Tuning device, tuning system, tuning method, and tuning program

Abstract

An acoustic device can be easily tuned regardless of the environment in which the acoustic device is used.　A tuning device for tuning an acoustic device connected to a first sound detection sensor for detecting a sound generated in a closed space and a first sound generation unit for outputting a sound into the closed space comprises: a signal output unit for outputting a signal to the first sound generation unit; a sound acquisition unit for acquiring a sound collected by the first sound detection sensor and a sound collected by a second sound detection sensor disposed inside the closed space; and an adjustment unit for adjusting a parameter used for acoustic processing of a signal transmitted through a reception-side signal path for transmitting a signal to the first sound generation unit or a transmission-side signal path for transmitting a signal recorded by the first sound detection sensor. The adjustment unit includes: a reception-side equalizer adjustment unit that, on the basis of the signal output by the signal output unit and the sound acquired by the sound acquisition unit, adjusts a filter coefficient or a gain for each frequency of an equalizer filter used in a reception-side equalizer for adjusting the frequency of a signal transmitted through the reception-side signal path; a reception volume adjustment unit that, on the basis of the signal output by the signal output unit and the sound acquired by the sound acquisition unit, adjusts a target value of a reception-side gain adjustment unit for adjusting the gain of an input signal, the reception volume adjustment unit performing processing immediately after the reception-side equalizer adjustment unit; and an adaptive filter adjustment unit that, on the basis of the signal output by the signal output unit, sets an initial value of an adaptive filter used for removing an echo generated by the sound output from the first sound generation unit, the adaptive filter adjustment unit performing processing immediately after the reception volume adjustment unit.

Description

Tuning device, tuning system, tuning method, and tuning program

[0001] The present invention relates to a tuning device, a tuning system, a tuning method, and a tuning program.

[0002] Patent Document 1 discloses an integrated microphone / sound generator device that can ensure the quality of sound collected by the microphones with a simple configuration. This integrated microphone / sound generator device includes two microphones, a sound generator disposed between the two microphones, and a control unit that combines the sounds collected from the two microphones. The distance between the two microphones is set so that when speech recognition is performed based on the sound combined by the control unit, the microphones are within a range that can collect frequencies required for speech recognition and the amount of noise caused by vibrations of the sound generator is within an acceptable range.

[0003] Patent No. 7184527

[0004] When the microphone and sound generating unit are fixed in position, tuning is required for each space (e.g., the size and shape of the vehicle's interior) in which the integrated microphone / sound generating unit device is installed in order to ensure sound quality, which increases the number of adjustment, evaluation, and management steps, resulting in increased time and costs.

[0005] The integrated microphone / sound generator device of Patent Document 1 reduces the effort required for tuning by appropriately setting the distance between the microphones. However, ensuring sound quality by adjusting the distance between the microphones limits the distance between the microphones, which in turn limits the placement, size, and shape of the sound generator and microphone of the integrated microphone / sound generator device. As a result, this may impair the uniformity and aesthetics of the design of the part where the integrated microphone / sound generator device is installed, force the repositioning of other systems, or make integration with other systems that use speakers or microphones difficult. Therefore, the integrated microphone / sound generator device of Patent Document 1 is difficult to apply to in-vehicle integrated microphone / sound generator devices and has low versatility.

[0006] The present invention has been made in consideration of the above circumstances, and aims to provide a tuning device, tuning system, tuning method, and tuning program that can easily tune an audio device regardless of the environment in which the audio device is used.

[0007] In order to solve the above-described problems, a tuning device according to the present invention is, for example, a tuning device for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, and includes a signal output unit that outputs a signal to the first sound generation unit, a sound acquisition unit that acquires the sound collected by the first sound detection sensor and the sound collected by a second sound detection sensor arranged inside the closed space, and an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiving side signal path that transmits signals to the first sound generation unit or a transmitting side signal path that transmits signals collected by the first sound detection sensor, and the adjustment unit adjusts parameters used in acoustic processing for signals transmitted through a receiving side signal path that transmits signals to the first sound generation unit or a transmitting side signal path that transmits signals collected by the first sound detection sensor, and a receiver-side equalizer adjustment unit that adjusts the filter coefficient or gain for each frequency of an equalizer filter used in a receiver-side equalizer that adjusts the frequency of a signal transmitted through the receiver-side signal path based on the signal output from the signal output unit and the sound acquired by the sound acquisition unit, the receiver-side volume adjustment unit adjusting a target value of a receiver-side gain adjustment unit that adjusts the gain of an input signal based on the signal output from the signal output unit and the sound acquired by the sound acquisition unit, the receiver-side volume adjustment unit performing processing after the receiver-side equalizer adjustment unit; and an adaptive filter adjustment unit that sets an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generation unit based on the signal output from the signal output unit, the adaptive filter adjustment unit performing processing after the receiver-side volume adjustment unit.

[0008] In order to solve the above-mentioned problems, a tuning system according to the present invention is a tuning system including, for example, a terminal having a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound inside the closed space, an acoustic device connected to the terminal, and a tuning device that tunes the acoustic device, wherein the tuning device includes a signal output unit that outputs a signal to the first sound generation unit, a sound acquisition unit that acquires sound collected by the first sound detection sensor and sound collected by a second sound detection sensor arranged inside the closed space, and an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiving-side signal path that transmits a signal to the first sound generation unit or a transmitting-side signal path that transmits a signal recorded by the first sound detection sensor, based on the data acquired by the sound acquisition unit. The adjustment unit is characterized by comprising: a receiver-side equalizer adjustment unit that adjusts the filter coefficient or gain for each frequency of an equalizer filter used in a receiver-side equalizer that adjusts the frequency of a signal transmitted through the receiver-side signal path based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit; a receiver-side volume adjustment unit that adjusts a target value of a receiver-side gain adjustment unit that adjusts the gain of an input signal based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit, the receiver-side volume adjustment unit performing processing after the receiver-side equalizer adjustment unit; and an adaptive filter adjustment unit that sets an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generation unit based on the signal output by the signal output unit, the adaptive filter adjustment unit performing processing after the receiver-side volume adjustment unit.

[0009] In order to solve the above-mentioned problems, a tuning method according to the present invention is, for example, a tuning method for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that outputs sound into the closed space, and includes an adjustment step of adjusting parameters used in acoustic processing for a signal transmitted through a receiver-side signal path that transmits a signal to the first sound generation unit or a transmitter-side signal path that transmits a signal collected by the first sound detection sensor, and the adjustment step includes adjusting parameters used in acoustic processing for a receiver-side equalizer that adjusts the frequency of a received signal based on a signal output by a signal output unit to the first sound generation unit and a sound collected by a second sound detection sensor arranged inside the closed space. The method includes a receiving-side equalizer adjustment step of adjusting the filter coefficient of an equalizer filter or a gain for each frequency, a receiving-side volume adjustment step of adjusting a target value of a receiving-side gain adjustment unit that adjusts the gain of an input signal based on the signal output by the signal output unit and the sound picked up by the first sound detection sensor, the receiving-side volume adjustment step being performed after the receiving-side equalizer adjustment step, and an adaptive filter adjustment step of setting an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generation unit, based on the signal output by the signal output unit, the adaptive filter adjustment step being performed after the receiving-side volume adjustment step.

[0010] In order to solve the above-mentioned problems, a tuning program according to the present invention is, for example, a tuning program for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, and includes a computer, a signal output unit that outputs a signal to the first sound generation unit, a sound acquisition unit that acquires sound collected by the first sound detection sensor and sound collected by a second sound detection sensor disposed inside the closed space, and an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiving side signal path that transmits signals to the first sound generation unit or a transmitting side signal path that transmits signals collected by the first sound detection sensor, based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit. and an adaptive filter adjustment unit that sets an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generating unit, based on the signal output from the signal output unit.

[0011] According to the above aspect of the present invention, the adjustment unit automatically adjusts the filter coefficient or frequency gain of the equalizer filter used in the receiving-side equalizer, the target value of the receiving-side gain adjustment unit, and the initial value of the adaptive filter used to remove echoes generated by the sound output from the first sound generation unit, based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit. This makes it possible to easily tune the acoustic device regardless of the environment in which the acoustic device is used.

[0012] The receiver-side equalizer adjustment unit may perform coefficient estimation using an autoregressive model on a first signal, which is a result of performing frequency analysis on a result of processing an arbitrary signal output to the closed space in a state where the filter coefficients of the receiver-side equalizer are initialized, or on a second signal, which is generated based on the first signal, and use the estimated coefficients as the filter coefficients of the equalizer filter, which are the parameters. This makes it possible to adjust the filter coefficients easily and accurately.

[0013] The receiver equalizer adjustment unit may generate the second signal based on a result of multiplying the first signal by a signal obtained by converting white noise into a frequency function, and perform coefficient estimation for the second signal using an autoregressive model. In this way, desired characteristics can be achieved by performing coefficient estimation for the second signal obtained by multiplying the first signal by white noise.

[0014] The receiver equalizer adjustment unit may multiply the first signal by a signal obtained by converting white noise into a frequency function, and the result may be multiplied by a signal obtained by converting a filter having an inverse characteristic of a desired characteristic into a frequency function, to obtain the second signal. This allows the equalizer filter to have the desired characteristic.

[0015] The adjustment unit may include a double-talk detection adjustment unit that sets a threshold value used in double-talk detection for detecting whether or not a double-talk state is occurring, and the double-talk detection adjustment unit may acquire information on the magnitude of a reference signal transmitted through the receiver signal path and the magnitude of a residual echo signal obtained by removing a linear echo component from the transmitter signal acquired by the first sound detection sensor, use data with the largest magnitude of the residual echo signal for each reference signal magnitude as training data, and use the coefficient of a regression curve obtained by regressing the training data with a cubic function as the threshold value. This makes it possible to automatically adjust the threshold value for detecting a double-talk state.

[0016] The double-talk detection adjustment unit may calculate the coefficients by shifting the regression curve in a direction that increases the magnitude of the residual echo signal based on a value indicating the variance of the learning data, thereby obtaining appropriate coefficients.

[0017] The double-talk detection and adjustment unit may increase the value input to the receiving-side gain adjustment unit from an initial value to obtain information about the magnitude of the reference signal and the residual echo signal, thereby improving robustness.

[0018] The adjustment unit may include a delay adjustment unit that adjusts a setting related to a delay time of the reference signal relative to the transmission signal based on a reference signal input to the first sound generation unit and a transmission signal input from the first sound detection sensor, thereby enabling automatic adjustment of the echo path delay.

[0019] The adjustment unit may include a transmitter equalizer adjustment unit that adjusts a gain used in a transmitter equalizer that adjusts the frequency of the signal output from the acoustic device, thereby enabling automatic adjustment of the gain of an equalizer filter used in the transmitter equalizer.

[0020] The adjustment unit may include a noise level estimation unit that estimates the loudness of a noise relative to the loudness of the sound acquired by the first sound detection sensor, thereby making it possible to automatically adjust the gain of the dynamic range control.

[0021] In order to solve the above problem, a tuning device according to the present invention is, for example, a tuning device for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, and is characterized by comprising: a signal output unit that outputs a signal to the first sound generation unit; a sound acquisition unit that acquires the sound picked up by the first sound detection sensor and the sound picked up by a second sound detection sensor arranged inside the closed space; an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiving-side signal path that transmits signals to the first sound generation unit or a transmitting-side signal path that transmits signals recorded by the first sound detection sensor; an evaluation unit that measures the performance of the acoustic device using the results of the tuning process performed by the adjustment unit and evaluates the measurement results; a past log storage unit that stores the results of the tuning process performed by the adjustment unit and the evaluation results by the evaluation unit; and a common parameter calculation unit that determines whether parameter sets measured under multiple conditions can be standardized based on the information stored in the past log storage unit and information about past performance, and calculates common parameters if standardization is possible. This makes it possible to standardize parameters under multiple conditions using past results. Furthermore, by standardizing parameters under multiple conditions, it is possible to use the same parameters for different vehicle models, for example, which has cost benefits.

[0022] In order to solve the above problems, a tuning device according to the present invention is a tuning device for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, the tuning device comprising: a signal output unit that outputs a signal to the first sound generation unit; a sound acquisition unit that acquires the sound picked up by the first sound detection sensor and the sound picked up by a second sound detection sensor disposed in the closed space; an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiving-side signal path that transmits signals to the first sound generation unit or a transmitting-side signal path that transmits signals recorded by the first sound detection sensor; an evaluation unit that measures the performance of the acoustic device using the results of the tuning process performed by the adjustment unit and evaluates the measurement results; a past log storage unit that stores the results of the tuning process performed by the adjustment unit and the evaluation results by the evaluation unit; and a compatibility prediction unit that predicts whether parameters under certain conditions can be used under other conditions based on information stored in the past log storage unit and information related to past performance, thereby making it possible to estimate whether the same parameters can be used under new conditions.

[0023] According to the present invention, an audio device can be easily tuned regardless of the environment in which the audio device is used.

[0024] 1 is a diagram schematically illustrating a tuning system 100 including a tuning device 1 according to a first embodiment. FIG. 1 is a diagram schematically illustrating the functional blocks of an acoustic device 56. FIG. 2 is a diagram schematically illustrating the functional blocks of the tuning device 1. FIG. 3 is a flowchart illustrating the flow of processing performed by a delay adjustment unit 131. (A) shows an example of the results recorded in step S13, and (B) shows the results of calculating the cross-correlation when the signal of (A) is obtained. FIG. 4 is a graph illustrating the comparison of multiple convergence values. FIG. 5 is a flowchart illustrating the flow of processing performed by a transmission volume adjustment unit 132. FIG. 6 is a diagram schematically illustrating the flow of processing in steps S23, S24, and S25. FIG. 7 is a flowchart illustrating the flow of processing performed by a receiving-side equalizer adjustment unit 133. FIG. 8 is a diagram schematically illustrating the flow of sound. FIG. 9 is a diagram schematically illustrating the flow of processing in steps S31 to S34 in the receiving-side equalizer adjustment unit 133. FIG. 10 is a flowchart illustrating the flow of processing performed by a receiving volume adjustment unit 134. FIG. 11 is a diagram schematically illustrating the flow of processing in steps S42 to S45. FIG. 12 is a flowchart illustrating the flow of processing performed by an adaptive filter adjustment unit 135. 1 is a diagram schematically showing the data flow of the process of extracting the convergence value of the adaptive filter. FIG. 2 is a flowchart showing the process flow performed by the transmitting-side equalizer adjustment unit 136. FIG. 3 is a diagram schematically showing the signal flow of steps S61 to S64. FIG. 4 is a diagram showing an example of the evaluation result of frequency characteristics. FIG. 5 is an example showing the result of processing performed by the transmitting-side equalizer 67. FIG. 6 is a flowchart showing the process flow performed by the double-talk detection adjustment unit 137. FIG. 7 is a diagram schematically showing the signal flow of steps S72 and S73. FIG. 8 is an example of a scatter plot of the residual echo signal [i] with respect to the reference signal [i] at a certain time. FIG. 9 is a flowchart showing the process flow performed by the noise level estimation unit 138. FIG. 10 is a diagram schematically showing the process flow of steps S81 to S83. FIG. 11 is a graph showing an example of the relationship between the sum of the power of estimated noise and the loudness of noise near the sound detection sensor 51, and an example of a regression curve. FIG. 11 is a diagram schematically showing the flow of sound acquired by the receiving-side equalizer adjustment unit 133A. FIG. 12 is a flowchart showing the process flow performed by the receiving-side equalizer adjustment unit 133A. 10 is a graph showing an example of a receiving frequency response and a target value. FIG. 11 is a diagram showing a model of a sound flow acquired by a receiving-side equalizer adjustment unit 133B.1 is a flowchart showing the flow of processing performed by the receiving-side equalizer adjustment unit 133B. FIG. 2 is a diagram schematically showing the flow of sound acquired by the receiving-side equalizer adjustment unit 133C. FIG. 3 is a flowchart showing the flow of processing performed by the receiving-side equalizer adjustment unit 133C. FIG. 4 is a diagram showing an outline of the functional blocks of the tuning device 2. FIG. 5 is a diagram schematically showing a series of processing flows of the tuning device 2. FIG. 6 is a flowchart showing the flow of processing performed by the error processing unit 190. FIG. 7 is a diagram showing detection conditions for poor connection for each tuning item. FIG. 8 is a diagram showing an outline of the functional blocks of the tuning device 3. FIG. 9 is a diagram showing an outline of the functional blocks of the common parameter calculation unit 185. FIG. 10 is a flowchart showing the processing flow of the common parameter calculation unit 185 predicting the standardization of the transmitted-voice volume parameter 132p. FIG. 11 is a flowchart showing the processing flow of the common parameter calculation unit 185 predicting the standardization of the filter coefficients 133p of the receiving-side equalizer 72. FIG. 12 is a flowchart showing the processing flow of the common parameter calculation unit 185 predicting the standardization of the received-voice volume parameter 134p. FIG. 13 is a diagram showing an outline of the functional blocks of the tuning device 4. 1 is a diagram schematically showing the flow of data in the support possibility prediction unit 186 and the information acquisition unit 187. FIG. 2 is a flowchart showing the process flow in which the support possibility prediction unit 186 predicts the commonality of a parameter (outgoing-voice volume parameter 132p) affected by the sound detection sensor 51. FIG. 3 is a diagram showing an example of past results stored in the database 200. FIG. 4 is a flowchart showing the process flow in which the support possibility prediction unit 186 predicts the commonality of a parameter (received-voice volume parameter 134p) affected by the sound generation unit 52. FIG. 5 is a flowchart showing the process flow in which the support possibility prediction unit 186 predicts the commonality of a parameter (adaptive filter initial value 135p) affected by the reflection level of the closed space 90.

[0025] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A tuning device according to an embodiment of the present invention will now be described in detail with reference to the accompanying drawings. The tuning device is a device for suppressing echoes that occur during a call in a tuning system.

[0026] 1 is a diagram schematically illustrating a tuning system 100 including a tuning device 1 according to a first embodiment. Tuning system 100 mainly comprises terminals 50 and 53 provided in a closed space 90, tuning device 1, and an acoustic device 56. Terminal 50 has a sound detection sensor 51 (corresponding to the first sound detection sensor of the present invention) and a sound generation unit 52 (corresponding to the first sound generation unit of the present invention), and terminal 53 has a sound detection sensor 54 (corresponding to the second sound detection sensor of the present invention) and a sound generation unit 55.

[0027] The sound detection sensors 51 and 54 are sensors capable of detecting sound, such as microphones or vibration sensors. The sound generation units 52 and 55 are components that generate sound through vibration or the like, such as speakers or exciters that vibrate a contact surface to produce sound. In this embodiment, the terminals 50 and 53 each have one sound detection sensor and one sound generation unit, but the number of sound detection sensors and sound generation units that the terminals 50 and 53 have is not limited to this.

[0028] The closed space 90 is an unopened, closed indoor space, such as a vehicle cabin, an office, or a room in a home. In this embodiment, the closed space 90 is a vehicle cabin. The terminals 50 and 53 are provided inside the closed space 90 and emit and detect sounds within the closed space 90.

[0029] The sound detection sensor 51 and the sound generation unit 52 of the terminal 50 are each connected to an acoustic device 56. The acoustic device 56 is a device including a processing unit that executes a memory program in which programs for operating the terminal 50 and the like are stored, and is, for example, a DSP (digital signal processor). The acoustic device 56 may be integrated with the terminal 50 or may be provided in a location separate from the terminal 50. The acoustic device 56 may also be provided in a location other than the closed space 90.

[0030] The acoustic device 56 is also included in a voice communication system in which a near-end talker using the terminal 50 performs voice communication with a far-end talker using a terminal (far-end terminal) located far away in the closed space 90. The voice signal of the far-end talker input via the far-end terminal is amplified and output by the sound generation unit 52, and the voice emitted by the user on the near-end side is collected by the sound detection sensor 51 and transmitted to the far-end talker, thereby enabling the near-end talker to make a voice call (hands-free call) without holding the terminal 50. The terminal 50 and the far-end terminal are connected by a general telephone line or the like.

[0031] Tuning device 1 is connected to terminal 50, outputs sound to terminal 50, and acquires sound from terminal 50. Tuning device 1 is also connected to terminal 53, outputs sound to terminal 53, and acquires sound from terminal 53. Note that tuning device 1 may include the function of a base station simulator that reproduces the function of a base station that transmits and receives radio waves for communication, and audio device 56 may include the function of a communication device, and tuning device 1 may be connected to terminal 50 via the base station simulator and the communication device.

[0032] The tuning device 1 may be configured, for example, by a computer system including an arithmetic unit such as a CPU (Central Processing Unit) for executing information processing, storage devices such as a RAM (Random Access Memory) and a ROM (Read Only Memory), a dedicated board mounted on various terminals (e.g., an in-vehicle device, a conference system, or a mobile terminal), and software (tuning program). The tuning program may be stored in advance in a storage medium such as an HDD built into a computer or a ROM in a microcomputer having a CPU, and then installed into the computer from there. The tuning program may also be temporarily or permanently stored in a removable storage medium such as a semiconductor memory, a memory card, an optical disk, a magneto-optical disk, or a magnetic disk.

[0033] The tuning device 1 may include part or all of the acoustic device 56, and the acoustic device 56 may include part or all of the tuning device 1. The tuning device 1 will be described in detail later.

[0034] 2 is a diagram showing an outline of the functional blocks of the acoustic device 56. Functionally, the acoustic device 56 mainly includes an echo canceller 61, a frequency analyzer (FFT unit) 62, a noise estimator 63, a noise suppressor 64, a nonlinear echo suppressor 65, a noise superimposing unit 66, a transmitter equalizer 67, a restoration unit (IFFT unit) 68, gain adjusters 69 and 70, a dynamic range controller (DRC) 71, a receiver equalizer 72, an FFT unit 73, and a double-talk detector 74. In FIG. 2, the upper signal path is a transmitter signal path that transmits an input signal input from the sound detection sensor 51, and the lower signal path is a receiver signal path that transmits a signal to the sound generator 52. Note that the functional components of the acoustic device 56 may be further divided into more components depending on the processing content, or one component may perform the processing of multiple components.

[0035] The echo canceller 61 is a functional unit that cancels echoes using an adaptive filter. The echo canceller 61 updates filter coefficients according to a given procedure, generates a pseudo echo signal from a signal transmitted through the receiving side signal path, and cancels echoes by subtracting the pseudo echo signal from a signal transmitted through the transmitting side signal path. The adaptive filter and the processing of the echo canceller 61 are already known, so a description thereof will be omitted.

[0036] The FFT units 62 and 73 perform a fast Fourier transform (FFT) on the signals. The FFT unit 62 performs a fast Fourier transform on the signal transmitted through the transmitting side signal path, in this case the signal that has passed through the echo canceller 61, and the FFT unit 73 performs a fast Fourier transform on the reference signal transmitted through the receiving side signal path. The FFT units 62 and 73 convert a time function into a frequency function.

[0037] The noise estimation unit 63 and the noise suppression unit 64 are functional units for suppressing noise from the echo-free signal to generate a suppressed signal. In this embodiment, the noise estimation unit 63 and the noise suppression unit 64 suppress noise using a spectral subtraction method, but this is an example. The noise estimation unit 63 is a functional unit that estimates noise components contained in the echo-free signal that has been converted into a frequency-domain function, i.e., an estimated noise signal, for each frequency domain. The noise suppression unit 64 is a functional unit that suppresses the noise signal from the echo-free signal based on the power spectral density S(ω) of the estimated noise signal estimated by the noise estimation unit 63 to generate a suppressed signal. The processing of the noise estimation unit 63 and the noise suppression unit 64 is already known, so description thereof will be omitted.

[0038] The nonlinear echo suppressor 65 performs echo suppression processing (processing to strongly suppress echoes) on the input signal that has passed through the noise suppressor 64. The nonlinear echo suppressor 65 changes the processing based on the detection result of the double-talk detector 74 (described in detail later). For example, the nonlinear echo suppressor 65 enables echo suppression processing when there is single talk from the far end only, and disables echo suppression processing in other cases. Furthermore, for example, the nonlinear echo suppressor 65 strongly suppresses echoes when there is single talk from the far end only, and weakly suppresses echoes in other cases. Echo suppression processing is already well known, so a detailed description will be omitted. The double-talk detector 74 inputs the result of detecting whether or not there is a double-talk state to the nonlinear echo suppressor 65 every unit time. Therefore, the nonlinear echo suppressor 65 switches between enabling and disabling echo suppression processing every unit time.

[0039] The noise superimposing unit 66 is a functional unit that generates comfort noise based on the power spectral density of the estimated noise signal estimated by the noise estimating unit 63 and superimposes the comfort noise on the signal output from the nonlinear echo suppressing unit 65. The noise superimposing unit 66 adds the power spectral density of the suppressed signal to the power spectral density of the comfort noise for each frequency domain. In other words, the noise superimposing unit 66 superimposes the two in the frequency domain. The processing performed by the noise superimposing unit 66 is already known, so a description thereof will be omitted.

[0040] The transmitting-side equalizer 67 is a functional unit that adjusts the volume of a sound of a specific frequency. The transmitting-side equalizer 67 adjusts the frequency characteristics of the sound output from the noise superimposing unit 66 and inputs the sound to the IFFT unit 68. The processing of the transmitting-side equalizer 67 is already known, so a description thereof will be omitted.

[0041] The IFFT unit 68 performs an inverse FFT (IFFT) on the input signal that has passed through the transmitting-side equalizer 67 .

[0042] The gain adjustment units 69 and 70 are functional units that adjust the gain of the input signal. Specifically, the gain adjustment units 69 and 70 adjust the degree to which the input signal is amplified (gain) to output a signal of a desired level (magnitude). The gain adjustment unit 69 (corresponding to the transmitting side gain adjustment unit of the present invention) adjusts the gain according to the magnitude of the signal output from the IFFT unit 68, and the gain adjustment unit 70 (corresponding to the receiving side gain adjustment unit of the present invention) adjusts the gain according to the magnitude of the input signal on the receiving side. The gain adjustment units 69 and 70 are, for example, AGC (auto gain control), and since they are already known, their description will be omitted.

[0043] The DRC 71 is a functional unit that adjusts the dynamic range (acoustic variation range) of the signal output from the gain adjustment unit 70. The DRC 71 also changes the compression characteristics depending on noise and the like in the environment in which the terminal 50 is installed. The DRC 71 is already known, and therefore a description thereof will be omitted.

[0044] That is, the gain adjustment unit 70 changes the input voice depending on the magnitude of the input signal from the receiving side, and the DRC 71 changes the input signal depending on the magnitude of the noise.

[0045] The receiver-side equalizer 72 is a functional unit that adjusts the volume of sound at a specific frequency. The receiver-side equalizer 72 adjusts the frequency characteristics of the sound input from the DRC 71 and outputs the sound to the sound generation unit 52. The sound whose frequency characteristics have been adjusted by the receiver-side equalizer 72 is output from the sound generation unit 52 toward the closed space 90. The receiver-side equalizer 72 is already known, and therefore a description thereof will be omitted. Note that a functional unit that performs codec or an amplifier may be provided between the receiver-side equalizer 72 and the sound generation unit 52.

[0046] The double-talk detection unit 74 is a functional unit that detects whether or not a double-talk state is occurring based on the reference signal input from the FFT unit 73 and the signal input from the noise estimation unit 63. Here, the double-talk state refers to a state in which both the near-end talker and the far-end talker are speaking, i.e., signals are transmitted simultaneously through the receiving signal path and the transmitting signal path.

[0047] For example, the double-talk detector 74 holds a threshold value set for each frequency band, compares the power spectrum value of the reference signal with the threshold value for each frequency band, and detects that the far-end talker is speaking (there is far-end speech) if the number of frequency bands in which the power spectrum value of the reference signal exceeds the threshold value is equal to or greater than a certain number.Furthermore, for example, the double-talk detector 74 detects that the near-end talker is speaking (there is near-end speech) if the signal level of the transmission signal is equal to or greater than a threshold.

[0048] Returning to the explanation of Fig. 1, the tuning device 1 functionally comprises a signal output unit 110, a sound acquisition unit 120, an adjustment unit 130, and a modification unit 140. The signal output unit 110 and the sound acquisition unit 120 are connected to an acoustic device 56. Furthermore, the signal output unit 110 is connected to a sound generation unit 55, and the sound acquisition unit 120 is connected to a sound detection sensor 54. The adjustment unit 130 and the modification unit 140 are connected to the acoustic device 56.

[0049] Functionally, the acoustic device 56 has an acoustic processing unit 561 and a parameter storage unit 562. The acoustic processing unit 561 is a functional unit that runs a program that operates each functional block (see FIG. 2) of the acoustic device 56. The parameter storage unit 562 stores initial values ​​of various parameters used in the acoustic processing unit 561. The tuning device 1 adjusts the various parameters used in the acoustic processing unit 561 and rewrites the initial values ​​of the various parameters stored in the parameter storage unit 562 with the adjusted parameters.

[0050] 3 is a diagram showing an outline of the functional blocks of the tuning device 1. Functionally, the tuning device 1 mainly includes a signal output unit 110, a sound acquisition unit 120, an adjustment unit 130, a change unit 140, and a storage unit 150. Functionally, the adjustment unit 130 mainly includes a delay adjustment unit 131, a transmission volume adjustment unit 132, a reception-side equalizer adjustment unit 133, a reception-side volume adjustment unit 134, an adaptive filter adjustment unit 135, a transmission-side equalizer adjustment unit 136, a double-talk detection adjustment unit 137, and a noise level estimation unit 138. Note that the functional components of the tuning device 1 may be further classified into more components depending on the processing content, or one component may perform the processing of multiple components.

[0051] The signal output unit 110 and the sound acquisition unit 120 are each connected to the adjustment unit 130. The adjustment unit 130 is also connected to the change unit 140.

[0052] The signal output unit 110 is a functional unit that outputs a signal to the sound generation units 52 and 55. The sound output to the sound generation unit 52 may be a signal generated by the adjustment unit 130, or may be a preset sound (for example, stored in the storage unit 150).

[0053] The sound acquisition unit 120 is a functional unit that acquires sounds picked up by the sound detection sensors 51 and 54. The sound acquisition unit 120 outputs the acquired sounds to the adjustment unit 130.

[0054] The change unit 140 is a functional unit that outputs the adjustment results obtained by the adjustment unit 130 to the acoustic device 56. In this embodiment, the change unit 140 outputs parameters that are the adjustment results obtained by the adjustment unit 130 to the acoustic device 56, and the acoustic device 56 stores the acquired parameters in the parameter storage unit 562.

[0055] The adjustment unit 130 is a functional unit that adjusts parameters used in acoustic processing of a signal transmitted through a receiving-side signal path that transmits a signal to the sound generation unit 52 or a transmitting-side signal path that transmits a signal recorded by the sound detection sensor 51, based on the data acquired by the sound acquisition unit 120. The adjustment unit 130 performs processing in the order of a delay adjustment unit 131, a transmitting-side volume adjustment unit 132, a receiving-side equalizer adjustment unit 133, a receiving-side volume adjustment unit 134, an adaptive filter adjustment unit 135, a transmitting-side equalizer adjustment unit 136, a double-talk detection adjustment unit 137, and a noise level estimation unit 138. Each functional unit of the adjustment unit 130 will be described in detail below.

[0056] The delay adjustment unit 131 is a functional unit that adjusts the delay time from when the signal is output from the DRC 71 until when it is input to the echo removal unit 61, that is, the setting related to the delay time of the reference signal relative to the transmission signal. Fig. 4 is a flowchart showing the flow of processing performed by the delay adjustment unit 131.

[0057] The delay adjustment unit 131 outputs a signal to the sound generation unit 52 via the signal output unit 110 (step S11). As a result, sound is output from the sound generation unit 52 to the closed space 90. The delay adjustment unit 131 also acquires sound picked up by the sound detection sensor 51 via the sound acquisition unit 120 (step S12). The delay adjustment unit 131 also stores (records) the signal output to the sound generation unit 52 and the sound picked up by the sound detection sensor 51 (step S13). The delay adjustment unit 131 simultaneously performs the processes of steps S11 to S13.

[0058] 5A shows an example of the results recorded in step S13. The top row shows the transmission signal (sound picked up by the sound detection sensor 51), and the bottom row shows the reference signal (signal output to the sound generation unit 52). As shown, the transmission signal is delayed by a delay time Δt relative to the reference signal. In the graph shown in FIG. 5A, when the signal is delayed, i.e., when time advances, the graph shifts in parallel to the right.

[0059] Returning to the description of FIG. 4 , the delay adjustment unit 131 then calculates the cross-correlation between the transmission signal and the reference signal and obtains the time when the correlation is maximized (step S14). FIG. 5B shows the results of calculating the cross-correlation when the signal in FIG. 5A is obtained. Here, the cross-correlation indicates the similarity between two functions (the reference signal and the transmission signal) by relatively shifting the two functions. From FIG. 5B, it can be seen that the similarity between the reference signal and the transmission signal is greatest when the delay is 207 samples, i.e., when the reference signal is delayed by 207 samples (the reference signal is translated to the right by 207 samples in FIG. 5A). This means that the delay time Δt in FIG. 5A is 207 samples.

[0060] Returning to the description of FIG. 4 , the delay adjustment unit 131 then subtracts an arbitrary correction value from the time at which the correlation obtained in step S14 is maximized, and sets the result as a new parameter (step S15). For example, if the time at which the correlation is maximized is 207 samples and the correction value is 15, the delay adjustment unit 131 subtracts 15 samples from 207 samples to set 192 samples as a new parameter. The arbitrary correction value may be set in advance and stored in the storage unit 150.

[0061] The delay adjustment unit 131 outputs the set parameters to the change unit 140, and the change unit 140 stores the set parameters in the parameter storage unit 562 (step S16). As a result, the parameters related to the delay (delay parameters 131p (see FIG. 2)) in the parameter storage unit 562 are set.

[0062] The parameter takes a positive value when the transmission signal is delayed relative to the reference signal (as shown in FIG. 5 ). Although unlikely, it is also possible that the reference signal is delayed relative to the transmission signal, in which case the parameter takes a negative value. If the reference signal is delayed relative to the transmission signal, this can be addressed by delaying the transmission signal. However, if the acoustic device 56 is not configured to delay the transmission signal, the transmission signal will be received first and echo suppression processing will not be performed. In this case, it is desirable for the delay adjustment unit 131 to output a warning via an output unit (not shown) or the like during the parameter calculation process (e.g., step S14 or S15).

[0063] The process of step S15 is not essential. However, if a correction value is not set, problems may occur in the learning process in the echo removal unit 61. The learning process and problems that may occur therein will be described below.

[0064] In the learning process, a reference signal is delayed by times t1, t2, ... tn to generate multiple delayed reference signals, and the adaptive filter's convergence values ​​are obtained based on each delayed reference signal. The adaptive filter is then corrected based on the results of comparing these convergence values. Figure 6 is a graph showing how multiple convergence values ​​are compared. The horizontal axis of Figure 6 represents the index of the filter coefficients, and the vertical axis represents the magnitude of the convergence value. Figure 6 shows values ​​for delay times tn (n is a natural number) of 160, 170, 180, 190, and 200 (units: samples). The sampling frequency used to create the graph in Figure 6 was 16 kHz, and the sampling interval for 160 samples, for example, was 10 ms.

[0065] In the example shown in Figure 6, the order of the adaptive filter is 50, and when no correction is made (graph L160), only 10 reflected components can be estimated from P160, which corresponds to the direct wave. On the other hand, when a delay is inserted (corrected) (graph L180), 30 reflected components can be estimated from P180, which corresponds to the direct wave. In other words, in Figure 6, the direct wave (the peak in Figure 5(B)) should be within the order of the adaptive filter and should be located further to the left. To achieve this, it is desirable for the delay adjustment unit 131 to perform a process of subtracting a correction value (step S15).

[0066] Returning to the explanation of Fig. 3, the transmission volume adjustment unit 132 is a functional unit that adjusts the target value of the gain adjustment unit 69, in other words, the target value of the magnitude of the signal output from the transmission side terminal 76 (see Fig. 2). Fig. 7 is a flowchart showing the flow of processing performed by the transmission volume adjustment unit 132.

[0067] The outgoing voice volume adjustment unit 132 outputs a signal to the sound generation unit 55 (see FIG. 1) via the signal output unit 110 (step S21). As a result, sound is output from the sound generation unit 55 to the closed space 90. Next, the outgoing voice volume adjustment unit 132 acquires the sound that has been picked up by the sound detection sensor 51 and passed through the gain adjustment unit 69 (step S22).

[0068] Next, the send volume adjustment unit 132 acquires a measurement result of the send loudness rating (SLR) based on the result acquired in step S22 (step S23). The SLR can be measured using a known technique (ITU-T Recommendation P.79, etc.). Next, the send volume adjustment unit 132 calculates a difference D1 between the SLR result acquired in step S24 and a target (step S24).

[0069] 8 is a diagram schematically showing the processing flow of steps S23, S24, and S25. The sound picked up by the sound detection sensor 51 is adjusted by the gain adjustment unit 69 to a signal of a desired level, and is output from the transmission end 76. At this time, the gain adjustment unit 69 uses the initial values ​​stored in the parameter storage unit 562 as parameters. The transmission volume adjustment unit 132 acquires the SLR of the sound output from the transmission end 76. For example, if the acquired SLR is 10 dB and the target SLR is 13 dB, the transmission volume adjustment unit 132 calculates that the difference between the acquired SLR result and the target is −3 dB.

[0070] Returning to the explanation of Fig. 7, next, the transmission volume adjustment unit 132 sets a target value (new transmission volume) for the gain adjustment unit 69 based on the difference calculated in step S24 (step S25). For example, the transmission volume adjustment unit 132 calculates the new transmission volume using the following formula (1). In the example shown in Fig. 8, since the difference D1 is -3 dB, if the initial value is 1, the new transmission volume is calculated as 0.71. New transmission volume = old transmission volume (initial value) x 10.0^(difference D1 / 20.0) ... (1)

[0071] Thereafter, the transmission volume adjustment unit 132 outputs the new transmission volume to the change unit 140, and the change unit 140 changes the initial value stored in the parameter storage unit 562 to the new transmission volume (step S26). As a result, the parameter (transmission volume parameter 132p (see FIG. 2)) related to the gain adjustment unit 69 in the parameter storage unit 562 is set.

[0072] Returning to the explanation of Fig. 3, the receiving-side equalizer adjustment unit 133 is a functional unit that adjusts the filter coefficient or gain for each frequency of the equalizer filter used in the receiving-side equalizer 72 that adjusts the frequency of the received signal (reference signal). In this embodiment, the receiving-side equalizer adjustment unit 133 adjusts the filter coefficient of the equalizer filter of the receiving-side equalizer 72. Fig. 9 is a flowchart showing the flow of processing performed by the receiving-side equalizer adjustment unit 133.

[0073] The receiving-side equalizer adjustment unit 133 outputs an arbitrary signal to the sound generation unit 52 via the signal output unit 110 (step S31). The receiving-side equalizer adjustment unit 133 acquires the sound output from the sound generation unit 52 to the closed space 90 via the sound detection sensor 54 and the sound acquisition unit 120, and acquires the frequency characteristics of this sound (hereinafter referred to as the receiving frequency characteristic) (step S32).

[0074] FIG. 10 is a diagram illustrating a schematic flow of sound. Sound input from the signal output unit 110 is input to the receiving-side signal path from the receiving-side terminal 75 (see FIG. 2), transmitted through the receiving-side signal path, and input to the receiving-side equalizer 72. In the receiving-side equalizer 72, the signal transmitted through the receiving-side signal path is multiplied by an equalizer filter, and the result is output from the sound generating unit 52. An amplifier may be provided upstream of the sound generating unit 52. The equalizer filter amplifies or attenuates a specific band from a flat state across the entire frequency band. At this stage, the equalizer filter is set to an initial value. In this embodiment, an FIR (Finite Impulse Response) filter is used as the equalizer filter. The receiving-side equalizer adjustment unit 133 adjusts the filter coefficients of the FIR filter.

[0075] Sound is output from the sound generating unit 52 to the closed space 90, and the sound output to the closed space 90 is acquired by the sound detection sensor 54. The sound acquired by the sound detection sensor 54 is input to the receiving-side equalizer adjusting unit 133 via the sound acquiring unit 120. The receiving-side equalizer adjusting unit 133 performs frequency analysis using octave analysis or the like to acquire a receiving frequency response (corresponding to the first signal of the present invention).

[0076] For example, if the frequency characteristics of the transfer function of the closed space 90 are those shown in graph F1, and if the frequency characteristics of the sound that has passed through the receiving-side equalizer 72 are flat as shown in graph F2, the receiving F-characteristics will be the characteristics obtained by multiplying graph F2 by graph F1, i.e., the characteristics shown in graph F1 here. On the other hand, if the frequency characteristics of the transfer function of the closed space 90 are those shown in graph F1, and if the frequency characteristics of the sound that has passed through the receiving-side equalizer 72 have the characteristics shown in graph F3 (which has the opposite characteristics to graph F2), the receiving F-characteristics will be the characteristics obtained by multiplying graph F3 by graph F1, i.e., the flat characteristics (the characteristics shown in graph F2) here. In graphs F1, F2, and F3, the horizontal axis represents frequency [Hz] and the vertical axis represents magnitude [dB].

[0077] In this embodiment, graph F3 is represented by an FIR filter. The series of constants (a1, a2...ap) given to the constant multipliers that make up the FIR filter are the filter coefficients, which are automatically calculated by the receiver equalizer adjustment unit 133. The frequency characteristics of the FIR filter are determined by determining the filter coefficients. Furthermore, if p for the filter coefficients a1, a2...ap is large, the amount of calculation increases and ripple increases, while if p is small, the desired characteristics are not achieved, so an appropriate p must be set. In this embodiment, p is set to 41.

[0078] Returning to the description of Fig. 9, the receiving-side equalizer adjustment unit 133 generates a signal x(n) for coefficient estimation based on the received frequency response (step S33), and performs coefficient estimation processing on the signal generated in step S33 (step S34).

[0079] FIG. 11 is a diagram illustrating the processing flow of steps S31 to S34 in the receiver equalizer adjustment unit 133. With the filter coefficients initialized (filter coefficients: a1 = 1, a2, ..., ap = 0), white noise is output from the sound generation unit 52 via the signal output unit 110 to acquire the received frequency response (corresponding to the first signal of the present invention) (steps S31 and S32). The arbitrary signal may be, for example, a signal containing speech or a sweep signal whose frequency continuously changes within a certain range over time. The received frequency response h(n) is then multiplied by white noise w(n) (a signal obtained by converting white noise into a frequency function). For example, in the FFT process for converting the signal into a frequency function, a window function (e.g., a Hanning window) is used to extract a portion of the signal (e.g., N = 2048 samples). The frame shift amount S is N / 2 (50% overlap). However, if the number of samples is significantly less than N=2048, there is a risk that the frequency resolution will decrease, so it is desirable to set the number of samples to around 2048.

[0080] Furthermore, the result of multiplying the received frequency response h(n) by the white noise w(n) is multiplied by a signal obtained by converting g(n) into a frequency function using FFT processing. By performing an inverse FFT on this result, a signal x(n) (corresponding to the second signal of the present invention) for coefficient estimation is obtained. Note that g(n) is a filter having the inverse characteristics of the desired characteristics.

[0081] Note that the process of multiplying g(n) by a signal converted into a frequency function is not essential. In this case, coefficient estimation is performed on signal x'(n) obtained by inverse FFT of the result (corresponding to the second signal of the present invention) of multiplying the received frequency response h(n) by white noise w(n). However, signal x'(n) that is not multiplied by g(n) has flat characteristics, and if an attempt is made to obtain a signal in a specific frequency range using signal x'(n), it may not be able to respond to sudden changes and ripples may occur. In such cases, it is desirable to use a filter (e.g., a low-pass filter or a high-pass filter) that attenuates signals of frequency components near at least one end of the desired frequency range.

[0082] Next, coefficient estimation is performed on the obtained signal x(n) using an autoregressive model (AR model). Various estimation methods, such as the Burg method, can be used. This results in the values ​​of the filter coefficients a1, a2...ap.

[0083] 11, only the processing enclosed by the dotted line is essential, and the other processing is not essential. When only the processing enclosed by the dotted line is performed, the received signal frequency characteristic h(n) is used for the signal x(n), and coefficient estimation using an autoregressive model (AR model) is performed for the received signal frequency characteristic h(n).

[0084] However, in reality, the received F characteristic h(n) takes on discrete values, and there is a risk that the filter coefficients a1, a2...ap cannot be estimated with high accuracy due to characteristics that are not visible in the received F characteristic h(n). The previous processing in Figure 11 (processing before the processing surrounded by the dotted line) converts the received F characteristic h(n) into a smooth signal x(n), and using signal x(n) makes it possible to estimate the filter coefficients a1, a2...ap with high accuracy. Therefore, it is desirable to perform all of the processing shown in Figure 11, not just the processing surrounded by the dotted line.

[0085] Returning to the description of Fig. 9, the receiving-side equalizer adjustment unit 133 sets the values ​​of the filter coefficients a1, a2 ... ap estimated by the AR model as parameters (step S35) and outputs these to the modification unit 140, which then modifies the initial values ​​of the filter coefficients a1, a2 ... ap stored in the parameter storage unit 562 to the values ​​of the filter coefficients a1, a2 ... ap estimated by the AR model (step S36). This sets the filter coefficients of the equalizer filter used in the receiving-side equalizer 72 in the parameter storage unit 562 (filter coefficients 133p (see Fig. 2)).

[0086] Returning to the explanation of Fig. 3, the received-voice volume adjustment unit 134 is a functional unit that adjusts the target value of the gain adjustment unit 70, in other words, the target value of the magnitude of the signal input to the DRC 71 and the receiving-side equalizer 72. Fig. 12 is a flowchart showing the flow of processing performed by the received-voice volume adjustment unit 134.

[0087] The received-voice volume adjustment unit 134 acquires the initial value stored in the parameter storage unit 562 (step S41). Next, the received-voice volume adjustment unit 134 inputs a signal to the receiving end 75 (see FIG. 2 ) via the signal output unit 110, and outputs sound from the sound generation unit 52 (step S42). As a result, sound is output from the sound generation unit 52 into the closed space 90.

[0088] Next, the received volume adjustment unit 134 acquires sound from the sound detection sensor 54 via the sound acquisition unit 120 (step S43), and acquires a measurement result of the receive loudness rating (RLR) based on the result acquired in step S43 (step S44). The RLR can be measured using a known technique, as with the SLR. Next, the received volume adjustment unit 134 calculates a difference D2 between the RLR result acquired in step S44 and a target (step S45).

[0089] 13 is a diagram showing a schematic flow of the process of steps S42 to S45. A signal input to the receiving end 75 is adjusted by the gain adjustment unit 70 to a desired level, and is output from the sound generation unit 52. The receiving volume adjustment unit 134 acquires the RLR of the sound output from the sound detection sensor 54. For example, if the acquired RLR is 0 dB and the target RLR is −3 dB, the receiving volume adjustment unit 134 calculates that the difference D2 between the acquired RLR result and the target is 3 dB.

[0090] Returning to the explanation of Fig. 12, next, the received voice volume adjustment unit 134 sets a target value (new received voice volume) for the gain adjustment unit 70 based on the difference calculated in step S45 (step S46). For example, the received voice volume adjustment unit 134 calculates the new received voice volume using the following formula (2). In the example shown in Fig. 8, since the difference D2 is 3 dB, if the initial value is 1, the new received voice volume is calculated as 1.41. New received voice volume = old received voice volume (initial value) × 10.0^(difference D2 / 20.0) ... (2)

[0091] Thereafter, the received-voice volume adjustment unit 134 outputs the new received-voice volume to the change unit 140, and the change unit 140 changes the initial value stored in the parameter storage unit 562 to the new received-voice volume (step S47). As a result, the parameter related to the gain adjustment unit 70 in the parameter storage unit 562 (the received-voice volume parameter 134p (see FIG. 2)) is set.

[0092] Returning to the explanation of Fig. 3, the adaptive filter adjustment unit 135 is a functional unit that sets the initial value of an adaptive filter used to remove echoes generated by the sound output from the sound generation unit 52. Fig. 14 is a flowchart showing the flow of processing performed by the adaptive filter adjustment unit 135.

[0093] The adaptive filter adjustment unit 135 inputs a signal to the receiving end 75 (see FIG. 2) via the signal output unit 110 continuously for several seconds, causing the sound generation unit 52 to output sound (step S51). As a result, sound is output continuously for several seconds from the sound generation unit 52 into the closed space 90. Then, the adaptive filter adjustment unit 135 acquires sound from the sound detection sensor 51 via the sound acquisition unit 120 (step S52).

[0094] Next, the adaptive filter adjuster 135 extracts the convergence value of the adaptive filter from the echo remover 61 (step S53). Fig. 15 is a diagram schematically showing the data flow of the process of extracting the convergence value of the adaptive filter.

[0095] The echo canceller 61 functionally includes a filter generator 61 a and an echo canceller 61 b. The filter generator 61 a is a functional unit that obtains a convergence value of an adaptive filter by delaying a reference signal. The echo canceller 61 b ​​is a functional unit that removes linear echoes from the input signal collected by the sound detection sensor 51.

[0096] The filter generating unit 61a is configured to be able to execute a learning algorithm using linear processing. The learning algorithm using linear processing is, for example, NLMS or LMS, but is not limited to these, and any known algorithm can be applied.

[0097] The filter generation unit 61a acquires multiple delayed reference signals and obtains convergence values ​​of multiple adaptive filters based on the delayed reference signals. The multiple delayed reference signals are signals generated by delaying reference signals by times t1, t2, . . . tn, respectively. The filter generation unit 61a obtains the convergence value (see FIG. 6) of the adaptive filter trained based on the delayed reference signals. Next, the filter generation unit 61a compares the convergence values ​​of the multiple adaptive filters and modifies the adaptive filter based on the comparison results. For example, for an element (index I1) that does not change depending on the delay time, the adaptive filter is modified to make the magnitude of the filter coefficient smaller than the convergence value (the learning update width is reduced). This allows the filter generation unit 61a to determine the convergence value. The adaptive filter adjustment unit 135 acquires the convergence value. Note that the adaptive filter adjustment unit 135 may include the filter generation unit 61a.

[0098] Next, the adaptive filter adjuster 135 sets the acquired convergence values ​​of the adaptive filter as new parameters (step S54) and outputs them to the changer 140, which then changes the initial values ​​of the adaptive filter stored in the parameter storage unit 562 to the acquired convergence values ​​(step S55). This sets the initial values ​​of the adaptive filter (adaptive filter initial values ​​135p (see FIG. 2)) used in the echo remover 61 in the parameter storage unit 562.

[0099] Returning to the explanation of Fig. 3, the transmitting-side equalizer adjustment unit 136 is a functional unit that adjusts the filter coefficient or gain for each frequency of the equalizer filter used in the transmitting-side equalizer 67 that adjusts the frequency of the transmitting signal. Fig. 16 is a flowchart showing the flow of processing performed by the transmitting-side equalizer adjustment unit 136.

[0100] The transmitting-side equalizer adjustment unit 136 outputs sound from the sound generation unit 55 via the signal output unit 110 (step S61). As a result, sound is output from the sound generation unit 55 into the closed space 90. Note that the sound output in step S61 may be from a sound generation unit other than the sound generation unit 55. Then, the transmitting-side equalizer adjustment unit 136 acquires sound from the sound detection sensor 51 via the sound acquisition unit 120 (step S62). Next, the transmitting-side equalizer adjustment unit 136 acquires an evaluation result of the frequency characteristics of the transmission signal acquired from the sound detection sensor 51 (step S63), and calculates the difference between the evaluation result and a target value (step S64).

[0101] 17 is a diagram showing a schematic diagram of the signal flow in steps S61 to S64. The sound output from the sound generating unit 55 to the closed space 90 is acquired by the sound detection sensor 51, transmitted through the transmitting-side signal path, passes through the transmitting-side equalizer 67 and the IFFT unit 68, and is output from the transmitting-side terminal 76. At this time, the transmitting-side equalizer 67 performs processing using the initial values ​​stored in the parameter storage unit 562. The transmitting-side equalizer adjustment unit 136 acquires the evaluation result of the frequency characteristics of the signal output from the transmitting-side terminal 76, i.e., the signal magnitude for each frequency.

[0102] Fig. 18 is a diagram showing an example of the evaluation results of frequency characteristics. In Fig. 18, the solid line represents the evaluation results of frequency characteristics, and the dotted line represents the target value. The transmitter equalizer adjustment unit 136 calculates the difference D3 between the evaluation results of frequency characteristics and the target value using the following equation (3). The difference D3 is calculated for each frequency (frequency band), and the difference D3 is calculated for each frequency bin. Difference D3 = Target value - Evaluation result of frequency characteristics (3)

[0103] If the evaluation result of the frequency characteristics is greater than the target value, the difference D3 will be a negative value (a downward arrow in the graph of FIG. 18), and if the target value is greater than the evaluation result of the frequency characteristics, the difference D3 will be a positive value (an upward arrow in the graph of FIG. 18).

[0104] Returning to the explanation of Fig. 16, the transmitting-side equalizer adjustment unit 136 then sets a gain for each frequency based on the difference calculated in step S64 (step S65). For example, the transmitting-side equalizer adjustment unit 136 calculates a parameter (new gain) using the following equation (4). The number of new gains calculated is the same as the number of differences D3, i.e., the number of frequency bins. New gain = initial value × 10.0^(difference D3 / 20.0) (4)

[0105] Thereafter, the transmitting-side equalizer adjustment unit 136 outputs the new gain to the change unit 140, and the change unit 140 changes the initial value stored in the parameter storage unit 562 to the new gain (step S66). As a result, the parameter (gain 136p (see FIG. 2)) related to the transmitting-side equalizer 67 in the parameter storage unit 562 is set.

[0106] 19 shows an example of the results of processing performed by the transmitting-side equalizer 67. In FIG. 19, the solid line indicates the processing results using the initial value, and the dashed-dotted line indicates the processing results using the new gain calculated in step S65. By using the new gain, the transmitting-side equalizer 67 adjusted the magnitude of the transmission signal so that it falls within the desired range (the area outside the shaded area).

[0107] Returning to the explanation of Fig. 3, the double-talk detection adjustment unit 137 is a functional unit that sets a threshold value used in double-talk detection to detect whether or not a double-talk state is occurring, and a suppression amount used in nonlinear echo suppression processing. Fig. 20 is a flowchart showing the flow of processing performed by the double-talk detection adjustment unit 137.

[0108] The double-talk detection adjustment unit 137 outputs a signal (received single talk) to the receiving end 75 via the signal output unit 110 and outputs sound from the sound generation unit 52 under the condition that no sound is generated in the closed space 90 (step S71). At this time, it is desirable that the double-talk detection adjustment unit 137 instructs the acoustic device 56 to increase the parameter in the gain adjustment unit 70 (amplification of received sound volume) from its initial value, thereby improving robustness.

[0109] Next, the double-talk detection adjustment unit 137 acquires the magnitude of the reference signal input to the receiving end 75 and the magnitude of the signal (residual echo signal) output to the sound generation unit 52 and obtained by the sound detection sensor 51 after the echo cancellation process (removal of the linear echo component) has been performed by the echo removal unit 61 on the transmitted signal (step S72). Because the reference signal and the residual echo signal are both time-dependent signals, the double-talk detection adjustment unit 137 converts the reference signal and the residual echo signal into a reference signal [i] and a residual echo signal [i], which are frequency-dependent signals, at regular intervals, and calculates the power of each frequency band.

[0110] 21 is a diagram schematically illustrating the signal flow in steps S72 and S73. The double-talk detection unit 74 functionally includes FFT units 74a and 74b and a calculation unit 74c. The reference signal input to the receiving terminal 75 is converted into a reference signal [i] by the FFT unit 74a and input into the calculation unit 74c. The transmission signal acquired by the sound detection sensor 51 passes through the echo removal unit 61 and becomes a residual echo signal. The residual echo signal is then converted into a residual echo signal [i] by the FFT unit 74b and input into the calculation unit 74c. The calculation unit 74c calculates the power spectrum of the reference signal [i] and the residual echo signal [i].

[0111] Next, the calculation unit 74c obtains basic training data based on the results obtained in step S72, and selects only the data with the largest residual echo signal magnitude for each reference signal magnitude from the basic training data (step S73).The calculation unit 74c outputs the training data to the double-talk detection adjustment unit 137.

[0112] The learning data will be described with reference to FIG. 22. FIG. 22 is an example of a scatter plot of the residual echo signal [i] versus the reference signal [i] at a certain time (e.g., time t1). The horizontal axis of FIG. 22 represents the logarithm of the power of the reference signal, and the vertical axis represents the logarithm of the residual echo signal. The double-talk detection unit 74 acquires a sufficient amount of reference signal and residual echo signal in advance. The calculation unit 74c then creates a scatter plot as shown in FIG. 22 using the acquired data. In FIG. 22, all of the data acquired in step S72 is plotted. All of these plots are basic learning data.

[0113] Thereafter, the calculation unit 74c calculates, for example, a point P at which the logarithm of the power of the reference signal (horizontal axis in FIG. 22) is 0 and the logarithm of the residual echo signal (vertical axis in FIG. 22) is the largest in the basic training data. ０ is left as learning data when the logarithm of the power of the reference signal is 0. In addition, the calculation unit 74c may, for example, select a point P ―５ is left as learning data when the logarithm of the reference signal power is -5. Similarly, the calculation unit 74c leaves the point where the logarithm of the residual echo signal is largest for all logarithms of the reference signal power, and sets these as learning data. In Figure 22, the learning data are represented by light-colored plots. Note that the dark-colored plots in Figure 22 are unnecessary and will not be used in the future. Note that part of the functions of the double-talk detection unit 74 may be included in the double-talk detection adjustment unit 137.

[0114] Returning to the explanation of Fig. 20, the double-talk detection adjustment unit 137 then performs regression on the learning data obtained in step S73 using a cubic function (step S74), and sets the regression coefficient as a parameter to be used by the double-talk detection unit 74 (step S75). This parameter serves as a threshold that the double-talk detection unit 74 compares with the value of the power spectrum of the reference signal to determine whether or not a double-talk state exists.

[0115] Steps S74 and S75 will be described with reference to Fig. 22. The learning data can be regressed with a cubic function, and in Fig. 22, the result of regression with a cubic function is shown as a regression curve c. The regression curve c is expressed by the following equation (5), where the coefficient a ３ , a ２ , a １ , a ０ are the parameters. Regression curve c = a ３ x ３ +a ２ x ２ +a １ x + a ０ ...(5)

[0116] Returning to the description of Fig. 20, the double-talk detection adjustment unit 137 then calculates the parameter a ３ , a ２ , a １ , a ０ to the modification unit 140, and the modification unit 140 changes the initial value stored in the parameter storage unit 562 to the parameter a ３ , a ２ , a １ , a ０ (Step S76) As a result, the parameter (threshold 137p (see FIG. 2)) related to the double-talk detector 74 in the parameter storage unit 562 and the parameter (suppression amount 139p) used in the nonlinear echo suppressor 65 are set.

[0117] The coefficient a of the regression curve c (see FIG. 22) ３ , a ２ , a １ , a ０ When calculating the coefficient a, the regression curve c' is obtained by moving the regression curve c upward (towards the direction in which the magnitude of the residual echo signal increases) based on the value indicating the variance of the learning data. ３ , a ２ , a １ , a ０ The regression curve c' can be expressed by the following equation (6). This method is particularly effective when determining the threshold value 137p used in the double-talk detection unit 74. Regression curve c'=a ３ x ３ +a ２ x２ +a １ x + (a ０ + β) ... (6)

[0118] In addition, when calculating the suppression amount 139p used in the nonlinear echo suppressor 65, the coefficient a ３ , a ２ , a １ , a ０ When calculating the residual echo signal, the regression curve c may be shifted downward (toward a smaller magnitude of the residual echo signal). For example, by setting the regression curve c'' to the regression curve c'-γ (γ is a constant), excessive suppression can be prevented.

[0119] Also, different regression curves may be used for setting the threshold 137p used for speech detection and for setting the suppression amount 139p used for echo suppression. For example, the threshold 137p may be calculated using the regression curve c', and the suppression amount 139p may be calculated using the regression curve c', or the threshold 137p may be calculated using the regression curve c', and the suppression amount 139p may be calculated using the regression curve c''.

[0120] Returning to the explanation of Fig. 3, the noise level estimation unit 138 is a functional unit that estimates the loudness of a noise relative to the loudness of a sound acquired by the sound detection sensor 51. Fig. 23 is a flowchart showing the flow of processing performed by the noise level estimation unit 138.

[0121] The noise level estimation unit 138 reproduces noise from the sound generation unit 55 or a playback device (not shown) (step S81). This noise may be white noise or pink noise, or may be a sound simulating noise generated by a moving vehicle. At this time, it is desirable to determine in advance the volume of the noise to be reproduced from the sound generation unit 55 or a playback device (not shown) so that the volume of the noise near the sound detection sensor 51 is a desired volume.

[0122] Next, the noise level estimator 138 acquires a signal from the sound detection sensor 51 via the sound acquisition unit 120 (step S82). That is, the sound detection sensor 51 acquires the noise played back in step S81. Next, the noise level estimator 138 acquires the sum of the loudnesses of the sounds acquired in step S82 (step S83). The noise level estimator 138 performs the processes of steps S81 to S83 on multiple noises of different loudness.

[0123] 24 is a diagram schematically illustrating the processing flow in steps S81 to S83. First, noise is reproduced from the sound generator 55 or the like and acquired by the sound detection sensor 51 (steps S81 and S82). It is preferable that the noise is reproduced from multiple sound generators. The noise acquired by the sound detection sensor 51 passes through the echo removal unit 61 and the FFT unit 62 and is input to the noise estimation unit 63. The noise estimation unit 63 calculates the power spectrum of the input signal (analytic signal) and calculates its sum (step S83). It is also possible to perform A-weighting correction on the analytic signal and calculate the sum based on the result.

[0124] Returning to the explanation of Fig. 23, the noise level estimator 138 acquires the noise acquired in step S83 and the magnitude of the noise near the sound detection sensor 51, calculates a regression curve showing the relationship between them (step S84), and sets parameters based on this regression curve (step S85).

[0125] Fig. 25 is a graph showing an example of the relationship between the sum of the power of the estimated noise calculated by the noise estimation unit 63 and the volume of the noise near the sound detection sensor 51, and a regression curve. In Fig. 25, the horizontal axis represents the volume of the noise near the sound detection sensor 51 (after correction using A-weighting), and the vertical axis represents the sum of the power of the estimated noise (for example, the time average of the estimated noise, or the minimum value of the estimated noise over a certain period of time). The regression curve shows that when the noise near the sound detection sensor 51 is 50 dBA, the noise level is -27 dB. The parameters may be the equation of the regression curve, or may be one or more numerical values ​​that associate the noise with the noise level.

[0126] In order to obtain accurate results in estimating the noise level, it is desirable not to generate sound from the sound generation unit 52 when acquiring the reproduced noise (steps S81 and S82). When the sound generation unit 52 does not generate sound, the echo removal unit 61 in Fig. 24 is not necessary.

[0127] Returning to the description of Fig. 23, the noise level estimator 138 then outputs the parameters found in step S85 to the changer 140, which then changes the initial values ​​stored in the parameter storage unit 562 to the found parameters (step S86). This sets the parameters related to the DRC 71 (noise level estimation 138p (see Fig. 2)) in the parameter storage unit 562.

[0128] Returning to the explanation of Fig. 2, as described above, the adjustment unit 130 changes the parameters in the parameter storage unit 562 in the following order: a parameter related to delay (delay parameter 131p), a parameter related to the gain adjustment unit 69 (transmission volume parameter 132p), a filter coefficient of the equalizer filter used in the reception-side equalizer 72 (filter coefficient 133p), a parameter related to the gain adjustment unit 70 (reception volume parameter 134p), an initial value of the adaptive filter used in the echo cancellation unit 61 (adaptive filter initial value 135p), a parameter related to the transmission-side equalizer 67 (gain 136p), a parameter related to the double-talk detection unit 74 (threshold value 137p), and a parameter related to the DRC 71 (noise level estimation 138p).

[0129] According to this embodiment, the adjustment unit 130 can automatically adjust the parameters used by the acoustic device 56 .

[0130] If the tuning device 1 were not used, an operator would have to adjust (tune) each parameter one by one, which requires a large number of measurement items and specialized knowledge of signal processing, requiring highly skilled personnel and time. Furthermore, the order in which parameters are set is important for some items (e.g., setting the receiver volume parameter 134p followed by the adaptive filter initial value 135p), and adjusting the parameters in the wrong order would require readjustment. Furthermore, if the closed space 90 is a vehicle cabin, tuning would need to be performed for each vehicle model, which would make it difficult to secure sufficient time, vehicles, equipment, and personnel.

[0131] In contrast, by using the tuning device 1, even an operator without specialized knowledge can perform tuning. Furthermore, because tuning can be performed automatically, it can easily be adapted to a variety of vehicle models. In other words, the acoustic device 56 can be easily tuned regardless of the environment in which it is used.

[0132] Furthermore, according to this embodiment, the filter coefficients of the equalizer filter are obtained by performing coefficient estimation using an autoregressive model for the received frequency response, so that the filter coefficients can be adjusted easily and accurately. Furthermore, by estimating the filter coefficients of the equalizer filter using an autoregressive model, there is an advantage that the filter coefficients can be easily tuned to obtain desired characteristics.

[0133] In the embodiment of the present invention, the adjustment unit 130 includes the delay adjustment unit 131, the transmission volume adjustment unit 132, the reception side equalizer adjustment unit 133, the reception volume adjustment unit 134, the adaptive filter adjustment unit 135, the transmission side equalizer adjustment unit 136, the double-talk detection adjustment unit 137, and the noise level estimation unit 138, but the configuration of the adjustment unit 130 is not limited to this. Furthermore, in the embodiment of the present invention, the acoustic device 56 includes an echo removal unit 61, a frequency analyzer (FFT unit) 62, a noise estimation unit 63, a noise suppression unit 64, a nonlinear echo suppression unit 65, a noise superposition unit 66, a transmitting-side equalizer 67, a restoration unit (IFFT unit) 68, gain adjustment units 69 and 70, a DRC 71, a receiving-side equalizer 72, an FFT unit 73, and a double-talk detection unit 74, but the configuration of the acoustic device 56 is not limited to this.

[0134] For example, the adjustment unit 130 may at least include a receiving-side equalizer adjustment unit 133, a receiving-side volume adjustment unit 134, and an adaptive filter adjustment unit 135. The acoustic device 56 may at least include an echo cancellation unit 61, a gain adjustment unit 70, and a receiving-side equalizer 72. This makes it possible to adjust the minimum necessary tuning items (the filter coefficients (filter coefficients 133p) of the equalizer filter used in the receiving-side equalizer 72, the parameters (received-side volume parameters 134p) related to the gain adjustment unit 70, and the initial values ​​of the adaptive filter used in the echo cancellation unit 61 (adaptive filter initial values ​​135p)). Also, for example, the adjustment unit 130 may have a transmission volume adjustment unit 132 in addition to a receiving-side equalizer adjustment unit 133, a receiving-side volume adjustment unit 134, and an adaptive filter adjustment unit 135, and the acoustic device 56 may have a gain adjustment unit 69 in addition to an echo removal unit 61, a gain adjustment unit 70, and a receiving-side equalizer 72. Although it is desirable to perform adjustment by the transmission volume adjustment unit 132, the parameter related to the gain adjustment unit 69 (transmission volume parameter 132p) is easy to estimate empirically, and therefore the transmission volume adjustment unit 132 can be omitted from the essential components.

[0135] In the embodiment of the present invention, the acoustic device 56 includes the DRC 71, but the configuration of the acoustic device 56 is not limited to this. For example, if the closed space 90 is a vehicle interior and the acoustic device 56 is an in-vehicle device, a road noise volume control (RVC) may be provided before or after the DRC 71. The RVC is a functional unit that cancels out noise generated by traveling on a road. Like the DRC 71, the RVC performs processing using the noise level estimation 138p.

[0136] Furthermore, in the embodiment of the present invention, the parameters set by the adjustment unit 130 are output to the acoustic device 56, and the acoustic device 56 stores the acquired parameters in the parameter storage unit 562, but the method of storing the parameters is not limited to this. For example, a build environment may be prepared in the tuning device 1, and an object including the parameters set by the adjustment unit 130 may be output from the tuning device 1 to the acoustic device 56, and the acoustic device 56 may operate in accordance with the output object. Even with this method, the parameters adjusted by the adjustment unit 130 can be stored in the acoustic device 56.

[0137] Furthermore, in the present invention, the adjustment unit 130 can adjust the parameters using a wideband signal, or the adjustment unit 130 can adjust the parameters using a signal obtained by narrowing the frequency band of a wideband signal to a narrowband signal. By limiting the frequency of the wideband signal using a filter to make it a narrowband signal, it is possible to eliminate the process of preparing an evaluation system for the narrowband.

[0138] <Modification 1 of the Receiver-Side Equalizer Adjustment Unit> In this embodiment, a second-order FIR filter is used as the equalizer filter, and the values ​​of the filter coefficients a1, a2...ap are estimated using an AR model, but the method of determining the filter and parameters of the equalizer filter is not limited to this. Below, we will explain the processing of the receiver-side equalizer 72A and receiver-side equalizer adjustment unit 133A according to this modification. Everything except the receiver-side equalizer and receiver-side equalizer adjustment unit is the same as in tuning device 1.

[0139] The receiving-side equalizer adjustment unit 133A is a functional unit that adjusts the filter coefficients of the equalizer filter used in the receiving-side equalizer 72A that adjusts the frequency of the received signal (reference signal). Fig. 26 is a diagram schematically showing the flow of sound acquired by the receiving-side equalizer adjustment unit 133A.

[0140] The sound input from the signal output unit 110 is input to the receiving-side signal path from the receiving-side terminal 75, transmitted through the receiving-side signal path, and input to the receiving-side equalizer 72A. In the receiving-side equalizer 72A, the signal transmitted through the receiving-side signal path is multiplied by an equalizer filter, and the result is output from the sound generation unit 52. In this modification, N peaking filters are used as the equalizer filters, and the characteristics of the receiving-side equalizer 72A are determined by superimposing the N peaking filters. In this modification, an IIR (Infinite Impulse Response) filter is used as the peaking filter, and the coefficients of the IIR filter are determined by determining the frequency, bandwidth, and gain. The sound output from the sound generation unit 52 is acquired by the sound detection sensor 54 and input to the receiving-side equalizer adjustment unit 133A.

[0141] 27 is a flowchart showing the flow of processing performed by the receiving-side equalizer adjustment unit 133A. The receiving-side equalizer adjustment unit 133A outputs a signal to the sound generation unit 52 via the signal output unit 110 (step S101). At this stage, the coefficients of the equalizer filter are set to initial values ​​(gain1=1, gain2=1, ..., gain N = 1). The receiving-side equalizer adjustment unit 133A acquires the sound output from the sound generation unit 52 to the closed space 90 via the sound detection sensor 54 and the sound acquisition unit 120, and performs frequency analysis on the sound to acquire the receiving FC (step S102). Next, the receiving-side equalizer adjustment unit 133A calculates the difference (gain) between the receiving FC acquired in step S102 and a target value (step S103).

[0142] Next, the receiving-side equalizer adjustment unit 133A groups similar gains (±Y [dB]) at nearby frequencies (±X [Hz]) with respect to the gain calculated in step S103 (step S104). Note that X and Y can be set arbitrarily. Then, the receiving-side equalizer adjustment unit 133A sets parameters based on the grouping results (step S105).

[0143] The processing of steps S104 and S105 will be described with reference to FIG. 28. FIG. 28 is a graph showing an example of the received FC and the target value. In FIG. 28, the solid line represents the received FC, and the two-dot chain line represents the target value. In step S104, for example, in a frequency band FB1 where the frequency F1 is near 200 Hz (a 100 Hz range centered on 200 Hz), the difference between the received FC and the target value is similar, so the gain G1 of the frequency band FB1 is grouped and the gain G1 is set to +10 dB. Also, in step S104, for example, in a frequency band FB2 where the frequency F1 is near 1200 Hz (a 10 Hz range centered on 1200 Hz), the difference between the received FC and the target value is similar, so the gain G2 of the frequency band FB2 is grouped and the gain G2 is set to +9 dB.

[0144] In step S105, the top N peaking filters are selected in descending order of absolute gain value (vertical difference in FIG. 28 ), and these are set as N peaking filters. For example, the gain G1 at frequency F1=200 Hz has the largest absolute gain value, so the first peaking filter is set as follows: frequency f1=frequency F1, bandwidth BW1=bandwidth FB1 (100 Hz), and gain Gain1=gain G1 (+10 dB). For example, the second peaking filter is set as follows: gain G2 at frequency F2=1200 Hz, so the second peaking filter is set as follows: frequency f2=frequency F2, bandwidth BW2=bandwidth FB2 (10 Hz), and gain Gain2=gain G2 (+9 dB).

[0145] Returning to the explanation of Fig. 27, the receiving-side equalizer adjustment unit 133A sets IIR filter coefficients as parameters based on the set N peaking filters, their frequencies, bandwidths, and gains (step S105), and outputs these to the change unit 140. The change unit 140 then changes the filter coefficients stored in the parameter storage unit 562 to the filter coefficients determined in step S105 (step S106). In this way, the filter coefficients of the equalizer filters used in the receiving-side equalizer 72A are set in the parameter storage unit 562.

[0146] <Modification 2 of the receiving-side equalizer adjustment unit> The following describes the processing of the receiving-side equalizer 72B and receiving-side equalizer adjustment unit 133B according to this modification. Everything except the receiving-side equalizer and receiving-side equalizer adjustment unit is the same as in the tuning device 1. Note that parts that are the same as those in the receiving-side equalizer adjustment unit 133A are given the same reference numerals and descriptions thereof will be omitted.

[0147] The receiving-side equalizer adjustment unit 133B is a functional unit that adjusts the filter coefficients of the equalizer filter used in the receiving-side equalizer 72B that adjusts the frequency of the received signal (reference signal). Fig. 29 is a diagram schematically showing the flow of sound acquired by the receiving-side equalizer adjustment unit 133B.

[0148] The sound input from the signal output unit 110 is input from the receiving terminal 75 and transmitted through the receiving signal path, and the signal transmitted through the receiving signal path is multiplied by an equalizer filter in the receiving equalizer 72B, and output from the sound generation unit 52. In this modification, N peaking filters are used as the equalizer filters, and the peaking filters are second-order IIR filters (biquadratic filters) with both the numerator and denominator. The sound output from the sound generation unit 52 is acquired by the sound detection sensor 54 and input to the receiving equalizer adjustment unit 133B.

[0149] 30 is a flowchart showing the flow of processing performed by the receiving-side equalizer adjustment unit 133B. The receiving-side equalizer adjustment unit 133B sets the coefficients of the equalizer filter to initial values ​​(gain1=1, gain2=1, ..., gain N With the frequency response set to 0 (=1), sound is output from the sound generation unit 52 (step S101). The receiving-side equalizer adjustment unit 133B acquires the receiving FC via the sound detection sensor 54 and the sound acquisition unit 120 (step S102). Next, the receiving-side equalizer adjustment unit 133B calculates the difference D4 between the receiving FC and the target value (step S103).

[0150] Next, the receiver equalizer adjustment unit 133B modifies the provisional parameters using an optimization algorithm (e.g., gradient descent, Newton's method, or genetic algorithm) so as to minimize a cost function based on the difference D4 and the frequency response of the provisional parameters, and adopts the parameters that minimize the cost function as the final parameters (step S107). Here, the cost function can be expressed by the following equation (7): Cost function = difference D4 - frequency response of provisional parameters (6)

[0151] The frequency response can be calculated as follows: Since the transfer function of each filter is known (see formula (7)), the frequency response is calculated by substituting z = exp(jf / 2π) (j is the imaginary unit) into the product of each filter (see formula (8)). H1(z) = (b0 + b1*z^-1 + b2*z^-2) / (1 + a1*z^-1 + a2*z^-2) ・・・(7) Here, b0 = gain1 * (1 + (BW1 / f1)^2) b1 = -2 * gain1 * (1 + (BW1 / f1)^2) b2 = gain1 * (1 + (BW1 / f1)^2) a1 = -2 * (1 - (BW1 / f1)^2) a2 = (1 - (BW1 / f1)^2) H_total(f) = H1(f) * H2(f) * ... * HN(f) ... (8)

[0152] Alternatively, the frequency response may be determined from past results that are most similar to the desired frequency (for example, the frequency with the largest vertical axis of the desired frequency response). This reduces the amount of calculation required. If no similar result can be obtained from past results, all gains may be set to 1 and BW=100, and the frequencies may be set in descending order of vertical value.

[0153] The use of an IIR filter, such as in the receiver-side equalizer adjustment units 133A and 133B, has the advantage of being able to reduce the amount of processing and to perform processing even with an inexpensive CPU or the like. Furthermore, since tuning using an IIR filter has been common for some time, the use of an IIR filter, such as in the receiver-side equalizer adjustment units 133A and 133B, is highly versatile. The use of an IIR filter also has the advantage of small delay. However, in order to perform tuning to achieve desired characteristics, it is desirable to employ a receiver-side equalizer adjustment unit 133 that uses an autoregressive model.

[0154] <Modification 3 of the receiving-side equalizer adjustment unit> The following describes the processing of the receiving-side equalizer 72C and receiving-side equalizer adjustment unit 133C according to this modification. Everything except the receiving-side equalizer and receiving-side equalizer adjustment unit is the same as in the tuning device 1. Note that parts that are the same as those in the receiving-side equalizer adjustment unit 133 are given the same reference numerals and descriptions thereof will be omitted.

[0155] The receiving-side equalizer adjustment unit 133C is a functional unit that adjusts the gain for each frequency of the equalizer filter used in the receiving-side equalizer 72C that adjusts the frequency of the received signal (reference signal). Fig. 31 is a diagram schematically showing the flow of sound acquired by the receiving-side equalizer adjustment unit 133C.

[0156] The sound input from the signal output unit 110 is input from the receiving terminal 75, transmitted through the receiving signal path, and input to the receiving equalizer 72C. The receiving equalizer 72C has an FFT unit 72a before the receiving equalizer 72 and an IFFT unit 72b after the receiving equalizer 72. The signal that has passed through the receiving equalizer 72C is output from the sound generation unit 52 to the closed space 90. The sound output to the closed space 90 is acquired by the sound detection sensor 54 and input to the receiving equalizer adjustment unit 133C.

[0157] 32 is a flowchart showing the flow of processing performed by the receiving-side equalizer adjustment unit 133C. The receiving-side equalizer adjustment unit 133C initializes the gain of the receiving-side equalizer 72C to 0 dB and outputs sound from the sound generation unit 52 (step S109). The receiving-side equalizer adjustment unit 133C acquires sound via the sound detection sensor 54 and the sound acquisition unit 120, acquires an evaluation result of the frequency characteristics of the acquired sound (step S110), and calculates a difference D5 between the evaluation result and a target value (step S111). The difference D5 is calculated for each frequency (frequency band), and the difference D5 is calculated for the number of frequency bins using the following mathematical formula (9): Difference D5 = Target Value - Evaluation Result of Frequency Characteristics (9)

[0158] Next, the receiving-side equalizer adjustment unit 133C sets a gain for each frequency based on the difference calculated in step S111 (step S112). For example, the receiving-side equalizer adjustment unit 133C calculates a parameter (new gain) using the following formula (10). The new gain is calculated for each difference D5, that is, for each frequency bin. New Gain = Initial Value × 10.0^(Difference D5 / 20.0) (10)

[0159] Thereafter, the receiving-side equalizer adjustment unit 133C outputs the new gain to the change unit 140, and the change unit 140 changes the initial value stored in the parameter storage unit 562 to the new gain (step S113). Note that the processing of steps S110 to S113 is similar to the processing of steps S62 to S66 by the transmitting-side equalizer adjustment unit 136, and therefore detailed description thereof will be omitted.

[0160] In this way, the receiver equalizer adjustment unit can adjust the gain for each frequency used in the receiver equalizer. However, this increases the amount of processing compared to adjusting the filter coefficients of the equalizer filter used in the receiver equalizer, and delays can easily become a problem.

[0161] Second Embodiment A tuning device 2 according to a second embodiment will be described below. Note that the same components as those in the tuning device 1 according to the first embodiment are given the same reference numerals, and descriptions thereof will be omitted.

[0162] 33 is a diagram showing an outline of the functional blocks of the tuning device 2. Functionally, the tuning device 2 mainly includes a signal output unit 110, a sound acquisition unit 120, an adjustment unit 130, a change unit 140, a storage unit 150, a user instruction unit 160, an evaluation unit 170, a past log storage unit 180, and an error processing unit 190. Note that the functional components of the tuning device 2 may be further classified into more components depending on the processing content, or one component may execute the processing of multiple components.

[0163] The user instruction unit 160 is a functional unit (described in detail below) that issues instructions to the user using the tuning device 2 and outputs the progress and results of processing. The user instruction unit 160 outputs the progress, results, etc. to a display device (not shown) or a sound display unit (not shown). The evaluation unit 170 is a functional unit that measures the performance of the acoustic device 56 using the results of the tuning process performed by the adjustment unit 130 and evaluates the measurement results. The past log storage unit 180 is a functional unit that stores the results of the tuning process performed by the adjustment unit 130 and the evaluation results performed by the evaluation unit 170. The error processing unit 190 is a functional unit that performs a stop process or a recovery process when an error occurs during processing by the adjustment unit 130 (described in detail below). An error could be, for example, a case where processing stops midway due to a poor connection or the like.

[0164] 34 is a diagram showing a schematic flow of a series of processes performed by the tuning device 2. The display device has an image processing unit, and generates and displays an image on a monitor based on a signal output by the user instruction unit 160. Display screens M1 to M9 in FIG. 23 are examples of display screens displayed on the monitor by the user instruction unit 160.

[0165] First, the user instruction unit 160 displays a display screen M1 on the monitor, which prompts the user to select whether or not to start tuning processing. When the user presses the soft button m1, the user instruction unit 160 displays a display screen M2 on the monitor, which prompts the user to input conditions (such as the vehicle model).

[0166] When the user inputs data into box m2 and presses soft button m3, user instruction unit 160 displays display screen M3 on the monitor, which instructs the user to set the environment. Display screen M3 includes an instruction for the placement position of terminal 53. When the user presses soft button m4, user instruction unit 160 issues an instruction to adjustment unit 130, which then starts processing.

[0167] Furthermore, the user instruction unit 160 displays on the monitor display screens M4, M5, etc., which show the progress of the processing by the adjustment unit 130. The adjustment unit 130 performs processing (calibration processing) in the following order: adjustment of the delay parameter 131p by the delay adjustment unit 131, adjustment of the transmission volume parameter 132p by the transmission volume adjustment unit 132, adjustment of the filter coefficient 133p by the reception-side equalizer adjustment unit 133, adjustment of the reception volume parameter 134p by the reception-side equalizer adjustment unit 134, adjustment of the adaptive filter initial value 135p by the adaptive filter adjustment unit 135, adjustment of the gain 136p by the transmission-side equalizer adjustment unit 136, adjustment of the threshold 137p and the suppression amount 139p by the double-talk detection adjustment unit 137, and adjustment of the noise level estimation unit 138, and outputs information regarding the progress of the processing (information indicating what processing is currently being performed) to the user instruction unit 160. The user instruction unit 160 then displays on the monitor the processing currently being performed by the adjustment unit 130 and its progress.

[0168] For example, when the delay adjustment unit 131 is performing the adjustment process of the delay parameter 131p, the user instruction unit 160 displays on the monitor a display screen M4 indicating that the first process (STEP 1) is being performed, and also displays the progress of the process by the delay adjustment unit 131 on a progress display bar m5 on the display screen M4.

[0169] When the calibration process by the adjustment unit 130 is completed, the adjustment unit 130 issues instructions to the user instruction unit 160 and the evaluation unit 170, the user instruction unit 160 displays the display screen M6 on the monitor, and the evaluation unit 170 starts the measurement process. Note that the display screen M6 is a screen indicating that the tuning process has been completed and that the measurement process will begin.

[0170] The evaluation unit 170 outputs information regarding the progress of the measurement process and the evaluation process to the user instruction unit 160. Then, the user instruction unit 160 displays a display screen M87 on the monitor that shows the progress of the processing by the evaluation unit 170. The evaluation unit 170 performs the evaluation process after the measurement process and outputs the evaluation result to the adjustment unit 130 and the user instruction unit 160. When the user instruction unit 160 obtains the evaluation result from the evaluation unit 170, it displays a display screen M8 on the monitor that shows the evaluation result.

[0171] The evaluation unit 170 may output an evaluation report to the user instruction unit 160 instead of the evaluation result. The evaluation unit 170 can create an evaluation report with specifications previously specified by a user or the like based on the evaluation result. The evaluation unit 170 may also generate acoustic analysis data used during adjustment, information indicating progress (timestamp, error log, etc.), and output these to the user instruction unit 160.

[0172] Furthermore, upon acquiring the evaluation results from the evaluation unit 170, the adjustment unit 130 associates the determined parameters with the evaluation results and stores them in the past log storage unit 180. Furthermore, the adjustment unit 130 may acquire information input to the display screen M2 before measurement from the user instruction unit 160, and store this information in association with the parameters and evaluation results in the past log storage unit 180. This completes the series of processes of the tuning device 2.

[0173] It is sufficient if the processing of the tuning device 2 proceeds normally, but in some cases, the processing of the adjustment unit 130 may be interrupted due to reasons such as the connection between the tuning device 2 (base station simulator) and the acoustic device 56 being cut off during the processing of the adjustment unit 130. If the processing of the adjustment unit 130 is interrupted, the adjustment unit 130 issues an instruction to the error processing unit 190, and the error processing unit 190 performs the processing.

[0174] 35 is a flowchart showing the flow of processing performed by the error processing unit 190 when the connection between the tuning device 2 and the sound device 56 is cut off. The error processing unit 190 detects whether the cause of the error is a poor connection, i.e., whether a poor connection has actually occurred between the tuning device 2 and the sound device 56 (step S120). Because a poor connection prevents sound from being produced, this detection is possible for tuning items that produce sound, i.e., all tuning items.

[0175] 36 is a diagram showing an example of a detection condition for a poor connection for each tuning item. For example, when the transmission volume adjustment unit 132 is performing processing (item: transmission volume) and when the reception volume adjustment unit 134 is performing processing (item: reception volume), the error processing unit 190 detects that a poor connection has occurred when the SLR calculated by the transmission volume adjustment unit 132 and the RLR calculated by the reception volume adjustment unit 134 become large values ​​(e.g., 30 dB or more).

[0176] Returning to the explanation of Fig. 35, if the occurrence of a poor connection cannot be detected (NO in step S120), the error is due to a reason other than the disconnection of the tuning device 2 and the audio device 56, and therefore the error processing unit 190 instructs the adjustment unit 130 to stop the processing, and outputs information to the user instruction unit 160 recommending that the processing be restarted from the beginning (step S124).

[0177] If a connection failure is detected (YES in step S120), the error processing unit 190 requests the base station simulator of the tuning device 2 to reconnect (step S121) and requests the communication device of the acoustic device 56 to connect (step S122).

[0178] Next, the error processing unit 190 instructs the adjustment unit 130 to resume the adjustment by returning to the adjustment immediately before the adjustment that was being performed when the error occurred (step S123). For example, if an error occurs while the adaptive filter adjustment unit 135 is performing processing, the error processing unit 190 instructs the adjustment unit 130 to resume from the processing of the received voice volume adjustment unit 134, which performed processing before the adaptive filter adjustment unit 135. Note that in step S123, the error processing unit 190 may instruct the adjustment unit 130 to restart the adjustment processing from the beginning.

[0179] According to this embodiment, the user can easily use the tuning device 2. Furthermore, tuning can be resumed even if an error occurs during tuning.

[0180] Third Embodiment A tuning device 2 according to a third embodiment will be described below. Note that the same components as those in the tuning devices 1 and 2 according to the first and second embodiments are given the same reference numerals, and descriptions thereof will be omitted.

[0181] 37 is a diagram showing an outline of the functional blocks of the tuning device 3. Functionally, the tuning device 3 mainly includes a signal output unit 110, a sound acquisition unit 120, an adjustment unit 130, a change unit 140, a storage unit 150, a user instruction unit 160, an evaluation unit 170, a past log storage unit 180, a common parameter calculation unit 185, and an error processing unit 190. Note that the functional components of the tuning device 3 may be further classified into more components depending on the processing content, or one component may execute the processing of multiple components.

[0182] The common parameter calculation unit 185 is a functional unit that examines whether parameter sets measured under multiple conditions can be standardized, and calculates common parameters if they can be standardized. A parameter set is a set of parameters measured by the adjustment unit 130 under certain conditions, and includes a delay-related parameter (delay parameter 131p), a parameter related to the gain adjustment unit 69 (transmitted-voice volume parameter 132p), a filter coefficient (filter coefficient 133p) of an equalizer filter used in the receiver-side equalizer 72, a parameter related to the gain adjustment unit 70 (received-voice volume parameter 134p), an initial value (adaptive filter initial value 135p) of an adaptive filter used in the echo cancellation unit 61, a parameter related to the transmitter-side equalizer 67 (gain 136p), a parameter related to the double-talk detection unit 74 (threshold value 137p), and a parameter related to the DRC 71 (noise level estimation 138p). Note that the parameter set does not need to include all of the above parameters.

[0183] 38 is a diagram showing an outline of the functional blocks of the common parameter calculation unit 185. Functionally, the common parameter calculation unit 185 has a variation range calculation unit 185a, an evaluation result examination unit 185b, and a new parameter calculation unit 185c.

[0184] The change range calculation unit 185a is a functional unit that calculates the change range or average value of a parameter based on multiple parameter sets. Multiple parameter sets (e.g., parameter sets for car 1, car 2, ... car N) stored in the past log storage unit 180 are input to the change range calculation unit 185a. The change range calculation unit 185a outputs the calculation result to the new parameter calculation unit 185c.

[0185] The evaluation result review unit 185b is a functional unit that reviews evaluation result sets. Multiple evaluation result sets (e.g., evaluation result sets for car 1, car 2, ..., car N) stored in the past log storage unit 180 are input to the evaluation result review unit 185b. The evaluation result sets are the results of evaluation performed by the evaluation unit 170 using parameter sets measured by the adjustment unit 130, and are stored in the past log storage unit 180 for each parameter set. For example, the parameter set when the cabin of car 1 is a closed space 90 (condition: car 1) is parameter set 1, and the evaluation result set when parameter set 1 is used is evaluation result set 1. The evaluation result review unit 185b also acquires the evaluation criteria for each parameter from the past log storage unit 180 along with the evaluation result sets. The evaluation result review unit 185b outputs the review results to the new parameter calculation unit 185c.

[0186] The new parameter calculation unit 185c is a functional unit that examines whether parameters can be standardized based on information from the variation range calculation unit 185a and the evaluation result examination unit 185b, and calculates new standardized parameters if standardization is possible. A database 200 in which past performance data is stored is connected to the new parameter calculation unit 185c, and the new parameter calculation unit 185c acquires information on the past performance data from the database 200.

[0187] Next, a description will be given of the processing of the common parameter calculation unit 185. Below, an example will be described in which the common parameter calculation unit 185 acquires the parameter sets and evaluation result sets of cars 1, 2, and 3, and considers whether the parameter sets of cars 1, 2, and 3 can be made common.

[0188] 39 is a flowchart showing the flow of processing by the common parameter calculation unit 185 to predict the commonality of the transmission volume parameter 132p. First, the change range calculation unit 185a acquires multiple parameter sets, acquires the transmission volume parameter 132p included in the parameter set, and calculates the maximum value V1max and minimum value V1min of the transmission volume parameter 132p (step S131). For example, if the transmission volume parameter 132p of car 1 is +0 dB, the transmission volume parameter 132p of car 2 is +3.5 dB, and the transmission volume parameter 132p of car 3 is +3.5 dB, the change range calculation unit 185a sets the maximum value V1max of the transmission volume parameter 132p to +3.5 dB and the minimum value V1min to +0 dB. The change range calculation unit 185a outputs the result of step S131 to the new parameter calculation unit 185c.

[0189] Furthermore, the evaluation result review unit 185b acquires multiple evaluation result sets and parameter evaluation criteria, and calculates the minimum margin of the evaluation results (step S132). For example, if the evaluation result of the transmission volume of car 1 is 10 dB, the evaluation result of car 2 is 11 dB, and the evaluation result of car 3 is 12 dB, and the evaluation criteria (spec) of the transmission volume are 9 dB or more, the evaluation result review unit 185b calculates that the minimum margin is 1 dB. The evaluation result review unit 185b outputs the results of step S132 to the new parameter calculation unit 185c.

[0190] The order of the processing in step S131 and the processing in step S132 does not matter.

[0191] Next, the new parameter calculation unit 185c calculates the result C1 by subtracting the required margin from the minimum margin calculated in step S132 (step S133). For example, if the minimum margin is 1 dB and the required margin is 1 dB, the evaluation result review unit 185b calculates the subtraction result C1 to be 0 dB. The required margin may be set in advance or may be calculated based on information obtained from the database 200. For example, if the database 200 stores information indicating that the measurement variation during evaluation of the transmitted voice volume is ±0.5 dB, the new parameter calculation unit 185c determines that a minimum margin of 1 dB from the specifications is required.

[0192] Next, the new parameter calculation unit 185c calculates the maximum value C2 by which the parameter can be changed based on the information obtained from the database 200 (step S134). For example, if the database 200 stores information that lowering the transmission volume by 1 dB will increase (worse) the evaluation result by 1 dB, and that raising the transmission volume will worsen the noise level, and that a noise level of +4.0 dB is acceptable but +4.5 dB is unacceptable, the new parameter calculation unit 185c sets the maximum value C2 by which the parameter can be changed to +4.0 dB.

[0193] Next, the new parameter calculation unit 185c determines whether or not the parameters can be consolidated into one based on the subtraction result C1 obtained in step S133 and the maximum value C2 to which the parameters can be changed obtained in step S134 (step S135). Specifically, the new parameter calculation unit 185c determines whether or not the parameters can be consolidated into one by comparing the sum C3 of the minimum value V1min and the maximum value C2 with the sum C4 of the maximum value V1max and the subtraction result C1.

[0194] If the sum C3 is equal to or greater than the sum C4 (maximum value V1max+subtraction result C1) (sum C3≧sum C4) (YES in step S135), the new parameter calculation unit 185c determines that the parameters can be consolidated into one, and sets the sum C4 (maximum value V1max+subtraction result C1) as the new transmitted voice volume parameter 132p (common parameter) (step S136).If the sum C3 is smaller than the sum C4 (sum C3<sum C4) (NO in step S135), the new parameter calculation unit 185c determines that the parameters cannot be consolidated into one, and obtains this determination result as the result (step S137).

[0195] For example, if the maximum value V1max is +3.5 dB, the minimum value V1min is +0 dB, the subtraction result C1 is 0 dB, and the maximum value C2 is +4.0 dB, the sum C3 (minimum value V1min + maximum value C2) is +4.0 dB, and the sum C4 (maximum value V1max + subtraction result C1) is +3.5 dB. In other words, since the sum C3 (+4.0 dB) is greater than the sum C4 (+3.5 dB), the new parameter calculation unit 185c determines that the parameters can be consolidated into one, and sets the new transmitted speech volume parameter 132p to +3.5 dB (sum C4).

[0196] 40 is a flowchart showing the flow of processing by the common parameter calculation unit 185 to predict the commonality of the filter coefficients 133p of the receiving-side equalizer 72. First, the variation range calculation unit 185a acquires multiple parameter sets, calculates the average value of the filter coefficients 133p included in the parameter sets, and sets this as a new parameter candidate (step S141). Also in step S141, the variation range calculation unit 185a calculates, for each parameter set, a difference D6 between the filter coefficients 133p of the acquired parameter set and the new parameter candidate. The variation range calculation unit 185a outputs the new parameter candidate and the difference D6 to the new parameter calculation unit 185c.

[0197] In step S141, the change range calculation unit 185a may acquire from the evaluation result examination unit 185b the conditions of the evaluation result set that have the smallest margin for the specifications of the evaluation results of the F characteristic evaluation (for example, car 1), and may set the filter coefficients 133p included in the parameter set acquired under those conditions (for example, the filter coefficients 133p of car 1) as candidates for new parameters. Note that the F characteristic evaluation is an evaluation result of the frequency characteristics of a signal that is input from the receiving end 75 and passed through the receiving-side equalizer 72 during evaluation by the evaluation unit 170.

[0198] Next, the evaluation result review unit 185b acquires the evaluation result set and reflects the difference D6 calculated in step S141 in the FC evaluation included in the evaluation result set (step S142). Specifically, the evaluation result review unit 185b calculates a reflection value C5 by subtracting the difference D6 from the FC evaluation included in the evaluation result set. Because the FC evaluation is calculated for each frequency band, the reflection value C5 is also calculated for each frequency band. The evaluation result review unit 185b outputs the reflection value C5 to the new parameter calculation unit 185c.

[0199] Next, the new parameter calculation unit 185c determines whether all of the reflected values ​​C5 of the F characteristic evaluation fall within a desired range (for example, the shaded area in FIG. 28) (step S143). If all of the evaluation results fall within the desired range (for example, the shaded area in FIG. 28) (YES in step S143), the new parameter calculation unit 185c sets the parameter candidates found in step S141 as new parameters (common parameters) (step S144).

[0200] If none of the evaluation results fall within the desired range (e.g., the shaded area in FIG. 28 ) (NO in step S143), the new parameter calculation unit 185c modifies the difference D6 for the frequency bands that did not fall within the range by a desired amount toward the range (step S145). For example, if the difference D6 at 1000 Hz is +4 dB, and the reflected values ​​C5 for cars 2 and 3 are within the range, but the reflected value C5 for car 1 is -3 dB outside the range, the new parameter calculation unit 185c moves the difference D6 at 1000 Hz by a predetermined value (e.g., 4 dB) toward the range (here, the positive direction). In other words, the new difference D6′ at 1000 Hz is set to the difference D6+4 dB.

[0201] Next, the new parameter calculation unit 185c determines whether the process of step S145 has been performed a predetermined number of times (e.g., three times) (step S146). If the process of step S145 has not been performed the predetermined number of times (NO in step S146), the new parameter calculation unit 185c returns the process to step S142. As a result, a reflection value C5' is calculated using the new difference D6' (step S142), and it is determined whether the reflection value C5' falls within a desired range (step S143).

[0202] For example, for car 1, whose reflection value C5 was outside the -3 dB range when difference D6 was used, the reflection value C5' for car 1 when difference D6' (difference D6 + 4 dB) becomes +0.1 dB, falling within the desired range. In this case, the reflection values ​​C5' for cars 2 and 3, whose reflection values ​​C5 were within the range when difference D6 was used, are checked to see if they are within the desired range. For example, if the reflection value C5 for car 2 is 1 dB, the reflection value C5' becomes +4.1 dB, falling within the desired range. Also, if the reflection value C5 for car 3 is 3 dB, the reflection value C5' becomes +6.1 dB, falling within the desired range. In this way, if the reflection value C5' falls within the desired range (YES in step S143), the new parameter calculation unit 185c adds the difference D6' to the parameter candidate calculated in step S141 and sets the result as a new parameter (common parameter) (step S144).

[0203] If the process of step S145 has been performed a predetermined number of times (YES in step S146), the new parameter calculation unit 185c determines that the parameters cannot be consolidated into one (step S147).

[0204] 41 is a flowchart showing the process flow of the common parameter calculation unit 185 predicting the commonality of the received voice volume parameter 134p. First, the variation range calculation unit 185a acquires a plurality of parameter sets, acquires the received voice volume parameter 134p included in the parameter set, and calculates the maximum value V1max and minimum value V1min of the received voice volume parameter 134p (step S131). The process of step S151 is the same as the process of step S131.

[0205] The evaluation result review unit 185b also acquires multiple evaluation result sets and parameter evaluation criteria, and calculates the minimum margin of the evaluation results (step S152). The process of step S152 is the same as the process of step S132. Note that the order of the processes of steps S151 and S152 does not matter.

[0206] Next, the new parameter calculation unit 185c calculates the result C1 by subtracting the necessary margin from the minimum margin calculated in step S132 (step S153). The process of step S153 is the same as the process of step S133.

[0207] Next, the new parameter calculation unit 185c calculates the maximum value C2 by which the parameter can be changed based on the information acquired from the database 200 (step S154). The process of step S154 is the same as the process of step S134.

[0208] Next, the new parameter calculation unit 185c determines whether the parameters can be consolidated into one based on the sum C3 of the minimum value V1min and the maximum value C2 and the sum C4 of the maximum value V1max and the subtraction result C1 (step S155). The process of step S155 is the same as the process of step S135.

[0209] If the sum C3 is greater than or equal to the sum C4 (YES in step S155), the new parameter calculation unit 185c sets the sum C4 (maximum value V1max+subtraction result C1) as the new transmitted voice volume parameter 132p (step S156). If the sum C3 is less than the sum C4 (NO in step S155), the new parameter calculation unit 185c determines as a result that the parameters cannot be consolidated into one (step S157).

[0210] The new parameter calculation unit 185c outputs the results thus obtained (the new parameters or the fact that the parameters cannot be consolidated into one) via the user instruction unit 160 to a display device (not shown) or the like.

[0211] According to this embodiment, it is possible to standardize parameters under a plurality of conditions using past results. By standardizing parameters under a plurality of conditions, it is possible to use the same parameters for different vehicle models, for example, which has a cost advantage.

[0212] In this embodiment, the common parameter calculation unit 185 considers whether parameters can be standardized and calculates standardized parameters based on the parameter sets and evaluation result sets stored in the past log storage unit 180, but the common parameter calculation unit 185 can also perform similar processing on parameter sets and evaluation result sets other than those stored in the past log storage unit 180. In this case, the common parameter calculation unit 185 can acquire the parameter sets and evaluation result sets from an external storage device (not shown) or the like.

[0213] <Fourth embodiment> A tuning device 4 according to a fourth embodiment will be described below. Note that the same components as those in the tuning devices 1 to 3 according to the first to third embodiments are given the same reference numerals, and descriptions thereof will be omitted.

[0214] 42 is a diagram showing an outline of the functional blocks of the tuning device 4. Functionally, the tuning device 4 mainly includes a signal output unit 110, a sound acquisition unit 120, an adjustment unit 130, a change unit 140, a storage unit 150, a user instruction unit 160, an evaluation unit 170, a past log storage unit 180, a compatibility prediction unit 186, an information acquisition unit 187, and an error processing unit 190. Note that the functional components of the tuning device 4 may be further classified into more components depending on the processing content, or one component may execute the processing of multiple components.

[0215] The compatibility prediction unit 186 is a functional unit that predicts whether a parameter set (base parameter set) under certain conditions is compatible with other conditions. The information acquisition unit 187 is a functional unit that acquires specifications that affect the parameters. For example, the information acquisition unit 187 acquires specifications of vehicle models A and B, and the compatibility prediction unit 186 predicts whether the parameters acquired for vehicle model A are applicable to vehicle model B. The specifications that affect the parameters are specifications of the configuration of the voice communication system, and are specifications that may change signals transmitted through the receiving-side signal path or the transmitting-side signal path, such as the distance between the terminals 50 and 53, the performance of the sound detection sensor 51 and the sound generation unit 52, the size of the closed space 90, and interior components.

[0216] FIG. 43 is a diagram showing a schematic diagram of the data flow in the compatibility possibility prediction unit 186 and the information acquisition unit 187.

[0217] The information acquisition unit 187 acquires specifications via an input device, a communication network (not shown), etc. The specifications include, for example, the distance between the sound detection sensor 51 and the terminal 53, the distance between the sound generation unit 52 and the terminal 53, the performance of the sound generation unit 52, and the reflection level of the closed space 90. The information acquisition unit 187 outputs the acquired specifications to the compatibility possibility prediction unit 186.

[0218] The support possibility prediction unit 186 obtains the difference between the specifications (base specifications) when the base parameter set was obtained and the specifications (study specifications) to be predicted as to whether parameters can be standardized. The support possibility prediction unit 186 also obtains the base parameter set and the corresponding evaluation result set from the past log storage unit 180. Furthermore, a database 200 in which past performance data is stored is connected to the support possibility prediction unit 186, and the support possibility prediction unit 186 obtains information on past performance data from the database 200.

[0219] Next, a description will be given of the processing of the compatibility prediction unit 186. Hereinafter, a case will be described in which the compatibility prediction unit 186 acquires the parameter set and specifications of car 2 and the specifications of a new car 4, and examines whether the parameter set of car 2 can be applied to car 4.

[0220] 44 is a flowchart showing the process flow of the compatibility prediction unit 186 predicting the commonality of the transmission volume parameter 132p, which is a parameter affected by the sound detection sensor 51. First, the compatibility prediction unit 186 obtains the distances between the sound detection sensors 51 of cars 2 and 4 and the terminal 53, and estimates the evaluation result for the transmission volume of car 4 based on information related to past performance stored in the database 200 (step S161).

[0221] First, information regarding past performance data related to the transmission volume will be described. FIG. 45 shows an example of information regarding past performance data stored in the database 200. The horizontal axis of FIG. 45 represents the distance between the terminal 50 and the terminal 53, and the vertical axis represents the SLR. FIG. 45 shows the SLR for each seat in a vehicle with a row of seats. The distance between the terminal 50 and the terminal 53 varies depending on the vehicle, and the SLR is affected by the principle of sound attenuation with distance. For vehicles with similar conditions within the closed space 90 (vehicle interior) (e.g., the relationship between the distance between the terminal 50 and the seat and the vehicle width), the SLR for the driver's seat can be estimated even if the vehicle size changes due to factors such as vehicle class. On the other hand, changing the transmission volume also affects other test items. For example, in noise measurement, the lower the transmission volume, the more desirable results are obtained. Note that the smaller the SLR, the higher the output volume.

[0222] Step S161 will be described using a specific example. Because the distance between the sound detection sensor 51 and the terminal 53 affects the volume of the transmitted voice, the response possibility prediction unit 186 acquires the distance between the sound detection sensor 51 of vehicle 2 and the terminal 53, which is 50 cm in this case. Furthermore, based on past results stored in the database 200, the response possibility prediction unit 186 calculates a regression line (regression line L in this case) that passes through the conditions of vehicle 2 (a volume evaluation of 15 dB when the distance between the sound detection sensor 51 and the terminal 53 is 50 cm). Furthermore, based on the regression line L, the response possibility prediction unit 186 estimates the volume evaluation result e when the distance between the sound detection sensor 51 of vehicle 4 and the terminal 53 is 75 cm in this case. From the results shown in FIG. 45 , it is estimated that the volume evaluation result e is 18 dB.

[0223] Returning to the explanation of Fig. 44, the compatibility prediction unit 186 determines whether the result obtained in step S161 satisfies the specifications (step S162). If the specifications are satisfied (YES in step S162), the compatibility prediction unit 186 predicts that the same transmission volume parameter 132p can be used for the car 4 (step S163).

[0224] If the specifications are not met (NO in step S162), the compatibility prediction unit 186 determines whether the result (transmitted voice volume) obtained in step S161 is equal to or less than the specifications (step S164).

[0225] If the transmission volume is below the specification (YES in step S164), that is, if the transmission volume is low, the compatibility prediction unit 186 compares the sum A (see formula (11)) of the transmission volume and X with the sum B (see formula (12)) of the transmission volume, the shortage of the transmission volume relative to the specification, and a margin, and determines whether the sum A is equal to or greater than the sum B (step S165). Note that X is a value found from past results that show that increasing the transmission volume worsens the noise level, and is, for example, +4 dB. The margin is a value found from past results that show a fluctuation of ±0.5 dB due to the influence of measurement errors, etc., and is, for example, 1 dB. Sum A = Transmission Volume + X (11) Sum B = Transmission Volume + Shortfall of Transmission Volume from the Specification + Margin (12)

[0226] If sum A is greater than or equal to sum B (sum A ≥ sum B) (YES in step S165), the compatibility prediction unit 186 estimates that the transmission volume parameters 132p of cars 2 and 4 can be standardized, and sets sum B as the standardized transmission volume parameter 132p (step S166).

[0227] If the transmission volume is not below the specification (NO in step S164), that is, if the transmission volume is high, the compatibility prediction unit 186 compares a value A' (see formula (13)) obtained by subtracting X from the transmission volume with the sum B (see formula (12)) of the transmission volume, the shortage of the transmission volume relative to the specification, and the margin, and determines whether the value A' is below the sum B (step S167). Value A' = Transmission volume - X ... (13)

[0228] If value A' is less than or equal to sum B (value A'≦sum B) (YES in step S165), the compatibility prediction unit 186 estimates that the transmission volume parameters 132p of cars 2 and 4 can be standardized, and sets sum B as the standardized transmission volume parameter 132p (step S168).

[0229] If sum A is not equal to or greater than sum B (sum A<sum B) (NO in step S165) and value A' is not equal to or less than sum B (value A'>B) (NO in step S167), the compatibility possibility prediction unit 186 determines as a result that the parameters cannot be consolidated into one (step S169). This causes the compatibility possibility prediction unit 186 to end the series of processes.

[0230] 46 is a flowchart showing the process flow of the compatibility prediction unit 186 predicting the commonality of the received voice volume parameter 134p, which is a parameter affected by the sound generation unit 52. First, the compatibility prediction unit 186 acquires information about the performance of the sound generation units 52 of vehicles 2 and 4, and estimates the evaluation result of the transmitted voice volume of vehicle 4 based on the results stored in the database 200 (step S171). For example, when information is acquired from the information acquisition unit 187 that the volume of the sound generation unit 52 of vehicle 2 is the same as the reference value, the volume of the sound generation unit 52 of vehicle 4 is 2 dB lower than the reference value, and the evaluation result of the received voice volume of vehicle 2 is 1 dB, and when information is acquired from the database 200 that a 1 dB decrease in volume increases the volume evaluation by 1 dB, the compatibility prediction unit 186 estimates that the evaluation result of vehicle 4 is 3 dB. Note that the information about past performance regarding the received voice volume is the same as the information about past performance regarding the transmitted voice volume.

[0231] Next, the compatibility prediction unit 186 determines whether the evaluation result estimated in step S171 satisfies the specifications (step S172). If the estimated evaluation result satisfies the specifications (YES in step S172), the compatibility prediction unit 186 estimates that the parameters of car 2 can be applied to car 4 (step S173).

[0232] If the estimated evaluation result does not satisfy the specifications (NO in step S172), that is, if the received voice volume is low, the compatibility prediction unit 186 compares the sum E (see formula (13)) of the received voice volume and Y with the sum F (see formula (14)) of the received voice volume, the shortfall of the received voice volume from the specifications, and the margin, and determines whether the sum E is equal to or greater than the sum F (step S174). Note that Y is a value found from past results that show that double-talk detection performance deteriorates when the received voice volume is increased, and is, for example, +4 dB. The margin is a value found from past results that show a fluctuation of ±0.5 dB due to the influence of measurement errors, etc., and is, for example, 1 dB. Sum E = Received Voice Volume + Y (13) Sum F = Received Voice Volume + Shortfall of the received voice volume from the specifications + Margin (14)

[0233] If the sum E is greater than or equal to the sum F (Sum E ≧ Sum F) (YES in step S174), the compatibility prediction unit 186 estimates that the received voice volume parameters 134p of cars 2 and 4 can be standardized, and sets the sum F to the standardized received voice volume parameter 134p (step S175).

[0234] If the sum E is not equal to or greater than the sum F (sum E<sum F) (NO in step S174), the compatibility prediction unit 186 determines that the parameters cannot be consolidated into one (step S176). This causes the compatibility prediction unit 186 to end the series of processes.

[0235] 47 is a flowchart showing the process flow of the compatibility prediction unit 186 for predicting the commonality of the adaptive filter initial value 135p, which is a parameter affected by the reflection level of the closed space 90. First, the compatibility prediction unit 186 acquires information about the reflection levels of cars 2 and 4 and estimates the evaluation result of the transmission volume of car 4 based on the results stored in the database 200 (step S181). The reflection level is affected, for example, by the seat material, and the information about the reflection level is, for example, reverberation time. For example, when the information acquisition unit 187 acquires information that the reverberation time of car 2 is a reference value, the reverberation time of car 4 is the reference value plus 1 ms, and the evaluation result of the reception volume of car 2 is 1 dB, the evaluation result of the echo suppression is 50 dB, and the double-talk detection performance is 3 dB, and when the information acquired from the database 200 indicates that the evaluation result of the echo suppression decreases by 3 dB when the reverberation time increases by 1 ms, the compatibility prediction unit 186 estimates that the evaluation result of the echo suppression of car 4 is 47 dB.

[0236] Next, the compatibility prediction unit 186 determines whether the evaluation result estimated in step S181 satisfies the specifications (step S182). If the estimated evaluation result satisfies the specifications (YES in step S182), the compatibility prediction unit 186 estimates that the parameters of car 2 can be applied to car 4 (step S183).

[0237] If the estimated evaluation result does not satisfy the specifications (NO in step S132), the compatibility prediction unit 186 sets K so that the echo suppression evaluation result -X + Y * K satisfies the specifications (step S184). Here, X is the amount of decrease in the echo suppression evaluation result due to reverberation time, and is calculated by multiplying the reverberation time by 3 dB (reverberation time * 3 dB). Also, Y is the amount of increase in the echo suppression evaluation result due to an increase in the received voice volume, and is calculated by multiplying the decrease in the received voice volume by 3 dB (increase in the received voice volume * 3 dB). Note that it is possible to obtain from the database 200 that a 1 dB decrease in the received voice volume results in a 3 dB decrease in the echo suppression evaluation result.

[0238] Next, the compatibility prediction unit 186 compares the difference G (see formula (13)) between the received voice volume and K with the sum F of the received voice volume, the shortage of the received voice volume relative to the specifications, and the margin, and determines whether the sum F is equal to or greater than the sum F (step S185). Difference G = Received voice volume - K (13)

[0239] If the difference G is greater than or equal to the sum F (difference G ≥ sum F) (YES in step S185), the compatibility prediction unit 186 estimates that the adaptive filter initial values ​​135p of cars 2 and 4 can be standardized, and sets the sum F as the standardized adaptive filter initial value 135p (step S186).

[0240] If the difference G is not equal to or greater than the sum F (difference G<sum F) (NO in step S184), the compatibility prediction unit 186 determines that the parameters cannot be consolidated into one (step S187). This causes the compatibility prediction unit 186 to end the series of processes.

[0241] According to this embodiment, it is possible to estimate whether the same parameters can be applied under new conditions, i.e., whether the acoustic test (evaluation performed by the evaluation unit 170) satisfies the standard even when the same parameters are used. For example, when selling a new car model, it is possible to predict whether the parameters of an old car model can be applied to the new car model based on the change points between the new car model and the old car model. Furthermore, when determining the specifications of a new car, it is possible to predict the range of change relative to the specifications of the old car so that the parameters of the old car model can be applied.

[0242] In the third and fourth embodiments, the common parameter calculation unit 185 considers whether parameters can be standardized and calculates the standardized parameters, and the support possibility prediction unit 186 estimates whether the same parameter set can be used, based on the parameter sets and evaluation result sets stored in the past log storage unit 180, but the common parameter calculation unit 185 and the support possibility prediction unit 186 can perform similar processing on parameter sets and evaluation result sets other than the parameter sets and evaluation result sets stored in the past log storage unit 180. In this case, the common parameter calculation unit 185 and the support possibility prediction unit 186 can simply acquire the parameter sets and evaluation result sets from an external storage device (not shown) or the like. Furthermore, the device may be provided as a common parameter calculation device having a common parameter calculation unit that considers whether parameter sets measured under multiple conditions can be made compatible based on the parameter sets and evaluation result sets acquired from an external storage device or the like and information on past performance, and calculates common parameters if they can be made compatible; or as a compatibility prediction device having a compatibility prediction unit that predicts whether parameters under certain conditions can be made compatible with other conditions based on the parameter sets and evaluation result sets acquired from an external storage device or the like and information on past performance, as well as the information stored in the past log storage unit and information on past performance.

[0243] Although an embodiment of the present invention has been described above in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and design changes and the like are also included within the scope that does not deviate from the gist of the present invention.

[0244] 1, 2, 3, 4: Tuning device 50, 53: Terminal 51, 54: Sound detection sensor 52, 55: Sound generation unit 56: Acoustic device 61: Echo removal unit 61a: Filter generation unit 61b: Echo removal unit 62, 73: FFT unit 63: Noise estimation unit 64: Noise suppression unit 65: Nonlinear echo suppression unit 66: Noise superposition unit 67: Transmitting side equalizer 68: IFFT unit 69, 70: Gain adjustment unit 71: Dynamic range control (DRC) 72, 72A, 72B, 72C: Receiving side equalizer 74: Double talk detection unit 72a, 74a, 74b: FFT unit 72b: IFFT unit 74c: Calculation unit 75: Receiving side terminal 76: Transmitting end 90: Closed space 91: Communication device 100: Tuning system 110: Signal output unit 120: Sound acquisition unit 130: Adjustment unit 131: Delay adjustment unit 131p: Delay parameter 132: Transmitting voice volume adjustment unit 132p: Transmitting voice volume parameter 133, 133A, 133B, 133C: Received voice equalizer adjustment unit 133p: Filter coefficient 134: Received voice volume adjustment unit 134p: Received voice volume parameter 135: Adaptive filter adjustment unit 135p: Adaptive filter initial value 136: Transmitting voice equalizer adjustment unit 136p: Gain 137: Double talk detection adjustment unit 137p: Threshold value 138: Noise level estimation unit 138p: Noise level estimation 139p : Suppression amount 140 : Change unit 150 : Storage unit 160 : User instruction unit 170 : Evaluation unit 180 : Past log storage unit 185 : Common parameter calculation unit 185 a : Change range calculation unit 185 b : Evaluation result examination unit 185 c : New parameter calculation unit 186 : Possibility of response prediction unit 187 : Information acquisition unit 190 : Error processing unit 200 : Database 561 : Acoustic processing unit 562 : Parameter storage unit

Claims

1. A tuning device for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, comprising: a signal output unit that outputs a signal to the first sound generation unit; a sound acquisition unit that acquires the sound picked up by the first sound detection sensor and the sound picked up by a second sound detection sensor arranged inside the closed space; and an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiver signal path that transmits signals to the first sound generation unit or a transmitter signal path that transmits signals recorded by the first sound detection sensor, wherein the adjustment unit comprises: a receiver equalizer adjustment unit that adjusts the filter coefficient or gain for each frequency of an equalizer filter used in a receiver equalizer that adjusts the frequency of signals transmitted through the receiver signal path, based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit; a receiver volume adjustment unit that adjusts a target value of a receiver gain adjustment unit that adjusts the gain of an input signal based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit, the receiver volume adjustment unit performing processing after the receiver equalizer adjustment unit; and an adaptive filter adjustment unit that sets an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generation unit, based on the signal output by the signal output unit, the adaptive filter adjustment unit performing processing after the receiver volume adjustment unit.

2. The tuning device described in claim 1, characterized in that the receiver equalizer adjustment unit performs coefficient estimation using an autoregressive model on a first signal that is the result of performing frequency analysis on the result of processing an arbitrary signal output to the closed space with the filter coefficients of the receiver equalizer initialized, or a second signal that is generated based on the first signal, and uses the estimated coefficients as the filter coefficients of the equalizer filter, which are the parameters.

3. The tuning device according to claim 2, characterized in that the receiver equalizer adjustment unit generates the second signal based on the result of multiplying the first signal by a signal obtained by converting white noise into a function of frequency, and performs coefficient estimation for the second signal using an autoregressive model.

4. The tuning device according to claim 3, characterized in that the receiver equalizer adjustment unit multiplies the first signal by a signal obtained by converting white noise into a frequency function, and multiplies the result by a signal obtained by converting a filter having the inverse characteristics of a desired characteristic into a frequency function, to obtain the second signal.

5. The tuning device according to any one of claims 1 to 4, characterized in that the adjustment section has a double-talk detection adjustment section that sets a threshold value used in double-talk detection to detect whether or not a double-talk state is occurring, the double-talk detection adjustment section acquires information regarding the magnitude of a reference signal transmitted through the receiving-side signal path and the magnitude of a residual echo signal obtained by removing a linear echo component from the transmitting signal acquired by the first sound detection sensor, uses the data with the largest magnitude of the residual echo signal for each reference signal magnitude as learning data, and uses the coefficient of a regression curve obtained by regressing the learning data with a cubic function as the threshold value.

6. The tuning device according to claim 5, characterized in that the double-talk detection adjustment unit calculates the coefficient by shifting the regression curve in the direction in which the magnitude of the residual echo signal increases based on a value indicating the variance of the learning data.

7. A tuning device as described in claim 5 or 6, characterized in that the double-talk detection adjustment unit increases the value input to the receiving side gain adjustment unit from the initial value to obtain information regarding the magnitude of the reference signal and the residual echo signal.

8. A tuning device as described in any one of claims 1 to 7, characterized in that the adjustment unit has a delay adjustment unit that adjusts a setting regarding the delay time of the reference signal relative to the transmission signal based on the reference signal input to the first sound generation unit and the transmission signal input from the first sound detection sensor.

9. A tuning device according to any one of claims 1 to 8, characterized in that the adjustment unit has a transmitting-side equalizer adjustment unit that adjusts the gain used in a transmitting-side equalizer that adjusts the frequency of the signal output from the acoustic device.

10. A tuning device as described in any one of claims 1 to 9, characterized in that the adjustment unit has a noise level estimation unit that estimates the loudness of noise relative to the loudness of sound acquired by the first sound detection sensor.

11. The tuning device according to any one of claims 1 to 10, further comprising an error processing section that performs a stop process or a return process when an error occurs during processing by the adjustment section.

12. A tuning system comprising: a terminal having a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space; an acoustic device connected to the terminal; and a tuning device that tunes the acoustic device, wherein the tuning device comprises: a signal output unit that outputs a signal to the first sound generation unit; a sound acquisition unit that acquires sound picked up by the first sound detection sensor and sound picked up by a second sound detection sensor arranged inside the closed space; and an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiver signal path that transmits signals to the first sound generation unit or a transmitter signal path that transmits signals recorded by the first sound detection sensor, based on the data acquired by the sound acquisition unit, wherein the adjustment unit comprises: a receiver equalizer adjustment unit that adjusts the filter coefficient or gain for each frequency of an equalizer filter used in a receiver equalizer that adjusts the frequency of signals transmitted through the receiver signal path, based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit; a receiver volume adjustment unit that adjusts a target value of a receiver gain adjustment unit that adjusts the gain of an input signal based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit, the receiver volume adjustment unit performing processing after the receiver equalizer adjustment unit; and an adaptive filter adjustment unit that sets an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generation unit, based on the signal output by the signal output unit, the adaptive filter adjustment unit performing processing after the receiver volume adjustment unit.

13. A tuning method for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, comprising an adjustment step of adjusting parameters used in acoustic processing for signals transmitted through a receiver signal path that transmits signals to the first sound generation unit or a transmitter signal path that transmits signals collected by the first sound detection sensor, the adjustment step comprising: a receiver equalizer adjustment step of adjusting a filter coefficient or a gain for each frequency of an equalizer filter used in a receiver equalizer that adjusts the frequency of a received signal, based on a signal output by a signal output unit to the first sound generation unit and sound collected by a second sound detection sensor arranged inside the closed space; and a receiver volume adjustment step of adjusting a target value of a receiver gain adjustment unit that adjusts the gain of an input signal, based on the signal output by the signal output unit and the sound collected by the first sound detection sensor, the receiver volume adjustment step being performed after the receiver equalizer adjustment step. an adaptive filter adjustment step of setting an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generation unit, based on the signal output from the signal output unit, the adaptive filter adjustment step being performed after the received volume adjustment step.

14. A tuning program for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, comprising: a computer; a signal output unit that outputs a signal to the first sound generation unit; a sound acquisition unit that acquires the sound collected by the first sound detection sensor and the sound collected by a second sound detection sensor disposed inside the closed space; an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiver signal path that transmits signals to the first sound generation unit or a transmitter signal path that transmits signals recorded by the first sound detection sensor, the receiver equalizer adjustment unit adjusting the filter coefficient or gain for each frequency of an equalizer filter used in a receiver equalizer that adjusts the frequency of signals transmitted through the receiver signal path, based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit; a receiver volume adjustment unit that adjusts a target value of a receiver gain adjustment unit that adjusts the gain of an input signal based on the signal output by the signal output unit and the sound acquired by the sound acquisition unit, the receiver volume adjustment unit performing processing after the receiver equalizer adjustment unit; and an adaptive filter adjustment unit that sets an initial value of an adaptive filter used to remove echoes generated by the sound output from the first sound generation unit, based on the signal output by the signal output unit, the adaptive filter adjustment unit performing processing after the receiver volume adjustment unit.

15. A tuning device for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, comprising: a signal output unit that outputs a signal to the first sound generation unit; a sound acquisition unit that acquires sound picked up by the first sound detection sensor and sound picked up by a second sound detection sensor arranged inside the closed space; an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiving side signal path that transmits signals to the first sound generation unit or a transmitting side signal path that transmits signals recorded by the first sound detection sensor; an evaluation unit that measures the performance of the acoustic device using the results of the tuning process performed by the adjustment unit and evaluates the measurement results; a past log storage unit that stores the results of the tuning process performed by the adjustment unit and the evaluation results by the evaluation unit; and a common parameter calculation unit that examines whether parameter sets measured under multiple conditions can be standardized based on the information stored in the past log storage unit and information related to past performance, and calculates common parameters if standardization is possible.

16. A tuning device for tuning an acoustic device connected to a first sound detection sensor that detects sound generated in a closed space and a first sound generation unit that emits sound into the closed space, comprising: a signal output unit that outputs a signal to the first sound generation unit; a sound acquisition unit that acquires sound picked up by the first sound detection sensor and sound picked up by a second sound detection sensor arranged inside the closed space; an adjustment unit that adjusts parameters used in acoustic processing for signals transmitted through a receiving side signal path that transmits signals to the first sound generation unit or a transmitting side signal path that transmits signals recorded by the first sound detection sensor; an evaluation unit that measures the performance of the acoustic device using the results of the tuning process performed by the adjustment unit and evaluates the measurement results; a past log storage unit that stores the results of the tuning process performed by the adjustment unit and the evaluation results by the evaluation unit; and a compatibility prediction unit that predicts whether parameters under certain conditions can be adapted to other conditions based on the information stored in the past log storage unit and information related to past performance.

Images