JP2002051392A - In-vehicle conversation assisting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable crews of a vehicle to clearly hear the speaking voices of their partners, even under a noisy conditions or while a music is being reproduced. SOLUTION: A voice-correcting section 201 separates speaking voice signals from the interior sounds of the vehicle collected through microphones 105-108, provided at each seat and performs gain correction on the voice signals according to the sound pressure levels of audio sounds and noise. A speaking voice and sound field control section 202 determines the seats of persons speaking from the magnitudes of input levels at each microphone and outputs the corrected speaking voices to a synthesizer 204 by changing the output levels of the speaking voices at each loudspeaker channel, so that the corrected speaking voices may be heard from the seats of the speaking persons. The synthesizer 204 synthesizes the speaking voices with the audio sounds and outputs synthesized signals from loudspeakers 101-105, after amplifying the signals by means of an amplifier 205.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an in-vehicle conversation assistance device, and more particularly to an in-vehicle conversation assistance device capable of outputting a conversation sound acquired by a microphone attached to each seat from an audio reproduction sound output speaker together with the audio reproduction sound. About.

[0002]

2. Description of the Related Art Various sounds are present in the interior of an automobile. This includes music played by the audio device, noise from the engine, noise from the tires while traveling, and the like. In such a state, a plurality of occupants had a conversation.

[0003]

For this reason, the sound uttered by the person in the front seat is sometimes drowned out by the audio sound or ambient sound, and cannot be heard in the rear seat. Also,
In conversations while driving on a highway, not only the reproduction sound output from the audio device, but also the engine noise, road noise from the tires, wind noise, and the like increase, making it more difficult to hear conversation sounds. Therefore, when talking, it is necessary to lower the volume of the audio device for conversation, or to speak at a size that can be heard by the other party even when listening to music with the audio device. It is conceivable that a microphone is attached to each seat and the sound emitted by the occupant is output from the speaker together with the audio sound.However, since the audio sound and the conversation sound are simply output as they are, the audio sound There was a problem that conversational voice could not be heard due to being masked.

In addition, when a plurality of people are in the vehicle, simply listening to the conversation sound output from the speaker masks the conversation sound in the audio reproduction sound and the surrounding sound, which makes it difficult for the person to hear. There was a problem that it was difficult to understand whether it was.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an in-vehicle conversation assistance device capable of clearly listening to conversation voices between occupants even under noise or during reproduction of music or the like.

It is another object of the present invention to provide an in-vehicle conversation assisting device that can clearly hear conversation voices between occupants and easily understand who is talking, even in the presence of noise or during reproduction of music or the like. It is to provide.

[0007]

SUMMARY OF THE INVENTION According to the present invention, there is provided an in-vehicle conversation assistance device capable of outputting a conversation voice of an occupant in a vehicle together with audio sounds from a plurality of audio sound output speakers installed in the vehicle. Sound collecting means provided for each seat, audio sound removing means for removing a component corresponding to audio sound from the output signal of the sound collecting means, and corresponding to ambient noise from the output signal of the sound collecting means And a component corresponding to the audio noise from the output signal of the sound collecting unit, the component corresponding to the audio sound being removed by the audio sound removing unit, and the component corresponding to the peripheral noise being removed by the noise removing unit. Voice correction means for performing a predetermined gain correction on a component corresponding to the conversation voice after the noise has been removed, and gain correction by the voice correction means. The output level of conversational speech is achieved by vehicle conversation assistance device that includes a conversational speech sound field control means for setting for each speaker.

That is, with the above configuration, the conversational voice is separated from the in-vehicle sound collected for each seat, the gain of the conversational voice is corrected, and the output level of each speaker is changed to change the conversational voice. Can be output.

Further, according to the present invention, the sound correction means outputs the sound from the speaker based on a sound pressure level of the conversation sound and a sound pressure level of the audio sound and the surrounding noise. Gain calculating means for calculating a correction gain for each of a plurality of divided frequency bands of the conversation voice so that the conversation voice can be clearly heard, and the correction gain calculated by the gain calculation means for the conversation voice This is achieved by an in-vehicle conversation assistance device including:

That is, with the above-described configuration, the conversation sound output from the speaker can be clearly heard based on the sound pressure level of the conversation sound and the sound pressure levels of the audio sound and the surrounding noise. A correction gain for each frequency band of the voice is calculated by the gain correction means, and correction based on the correction gain is performed, so that a clear voice can be output from the speaker. The amount of gain correction that can be heard clearly is not determined uniformly in all frequency ranges, but differs for each frequency component of audio sound, ambient noise, or conversational voice. By calculating the gain and performing the gain correction, a clearer sound can be output from the speaker.

Further, according to the present invention, there is provided a speaker seat discriminating means for discriminating a speaker's seat, wherein the conversational sound field control means includes a speaker seat discriminated by the speaker seat discriminating means. This is achieved by an in-vehicle conversation assistance device that sets the output level of the corrected conversation voice accordingly for each speaker.

That is, with the above-described configuration, the position of the speaker is determined, the speech sound output level of each speaker is adjusted according to the position of the speaker, and the direction in which the conversation sound is heard is changed. Can be.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (a) Schematic Description FIG. 1 is a schematic explanatory diagram of an in-vehicle conversation assistance device according to the present invention.
In the figure, 101 to 104 are speakers installed in the vehicle, 101 is a front light speaker, 102
Denotes a front left speaker, 103 denotes a rear right speaker, and 104 denotes a rear left speaker. 105-10
Reference numeral 8 denotes a microphone installed near each seat, reference numeral 105 denotes a driver's seat microphone, reference numeral 106 denotes a passenger seat microphone, reference numeral 107 denotes a rear right seat microphone, and reference numeral 108 denotes a rear left seat microphone.

Reference numeral 200 denotes a control unit for controlling the entire in-car conversation assistance device, 201 denotes a voice correction unit, 202 denotes a conversation voice sound field control unit, 203 denotes an audio device, 204 denotes a synthesizer, 20
5 is an amplifier.

The audio correction unit 201 is provided for each of the microphones 105, 1
A device which removes audio sounds and noises from the sounds in the vehicle collected at 06, 107 and 108, extracts the conversation voice of the occupant, performs a loudness correction process on the conversation voice based on the audio sound and noise, and outputs the result. It is.

The conversation sound field control unit 202 adjusts the conversation sound signal level to be output from each speaker according to the configuration of the speaker's seat and the occupant, and adjusts the conversation sound signal corrected by the speech correction unit 201. It controls the sound field of the conversation voice.

The synthesizer 204 synthesizes the audio sound signal input from the audio device 203 and the conversation sound signal input from the conversation sound field control unit 202 for each speaker channel, and inputs the synthesized signal to the amplifier 205. I do. The amplifier 205 amplifies the synthesized signal of the audio sound signal and the conversation sound signal input from the synthesizer 204 and outputs the amplified signal from each of the speakers 101 to 104.

(B) Detailed Description of the Audio Correction Unit FIG. 2 is a diagram showing the configuration of the audio correction unit 201. The audio correction unit 201 includes an adaptive filter 10, a filter 12, operation units 20, 22, a peripheral noise removal unit 30, a loudness compensation operation unit 40, and an audio correction filter 42.

The adaptive filter 10 is for simulating the transfer characteristic of the acoustic space in the vehicle compartment, is an FIR digital filter having a filter coefficient (tap coefficient) W1, and is input from the audio device 203. A predetermined adaptive signal processing is performed using the audio sound signal as a reference signal. This filter coefficient W1 is determined by LMS (Least Mean S
quare) The algorithm is updated so that the power of the difference signal (described later) output from the calculation unit 22 is minimized. The filter 12 is for simulating the transfer characteristic of the acoustic space in the vehicle interior, like the adaptive filter 10, and has a filter coefficient W2. Filter coefficient W
2, the filter coefficient W1 of the adaptive filter 10 is copied at a predetermined timing.

The arithmetic unit 20 receives the output signal of the microphone 105 and the output signal of the filter 12, and calculates the difference between these two signals. The filter 12 removes the echo component of the conversation voice that is output from the speaker 101 and goes around the microphone 105 and is collected. The arithmetic unit 22 receives the difference signal output from the arithmetic unit 20 and the output signal of the adaptive filter 10, and calculates the difference between these two signals.

The peripheral noise elimination unit 30 is connected to the operation unit 2 in the preceding stage.
The component corresponding to the peripheral noise included in the difference signal output from the second signal is removed. From the peripheral noise removing unit 30, only the component corresponding to the conversation voice included in the signal output from the microphone 105 is extracted and output. In addition,
The peripheral noise removing unit 30 can be configured by, for example, a high-pass filter. Since the noise in the vehicle compartment has a large low-frequency component, the noise can be reduced by a high-pass filter.

An audio sound signal, a surrounding noise signal, and a conversation voice signal are input to the loudness compensation calculation unit 40. Based on these signals, each of the frequency components necessary for outputting the conversation voice from the speaker 101 is output. Is calculated. The voice correction filter 42 performs gain correction on the conversation voice signal output from the peripheral noise removal unit 30 based on the correction gain of each frequency component calculated by the loudness compensation calculation unit 40. The detailed configurations of the loudness compensator 40, the audio correction filter 42, and the audio correction filter 42 will be described later.

The audio sound signal output from the audio device 203 is input to the adaptive filter 10, and is output to the speaker 10 via the synthesizer 204 and the amplifier 205.
1 to the vehicle interior. Since the audio sound output from the speaker 101 is collected by the microphone 105 installed at a predetermined position where the speaker can hear the uttered voice of the speaker, the audio sound is output from the microphone 105 and is output to the arithmetic unit 22 through the arithmetic unit 20. The signal input to one of the input terminals includes a component for the audio sound. Further, since the audio sound is output by the microphone 105 after being output to the vehicle interior, the audio sound reflects the transfer characteristics of the acoustic space in the vehicle interior.

Therefore, the output signal of the microphone 105 including the component corresponding to the audio sound reflecting the transfer characteristic of the acoustic space in the vehicle cabin and the audio sound signal directly input from the audio device 203 are passed through the adaptive filter 10. The difference between the difference signal and the calculated signal is calculated by the calculation unit 22, and the filter coefficient W1 of the adaptive filter 10 is updated so that the power of the difference signal is minimized. This simulates the transfer characteristics. That is, the signal input to one input terminal of the arithmetic unit 22 includes a component corresponding to the audio sound output from the audio device 203 and output to the actual acoustic space in the vehicle interior, and the other The signal input to the input terminal includes an adaptive filter 10 simulating the characteristics of the acoustic space.
A component corresponding to the audio sound after passing through is included, and a difference corresponding to the audio sound is calculated by the calculation unit 22 to remove the component corresponding to the audio sound. A peripheral noise removing unit 30 is provided after the arithmetic unit 22.
Are arranged, and the peripheral noise included in the output signal of the arithmetic unit 22 is removed.

As described above, the adaptive filter 10 and the arithmetic unit 2
2 removes the component corresponding to the audio sound, and further removes the component corresponding to the peripheral noise by the peripheral noise removing unit 30. Therefore, the conversation voice of the crew and
The audio sound output from the speaker 101 and the surrounding noise such as running noise and engine noise are simultaneously output from the microphone 10.
5, the components corresponding to the audio sound and the surrounding noise are removed from the output signal of the microphone 105 on which these are superimposed, and the surrounding noise 30 corresponds to the conversation voice of the occupant. It is possible to output only the component that performs

The component corresponding to the speech voice output from the surrounding noise 30 is input to the voice correction filter 42 and voice-corrected according to the correction gain from the loudness compensation calculation unit 40. It is output to the control unit 202.

(C) Details of Loudness Compensation Calculation Unit (c-1) Principle of Gain Calculation First, the principle of gain calculation of the loudness compensation calculation unit 52 will be described. FIG. 3 is a diagram showing a correspondence relationship (loudness curve) between a physical sound pressure level and a loudness (loudness) of a sound felt when a person hears the sound.
In the figure, the horizontal axis is the sound pressure level (unit: dB-SP)
L), the vertical axis is the loudness (unit: sound) indicating the loudness of the sound perceived by humans, the curve is a loudness curve under silence, and the curve is a loudness curve under noise.
However, the curve changes according to the noise level.

In the figure, if the loudness values are the same, a human perceives sounds of the same volume. Therefore, for example, the sound that a person perceives as having a magnitude of 0.1 sound has a sound pressure level of about 12 dB-SPL under silence, but about 37 dB-SPL under curve noise.
This is a sound having a sound pressure level of SPL. In other words, in order to make the sound output at about 12 dB-SPL in silence to be the same volume under the noise of the curve, it is necessary to have about 37 dB-SPL.
It is necessary to output PL sound. That is, under the noise of the curve, it is necessary to add a gain of about 25 dB. In addition, the sound that a person feels the size of one sound is a sound having a sound pressure level of about 42 dB-SPL under silence, but a sound having a sound pressure level of about 49 dB-SPL under a curve noise. Below, it is necessary to add a gain of about 7 dB. Therefore, even under the same noise, it is necessary to change the gain to be applied according to the sound pressure level of the output sound.

FIG. 4 is a diagram showing how much gain needs to be added to the sound pressure level under silence in order to feel the same loudness as under silence. In the figure, the horizontal axis represents the sound pressure level of the sound output in silence, and the vertical axis represents the gain value that needs to be added in order to feel the same loudness as in quiet under noise. For example, a sound output at a sound pressure level of 20 dB-SPL under silence can be perceived by a human being as having the same loudness as under silence by adding a gain of about 19 dB under noise.

The loudness compensation calculating section 40 preliminarily calculates the sound pressure level of the sound signal (the sound pressure level of the voice output by the user output from the peripheral noise removing section 30) at various noise levels as shown in FIG. A relationship with the gain to be added (hereinafter, referred to as a gain table) is stored in an internal memory, and an optimum gain table is selected based on the input audio sound signal and the ambient noise signal, and the selected gain table is selected. An optimum gain is calculated based on the table and the audio signal output from the peripheral noise removing unit 30. The loudness compensation calculation unit 40 outputs the calculated gain to the sound correction filter 42 to give an optimum gain to the sound signal.

Generally, audio sound signals and noise signals have various frequency components, and the sound pressure level differs for each frequency component. Therefore, there is an imbalance that the difficulty of listening to the conversation voice signal varies depending on the sound pressure level of each frequency component of the conversation voice signal, the audio sound, and the ambient noise. In addition, since each frequency component of the audio sound and the surrounding noise exerts a masking effect on the speech signal of the high frequency component, it is necessary to consider this.

Therefore, it is desirable to provide an optimum gain for each frequency component of the speech signal. That is, the speech signal, the audio sound signal, and the surrounding noise signal are each divided into predetermined frequency bands, and an optimum gain table is selected for each frequency band based on the frequency components of the audio sound signal and the surrounding noise signal. It is desirable to calculate the optimum gain based on the selected gain table and the frequency component of the audio signal.

(C-2) Configuration of Loudness Compensation Computing Unit FIG. 5 is a diagram showing the configuration of the loudness compensation computing unit 40. As shown in the figure, the loudness compensation calculating section 40 is configured to include a frequency band level averaging section 61, a loudness calculating section 64, a frequency band gain table selecting section 65, a frequency band level averaging section 66, and a gain table 69. .

The frequency band level averaging unit 61 calculates the audio sound signal input from the adaptive filter 10 and the arithmetic unit 22.
A known FFT is performed for each predetermined time block with respect to a surrounding noise signal (hereinafter, referred to as noise or the like) input from
(Fast Fourier Transform) calculation is performed, and the average of the sound pressure level is calculated for each predetermined frequency band. Noise etc.
For example, the frequency is divided into 1/3 octaves in consideration of the characteristic that human hearing can recognize a difference in loudness approximately every 1/3 octave.

When the occupant has a conversation, the difference signal output from the calculation unit 22 includes a component corresponding to the occupant's speech signal. Therefore, the loudness compensation calculation unit 40 performs the operation when the occupant is not speaking. Only the peripheral noise signal output from the arithmetic unit 22 is taken in. For example, it is determined that the utterance is made when the output signal level of the arithmetic unit 22 exceeds a predetermined level. In this case, the surrounding noise may be taken in by pressing the talk switch.

The loudness calculating section 64 is a well-known Zwick.
ker loudness calculation method (ISO 532B) and S
Evens loudness calculation method (ISO 532
A) is used to adjust the sound pressure level of noise or the like output from the frequency level averaging unit 61 for each frequency band. Specifically, the adjustment is performed as follows. That is, when there is noise or the like of a certain frequency component, this noise or the like not only affects the difficulty of hearing the conversation voice signal of the same frequency component, but also the speech voice of the frequency component adjacent to the high frequency side as described above. It also affects the difficulty of hearing the signal. In consideration of this, the loudness calculation unit 64 adjusts the sound pressure level of each frequency component such as noise according to the magnitude of the sound pressure level of the frequency component such as noise adjacent to the low frequency side. That is, when the sound pressure level of the adjacent low frequency component is large, the sound pressure level of the frequency component adjacent to the high frequency side is corrected to be higher. By making such adjustments, when selecting a gain table for each frequency band, it is sufficient to focus only on the sound pressure level of noise or the like in the corresponding frequency band, and the frequency band adjacent to the low frequency side is sufficient. It is not necessary to perform a complicated process of considering noise or the like.

The frequency band gain table selection section 65
An optimum gain table 69 is selected for each frequency band based on the adjusted sound pressure level of noise or the like for each frequency band output from the loudness calculation unit 64.

The frequency band level averaging unit 66 performs a well-known FFT operation on the speech signal output from the peripheral noise removing unit 30 for each short-time block, and calculates the sound pressure level for each predetermined frequency band. Calculate the average. A conversation voice signal is divided into frequency bands similar to noise and the like. The speech signal divided for each frequency band output from the frequency band level averaging unit 66 is converted into a gain table 69 selected by the frequency band gain table selecting unit 65.
And an appropriate gain value is calculated for each frequency band.

As described above, by dividing noise or the like or an audio signal into predetermined frequency bands, it becomes possible to select a gain table for each frequency band and add an optimum gain to the audio signal.

In the above-mentioned loudness compensation calculating section 40,
Although the average of the sound pressure level for each frequency band of the audio signal, noise, and the like was obtained using the frequency band level averaging units 61 and 66, a filter bank and a block averaging unit were used instead of these frequency band level averaging units. The average of the sound pressure levels for each frequency band may be obtained.

FIG. 6 is a diagram showing a configuration of a modified example of the loudness compensation calculating section 40. The loudness compensation calculation unit 40 shown in FIG.
3. Loudness calculating section 64, frequency band gain table selecting section 65, filter bank 67, block averaging section 6
8, a gain table 69 is included.

The filter bank 62 is a group of band-pass filters having a predetermined frequency bandwidth. These band-pass filter groups are used to filter audio sound output from the adaptive filter 10 and noise output from the arithmetic unit 22 in a predetermined manner. Is divided for each frequency band. Block averaging unit 63
Averages sound pressure levels such as noise divided for each frequency band output from the filter bank 62 for each predetermined time block, and outputs the averaged sound pressure level to the loudness calculation unit 64.

The filter bank 67 includes the filter bank 6
Similarly to 2, the audio signal output from the peripheral noise removing unit 30 is divided for each predetermined frequency band by a band-pass filter group having a predetermined frequency bandwidth. The block averaging unit 68 averages the sound pressure level of the audio signal divided for each frequency band output from the filter bank 67 for each predetermined time block, and outputs the average sound pressure level to the gain table 69.

As described above, by using the filter bank and the block averaging unit, noise or the like or a speech signal can be divided into predetermined frequency bands, and a gain table is selected for each frequency band to produce speech speech. An optimum gain can be added.

(D) Details of Audio Correction Filter Next, details of the audio correction filter 42 will be described. The audio correction filter 42 may be of any type as long as it can be corrected (addition of gain) according to the gain characteristic calculated by the above-described loudness compensation calculation unit 40, and various configurations are conceivable. Three configurations will be described.

(D-1) When Using a Filter Bank and a Variable Gain Section FIG. 7 is a diagram showing a configuration of a sound correction filter 42 using a filter bank and a variable gain section. The filter 42 for audio correction shown in the figure includes a filter bank 71, a variable gain unit 72, and an adder 73.

The filter bank 71 is a group of band-pass filters having a predetermined frequency bandwidth, and divides an audio signal into each frequency band by these band-pass filters. The variable gain unit 72 calculates the gain for each frequency band calculated by the loudness compensation calculation unit 40,
The gain is adjusted by giving the sound pressure level of the audio signal divided for each frequency band output from the filter bank 71. The adder 73 realizes a desired gain correction by adding the audio signals whose gain has been adjusted for each frequency band. According to this configuration, the audio correction filter 42 can be configured at low cost with an analog circuit.

(D-2) When a Frequency Sampling Filter is Used FIG. 8 is a diagram showing a configuration of a sound correction filter 42 using a frequency sampling filter. The voice correction filter 42 shown in FIG.
An FFT operation unit 82 and an FIR filter 83 are included.

The spline function interpolation unit 81 uses the gain of each frequency band calculated by the loudness compensation calculation unit 40 as the gain of the center frequency of each frequency band, and uses a well-known spline function between the gain values. By performing interpolation, a smooth gain characteristic in a frequency domain is obtained. The IFFT operation unit 82 converts the gain characteristic output from the spline function interpolation unit 81 into a well-known I
The frequency domain is converted to the time domain using FFT (Inverse Fast Transform) operation, and the value of the tap coefficient of the FIR filter 83 is set. The FIR filter 83 performs filtering processing on the time axis with respect to the audio signal, and realizes desired gain correction. According to this configuration, a linear phase filter can be realized, and the correction for the audio signal can be performed not for each frequency band but for each frequency component.

(D-3) When a Frequency Domain Filter is Used FIG. 9 is a diagram showing a configuration of a sound correction filter 42 using a frequency domain filter. The filter 42 for audio correction shown in the figure includes a spline function interpolation unit 91, an FFT operation unit 92, a frequency band filtering unit 93, and an IFFT operation unit 94.

The spline function interpolation unit 91 uses the gain of each frequency band calculated by the loudness compensation calculation unit 40 as the gain of the center frequency of each frequency band, and uses a well-known spline function between the respective gain values. By performing interpolation, a smooth gain characteristic in a frequency domain is obtained. The FFT operation unit 92 performs an FFT operation on the audio signal, and converts the time domain into the frequency domain. The frequency band filtering unit 93 performs a filtering process on the audio signal in the frequency domain output from the FFT operation unit 91 by using a smooth gain characteristic output from the spline function interpolation unit 91. I
The FFT operation unit 94 includes a frequency band filtering unit 93
I for the audio signal in the frequency domain output from
An FFT operation is performed to convert from the frequency domain to the time domain, thereby achieving desired gain correction. In the process of the IFFT operation, a well-known overlap-add method or overlap-save method may be used to realize linear filtering. With this configuration, the amount of calculation can be relatively reduced even when the number of taps of the filter is large.

In any of the above three types of sound correction filters 40, if the gain changes abruptly, the output waveform becomes discontinuous, so that G (n) = αG (n−1). It is preferable to gradually update the gain characteristic using + βGm. Here, G (n) is a gain characteristic at time n,
m is a gain characteristic calculated by the loudness compensation calculation unit 52 and the spline function interpolation units 81 and 91. α,
β is a coefficient and has a relation of α + β = 1.

In the description of the audio correction unit 201, for simplicity, one microphone and one speaker have been described. However, in the conversation assisting apparatus of the present invention, the microphone is used. A voice correction unit is provided for one of the microphones, and the voice correction processing is performed on the conversation voice collected by each microphone.

(E) Details of the Conversational Sound Field Control Unit FIG. 10 is a detailed explanatory diagram of the conversational sound field control unit 202. In the figure, 200 is a control unit, 201 is a voice correction unit, 202 is a conversation voice sound field control unit, 203 is an audio device, and 204 is a synthesizer.

The microphones 105, 1 attached to each seat
The in-vehicle sound collected at 06, 107 and 108 is the sound correction unit 2
01 is input. The voice correction unit 201 performs voice correction on the conversation voice signal input as described above, and outputs the corrected voice signal to the conversation voice sound field control unit 202.

The conversational sound field control unit 202 adjusts the output level of the corrected conversational sound signal input from the sound correction unit 201 for each output channel by the fader / balance control units 202a to 202d. ~ 202
h and outputs the result to the synthesizer 204.

In the synthesizer 204, the audio device 203
And the conversational sound signal input from the conversational sound field control unit 202 for each channel of the output speaker, and outputs the result to the amplifier 205.

The control unit 200 determines the seat where the speaking person sits from the difference between the conversational voice input levels collected by the microphones,
When outputting the sound, the sound output level of each speaker is adjusted, and an instruction is given to the conversation sound field control unit 202 so that the sound can be heard from the seat direction of the speaker. For example, by comparing the level of the conversation voice collected by each microphone, if the sound pressure level of the audio signal obtained by the microphone attached to the front left seat is high, it is determined that the speaker is an occupant of the front left seat. An instruction to adjust the conversation sound output level from each speaker so that sound can be heard from the front left direction is given. The conversation sound field control unit 202 adjusts the output level of the sound for each speaker channel according to the instruction from the control unit 200.

By doing so, it is possible to output the voice in the vehicle so that it can be heard clearly from the direction of the speaker, and when talking in the audio playback sound or noise, the content of the conversation can be clearly heard. In addition to this, it is possible to make it easy to understand which person is speaking the conversation voice.

As described above, in the conversation assisting apparatus of the present invention, the sound pressure level of each predetermined frequency band such as an audio sound signal and a peripheral noise signal is obtained, and the gain table suitable for the sound pressure level of each frequency band is obtained. Is selected, and optimal gain correction can be performed for each frequency band of the input conversation voice signal. Therefore, when a conversation is made, by adjusting the gain of the conversation sound output from the speaker for each frequency band according to the level of ambient noise or the like, the conversation sound can always be clearly heard.

In the description of the above embodiment, the determination of the seat position of the occupant who made the utterance is made based on the level difference of the conversation voice signal collected by each microphone. However, a talk switch is provided in each seat. Alternatively, it may be determined that the seat on which the talk switch is pressed is the seat of the occupant who spoke.

As a modified example, a sensor (for example, a weight sensor, a temperature sensor, etc.) capable of determining whether or not there is an occupant in each seat is installed, and the control unit 200 controls the number of occupants and the number of occupant seats. It is also possible to change the sound field setting of the conversation sound output from the speaker of the conversation sound field control unit 202. For example, when a person is only in the driver's seat and the passenger's seat and there is no person in the rear seat, the conversation sound field control unit 202 controls the conversation sound field control unit 202 so that the conversation sound can be heard loudly from a speaker near the non-speaker. When the setting is made and there are three or more occupants, the setting of the conversation sound field control unit 202 can be made so that the conversation sound can be heard from the direction of the seat where the speaker is present.

In this way, when there are two occupants, it is possible to easily determine who the speaker is, so that the conversation voice can be heard more clearly by listening to the conversation content from a nearby speaker. On the other hand, when there are three or more occupants, it is possible to easily determine who the speaker is.

Although the present invention has been described with reference to the embodiments, the present invention can be variously modified in accordance with the gist of the present invention described in the claims, and the present invention does not exclude these.

[0065]

As described above, according to the present invention, a microphone is provided for each seat, a conversation voice is separated from a sound collected by each microphone, and the conversation voice is determined according to the sound pressure level of the audio sound and the surrounding noise. The gain correction is performed so that the conversation sound is output together with the audio sound from the speaker, so that the audio sound in the car and the conversation sound in the noise can be clearly heard, and the burden of talking in the car is reduced. It is reduced.

Further, according to the present invention, the seat of the speaker is determined, and the sound field of the conversation voice output from the speaker is changed according to the speaker's seat, so that the conversation voice is output from the direction of the speaker's seat. Since it is made audible, when there are a plurality of occupants, it is possible to easily determine who spoke the content. In addition, the sound field setting of the conversation voice is changed according to the number of occupants and the boarding position. Therefore, when the number of occupants is two, the conversation voice can be heard more clearly, and when the number of occupants is three or more. Can easily determine who is speaking.

[Brief description of the drawings]

FIG. 1 is a schematic explanatory view of an in-vehicle conversation assistance device according to the present invention.

FIG. 2 is a diagram illustrating a configuration of an audio correction unit.

FIG. 3 is a diagram showing a correspondence relationship between a sound pressure level and a loudness of a sound felt when a person hears the sound.

FIG. 4 is a diagram showing how much gain needs to be added to a sound pressure level in silence in order to feel a sound of the same volume as in silence under noise.

FIG. 5 is a diagram illustrating a configuration of a loudness compensation gain calculator.

FIG. 6 is a diagram showing a modification of the loudness compensation gain calculation unit.

FIG. 7 is a diagram showing a configuration of an audio correction filter using a filter bank and a variable gain unit.

FIG. 8 is a diagram showing a configuration of an audio correction filter using a frequency sampling filter.

FIG. 9 is a diagram showing a configuration of a sound correction filter using a frequency domain filter.

FIG. 10 is a detailed explanatory diagram of a conversation sound field control unit.

[Explanation of symbols]

 101, 102, 103, 104: Microphone 105, 106, 107, 108: Speaker 200: Control unit 201: Voice correction unit 202: Conversation voice sound field control unit 203: Audio Device 204 ・ ・ ・ Synthesizer 205 ・ ・ ・ Amplifier

Claims (4)

[Claims] 1. An in-vehicle conversation assistance device capable of outputting a conversation voice of an occupant in a vehicle together with audio sounds from a plurality of audio sound output speakers installed in the vehicle, comprising: a sound collection means provided for each seat; An audio sound removing unit that removes a component corresponding to an audio sound from an output signal of a sound unit; a noise removing unit that removes a component corresponding to a peripheral noise from an output signal of the sound collecting unit; From the output signal of the means, the component corresponding to the audio sound is removed by the audio sound removing means, and the component corresponding to the speech voice after the component corresponding to the surrounding noise is removed by the noise removing means. On the other hand, an audio correction unit for performing a predetermined gain correction, and an output level of the conversational sound whose gain has been corrected by the audio correction unit for each speaker. Car conversation assistance device is characterized in that a conversational speech sound field control means for constant. 2. The speech correction means according to claim 1, wherein said speech sound output from said speaker is clearly heard based on a sound pressure level of said speech sound and a sound pressure level of said audio sound and said peripheral noise. Gain calculation means for calculating a correction gain for each of a plurality of divided frequency bands of the conversation voice; gain correction means for performing a gain correction on the conversation voice based on the correction gain calculated by the gain calculation means; The in-vehicle conversation assistance device according to claim 1, further comprising: 3. A speaker voice discriminating means for discriminating a speaker's seat, wherein the conversation voice sound field control means determines the speech voice corrected according to the speaker's seat determined by the voice speaker's seat discriminating means. 3. The in-vehicle conversation assistance device according to claim 1, wherein the output level is set for each speaker. 4. An occupant detection means for detecting whether or not an occupant is in each seat, wherein the conversational sound field control means is based on the number of occupants detected by the occupant detection means and the boarding position. The in-vehicle conversation assistance device according to claim 1, wherein an output bell of the corrected conversation voice is set for each speaker.