JP2010141468A - Onboard acoustic apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an onboard acoustic apparatus for preventing the effect of sounds occurring around a vehicle. <P>SOLUTION: The onboard acoustic apparatus 3 for output of a sound signal to respective speakers 4 installed at a plurality of places in the vehicle includes: an amplifying processing means 6 for amplifying the sound signal when the sound signal to be a sound source is input, and for output of the amplified sound signal to the respective speakers 4; a storage means 19 for storing a first acoustic model 31 comprising a plurality of models which are the models of sounds, arriving at a vehicle 2, around the vehicle and which are prepared in accordance with the arrival directions of sounds; and a volume adjusting means 9 for controlling the amplifying processing means 6 in accordance with the arrival directions of the sounds around the vehicle 2, on the basis of a signal from a microphone 5 for detecting sounds around the vehicle 2 and a first acoustic model 31 stored by the storage means 19. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an on-vehicle acoustic device.

In the field of in-vehicle audio equipment, a technique for adjusting the volume according to the noise around the vehicle has been proposed. For example, Patent Document 1 discloses a vehicle voice recognition technology that directs the directivity of a microphone toward a driver when a proximity vehicle is detected by a laser radar. Patent Document 2 discloses a technique for increasing the volume of a loudspeaker having a larger nearby noise among a front seat speaker and a rear seat speaker. Patent Document 3 estimates the noise level felt by the listener from the detected vehicle interior noise level and the vehicle speed signal, and increases the volume and dynamic range of the reproduced audio signal as the running noise level increases. Technology is disclosed.
JP 2001-154688 A Japanese Utility Model Publication No. 3-42447 JP-A-6-319192

  When listening to audio in the vehicle, if noise is generated around the vehicle, it is difficult to hear the sound in the direction in which the noise is generated, and the way of hearing is biased. Further, even if the volume is increased, the volume of all speakers is increased, so that the balance of how the sound is heard remains biased. In particular, when a person is seated in each seat, the sound volume of a person who is difficult to hear sounds changes the way other people hear the sound, resulting in a sound bias. This invention is made | formed in view of such a subject, and makes it a subject to provide the vehicle-mounted audio apparatus which suppresses the influence of the sound which generate | occur | produced around the vehicle.

  In order to solve the above-described problems, the present invention adjusts the sound balance of each speaker according to the arrival direction of the sound around the vehicle based on the acoustic model of the sound from each direction prepared in advance and the actually measured sound. .

  Specifically, it is an in-vehicle acoustic device that outputs a sound signal to each speaker installed in a plurality of locations in the vehicle, and when a sound signal serving as a sound source is input, the sound signal is amplified and output to each speaker. Amplifying processing means for storing, a storage means for storing a first acoustic model composed of a plurality of models prepared according to the direction of sound arrival, which is a model of sound around the vehicle arriving at the vehicle, and the vehicle Volume adjustment means for performing control according to the direction of arrival of sound around the vehicle to the amplification processing means based on a signal from a microphone that detects surrounding sounds and the first acoustic model stored in the storage means And comprising.

  Here, each speaker from which the in-vehicle acoustic device outputs an audio signal is installed at a plurality of locations in the vehicle. Therefore, by adjusting the volume balance of each speaker, the center point of the sound felt by the passenger can be changed. Is possible. The amplification processing means can adjust the amplification factor of the audio signal output to each speaker for each speaker, and the balance thereof is appropriately adjusted by the amplification processing means.

By the way, when external noise arrives, the speaker sound heard from the direction close to the arrival direction is erased by the noise and becomes difficult to hear, so that the center point of the sound felt by the occupant shifts. For this reason, it is necessary to adjust the volume balance between the speakers so that the volume of the speakers in the direction of arrival becomes higher than the other when external noise arrives. In the on-vehicle acoustic device, in order to make the volume of the speaker in the direction of arrival of the sound generated around the vehicle larger than others, a plurality of models prepared in advance according to the direction of arrival of the sound and the measured sound The balance of the sound is adjusted by comparing the signal and controlling the amplification processing means according to the comparison result.

  According to the vehicle-mounted acoustic device configured as described above, it is possible to suppress the influence of sound generated around the vehicle.

  Further, the sound volume adjusting means calculates the likelihood of the signal from the microphone and the first acoustic model for each model, and is installed at a position corresponding to the direction of arrival of the sound defined in each model. The amplification processing means may be controlled so that the volume of the speaker becomes a volume corresponding to the calculated likelihood. According to this, since the volume can be adjusted according to the likelihood that is a numerical parameter, the volume of the speaker according to the sound arrival direction is finely adjusted. For this reason, it is possible to suppress the influence of the sound generated around the vehicle more appropriately.

  Further, the storage means further stores a second acoustic model that is a model of sounds around the vehicle that arrives at the vehicle, and includes a plurality of models prepared according to the type of sound, and the volume adjustment The means may determine the type of sound around the vehicle based on the signal from the microphone and the second acoustic model, and control the amplification processing means according to the type of sound. Good. With such a configuration, it is possible to balance the sound of each speaker according to the type of sound around the vehicle, and the processing content when suppressing the influence of the sound generated around the vehicle is appropriately determined. It is possible to set.

  It is possible to suppress the influence of sound generated around the vehicle.

  Hereinafter, the best mode for carrying out the present invention will be described by way of example. Embodiment shown below is an illustration and this invention is not limited to these.

<System configuration>
FIG. 1 is a diagram showing a vehicle 2 equipped with an audio system 1 according to an embodiment of the present invention. As shown in FIG. 1, the audio system 1 includes a main unit 3 (corresponding to an on-vehicle acoustic device in the present invention), a speaker 4 (FR, FL, RR, RL), and a microphone 5 (in the present invention, Equivalent to a microphone). The main unit 3 has an operation panel (not shown) on the front, and is arranged at a position where a driver sitting on the driver's seat can easily operate. The main unit 3 is connected to a microphone 5 attached to the front of the vehicle 2 and four speakers 4 arranged in the vehicle interior. The vehicle 2 is an ordinary car having 4 to 5 passengers, and four speakers 4 are located near the right front seat (speaker 4FR), near the left front seat (speaker 4FL), and near the right rear seat (speaker 4RR). ), Near the left rear seat (speaker 4RL).

  Each speaker 4 is disposed on an interior panel, a rear parcel board, or the like in the vehicle interior, and converts an electrical audio signal output from the main unit 3 into a sound wave. As is well known, the speaker 4 has a built-in coil that changes the amplitude and period according to the intensity and frequency of the electric audio signal output from the amplifier, and the cone connected to the coil vibrates in response to the movement of the coil. Thus, the electric signal is converted into a sound wave.

The microphone 5 is disposed near the front bumper in front of the vehicle 2 and converts sounds around the vehicle 2 into electric signals. As is well known, the microphone 5 has a built-in capacitor for converting a sound wave into an electric signal, and the capacitor generates an electric sound signal by vibration of the sound wave. The microphone 5 applies electric power supplied from the main unit 3 to both poles of the capacitor, and converts sound waves into electrical signals by potential fluctuation between both poles caused by sound waves.

  FIG. 2 is a system configuration diagram of the audio system 1. As shown in FIG. 2, the main unit 3 of the audio system 1 includes an amplifier unit 6 (corresponding to amplification processing means in the present invention), a selector 7 and a source unit 8 (corresponding to sound sources in the present invention). The control unit 9 (corresponding to the sound volume adjusting means in the present invention) and the operation switch panel 10 are connected. The speaker 4 is connected to the amplifier unit 6 and the microphone 5 is connected to the control unit 9. ing.

  The amplifier unit 6 is a four-channel power amplifier incorporating four amplifiers 11 (amplifiers 11FL, FR, RL, RR). Each amplifier 11 is a source via a selector 7 according to a control signal from the control unit 9. The audio signal from the unit 8 is amplified and output to the corresponding speaker 4. The four amplifiers 11 are configured such that the amplification circuits are independent from each other, and the control unit 9 can arbitrarily set the amplification factor of each amplifier 11. That is, the control unit 9 can make the amplification factor of a specific amplifier out of the four amplifiers 11 higher or lower than other amplifiers.

  The selector 7 includes a selector switch 12 that switches sound sources of a plurality of sources, selects an audio signal of any one of the plurality of sources of the connected source unit 8, and sends the selected audio signal to the amplifier unit 6. Note that since the audio signal output from the source is the left (L) audio signal and the right (R) audio signal, the two selector switches 12 are provided in parallel, and the left and right audio signals are provided. The audio signals can be switched in parallel. Of the audio signals from the source, the left audio signal is sent to the amplifiers 11FL and RL, and the right audio signal is sent to the amplifiers 11FR and RR. The selector switch 12 of the selector 7 is an electronic switch that can operate according to a control signal from the control unit 9.

  The source unit 8 includes a CD player 13 (CD: Compact Disc), an FM tuner 14 (FM: Frequency Modulation), an AM tuner 15 (AM: Amplitude Modulation), and an external input 16. The CD player 13 is a playback device capable of playing back audio content recorded on a CD, converts the played back audio data into an electrical audio signal, and outputs it to the selector 7. Further, the FM tuner 14 converts an FM broadcast radio wave received via an antenna (not shown) into an electric audio signal and outputs the electric sound signal to the selector 7. As with the FM tuner, the AM tuner 15 also converts the AM broadcast radio wave received via an antenna (not shown) into an electrical audio signal and outputs it to the selector 7. The external input 16 is a circuit that outputs an audio signal input via an input terminal (not shown) provided on the exterior of the main unit 3 to the selector 7. Operations such as playback and fast-forwarding of the CD player 13, channel selection operation in the FM tuner 14, and channel selection operation of the AM tuner 15 are realized by control signals from the control unit 9. In this embodiment, a source such as a CD, FM, or AM is illustrated, but the present invention is not limited to such a sound source. For example, a portable music player having a built-in flash memory, etc. It may be a sound source.

  The control unit 9 includes a microphone amplifier 17, an A / D converter 18 (A / D: ANALOG-to-DIGITAL), a memory 19 (corresponding to storage means in the present invention), an operation switch panel 10, and a CPU 20 (CPU : Central Processing Unit), and mainly controls operations of the amplifier unit 6, the selector 7, and the source unit 8.

  The CPU 20 is an arithmetic processing unit that implements various functional units by executing computer programs. When the main power supply of the main unit 3 is turned on by a user operation, the noise determination unit 21 cooperates with the memory 19. The functions of the likelihood calculation unit 22, the amplifier control unit 23, the source control unit 24, the selector control unit 25, the operation processing unit 26, and the amplification degree determination unit 27 are realized. These various functional units are not all realized in parallel with the start of the operation of the CPU 20, and program modules that implement these various functional units are task-processed by the CPU 20. The operations of these functional units will be described in detail in the processing flow described later.

  The microphone amplifier 17 is a signal amplification circuit that amplifies the electrical audio signal from the microphone 5 and sends it to the A / D converter 18. The electric audio signal from the microphone 5 amplified by the microphone amplifier 17 is converted by the A / D converter 18 into a digital signal that can be numerically processed by the CPU 20 and sent to the CPU 20. As is well known, the A / D converter 18 is a converter that converts an analog signal into a digital signal.

  The operation switch panel 10 is a panel including various operation buttons, a status indicator, a liquid crystal display, and the like, and these buttons and indicators are arranged in front of the main unit 3. When the operation button is pressed, the operation switch panel 10 transmits an electrical signal to the CPU 20 and turns on / off the status display lamp based on a control signal sent from the CPU 20 or displays information on the liquid crystal display. To do.

  The memory 19 is a flash memory that does not require a memory holding operation, and stores or erases data when accessed by the CPU 20. The memory 19 stores an acoustic model 31 (corresponding to the first acoustic model referred to in the present invention), a balance map 28, a coefficient map 29, and a control parameter 30, which will be described later. To access.

  FIG. 3 is a diagram showing the contents of the acoustic model 31 stored in the memory 19. As shown in FIG. 3, the acoustic model 31 includes five acoustic models 31-30, 60, 90, which are created in advance at the time of factory shipment of the audio system 1 and stored in the memory 19 according to the direction of arrival of noise. 120 and 150. The numerical value of 2 to 3 digits attached to the reference sign indicates the direction of noise arrival to the microphone 5 as an angle, and as shown in FIG. It is set to be °, 90 °, 120 °, and 150 °. These five acoustic models 31-30... 150 are noise models (GMM: Gaussian Mixture Model) that are recorded on the basis of the recorded audio signals, recording noise coming from each direction toward the microphone 5. It is an acoustic model used when identifying the direction of arrival of the sound input to the microphone 5 in the process of the third operation flow described later. The acoustic model is not limited to the above angle, and models of 210 °, 240 °, 270 °, 300 °, and 330 ° can also be used.

FIG. 5 is a diagram showing the contents of the balance map 28. The balance map 28 defines the sound balance in the passenger compartment. The balance map 28 is a map that defines the relationship between the seating state of the occupant of the vehicle 2 and the amplification factor of each amplifier (amplifiers 11FR, FL, RR, RL). In other words, the balance map 28 is configured so that each seat (not shown) embedded in the seat surface of the right front, left front, right rear, and left rear seats so that the balance of the volume of sound reaching the passenger's ear of the vehicle 2 is as uniform as possible. It is the map which defined the amplification factor of each amplifier according to the detection state of a sensor (FR, FL, RR, RL). For example, if the occupant is sitting only in the front right seat and no one is sitting in the other seats, the center of the sound that can be heard in the passenger compartment is near the front right seat, so that the left front, left rear, and right rear The amplification factors of the amplifiers 11FL, RR, and RL that amplify the sound of the speaker 4 are set to be larger than those of the amplifier 11FR. The seating sensor is a sensor that detects the presence or absence of an occupant based on a load, and detects “ON” when seated, and “OFF” when not seated. The balance map 28 does not define the case where the FR seating sensor is OFF. This is because when the audio system 1 is operated, it is difficult to consider the situation in the vehicle interior where there is no driver.

  FIG. 6 is a diagram showing the contents of the coefficient map 29. The coefficient map 29 defines coefficients for adjusting the amplification factor of each amplifier based on the arrival direction of noise identified by arithmetic processing described later. That is, the coefficient map 29 defines coefficients to be multiplied by the control parameter of the amplification factor of the corresponding amplifier so that the sound of the speaker in the direction in which the noise comes increases, for example, the direction of 30 to 60 ° When the external noise is generated, the gain on the right front side is increased, and when the external noise is generated in the direction of 120 to 150 °, the left front side is increased. A coefficient for multiplying the parameter is defined (see FIG. 7).

  FIG. 8 is a diagram showing the contents of the control parameter 30. In the memory 19, in order to control the amplification factor of each amplifier 11, a storage area for storing the control parameter 30 is prepared as shown in FIG. The parameter “VOL” is a main volume parameter that increases or decreases when the user operates the volume knob of the operation switch panel 10. The parameter “VFR” is an amplifier 11FR, “VFL” is an amplifier 11FL, “VRR” is an amplifier 11RR, and “VRL” is a parameter for individually determining the amplification factor of the amplifier 11RL. A value obtained by multiplying other variables is the amplification factor of the corresponding amplifier 11. Details will be described later.

<Audio system operation flow>
When the main power supply of the audio system 1 is turned on, the control unit 9 of the main unit 3 executes the following three operation flows in parallel, and controls the amplifier unit 6, the selector 7, and the source unit 8. In each operation flow shown below, the CPU 20 that has started to execute the computer program upon receiving power supply refers to the data in the memory 19 and receives signals from the A / D converter 18 and the operation switch panel 10. Realized by processing. Hereinafter, each operation flow will be described in order. For convenience of explanation, each operation flow will be described in order. However, the audio system 1 does not have a correlation with each other when the operation flows are executed in parallel. For example, the processing in step S105 of the first operation flow is performed. Even if step S305 of the third operation flow is executed and the control parameter is changed immediately before execution of step S105, the process of step S105 is not interrupted.

  FIG. 9 is a diagram showing a first operation flow. Hereinafter, the first operation flow realized by the audio system 1 will be described with reference to FIG.

  (Step S101) When the main power supply of the audio system 1 is turned on, the CPU 20 executes the computer program, thereby realizing the amplifier control unit 23. The amplifier control unit 23 scans input points from four seating sensors (not shown) to detect the right front, left front, right rear, and left rear seating states. Then, the amplifier controller 23 refers to the balance map 28 stored in the memory 19 and searches for a mode that matches the detected seating state. For example, if it is detected that one occupant is seated in the front right seat, “mode A” is searched from the balance map 28. Then, the amplification factor of each amplifier defined in “mode A” of the balance map 28 is read. By determining the amplification factor of each amplifier based on such a balance map 28, the sound balance is best for all members, regardless of whether the number of passengers is one or multiple. The center point of the sound is adjusted.

  (Step S102) When the process of Step S101 is completed, the amplifier control unit 23 refers to the control parameter 30 in the memory 19 next. That is, the amplifier control unit 23 reads parameters of the variables VFR, VFL, VRR, VRL, and VOL. If the coefficient adjustment process (step S305) in the third operation flow to be described later has not yet been executed after the main unit 3 is activated, “1.00” is set to the variables VFR, VFL, VRR, and VRL. Yes.

  (Step S103) The amplifier control unit 23 multiplies the amplification factor of each amplifier read in the process of Step S101 by the parameters VFR, VFL, VRR, and VRL corresponding to the amplifiers read in the process of Step S102. Further, VOL is multiplied. Thereby, the control target value of the amplification factor of each amplifier is determined. For example, when one occupant is seated in the right front seat, the variable VFR is 1.00 and the VOL is 1, the control target value of the amplification factor of the amplifier 11FR is determined to be 50. Further, as a result of the execution of the process in step S305 of the third operation flow executed in parallel with the first operation flow, the VFR of the control parameter 30 stored in the memory 19 is 1.2, for example. Furthermore, if the VOL is 1.05 by the user operation, the control target value of the amplification factor of the amplifier 11FR is determined to be 50 × 1.2 × 1.05 = 63.

  (Step S <b> 104) The amplifier control unit 23 notifies the amplification factor of each amplifier 11 determined in the process of Step S <b> 103 to each amplifier 11 of the amplifier unit 6 with a control signal. Each amplifier 11 notified of the amplification factor from the amplifier control unit 23 uses this amplification factor as a control target value, amplifies the electrical audio signal sent from the source unit 8 via the selector 7, and sends it to the corresponding speaker 4.

  (Step S <b> 105) If the main power supply of the audio system 1 remains on, the amplifier control unit 23 executes the processes after step S <b> 101 again. On the other hand, if the main power supply of the audio system 1 is turned off, execution of a series of processes is terminated.

  The first operation flow is as described above. Next, the second operation flow will be described. FIG. 10 is a diagram showing a second operation flow. Hereinafter, the second operation flow realized by the audio system 1 will be described with reference to FIG.

  (Step S201) When the main power supply of the audio system 1 is turned on, the CPU 20 executes the computer program, thereby realizing the source control unit 24, the selector control unit 25, and the operation processing unit 26. The operation processing unit 26 scans the energized state of the buttons on the operation switch panel 10 and checks whether there is an operation by the user. When detecting a CD playback operation, FM reception, AM reception, or external input acceptance operation by the user, the operation processing unit 26 executes the processes in and after step S202.

  (Step S202) When the operation processing unit 26 detects a CD playback operation, FM reception, AM reception, or external input reception operation, the selector control unit 25 sends a control signal to the selector 7, and the user The selector switch 12 is set to the requested source.

  (Step S203) After the selector switch 12 is controlled by the selector control unit 25, the source control unit 24 controls the source unit 8 to activate the source requested by the user. That is, in step S201, the CD player 13 is activated if a CD playback operation is performed, the FM tuner 14 is activated if an FM reception operation is performed, and the AM tuner 15 is activated if an AM reception operation is performed. When the external input acceptance operation is performed, the circuit of the external input 16 is turned on.

  (Step S204) After the execution of the process of step S203 by the source control unit 24 is completed, if the main power supply of the audio system 1 remains on, the operation processing unit 26 executes the process of step S201 again. On the other hand, if the main power supply of the audio system 1 is turned off, execution of a series of processes is terminated.

  The second operation flow is as described above. Next, the third operation flow will be described. FIG. 11 is a diagram showing a third operation flow. Hereinafter, the third operation flow realized by the audio system 1 will be described with reference to FIG. In the third operation flow, when there is noise around the vehicle 2, the acoustic transfer characteristic estimated based on the sound signal of the noise and the transfer characteristic in each direction stored in advance in the memory 19 are shown. By obtaining the likelihood with the acoustic model 31, the direction of noise arrival is estimated, and the amplification factor of each amplifier 11 is adjusted accordingly.

  (Step S301) When the main power supply of the audio system 1 is turned on, the CPU 20 executes the computer program, whereby the noise determination unit 21, the likelihood calculation unit 22, and the amplification degree determination unit 27 are realized. The noise determination unit 21 embodied by the execution of the computer program acquires an audio signal from the microphone 5 that is amplified by the microphone amplifier 17 and converted into a digital signal by the A / D converter 18.

  (Step S302) The noise determination unit 21 analyzes the audio signal from the microphone 5, and if the power of the signal observed by the microphone 5 exceeds a predetermined threshold value continuously for, for example, 1 sec or longer (this condition is defined in the present invention). This corresponds to a predetermined condition), and the likelihood calculation unit 22 is caused to execute the processing after step S303, which is an algorithm for volume adjustment. The predetermined threshold is the volume of the ambient noise when the audio sound becomes difficult for the passenger to hear due to the ambient noise, and is configured to be arbitrarily set according to the soundproof performance of the vehicle 2 and the like. Further, the duration of the predetermined threshold required for the condition determination may be a value other than 1 sec. For example, it is set to a time shorter than when the user starts to feel uncomfortable due to external noise.

(Step S <b> 303) The likelihood calculating unit 22 calculates the MFCC (Mel Frequency Cepstrum Coefficient) from the sound signal of the noise observed with the microphone 5 by executing the following processing. The MFCC is a parameter that is widely used in speech analysis such as speech recognition, and is a parameter that expresses the spectrum of the characteristics of the acoustic transfer characteristics from the sound source (noise source) to the microphone 5 on the Mel scale frequency axis. It is. The likelihood calculating unit 22 performs high-frequency emphasis (pre-emphasis) on the input voice data obtained by the microphone 5 using Formula 1 shown below in order to calculate the MFCC.

Next, the likelihood calculating unit 22 applies the window function of Expression 2 shown below to the voice data in which the high frequency is emphasized in Expression 1 above. As a result, only the audio data in the finite section remains, and subsequent numerical analysis becomes easy.

  Next, the likelihood calculating unit 22 performs a Fourier transform (FFT: Fast Fourier Transform) on the audio data to which the window function shown in Equation 2 is applied. Thereby, the audio data is decomposed into the frequency domain.

Next, the likelihood calculating unit 22 performs mel filter bank processing of a triangular filter array on the speech data (frequency domain) after Fourier transform. FIG. 12 illustrates the outline of the mel filter bank processing. By performing Mel filter bank processing on the audio data after Fourier transform, p outputs m _j (j = 1, 2,..., P) equal to the number of triangular filters are calculated. m _j represents the power spectrum of the j th region.

The likelihood calculating unit 22 takes a natural logarithm for each calculated power spectrum value (m _j (j = 1... P)), and further performs a discrete cosine transform (DCT (Discrete Cosine Transform) according to Equation 3 below. ) -II), a cepstrum value that is a vector quantity representing the absolute position of the speech spectrum is obtained.

The likelihood calculating unit 22 further applies the following mathematical formula 4 to the cepstrum value C _n (n = 1, 2,..., 16) and performs rescaling.

The likelihood calculating unit 22 uses the n-dimensional C _n ′ calculated by the above calculation process as a MFCC parameter, and completes the process of this step.

  (Step S304) The likelihood calculating unit 22 that has calculated the MFCC parameters by executing the processing of Step S303 described above and the MFCC parameters calculated in Step S303 and the acoustic models (acoustic models 31) stored in the memory 19 (Step S304). -30... 150) and the likelihood for each direction is calculated. FIG. 13 illustrates the processing contents when calculating the likelihood for each direction. For example, when noise comes from a direction of 120 to 150 °, the likelihood for each direction is a value as shown in FIG.

In addition, the calculation method of the likelihood regarding an audio | voice signal is as follows. That is, when the feature amount of MFCC is set to o, the speech likelihood P (o) is obtained as the sum of W weighted normal distributions as shown in Equation 5 below.

Of the W normal distributions, the w-th average is μw and the variance is σw. Also, λw is a weighting coefficient that satisfies Equation 6 below.

  (Step S305) When the likelihood calculating unit 22 calculates the likelihood in each direction, the amplification determining unit 27 calculates the amplification of each amplifier 11 based on the likelihood calculated in step S304 and the coefficient map 29. To do. Specifically, the amplification degree determination unit 27 accesses the memory 19 and refers to VFR, VFL, VRR, and VRL of the current control parameter 30. Then, the likelihood of each direction calculated in step S304 is multiplied by the value corresponding to each direction of the coefficient map 29, and the sum of these is multiplied by the control parameter corresponding to each speaker.

For example, “1.00” is stored in the VFR of the memory 19, and as a result of the calculation in step S304, the likelihood in the 30-degree direction is “1”, the likelihood from the 60-degree direction is “2”, 90 Consider a case in which the likelihood from the direction of “3” is “3”, the likelihood from the direction of 120 degrees is “20”, and the likelihood from the direction of 150 degrees is “15”. In this case, since the FR of the coefficient map 29 is defined as “5 (30 degrees)” “5 (60 degrees)” “4 (90 degrees)” “4 (120 degrees)” “3 (150 degrees)”. The likelihood kFR of each direction is multiplied by the value corresponding to each direction of the coefficient map 29, and the total value kFR is as follows.
kFR = 5 * 1 + 5 * 2 + 4 * 3 + 4 * 20 + 3 * 15 = 138
The amplification degree determination unit 27 overwrites the value obtained by multiplying the calculated kFR by a predetermined correction coefficient (for example, 0.01) on the VFR of the memory 19 as a new value of VFR. Such a series of arithmetic processing is performed for each parameter, and the variables VFR, VFL, VRR, VRL stored in the memory 19 are sequentially overwritten and updated (in this example, VFR is 1.38, VFL is overwritten with 1.98, VRR with 0.83, and VRL with 0.97).

  During execution of the third operation flow, the first operation flow is also executed in parallel. Therefore, when the variables VFR, VFL, VRR, and VRL in the memory 19 are overwritten and updated by the processing of this step S305, the first operation flow is executed. The control target value of the amplification factor of each amplifier determined in step S103 of the operation flow changes before and after execution of step S305, and the volume of each speaker changes. Therefore, as shown in FIG. 13, when noise comes from a direction of 120 to 150 degrees that is the left front of the vehicle 2, VFR is changed from 1.00 to 1.38, and VFL is changed from 1.00 to 1. .98 is overwritten and updated, the amplification factors of the amplifiers 11FL and FR are increased, and the sounds of the speakers 4FL and FR are increased. As described above, by continuously updating each parameter in accordance with the arrival direction of the sound arriving at the vehicle 2, it is possible to realize sound in which the volume of each speaker 4 is balanced.

  (Step S306) After the process of step S305 by the amplification degree determination unit 27 is completed, if the main power supply of the audio system 1 remains on, the noise determination unit 21 executes the processes after step S301 again. On the other hand, if the main power supply of the audio system 1 is turned off, execution of a series of processes is terminated.

  As described above, according to the audio system 1 according to the present embodiment, the volume of each speaker is adjusted so that the volume of audio from the approach direction is increased even when noise comes from the vicinity of the vehicle 2 into the vehicle interior. Therefore, even if ambient noise occurs, the audio sound does not become difficult to hear. In particular, according to the present embodiment, since the direction of the noise source is estimated by the single microphone 5, it is much larger than that in which the direction of the noise source is estimated by the time difference or the sound difference of the sound sources by a plurality of microphones. Cost reduction and system downsizing are possible. The downsizing of the system is particularly effective in a car or the like where there is not enough room due to the installation of equipment, and the single microphone 5 makes it easy to handle wiring.

In the above embodiment, the mode selection according to the occupant's seating state is performed in step S101, and the amplification factor of each amplifier is read. However, the present invention is not limited to this, and the seated state of the passenger may be ignored. In this case, the amplification factors of all the amplifiers are made to be the same in the process in step S101 (for example, mode D is selected). According to this, since it is possible to omit the installation of the seating sensor, the system configuration can be further simplified.

  In the above embodiment, the system configuration is such that a 2ch stereo audio signal is processed. However, the present invention is not limited to this, and may be a system that processes, for example, a 1ch monaural audio signal or a 5.1ch surround audio signal.

<Modification 1>
In the above embodiment, in the process of step S302 of the third operation flow, the branch process is performed based on the power of the audio signal and the duration of the noise. However, the noise determination unit 21 performs the following noise determination. You may do as follows.

  In other words, the noise determination unit 21 uses a GMM (corresponding to the second acoustic model in the present invention) in which noise models such as running noise and horn are prepared in advance, and the sound signal observed by the microphone 5. The likelihood of GMM is calculated, and the noise type is determined. The noise type determination is performed by comparing the GMM of a plurality of types of noise prepared in advance and stored in the memory 19 with the sound signal of the microphone 5 to calculate the respective likelihoods, and specifying the noise model with the highest likelihood. . Then, processing corresponding to the identified noise model is performed.

  The process corresponding to the identified noise model is a volume adjustment process for each speaker according to the type of noise.For example, if the noise belongs to a transient type, the volume of the speaker is not adjusted. This refers to adjusting the volume of each speaker in the case of noise that continues for a long time. FIG. 14 is an operation flowchart according to this modification. Since this modification is a modification of step S302, in FIG. 14, the flow after step S301 and step S303 described above is omitted.

  As shown in FIG. 14, the noise determination unit 21 calculates the MFCC feature amount from the audio signal observed by the microphone 5 (S302A). Then, the noise determination unit 21 calculates the likelihood between the calculated MFCC feature amount and each noise model stored in advance in the memory 19. And the noise determination part 21 specifies the noise model with the highest likelihood (S302B). After identifying the noise type, the noise determination unit 21 determines the necessity of volume control for each speaker from the noise type. Determination conditions are defined in the map shown in FIG. 15 and stored in the memory 19 in advance. The noise determination unit 21 determines the necessity of volume control for each speaker by referring to the map shown in FIG. 15 (S302C).

  According to such a modification, since it is determined whether or not the speaker volume needs to be controlled according to the type of noise, the volume of the speaker does not fluctuate due to transient noise, and the occupant can hear the audio sound. You can listen comfortably.

<Modification 2>
Moreover, you may deform | transform the said embodiment and modification as follows. In other words, in the above-described embodiments and modifications, the microphone 5 is arranged at the central portion in front of the vehicle. However, in the above-described embodiment or modification, the microphone 5 may be disposed on the right side of the vehicle front (see FIG. 16). In this case, the acoustic model 55 stored in the memory 19 and the coefficient map shown in FIG. 6 need to be created with a microphone arranged on the right front. In the above-described embodiment and modification, the microphone 5 can be arranged at an arbitrary position as described above, and therefore, the attachment location of the microphone can be selected according to the vehicle type.

<Modification 3>
Moreover, you may deform | transform the said embodiment and modification as follows. That is, in the above-described embodiments and modifications, one microphone 5 is used. However, there are cases where the microphone 5 is difficult to pick up noise or the like on the side opposite to the side where the microphone 5 is disposed. In order to detect such noise more reliably, the microphones 5 may be installed at a plurality of locations (see FIG. 17). By doing in this way, for example, even when the size of the vehicle is large, such as a large vehicle, noise around the vehicle can be reliably picked up by the microphone 5. In this modification, a large number of microphones 5 are provided, but it is not necessary to provide a plurality of microphone amplifiers 17 and A / D converters 18. That is, after the signals from the plurality of microphones 5 are mixed, they are processed as one signal by the microphone amplifier 17 and sent to the CPU 20 via the A / D converter 18. However, the acoustic model 55 stored in the memory 19 and the coefficient map shown in FIG. 6 need to be created based on the plurality of microphones 5.

<Preparation of acoustic model>
Hereinafter, the acoustic model 31 used in the embodiment and each modification will be supplementarily described. The acoustic model 31 of the audio system 1 is previously stored in the memory 19 at the time of factory shipment. As in the above embodiment, the acoustic model 31 is created based on the recorded sound by recording sounds from each direction (ie, 30, 60, 90, 120, and 150 degrees) with a microphone attached to the front of the vehicle. Model. Since the vehicle for preparing the acoustic model and the user's vehicle 2 on which the audio system 1 is mounted are not the same, the acoustic model 31 stored in the memory 19 does not always match the acoustic characteristics around the vehicle 2. However, in general, in the case of a normal vehicle, the outer shape and size of the vehicle are almost the same, and it is considered that the difference in the outer shape and size of each vehicle hardly affects the acoustic model. For this reason, in the said embodiment, the said series of processes are performed using the acoustic model 31 previously stored in the memory 19 at the time of factory shipment.

The figure which shows the vehicle carrying an audio system. The system block diagram of an audio system. The figure which shows the content of the acoustic model. The figure which shows the angle of the arrival direction of noise. The figure which shows the content of the balance map. The figure which shows the content of a coefficient map. The figure which shows how to define a coefficient. The figure which shows the content of the parameter for control. The 1st operation | movement flowchart of an audio system. The 2nd operation | movement flowchart of an audio system. The 3rd operation | movement flowchart of an audio system. The figure which shows the outline | summary of a mel filter bank process. The figure which showed the outline | summary of the 3rd operation | movement flow. The operation | movement flowchart which concerns on a modification. The figure which shows the content of the determination map. The figure which shows the installation position of the microphone which concerns on a modification. The figure which shows the installation position of the microphone which concerns on a modification.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Audio system 2 ... Vehicle 3 ... Main unit 4 ... Speaker 5 ... Microphone 6 ... Amplifier unit 7 ... Selector 8 ... Source unit 9 ... Control unit 10. Operation switch panel

Claims (3)

An in-vehicle audio device that outputs an audio signal to each speaker installed in a plurality of locations in a vehicle,
When an audio signal serving as a sound source is input, amplification processing means for amplifying the audio signal and outputting it to each speaker;
Storage means for storing a first acoustic model composed of a plurality of models that are models of sound around the vehicle that arrives at the vehicle and are prepared according to the direction of arrival of the sound;
Volume adjustment for performing control according to the direction of arrival of sound around the vehicle to the amplification processing unit based on a signal from a microphone that detects sound around the vehicle and the first acoustic model stored in the storage unit Means,
In-vehicle acoustic device. The sound volume adjusting means calculates the likelihood of the signal from the microphone and the first acoustic model for each model, and a speaker installed at a position corresponding to the sound arrival direction defined in each model. Controlling the amplification processing means so that the volume is a volume according to the calculated likelihood,
The on-vehicle acoustic device according to claim 1. The storage means is a model of sound around the vehicle that arrives at the vehicle, and further stores a second acoustic model composed of a plurality of models prepared according to the type of sound,
The volume adjustment means determines the type of sound around the vehicle based on the signal from the microphone and the second acoustic model, and controls the amplification processing means according to the type of sound.
The on-vehicle acoustic device according to claim 1 or 2.

Images