JP2012025270A - Apparatus for controlling sound volume for vehicle, and program for the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a novel method for obtaining voice originating from a vehicle compartment separating from noise originating from a vehicle exterior, in a technique that it is determined whether an occupant utters on the basis of a comparison between the voice originating from a vehicle interior and an output voice of an on-board acoustic device to reduce a sound volume of the on-board acoustic device when determined that the occupant utters.SOLUTION: The apparatus includes: a vehicle parameter box 4e that corrects a frequency spectrum data of voice originating from the vehicle exterior, on the basis of a vehicle parameter box data indicating a damping factor per frequency for the frequency spectrum data of voice originating from the vehicle exterior in which a microphone 6 for the vehicle exterior has received, to estimate the frequency spectrum data when the voice originating from the vehicle exterior transmits into the vehicle compartment, attenuates, and is received by a microphone 5 for the vehicle interior; a difference detection unit 4f for removing an element of the frequency spectrum data of the voice originating from the vehicle exterior after corrected by the vehicle parameter box 4e from the frequency spectrum data of the voice of the vehicle interior in which the microphone 5 for the vehicle interior has received.

Description

  The present invention relates to a volume control device for a vehicle and a program for the volume control device.

  2. Description of the Related Art Conventionally, a technique for adjusting the volume of an in-vehicle acoustic device so as not to hinder the conversation of a vehicle occupant is known. Specifically, in Patent Document 1, it is determined whether or not the passenger is talking by comparing the voice acquired by the microphone in the vehicle and the output voice of the in-vehicle acoustic device, and determining that the conversation is being performed. In such a case, a technique for reducing the volume of the in-vehicle acoustic device is described.

  At this time, Patent Document 1 employs a directional microphone array that can switch the sound acquisition direction as a microphone in the vehicle in order to prevent noise derived from outside the vehicle from entering the voice acquired by the microphone in the vehicle. The direction of sound acquisition is set to the direction of the driver's seat or the front passenger seat using a directional microphone array.

JP 2002-171588 A

  However, in the method using the directional microphone array as in the cited document 1, it is difficult to sufficiently separate the voice derived from the vehicle interior from the noise derived from the outside of the vehicle.

  In view of the above points, the present invention determines whether or not an occupant is speaking based on a comparison between in-vehicle audio and output audio from an in-vehicle acoustic device. In the technology for reducing the volume, an object is to provide a new method of acquiring sound derived from the interior of a vehicle by separating it from noise derived from the outside of the vehicle.

  In order to achieve the above object, the invention according to claim 1 includes a speaker (1a to 1f) that outputs sound into a vehicle compartment, an acoustic device (3) that outputs an audio signal, and the acoustic device (3). An audio output unit (2) that amplifies an audio output audio signal that is an output audio signal and outputs the amplified audio signal to the speakers 1a to 1f, an in-vehicle microphone (5) that receives in-vehicle audio as in-vehicle audio, and outside the vehicle An on-vehicle microphone (6) that receives the generated outside-derived sound and a volume control device for a vehicle that outputs frequency spectrum data of the in-vehicle sound received by the on-board microphone (5). Voice spectrum output means (4a, 4b) and vehicle voice spectrum output means for outputting frequency spectrum data of the voice derived from outside the vehicle received by the vehicle microphone (6). c, 4d), and vehicle parameter box data indicating the attenuation rate for each frequency with respect to the frequency spectrum data of the outside-speech-derived sound output by the outside-speech spectrum output means (4c, 4d). Based on the parameter box data, by correcting the frequency spectrum data of the vehicle-derived sound output by the vehicle exterior sound spectrum output means (4c, 4d), the vehicle-derived sound is transmitted to the vehicle interior and attenuated to be used for the vehicle interior. The vehicle parameter box means (4e) for estimating the frequency spectrum data received by the microphone (5) and outputting the corrected frequency spectrum data, and the vehicle interior output by the vehicle interior sound spectrum output means (4a, 4b) From the frequency spectrum data of the voice, the vehicle parameters Difference detecting means (4f) for removing the component of the corrected frequency spectrum data of the outside-derived sound output from the box means (4e) and outputting the frequency spectrum data after the removal as frequency spectrum data of the in-vehicle derived sound. And frequency spectrum data of the sound output sound signal output by the sound device (3), the frequency spectrum data of the generated sound output sound, and the in-vehicle sound output by the difference detection means (4f) And a comparison means (4g) for determining whether or not both are equal within a predetermined error range, and based on the determination result of the difference detection means (4f), When the frequency spectrum data and the frequency spectrum data of the voice derived from the vehicle are not equal within a predetermined error range, The audio output unit (2) is controlled to reduce the amplification factor of the audio output to the speakers (1a to 1f) with respect to the audio output audio signal received from the audio device (3) below the first value, When the frequency spectrum data of the sound output sound and the frequency spectrum data of the sound derived from the vehicle are equal within a predetermined error range, the sound output unit (2) is controlled to control the speakers (1a to 1f). ), And a control means (4k) for returning the amplification factor of the sound output to the sound output sound signal received from the sound device (3) to the first value.

  Thus, based on the vehicle parameter box data indicating the attenuation rate for each frequency with respect to the frequency spectrum data of the sound derived from outside the vehicle, the frequency spectrum data of the sound derived from outside the vehicle is corrected, and the components of the frequency spectrum data of the sound derived from outside the vehicle after correction Is removed from the frequency spectrum data of the in-vehicle sound received by the in-vehicle microphone (5), and the frequency spectrum data after the removal is used as the frequency spectrum data of the in-vehicle sound, so that the out-of-vehicle sound is converted from the in-vehicle sound. It comes to separate.

  Thus, by storing and using the attenuation frequency characteristics suitable for the vehicle as vehicle parameter box data, it is possible to obtain the sound derived from the interior of the vehicle separately from the noise derived from the outside of the vehicle by a method suitable for the vehicle. it can.

  According to a second aspect of the present invention, in the volume control device according to the first aspect, the speakers (1a to 1f) are a plurality of speakers provided at different positions in a vehicle, and the volume control device. (4) extracts the image of each occupant's face in the vehicle interior based on the captured images sequentially received from the in-vehicle camera (9) mounted on the vehicle, and each occupant based on the extracted face image Image analyzing means (4j) for determining whether or not the voice is spoken, and the control means (4k) is configured such that the frequency spectrum data of the sound output sound and the frequency spectrum data of the sound derived from the vehicle are predetermined. Based on the determination result of the image analysis unit (4j), the speaker near the occupant determined to be speaking is controlled among the speakers (1a to 1f) based on the determination result of the image analysis unit (4j). No Amplification factor for the sound output sound signal received from the sound device (3) of the sound selected as a speaker and controlled by the sound output unit (2) to be output to the control target speakers (1a to 1f) Is lower than the first value.

  For example, when speaking for the purpose of voice recognition or mobile phone call, there is no need for other occupants to listen to the utterance, so there is no need to reduce the volume of the speakers near other occupants. Only the problem that (for example, music) cannot be comfortably heard occurs. On the other hand, if you lower the volume of the speaker near the speaking occupant, you can prevent unnecessary noise from entering the conversation, for example during a mobile phone conversation, and you can hear your voice from the other person speaking on the mobile phone. It becomes easy. In the case of speech for speech recognition, speech recognition accuracy is improved.

  According to a third aspect of the present invention, in the sound volume control device according to the first or second aspect, the vehicle parameter box data includes the frequency of the vehicle-derived sound output by the vehicle sound spectrum output means (4c, 4d). As the data indicating the attenuation rate for each frequency with respect to the spectrum data, the vehicle parameter box means (4e) has a plurality of attenuation characteristic data (21, 22) indicating the attenuation rates caused by different parts of the vehicle. Then, using each of the plurality of attenuation characteristic data (21, 22) in the read vehicle parameter box data, the frequency spectrum data of the vehicle-derived sound output by the vehicle sound spectrum output means (4c, 4d) is corrected. It is characterized by doing.

  In this way, in order to correct the frequency spectrum data of the sound derived from outside the vehicle, if a plurality of attenuation characteristic data (21, 22) indicating the attenuation rate due to different parts of the vehicle is used, for example, the vehicle For vehicles with the same part (for example, the vehicle body) but different parts of the vehicle (for example, seats), only the damping characteristic data for the other part may be used. Compared with the case where the attenuation characteristic data of the entire vehicle is created corresponding to any combination, the creation of the vehicle parameter box data can be divided for each part, so the convenience in creating the vehicle parameter box data is improved.

  According to a fourth aspect of the present invention, in the volume control device according to the third aspect, one of the plurality of attenuation characteristic data (21, 22) indicates an attenuation rate caused by a seat of the vehicle. Sheet attenuation characteristic data (22), and the volume control device corrects the attenuation rate of the sheet attenuation characteristic data (22) at the target frequency with respect to each target frequency while sequentially shifting the target frequency. The correction means (210-250) outputs the frequency sound, which is a sound including only the target frequency component, to the speakers (1a to 1f), and the speaker (1 1a to 1f) based on the signal level A of the sound signal of the frequency sound to be output and the signal level B of the sound signal of the frequency sound received and output by the in-vehicle microphone (5), Characterized by modifying the attenuation factor of the seat damping characteristic data in the frequency of the elephant (22).

  Thus, since the seat attenuation characteristic data (22) in the vehicle parameter box data can be corrected, it is possible to reflect a subtle difference in the attenuation characteristic for each vehicle, and the versatility of the vehicle parameter box is accordingly increased. spread. That is, the range of vehicles to which the same vehicle parameter box can be applied is expanded.

  Further, as described in claim 5, the features of the invention described in claim 1 can also be understood as a program. In addition, the code | symbol in the bracket | parenthesis in the said and the claim shows the correspondence of the term described in the claim, and the concrete thing etc. which illustrate the said term described in embodiment mentioned later. .

It is a block diagram of a vehicle-mounted system. 2 is a diagram illustrating positions of speakers 1a to 1f and in-vehicle microphone 5 in vehicle 100. FIG. 3 is a diagram illustrating a functional configuration of a volume control device 4. FIG. It is a graph which shows an example of vehicle parameter box data. It is a flowchart of the process which the control part of a volume control apparatus performs. It is a flowchart of the process which the control part of a volume control apparatus performs.

  Hereinafter, an embodiment of the present invention will be described. FIG. 1 shows a configuration of an in-vehicle system according to the present embodiment. This in-vehicle system is mounted on a vehicle, and includes speakers 1a to 1f, an audio output unit 2, an audio device 3, a volume control device 4, an in-vehicle microphone 5, an in-vehicle microphone 6, a seating sensor 7, a door opening / closing sensor 8, and an in-vehicle camera. 9. A voice recognition device 10 is provided.

  The plurality of speakers 1a to 1f are attached to different positions in the vehicle interior (or in the wall facing the vehicle interior) of the vehicle 100, and mainly output sound into the vehicle interior. For example, the speaker 1a to 1f may be attached as shown in FIG.

  The sound output unit 2 amplifies the sound output sound signal received from the sound device 3 and outputs it to the speakers 1a to 1f. The amplification factor of the sound signal output to the speakers 1a to 1f with respect to the sound output sound signal received from the sound device 3 can be set for each of the speakers 1a to 1f.

  The sound device 3 reads audio data (for example, music data) recorded on a CD medium, a DVD medium, a flash memory, a magnetic storage medium, etc., and generates an audio signal from the read audio data in accordance with a user operation. An audio signal (sound output audio signal) is output to the audio output unit 2 and the volume control device 4 at the same timing. The acoustic device 3 receives radio and television broadcast radio waves in accordance with user operations, reads audio data (for example, music data) included in the received broadcast radio waves, generates an audio signal from the read audio data, and generates The audio signal (sound output audio signal) is output to the audio output unit 2 and the volume control device 4 at the same timing.

  The volume control device 4 is a device that controls the output volume of the speakers 1 a to 1 f by controlling each amplification factor in the audio output unit 2. Specifically, the audio signal output from the acoustic device 3 and the audio signal received from the in-vehicle microphone 5 are sequentially monitored, and based on a comparison between the two, an occupant's speech (conversation, phone call using a mobile phone, voice recognition) Utterance utterance etc.) is determined, and if it is determined that the utterance is being made, the utterance position where the utterance is made is specified using the captured image of the in-vehicle camera 9, and each amplification factor of the audio output unit 2 is determined. Is controlled, the volume of only the speaker close to the utterance position among the speakers 1a to 1f is reduced. Note that the volume control device 4 operates regardless of whether or not a main power switch of the vehicle 100 (for example, an IG switch in the case of a gasoline vehicle) is on.

  The in-vehicle microphone 5 is mounted in the vehicle interior (or in the wall facing the vehicle interior), receives the sound in the vehicle interior (in-vehicle sound), converts it into a sound signal, and transmits it to the volume control device 4 and the speech recognition device 10. Output. The mounting position of the in-vehicle microphone 5 is, for example, arranged at a position closest to the driver's seat 11 among the seat 11, the passenger seat 12, and the rear seat 13 so that the driver's sound can be better received among the passengers. A plurality of in-vehicle microphones 5 may be provided.

  The outside microphone 6 is attached to a wall facing the outside of the vehicle, converts outside-derived sound (road noise, etc.) generated outside the vehicle into a sound signal and outputs the sound signal to the volume control device 4.

  The seating sensor 7 is a well-known device that detects whether or not the vehicle is seated on the driver's seat 11 (by a capacitance method, a contact method, etc.) and outputs a detection signal of the detection result to the volume control device 4. The door opening / closing sensor 8 is a known device that detects opening / closing of the door of the vehicle 100 and outputs a detection signal of the detection result to the volume control device 4.

  The in-vehicle camera 9 is a device that captures a vehicle interior at a predetermined cycle (for example, 30 frames per second), and sequentially outputs a captured image of the captured result to the volume control device 4. The installation position and shooting direction of the in-vehicle camera 9 are set in advance so that the face of the occupant sitting in the driver's seat 11, the passenger seat 12, and the rear seat 13 enters the angle of view. For example, the in-vehicle camera 9 may be attached to a vehicle rearview mirror.

  The voice recognition device 10 performs a known voice recognition process on the voice signal output from the in-vehicle microphone 5 based on the user performing a voice recognition start operation on the voice recognition device 10, thereby performing voice recognition. As a result of the processing, a phrase spoken by an occupant (mainly a driver) in the passenger compartment is specified, and processing corresponding to the specified word is executed. As the speech recognition device 10, for example, there is a navigation device that searches a destination by using a phrase specified by speech recognition processing as a search word.

  Hereinafter, the volume control device 4 will be described in more detail. The volume control device 4 is a well-known microcontroller including a CPU, RAM, ROM, flash memory, and the like, and realizes various functions when the CPU executes a program recorded in the ROM or flash memory. FIG. 3 shows various functions realized by the volume control device 4 in a block configuration.

  The volume control device 4 includes an in-vehicle sound detection unit 4a, a frequency component calculation unit 4b, a vehicle exterior noise detection unit 4c, a frequency component calculation unit 4d, a vehicle parameter box unit 4e, a difference detection unit 4f, a comparison unit 4g, a seating detection unit 4h, A door opening / closing detection unit 4i, an image analysis unit 4j, and a control unit 4k are provided.

  The in-vehicle sound detection unit 4a samples the sound signal of the in-vehicle sound output from the in-vehicle microphone 5 at a predetermined sampling rate, converts it into digital sound data, and outputs it. The frequency component calculator 4b calculates frequency spectrum data (data indicating the intensity of each frequency component of the audio data) of the audio data by performing Fourier transform or the like on the audio data output from the in-vehicle audio detector 4a. Output. Thus, the in-vehicle sound detection unit 4a and the frequency component calculation unit 4b output the frequency spectrum data of the in-vehicle sound received by the in-vehicle microphone 5.

  The vehicle exterior noise detection unit 4c samples the sound signal of the sound derived from the vehicle output from the vehicle exterior microphone 6 at a predetermined sampling rate, converts it into digital sound data, and outputs it. The frequency component calculation unit 4d calculates frequency spectrum data (data indicating the intensity of each frequency component of the audio data) of the audio data by performing Fourier transform or the like on the audio data output from the vehicle exterior noise detection unit 4c. Output. Thus, the vehicle exterior noise detection unit 4c and the frequency component calculation unit 4d output the frequency spectrum data of the vehicle-derived sound received by the vehicle exterior microphone 6.

  The vehicle parameter box unit 4e corrects the frequency spectrum data of the sound derived from the outside of the vehicle output from the frequency component calculation unit 4d, so that the frequency spectrum when the sound derived from the outside of the vehicle is transmitted to the vehicle interior and attenuated and received by the in-vehicle microphone 5 is obtained. Estimate the data and output the corrected frequency spectrum data.

  This correction is performed based on the read vehicle parameter box data by reading the vehicle parameter box data recorded in the flash memory of the volume control device 4 which is a nonvolatile writable storage medium. FIG. 4 shows an example of vehicle parameter box data. The vehicle parameter box data has two pieces of attenuation characteristic data indicating an attenuation rate for each frequency (represented by a reduction amount in dB) with respect to the frequency spectrum data output by the frequency component calculation unit 4d. That is, it has body damping characteristic data 21 and seat damping characteristic data 22. The body damping characteristic data 21 and the seat damping characteristic data 22 indicate the damping rate caused by different parts of the vehicle.

  That is, the body attenuation characteristic data 21 is data indicating an attenuation rate for each frequency when sound derived from the outside of the vehicle is attenuated by the body of the vehicle 100 before reaching the vehicle interior microphone 5 after entering the vehicle interior. The body attenuation characteristic data 21 is set by experiments or the like according to the material and shape of the vehicle body of the vehicle 100, the installation position of the in-vehicle microphone 5, etc., and is previously set (for example, when the volume control device 4 is installed). Flash memory is recorded.

  The seat attenuation characteristic data 22 is data indicating an attenuation rate for each frequency when sound derived from outside the vehicle is attenuated by the seats 11 to 13 of the vehicle 100 before reaching the interior microphone 5 after entering the vehicle interior. The seat attenuation characteristic data 22 is set by experiments or the like according to the number, material, shape, arrangement, seat position of the in-vehicle microphone 5 and the like of the seats 11 to 13, and the volume (for example, when the volume control device 4 is installed) is set in advance. The flash memory of the control device 4 is recorded. As will be described later, the seat attenuation characteristic data 22 can be corrected after the volume control device 4 is mounted on the vehicle.

The vehicle parameter box unit 4e performs the following calculation using such vehicle parameter box data. The damping factor of each frequency f indicated by the body damping characteristic data 21 is D1 (f), the damping factor of each frequency f indicated by the seat damping characteristic data 22 is D2 (f), and the frequency component calculation unit 4d outputs the vehicle-derived origin. Assuming that the amplitude of each frequency of the frequency spectrum data of speech is I (f), the amplitude O (f) of each frequency of the corrected frequency spectrum data is
O (f) = I (f) 10− ^{(D1 (f) + D2 (f)) / 10}
Calculated by the formula

  The difference detection unit 4f removes the frequency spectrum data of the in-vehicle sound output from the frequency component calculation unit 4b by subtracting the corrected frequency spectrum data component of the vehicle-derived sound output from the vehicle parameter box unit 4e, Output frequency spectrum data after removal. Since the frequency spectrum data after the removal is obtained by removing the sound derived from outside the vehicle from the sound received by the in-vehicle microphone 5, the sound derived from the vehicle (that is, the utterance sound of the passenger in the vehicle and the outputs of the speakers 1a to 1f). (Sound synthesis) frequency spectrum data.

  The comparison unit 4g performs the same processing as the in-vehicle sound detection unit 4a and the frequency component calculation unit 4b on the sound output sound signal output from the sound device 3, thereby outputting the sound output sound signal output from the sound device 3. Frequency spectrum data is generated. Further, the difference detection unit 4f performs correction corresponding to the change in the sound output sound when propagating from the speakers 1a to 1f to the in-vehicle microphone 5 on the frequency spectrum data of the sound output sound. The correction amount of this correction is experimentally obtained in advance based on the arrangement of the speakers 1a to 1f and the in-vehicle microphone 5, and the material, position, and shape of the body of the vehicle 100 and the seats 11 to 13, and the volume control device 4 It is determined using the sound field correction data recorded in the flash memory.

  Furthermore, the comparison unit 4g compares the frequency spectrum data of the corrected sound output sound with the frequency spectrum data of the in-car sound output from the difference detection unit 4f, and determines whether or not both are equal within a predetermined error range. The determination is made, and the determination result is output to the control unit 4k.

  Specifically, according to the frequency spectrum data of the corrected sound output sound, the intensity (signal level) for each frequency f of the corrected sound output sound is A (f). Yes, according to the frequency spectrum data of the in-vehicle-derived sound, if the intensity (signal level) for each frequency f of the in-vehicle sound is B (f), | A (f) -B (f) | It is determined whether or not both values are equal within a predetermined error range depending on whether or not the integrated value is smaller than a predetermined value.

  If there is no utterance in the car (conversation, phone call using a mobile phone, utterance for voice recognition, etc.), if the acoustic device 3 is in an off state and does not output a sound output voice signal, The frequency spectrum data becomes a minute voltage in each frequency band. On the other hand, if the in-vehicle acoustic device 3 is outputting the sound output sound signal in the on state, the difference from the frequency spectrum data of the sound derived from the vehicle includes a component derived from the sound output sound of the sound device 3. .

  The comparison unit 4g also outputs the frequency spectrum data of the corrected sound output sound and the frequency spectrum data of the in-vehicle-derived sound output by the difference detection unit 4f to the control unit 4k.

  The seating detection unit 4h determines whether an occupant is seated in the driver's seat 11 based on the detection signal of the seating sensor 7, and outputs the determination result to the control unit 4k. The door opening / closing detection unit 4i determines the opening / closing state of the vehicle door based on the detection signal of the door opening / closing sensor 8, and outputs the determination result to the control unit 4k.

  The image analysis unit 4j extracts the image of each occupant's face in the passenger compartment based on the captured images sequentially received from the in-vehicle camera 9, using a well-known image recognition technique, and determines each occupant based on the extracted face image. Then, it is determined whether or not the user is talking, whether or not the user is talking on the mobile phone, and whether or not the user is speaking for voice recognition, and the determination result is output to the control unit 4k.

  Specifically, for each occupant, if the temporal change in the area of the mouth portion in the extracted image of the occupant's face is equal to or greater than a predetermined reference, it is determined that the occupant is speaking. Further, when it is determined that the occupant is speaking, whether or not the occupant is using a mobile phone is determined, the extracted face image of the occupant and the face template when the arm is applied to the ear Based on the image, determination is made using a well-known template matching technique. Further, when the occupant is an occupant at the position of the driver's seat 11 (ie, a driver) and the voice recognition start operation is performed, it is determined that the occupant is speaking for voice recognition. To do. A signal indicating whether or not a voice recognition start operation has been performed is acquired from the voice recognition device 10.

  Note that the position of each occupant is determined by, for example, dividing the captured image into four equal parts vertically and horizontally, and if the occupant's face image is included in the lower right part of the occupant, Is determined to be seated on the left side of the front seat, and if included in the lower left part of the front seat, the passenger is determined to be seated on the right side of the front seat and included in the upper right part of the split. For example, it may be determined that the occupant is seated on the left side of the rear seat, and if the occupant is included in the divided upper left portion, the occupant may be determined to be seated on the right side of the rear seat.

  The control unit 4k specifies the utterance position where the utterance is performed when the occupant is speaking based on the outputs of the comparison unit 4g, the seating detection unit 4h, the door opening / closing detection unit 4i, the image analysis unit 4j, and the like. Then, by controlling each amplification factor of the audio output unit 2, the volume of only the speaker close to the speech position among the speakers 1a to 1f is reduced.

  FIG. 5 shows a flowchart of processing executed by the control unit 4k. The control unit 4k is configured to always execute the process of FIG. 4 while the volume control device 4 is in operation. Hereinafter, the operation of the control unit 4k will be described with reference to FIG.

  First, when the driver is not in the passenger compartment, the control unit 4k first determines in step 110 whether the driver is in the vehicle based on the determination results output from the seating detection unit 4h and the door opening / closing detection unit 4i. Determine whether or not. In this case, it is determined that the vehicle is not on board, and then step 120 is executed and then step 110 is executed again. In step 120, by controlling the audio output unit 2, the respective amplification factors of the speakers 1a to 1f (the amplification factors for the audio output audio signals received from the audio devices 3 of the audio to be output to the speakers 1a to 1f) are determined in advance. Is set to a default value (corresponding to an example of the first value), which is an initial set value. Steps 110 and 120 are repeatedly executed until the driver gets in the passenger compartment.

  After that, when an occupant including a driver gets on the vehicle, if it is determined in step 110 that the driver has boarded, then in step 130, based on the determination result output by the difference detection unit 4f, the corrected sound output sound It is determined whether the frequency spectrum data of and the frequency spectrum data of the voice derived from the vehicle are equal within a predetermined error range. If no occupant in the vehicle is speaking, it is determined that both are equal within a predetermined error range regardless of whether or not the acoustic device 3 outputs the acoustic output sound to the sound output unit 2. Then, after step 120 is executed, step 110 is executed again.

  Thereafter, it is assumed that the occupant operates the audio device 3 and, as a result, the audio device 3 starts to output a signal of an audio output audio signal (for example, a music signal) to the audio output unit 2 and the volume control device 4. Then, the sound output sound signal is amplified by the sound output unit 2 with a default amplification factor, and the sound output sound after amplification is output from the speakers 1a to 1f. Even in this state, when the occupant does not speak, the processing of steps 110, 130, and 120 is repeated.

  When any occupant starts speaking while the sound output sound signal is being output from the sound device 3, the output from the comparison unit 4g is the frequency spectrum data of the sound output sound after correction and the frequency of the sound derived from the vehicle interior. It is determined that the difference from the spectrum data exceeds a predetermined error range. Therefore, the control unit 4 also determines in step 130 that both exceed the predetermined error range, and then proceeds to step 140 in order to control the amplification factor of the audio output unit 2.

  In step 140, a speaker to be controlled is selected from the speakers 1a to 1f based on the determination result output by the image analysis unit 4j. Specifically, in the determination result output by the image analysis unit 4j, if there is even one occupant having a conversation, all the speakers 1a to 1f are controlled. This is because in the case of conversation, it is necessary for other occupants to listen to the utterance, and for this purpose, it is desirable to reduce the volume of all the speakers 1a to 1f in the vehicle interior so as not to disturb the conversation.

  In addition, in the determination result output by the image analysis unit 4j, there is no occupant having a conversation, and the occupant talking on the mobile phone or speaking for voice recognition If so, only the speakers near the passenger are controlled. For example, when a passenger in the driver's seat is speaking for voice recognition, only the speakers 1a and 1c near the driver's seat 11 are controlled. Further, for example, when an occupant seated on the right side of the rear seat 13 is talking on a mobile phone, only the speaker 1f near the occupant is controlled.

  In this way, when speaking for the purpose of voice recognition other than conversation or mobile phone calls, it is not necessary for other occupants to listen to the utterance, so it is not necessary to lower the volume of the speakers near other occupants, and it can be lowered. On the contrary, only the problem that the sound output sound (for example, music) cannot be comfortably heard occurs. On the other hand, if you lower the volume of the speaker near the speaking occupant, you can prevent unnecessary noise from entering the conversation, for example during a mobile phone conversation, and you can hear your voice from the other person speaking on the mobile phone. It becomes easy. In the case of speech for speech recognition, speech recognition accuracy is improved.

  Subsequently, in step 150, the corrected acoustic output based on the frequency spectrum data of the corrected sound output sound output from the comparison unit 4g and the frequency spectrum data of the in-vehicle sound output from the difference detection unit 4f. A ratio R of the volume level of the voice to the volume level of the voice derived from the vehicle is calculated.

  Since the in-vehicle-derived voice is a synthesis of the acoustic output voice and the occupant's voice, if the occupant's voice is much louder than the acoustic output voice, the ratio R becomes almost zero. In such a case, the control target Even a speaker does not need to be turned down. On the other hand, if the sound output sound is much louder than the occupant's voice, the ratio R approaches one. In such a case, the effect of lowering the volume of the speaker to be controlled is great. For example, 0.3 is used as the threshold Th.

  If the ratio R is larger than the threshold value Th, then in step 160, the audio output unit 2 is controlled to receive the amplification factor of the control target speaker (the audio output to the control target speaker is received from the acoustic device 3). (Amplification rate for sound output sound signal) is reduced from the default value. For example, the amplification factor is reduced by 10 dB from the current value. Alternatively, the amplification factor may be lowered so that the output volume becomes zero. Following step 160, the process returns to step 110.

  After returning to step 110, while the same occupant's utterance continues, the process proceeds with steps 110, 130, and 140. In step 140, the same speaker as the previous one is selected as the control target. If the ratio R as a result of the amplification factor reduction in step 160 is equal to or less than the threshold Th, the process bypasses step 160 and returns to step 110. Therefore, while the same occupant is speaking, the volume of the speaker to be controlled remains reduced.

  When the occupant's utterance ends and the state returns to the state where no occupant is speaking, the process proceeds to Steps 110 and 130. In Step 130, it is determined that the in-vehicle sound and the acoustic output sound are equal within an error range, Thereafter, in step 120, the amplification factors of all the speakers 1a to 1f are returned to the default values (corresponding to the first value). Thereby, the volume of the sound output sound from the sound device 3 is restored.

  As described above, the voice recognition device 10 corrects the frequency spectrum data of the voice derived from the outside of the vehicle based on the vehicle parameter box data indicating the attenuation rate for each frequency with respect to the frequency spectrum data of the voice derived from the outside of the car. The component of the audio frequency spectrum data is removed from the frequency spectrum data of the in-vehicle sound received by the in-vehicle microphone 5, and the frequency spectrum data after the removal is used as the frequency spectrum data of the in-vehicle-derived sound. Is to be separated from in-car audio.

  Thus, by storing and using the attenuation frequency characteristics suitable for the vehicle as vehicle parameter box data, it is possible to obtain the sound derived from the interior of the vehicle separately from the noise derived from the outside of the vehicle by a method suitable for the vehicle. it can.

  Further, the vehicle parameter box data indicates the attenuation rate due to different parts (vehicle body, seat) of the vehicle as data indicating the attenuation rate for each frequency with respect to the frequency spectrum data of the voice derived from outside the vehicle output by the frequency component calculation unit 4d. The vehicle parameter box 4e has a plurality of attenuation characteristic data 21 and 22, and the vehicle parameter box 4e uses each of the plurality of attenuation characteristic data 21 and 22 in the read vehicle parameter box data to obtain the frequency spectrum data of the voice derived from outside the vehicle. It is to be corrected.

  As described above, if the speech recognition apparatus 10 uses the plurality of attenuation characteristic data 21 and 22 indicating the attenuation rate caused by different parts of the vehicle in order to correct the frequency spectrum data of the voice derived from outside the vehicle. For example, for vehicles with the same car body but different seats, only different damping characteristic data for the other parts may be used, and the damping characteristics of the entire vehicle corresponding to any combination of vehicle parts. Compared with the case of creating data, the creation of vehicle parameter box data can be divided for each part, so the convenience in creating the vehicle parameter box data is improved.

  Next, an operation for correcting the vehicle parameter seat attenuation characteristic data 22 after the volume control device 4 is mounted on the vehicle will be described. In order to realize this operation, the volume control device 4 executes a predetermined program regardless of whether or not a main power switch of the vehicle 100 (for example, an IG switch in the case of a gasoline vehicle) is on, The process shown in FIG. 6 is executed.

  In this process, it is determined in step 210 whether or not the vehicle 100 is traveling based on the detection signal of the vehicle speed sensor so that it can be executed when the vehicle 100 is parked in a garage or the like. Then, in step 220, it is determined whether or not the driver is in the vehicle based on the seating sensor 7 and the door opening / closing sensor 8. If it is determined that the driver is not in the vehicle, the target frequency is sequentially shifted, By executing steps 230 to 250 for each target frequency, the attenuation rate of the sheet attenuation characteristic data 22 at the target frequency is corrected. If it is determined in step 210 that the vehicle is traveling, or if it is determined in step 220 that the driver is riding, the process returns to step 210 again.

  Hereinafter, the processing of steps 230 to 250 for one target frequency will be described. First, in step 230, a specific speaker is selected from the speakers 1a to 1f in the vehicle interior, and a sound (frequency sound) including only a target frequency component is output to the selected speaker.

  The speaker to be selected is a speaker located on the side opposite to the in-vehicle microphone 5 when viewed from the center of the passenger compartment among the speakers 1a to 1f in the passenger compartment. In the case of the present embodiment, since the in-vehicle microphone 5 is on the front side in the vehicle interior, the speakers to be selected are the speakers 1e and 1f on the rear side in the vehicle interior. The reason for this is to make it easier for the sound coming out of the selected speakers 1e, 1f and reaching the in-vehicle microphone 5 to reflect the attenuation effect of the seats 11, 12, 13.

  In order to output the frequency sound to the selected speakers 1e and 1f, the volume control device 4 instructs the sound device 3 to output the frequency sound at a predetermined output level via a control line (not shown). At the same time, the audio output unit 2 is controlled to set the amplification factors of the speakers 1e and 1f to default values so that the volume of the speakers other than the speakers 1e and 1f becomes zero.

  Then, the acoustic device 3 outputs a sound signal of the frequency sound at the output level to the sound output unit 2 and the volume control device 4 according to the command. Then, the frequency sound is output from only the speakers 1e and 1f among the speakers 1a to 1f.

  Subsequently, in step 240, by performing frequency analysis of the sound signal of the frequency sound received from the acoustic device 3, the signal level (dB unit) of the target frequency is specified, and the sound output unit 2 is set to the specified signal level. The signal level A is calculated by performing correction according to the default amplification factor and the number of selected speakers 1e, 1f (two in this case). This signal level A is a signal level of a sound signal of a frequency sound input to the speakers 1e and 1f.

  Furthermore, the sound volume control device 4 specifies the signal level B (in dB units) of the target frequency based on the frequency spectrum data of the in-vehicle sound output from the frequency component calculation unit 4b. This signal level B is a signal level of a sound signal of a frequency sound that is received and output by the in-vehicle microphone 5.

  Based on the comparison between the signal levels A and B, the attenuation rate after correction of the sheet attenuation characteristic data 22 at the target frequency is determined. The set of signal levels A and B is data indicating the attenuation characteristics of the seats 11 to 13 at the target frequency.

  Subsequently, in step 250, the attenuation rate of the sheet attenuation characteristic data 22 at the target frequency is corrected based on the attenuation characteristic specified as described above. Specifically, the result C obtained by subtracting the signal level B from the signal level A may be adopted as the attenuation rate after correction of the sheet attenuation characteristic data 22 at the target frequency, or the sheet attenuation characteristic data at the target frequency. The average value of the current attenuation rate 22 and the subtraction result C may be adopted as the corrected attenuation rate of the sheet attenuation characteristic data 22 at the target frequency.

  By executing the processes in steps 230 to 250 once for each frequency and for each frequency, the entire correction of the sheet attenuation characteristic data 22 is completed. Then, the corrected seat damping characteristic data 22 is replaced with the current seat damping characteristic data 22 and recorded in the vehicle parameter box data.

  The seat attenuation characteristic data 22 is data for correcting the sound derived from outside the vehicle received by the microphone 6 for outside the vehicle, but in order to correct it, the sound is output from the speakers 1e and 1f that output sound into the vehicle. The result is used because the seats 11 to 13 are in the passenger compartment. That is, since the seats 11 to 13 are in the vehicle interior, the sound attenuation effect by the seats 11 to 13 is almost the same for the sound originating from outside the vehicle and the sound coming out of the speakers 1e and 1f and entering the in-vehicle microphone 5. It is possible.

  Thus, since the seat damping characteristic data 22 in the vehicle parameter box data can be corrected even after the volume control device 4 is mounted on the vehicle, a subtle difference in the damping characteristic for each vehicle can be reflected. The range of versatility of the vehicle parameter box is expanded. That is, the range of vehicles to which the same vehicle parameter box can be applied is expanded.

  In the above embodiment, the in-vehicle sound detection unit 4a, the frequency component calculation unit 4b, the outside noise detection unit 4c, the frequency component calculation unit 4d, the vehicle parameter box unit 4e, the difference detection unit 4f, the comparison unit 4g, and the seating detection unit 4h. , Door opening / closing detection unit 4i, image analysis unit 4j, and control unit 4k respectively include in-vehicle sound detection means, frequency component calculation means, vehicle exterior noise detection means, frequency component calculation means, vehicle parameter box means, difference detection means, and comparison means. , Corresponding to an example of a seating detection means, a door opening / closing detection means, an image analysis means, and a control means, and the sound volume control device 4 functions as an example of a correction means by executing the processing of FIG.

(Other embodiments)
As mentioned above, although embodiment of this invention was described, the scope of the present invention is not limited only to the said embodiment, The various form which can implement | achieve the function of each invention specific matter of this invention is included. It is.

  For example, in the above embodiment, the number of speakers 1a to 1f in the vehicle compartment is six, but the number of speakers 1a to 1f may be one or ten.

  In the above embodiment, the vehicle parameter box data includes the two attenuation characteristic data 21 and 22 for each part. However, the vehicle parameter box data may include three or more attenuation characteristic data. There may be only one. For example, the vehicle parameter box unit 4e further includes attenuation characteristic data indicating an attenuation factor caused by the engine of the vehicle. The vehicle parameter box unit 4e corrects the frequency spectrum data of the voice derived from outside the vehicle using the attenuation characteristic data. Also good.

  Moreover, in the said embodiment, when it determines with a certain passenger | crew having conversation, the image-analysis part 4j specifies the direction of the face of the passenger | crew by a known image recognition technique, and is in the direction of the direction of the face. The occupant may be specified as the occupant of the conversation partner, and data indicating the occupant who is talking and the occupant of the conversation partner of the occupant may be output to the control unit 4k. In this case, the control unit 4k selects both the speaker close to the talking passenger and the speaker close to the passenger who is talking to the passenger as the control target speakers in Step 140 of FIG. Also good. For example, when the passenger on the right side of the rear seat 13 is talking to the passenger on the left side of the rear seat 13, the speakers 1e and 1f are selected as the speakers to be controlled.

  The vehicle 100 may be equipped with a normal vehicle interior noise canceling device for eliminating road (running) noise.

  When the vehicle 100 is a vehicle having a large body (vehicle body) sound shielding effect, the sound volume control device 4 uses the sound received by the outside microphone 6 to be close to the driver's seat 11 among the speakers 1a to 1f. You may make it output only to the speaker 1c. In this way, it is also possible to pick up the sound outside the vehicle and provide the sound outside the vehicle only to the driver inside the vehicle.

  Further, the installation position of the in-vehicle microphone 5 is not limited to that in the above embodiment, and may be anywhere in the vehicle interior. The installation position of the outside microphone 6 is not limited to that in the above embodiment, and may be anywhere outside the vehicle.

  In the above-described embodiment, each function realized by the volume control device 4 executing a program uses hardware having those functions (for example, an FPGA capable of programming a circuit configuration). It may be realized.

DESCRIPTION OF SYMBOLS 1a-1f Speaker 2 Audio | voice output part 3 Acoustic apparatus 4 Volume control apparatus 5 Microphone for vehicle interior 6 Microphone for vehicle exterior Car-mounted camera 10 Voice recognition apparatus 11-13 Seat 21 Body attenuation characteristic data 22 Seat attenuation characteristic data 4a Car interior voice detection part 4b Frequency component calculation unit 4c Vehicle exterior noise detection unit 4d Frequency component calculation unit 4e Vehicle parameter box unit 4f Difference detection unit 4g Comparison unit 4j Image analysis unit 4k Control unit

Claims (5)

Speakers (1a to 1f) that output sound into the passenger compartment, acoustic devices (3) that output sound signals, and sound output sound signals that are sound signals output from the sound devices (3) An audio output unit (2) that outputs to 1a to 1f, an in-vehicle microphone (5) that receives in-vehicle sound that is sound inside the vehicle, and an out-of-vehicle microphone (6) that receives outside-origin sound generated outside the vehicle. A volume control device for an onboard vehicle,
In-vehicle sound spectrum output means (4a, 4b) for outputting frequency spectrum data of the in-vehicle sound received by the in-vehicle microphone (5);
An external audio spectrum output means (4c, 4d) for outputting frequency spectrum data of the external audio received by the external microphone (6);
Vehicle parameter box data indicating the attenuation rate for each frequency with respect to the frequency spectrum data of the outside-vehicle-derived sound output by the outside-vehicle sound spectrum output means (4c, 4d) is read from the storage medium, and based on the read vehicle parameter box data By correcting the frequency spectrum data of the sound derived from outside the vehicle output by the outside sound spectrum output means (4c, 4d), the sound derived from outside the vehicle is transmitted to the inside of the vehicle and attenuated and received by the microphone for in-vehicle (5). Vehicle parameter box means (4e) for estimating frequency spectrum data when output and outputting the corrected frequency spectrum data;
A component of the corrected frequency spectrum data of the outside-derived sound output by the vehicle parameter box means (4e) is removed from the frequency spectrum data of the in-vehicle sound output by the in-vehicle sound spectrum output means (4a, 4b). And the difference detection means (4f) which outputs the frequency spectrum data after removal as frequency spectrum data of the voice derived from the vehicle,
Frequency spectrum data of the sound output sound signal output by the sound device (3) is generated, and the frequency spectrum data of the generated sound output sound and the frequency of the in-vehicle sound output by the difference detection means (4f) Comparing means (4g) for comparing the spectrum data and determining whether or not both are equal within a predetermined error range;
If the frequency spectrum data of the sound output sound and the frequency spectrum data of the sound derived from the vehicle are not equal within a predetermined error range based on the determination result of the difference detection means (4f), the sound output unit ( 2), the amplification factor of the sound output to the speaker (1a to 1f) with respect to the sound output sound signal received from the sound device (3) is decreased below a first value, and the sound output sound When the frequency spectrum data of the vehicle and the frequency spectrum data of the in-car sound are equal within a predetermined error range, the sound output unit (2) is controlled and output to the speakers (1a to 1f). A volume control device comprising: control means (4k) for returning the amplification factor of the sound with respect to the sound output sound signal received from the sound device (3) to the first value. The speakers (1a to 1f) are a plurality of speakers provided at different positions in the vehicle,
The volume control device (4) extracts an image of each occupant's face in the passenger compartment based on captured images sequentially received from the in-vehicle camera (9) mounted on the vehicle, and based on the extracted face image. The image analysis means (4j) for determining whether or not each occupant is speaking,
When the frequency spectrum data of the sound output sound and the frequency spectrum data of the sound derived from the vehicle are not equal within a predetermined error range, the control means (4k) outputs a determination result of the image analysis unit (4j). On the basis of the speakers (1a to 1f), a speaker near an occupant determined to be speaking is selected as a speaker to be controlled, and the audio output unit (2) is controlled to perform the control. The amplification factor of the sound output to the target speaker (1a to 1f) with respect to the sound output sound signal received from the sound device (3) is lower than the first value. Volume control device. The vehicle parameter box data is an attenuation rate caused by a different portion of the vehicle as data indicating an attenuation rate for each frequency with respect to the frequency spectrum data of the outside-vehicle-derived sound output by the outside-vehicle sound spectrum output means (4c, 4d). A plurality of attenuation characteristic data (21, 22) indicating
The vehicle parameter box means (4e) uses the each of the plurality of attenuation characteristic data (21, 22) in the read vehicle parameter box data to output the vehicle sound spectrum output means (4c, 4d). The volume control apparatus according to claim 1 or 2, wherein the frequency spectrum data of the voice derived from outside the vehicle is corrected. One of the plurality of attenuation characteristic data (21, 22) is seat attenuation characteristic data (22) indicating an attenuation factor caused by the seat of the vehicle,
The volume control device includes correction means (210 to 250) for correcting the attenuation rate of the sheet attenuation characteristic data (22) at the target frequency with respect to each target frequency while sequentially shifting the target frequency. ,
The correcting means (210-250) causes the speakers (1a-1f) to output frequency sounds, which are sounds including only the target frequency component, and causes the speakers (1a-1f) to output frequency sound sounds. Based on the signal level A of the signal and the signal level B of the sound signal of the frequency sound received and output by the in-vehicle microphone (5), the attenuation rate of the seat attenuation characteristic data (22) at the target frequency is determined. The sound volume control device according to claim 3, wherein the sound volume control device is corrected. Speakers (1a to 1f) that output sound into the passenger compartment, acoustic devices (3) that output sound signals, and sound output sound signals that are sound signals output from the sound devices (3) An audio output unit (2) that outputs to 1a to 1f, an in-vehicle microphone (5) that receives in-vehicle sound that is sound inside the vehicle, and an out-of-vehicle microphone (6) that receives outside-origin sound generated outside the vehicle. A program for use in a volume control device for an onboard vehicle,
In-vehicle sound spectrum output means (4a, 4b) for outputting frequency spectrum data of the in-vehicle sound received by the in-vehicle microphone (5),
Vehicle sound spectrum output means (4c, 4d) for outputting frequency spectrum data of the sound derived from the vehicle received by the microphone for vehicle outside (6),
Vehicle parameter box data indicating the attenuation rate for each frequency with respect to the frequency spectrum data of the outside-vehicle-derived sound output by the outside-vehicle sound spectrum output means (4c, 4d) is read from the storage medium, and based on the read vehicle parameter box data By correcting the frequency spectrum data of the sound derived from outside the vehicle output by the outside sound spectrum output means (4c, 4d), the sound derived from outside the vehicle is transmitted to the inside of the vehicle and attenuated and received by the microphone for in-vehicle (5). Vehicle parameter box means (4e) for estimating frequency spectrum data when output and outputting corrected frequency spectrum data;
A component of the corrected frequency spectrum data of the outside-derived sound output by the vehicle parameter box means (4e) is removed from the frequency spectrum data of the in-vehicle sound output by the in-vehicle sound spectrum output means (4a, 4b). And the difference detection means (4f) which outputs the frequency spectrum data after removal as frequency spectrum data of the voice derived from inside the vehicle,
Frequency spectrum data of the sound output sound signal output by the sound device (3) is generated, and the frequency spectrum data of the generated sound output sound and the frequency of the in-vehicle sound output by the difference detection means (4f) Comparison means (4g) for comparing the spectrum data and determining whether or not both are equal within a predetermined error range, and based on the determination result of the difference detection means (4f), the frequency spectrum of the sound output sound When the data and the frequency spectrum data of the in-car sound are not equal within a predetermined error range, the sound output unit (2) is controlled to output the sound to be output to the speakers (1a to 1f). The amplification factor for the sound output sound signal received from the sound device (3) is decreased below the first value, the frequency spectrum data of the sound output sound, When the frequency spectrum data of the voice derived from the inside of the vehicle becomes equal within a predetermined error range, the audio device for controlling the voice output unit (2) and outputting the voice to the speakers (1a to 1f) A program for causing the volume control device to function as control means (4k) for returning the amplification factor for the sound output sound signal received from (3) to the first value.

Images