JPH05276599A - In-cabin sound field correction device - Google Patents

Abstract

PURPOSE:To provide the in-cabin sound field correction device in which the scale of the hardware is remarkably reduced and the correction accuracy is enhanced. CONSTITUTION:In the in-cabin sound field correction device provided with a seat position detection means 5 detecting a seat position of a listener in a cabin and processing the position as a numeral, a parameter decision means 2 receiving the seat position processed as a numeral and deciding a correction parameter, and a means 4 receiving the correction parameter and correcting an output of a stereo reproduction means 1, the parameter decision means 5 has a hierarchical neural network entering an output of the digitized means to the input stage.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle interior sound field correction device, which can significantly reduce the scale of hardware and can prevent an error in parameters and a decrease in asymmetry correction accuracy due to the error. The present invention relates to a sound field correction device.

[0002]

2. Description of the Related Art Generally, in a vehicle-mounted stereo reproducing apparatus, the positions of left and right speakers with respect to a listener are asymmetrical.
There is a problem in that the sound distribution between the left and right speakers is biased toward the speakers closer to the listener, and the original stereo reproduction effect cannot be obtained.

As a method for solving this problem, for example, a collection of lecture papers of the Japan Acoustic Banquet (March 1988) 1-5-1
4 (pages 357 to 358) etc., an apparent sound source is created at a position symmetrical to the speaker farther from the listener, and the sound image is evenly spread left and right while maintaining the correct sound image localization direction. The sound field is generated.

However, since this method is premised on that the listening position is constant, the listening position can be adjusted by, for example, shifting the seat back and forth, changing the height of the seat, or changing the angle of the backrest. The effect diminishes when the deviation occurs.

As a means for performing asymmetric correction for such a change in the listening position, for example, Japanese Patent Laid-Open No. 2-1580.
As disclosed in Japanese Patent Laid-Open No. 0-202, a digitizing means for digitizing the state of an arbitrary position to be corrected in the vehicle interior and a parameter determination for determining a correction parameter by inputting the digitized state of the position to be corrected It has been proposed to include means and means for inputting a correction parameter and correcting the output of the stereo reproduction means.

[0006]

In this prior art example, a Neumann type computer is used as the parameter determining means, the output of the digitizing means is quantized, and the front and rear position of the seat, the height, and the inclination of the backrest are quantized with respect to the quantized value. Since a combination pattern such as is determined and a correction parameter determined from the determined pattern is output, it is necessary to set parameters for a huge number of patterns, and a huge scale of hardware is required.

Further, there is a problem that a calculation error at the time of quantization causes a pattern determination error and a parameter error, and the accuracy of the asymmetry correction is lowered.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a vehicle interior sound field correction apparatus capable of significantly reducing the scale of hardware and enhancing the correction accuracy. To do.

[0009]

SUMMARY OF THE INVENTION In order to solve the above problems, the present invention provides a digitizing means for digitizing the state of an arbitrary position to be corrected in a vehicle compartment, and a state of the digitized position to be corrected. In a vehicle interior sound field correction device including a parameter determination unit that inputs and determines a correction parameter and a unit that inputs a correction parameter and corrects the output of the stereo reproduction unit, the parameter determination unit outputs the output of the digitization unit. It is characterized by having a hierarchical neural network for inputting to the input layer.

[0010]

In the present invention, the digitized data input to the computer having the hierarchical neural network include, for example, the seat on which the listener is seated, the front and rear position of the driver's seat, the height, the angle of the backrest, The front and rear position of the passenger seat, height, angle of the backrest, etc. are input.

Since the hierarchical neural network has a learning function, it is possible to perform learning so that the optimum correction parameters are output from the output layer with respect to these inputs, determine the combination pattern of these input data, and It is not necessary to store a pattern or a correction parameter corresponding to the pattern.

Further, since it is not necessary to set a combination pattern of a large number of input data, it is possible to learn a larger number of patterns and realize a more accurate asymmetry correction.

Further, since the hierarchical neural network has a function of estimating the correction parameter with considerably high accuracy from the learned data close to it even when the untrained data is input, the parameter associated with the quantization is The error can be reduced or eliminated.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS As shown in the block diagram of FIG. 1, a vehicle interior sound field correction apparatus according to an embodiment of the present invention inputs a right channel output R and a left channel output L of a stereo reproduction apparatus 1, and A correction unit 4 that corrects the outputs of the left and right speakers 3L and 3R based on the correction parameter output from the parameter determination unit 2 and a predetermined seat in the passenger compartment, that is, the seat positions of the driver's seat D and the passenger seat N are detected. And a seat position detecting means 5 for digitizing the position and outputting it to the parameter determining means 2.

The seat position detecting means 5 determines whether or not each seat D, N is seated, the longitudinal position x of each seat D, N, the reclining angle θ of the backrest, the height y of each seat D, N, etc. It is digitized and output to the parameter determining means 2.

As shown in FIG. 2, the parameter determining means 2
Includes an input layer 21, a middle layer 22 and an output layer 23, and each neuron mimicking element 211, 212 of the input layer 21,
The numerical data x, θ, y, etc. are individually input to 213, ..., And each neuron mimicking element 23 of the output layer 23 is input.
The correction parameters expected from 1, 232, ..., 23i, ..., For example, θi, xi, θ ′, x, etc. are output.

As shown in FIG. 3, the correction means 4 inputs the correction parameter x to the right channel output R to adjust the attenuation factor, and the correction parameter θ'to adjust the delay time. The delay device 7 is connected to correct the difference in distance from the left and right speakers 3L and 3R to the listening position (distance correction effect), and then the right channel output R is input to the band-splitting filter 8 and the correction parameter θi. And a phase shifter 9 for adjusting the amount of phase shift with respect to the right channel output R, and an attenuator 10 for inputting the correction parameter xi to adjust the attenuation rate.
And a level difference between both ears and a phase difference are set to values close to the sound image in the symmetrical position direction of the left speaker 3L (movement in the sound image localization direction). ..

That is, as shown in FIG. 4, the correcting means 4 first sets the right speaker R to the left speaker 3L.
The apparent sound source R ′ located at the same distance as the left speaker 3L from the listening position in the direction of was produced, and then the output of the left speaker 3L was given a phase difference and a level difference with respect to the output of the right speaker 3R. The principle is used that, by adding signals, an apparent sound source R ″ is created at a position symmetrical to the left speaker 3L with respect to the listener.

In this vehicle interior sound field correction device, since the parameter determining means 2 is composed of a neural network, it is expected for a combination pattern of a large number of digitized data output from the seat position detecting means 5. By learning the output value for each pattern, it is not necessary to store a large number of patterns and the output values corresponding to the patterns, and a single neurochip can perform complex and highly accurate sound field correction. ..

Further, since it is not necessary to store the above-mentioned patterns and the correction parameters corresponding thereto, it is possible to improve the accuracy of sound field correction by learning more diverse patterns.

Further, for a pattern that has not been learned in advance, there is a function of estimating the correction parameter from the learned data close to it in a considerable accuracy, so that the error of the parameter associated with the digitization or the quantization is not reduced. It can be resolved.

In the above embodiment, the sound field correction for the driver's seat D (or the passenger seat N) is explained.
The present invention is also applied to the case of performing sound field correction for both the driver's seat D and the passenger seat N, and further for correcting the sound field of the driver seat D and the passenger's seat N as well as the sound fields of the left and right passenger seats behind them. It is possible.

Further, although the sound field correction for 2-channel stereo has been described in the above embodiment, the present invention can be applied to correction of a sound field further multiplexed such as a 3-speaker system or 4-channel stereo. Is.

[0024]

As described above, according to the present invention, even when the listening position deviates due to changes in the front and rear positions of the seat, the height, the angle of the backrest, etc., these changes are detected and the changes are dealt with. By adjusting the correction parameters, the optimum sound field can be corrected.

Moreover, since the parameter determining means for inputting the listening position changing data and outputting the optimum correction parameter is composed of a neuronet, the listening position changing data combination pattern and the correction parameter corresponding to the pattern are combined. It is possible to determine the correction parameters for various patterns with one neurochip without having to store the above, and it is possible to significantly reduce the scale of the hardware and at the same time improve the correction accuracy.

[Brief description of drawings]

FIG. 1 is a block diagram of the present invention.

FIG. 2 is a schematic diagram showing a configuration of a parameter determining means of the present invention.

FIG. 3 is a block circuit diagram of correction means of the present invention.

FIG. 4 is a principle diagram of sound field correction of the present invention.

[Explanation of symbols]

 1 Stereo Playback Means 2 Parameter Determining Means 3L Left Speaker 3R Right Speaker 4 Correcting Means 5 Seat Position Detecting Means

Claims (1)

[Claims] 1. A digitizing means for digitizing a state of an arbitrary position to be corrected in a vehicle compartment, a parameter determining means for inputting the digitized state of the position to be corrected and a correction parameter, and a correction parameter. In the vehicle interior sound field correction apparatus including means for inputting and correcting the output of the stereo reproducing means, the parameter determining means has a hierarchical neural network for inputting the output of the digitizing means into the input layer. Vehicle interior sound field correction device.