice such as background noise. Therefore, development of a voice input device having a function of removing noise has been advanced.

  As a technology to remove noise in usage environments where noise exists, the microphone has sharp directivity, or the arrival direction of the sound wave is identified using the difference in the arrival time of the sound wave, and the noise is removed by signal processing. The method is known.

In recent years, electronic devices have been downsized, and technology for downsizing a voice input device has become important.
JP 7-312638 A Japanese Patent Laid-Open No. 9-331377 JP 2001-186241 A

  In order to give the microphone a sharp directivity, it is necessary to arrange a large number of vibrating membranes, and miniaturization is difficult.

  In addition, in order to accurately detect the direction of arrival of sound waves using the difference in arrival times of sound waves, it is necessary to install a plurality of vibrating membranes at intervals of about one-several wavelengths of audible sound waves. Is difficult.

  An object of the present invention is to provide a close-talking voice input device having a function of removing a noise component, a manufacturing method thereof, and an information processing system.

(1) A voice input device according to the present invention includes:
A close-talking voice input device,
A first microphone having a first vibrating membrane;
A second microphone having a second vibrating membrane;
A difference signal generation unit that generates a difference signal indicating a difference between the first voltage signal acquired by the first microphone and the second voltage signal acquired by the second microphone;
Including
The first and second diaphragms have a noise intensity ratio indicating a ratio of the intensity of the noise component included in the differential signal to the intensity of the noise component included in the first or second voltage signal. The input audio component included in the difference signal is arranged to be smaller than an input audio intensity ratio indicating a ratio of the intensity of the input audio component included in the first or second voltage signal to the intensity of the input audio component.

  According to this voice input device, the first and second microphones (first and second diaphragms) are arranged so as to satisfy a predetermined condition. According to this, the difference signal indicating the difference between the first and second voltage signals acquired by the first and second microphones can be regarded as a signal indicating the input sound from which the noise component has been removed. Therefore, according to the present invention, it is possible to provide a voice input device capable of realizing a noise removal function with a simple configuration that only generates a differential signal.

  In this voice input device, the difference signal generation unit generates a difference signal without performing analysis processing (Fourier analysis processing or the like) on the first and second voltage signals. Therefore, it is possible to reduce the signal processing load of the differential signal generation unit, or to realize the differential signal generation unit with a very simple circuit.

  Therefore, according to the present invention, it is possible to provide a voice input device that can be miniaturized and can realize a highly accurate noise removal function.

  In this voice input device, the first and second diaphragms may be arranged such that the intensity ratio based on the phase difference component of the noise component is smaller than the intensity ratio based on the amplitude of the input voice component. Good.

(2) In this voice input device,
It further includes a base having a recess formed on the main surface,
The first vibrating membrane is installed on the bottom surface of the recess,
The second vibrating membrane may be installed on the main surface.

(3) In this voice input device,
The base may be installed such that an opening communicating with the concave portion is disposed closer to the model sound source of the input sound than the formation region of the second vibration film on the main surface.

  According to this voice input device, the phase shift of the input voice incident on the first and second diaphragms can be reduced. Therefore, a differential signal with less noise can be generated, and a voice input device having a highly accurate noise removal function can be provided.

(4) In this voice input device,
The recess may be shallower than a distance between the opening and the formation region of the second vibration film.

(5) In this voice input device,
The main surface further includes a base formed with a first recess and a second recess shallower than the first recess,
The first diaphragm is installed on a bottom surface of the first recess;
The second vibrating membrane may be installed on the bottom surface of the second recess.

(6) In this voice input device,
The base is installed such that the first opening communicating with the first recess is disposed closer to the model sound source of the input sound than the second opening communicating with the second recess. May be.

  According to this voice input device, the phase shift of the input voice incident on the first and second diaphragms can be reduced. Therefore, a differential signal with less noise can be generated, and a voice input device having a highly accurate noise removal function can be provided.

(7) In this voice input device,
The difference between the depths of the first and second recesses may be smaller than the distance between the first and second openings.

(8) In this voice input device,
The base may be installed such that the input sound arrives at the first and second diaphragms simultaneously.

  According to this, since a differential signal that does not include a phase shift of the input speech can be generated, it is possible to provide a speech input device having a highly accurate noise removal function.

(9) In this voice input device,
The first and second vibrating membranes may be arranged so that normals are parallel.

(10) In this voice input device,
The first and second vibrating membranes may be arranged such that the normal lines are not the same straight line.

(11) In this voice input device,
The first and second microphones may be configured as semiconductor devices.

  For example, the first and second microphones may be silicon microphones (Si microphones). The first and second microphones may be configured as one semiconductor substrate. At this time, the first and second microphones and the differential signal generation unit may be configured as one semiconductor substrate. The first and second microphones and the differential signal generation unit may be configured as so-called MEMS (Micro Electro Mechanical Systems).

(12) In this voice input device,
The distance between the centers of the first and second vibrating membranes may be 5.2 mm or less.

  Note that the first and second vibrating membranes may be arranged so that the normal lines are parallel to each other and the interval between the normal lines is 5.2 mm or less.

(13) An information processing system according to the present invention includes:
A first microphone having a first diaphragm, a second microphone having a second diaphragm, a first voltage signal obtained by the first microphone, and obtained by the second microphone. A close-talking voice input device including a difference signal generation unit that generates a difference signal indicating a difference from the second voltage signal;
Based on the difference signal, an analysis processing unit that performs analysis processing of voice information input to the voice input device;
Including
The voice input device includes:
The first and second diaphragms have a noise intensity ratio indicating a ratio of the intensity of the noise component included in the differential signal to the intensity of the noise component included in the first or second voltage signal, The input audio component included in the difference signal is arranged to be smaller than an input audio intensity ratio indicating a ratio of the intensity of the input audio component included in the first or second voltage signal to the intensity of the input audio component.

  According to this information processing system, voice information analysis processing is performed based on a differential signal acquired by a voice input device arranged such that the first and second diaphragms satisfy a predetermined condition. According to this voice input device, the difference signal becomes a signal indicating the voice component from which the noise component has been removed. Therefore, various information processing based on the input voice can be performed by analyzing the difference signal.

  The information processing system according to the present invention may be a system that performs voice recognition processing, voice authentication processing, or voice-based command generation processing.

(14) An information processing system according to the present invention includes:
A first microphone having a first diaphragm, a second microphone having a second diaphragm, a first voltage signal obtained by the first microphone, and a second microphone obtained by the second microphone. A close-talking type voice input device including a difference signal generation unit that generates a difference signal indicating a difference from the second voltage signal, and a communication processing unit;
Based on the difference signal, a host computer that performs analysis processing of voice information input to the voice input device;
Including
The voice input device includes:
The first and second diaphragms have a noise intensity ratio indicating a ratio of the intensity of the noise component included in the differential signal to the intensity of the noise component included in the first or second voltage signal, Arranged so as to be smaller than an input voice intensity ratio indicating a ratio of the intensity of the input voice component included in the difference signal to the intensity of the input voice component included in the first or second voltage signal;
The communication processing unit performs communication processing with the host computer via a network.

  According to this information processing system, voice information analysis processing is performed based on a differential signal acquired by a voice input device arranged such that the first and second diaphragms satisfy a predetermined condition. According to this voice input device, the difference signal becomes a signal indicating the voice component from which the noise component has been removed. Therefore, various information processing based on the input voice can be performed by analyzing the difference signal.

  The information processing system according to the present invention may be a system that performs voice recognition processing, voice authentication processing, or voice-based command generation processing.

(15) A method of manufacturing a voice input device according to the present invention includes:
A first microphone having a first diaphragm, a second microphone having a second diaphragm, a first voltage signal obtained by the first microphone, and a second microphone obtained by the second microphone. And a differential signal generation unit that generates a differential signal indicating a difference from the second voltage signal, and a method for manufacturing a close-talking voice input device having a function of removing a noise component,
The first or second value of Δr / λ indicating the ratio between the center-to-center distance Δr of the first and second vibrating membranes and the noise wavelength λ and the intensity of the noise component included in the difference signal. A procedure for preparing data indicating a correspondence relationship with a noise intensity ratio indicating a ratio to the intensity of the noise component included in the voltage signal;
A procedure for setting the value of Δr / λ based on the data;
A procedure for setting the center-to-center distance based on the set value of Δr / λ and the wavelength of the noise;
including.

  According to the present invention, it is possible to provide a method for manufacturing a voice input device that can be miniaturized and has a highly accurate noise removal function.

(16) In the method of manufacturing the voice input device,
In the procedure of setting the value of Δr / λ,
Based on the data, the noise intensity ratio indicates the ratio of the intensity of the input audio component included in the difference signal to the intensity of the input audio component included in the first or second voltage signal. The value of Δr / λ may be set so as to be smaller than the ratio.

(17) In the method of manufacturing the voice input device,
The input voice intensity ratio may be an intensity ratio based on an amplitude component of the input voice.

(18) In this method of manufacturing a voice input device,
The noise intensity ratio may be an intensity ratio based on a phase difference of the noise component.

  Embodiments to which the present invention is applied will be described below with reference to the drawings. However, the present invention is not limited to the following embodiments. Moreover, this invention shall include what combined the following content freely.

1. Configuration of Voice Input Device According to First Embodiment First, the configuration of a voice input device 1 according to an embodiment to which the present invention is applied will be described with reference to FIGS. Note that the voice input device 1 described below is a close-talking voice input device, for example, a voice communication device such as a mobile phone or a transceiver, or an information processing system using technology for analyzing input voice. (Voice authentication system, voice recognition system, command generation system, electronic dictionary, translator, voice input remote controller, etc.), recording equipment, amplifier system (loudspeaker), microphone system, etc. .

  The voice input device according to the present embodiment includes a first microphone 10 having a first diaphragm 12 and a second microphone 20 having a second diaphragm 22. Here, the microphone is an electroacoustic transducer that converts an acoustic signal into an electrical signal. The first and second microphones 10 and 20 may be converters that output vibrations of the first and second diaphragms 12 and 22 (diaphragm) as voltage signals, respectively.

  In the voice input device according to the present embodiment, the first microphone 10 generates a first voltage signal. The second microphone 20 generates a second voltage signal. That is, the voltage signals generated by the first and second microphones 10 and 20 may be referred to as first and second voltage signals, respectively.

  The mechanism of the first and second microphones 10 and 20 is not particularly limited. FIG. 2 shows a structure of a condenser microphone 100 as an example of a microphone applicable to the first and second microphones 10 and 20. The condenser microphone 100 has a vibration film 102. The vibrating membrane 102 is a membrane (thin film) that vibrates in response to sound waves, has conductivity, and forms one end of the electrode. The condenser microphone 100 also has an electrode 104. The electrode 104 is disposed to face the vibration film 102. Thereby, the vibrating membrane 102 and the electrode 104 form a capacitance. When a sound wave enters the condenser microphone 100, the vibration film 102 vibrates, the distance between the vibration film 102 and the electrode 104 changes, and the capacitance between the vibration film 102 and the electrode 104 changes. By outputting this change in capacitance as, for example, a change in voltage, a sound wave incident on the condenser microphone 100 can be converted into an electrical signal. In the capacitor microphone 100, the electrode 104 may have a structure that is not affected by sound waves. For example, the electrode 104 may have a mesh structure.

  However, the microphone applicable to the present invention is not limited to the condenser microphone, and any microphone that is already known can be applied. For example, as the first and second microphones 10 and 20, electrodynamic (dynamic), electromagnetic (magnetic), piezoelectric (crystal), etc. microphones may be applied.

  The first and second microphones 10 and 20 may be silicon microphones (Si microphones) in which the first and second vibrating membranes 12 and 22 are made of silicon. By using the silicon microphone, the first and second microphones 10 and 20 can be reduced in size and performance. At this time, the first and second microphones 10 and 20 may be configured as one integrated circuit device. That is, the first and second microphones 10 and 20 may be configured on one semiconductor substrate. At this time, a differential signal generation unit 30 to be described later may also be formed on the same semiconductor substrate. That is, the first and second microphones 10 and 20 may be configured as so-called MEMS (Micro Electro Mechanical Systems). However, the first microphone 10 and the second microphone 20 may be configured as separate silicon microphones.

  In the voice input device according to the present embodiment, as described later, a function of removing a noise component is realized by using a differential signal indicating a difference between the first and second voltage signals. In order to realize this function, the first and second microphones (first and second vibrating membranes 12 and 22) are arranged so as to satisfy certain restrictions. Although details of the constraints to be satisfied by the first and second vibrating membranes 12 and 22 will be described later, in the present embodiment, the first and second vibrating membranes 12 and 22 (the first and second microphones 10 and 10 20) is arranged such that the noise intensity ratio is smaller than the input voice intensity ratio. As a result, the difference signal can be regarded as a signal indicating the audio component from which the noise component has been removed. For example, the first and second vibrating membranes 12 and 22 may be arranged such that the center-to-center distance is 5.2 mm or less.

  In the voice input device according to the present embodiment, the directions of the first and second vibrating membranes 12 and 22 are not particularly limited. The 1st and 2nd vibrating membranes 12 and 22 may be arrange | positioned so that a normal line may become parallel. At this time, the 1st and 2nd vibrating membranes 12 and 22 may be arrange | positioned so that a normal line may not become the same straight line. For example, the first and second vibrating membranes 12 and 22 may be arranged on the surface of a base (not shown) (for example, a circuit board) with a space therebetween. Alternatively, the first and second vibrating membranes 12 and 22 may be arranged so as to be shifted in the normal direction. However, the 1st and 2nd vibrating membranes 12 and 22 may be arrange | positioned so that a normal line may not become parallel. The first and second vibrating membranes 12 and 22 may be arranged so that the normal lines are orthogonal to each other.

  The voice input device according to the present embodiment includes a difference signal generation unit 30. The difference signal generation unit 30 generates a difference signal indicating a difference (voltage difference) between the first voltage signal acquired by the first microphone 10 and the second voltage signal acquired by the second microphone 20. To do. The difference signal generation unit 30 performs a process of generating a difference signal indicating a difference between the first and second voltage signals without performing an analysis process such as Fourier analysis. The function of the difference signal generation unit 30 may be realized by a dedicated hardware circuit (difference signal generation circuit) or may be realized by signal processing by a CPU or the like.

  The voice input device according to the present embodiment may further include a signal amplifying unit that amplifies the differential signal. The difference signal generation unit 30 and the signal amplification unit may be realized by a single control circuit. However, the voice input device according to the present embodiment may be configured not to have a signal amplifying unit therein.

  FIG. 3 shows an example of a circuit that can realize the differential signal generation unit 30 and the signal amplification unit. According to the circuit shown in FIG. 3, the first and second voltage signals are received, and a signal obtained by amplifying the difference signal indicating the difference by 10 times is output. However, the circuit configuration for realizing the differential signal generation unit 30 and the signal amplification unit is not limited to this.

  The voice input device according to the present embodiment may include a housing 40. At this time, the outer shape of the voice input device may be configured by the housing 40. A basic posture may be set in the housing 40, thereby restricting the travel path of the input voice. The first and second vibrating membranes 12 and 22 may be formed on the surface of the housing 40. Alternatively, the first and second vibrating membranes 12 and 22 may be disposed inside the housing 40 so as to face an opening (sound entrance) formed in the housing 40. And the 1st and 2nd vibrating membranes 12 and 22 may be arrange | positioned so that the distance from a sound source (model sound source of incident sound) may differ. For example, as shown in FIG. 1, the basic posture of the housing 40 may be set so that the travel path of the input voice is along the surface of the housing 40. And the 1st and 2nd vibrating membranes 12 and 22 may be arrange | positioned along the advancing path | route of an input audio | voice. The vibration film disposed on the upstream side of the traveling path of the input voice may be the first vibration film 12 and the vibration film disposed on the downstream side may be the second vibration film 22.

  The voice input device according to the present embodiment may further include an arithmetic processing unit 50. The arithmetic processor 50 performs various arithmetic processes based on the difference signal generated by the difference signal generator 30. The arithmetic processing unit 50 may perform analysis processing on the difference signal. The arithmetic processing unit 50 may perform processing (so-called voice authentication processing) for identifying a person who has emitted the input voice by analyzing the difference signal. Or the arithmetic processing part 50 may perform the process (what is called speech recognition process) which specifies the content of an input audio | voice by analyzing a difference signal. The arithmetic processing unit 50 may perform processing for creating various commands based on the input voice. The arithmetic processing unit 50 may perform processing for amplifying the difference signal. The arithmetic processing unit 50 may control the operation of the communication processing unit 60 described later. Note that the arithmetic processing unit 50 may realize the above functions by signal processing using a CPU or a memory.

  The arithmetic processing unit 50 may be disposed inside the housing 40, but may be disposed outside the housing 40. When the arithmetic processing unit 50 is disposed outside the housing 40, the arithmetic processing unit 50 may acquire a difference signal via the communication processing unit 60 described later.

  The voice input device according to the present embodiment may further include a communication processing unit 60. The communication processing unit 60 controls communication between the voice input device and another terminal (such as a mobile phone terminal or a host computer). The communication processing unit 60 may have a function of transmitting a signal (difference signal) to another terminal via a network. The communication processing unit 60 may also have a function of receiving signals from other terminals via a network. For example, the host computer may analyze the differential signal acquired via the communication processing unit 60 and perform various information processing such as voice recognition processing, voice authentication processing, command generation processing, and data storage processing. Good. That is, the voice input device may constitute an information processing system in cooperation with other terminals. In other words, the voice input device may be regarded as an information input terminal that constructs an information processing system. However, the voice input device may not have the communication processing unit 60.

  The audio input device according to the present embodiment may further include a display device such as a display panel and an audio output device such as a speaker. In addition, the voice input device according to the present embodiment may further include an operation key for inputting operation information.

  The voice input device according to the present embodiment may have the above configuration. According to this voice input device, a signal (voltage signal) indicating the voice component from which the noise component has been removed is generated by a simple process that simply outputs the difference between the first and second voltage signals. Therefore, according to the present invention, it is possible to provide a voice input device that can be miniaturized and has an excellent noise removal function. The principle will be described later in detail.

2. Noise Removal Function Hereinafter, the voice removal principle adopted by the voice input device according to the present embodiment and the conditions for realizing this will be described.

(1) Noise removal principle First, the noise removal principle of the voice input device according to the present embodiment will be described.

  The sound wave attenuates as it travels through the medium, and the sound pressure (the intensity and amplitude of the sound wave) decreases. Since the sound pressure is inversely proportional to the distance from the sound source, the sound pressure P is related to the distance r from the sound source.

と表すことができる。なお、式（１）中、ｋは比例定数である。図４には、式（１）を表すグラフを示すが、本図からもわかるように、音圧（音波の振幅）は、音源に近い位置（グラフの左側）では急激に減衰し、音源から離れるほどなだらかに減衰する。本実施の形態に係る音声入力装置では、この減衰特性を利用して雑音成分を除去する。

It can be expressed as. In equation (1), k is a proportionality constant. FIG. 4 shows a graph representing the expression (1). As can be seen from FIG. 4, the sound pressure (the amplitude of the sound wave) is abruptly attenuated at a position close to the sound source (left side of the graph). Attenuates gently as you move away. In the voice input device according to the present embodiment, noise components are removed using this attenuation characteristic.

  That is, in the close-talking sound input device, the user emits sound from a position closer to the first and second microphones 10 and 20 (first and second vibrating membranes 12 and 22) than a noise source. . Therefore, the user's voice is greatly attenuated between the first and second vibrating membranes 12 and 22, and a difference appears in the intensity of the user voice included in the first and second voltage signals. On the other hand, the noise component is hardly attenuated between the first and second vibrating membranes 12 and 22 because the sound source is farther than the user's voice. Therefore, it can be considered that no difference appears in the intensity of noise included in the first and second voltage signals. From this, noise is eliminated if the difference between the first and second voltage signals is detected, and therefore, a voltage signal (difference signal) indicating only the user's voice component that does not include the noise component can be acquired. it can. That is, the differential signal can be regarded as a signal indicating the user's voice from which the noise component has been removed.

  However, the sound wave has a phase component. Therefore, in order to realize a highly reliable noise removal function, it is necessary to consider the phase difference between the speech component and the noise component included in the first and second voltage signals.

  Hereinafter, specific conditions to be satisfied by the voice input device in order to realize the noise removal function by generating the difference signal will be described.

(2) Specific conditions to be satisfied by the voice input device As described above, the voice input device according to the present embodiment does not include noise as the difference signal indicating the difference between the first and second voltage signals. It is considered as an input audio signal. According to this voice input device, if the noise component included in the differential signal is smaller than the noise component included in the first or second voltage signal, it can be evaluated that the noise removal function has been realized. Specifically, the noise intensity ratio indicating the ratio of the intensity of the noise component included in the difference signal to the intensity of the noise component included in the first or second voltage signal is equal to the intensity of the audio component included in the difference signal. If the ratio is smaller than the voice intensity ratio indicating the ratio of the voice component included in the first or second voltage signal, it can be evaluated that the noise removal function is realized.

  Hereinafter, specific conditions to be satisfied by the voice input device (first and second vibrating membranes 12 and 22) in order to realize this noise removal function will be described.

  First, the sound pressure of sound incident on the first and second microphones 10 and 20 (first and second vibrating membranes 12 and 22) will be examined. If the distance from the sound source of the input voice (user's voice) to the first diaphragm 12 is R, and the phase difference is ignored, the sound pressure of the input voice acquired by the first and second microphones 10 and 20 (Strength) P (S1) and P (S2) are

と表すことができる。

It can be expressed as.

と表される。 Therefore, a voice intensity ratio ρ (P) indicating the ratio of the intensity of the input voice component included in the difference signal to the intensity of the input voice component acquired by the first microphone 10 when the phase difference of the input voice is ignored. Is

It is expressed.

  Here, the voice input device according to the present embodiment is a close-talking voice input device, and Δr can be considered to be sufficiently smaller than R.

と変形することができる。 Therefore, the above equation (4) is

And can be transformed.

  That is, it can be seen that the voice intensity ratio when the phase difference of the input voice is ignored is expressed by the equation (A).

と表すことができる。なお、式中、αは位相差である。
このとき、音声強度比ρ（Ｓ）は、

と表される。式（７）を考慮すると、音声強度比ρ（Ｓ）の大きさは、

と表すことができる。 By the way, considering the phase difference of the input voice, the sound pressures Q (S1) and Q (S2) of the user voice are

It can be expressed as. In the formula, α is a phase difference.
At this time, the voice intensity ratio ρ (S) is

It is expressed. Considering equation (7), the magnitude of the voice intensity ratio ρ (S) is

It can be expressed as.

  By the way, in equation (8), the term sinωt−sin (ωt−α) indicates the intensity ratio of the phase component, and the Δr / R sinωt term indicates the intensity ratio of the amplitude component. Even if it is an input audio component, the phase difference component becomes noise with respect to the amplitude component. Therefore, in order to accurately extract the input audio (user's audio), the intensity ratio of the phase component is greater than the intensity ratio of the amplitude component. Must be sufficiently small. That is, sinωt−sin (ωt−α) and Δr / R sinωt are

の関係を満たしていることが必要である。

It is necessary to satisfy the relationship.

と表すことができるため、上述の式（Ｂ）は、

と表すことができる。 here,

Therefore, the above formula (B) can be expressed as

It can be expressed as.

を満たす必要があることがわかる。 Considering the amplitude component of Equation (10), the voice input device according to the present embodiment is

It turns out that it is necessary to satisfy.

と近似することができる。 As described above, since Δr can be regarded as sufficiently small as compared with R, sin (α / 2) can be regarded as sufficiently small.

And can be approximated.

と変形することができる。 Therefore, the formula (C) is

And can be transformed.

と表せば、式（Ｄ）は、

と変形することができる。 Also, the relationship between α and Δr, which are phase differences, is

The expression (D) can be expressed as

And can be transformed.

  That is, in this embodiment, in order to accurately extract the input voice (user's voice), it is necessary to manufacture the voice input device so as to satisfy the relationship represented by the equation (E).

  Next, the sound pressure of noise incident on the first and second microphones 10 and 20 (first and second vibrating membranes 12 and 22) will be examined.

と表すことができ、第１のマイクロフォン１０で取得される雑音成分の強度に対する、差分信号に含まれる雑音成分の強度の比率を示す雑音強度比ρ（Ｎ）は、

と表すことができる。 Assuming that the amplitudes of the noise components acquired by the first and second microphones are A and A ′, the sound pressures Q (N1) and Q (N2) of the noise considering the phase difference component are

The noise intensity ratio ρ (N) indicating the ratio of the intensity of the noise component included in the difference signal to the intensity of the noise component acquired by the first microphone 10 is expressed as follows:

It can be expressed as.

と変形することができる。 As described above, the amplitudes (intensities) of the noise components acquired by the first and second microphones are almost the same, and can be handled as A = A ′. Therefore, the above equation (15) is

And can be transformed.

と表すことができる。 And the magnitude of the noise intensity ratio is

It can be expressed as.

と変形することができる。 Here, considering the above equation (9), equation (17) is

And can be transformed.

と変形することができる。 And considering equation (11), equation (18) is

And can be transformed.

と表すことができる。なお、Δｒ／Ｒとは、式（Ａ）に示すように、入力音声（ユーザ音声）の振幅成分の強度比である。式（Ｆ）から、この音声入力装置では、雑音強度比が入力音声の強度比Δｒ／Ｒよりも小さくなることがわかる。 Here, referring to equation (D), the noise intensity ratio is

It can be expressed as. Note that Δr / R is the intensity ratio of the amplitude component of the input voice (user voice) as shown in Expression (A). From the expression (F), it can be seen that in this voice input device, the noise intensity ratio is smaller than the intensity ratio Δr / R of the input voice.

  From the above, according to the voice input device designed so that the intensity ratio of the phase component of the input voice is smaller than the intensity ratio of the amplitude component (see equation (B)), the noise intensity ratio is the input voice intensity ratio. (See formula (F)). In other words, according to the voice input device designed so that the noise intensity ratio is smaller than the input voice intensity ratio, a highly accurate noise removal function can be realized.

  That is, according to the present embodiment, the first and second vibrating membranes 12 and 22 (first and second microphones 10 and 20) are arranged so that the noise intensity ratio is smaller than the input voice intensity ratio. According to the voice input device, it is possible to realize a highly accurate noise removal function.

3. Method for Manufacturing Voice Input Device Hereinafter, a method for manufacturing the voice input device according to the present embodiment will be described. In the present embodiment, the value of Δr / λ indicating the ratio between the center-to-center distance Δr of the first and second vibrating membranes 12 and 22 and the noise wavelength λ and the noise intensity ratio (the intensity based on the phase component of the noise). The voice input device is manufactured using the data indicating the correspondence relationship with the ratio.

と表すことができる。 The intensity ratio based on the phase component of noise is expressed by the above-described equation (18). Therefore, the decibel value of the intensity ratio based on the phase component of noise is

It can be expressed as.

  Then, by assigning each value to α in Expression (20), it is possible to clarify the correspondence between the phase difference α and the intensity ratio based on the phase component of noise. FIG. 5 shows an example of data representing the correspondence between the phase difference and the intensity ratio when the horizontal axis is α / 2π and the vertical axis is the intensity ratio (decibel value) based on the phase component of noise. .

  The phase difference α can be expressed as a function of Δr / λ, which is the ratio of the distance Δr to the wavelength λ, as shown in Equation (12), and the horizontal axis in FIG. 5 is regarded as Δr / λ. Can do. That is, FIG. 5 can be said to be data indicating a correspondence relationship between the intensity ratio based on the phase component of noise and Δr / λ.

  In the present embodiment, a voice input device is manufactured using this data. FIG. 6 is a flowchart for explaining the procedure for manufacturing the voice input device using this data.

  First, data (see FIG. 5) indicating the correspondence between the noise intensity ratio (intensity ratio based on the noise phase component) and Δr / λ is prepared (step S10).

  Next, a noise intensity ratio is set according to the application (step S12). In the present embodiment, it is necessary to set the noise intensity ratio so that the noise intensity decreases. Therefore, in this step, the noise intensity ratio is set to 0 dB or less.

  Next, a value of Δr / λ corresponding to the noise intensity ratio is derived based on the data (step S14).

  Then, a condition to be satisfied by Δr is derived by substituting the wavelength of the main noise into λ (step S16).

  As a specific example, consider the case of manufacturing a voice input device in which the noise intensity is reduced by 20 dB in an environment where the main noise is 1 kHz and the wavelength is 0.347 m.

  First, as a necessary condition, a condition for the noise intensity ratio to be 0 dB or less is examined. Referring to FIG. 5, it can be seen that the value of Δr / λ may be 0.16 or less in order to make the noise intensity ratio 0 dB or less. That is, it can be seen that the value of Δr may be 55.46 mm or less, and this is a necessary condition for the voice input device.

  Next, a condition for reducing the intensity of 1 kHz noise by 20 dB will be considered. Referring to FIG. 5, it can be seen that the value of Δr / λ may be 0.015 in order to reduce the noise intensity by 20 dB. When λ = 0.347 m, it can be seen that this condition is satisfied when the value of Δr is 5.199 mm or less. That is, if Δr is set to about 5.2 mm or less, a close-talking voice input device having a noise removal function can be manufactured.

  Note that the voice input device according to the present embodiment is a close-talking voice input device, and the distance between the sound source of the user's voice and the first or second diaphragm 12, 22 is usually 5 cm or less. Further, the distance between the sound source of the user voice and the first and second vibrating membranes 12 and 22 can be controlled by the design of the housing 40. Therefore, the value of Δr / R, which is the intensity ratio of the input voice (user's voice), becomes larger than 0.1 (noise intensity ratio), and it can be seen that the noise removal function is realized.

  Normally, noise is not limited to a single frequency. However, since noise having a frequency lower than that of noise assumed as the main noise has a longer wavelength than the main noise, the value of Δr / λ becomes small and is removed by the voice input device. In addition, the sound wave decays faster as the frequency is higher. For this reason, noise having a higher frequency than the noise assumed as the main noise attenuates faster than the main noise, so that the influence on the voice input device can be ignored. Thus, the voice input device according to the present embodiment can exhibit an excellent noise removal function even in an environment where noise having a frequency different from that assumed as main noise exists.

  Further, in this embodiment, as can be seen from the equation (12), it is assumed that noise is incident from a straight line connecting the first and second vibrating membranes 12 and 22. This noise is a noise in which the apparent distance between the first and second vibrating membranes 12 and 22 is the largest, and is a noise in which the phase difference is the largest in an actual use environment. That is, the voice input device according to the present embodiment is configured to be able to remove the noise having the largest phase difference. Therefore, according to the voice input device according to the present embodiment, noise incident from all directions is removed.

4). Effects Hereinafter, effects achieved by the voice input device according to the present embodiment will be described.

  As described above, according to the voice input device according to the present embodiment, the noise component is generated only by generating a differential signal indicating the difference between the voltage signals acquired by the first and second microphones 10 and 20. The removed audio component can be acquired. That is, in this voice input device, a noise removal function can be realized without performing complicated analysis calculation processing. Therefore, according to the present embodiment, it is possible to provide a voice input device capable of realizing a highly accurate noise removal function with a simple configuration.

  Also, this voice input device realizes a noise removal function by making the noise intensity ratio based on the phase difference smaller than the input voice intensity ratio. Incidentally, the noise intensity ratio based on the phase difference varies depending on the arrangement direction of the first and second vibrating membranes 12 and 22 and the noise incident direction. That is, as the interval (apparent interval) between the first and second vibrating membranes 12 and 22 with respect to noise increases, the phase difference of noise increases and the noise intensity ratio based on the phase difference increases. By the way, in this Embodiment, the voice input device can remove the noise with which the apparent interval of the 1st and 2nd vibrating membranes 12 and 22 becomes the widest so that Formula (12) may show. It is configured as follows. In other words, in the present embodiment, the first and second vibrating membranes 12 and 22 are arranged so that incident noise can be removed so that the noise intensity ratio based on the phase difference is maximized. . Therefore, according to this voice input device, noise incident from all directions is removed. That is, according to the present invention, it is possible to provide a voice input device capable of removing noise incident from all directions.

  In addition, according to this voice input device, the user voice component incident on the voice input device after being reflected by a wall or the like can also be removed. Specifically, the sound source of the user voice reflected by a wall or the like can be regarded as being farther than the sound source of the normal user voice, and the energy is largely lost due to the reflection. In addition, the sound pressure is not greatly attenuated between the second vibrating membranes 12 and 22. Therefore, according to this voice input device, the user voice component that is incident on the voice input device after being reflected by a wall or the like is also removed (as a kind of noise).

  And if this audio | voice input apparatus is utilized, the signal which shows the input audio | voice which does not contain noise can be acquired. Therefore, by using this voice input device, highly accurate voice recognition, voice authentication, and command generation processing can be realized.

  Moreover, if this voice input device is applied to a microphone system, the user's voice output from the speaker is also removed as noise. Therefore, it is possible to provide a microphone system in which howling hardly occurs.

5. Voice Input Device According to Second Embodiment Next, a voice input device according to a second embodiment to which the present invention is applied will be described with reference to FIG.

  The voice input device according to the present embodiment includes a base 70. A recess 74 is formed in the main surface 72 of the base 70. In the voice input device according to the present embodiment, the first vibration film 12 (first microphone 10) is disposed on the bottom surface 75 of the recess 74, and the second vibration film 22 ( A second microphone 20) is arranged. The recess 74 may extend perpendicular to the main surface 72, and the bottom surface 75 of the recess 74 may be a surface parallel to the main surface 72. The bottom surface 75 may be a surface orthogonal to the recess 74. Further, the recess 74 may have the same outer shape as the first vibrating membrane 12.

  In the present embodiment, the recess 74 may be shallower than the distance between the region 76 and the opening 78. That is, if the depth of the recess 74 is d and the distance between the region 76 and the opening 78 is ΔG, the base portion 70 may satisfy d ≦ ΔG. The base 70 may satisfy 2d = ΔG. Note that ΔG may be 5.2 mm or less. Alternatively, the base 70 may be configured such that a linear distance connecting the centers of the first and second vibrating membranes 12 and 22 is 5.2 mm or less.

  The base 70 is installed such that the opening 78 communicating with the recess 74 is disposed at a position closer to the sound source of the input sound than the region 76 in the main surface 72 where the second vibration film 22 is disposed. The base portion 70 may be installed so that the input sound arrives at the first and second vibrating membranes 12 and 22 at the same time. For example, the base 70 may be installed such that the distance between the input sound source (model sound source) and the first diaphragm 12 is the same as the distance between the model sound source and the second diaphragm 22. . The base unit 70 may be installed in a housing in which a basic posture is set so as to satisfy the above-described conditions.

  According to the voice input device according to the present embodiment, it is possible to reduce a shift in incident time of input voices (user voices) incident on the first and second vibrating membranes 12 and 22. That is, since the difference signal can be generated so as not to include the phase difference component of the input sound, the amplitude component of the input sound can be accurately extracted.

  In addition, since the sound wave does not diffuse in the recess 74, the amplitude of the sound wave is hardly attenuated. Therefore, in this voice input device, the intensity (amplitude) of the input voice that vibrates the first diaphragm 12 can be regarded as the same as the intensity of the input voice in the opening 78. Therefore, even when the voice input device is configured so that the input voice reaches the first and second vibrating membranes 12 and 22 at the same time, the first and second vibrating membranes 12 and 22 are vibrated. A difference appears in the intensity of the input speech. Therefore, the input sound can be extracted by acquiring a differential signal indicating the difference between the first and second voltage signals.

  In summary, according to the voice input device, the amplitude component (difference signal) of the input voice can be acquired so as not to include noise based on the phase difference component of the input voice. Therefore, it is possible to realize a highly accurate noise removal function.

  In addition, since the resonant frequency of the recessed part 74 can be set high by making the depth of the recessed part 74 below (DELTA) G (5.2 mm or less), it can prevent that the resonant noise generate | occur | produces in the recessed part 74. .

  FIG. 8 shows a modification of the voice input device according to the present embodiment.

  The voice input device according to the present embodiment includes a base 80. A first recess 84 and a second recess 86 shallower than the first recess 84 are formed on the main surface 82 of the base 80. Δd, which is the difference between the depths of the first and second recesses 84, 86, is between the first opening 85 that communicates with the first recess 84 and the second opening 87 that communicates with the second recess 86. It may be smaller than ΔG which is the interval. The first vibration film 12 is disposed on the bottom surface of the first recess 84, and the second vibration film 22 is disposed on the bottom surface of the second recess 86.

  Even with this voice input device, the same effects as described above can be obtained, so that a highly accurate noise removal function can be realized.

  Finally, FIGS. 9 to 11 show a mobile phone 300, a microphone (microphone system) 400, and a remote controller 500 as examples of the voice input device according to the embodiment of the present invention. FIG. 12 is a schematic diagram of an information processing system 600 including a voice input device 602 as an information input terminal and a host computer 604.

  In addition, this invention is not limited to the above-mentioned embodiment, A various deformation | transformation is possible. The present invention includes configurations that are substantially the same as the configurations described in the embodiments (for example, configurations that have the same functions, methods, and results, or configurations that have the same objects and effects). In addition, the invention includes a configuration in which a non-essential part of the configuration described in the embodiment is replaced. In addition, the present invention includes a configuration that achieves the same effect as the configuration described in the embodiment or a configuration that can achieve the same object. Further, the invention includes a configuration in which a known technique is added to the configuration described in the embodiment.

The figure for demonstrating an audio | voice input apparatus. The figure for demonstrating an audio | voice input apparatus. The figure for demonstrating an audio | voice input apparatus. The figure for demonstrating an audio | voice input apparatus. The figure for demonstrating the method to manufacture an audio | voice input apparatus. The figure for demonstrating the method to manufacture an audio | voice input apparatus. The figure for demonstrating an audio | voice input apparatus. The figure for demonstrating an audio | voice input apparatus. The figure which shows the mobile telephone as an example of a voice input device. The figure which shows the microphone as an example of an audio | voice input apparatus. The figure which shows the remote controller as an example of an audio | voice input apparatus. 1 is a schematic diagram of an information processing system.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Voice input device, 10 ... 1st microphone, 12 ... 1st diaphragm, 20 ... 2nd microphone, 22 ... 2nd diaphragm, 30 ... Differential signal generation part, 40 ... Housing, 50 ... Arithmetic processing unit 60 ... Communication processing unit 70 ... Base part 72 ... Main surface 74 ... Recessed part 75 ... Bottom surface 76 ... Area 78 ... Opening 80 ... Base part 82 ... Main surface 84 ... First concave part 85 ... 1st opening, 86 ... 2nd recessed part, 87 ... 2nd opening, 100 ... Capacitor-type microphone, 102 ... Vibrating membrane, 104 ... Electrode, 300 ... Mobile phone, 400 ... Microphone, 500 ... Remote controller 600 Information processing system 602 Information input terminal 604 Host computer

Claims (18)

A close-talking voice input device,
A first microphone having a first vibrating membrane;
A second microphone having a second vibrating membrane;
A difference signal generation unit that generates a difference signal indicating a difference between the first voltage signal acquired by the first microphone and the second voltage signal acquired by the second microphone;
Including
The first and second diaphragms have a noise intensity ratio indicating a ratio of the intensity of the noise component included in the differential signal to the intensity of the noise component included in the first or second voltage signal. A voice input arranged so as to be smaller than an input voice intensity ratio indicating a ratio of the intensity of the input voice component included in the difference signal to the intensity of the input voice component included in the first or second voltage signal. apparatus. The voice input device according to claim 1,
It further includes a base having a recess formed on the main surface,
The first vibrating membrane is installed on the bottom surface of the recess,
The second vibrating membrane is a voice input device installed on the main surface. The voice input device according to claim 2,
An audio input device in which the base is installed such that an opening communicating with the concave portion is arranged closer to the model sound source of the input audio than an area where the second vibration film is formed on the main surface. The voice input device according to claim 2 or 3,
The voice input device, wherein the concave portion is shallower than a distance between the opening and the formation region of the second vibration film. The voice input device according to claim 1,
The main surface further includes a base formed with a first recess and a second recess shallower than the first recess,
The first diaphragm is installed on a bottom surface of the first recess;
The second input diaphragm is a voice input device installed on the bottom surface of the second recess. The voice input device according to claim 5,
The base is installed such that the first opening communicating with the first recess is disposed closer to the model sound source of the input sound than the second opening communicating with the second recess. Voice input device. The voice input device according to claim 5 or 6,
A voice input device in which a difference in depth between the first and second recesses is smaller than an interval between the first and second openings. The voice input device according to any one of claims 2 to 7,
The voice input device in which the base is installed so that the input voice arrives at the first and second diaphragms simultaneously. The voice input device according to any one of claims 1 to 8,
The voice input device in which the first and second vibrating membranes are arranged so that normals thereof are parallel to each other. The voice input device according to any one of claims 1 to 9,
The voice input device in which the first and second vibrating membranes are arranged so that the normal lines are not the same straight line. The voice input device according to any one of claims 1 to 10,
The first and second microphones are voice input devices configured as semiconductor devices. The voice input device according to any one of claims 1 to 11,
A voice input device in which a distance between centers of the first and second diaphragms is 5.2 mm or less. A first microphone having a first diaphragm, a second microphone having a second diaphragm, a first voltage signal obtained by the first microphone, and a second microphone obtained by the second microphone. A close-talking voice input device including a difference signal generation unit that generates a difference signal indicating a difference from the second voltage signal;
Based on the difference signal, an analysis processing unit that performs analysis processing of voice information input to the voice input device;
Including
The voice input device includes:
The first and second diaphragms have a noise intensity ratio indicating a ratio of the intensity of the noise component included in the differential signal to the intensity of the noise component included in the first or second voltage signal, Information processing arranged so as to be smaller than an input sound intensity ratio indicating a ratio of the intensity of the input sound component included in the difference signal to the intensity of the input sound component included in the first or second voltage signal. system. A first microphone having a first diaphragm, a second microphone having a second diaphragm, a first voltage signal obtained by the first microphone, and obtained by the second microphone. A close-talking type voice input device including a difference signal generation unit that generates a difference signal indicating a difference from the second voltage signal, and a communication processing unit;
Based on the difference signal, a host computer that performs analysis processing of voice information input to the voice input device;
Including
The voice input device includes:
The first and second diaphragms have a noise intensity ratio indicating a ratio of the intensity of the noise component included in the differential signal to the intensity of the noise component included in the first or second voltage signal, The input audio component included in the difference signal is arranged to be smaller than an input audio intensity ratio indicating a ratio of the intensity of the input audio component included in the first or second voltage signal to the intensity of the input audio component;
An information processing system in which the communication processing unit performs communication processing with the host computer via a network. A first microphone having a first diaphragm, a second microphone having a second diaphragm, a first voltage signal obtained by the first microphone, and a second microphone obtained by the second microphone. And a differential signal generation unit that generates a differential signal indicating a difference from the second voltage signal, and a method for manufacturing a close-talking voice input device having a function of removing a noise component,
The first or second value of Δr / λ indicating the ratio between the center-to-center distance Δr of the first and second vibrating membranes and the noise wavelength λ and the intensity of the noise component included in the difference signal. A procedure for preparing data indicating a correspondence relationship with a noise intensity ratio indicating a ratio to the intensity of the noise component included in the voltage signal;
A procedure for setting the value of Δr / λ based on the data;
A procedure for setting the center-to-center distance based on the set value of Δr / λ and the wavelength of the noise;
A method of manufacturing a voice input device including: In the manufacturing method of the voice input device according to claim 15,
In the procedure of setting the value of Δr / λ,
Based on the data, the noise intensity ratio indicates the ratio of the intensity of the input audio component included in the difference signal to the intensity of the input audio component included in the first or second voltage signal. A method for manufacturing a voice input device, wherein the value of Δr / λ is set to be smaller than the ratio. In the manufacturing method of the voice input device according to claim 15 or 16,
The voice input device manufacturing method, wherein the input voice intensity ratio is an intensity ratio based on an amplitude component of the input voice. In the manufacturing method of the voice input device according to any one of claims 15 to 17,
The method of manufacturing a voice input device, wherein the noise intensity ratio is an intensity ratio based on a phase difference of the noise component.

Images