JPH07167707A - Apparatus for detecting abnormal sound of electric acoustic converter - Google Patents

Abstract

PURPOSE:To detect an abnormal sound with high accuracy irrespective of the presence of higher harmonics, etc., by sequentially inputting test signals changed in a plurality of stages of power to a to-be-tested converter (speaker or the like). CONSTITUTION:A signal waveform to be used for testing of a speaker 20 is stored in a memory 11 of a signal generator 10. The signal waveform is read out and converted to an analog signal by a D/A converter 12. The analog signal is amplified first with a smaller amplification factor by an amplifier 13, and input to the speaker 20. Then, the same signal waveform is read out from the memory 11, amplified with a larger factor and input to the speaker 20. In this manner, both the signal of a smaller power and the signal of a larger power are input to the speaker 20 sequentially and an acoustic output at each time is detected by a microphone 30 to obtain a corresponding acoustic signal. The presence/absence of an abnormal sound is detected on the basis of the acoustic signals. A trembling sound or the like can be detected highly accurately irrespective of the presence/absence of higher harmonics.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormal sound detection device for an electroacoustic transducer such as a speaker or a headphone, which detects abnormal noise called so-called chattering noise and other abnormal noise. Regarding

[0002]

2. Description of the Related Art For example, in the final step of manufacturing a speaker, a signal generator is connected to the manufactured speaker, and a signal swept from the low frequency range to the high frequency range by the signal generator is input to the speaker. The inspector listened to the emitted sound with his / her ears to determine whether or not there was a clinging sound. This chattering sound has various causes, for example, foreign matter is mixed into the speaker during the manufacturing process of the speaker, a part of the diaphragm of the speaker is peeled off, or other parts of the speaker are defective. Occurs in.

[0003]

There is a demand for automating the above-described inspection process for the presence or absence of chattering noise, and proposals have been made accordingly (Japanese Patent Publication No. 5-16).
No. 738). In the method proposed in this publication, a signal whose frequency is swept is input to a speaker, a sound emitted from the speaker is received by a microphone to obtain an acoustic signal, and the acoustic signal is passed through a high-pass filter. Is to determine the power of the output signal and determine whether or not a clinging sound is present according to the power. This is based on the empirical rule that chattering noise often contains harmonic components.

By the above-mentioned method, it is possible to detect a certain kind of chattering sound, but at least when the chattering sound exists in a high frequency region due to the cause of the abnormal sound such as chattering sound. The sound reproduced by the speaker may not include a harmonic component or may include only a very small harmonic component, and an abnormal noise having a frequency lower than that of the input signal may occur.

In view of the above circumstances, the present invention detects an abnormal sound of an electroacoustic transducer such as a speaker or a headphone regardless of whether or not the abnormal sound such as so-called chatter noise contains a harmonic component. It is an object of the present invention to provide an abnormal sound detection device for an electroacoustic transducer that can be performed.

[0006]

An abnormal sound detecting device for an electroacoustic transducer according to the present invention which achieves the above object, comprises: (1) Each test in which the power is switched to at least two stages to the electroacoustic transducer to be detected. Signal generator for sequentially inputting signals (2) Receiving the acoustic output of the detected electroacoustic transducer and outputting each acoustic signal at each time when each test signal is input to the detected electroacoustic transducer Microphone (3) The above-mentioned acoustic signals output from the microphone are input, and based on the acoustic signals, a signal processor for detecting abnormal noise of the detected electroacoustic transducer is provided. It is preferable that the signal generator, as each of the test signals, sequentially outputs each of the sweep signals whose frequencies are continuously changed and whose powers are different from each other.
Further, the signal processor, after correcting the difference in the power of each acoustic signal due to the difference in the power of each test signal,
It is preferable to perform a calculation for obtaining an amount representing the difference between the acoustic signals and detect the abnormal sound of the electroacoustic transducer to be detected based on the calculation result.

[0007]

[Function] An electroacoustic transducer such as a speaker does not generate a sticking sound while inputting a low power signal to the electroacoustic transducer and obtaining a low acoustic output, or comparing with the power of a normal sound. Although only a very low power chattering sound is generated, if a high power signal is input to the electroacoustic transducer and a large sound output is generated, the power of the normal sound is greatly exceeded and the chattering sound is greatly exceeded. Has a non-linearity that the power of is increased.

The present invention has been completed paying attention to this point. That is, the present invention sequentially inputs both a low-power test signal and a high-power test signal to the electroacoustic transducer, receives the acoustic output at each time point with a microphone to obtain each acoustic signal, and outputs the acoustic signals. The present invention is provided with a configuration for detecting the presence or absence of abnormal noise based on an acoustic signal, which enables highly accurate detection of abnormal noise such as rattling noise regardless of the presence or absence of harmonics.

[0009]

EXAMPLES Examples of the present invention will be described below. FIG. 1 is a block diagram of an embodiment of the abnormal sound detecting device of the present invention. In the waveform memory 11 which constitutes the signal generator 10,
The signal waveform used for the test of the chattering sound of the speaker 20 is
It is stored for each type of speaker 20. For those signal waveforms, a sine waveform that sweeps a frequency band according to the speaker to be tested is typically used.
The waveform is not limited to this, and may be a rectangular wave and a waveform in which the repeating frequency sweeps, a pseudo random waveform, or the like. Of those signal waveforms, a signal waveform corresponding to the speaker 20 to be tested is read out, converted into an analog signal by the D / A converter 12, and first amplified by the amplifier 13 with a small amplification factor,
The output of the amplifier 13 is input to the speaker 20 as a first test signal.

A sound corresponding to the first test signal is generated from the speaker 20, and the sound is received by the condenser microphone 30. The first acoustic signal obtained by receiving the sound by the condenser microphone 30 is amplified by the microphone amplifier 31 and then input to the signal processor 40. The first acoustic signal input to the signal processor 40 is converted into a digital acoustic signal by the A / D converter 41 and temporarily stored in the internal memory of the arithmetic unit 42.

Next, in the signal generator 10, the same signal waveform is read again from the waveform memory 11 and converted into an analog signal, which is greatly amplified by the amplifier 13 this time, and is output to the speaker 20 as the second test signal. Then, the second acoustic signal is input to the calculation unit 42 in the same manner as above. In the calculation unit 42, the first data input as described above is input.
And the following calculation is performed based on the second acoustic signal. That is, when the first acoustic signal is y ₁ , its average power is <y ₁ ² >, the second acoustic signal is y ₂ , and its average power is <y ₂ ² >, A = √ {<y ₂ ² > / <y ₁ ² >}
Is obtained, the first acoustic signal y ₁ is multiplied by A, and the difference ε = Ay ₁ −y ₂ (1) between the first acoustic signal y ₁ multiplied by A and the second acoustic signal is obtained. To be The above A may be a preset constant value. Further, here, the first acoustic signal y multiplied by A is used.
_It suffices that the amount representing the difference between ₁ and the second acoustic signal y ₂ is obtained, and the amount is not limited to equation (1), and for example, ε ² = (Ay ₁ −y ₂ ) ² ε ′ = Ay ₁ It may be / y ₂ or the like. Moreover, you may standardize like ε ″ = (Ay ₁ −y ₂ ) / (Ay ₁ + y ₂ ).

FIG. 2 is a graph showing the power with respect to the frequency of the acoustic signal. The solid line graph in the figure is a graph obtained by multiplying the first acoustic signal obtained when the first test signal of low power is input to the speaker 20 by A times,
The broken line graph in the figure is a graph of the second acoustic signal obtained when the second test signal having high power is input to the speaker 20, and the dashed-dotted line in the figure is the graph when the abnormal noise occurs. It is a graph of 2 acoustic signals.

As described above, when an abnormal noise such as a chattering noise is generated, the difference ε in the equation (1) is partially increased, and therefore it is possible to know whether or not the abnormal noise is generated based on this difference. There is no particular limitation on how the difference ε is linked to the determination of the presence or absence of abnormal noise, and for example, the absolute value of the difference ε for each frequency (corresponding to time in the case of a sweep waveform) | ε | maximum value of | epsilon _p | look, the maximum value | enter into the determination unit 43 (see FIG. 1), the maximum value in the determination unit 43 | | epsilon _p a predetermined threshold value | epsilon _p Whether or not abnormal noise is generated may be determined by determining whether or not it exceeds, or the absolute value | ε | of the difference ε of each frequency is added over each frequency, and the added value Σ | ε | May be input to the determination unit 43, and determination may be made based on whether or not the added value Σ | ε | exceeds a threshold value.

Further, in order to improve the determination accuracy, if the difference ε _r exists even when there is no abnormal sound, the difference ε _r
r is measured in advance, the difference ε _r is corrected for each frequency in the actual test stage, and the maximum value | ε _p | and the added value Σ | ε | of the remaining part are calculated and used for judgment. May be. Alternatively, the determination unit 43 may be provided with a neural network that inputs the first and second acoustic signals and the difference between them and outputs the determination result of the presence or absence of abnormal noise.

As described above, for example, the presence of abnormal noise is determined in the final inspection step of the speaker 20 manufacturing line,
The speaker 20 that emits an abnormal noise is removed from the production line.

[0016]

As described above, the present invention detects abnormal noise based on the non-linearity of abnormal noise depending on the magnitude of the test signal, and accurately detects abnormal noise regardless of the presence of harmonics. Sound is detected.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of an abnormal sound detection device of the present invention.

FIG. 2 is a graph showing power with respect to frequency of an acoustic signal.

[Explanation of symbols]

 10 signal generator 20 speaker 30 condenser microphone 31 microphone amplifier 40 signal processor

Claims (2)

[Claims] 1. A signal generator for sequentially inputting each test signal whose power is switched in at least two stages to a detected electroacoustic transducer, and an acoustic output of the detected electroacoustic transducer, A microphone that outputs each acoustic signal at each time when each test signal is input to the detected electroacoustic transducer, and each acoustic signal output from the microphone is input, and based on those acoustic signals, An abnormal sound detection device for an electroacoustic transducer, comprising: a signal processor that detects abnormal sound of the detected electroacoustic transducer. 2. The signal generator sequentially outputs, as the test signals, respective sweep signals whose frequencies are continuously changed and which have different powers, and the signal processor outputs the test signals. After correcting the difference in power of each of the acoustic signals due to the difference in signal power,
The electroacoustic apparatus according to claim 1, wherein an operation for obtaining an amount representing a difference between the acoustic signals is performed, and abnormal noise of the detected electroacoustic transducer is detected based on the operation result. Abnormal noise detection device of the converter.