JP2007002393A - Sound deadening helmet, vehicle system equipped with the same and method for deadening noise in helmet - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a sound-deadening helmet capable of suppressing or preventing deterioration of sound-deadening effect by suppressing influence of utterance of a helmet wearer. <P>SOLUTION: Sound in a helmet body 30 is detected by a microphone 102. Control sound for canceling noise detected by the microphone 102 is generated from a speaker 104. Control signals from a control signal producing circuit 106 are given to the speaker 104. Utterance-detecting microphones 1 and 2 for detecting utterance of a user P are provided in the helmet body 30. An utterance detector 3 detects presence or absence of utterance based on these output signals. In non-speaking in which utterance is not detected, gain of the control signal-producing circuit 106 is controlled by the gain control circuit 121 in non-speaking. In speaking in which utterance is detected, gain of the control signal-producing circuit 106 is controlled by a gain control circuit 122 in utterance. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a silencing helmet, a vehicle system including the same, and a noise silencing method in the helmet.

An active silencing helmet is known that actively removes noise by arranging microphones and speakers at the left and right ears and generating a control sound from the speakers that cancels noise detected by the microphones (Patent Document 1). . Thereby, the noise (mainly wind noise) felt by the driver can be reduced, and a comfortable driving environment can be realized.
JP 2000-54219 A

However, when the helmet wearer speaks, the sound is detected by the noise detection microphone. This prevents proper volume adjustment of the control sound. As a result, the silencing effect may be deteriorated.
An object of the present invention is to provide a silencing helmet capable of suppressing or preventing the deterioration of the silencing effect by suppressing the influence of the utterance of the wearer, a vehicle system including the same, and a noise silencing method in the helmet. .

  A noise reduction helmet according to an embodiment of the present invention includes a noise detection means for detecting noise in the helmet body, a sound generation means for generating a control sound for canceling the noise detected by the noise detection means, and the noise detection A control signal generating means for computing the output signal of the means to generate a control signal corresponding to the control sound and giving it to the sounding means, and an utterance detecting means for detecting the utterance of the wearer wearing the helmet body A non-speech gain adjusting unit that adjusts the gain of the control signal generating unit during a period when the utterance is not detected by the utterance detection unit, and the non-speech gain just before the utterance is detected by the utterance detection unit. A non-utterance gain storage means for storing a gain set by the adjustment means; and a period during which an utterance is detected by the utterance detection means. And a speech during gain adjustment means for adjusting the gain of said control signal generating means on the basis of the gain stored in the non-speech time gain storing means.

According to this configuration, the control signal corresponding to the noise detected by the noise detecting means is generated by the control signal generating means and given to the sound generation means, so that a control sound for canceling the noise is generated from the sound generation means. Thus, mute is performed.
On the other hand, the presence or absence of the speech of the wearer of the helmet body is detected by the speech detection means. During a period in which no utterance is detected, the gain of the control signal generating means is adjusted by the non-speech gain adjusting means. During the period when the utterance is detected, even if gain adjustment similar to that at the time of non-utterance is performed, proper mute cannot be expected, and there is a possibility of causing oscillation (howling). Therefore, during utterance, another gain adjustment control is performed by the utterance gain adjusting means.

More specifically, the gain immediately before the utterance is detected is stored in the non-utterance gain storage means. Using this, the utterance gain adjusting means adjusts the gain of the control signal generating means. Therefore, since gain adjustment is performed without being greatly affected by the utterance, undesired deterioration of the silencing effect can be suppressed or prevented.
The non-speech gain storage unit only needs to store the gain of the previous non-speech when the stored data is referred to by the utterance gain adjustment unit. That is, for example, the stored data of the non-speech gain storage unit is updated with the gain set by the non-speech gain adjustment unit as needed during the non-speech period, and in response to the detection of the utterance by the utterance detection unit, Storage data update means for stopping the update of the storage data may be provided. Thus, when an utterance is detected, the non-speech gain storage means stores the non-speech gain immediately before.

It is preferable that the noise detection means is disposed in the helmet body so as to be disposed at the user's ear when the helmet body is mounted. Thereby, since the mute is performed based on the sound close to the sound heard by the user, the accuracy of the mute can be improved.
The control signal generation means preferably generates the control signal by inverting the phase of the output signal of the noise detection means.

The utterance gain adjusting means preferably adjusts the gain of the control signal generating means in accordance with the output signal of the noise detecting means and the gain stored in the non-utterance gain storage means.
Further, the utterance gain adjusting means includes a spectrum calculating means for obtaining a frequency spectrum of noise in the helmet body based on an output signal of the noise detecting means, and immediately before the utterance is detected by the utterance detecting means. A non-speech noise spectrum storage unit for storing a frequency spectrum calculated by a spectrum calculation unit; a noise frequency spectrum calculated by the spectrum calculation unit during a period in which an utterance is detected by the utterance detection unit; The control based on the spectrum comparison means for comparing the frequency spectrum stored in the utterance noise spectrum storage means, the comparison result by the spectrum comparison means, and the gain stored in the non-speech gain storage means Gain control means for adjusting the gain of the signal generation means. It is preferable.

  According to this configuration, the frequency spectrum of noise immediately before the utterance is detected is stored in the non-utterance noise spectrum storage means. The frequency spectrum of noise during non-speaking is compared with the frequency spectrum of noise during speaking, and the gain of the control signal generating means is adjusted in consideration of the comparison result. Thereby, even during the utterance period, the gain of the control signal generating means can be adjusted according to the change of the noise situation, and an excellent silencing effect can be obtained.

The frequency spectrum consists of an amplitude spectrum and a phase spectrum.
The spectrum comparing means may compare the amplitude spectrum or its equivalent value for one or more specific frequencies between non-speaking and speaking. The comparison of the amplitude spectrum or its equivalent value is preferably performed in a frequency range (for example, a band of several tens of hertz) where the influence of speech is small. More specifically, it is preferable to perform the comparison in a frequency band in which a peak due to speech does not substantially appear in the speech amplitude spectrum.

The specific frequency preferably includes a plurality of different frequencies.
The gain control means obtains the gain of the control signal generation means by performing correction according to the comparison result by the spectrum comparison means on the gain stored in the non-speech gain storage means. There may be.
The non-speech gain adjusting means uses the output signal of the noise detecting means to obtain a sound pressure ratio acquiring means for acquiring a ratio of sound pressures in different frequency bands, and the sound pressure ratio acquired by the sound pressure ratio acquiring means. And a gain control means for controlling the gain of the control signal generation means so that the spectrum of the output signal of the noise detection means approaches a predetermined target spectrum. The “sound pressure” means an index of sound volume. Specifically, it may be an average value of the amplitude of the sound waveform, or a mean square of the amplitude. Good.

  According to this configuration, the ratio of sound pressures in different frequency bands is acquired using the output signal of the noise detection means, and the shape of the spectrum of the output signal of the noise detection means is optimized using the obtained sound pressure ratio. In this way, the gain of the control signal generating means is adjusted. Thereby, it is possible to cope with individual differences in the ear transfer function depending on the space shape between the wearer and the inner wall of the helmet, and a sufficient silencing effect can be obtained regardless of the wearer of the helmet.

  In the non-speech gain adjusting means having such a configuration, if the wearer's voice is mixed with the wind noise in the ear space formed between the helmet inner wall and the wearer's ear periphery, the frequency spectrum is disturbed. Cannot adjust the gain properly. Therefore, in the present invention, at the time of utterance, the gain of the control signal generating means is determined by a process different from that at the time of non-utterance. As a result, it is possible to perform excellent mute control corresponding to individual differences while suppressing the influence of speech.

  The sound pressure ratio acquisition means calculates a plurality of filters that filter output signals of the noise detection means with different frequency characteristics, and calculates sound pressures of a plurality of different frequency bands by processing the output signals of the plurality of filters. It is preferable to include sound pressure calculation means for performing sound pressure ratio calculation, and sound pressure ratio calculation means for calculating a sound pressure ratio as an index for control using sound pressures of a plurality of frequency bands calculated by the sound pressure calculation means. Thereby, it is possible to acquire the sound pressure ratio that is an index for control with a relatively simple circuit.

  The sound pressure ratio acquisition means includes a first acquisition means for acquiring a sound pressure in a resonance frequency band using the output signal of the noise detection means, and a reference that is a reference for comparison using the output signal of the noise detection means. Second acquisition means for acquiring sound pressure; and sound pressure ratio calculation means for calculating a ratio between the sound pressure in the resonance frequency band acquired by each of the first and second acquisition means and the reference sound pressure for comparison. It is good also as a structure to have. With this configuration, it is possible to acquire a sound pressure ratio that is an index for control relatively easily.

  The second acquisition means acquires, as the reference sound pressure, a sound pressure in a reference frequency range that is less affected by the noise reduction and the resonance frequency band that is silenced by the sound produced by the sound generation means. It is preferable. As a result, the sound pressure ratio calculated by the sound pressure ratio calculation means becomes a value corresponding to the increase or decrease of the sound pressure in the resonance frequency band. Therefore, the sound pressure level in the resonance frequency band can be adjusted by adjusting the gain of the control signal generating means, and thereby a desired spectrum can be obtained.

The reference frequency range may be the entire frequency range. That is, the sound pressure level in the entire frequency range may be used as a reference. This is because the sound pressure level in the entire frequency range is considered to hardly depend on the shape of the spectrum.
The non-speech gain adjusting means adjusts the gain of the control signal generating means so that the sound pressure ratio acquired by the sound pressure ratio acquiring means approaches a target sound pressure ratio corresponding to the predetermined target spectrum. Is preferred. Thereby, the spectrum of the output signal of the noise detecting means can be brought close to the target spectrum with simple control, and a good silencing effect can be obtained.

The non-speech gain adjusting means preferably sets the gain to zero when there is no noise. According to this configuration, since the gain is set to zero when there is no noise, the gain is not increased unnecessarily, and unnecessary silencing can be avoided.
The muffling helmet may further include first and second sound detection means for detecting sound generated by a wearer of the helmet body at different positions in the helmet body and outputting sound signals. The speech detection means includes a correlation calculation means for calculating a correlation value representing a correlation between voice signals output from the first and second voice detection means, and a correlation calculated by the correlation calculation means. An utterance determination unit that determines whether or not the wearer is speaking according to a sex value may be included.

  A silencing helmet according to another embodiment of the present invention includes a noise detection means for detecting noise in the helmet body, a sound generation means for generating a control sound for canceling the noise detected by the noise detection means, and the noise. A control signal generating means for calculating the output signal of the detecting means to generate a control signal corresponding to the control sound and giving it to the sounding means, and a sound uttered by the helmet body wearer in the helmet body First and second sound detection means for detecting sound signals at different positions and outputting sound signals, and correlation for calculating a correlation value representing the correlation between the sound signals output by the first and second sound detection means. Utterance determination means for determining whether or not the wearer wearing the helmet body is speaking according to the correlation value calculated by the correlation calculation means An utterance detecting means including: a non-speech gain adjusting means for adjusting a gain of the control signal generating means during a period in which no utterance is detected by the utterance detecting means; and a utterance being detected by the utterance detecting means. And an utterance gain adjusting means for adjusting the gain of the control signal generating means in a manner different from the non-utterance gain adjusting means.

  According to such a configuration, it is accurately determined whether or not the wearer is speaking by using the correlation value of the output signals of the first and second sound detection means arranged at different positions in the helmet body. can do. Then, by adjusting the gain of the control signal generation means in a different manner between when speaking and when not speaking, it is possible to suppress or prevent undesired deterioration of the silencing effect during speaking.

Three or more voice detection means may be provided, and it may be determined whether or not the wearer is speaking based on the correlation value of the output signals of the three or more voice detection means.
It is preferable that the first and second sound detection means are respectively disposed at substantially equal distances from the mouth of the wearer wearing the helmet body.
With this configuration, the utterance of the wearer can be detected more reliably. In other words, when the first and second voice detection means detect the voice uttered by the wearer at a position approximately equidistant from the wearer's mouth, a large correlation is recognized in the output signal corresponding to this. On the other hand, noise is also detected by the first and second sound detection means, but no significant correlation appears in the output signals of the first and second sound detection means with respect to the noise. In this way, it is possible to detect whether or not the wearer is speaking based on the correlation between the output signals of the first and second sound detection means.

In the first and second sound detection means, the correlation value calculated by the correlation calculation means is equal to or greater than a predetermined threshold when the helmet body wearer speaks, and when the wearer does not speak It is preferable that they are respectively disposed in the helmet body so as to be less than the threshold value.
With this configuration, it is possible to more reliably detect the utterance period based on the correlation between the output signals of the first and second sound detection means.

  According to the inventor's study, the first and second sound detection means are better placed in the helmet body than in the vicinity of the wearer's temple or in the ear, rather than in the vicinity of the wearer's temple. Has been obtained. That is, by arranging the first and second voice detection means in the vicinity of the wearer's mouth, the output signals of the first and second voice detection means (voice signals in the utterance frequency band) regarding the voice of the wearer. A large correlation was obtained between the output signals of the first and second sound detection means in any frequency range with respect to noise. Therefore, by arranging the first and second voice detection means at a position in the helmet body corresponding to the wearer's mouth, the wearer's speech and noise can be well separated, and the wearer's speech Presence or absence can be accurately detected.

  Although all the components of the muffling helmet may be attached to the helmet body, it is not always necessary. For example, the noise detection means and the sound generation means (further, if necessary, voice detection means for detecting a voice uttered by a helmet wearer) are attached to the helmet body, and at least a part of the other parts is different from the helmet body. You may comprise as an apparatus.

  A vehicle system according to an embodiment of the present invention includes a vehicle body and the muffling helmet described above, and includes at least the noise detection unit and the sound generation unit (and a voice detection unit that detects a voice uttered by a helmet wearer as necessary. ) Is provided in the helmet body of the muffling helmet, and at least a part of the remaining components excluding the noise detection means and sound generation means (and the sound detection means as necessary) of the muffling helmet is provided in the vehicle body. And further includes communication means for transmitting and receiving signals between the vehicle body side apparatus, the noise detection means and the sound generation means (and the sound detection means as required). With this configuration, a part of the component part of the muffler helmet can be arranged on the vehicle body.

  A vehicle system according to another embodiment of the present invention includes a vehicle body, the muffling helmet described above, sound information generating means for generating sound information provided in the vehicle body, and sound information generated by the sound information generating means. Transmission means for transmitting to the helmet body of the muffling helmet; and sound information sound generation means for sounding sound information provided in the helmet body and transmitted by the transmission means.

With this configuration, sound information from the sound information generating means on the vehicle body side can be provided to the helmet wearer in a state where the noise in the helmet body is muted regardless of individual differences. Thereby, the helmet wearer can hear the provided sound information comfortably.
Examples of the sound information generating means include a navigation system that generates voice guidance information, a mobile phone such as a mobile phone, a radio, and an audio device.

In addition to the wired communication means for connecting the sound information generating means and the helmet body with a cable, wireless communication means by infrared communication or radio wave communication can be applied to the transmission means.
A typical example of the sound information generating means is a speaker equipped in the helmet body. For example, a speaker installed in the helmet body may be shared by the sound information sounding means and the sounding means for mute, or each speaker may be used as a sound information sounding means and a sounding means for mute. May be provided.

  According to one embodiment of the present invention, a method for silencing a noise in a helmet includes a noise detecting step for detecting noise in the helmet body by a noise detecting unit, and a step for generating a control sound for canceling the detected noise from a sounding unit. A step of amplifying the control signal obtained by calculating the output signal of the noise detecting means by the amplifying means and supplying it to the sounding means; and an utterance detecting step of detecting the utterance of the wearer wearing the helmet body. A non-speech gain adjustment step for adjusting the gain of the amplification means during a period when the wearer's utterance is not detected; and a non-speech gain adjustment immediately before the wearer's utterance is detected. A speech gain adjustment step of adjusting the gain of the amplification means based on the gain set in the step.

With this method, when speaking, the gain of the amplification means is adjusted based on the gain at the previous non-speaking time, so that the noise in the helmet can be satisfactorily silenced without being greatly affected by the speaking.
In addition, in the helmet noise reduction method according to another embodiment of the present invention, a noise detection step for detecting noise in the helmet body by the noise detection means and a control sound for canceling the detected noise are generated from the sound generation means. A step of amplifying a control signal obtained by calculating an output signal of the noise detecting means by an amplifying means and supplying the sound to the sounding means; and a voice uttered by a wearer of the helmet body. And a second voice detecting means for calculating a correlation value representing a correlation between a voice detecting step for detecting at different positions in the helmet body and a voice signal output from the first and second voice detecting means. And an utterance determination step for determining whether or not the wearer wearing the helmet body is speaking according to the calculated correlation value. A non-speech gain adjusting step for adjusting the gain of the amplification means during a period when it is determined that the wearer is not speaking, and a non-speech period during which the wearer is determined to be speaking A speech gain adjustment step of adjusting the gain of the amplification means in a manner different from the time gain adjustment step.

  By this method, the utterance of the helmet wearer can be accurately detected. And since the mode of gain adjustment is made different between when speaking and when not speaking, the noise in the helmet can be silenced while minimizing the influence of the speaking.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
FIG. 1A is a block diagram showing a system configuration of an active silencing helmet according to an embodiment of the present invention, and FIG. 1B is an external view of the active silencing helmet.
This active silencing helmet 100 is a feedback type active silencing device applied to a helmet, and a microphone 102 as noise detecting means for detecting noise (wind noise, etc.) in the helmet and active detected noise. A control signal for generating a control sound for muting (secondary sound) by computing the output signal of the speaker 104 and the microphone 102 as a sounding means for generating a control sound (secondary sound) to be canceled A control signal generation circuit 106 (control signal generation means) to be generated and a gain adjustment circuit 120 for adjusting the control gain of the control signal generation circuit 106 are provided. The control signal generation circuit 106 includes a mute control filter circuit 107 and an amplifier 108 as an amplifying means capable of changing the gain, and a control signal output from the amplifier 108 is supplied to the speaker 104. The gain adjustment circuit 120 adjusts the gain (control gain) of the amplifier 108.

  The microphone 102 and the speaker 104 are disposed at appropriate predetermined positions in the outer shell 31 of the helmet body 30. Specifically, as shown in FIG. 1A, the microphone 102 and the speaker 104 are respectively arranged so that the user (wearer) P is positioned in the ear space of the user P with the helmet body 30 attached. Is done. In particular, the microphone 102 is disposed between the ear of the user P and the speaker 104 and positioned at the ear of the user P so that a sound close to the sound heard by the user P can be detected. The position of the microphone 102 becomes a muffle point. In FIG. 1B, reference numeral 33 indicates a cover, and reference numeral 35 indicates a shield.

  The silencing control filter circuit 107 takes in the instantaneous value of the sound waveform at a predetermined position (silence point) in the ear space in the helmet measured by the microphone 102 so that the sound pressure level at the silencing point in the ear space is minimized. Then, a control signal for driving the speaker 104 is calculated. When this control signal is amplified by the amplifier 108 and given to the speaker 104, the control sound is radiated from the speaker 104 to the ear space. Thereby, the noise of the user's P ear is offset. In this way, the control signal generation circuit 106 adaptively controls the output of the speaker 104 so that the sound is minimized at the position of the microphone 102.

FIG. 2 is a diagram illustrating a configuration of a control system targeted by the active silencing helmet according to the present embodiment. In FIG. 2, “P” is a frequency transfer function (ear transfer function) to be controlled, “C” is a frequency transfer function of the mute control filter circuit 107, “K” is a control gain (gain of the amplifier 108), and “y”. Is the output of the microphone 102, and "w" is noise (wind noise).
In this active silencing helmet 100, since the output y of the microphone 102 is close to the sound heard by the user P, the objective is to reduce this value. According to a known automatic control theory, when the frequency transfer function C of the mute control filter circuit 107 is designed to be a negative inverse function of the ear transfer function P (that is, C = −P ⁻¹ ), the control gain K is increased. For example, the microphone output y should approach zero (0). However, it is difficult to design the control filter C to be a negative inverse function of the ear transfer function P over the entire frequency range. When the control gain K is increased, the sound gradually increases at a certain frequency (resonance frequency), and finally It diverges (howling). As described above, the mute and the sound increase are two sides of each other, and the control gain K needs to be adjusted to an appropriate value in order to sufficiently exhibit the effect of the sound mute while appropriately suppressing the sound increase.

  For example, as a result of experiments, it has been found that there is an active silencing effect in the region of 100 to 400 Hz (silence range) that is the frequency range of the silencing target, and that the silencing volume increases as the control gain K increases. . On the other hand, it is known that there is a resonance frequency in the vicinity of 2.5 kHz, and that the volume increase increases as the control gain K increases. In this way, when the control amount (here, the microphone output y) is lowered in a certain frequency band, the phenomenon of increasing in another frequency band is generally known as a “water betting effect”.

  On the other hand, it is known that there are individual differences in the ear transfer function. Specifically, there is little individual difference in the phase of the ear transfer function, and there is also little individual difference in the shape of the gain (frequency dependence), but the magnitude of the gain is shifted overall by the user. Therefore, when the control gain is uniformly adjusted without considering individual differences, as described above, depending on the user, the control gain K may be too effective to diverge, or conversely, the control gain K may not be effective. Even if it does not diverge, the muffled sound may be smaller than expected. Therefore, even when there is an individual difference in the gain to be controlled, it is necessary to adjust the control gain K accordingly.

  Reference is again made to FIGS. 1A and 1B. Inside the helmet body 30, a pair of utterance detection microphones 1 and 2 (first and second sound detection means) are located near the mouth of the user P so that the distance from the mouth of the user P is equal. Is arranged. The active silencing helmet 100 further includes an utterance detector 3 (speech detection means) that receives the output signals of the utterance detection microphones 1 and 2 and detects whether the user P is speaking.

  On the other hand, the gain adjusting circuit 120 includes a non-speech gain adjusting circuit 121 (non-speech gain adjusting means) for adjusting the control gain when the user P is not speaking, and the user P speaks. The control gain generated by the utterance gain adjustment circuit 122 (speech gain adjustment means) for adjusting the control gain during the utterance, the control gain generated by the non-utterance gain adjustment circuit 121, and the control generated by the utterance gain adjustment circuit 122 A gain adjustment switching circuit 123 that selects one of the gains and inputs it to the amplifier 108, and a switching control circuit 124 that controls the gain adjustment switching circuit 123 based on the detection result of the speech detector 3 are provided. Yes.

  The switching control circuit 124 controls the gain adjustment switching circuit 123 so that the control gain output from the non-speaking gain adjustment circuit 121 is set in the amplifier 108 when the user P does not detect the utterance. Further, the switching control circuit 124 is a gain adjustment switching circuit so that the control gain output from the utterance gain adjustment circuit 122 is set in the amplifier 108 when the utterance of the user P is detected by the utterance detector 3. 123 is controlled.

  The utterance gain adjustment circuit 122 includes a gain storage memory 5 (non-utterance gain storage means) for storing the gain output by the non-utterance gain adjustment circuit 121, and a control gain output by the non-utterance gain adjustment circuit 121. The gain update changeover switch 6 for switching between updating the control gain stored in the gain storage memory 5 or prohibiting such update, and the detection result of the utterance detector 3 on the gain update changeover switch 6. And a gain update control circuit 7 that performs control according to the above. As described above, the gain update changeover switch 6 and the gain update control circuit 7 constitute storage data update means for updating the storage data of the gain storage memory 5.

  The speech gain adjustment circuit 122 further includes a fast Fourier transform circuit (FFT) 8 as a spectrum calculation means for generating a frequency spectrum of the output signal of the microphone 102, and a frequency output from the fast Fourier transform circuit 8. The spectrum storage memory 9 for storing the spectrum (non-speech noise spectrum storage means) and the state in which the update of the stored contents of the spectrum storage memory 9 by the frequency spectrum output from the fast Fourier transform circuit 8 is permitted and the update is prohibited. And a spectrum update control switch 11 that controls the spectrum update changeover switch 10 in accordance with the detection result of the utterance detector 3. The frequency spectrum consists of an amplitude spectrum and a phase spectrum.

  The utterance gain adjustment circuit 122 further compares the frequency spectrum calculated by the fast Fourier transform circuit 8 with the frequency spectrum stored in the spectrum storage memory 9, and calculates a comparison value corresponding to the comparison result. Control for generating a control gain to be given to the amplifier 108 at the time of speech based on the spectrum comparison circuit 12 (spectrum comparison means) and the comparison value generated by the spectrum comparison circuit 12 and the control gain stored in the gain storage memory 5 And a gain generation circuit 13 (gain control means). The control gain generated by the control gain generation circuit 13 is input to the gain adjustment switching circuit 123.

  The gain update control circuit 7 controls the gain update changeover switch 6 to an update permission state when not speaking. Thereby, the control gain generated by the non-speech gain adjustment circuit 121 is given to the gain storage memory 5, and the control gain stored in the gain storage memory 5 is updated every moment. On the other hand, when the utterance by the user P is detected by the utterance detector 3, the gain update control circuit 7 switches the gain update changeover switch 6 to the update prohibited state. As a result, the gain storage memory 5 holds the control gain output from the non-speech gain adjustment circuit 121 immediately before the utterance is detected.

  On the other hand, the spectrum update control circuit 11 controls the spectrum update changeover switch 10 to the update permission state when no utterance is detected by the utterance detector 3. Thereby, the frequency spectrum generated by the fast Fourier transform circuit 8 is given to the spectrum storage memory 9, and the stored contents of the spectrum storage memory 9 are updated every moment. On the other hand, during the utterance in which the utterance of the user P is detected by the utterance detector 3, the spectrum update control circuit 11 controls the spectrum update changeover switch 10 to an update prohibition state that prohibits updating of the stored contents of the spectrum storage memory 9. To do. As a result, the spectrum storage memory 9 stores the frequency spectrum at the time of non-utterance immediately before the utterance is detected.

  Therefore, at the time of utterance, the spectrum comparison circuit 12 compares the frequency spectrum at the time of the previous non-utterance with the frequency spectrum during the utterance, and generates a comparison value corresponding to the comparison result. Then, the control gain generation circuit 13 generates a control gain according to the comparison value and the control gain at the time of non-utterance immediately before stored in the gain storage memory 5. As a result, for example, with respect to the control gain at the time of non-speaking immediately before the user P's utterance is detected (control gain stored in the gain storage memory 5), the frequency spectrum at the time of non-speaking immediately before and the helmet body The control gain is generated by performing a correction according to the comparison result with the frequency spectrum representing the current noise situation in 30. The control gain generation circuit 13 is configured by a gain map that receives the control gain stored in the gain storage memory 5 and the comparison value generated by the spectrum comparison circuit 12 and outputs the control gain to be given to the amplifier 108. May be.

In this way, it is possible to perform gain control corresponding to an ever-changing noise situation while suppressing the influence due to the utterance of the user P, and obtain an excellent silencing effect regardless of whether or not the utterance is made. be able to.
FIG. 3 is a block diagram for explaining a configuration example of the utterance detector 3. The speech detector 3 includes a coherence computing unit 21 (correlation computing unit) that computes coherence (correlation) between the audio signals output from the speech detection microphones 1 and 2, and a coherence computed by the coherence computing unit 21. A threshold memory 22 that stores the corresponding threshold value, a coherence comparison unit 23 that compares the coherence calculated by the coherence calculation unit 21 with the threshold value stored in the threshold memory 22, and the coherence comparison unit 23 And an utterance determination unit 24 (utterance determination means) for determining the presence or absence of an utterance based on the comparison result of

  FIG. 4 is a diagram illustrating an example of coherence calculated by the coherence calculation unit 21. The horizontal axis of the graph of FIG. 4 is the frequency of the audio signal, and the vertical axis is the coherence value (where “m” represents millimeters (1/1000). The coherence obtained when the main noise detected by the microphone 102 is not a wind noise but is not uttered, and shows the coherence calculated when the user P is speaking. Results are shown.

  As understood from FIG. 4, at the time of speaking, the coherence takes a large value in the frequency band (700 Hz to 1.5 kHz) of the voice uttered by the user P and exceeds the threshold value TH. On the other hand, since the wind noise is random noise, there is little correlation between the noises detected by the speech detection microphones 1 and 2 and the coherence becomes a small value at any frequency.

  Therefore, the coherence comparison unit 23 (see FIG. 3) compares the coherence of the output signals of the speech detection microphones 1 and 2 with a predetermined threshold value TH. When the coherence exceeds the threshold value TH in the predetermined frequency band, the speech determination unit 24 (see FIG. 3) determines that the user P is speaking. If the coherence is equal to or less than the threshold value TH, the utterance determination unit 24 determines that the user P is not speaking when it is not speaking.

  FIGS. 5A and 5B are diagrams showing other arrangement examples of the speech detection microphones 1 and 2 in the helmet body. As shown in FIG. 5 (a), when the speech detection microphones 1 and 2 are respectively arranged at the left and right ears of the user P, the output signals of the speech detection microphones 1 and 2 at the time of non-speech (only wind noise). The coherence of is as in the example shown in FIG. On the other hand, as shown in FIG. 5 (b), when the speech detection microphones 1 and 2 are arranged near the temple of the user P (near the head), at the time of non-speech (only wind noise), Coherence as shown in FIG. 6 (b) is obtained.

In either case, a large coherence peak appears at any frequency, even when not speaking. Therefore, it is understood that the speech detection microphones 1 and 2 are appropriately placed in the mouth of the user P as shown in FIG. 1A.
In other words, the arrangement of the utterance detection microphones 1 and 2 is such that a peak of coherence exceeding a predetermined threshold appears in a predetermined frequency band at the time of utterance (a situation where there is an utterance sound and a wind noise), and at the time of non-utterance ( In a situation where only wind noise is present, it is preferable to select so that a peak of coherence exceeding the threshold value does not appear.

  FIG. 7 is a flowchart collectively showing the overall operation of the gain adjustment circuit 120. The non-speech gain adjustment circuit 121 obtains a control gain by processing suitable for non-speech based on the output signal of the microphone 102 (step S1). On the other hand, the fast Fourier transform circuit 8 calculates the frequency spectrum of the output signal of the microphone 102 (step S2).

The speech detector 3 determines whether or not the user P is speaking based on the output signals of the speech detection microphones 1 and 2 (step S3).
When the user P's speech is not detected, the control gain generated by the non-speech gain adjustment circuit 121 is stored in the gain storage memory 5 (step S4), and the fast Fourier transform circuit 8 is stored in the spectrum storage memory 9. The frequency spectrum to be calculated (the spectrum of only wind noise) is stored (step S5). Then, the control gain generated by the non-speech gain adjustment circuit 121 is given to the amplifier 108 (step S6).

  On the other hand, when the utterance of the user P is detected by the utterance detector 3 (YES in step S3), the spectrum comparison circuit 12 uses the frequency spectrum (wind noise spectrum) stored in the spectrum storage memory 9 and the high speed. The frequency spectrum calculated by the Fourier transform circuit 8 is compared, and a comparison value corresponding to the comparison result is calculated (step S7). This comparison value may be, for example, an amplitude or power ratio at any frequency (one or more).

  Then, the control gain generation circuit 13 obtains a control gain based on the control gain stored in the gain storage memory 5 and the comparison value (step S8). The control gain generation circuit 13 may determine the control gain based on a predetermined gain map, or is calculated by the spectrum comparison circuit 12 with respect to the control gain stored in the gain storage memory 5. A control gain corresponding to the time of non-speaking may be obtained by performing correction based on the comparison value.

The non-speech control gain obtained in this way is provided to the amplifier 108 (step S9).
Such processing is repeatedly executed at a predetermined cycle.
FIG. 8 is a diagram for explaining an example of a specific function of the spectrum comparison circuit 12, and each example of the amplitude spectrum of the utterance noise and the amplitude spectrum of the non-utterance noise is drawn in an overlapping manner. The wind noise component and the utterance peak overlap the amplitude spectrum of utterance noise. Further, in this example, the wind noise component is smaller than that during non-speaking.

Several specific frequencies f ₁ , f ₂ , and f ₃ in the frequency band where the influence of the utterance is small (frequency band where the utterance peak does not appear) are selected in advance. The amplitude spectrum of speech noise at each frequency is represented as L _N1 , L _N2 and L _N3, and the amplitude spectrum of non-speech noise is represented as L _S1 , L _S2 and L _S3 .
In this case, the spectrum comparison circuit 12 obtains the next comparison values C ₁ , C ₂ , and C ₃ by, for example, taking the ratio of the amplitude spectrum at each frequency.

C ₁ = L _N1 / L _S1
C ₂ = L _N2 / L _S2
C ₃ = L _N3 / L _S3
FIG. 9 is a block diagram illustrating an example in which the spectrum comparison circuit 12 is configured using a bandpass filter. In this example, a plurality of (three in this example) band-pass filters 61, 62, and 63 each having a pass band with a predetermined width centered on the frequencies f ₁ , f ₂ , and f ₃ are spectrum comparison circuits 12. Is provided. These band pass filters 61, 62, and 63 extract and output the waveform of each frequency component from each noise waveform during speech and during non-speech.

By taking an average of absolute values, an RMS (Root Mean Square), and other average values for each waveform, amplitude spectrum equivalent values L _N1 , L _S1 , L _N2 , L _S2 , L _N3 , and L _S3 are obtained. Based on these, the comparison values C ₁ = L _N1 / L _S1 , C ₂ = L _N2 / L _S2 , and C ₃ = L _N3 / L _S3 are obtained by taking the ratio of the amplitude spectrum equivalent values at each frequency. It is done.
Since the frequencies f ₁ , f ₂ , and f ₃ are included in the frequency band where the influence of the speech is small, the comparison values C ₁ , C ₂ , and C ₃ at these frequencies represent only the fluctuation component of the wind noise. . In addition, the comparison values C ₁ , C ₂ , and C ₃ represent the amplitude ratio of the wind noise during utterance with respect to the time when the utterance is not performed, and thus are appropriate indexes for knowing the fluctuation of the wind noise.

One comparison value may be used, but if a plurality of comparison values are taken, fluctuations in wind noise can be accurately determined.
Next, an example of processing for obtaining the control gain by the control gain generation circuit 13 will be described. For example, the control gain generation circuit 13 calculates the control gain K according to the following calculation formula.
(a) K = K _old W ₁ C ₁ (calculated from a single frequency)
(b) K = K _old (W ₁ C ₁ + W ₂ C ₂ + W ₃ C ₃ ) (when calculating from multiple frequencies)
Here, K _old is a control gain stored in the gain storage memory 5, and W ₁ , W ₂ and W ₃ are weighting coefficients corresponding to a plurality of frequencies f ₁ , f ₂ and f ₃ . That is, by multiplying the comparison value and the weighting coefficient, the stored control gain K _old is corrected, and the control gain K corresponding to the time of speech is obtained.

  At a certain speed range, the higher the overall wind noise level, the greater the proportion of low frequency noise in the wind noise, and the lower the overall wind noise level, the lower frequency noise rate. Get smaller. Therefore, as will be described later, it is effective to increase the control gain K as the proportion of low frequency noise increases, and to decrease the control gain K as the proportion of low frequency noise decreases. The control gain K calculated by the calculation formula (a) or (b) behaves in this way and is valid.

When the control gain K is calculated using a gain map, for example, the gain map illustrated in FIG. 10 is applied. The gain map obtains a comparison value C ₁ based on an amplitude spectrum at a single frequency f ₁ (or an equivalent value of an amplitude spectrum such as an average of absolute values of the bandpass filter output), and uses this to calculate a control gain K. It is for seeking.
Most cell values (control gain K) are the same numerical values (when W ₁ = 1.0) as the calculation result by the calculation formula (a), and therefore have the above-described validity. . However, the control gain K is set to “8” in the shaded cells. If the control gain K is increased too much, howling occurs. Therefore, in the shaded portion, consideration is given to preventing howling by forcing the control gain K to a peak (by limiting it to a predetermined upper limit value).

  FIG. 11 is a block diagram for explaining a configuration example of the non-speech gain adjusting circuit 121, and FIG. 12 is a diagram for explaining active mute control realized by the circuit of FIG. FIG. 11 shows a circuit configuration during non-speech. In this configuration example, the amplifier 108 is a digital amplifier, and the control signal generated by the mute control filter circuit 107 is an A / D converter 202. To the digital amplifier 108. The digital amplifier 108 amplifies the control signal generated by the mute control filter circuit 107 with the control gain K, and then outputs the amplified signal to the speaker 104 via the D / A converter 204. The speaker 104 receives the amplified control signal as an input and radiates a sound for muting to the ear space to cancel the noise.

The non-speech gain adjustment circuit 121 includes filters 206-1, 206-2, sound pressure calculation units 210-1, 210-2 (sound pressure calculation unit), and sound pressure ratio calculation unit 212 (sound pressure ratio calculation unit). And an adjusting unit 214 as a gain control means.
The output signal of the microphone 102 (sound pressure value at the microphone position) is input to the filters 206-1 and 206-2. The filter 206-1 is a filter that selectively passes a signal of a specific frequency band (for example, the center frequency fr), and the filter 206-2 is a filter of another specific frequency band (for example, the center frequency fw). It is a filter that selectively passes a signal. The frequency fr is a center frequency in a frequency band that is less affected by active silence (ANC), and the frequency fw is a resonance frequency (see FIG. 12). In FIG. 12, “N ₁ ” indicates a noise spectrum when the ANC is turned off in the helmet, and “N ₂ ” indicates a noise spectrum when the ANC is turned on in the helmet.

  The signals Xr and Xw that have passed through the filters 206-1 and 206-2 are respectively input to the sound pressure calculation units 210-1 and 210-2 via the A / D converters 208-1 and 208-2. . The sound pressure calculation unit 210-1 calculates an average value (sound pressure) Lr of the amplitude of the signal Xr that has passed through the filter 206-1, and the sound pressure calculation unit 210-2 receives the signal Xw that has passed through the filter 206-2. Average value (sound pressure) Lw is calculated (see FIG. 12). Thus, the filter 206-2 and the sound pressure calculation unit 210-2 correspond to first acquisition means for acquiring the sound pressure in the resonance frequency band, and the filter 206-1 and the sound pressure calculation unit 210-1 are compared. This corresponds to a second acquisition means for acquiring a reference sound pressure as a reference for the above. The calculation method of the average value of the filter passing signal may be, for example, an effective value (RMS) or an absolute value average.

The sound pressures Lr and Lw calculated by the sound pressure calculation units 210-1 and 210-2 are input to the sound pressure ratio calculation unit 212, respectively. The sound pressure ratio calculation unit 212 calculates the ratio J (= Lw / Lr) of the input sound pressures Lr and Lw.
Then, the sound pressure ratio J calculated by the sound pressure ratio calculation unit 212 is input to the adjustment unit 214. The adjustment unit 214 uses the input sound pressure ratio J to adjust the control gain K (gain of the digital amplifier 108) by integral control (I control).

Specifically, as shown in the following (Expression 1), a target value J _d (target sound pressure ratio) of the sound pressure ratio J is determined in advance, and a deviation (J _d − of the sound pressure ratio J with respect to the target value J _{d is} determined. J) is integrated over time, and the control gain K is determined by taking its absolute value.

In short, in the active silencing control realized by this digital circuit, the sound pressures Lr and Lw of the specific frequency bands fr and fw are calculated by filtering and sound pressure calculation processing, and the ratio J (= Lw / Lr) between them is used. To adjust the control gain K.

In the circuit of FIG. 11, as an example, for example, the digital amplifier 108, the sound pressure calculation units 210-1 and 210-2, the sound pressure ratio calculation unit 212, and the adjustment unit 214 are a digital signal processor (DSP). ) 216.
The frequency band of the sound pressure that constitutes the sound pressure ratio J that is an index for control is not limited to the frequency bands fr and fw described above. For example, the control gain K may be adjusted using the following (Formula 2) to (Formula 5).

Here, “L ₁ ” is an absolute value average of signals obtained by applying the high-pass filter (center frequency fw) to the output signal y of the microphone 102 (corresponding to the sound pressure level in the resonance frequency range). “ ₂ ” is an absolute value average (corresponding to the sound pressure level in the entire frequency range as the reference frequency range) when the output signal y of the microphone 102 is passed as it is. The ratio J (= L ₁ / L ₂ ) between the _two means the ratio of high-frequency components (including resonance frequency) in the entire wind noise. J _d is the optimum value (target value) of the sound pressure ratio J, and k _p is an appropriate constant. Further, F ₁ in (Expression 2) represents an operator corresponding to the high-pass filter. That is, the result of filtering the signal y (t) with the high-pass filter is expressed as “F ₁ y (t)”.

There are two functions in (Equation 1) and (Equation 5) that define the control gain. The first function is to sound pressure ratio J adjusts the control gain K so as to approach the target value J _d, the second function is to keep the control gain K to zero (0) value greater than or equal to . The first function is realized by integral control (I control), and the second function is realized by absolute value calculation in (Expression 1) and (Expression 5). The reason for performing the integration control, during the sound pressure ratio J and the target value J _d, in order to prevent the proportional control (P control) and steady state error which can not be differentiated control (D control) in removal occurs is there. Therefore, at least integral control should be included, and proportional control and / or differential control can be combined with integral control.

The reason for calculating the absolute value is to prevent the malfunction (divergence phenomenon) due to the possibility that the value of K may be negative when the control gain K is adjusted digitally.
More specifically, since it is known in advance that the sound pressure ratio J monotonously increases with respect to the control gain K, the value obtained by integrating (J _d −J) over time is defined as the gain K. when the sound pressure ratio J is smaller than the target value J _d, the gain K gradually increases, also increases the sound pressure ratio J simultaneously. Conversely, when the sound pressure ratio J is larger than the target value J _d, the gain K gradually decreases, also decreases the sound pressure ratio J simultaneously. Thus, the sound pressure ratio J converges to the target value _Jd , and the spectrum of the output signal of the microphone 102 is optimized.

On the other hand, if the control gain K is small and takes a negative value, oscillation (howling) occurs. To prevent this, the absolute value of the integral value is set as the control gain K. As a result, the lower limit value of the control gain K can be limited to “0”.
Therefore, the control gain K can be adjusted to an optimum value by the integral control of (Expression 1) or (Expression 5).

13A, 13B, and 13C are diagrams for explaining the effect of the active mute control by the function of the non-speech gain adjustment circuit 121 as described above. FIG. 13A shows a state in which the wind noise is large. FIG. 13B is a diagram illustrating an effect in a state where the wind noise is low, and FIG. 13C is a diagram illustrating an effect in a state where there is no wind noise.
The active mute control by the function of the non-speech gain adjustment circuit 121 can eliminate individual differences in the ear transfer function, and can enable optimum control independent of the magnitude of wind noise. In other words, this active silencing control is aimed at converging the spectrum shape of noise (wind noise) to a certain target spectrum shape. As an example of the target spectrum shape, for example, the sound pressure L ₂ is set to 10 times the sound pressure L ₁ (+20 dB at the sound pressure level), that is, the target value J _d in Equation 5 is set to 1/10. become. Then, the control gain K is adjusted by (Equation 5) such that the current ratio J (= L ₁ / L ₂ ) of the sound pressures L ₁ and L ₂ matches the target value J _d . That is, since the ratio of sound pressures in different frequency bands is used, control independent of the absolute value of the microphone output signal can be performed.

More specifically, when the sound pressure L _{1 in the} resonance frequency band is increased by the active silence (ANC), the user P recognizes the “noisiness” by comparison with the sound pressure L ₁ before ANC. In other words, if the sound pressure in the frequency band f ₃ that is less influenced by ANC is L ₃ , the sound pressure L ₃ after ANC is also recognized by comparing it with the sound pressure L ₁ after ANC. This is because the sound pressure L ₃ is hardly changed by ANC (it is influenced by the overall noise level). As a result, the sound in the muffler range (frequency range subject to mute by ANC; the frequency range of the main wind noise) is moderately muted, and the sound in the resonance frequency band is not so “noisy”. There should be a relative relationship (sound pressure ratio after ANC) between the sound pressures L ₃ and L ₁ . Such modest relative relationship is not unique to the sound pressure L ₃ of a specific frequency band L _1, exists between all frequencies. Therefore, in general, it can be said that there is an appropriate spectrum shape in terms of comfort as the spectrum shape after ANC.

Since the sound pressure L ₂ representing the sound pressure level in the entire frequency range can be used as a value that hardly changes before and after ANC, the sound pressure ratio J = L ₁ / L ₂ is a spectrum shape that depends on the control gain K. Is a value representing Therefore, by adjusting the control gain K to approach the sound pressure ratio J to a target value J _d, we can achieve appropriate spectrum.
For example, in FIGS. 13A and 13B, the control gain K is increased when the noise in the low frequency range (silence range) is large (when the sound pressure ratio J is small). Thereby, as shown by the arrow A in the same figure, the noise of a low frequency region falls. On the other hand, when the noise in the high frequency range (resonance frequency range) is large (when the sound pressure ratio J is large), the control gain K is decreased. As a result, as indicated by an arrow B in FIG. Such control of the control gain K is automatically performed by the integral control according to the above (Formula 1) or (Formula 5), and thereby converges to an appropriate target spectrum shape.

Moreover, as shown in FIGS. 13A and 13B, the target spectrum shape does not depend on the overall noise level. That is not the shape of the target spectrum is changed by the magnitude of the wind noise, the target value J _d representative of the target spectrum may be regarded as a constant value. Therefore, by adjusting the control gain K by the above-described integral control using the sound pressure ratio J, optimum control can be performed regardless of the magnitude of the wind noise.

  As described above, the final goal of applying active mute to the wind noise is to converge the spectrum shape of the wind noise to an appropriate spectrum shape in terms of comfort. Any user P can adjust to the target spectrum shape by adjusting the control gain K. However, because of individual differences in the ear transfer function, the value of the control gain K for convergence is the user P. It depends on. Therefore, in order to eliminate individual differences, the control gain K should be adjusted so as to converge to an appropriate shape while directly monitoring the shape of the spectrum rather than simply adjusting the control gain. This is also realized by the integral control using the above-described sound pressure ratio J.

Further, when there is no wind noise, the control gain K is set to zero (0), and as shown in FIG. 13C, active silencing is not performed and the noise signal is not unnecessarily increased. That is, when there is no wind noise, background noise (mainly high frequency sound) is dominant in the microphone output signal as compared with the wind noise state. Therefore, the ratio of the high frequency component in the total noise is larger than that in the state where there is wind noise. Therefore, the value of the sound pressure ratio J (= L ₁ / L ₂ or Lw / Lr) exceeds the target value J _d , and for example, the control gain K continues to decrease according to (Equation 5). However, since the K value is prevented from becoming negative because of the absolute value calculation, the output finally converges to K = 0 and the output of the speaker 104 becomes zero (0). That is, active mute is not performed.

  As described above, the non-speech gain adjustment circuit 121 having the configuration shown in FIG. 11 is designed to mute by converging the wind noise frequency spectrum detected by the microphone 102 to the target spectrum shape. Therefore, when the frequency spectrum of the output signal of the microphone 102 is affected by the speech of the user P, the non-speech gain adjustment circuit 121 may not be able to set an appropriate control gain for silencing.

Therefore, in this embodiment, when the utterance of the user P is detected, the control gain of the amplifier 108 is set by the utterance gain adjustment circuit 122 instead of the non-utterance gain adjustment circuit 121. As a result, it is possible to perform a good silencing operation regardless of whether or not the user speaks.
FIG. 14 is a diagram for explaining another embodiment of the present invention, and shows the overall configuration of a vehicle system including the active silencing helmet as described above. FIG. 15 is a block diagram showing an electrical configuration of the vehicle system. In these FIG. 14 and FIG. 15, the part corresponding to each part shown by above-mentioned FIG. 1A and FIG. 1B is shown with the same referential mark.

  In this embodiment, among the components of the active silencing helmet, only the microphone 102, the speaker 104 (for example, a flat speaker), and the speech detection microphones 1 and 2 are attached to the helmet body 30, and the rest such as the control signal generation circuit 106 is left. This component is provided in an ANC controller amplifier 41 as a vehicle body side device attached to a vehicle body 40 of a two-wheeled vehicle as an example of a vehicle. The ANC controller amplifier 41 is connected to the microphone 102 and the speaker 104 via a wire harness 42 in which a plurality of cables are bundled.

  The wire harness 42 includes a microphone signal line 43 for inputting the output signal of the microphone 102 to the ANC controller amplifier 41 and an audio signal line 44 for supplying a control signal for muting from the ANC controller amplifier 41 to the speaker 104. And an utterance detection microphone signal line 45 for inputting the output signals of the utterance detection microphones 1 and 2 to the ANC controller amplifier 41.

The sound signal line 44 is further connected to a sound information generating device 50 (sound information generating means) provided in the vehicle body 40. The sound information generating device 50 includes a sound source unit 51 that generates an audio signal, and a preamplifier 52 that amplifies the audio signal generated by the sound source unit 51 and transmits the amplified signal to the control signal generation circuit 106.
In this embodiment, the control signal generation circuit 106 includes an ANC control gain changing unit 108-1, an adding circuit 109, and a power amplifier 108-2. The ANC control gain changing unit 108-1 is a preamplifier circuit that amplifies the control signal from the mute control filter circuit 107 in accordance with the control gain generated by the gain adjustment circuit 120. The control signal output by the ANC control gain changing unit 108-1 and the output signal of the preamplifier 52 of the sound information generating device 50 are added by the adding circuit 109. The signal after the addition is amplified with a constant gain by the power amplifier 108-2 and sent to the audio signal line 44.

Therefore, the audio signal line 44 also functions as a transmission unit that transmits the audio signal generated by the sound information generating device 50 to the helmet body 30.
Thus, the speaker 104 provided in the helmet body 30 always produces a control sound for muting, and the sound signal generated from the sound information generating device 50 is acousticized and output when necessary. It will be. That is, the speaker 104 also functions as sound information sounding means for sounding sound information. In this manner, the wearer of the helmet body 30 can listen to the sound information generated by the sound information generating device 50 in a situation where the wind noise is well silenced.

The sound information generating device 50 may be a navigation device that provides voice guidance, may be an audio device such as a radio or an audio player, or may be a mobile phone (for example, a mail function in addition to a basic function for conversation). May have a reading function).
The ANC controller amplifier 41 and the sound information generating device 50 and the helmet body 30 do not necessarily have to be connected by wire, and signals may be exchanged between them by wireless communication such as infrared communication.

Of course, this embodiment is applicable to a four-wheeled vehicle as long as a helmet is required.
In addition, various design changes can be made within the scope of matters described in the claims.

It is a block diagram which shows the system configuration | structure of the active silencing helmet which concerns on one Embodiment of this invention. It is an external view of the active silencing helmet. It is a figure which shows the structure of the control system which the active silencing helmet which concerns on the said embodiment makes object. It is a block diagram for demonstrating the structural example of a speech detector. It is a figure which shows an example of the coherence calculated by the coherence calculating part. FIG. 5A and FIG. 5B are diagrams showing an arrangement example of the speech detection microphones in the helmet body. 6 (a) and 6 (b) are diagrams showing examples of coherence corresponding to the arrangement of the speech detection microphones of FIGS. 5 (a) and 5 (b), respectively. It is a flowchart which shows collectively the operation | movement of the whole gain adjustment circuit. It is a figure for demonstrating an example of the specific function of a spectrum comparison circuit. It is a block diagram which shows the example which comprises a spectrum comparison circuit using a band pass filter. It is a figure which shows the example of the gain map for calculating a control gain. It is a block diagram for demonstrating the structural example of the gain adjustment circuit at the time of a non-utterance. It is a figure for demonstrating the active mute control implement | achieved by the circuit of FIG. 13A, 13B, and 13C are diagrams for explaining the effect of the active mute control by the function of the non-speech gain adjustment circuit, and FIG. 13A is a diagram showing the effect in a state where the wind noise is large. FIG. 13B is a diagram illustrating an effect in a state where the wind noise is low, and FIG. 13C is a diagram illustrating an effect in a state where there is no wind noise. It is a figure for demonstrating other embodiment of this invention, and the whole structure of the vehicle system provided with the said active silencing helmet is shown. It is a block diagram which shows the electric constitution of the said vehicle system.

DESCRIPTION OF SYMBOLS 1, 2 Utterance detection microphone 3 Utterance detector 5 Gain preservation memory 6 Gain update changeover switch 7 Gain update control circuit 8 Fast Fourier transform circuit 9 Spectrum preservation memory 10 Spectrum update changeover switch 11 Spectrum update control circuit 12 Spectrum comparison circuit 13 Control gain Generation circuit 21 Coherence calculation unit 22 Threshold memory 23 Coherence comparison unit 24 Speech determination unit 30 Helmet body 31 Outer 33 Cover 35 Shield 40 Car body 41 Controller amplifier 42 Wire harness 43 Microphone signal line 44 Audio signal line 45 Speech detection microphone signal line DESCRIPTION OF SYMBOLS 50 Sound information generator 51 Sound source part 52 Preamplifier 61-63 Band pass filter 100 Active silencer helmet 102 Microphone 104 Speaker 106 control signal generation circuit 107 mute control filter circuit 108 amplifier 108-1 ANC control gain changing unit 108-2 power amplifier 109 addition circuit 120 gain adjustment circuit 121 non-speech gain adjustment circuit 122 utterance gain adjustment circuit 123 gain adjustment switching circuit 124 switching control circuit 202 A / D converter 204 D / A converter 206-1, 206-2 filter 208-1, 208-2 A / D converter 210-1, 210-2 sound pressure calculation unit 212 sound pressure ratio Calculation unit 214 Adjustment unit P User

Claims (15)

Noise detection means for detecting noise in the helmet body;
Sound generation means for generating a control sound for canceling the noise detected by the noise detection means,
A control signal generating means for calculating the output signal of the noise detecting means to generate a control signal corresponding to the control sound and giving it to the sounding means;
Utterance detection means for detecting the utterance of the wearer wearing the helmet body;
Non-speech gain adjusting means for adjusting the gain of the control signal generating means during a period when no utterance is detected by the utterance detecting means;
A non-speech gain storage unit that stores a gain set by the non-speech gain adjustment unit immediately before an utterance is detected by the utterance detection unit;
An utterance gain adjusting means for adjusting a gain of the control signal generating means based on a gain stored in the non-utterance gain storage means during a period in which an utterance is detected by the utterance detection means. helmet.   The utterance gain adjusting means adjusts the gain of the control signal generating means in accordance with the output signal of the noise detecting means and the gain stored in the non-utterance gain storage means. The mute helmet described. The utterance gain adjustment means includes:
Spectrum calculation means for obtaining a frequency spectrum of noise in the helmet body based on the output signal of the noise detection means,
Non-speech noise spectrum storage means for storing the frequency spectrum calculated by the spectrum calculation means immediately before the utterance is detected by the utterance detection means;
Spectrum comparison means for comparing the frequency spectrum of the noise calculated by the spectrum calculation means during a period when the utterance is detected by the utterance detection means and the frequency spectrum stored in the non-speech noise spectrum storage means; ,
The silencing helmet according to claim 1, further comprising: gain control means for adjusting the gain of the control signal generation means based on the comparison result by the spectrum comparison means and the gain stored in the non-speech gain storage means. .   The spectrum comparison unit is configured to calculate an amplitude spectrum at a specific frequency in a frequency band where an utterance peak corresponding to the utterance of a wearer does not substantially appear or an equivalent value thereof during a period when the utterance is detected by the utterance detection unit. The silencing helmet according to claim 3, which compares the frequency spectrum of the noise calculated by the spectrum calculation means and the frequency spectrum stored in the non-speech noise spectrum storage means.   The silencing helmet according to claim 4, wherein the specific frequency includes a plurality of different frequencies.   The gain control means obtains the gain of the control signal generation means by performing a correction according to the comparison result by the spectrum comparison means on the gain stored in the non-speech gain storage means. The silencing helmet according to claim 4 or 5, wherein The non-speech gain adjusting means includes:
A sound pressure ratio acquisition means for acquiring a ratio of sound pressures in different frequency bands using the output signal of the noise detection means;
Gain control means for controlling the gain of the control signal generation means so that the spectrum of the output signal of the noise detection means approaches a predetermined target spectrum using the sound pressure ratio acquired by the sound pressure ratio acquisition means. The silencing helmet according to any one of claims 1 to 6. And further comprising first and second sound detection means for detecting sound emitted by a wearer of the helmet body at different positions in the helmet body and outputting sound signals;
The speech detection means includes
Correlation calculating means for calculating a correlation value representing the correlation of the audio signals output by the first and second sound detecting means;
The muffler helmet according to any one of claims 1 to 7, further comprising: an utterance determination unit that determines whether or not the wearer is speaking according to a correlation value calculated by the correlation calculation unit. . Noise detection means for detecting noise in the helmet body;
Sound generation means for generating a control sound for canceling the noise detected by the noise detection means,
A control signal generating means for calculating the output signal of the noise detecting means to generate a control signal corresponding to the control sound and giving it to the sounding means;
First and second sound detection means for detecting sound emitted by a wearer of the helmet body at different positions in the helmet body and outputting sound signals;
Correlation calculating means for calculating the correlation value representing the correlation between the audio signals output by the first and second sound detecting means, and the helmet according to the correlation value calculated by the correlation calculating means. Speech detection means including speech determination means for determining whether or not the wearer wearing the main body is speaking;
Non-speech gain adjusting means for adjusting the gain of the control signal generating means during a period when no utterance is detected by the utterance detecting means;
A muffling helmet, comprising: an utterance gain adjusting means that adjusts the gain of the control signal generating means in a manner different from the non-utterance gain adjusting means during a period in which an utterance is detected by the utterance detecting means. The first and second sound detection means include
The silencer helmet according to claim 8 or 9, wherein the silencer helmet is disposed at a position substantially equidistant from a mouth of a wearer wearing the helmet body. The first and second sound detection means include
The helmet value is calculated so that the correlation value calculated by the correlation calculation means is equal to or greater than a predetermined threshold value when the wearer of the helmet body speaks and less than the threshold value when the wearer does not speak. The silencing helmet according to any one of claims 8 to 10, which is disposed in the body. The car body,
Including the muffling helmet according to any one of claims 1 to 11,
At least the noise detection means and the sound generation means are provided in a helmet body of the silencing helmet,
At least a part of the remaining components excluding the noise detection means and sound generation means of the muffling helmet is provided in the vehicle body to form a vehicle body side device, the vehicle body side device, the noise detection means, and the sound generation means, A vehicle system further comprising communication means for transmitting and receiving signals between the two. The car body,
The silencing helmet according to any one of claims 1 to 11,
Sound information generating means provided on the vehicle body for generating sound information;
Transmitting means for transmitting sound information generated by the sound information generating means to the helmet body of the muffling helmet;
A vehicle system comprising sound information sounding means that is provided in the helmet body and that makes sound information transmitted by the transmitting means sound. A noise detection step for detecting noise in the helmet body by a noise detection means;
Generating a control sound for canceling the detected noise from the sound generation means;
Amplifying a control signal obtained by calculating the output signal of the noise detecting means by an amplifying means and giving the sounding means;
An utterance detection step of detecting an utterance of a wearer wearing the helmet body;
A non-speech gain adjustment step of adjusting the gain of the amplification means during a period when the wearer's utterance is not detected;
A noise adjustment in a helmet, including a gain adjustment step during speech adjustment that adjusts the gain of the amplification means based on the gain set in the gain adjustment step during the previous non-speech during a period when the wearer's speech is detected Method. A noise detection step for detecting noise in the helmet body by a noise detection means;
Generating a control sound for canceling the detected noise from the sound generation means;
Amplifying a control signal obtained by calculating the output signal of the noise detecting means by an amplifying means and giving the sounding means;
A voice detection step of detecting a voice uttered by a wearer of the helmet body at different positions in the helmet body by the first and second voice detection means;
A correlation calculation step for calculating a correlation value representing the correlation between the audio signals output by the first and second audio detection means;
An utterance determination step for determining whether or not the wearer wearing the helmet body is speaking according to the calculated correlation value;
A non-speech gain adjustment step of adjusting the gain of the amplification means during a period when it is determined that the wearer is not speaking;
A method for silencing noise in a helmet, comprising: a gain adjustment step during speech that adjusts a gain of the amplification means in a manner different from the gain adjustment step during non-speech during a period when it is determined that the wearer is speaking.