JP2008042390A - In-vehicle conversation support system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an in-vehicle conversation support system for allowing crews to share a comfortable conversion environment, when vehicle is moving. <P>SOLUTION: The in-vehicle conversation support system supports an in-vehicle conversation in an in-vehicle environment, where a speaker listens voice issued from another speaker and amplified voice from a loudspeaker. The system includes a plurality of individual microphones or a common microphone arranged in the vehicle; a common loudspeaker or a plurality of individual loudspeakers arranged in the vehicle; and an acoustic processing part where a correcting part, having a filter part for preventing howling and a delay part for a precedence sound effect, and an adjusting part are arranged, and the output of the individual microphones or the common microphone are corrected by the correcting part, amplified by the adjusting part, and connected to the common loudspeaker or the individual loudspeakers. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention enables in-vehicle conversation support system in a vehicle environment in which a speaker listens to a live sound (direct sound) produced by another speaker and a loud sound from a speaker, and further enables conversation while listening to the sound of other contents. The present invention relates to an in-car conversation support system.

One of the things that enhances the enjoyment of touring by multiple passengers is a comfortable conversation environment.
However, engine noise, tire noise, or noise from the outside of the vehicle when the vehicle is running as compared to when the vehicle is stopped makes it difficult to maintain a comfortable conversation environment because the in-car conversation is hindered.
In particular, when a large recreational vehicle (RV) or an occupant includes an elderly person, the distance between the occupants increases, and a sensory deafness (specific to the elderly person) is also added to distinguish between noise and conversation. It becomes difficult to “distinguish”.
In order to improve this, it is necessary to effectively amplify the in-vehicle conversation sound according to the magnitude and spectrum of the running noise, and to stably amplify the collected conversation sound without howling.

Many proposals have been made on the improvement of the in-car conversation environment or the technology necessary for the improvement.
The invention relating to the in-vehicle conversation assistance device described in Patent Document 1 discloses a roof module for in-vehicle conversation assistance in which a speaker and a microphone are arranged on the same plane. A structure that is not transmitted to the microphone) is provided.
The invention of the in-vehicle conversation assisting device described in Patent Literature 2 discloses that a microphone is provided for each speaker (that is, for each seat) and various adjustments are made.
The invention of the in-vehicle audio device described in Patent Document 3 proposes an arrangement of a microphone and a speaker in consideration of howling.
The invention of the acoustic system described in Patent Document 4 discloses the arrangement of microphones and speakers in an acoustic space (non-planar) for preventing howling.
The invention of the sound field control device described in Patent Document 5 shows an invention in which feedback from a speaker to a microphone is removed using a FIR filter to prevent howling.

  Since the in-vehicle conversation support system according to the present invention has much in common with the invention described in Patent Document 2 and the in-vehicle environment, the invention described in Patent Document 2 will be described in some detail. The invention described in Patent Document 2 is excellent in that it is designed so that conversational voices between passengers can be heard clearly even under noisy or playing music. The in-vehicle sound collected by a plurality of microphones installed for each seat is gain-corrected according to the level of audio sound and noise, and is amplified in the vehicle. The conversation sound is separated from the signal picked up by the microphone, and the conversation voice output level is set for each speaker channel so that the speaker's seat can be judged from the input level of each microphone and the sound can be heard from the direction of the speaker's seat. It is controlled and synthesized with audio sound and output from each speaker.

However, the following problems are expected.
(A) The voice of the speaker that is loudly output from the speaker returns to the microphone again, and there is a possibility that howling will occur if the voice is loudened.
(A) If the frequency characteristic of the system (loop gain of a loudspeaker system including a microphone-speaker) is flattened so as to avoid howling, reverberation may occur and eventually the clarity of the conversation may be hindered.
(C) Since the level of the audio sound is changed according to the level of conversation, there is a possibility that a sense of incongruity may occur as a listening feeling.
(D) Since the gain of the speaker at the speaker's position is increased, a large feedback sound enters the microphone of the own channel, and it becomes easier to perform howling (in order to prevent this, the gain of the speaker of the farthest seat is increased) If this is raised, the sound image localization may be out of the position of the speaker and become an unnatural loud sound.
(E) Since vehicle interior noise is picked up by the vehicle interior microphone, it may be mixed with conversational sound and difficult to separate, or a new feedback loop may be formed between the speakers and howling may be facilitated.
JP 2001-105989 A Japanese Patent Application Laid-Open No. 2002-05392 Japanese Patent Laid-Open No. 07-255099 JP 2006-109340 A Japanese Patent No. 3158468

An object of the present invention is to provide an in-vehicle conversation support system that can provide a conversational clarity / intelligibility equivalent to that when the vehicle is stopped even when traveling by a microphone and a speaker disposed in the vehicle.
Still another object of the present invention is to save on-vehicle equipment by sharing an amplifier and a speaker in combination with other devices other than the present system, and facilitate connectivity of audio signals with other devices. It is to provide an in-car conversation support system.
Still another object of the present invention is to provide an in-vehicle conversation support system capable of simultaneously recording and reproducing a camera image for imaging outside / inside the vehicle and in-vehicle sound obtained by the microphone.
Still another object of the present invention is to provide an in-car conversation support system capable of instructing a specific part of a camera image for imaging outside / inside the vehicle and performing image processing.
Still another object of the present invention is to have a function of tagging in the time axis direction as additional information of images and sounds during recording / playback of camera images and sounds. The object is to provide an in-car conversation support system that is reproducible.
Still another object of the present invention is to provide an in-vehicle conversation support system that can search a tag by voice recognition of a switch or a specific word limited to each passenger seat and can be viewed on a display device for each passenger seat. is there.

In order to achieve the object, an in-car conversation support system according to claim 1 according to the present invention comprises:
An in-car conversation support system in an in-car environment where a speaker listens to live sound from other speakers and loud sound from a speaker,
A plurality of individual microphones or common microphones arranged in the vehicle;
A common speaker or a plurality of individual speakers arranged in the vehicle;
A correction unit including a filter unit for preventing howling and a delay unit for effect of preceding sound, and an adjustment unit, wherein the output of the individual microphone or the common microphone is corrected by the correction unit and amplified by the adjustment unit, and the common speaker or And an acoustic processing unit connected to the individual speakers.

The in-vehicle conversation support system according to claim 2 according to the present invention comprises:
An in-car conversation support system in an in-car environment where a speaker listens to live sound from other speakers and loud sound from a speaker,
A plurality of in-car microphones arranged for each speaker in the car,
A common speaker installed in the car for each speaker,
A sound that includes a correction unit and an adjustment unit including a filter unit for preventing howling and a delay unit for preceding sound effect, corrects the output of the in-vehicle microphone by the correction unit, amplifies the adjustment unit, and connects to the common speaker And a processing unit.

The in-car conversation support system according to claim 3 according to the present invention comprises:
An in-car conversation support system in an in-car environment where a speaker listens to live sound from other speakers and loud sound from a speaker,
A plurality of individual speakers arranged in the vehicle;
A common microphone that is shared by each speaker in the car,
A sound that includes a correction unit and an adjustment unit including a filter unit for preventing howling and a delay unit for the preceding sound effect, and corrects the output of the common microphone by the correction unit, and amplifies and adjusts the output from the adjustment unit to the individual speaker. And a processing unit.

The in-vehicle conversation support system according to claim 4 according to the present invention is the system according to claim 2,
The common speaker is a ceiling speaker arranged on a ceiling in a vehicle,
The plurality of individual microphones are in-vehicle microphones arranged in relation to a speaker's seat,
The filter of the filter unit for preventing howling is a FIR having an acoustic transfer characteristic between a point on a closed curved surface surrounding the ceiling speaker and the arbitrary in-vehicle microphone, or an IIR filter,
The filter unit is arranged in a feedback cancellation loop in the acoustic processing unit.
The in-vehicle conversation support system according to claim 5 of the present invention is the system according to claim 3,
The common microphone is a ceiling microphone arranged on the ceiling in the vehicle,
The plurality of individual speakers are in-vehicle speakers arranged in relation to a speaker's seat;
The filter of the filter unit for preventing howling is a FIR having a sound transfer characteristic between a point on a closed curved surface surrounding the ceiling microphone and the arbitrary speaker in the vehicle, or an IIR filter,
The microphone output is filtered and added for each in-vehicle speaker channel.
The in-car conversation support system according to claim 6 of the present invention is the system according to claim 4 or 5,
The filter characteristics of the acoustic processing unit are set in advance by a parameter setting unit based on acoustic transmission characteristics between the individual microphones or individual speakers and points on the corresponding closed curved surface, or measured values of impulse responses.

The in-car conversation support system according to claim 7 of the present invention is the system according to any one of claims 1 to 6, further comprising a running state detection unit for detecting or estimating a driving environment based on data from a microphone outside the vehicle or an in-vehicle computer for driving control. Including
The detection unit is configured to control the adjustment unit of the acoustic processing unit.
The in-car conversation support system according to claim 8 of the present invention is the in-car conversation support system according to claims 1 to 7, further comprising an audio signal input terminal for inputting an audio signal to the amplifier of the adjustment unit of the acoustic processing unit, An audio signal external to the system can be connected.
In the combination of this system and another device (for example, a vehicle-mounted stereo device), the on-vehicle equipment can be saved by sharing the amplifier and the speaker, and the connectivity of audio signals with other devices can be facilitated. Furthermore, by providing the output terminal of the microphone amplifier, the connectivity of the audio signal with other devices (for example, a device capable of recording / reproducing the audio signal) other than this system can be made easier. In addition, by connecting the audio signal input and output of the communication device to the microphone amplifier output terminal and the audio signal input terminal of the system, it is possible to easily support in-car conversation between passengers of the accompanying vehicle. it can.

The in-car conversation support system according to claim 9 of the present invention is the system according to claims 1 to 8, further comprising two or more cameras for acquiring an omnidirectional image outside the car and an image inside the car,
A recording / reproducing apparatus capable of simultaneously recording and reproducing the output of the microphone, the output of the camera, and a related audio signal;
Instruction means for selecting recording / reproduction of the recording / reproducing apparatus;
A plurality of display devices for displaying the image output of the recording / reproducing device;
It is composed of
The related audio signal is configured to be connected to the audio signal input terminal directly or after reproduction.
The related audio signal is a signal in an audio band related during a conversation or camera image acquisition input from the outside of the system.
The in-vehicle conversation support system according to claim 10 according to the present invention is the system according to claim 9, further comprising an image processing device that performs image processing of the camera,
Instruction means for instructing which part of the camera image is to be cut out and processed;
Image selection means for selecting image display before / after image processing;
Is provided.
The in-vehicle conversation support system according to claim 11 according to the present invention is the system according to claim 9 or 10, wherein the recording / reproducing apparatus can tag in the time axis direction as additional information of an image or sound during recording / reproduction. It has a function and has a function that can search the recorded tag in the time axis direction,
Furthermore, it has a function that allows playback from the searched tag position,
Tagging search instructing means for instructing the recording / reproducing apparatus to instruct tagging or instructing to search for a recorded tag is provided.
A tag in this system is additional information related to an image / sound, and can be recorded / reproduced in the time axis direction in the same way as an image / sound. Tagging is recording an instructed tag.

The in-car conversation support system according to claim 12 of the present invention is the system according to claim 11,
The tagging search instruction means is a voice recognition device that recognizes a specific word or a plurality of limited dedicated switches, provided for each passenger seat,
The display device is configured to be provided for each passenger seat.
According to a thirteenth aspect of the present invention, there is provided an in-car conversation support system according to any one of the ninth to twelfth aspects of the present invention, in which an image / sound is recorded / reproduced before the time when the instruction is given by an instruction from a dedicated instruction means. It is configured to have a function that can be played back in a specific time range with a time delay (rewind playback at a specific time).
The in-vehicle conversation support system according to claim 14 according to the present invention is the system according to any one of claims 9 to 13, further comprising instruction means for selecting an image,
And an image composition conversion device for synthesizing and converting a plurality of images selected from the audio / image output of the recording / reproducing device into images for divided display.

According to the present invention, the sound quality and level of the loud sound and the reproduced sound can always be maintained comfortably and optimally by predicting the occurrence of noise in the vehicle interior and automatically adjusting the sound processing unit. Furthermore, since the feedback from the speaker to the microphone itself is removed or reduced, howling is prevented and stable sound amplification is possible, and speech intelligibility is not hindered by reverberation extension. Moreover, since the raw sound is preceded at all seats by inserting a delay into the loudspeaker system, the localization of the speaker can be clearly maintained at all the occupant positions even in the simultaneous utterance of a plurality of speakers. If the reproduction sound of a car stereo or BGM in the vehicle is processed as the above-mentioned noise, a comfortable conversation is possible even under joint listening such as BGM.
Furthermore, by providing a terminal for inputting an audio signal outside the system, the power amplifier and the speaker for the audio signal can be shared between the system and other devices, and in-vehicle equipment can be saved. . Also, the audio signal connection can be facilitated.
Furthermore, by enabling simultaneous recording and reproduction of the in-vehicle sound obtained from the microphone and the camera images inside and outside the vehicle, it is possible to provide an in-car conversation support system that can be used as a sound / image information source for content creation during touring. . Car interiors that can display images without a sense of incongruity by performing image processing such as distortion correction, perspective projection conversion, and dynamic range compression on specific parts of camera images, including ultra-wide-angle cameras, omnidirectional cameras, and wide dynamic range cameras A conversation support system can be provided.
Furthermore, by providing a tagging search function in the time axis direction, it is possible to provide an in-car conversation support system that enables confirmation of images / conversations by searching for a place to be seen and heard and reproducing from that position.
Furthermore, the tag search means and the display device provided for each passenger seat can confirm the image in front of the user, and can share information and enjoy a conversation.

  As the in-vehicle conversation support system according to the present invention, the inventors of the present invention plan two in-vehicle environments. One of them is a vehicle interior environment (vehicle interior environment A) in which a common ceiling speaker is provided in a vehicle interior (driver's seat, passenger seat, passenger seat) space, and a plurality of microphones are individually disposed in the passenger seat, etc., as a first embodiment. , Shown in FIG. 3A. On the other hand, FIG. 2B and FIG. 3B show a second embodiment of an in-vehicle environment (in-vehicle environment B) in which a common ceiling microphone and individual speakers are arranged in the in-vehicle space. In each figure, “Mic” indicates “microphone”, and “SP” indicates “speaker”. 3A and 3B is a basic form of the embodiment in which the circuit in the one-dot chain line frame is removed, and the extended form is obtained by adding a circuit in the one-dot chain line frame to this basic form.

FIG. 1 is a conceptual diagram of an in-vehicle conversation support system according to the present invention, and shows an in-vehicle environment A. The in-vehicle conversation support system according to the present invention is an in-vehicle conversation support system in an in-vehicle environment in which a speaker listens to a live sound produced by another speaker and a loud sound from a speaker.
In an in-vehicle environment A composed of a plurality of individual microphones arranged in the vehicle and a common speaker arranged in the vehicle, including a correction unit and an adjustment unit consisting of a filter unit for preventing howling and a delay unit for preceding sound effect The acoustic processing unit is provided, the individual microphone output is corrected by the correction unit by the acoustic processing unit, amplified by the adjustment unit, and connected to the common speaker. As a result, howling is prevented for a speaker in the vehicle, and a loud sound of another speaker localized in the direct sound (raw sound) direction can be heard.

In the in-vehicle environment A of the first embodiment, a plurality of individual microphones are arranged on the head of an occupant (speaker), that is, on the headrest of the seat or behind the front seat.
The acoustic transfer characteristics of the propagation path from the periphery of one speaker installed near the ceiling to each individual microphone are simulated by a FIR (Finite Impulse Response) filter, and the speaker installed near the ceiling wraps around the individual microphone. By canceling the sound, the system is configured to amplify and support in-car conversation with stable and natural localization without howling regardless of temperature / humidity changes in the car or movement of passengers. Not only the FIR (Finite Impulse Response) filter but also the IIR (Infinite Impulse Response) filter can be used in the same manner.
In the in-vehicle environment B of the second embodiment, a plurality of speakers for an existing car stereo are used, and an acoustic transfer characteristic of a propagation path from the plurality of speakers to the periphery of one microphone installed on the ceiling surface is expressed by FIR (Finite Impulse Response) The system can be configured to obtain the same effect by simulating with a filter and canceling the sound that circulates from a plurality of speakers for an existing car stereo to a microphone installed on the ceiling surface.

Hereinafter, an embodiment of an in-vehicle environment A of an in-vehicle conversation support system according to the present invention will be described with reference to the drawings.
FIG. 2A is a block diagram of an in-vehicle environment A (first embodiment) when a common ceiling speaker and an individual microphone are arranged in the vehicle.
FIG. 3A is a diagram simultaneously showing the arrangement in the vehicle and the circuit configuration of the sound processing unit in the first embodiment of the in-vehicle conversation support system according to the present invention, and will be described below in a basic form.
An individual microphone is prepared for each seat, and each microphone is installed at the headrest or an appropriate location near the front of each occupant (the back of the front seat, the steering wheel, the dashboard, etc.).
Each individual microphone (as described later) is turned on only when the occupant is seated, or the FIR filter parameters are changed according to the vacant seat / seat.

A common speaker is installed almost at the center of the interior ceiling of the vehicle, and a signal that has been subjected to filter processing / delay processing and power amplification in the adjustment unit is connected to the common speaker by the correction unit of the sound processing unit and emitted as sound. .
A filter (FIR filter) is inserted in each microphone channel of the acoustic processing unit.
In general, the speaker is placed at n points on the circumference of a radius of about r = λ / 2 with the speaker as the center, the impulse response r (x) is measured at each microphone point, and the parameters of the FIR filter are about −r ( x). Here, λ is a wavelength with respect to the center frequency f 1 of the voice main band.
The audio signals of each channel processed by the filter (FIR filter) are added to become one and output from the ceiling speaker via a delay circuit, a frequency equalizer, a gain adjustment unit, and a power amplifier. The delay time τ of the delay circuit is set so as to be further delayed by 5 to 20 (msec) (Δt) with respect to the time L / c with respect to the distance L between the farthest passenger seats (τ = L / c + Δt) . Where c is the speed of sound in the air.

A driving state that collects information such as vehicle speed, engine speed, and engine load from an in-vehicle computer (in-vehicle PC), predicts noise generated by the vehicle, and generates a signal that optimally sets the frequency equalizer and gain adjustment unit from moment to moment A detection unit (or noise detection control unit) is provided.
In addition, if necessary, noise outside the vehicle, particularly when the vehicle travels on various roads, is picked up from the tire house, etc., and in-vehicle noise is predicted from previously measured sound insulation data. The final control signal is output by combining the noise calculated from the engine speed and engine load. In addition, another microphone for detecting the final vehicle interior sound is not arranged in the vehicle.
A seat switch (seat SW) that detects the presence or absence of an occupant is attached to the seat, and the microphone is turned ON / OFF according to the presence or absence of the occupant (when the seat is empty, the microphone channel is turned OFF), or the FIR of each channel Prepare and change two sets of filter parameters. Control only those channels that are affected by the presence or absence of passengers.

  A frequency correction device (equalizer; EQ) and a gain adjustment volume (Vol) are connected in cascade to the output system of the acoustic processing unit of the in-car conversation support system according to the present invention, and the driving conditions (vehicle speed, load, engine speed) Etc.) and sound quality and level are automatically controlled (or automatically selected) so that conversation can be heard optimally according to external noise. In addition, in the loudspeaker system that is configured to radiate downward from the ceiling speaker to the audience seat through the delay circuit and power amplifier after adding the outputs of each system (FIR processing results), the voice of the speaker is always heard at all seats. It reaches first, and is further configured to reach a loud sound from the speaker with a delay of several msec to several tens of msec, and a sound image localization is secured at the position of the speaker.

The operation of the in-vehicle conversation support system in the in-vehicle environment A is as follows.
First, when the vehicle is stopped, there is little noise or the engine is idling, so the gain of this system is at a minimum or nearly zero.
When the vehicle is accelerated from this state toward a certain vehicle speed, the in-vehicle noise generally increases with acceleration, the vehicle speed, or the engine load, and listening to in-vehicle conversation (intelligibility or intelligibility) by a plurality of passengers decreases. This system minimizes this effect and helps maintain the comfort of in-car conversation.

The voice produced by the occupant is picked up by the microphone in each seat and enters the FIR filter through the microphone amplifier. Each FIR filter simulates the acoustic transfer characteristics of the propagation path from the speaker periphery (on the circumference of λ / 2 around the speaker) to each individual microphone, and wraps around the individual microphone from the speaker installed near the ceiling Since the parameters are set so as to cancel the sound, even if the loudness level is increased according to the vehicle speed and load, howling is not to some extent.
That is, as shown in Fig. 2A, if h _i (x) is the impulse response between microphone (i) and point (i) on the closed surface, and -h ' _i (x) is the impulse response of FIR-i, The parameters of the FIR filter (FIR-i) are set so that h ' _i (x) ≒ h _i (x), so by canceling the sound that circulates from the speakers installed near the ceiling to individual microphones Prevent howling.
The outputs of each FIR are added to enter a delay circuit, delayed by τ _i (msec), and sent to the equalizer. The value of τ _i is set so that the raw sound of the speaker first reaches all seats and is output from the speaker with a delay of several to several tens of msec (actually 5 to 20 msec). By this processing, natural localization of the sound image is always obtained in the direction of the speaker at all seats. Since each τ _i is substantially equal, it can be shared and placed in the P portion as τ.

In addition, the loudness level and equalizer settings at each point in time are extracted from the in-vehicle PC, such as the vehicle speed, load, and engine speed, and data such as vehicle exterior noise is measured as necessary. The vehicle interior noise is predicted and the necessary control amount is calculated, so that the optimum control is automatically performed every moment (for example, the vicinity of 4 kHz contributing to the intelligibility is emphasized).
By simulating the acoustic transfer characteristics of the propagation path from multiple points surrounding the speaker to individual microphones with a filter, the sound that wraps around the individual microphones from the speakers installed near the ceiling is canceled out. There is no such thing as howling with changes in temperature and humidity. However, if there is no occupant located immediately in front of the microphone (the seat is vacant), the effect is large, so the presence or absence of the occupant is detected with a seat switch (seat SW), and only the FIR parameter of the corresponding channel is vacant. The parameter can be automatically changed so that the predetermined position becomes the insensitive area.

According to the above-described embodiment, a natural loud sound of approximately 6 to 10 (dB) can be obtained when traveling based on the time when the vehicle is stopped. By simulating the acoustic transfer characteristics of the propagation path from multiple points surrounding the speaker to individual microphones using a filter, the sound that circulates from the speakers installed near the ceiling to the individual microphones is canceled, so the reverberation is extended by loud sound. Therefore, the speaker is placed at the center of a plurality of points surrounding it without any alienation of clarity and sound quality, so that it can operate stably even when the passenger moves or changes in temperature and humidity. The loudness level and sound quality are optimally and automatically controlled according to driving conditions and external noise.
In addition, since the loud sound from the speaker is delayed so as to arrive after the raw sound of the speaker is reached at all seats, the sound image is always localized at the position of the speaker, and natural conversation is promoted.

Next, referring to the drawings and the like, an embodiment of an in-vehicle environment B of the in-vehicle conversation support system according to the present invention will be described.
FIG. 2B is a block diagram of an in-vehicle environment B (second embodiment) when a common ceiling microphone is disposed in the vehicle and an individual speaker is disposed in the seat.
FIG. 3B is a diagram simultaneously showing the arrangement in the vehicle and the circuit configuration of the sound processing unit in the second embodiment of the in-vehicle conversation support system according to the present invention, and will be described below in a basic form.
As shown in FIG. 2B and FIG. 3B, one common microphone is installed at the approximate center of the interior ceiling surface. In this embodiment, a speaker is provided for each seat. In practice, the use of a plurality of speakers provided on the ceiling does not prevent the use of a plurality of speakers that do not accurately correspond to the seat. In other words, a car stereo can be used as a speaker, but an explanation will be given with reference to an example in which a speaker is arranged for a seat.

The sound of the passengers in each seat is picked up by the ceiling microphone with almost equal sensitivity, distributed to the FIR filter corresponding to each speaker output through the microphone amplifier, gain adjuster, equalizer, and delay circuit at the same rate for signal processing. Then, each speaker radiates indoors through a power amplifier. The configuration of the delay circuit will be described with reference to FIG. 2B. In the figure, h _i (x) represents the impulse response between the speaker (i) and the point (i) on the closed surface. Here, h ′ _i (x) ≈h _i (x). The localization control delay τ _i of each input system is set so that the live sound precedes the loud sound at all seats. Since each τ _i is substantially equal, it can be shared and placed in the P portion as τ.
Each FIR is parameter-set to simulate the acoustic transfer characteristics of the propagation path from each speaker to the periphery of one microphone installed on the ceiling. As in the previous example using a plurality of microphones, each FIR parameter is automatically changed or corrected by detecting the presence or absence of an occupant by the seat SW. The same is true in that the gain and equalizer are optimally adjusted according to the noise inside and outside the vehicle detected by the noise detection control unit.

Each FIR filter parameter can be determined by the LES method (Least Squares Error). In short, the acoustic transfer characteristics of the propagation path from multiple points to the individual microphones around the ceiling microphone (speaker in the above-mentioned interior environment A) (for example, the circumference of radius λ / 2) are simulated by the FIR filter. By canceling the sound that circulates from the speaker installed in the individual microphone to the individual microphone, it is possible to operate stably against slight changes in indoor acoustic conditions (occupant movement and temperature / humidity change).
Therefore, the FIR filter is basically a fixed one (the number of elements is small and an inexpensive one can be used).

(Parameter Setting Unit) Here, the parameter setting of the filter common to the above-described in-vehicle environment A will be described. Either of the following two can be used.
A) When using an all-sample convolution device for the filter, place the speaker actually used at the point (i) on the closed surface and measure the impulse response h _i (x) between the two points with the corresponding microphone (i). The sign is inverted and used as it is as the impulse response -h ' _i (x) ≒ -h _i (x) of the filter (i). Sequentially, each system is adjusted in the same procedure as i = 1, 2, 3,..., And the operation is started by returning the speaker to the center of the point on each closed curved surface.

A) When using an adaptive FIR filter (finite number of taps / discrete) as the filter, place the actual speaker at point (i) on the closed surface and output pink noise. Only the output of the corresponding microphone (i) and the output of the filter (adaptive FIR filter) (i) are added by the signal adder (at this time, other systems are off). The output of the signal adder at that time is input to the error detection input (the adaptive FIR filter has two inputs and one output) of the adaptive FIR filter, and is adaptively operated. When the output of the signal adder is minimized and converges, parameter identification is completed.
In this state, the adaptive operation is stopped and it is used as a fixed FIR filter with the identified parameters. Sequentially, each system is adjusted in the same procedure as i = 1, 2, 3,..., And the operation is started by returning the speaker to the center of the point on each closed curved surface.
In both cases, the impulse response h _i (x) changes with the passage of time, but the points on each closed curved surface are originally set off the center of the speaker, and are intended to exhibit the effect as a whole. Thus, individual temporal variations do not have a significant effect on system stability.

Also in the form of the in-vehicle environment B of the in-vehicle conversation support system according to the present invention, the same operation and effect as in the above-described in-vehicle environment A can be obtained.
In the basic embodiment described above, the circuit in the one-dot chain line frame in FIG. 3A and FIG. 3B is added to the extended embodiment, and as shown in the figure, the power amplifier of the adjustment unit of the acoustic processing unit By adding a circuit from the audio signal terminal to the signal adder at the input of the power amplifier on the input side (by providing the audio signal input terminal in advance from the viewpoint of the user), this system and other devices (for example, It is possible to share the power amplifier and the speaker in combination with the in-vehicle stereo apparatus), and the in-vehicle equipment can be saved. In addition, it is possible to easily input an audio signal (external audio signal viewed from the present system) of an in-vehicle radio, a television, a stereo device, or a portable music player so that the sound can be amplified by a speaker. In FIG. 3B, an external audio signal is input as a monaural signal. In the case of inputting a stereo signal or a multi-channel audio signal such as four channels, it can be realized by providing an additional circuit corresponding to each channel on the input side of each corresponding power amplifier.
In order to adjust the balance between the loudness level on the conversation support side and the loudness level of the audio signal input from the outside, the volume of the external audio signal input from the audio signal input terminal is adjusted. Note that if there is a volume adjustment mechanism for audio output in other devices, the volume adjustment mechanism immediately after the audio signal input terminal on this system side can be eliminated (in this case, the power amplifier is substantially The same as connecting an external audio signal to the input).

The in-car conversation support system according to the present invention performs image processing and storage in parallel in addition to the above-described voice.
FIG. 4 shows an example in which a plurality of video cameras are installed outside and inside the vehicle. When an all-around camera equipped with a fisheye camera and a hyperboloid mirror is used for the in-vehicle cameras 101 to 104 and the in-vehicle cameras 105 to 106 in FIGS. 5A and 5B, an omnidirectional image can be easily obtained. The number of installations can be reduced. In the example of the camera arrangement in the figure, the effective shooting range is indicated by an arrow range with respect to the cameras 101 to 106. When using an ultra-wide-angle camera or a normal camera as the outside camera, it is necessary to install a larger number of cameras (to obtain a camera image as part of the omnidirectional image) in order to capture the entire periphery of the vehicle without omission. . The in-vehicle camera 105 is installed so as to photograph a person sitting in the driver's seat and the passenger seat, and the in-vehicle camera 106 is installed so as to photograph a person sitting in the rear seat. If a wide-angle camera is used for the in-vehicle camera, a wide range can be imaged, and a landscape through the window can be imaged in addition to the passenger.

In addition, if a wide dynamic range camera is used as the camera, an image with less overexposure and underexposure can be taken compared to a normal camera, but the display device cannot be displayed as it is because the dynamic range of the display device is narrow. The captured image can be displayed on the display device by performing dynamic range compression (image processing).
In the case of an omnidirectional camera equipped with a super-wide-angle camera, a fish-eye camera, or a hyperboloidal mirror, the captured image is a distorted image as it is, so it is very difficult to see and it is difficult to grasp what is reflected. Therefore, by performing image processing such as distortion correction and perspective projection conversion and displaying an image taken with a normal camera, it is possible to eliminate a sense of incongruity when viewed.

FIG. 5 shows a block diagram of an additional device portion that can acquire a moving image of an outside / in-vehicle camera in the in-vehicle conversation support system, and record and reproduce it simultaneously with an audio signal. FIG. 6 shows an image processing device and an image composition conversion device. The block diagram of an internal schematic structure is shown.
In FIG. 5, moving image signals from the in-vehicle and out-of-vehicle cameras are input to the image processing device, subjected to image processing, and output to the image display device. The image display device can be configured by a device that displays an image, such as a liquid crystal display device.

An image processing instruction circuit that is an instruction means for instructing which part of a camera image is to be cut out and processed, an image selection circuit that is an image selection means for selecting an image display before / after image processing, and recording / reproduction of a recording / reproducing apparatus The recording / reproducing instruction circuit, which is an instruction means for selecting a signal, can be configured by a switch, and signals corresponding to the switch instructed by the operator as an image processing instruction signal, an image selection signal, and a recording / reproduction instruction signal, respectively. Output to the image processing device.
The recording / playback instruction signal is also output to the recording / playback apparatus.
The recording / reproducing apparatus can be composed of a magnetic tape recording / reproducing apparatus, a hard disk recording / reproducing apparatus, or a memory recording / reproducing apparatus. By using a plurality of these, a recording / reproducing apparatus having a multi-channel structure for both audio signals and image signals can be obtained, and a recording / reproducing / stopping function is executed by a recording / reproducing instruction signal. The image input / output unit of the recording / reproducing apparatus inputs a recorded image signal from the image processing apparatus and outputs the reproduced image signal to the image processing apparatus. The audio input / output unit of the recording / reproducing apparatus inputs the audio signal from the microphone amplifier output of FIGS. 3A and 3B and the related audio signal input to the audio signal input terminal 1 as a recording audio signal (mixed and recorded audio). The reproduced audio signal is output to the input of the power amplifier shown in FIGS. 3A and 3B or to the audio signal input terminal 2.
The audio signal from the individually recorded microphone amplifier output and the reproduced audio signal of the related audio signal are mixed and output, or the reproduced audio signal recorded after mixing is output. In the expanded embodiment of FIGS. 3A and 3B, one audio signal input terminal is provided. However, two circuits within a one-dot chain line frame are prepared, and a related audio signal is supplied to one audio signal input terminal 1. By inputting the reproduced audio signal of the recording / reproducing apparatus to the other audio signal input terminal 2 and adjusting the volume, the related audio signal and the reproduced audio signal can be easily amplified from the speaker.

As shown in FIG. 5, the tagging search instruction device, which is a tagging search instruction means, includes a tagging search instruction circuit and a voice tagging search instruction unit. The tagging search instruction circuit can be constituted by a plurality of limited switches, and outputs tagging search instruction information to the recording / reproducing apparatus. The voice tagging search instruction unit can be constituted by a voice recognition device and a voice input unit for this device. The voice input unit includes a plurality of voice signals from the microphone amplifier output, a plurality of dedicated microphones, and the output, and inputs the voice signals to the voice recognition device. The voice recognition apparatus recognizes a specific word limited from the input voice signal as a voice tag for recording / searching, and outputs tagging search instruction information to the recording / reproducing apparatus.
In the recording / playback device, tagging is performed by tagging instruction information during recording or playback of images or sound (tag information is recorded), or tag search is performed by tag search instruction information during playback or stop. Playback and execution of recorded images and sound from the result position.

In FIG. 6, the image signal from the camera is output as a recorded image signal to the reproduction / recording apparatus and also to the image signal selection circuit. The other input of this circuit is a reproduced image signal inputted from the recording / reproducing apparatus. In the same circuit, the image signal from the camera or the reproduced image signal is selected by the recording / reproducing instruction signal and output to the subsequent stage. There are selection circuits for the number of cameras (two-dot chain line portion in the figure).
The selected image signal is subjected to image processing for each camera image in accordance with the type of camera in order to make the display easy to see. An image processing process suitable for the camera is selected by the image processing instruction signal. This selection is performed at the time of camera installation since it is possible to know what kind of camera is installed at the time of camera installation.

The image processing unit includes through processing that does not process anything, distortion correction processing that corrects distortion of normal camera images and wide-angle camera images, perspective projection conversion processing of fish-eye camera images and camera images using hyperboloid mirrors, wide dynamics There is a dynamic range compression process for range camera images.
The image obtained with a fisheye camera or an omnidirectional camera with a hyperboloidal mirror is a circular image, and it is difficult to understand as it is. It is displayed as an image as if it were done. The image processing instruction signal indicates an image portion to be subjected to perspective projection conversion. For example, an image portion to be cut out and displayed by perspective projection can be specified by parameters of a vertical angle with respect to a horizontal plane, a direction angle on the horizontal plane, and a zoom level, and can be subjected to a perspective projection conversion process.

In the dynamic range compression processing, methods such as table conversion processing, calculation processing using line transformation, logarithmic conversion, wavelet conversion, and a combination of these are used so that the dynamic range is within the display dynamic range. .
Depending on the combination of the camera and the lens, the processing contents of the image processing unit can be combined.
For example, when a fisheye lens is attached to a wide dynamic range camera, the image processing process is a perspective projection conversion process and a dynamic range compression process.
In the image selection circuit, the image processing instruction signal is connected to the output of the same image processing process connected in the image processing selection circuit, and the image data of the through process and the processed image data are processed before the image processing by the instruction of the image selection signal. / Selection is made as a later image.

Image processing from the image processing selection circuit to the image selection circuit is performed for each camera image (the portion within the one-dot chain line frame in FIG. 6) and input to the display image selection circuit and the image composition conversion device.
The display image selection circuit includes a circuit unit that selects an image of the camera, and a circuit unit that selects a selected camera image and an image from the image composition conversion device, and the image selected by the image selection signal is displayed as a display image. Output to the display device.

The display image selection circuit is a circuit that selects an image instructed by an image selection signal, and can be configured by a switch.
The image composition conversion device is a device that synthesizes images for displaying a plurality of images simultaneously on one screen, and includes a divided display image selection circuit, an image size conversion circuit, and an image composition circuit. The instruction means for selecting an image to be added in order to select a plurality of images can be constituted by an addition switch or the like, and is added in the image selection circuit. An instruction including the additional switch is output from the circuit unit as an image selection signal.

The image input to the image composition conversion device is input to the split display image selection circuit and output to the image size conversion circuit. The divided display image selection circuit is a circuit for selecting an image instructed by an image selection signal (the number of images simultaneously displayed on one screen).
The image input to the image size conversion circuit is converted in image size and output to the image composition circuit. The image size conversion circuit is a circuit that generates a reduced image by converting the image size in order to display a plurality of images simultaneously on one screen, so that the image size can be converted using a thinning circuit or an interpolation circuit. Can be configured. For example, if the pixels are thinned every other pixel both horizontally and vertically, a 1/4 size image can be created, and four images can be displayed on one display screen.

The reduced image input to the image composition circuit synthesizes a predetermined number of images with a display image for one screen and outputs it to the display image selection circuit. The image synthesizing circuit is a circuit that collects input reduced images for one screen and synthesizes and converts the images as display images.
Further, in FIGS. 5 and 6, the display image output of the image processing apparatus is input to the recording / reproducing apparatus as a recorded image signal and recorded, and is input to the image processing apparatus as a reproduced image signal during reproduction and then input to the display image selection circuit. The display image recorded during playback can be selected and displayed. By doing in this way, it can be understood later which image was displayed and viewed.

It should be noted that the tagging search instruction device and the image display device are installed for each passenger seat, and the tagging search instruction device can also be configured in an integrated form with the display device. In addition, the tagging search instruction device and the image display device provided in the driver's seat are installed at a position that does not hinder the driver's driving. In this case, the tagging search instruction device may be configured to be incorporated in the steering wheel in the form of a limited number of switches or a dedicated microphone.
In this way, passengers other than the driver can easily instruct tagging with a tagging search instruction device such as a switch or voice at hand while talking with other people while viewing a plurality of images. On the other hand, the driver cannot gaze at the image, but can participate in the conversation, and it is easy with a few search switches and voice tagging search instruction devices provided at hand such as the steering wheel. You can instruct tagging.

It is possible to support and excite the conversation by rewinding and playing back a little (a short period of time) a conversation that has been missed or a conversation that has been missed. This function can be added in the recording / reproducing apparatus. Audio / image signals for recording / playback are shared in the recording / playback apparatus. The dedicated instruction means can be constituted by a switch, and is shown by a dedicated switch in FIG.
FIG. 7 shows a diagram for explaining the short-time rewind playback function. In the figure, a recording / reproducing ring buffer indicates a recording area for always recording audio / images within a predetermined time, and is configured as a ring buffer. Also, the arrow at the lower end of the ring buffer in the figure indicates the direction in which the recording area is used when recording in the buffer.

The recording / reproducing ring buffer has two areas A and B, and starts recording audio and images simultaneously when the power is turned on. The length of each ring buffer is set longer than the playback time of a short fixed time.
Initially, the buffer A side is used, and when the dedicated switch is pressed (set to buffer position t1), recording of the buffer A is stopped. In order to reproduce sound and images from a short period of time, for example, K seconds ago, reproduction of the buffer A is started with the buffer position t1-K as the reproduction start position K seconds ago, and after reproduction for K seconds (to the buffer position t1). (When it comes) Stop playback and resume recording. Buffers A and B are used alternately so that the first instruction uses the buffer A side and the next instruction uses buffer B. While one buffer is being played back, the other buffer is recorded as it is, so there is no non-recording period. In the drawing, since recording is continued on the buffer B side even during reproduction on the buffer A side, when the buffer A side resumes recording at the buffer position t1, the recording position on the buffer B side has advanced to t1 + K. Therefore, during the rewind playback on the buffer A side, the recording area on the buffer B side is from the B side buffer position t1 to t1 + K.

In addition, in order to perform fast rewind reproduction according to an instruction from the dedicated switch, a recording / reproduction ring buffer having a small rewind time delay such as a memory storage device such as a storage memory or a memory card or a hard disk device is used.
The image / sound / tag information recorded in this way can be played back in the car or in the room to watch with the passengers for further conversation, and can be quickly tagged at the point where the tag is desired. . You can search for recorded tags by tag information and play back images and sounds from the tagged position. It can also be used to edit the recorded audio / image / tag information and produce playback content for viewing.

Furthermore, audio / image information sources that are recorded and tagged simultaneously with in-vehicle audio and camera images inside and outside the vehicle can be used for content production for image processing and editing processing, and can be used as viewing content for viewing. .
In addition, even when the image is simply played back, it is possible to enjoy the conversation while watching the image that was not seen as the driver with the accompanying person, and tagging can be performed during the playback. Furthermore, by searching for a tag, it can be played back and enjoyed from the tagged position.

  The tagging search instruction device, which is a tagging search instruction means, performs tagging as a scene of a so-called near-hat that feels dangerous while driving, and a beautiful landscape such as a stunning mountain range or a wave of waves. A scene is mentioned. For the former dangerous spot scene, when the location is approached at a later date or the location is included on the driving route, information is presented to the driver and passengers as content, creating a risk avoidance effect be able to. In addition, the latter scenes such as beautiful scenery can be retrospectively reproduced by playing the content of the relevant part at home after driving, producing an effect that soaks in memories, and by introducing the content to others, It is also possible to convey the charm of the place and increase its value as a tourist spot.

Hereinafter, a first embodiment of the tagging search instruction means used in the in-vehicle conversation support system according to the present invention will be described with reference to FIG. 8A.
FIG. 8A shows an example where the driver performs tagging. Since the tagging operation performed by the driver is required to be a low-load operation that does not affect the driving operation for safe driving, the tagging operation by the tagging search instruction device composed of a small number of switches is not required. Is appropriate. In this example, a small child jumped on the road following the ball, and the driver who felt danger recognized that this area was a place that included a risk factor that a small child would jump out. The switch of the switch box is pushed and the tag "dangerous point" is given. Even with such a simple switch, it is equivalent to marking only meaningful places in enormous lengths of recorded data, just by giving tags of the general types of “dangerous spots” and “beautiful scenery”. , Can contribute to efficient content creation.

Next, a second embodiment of the tagging search instruction means used in the in-car conversation support system according to the present invention will be described with reference to FIG. 8B.
FIG. 8B shows an example where a passenger other than the driver performs tagging. The tagging operation performed by the passenger can be designed with a higher degree of freedom than the operation performed by the driver. Tagging with a switch box can greatly increase the number of switches, and more detailed information can be given as tags. Tagging by voice input is also possible by recognizing a passenger's utterance recorded by a seat microphone mounted on the seat with a voice recognition device of a tagging search instruction device. In this case, by recognizing the utterances of the tag types of “dangerous part” and “beautiful scenery”, in addition to being able to give the same tag as the instruction by the switch, by dictating the utterance, the explanation of the part can be used as a tag of text information It is also possible to grant. By using this text information for matching with a search keyword when browsing the content, it is possible to contribute to the efficiency and reliability of the content search.

  The in-car conversation support system according to the present invention can basically be widely used in the fields of the automobile industry, transportation, and touring.

1 is a conceptual diagram of an in-car conversation support system (first embodiment) according to the present invention. It is a block diagram of the in-vehicle conversation support system (first embodiment) in the case of an in-vehicle environment (in-vehicle environment A) of a shared ceiling speaker and an individual microphone. It is a block diagram of the in-vehicle conversation support system (2nd Embodiment) in the case of the in-vehicle environment (in-vehicle environment B) of a common ceiling microphone and an individual speaker. It is the figure which showed simultaneously the schematic of the arrangement | positioning in a vehicle of 1st Embodiment of the in-vehicle conversation assistance system by this invention, and the circuit diagram of a sound processing part. It is the figure which showed simultaneously the schematic of the arrangement | positioning in a vehicle of 2nd Embodiment of the in-vehicle conversation assistance system by this invention, and the circuit diagram of a sound processing part. It is a top view which shows the example which arrange | positions the several camera utilized with the system by this invention. It is a block diagram which shows the example of the apparatus added in the system by this invention. It is a block diagram for demonstrating the image processing apparatus and image composition conversion apparatus which are used with the system by this invention. It is a figure for demonstrating the rewind reproduction | regeneration function of the short fixed time of the apparatus utilized with the system by this invention. It is a figure for demonstrating 1st Embodiment of the tagging search instruction | indication means utilized with the system by this invention. It is a figure for demonstrating 2nd Embodiment of the tagging search instruction | indication means utilized with the system by this invention.

Explanation of symbols

101-104 Outside camera 105-106 Inside camera

Claims (14)

An in-car conversation support system in an in-car environment where a speaker listens to live sound from other speakers and loud sound from a speaker,
A plurality of individual microphones or common microphones arranged in the vehicle;
A common speaker or a plurality of individual speakers arranged in the vehicle;
A correction unit including a filter unit for preventing howling and a delay unit for effect of preceding sound, and an adjustment unit, wherein the output of the individual microphone or the common microphone is corrected by the correction unit and amplified by the adjustment unit, and the common speaker or An acoustic processing unit connected to individual speakers;
An in-car conversation support system consisting of An in-car conversation support system in an in-car environment where a speaker listens to live sound from other speakers and loud sound from a speaker,
A plurality of in-car microphones arranged for each speaker in the car,
A common speaker installed in the car for each speaker,
A sound that includes a correction unit and an adjustment unit including a filter unit for preventing howling and a delay unit for preceding sound effect, corrects the output of the in-vehicle microphone by the correction unit, amplifies the adjustment unit, and connects to the common speaker A processing unit;
An in-car conversation support system consisting of An in-car conversation support system in an in-car environment where a speaker listens to live sound from other speakers and loud sound from a speaker,
A plurality of individual speakers arranged in the vehicle;
A common microphone that is shared by each speaker in the car,
A sound that includes a correction unit and an adjustment unit including a filter unit for preventing howling and a delay unit for the preceding sound effect, and corrects the output of the common microphone by the correction unit, and amplifies and adjusts the output from the adjustment unit to the individual speaker. A processing unit;
An in-car conversation support system consisting of The common speaker is a ceiling speaker arranged on a ceiling in a vehicle,
The plurality of individual microphones are in-vehicle microphones arranged in relation to a speaker's seat,
The filter of the filter unit for preventing howling is a FIR having an acoustic transfer characteristic between a point on a closed curved surface surrounding the ceiling speaker and the arbitrary in-vehicle microphone, or an IIR filter,
The filter unit is arranged in a feedback cancellation loop in the acoustic processing unit,
The in-vehicle conversation support system according to claim 2. The common microphone is a ceiling microphone arranged on the ceiling in the vehicle,
The plurality of individual speakers are in-vehicle speakers arranged in relation to a speaker's seat;
The filter of the filter unit for preventing howling is a FIR having a sound transfer characteristic between a point on a closed curved surface surrounding the ceiling microphone and the arbitrary speaker in the vehicle, or an IIR filter,
The microphone output is filtered and added for each in-vehicle speaker channel.
The in-vehicle conversation support system according to claim 3. The in-car conversation support system according to claim 4 or 5,
The in-vehicle conversation support system in which the filter characteristic of the acoustic processing unit is preset by the parameter setting unit based on the acoustic transmission characteristic between the individual microphone or individual speaker and a point on the corresponding closed curved surface, or the measured value of the impulse response. The in-vehicle conversation support system according to any one of claims 1 to 6 further includes a traveling state detection unit that detects or estimates a driving environment based on data from an external vehicle microphone or an in-vehicle computer for driving control.
An in-vehicle conversation support system in which the detection unit controls the adjustment unit of the acoustic processing unit. The in-vehicle conversation support system according to claim 1, further comprising an audio signal input terminal for inputting an audio signal to an amplifier of the adjustment unit of the acoustic processing unit, and enabling connection of an audio signal outside the system. Support system. The in-vehicle conversation support system according to claim 1, further comprising two or more cameras for acquiring an omnidirectional image outside the vehicle and an image inside the vehicle,
A recording / reproducing apparatus capable of simultaneously recording and reproducing the output of the microphone, the output of the camera, and a related audio signal;
Instruction means for selecting recording / reproduction of the recording / reproducing apparatus;
A plurality of display devices for displaying the image output of the recording / reproducing device;
It is composed of
The related audio signal is connected to the audio signal input terminal directly or after reproduction,
In-car conversation support system configured as described above. The in-vehicle conversation support system according to claim 9, further comprising: an image processing device that performs image processing of the camera;
Instruction means for instructing which part of the camera image is to be cut out and processed;
Image selection means for selecting image display before / after image processing;
An in-car conversation support system. 11. The in-vehicle conversation support system according to claim 9 or 10, wherein the recording / reproducing apparatus has a function capable of tagging in the time axis direction as additional information of an image or sound during recording / reproduction, and the recorded tag Has a function to search for the time axis direction,
Furthermore, it has a function that allows playback from the searched tag position,
Tagging search instruction means for instructing the recording / reproducing apparatus to instruct tagging or to search for a recorded tag;
An in-car conversation support system. The in-vehicle conversation support system according to claim 11,
The tagging search instruction means is a voice recognition device that recognizes a specific word or a plurality of limited dedicated switches, provided for each passenger seat,
The display device is provided for each passenger seat,
In-car conversation support system configured as described above. The in-vehicle conversation support system according to any one of claims 9 to 12, further comprising a specific instruction means that delays images / sounds recorded / reproduced before the designated time by a specific instruction by an instruction from a dedicated instruction means. In-car conversation support system configured to have a function that can be played back in a range. The in-car conversation support system according to any one of claims 9 to 13, further comprising an instruction means for selecting an image,
An image synthesizing / converting device for synthesizing and converting a plurality of images selected from the audio / image output of the recording / reproducing device into images for split display;
An in-car conversation support system with a configuration with

Images