JP2506863B2 - Speaker harmonic distortion measurement method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a harmonic distortion measuring method for evaluating reproduction characteristics of a speaker of an audio device.

2. Description of the Related Art A conventional method for measuring the high frequency distortion of a speaker is to perform frequency analysis with an FFT analyzer. The high frequency distortion measuring method described above will be described below with reference to the drawings.

FIG. 7 shows a conventional measuring method for measuring harmonic distortion.
In FIG. 7, 1 is an oscillator, 2 is an amplifier for driving a speaker 3, 4 is a microphone, and 5 is an FFT analyzer for analysis using a fast Fourier transform (FFT) algorithm. The sine wave signal generated by the oscillator 1 is amplified by the amplifier 2 and drives the speaker 3. The sound pressure signal from the speaker 3 is converted into an electric signal by the microphone 4 and then FF
The T analyzer 5 calculates the fundamental frequency component and harmonic components.

Problems to be Solved by the Invention However, in the conventional harmonic distortion analysis method described above, the FF
Uses the T (Fast Fourier Transform) algorithm to analyze large amounts of data, typically one frequency.
There was a problem that we had to import 512 to 4096 points of data.

That is, when the sampling time is Δt (second), the acquisition time T (second) of one frequency is 10 data points.
With 24 points, T = Δt * 1024 (seconds) is required. When trying to measure 20Hz to 20KHz in the audible band by separating 301 points. If the measurement frequency range is divided into 200Hz, 2KHz, and 20KHz from the measurement frequency resolution, the measurement frequency range of the FFT analyzer is calculated up to the third harmonic. When it does, it becomes 600Hz, 6KHz, 60KHz. Frequency resolution is 600Hz / 10
24 = 0.586Hz, 5.86Hz, 58.6Hz, so T is 1 / 0.586 = 1.7
(Seconds), 170 (milliseconds), 17 (milliseconds). Therefore, even if the averaging process was not performed, the total acquisition time of 301 points was 188.7 seconds.

In view of the above problems, the present invention is to provide a method for measuring harmonic distortion of a speaker that can analyze harmonic distortion in a short time with a small amount of data.

Means for Solving the Problems In order to solve the above problems, the method for measuring harmonic distortion of a speaker according to the present invention uses a sine wave of a frequency as an input signal,
Is applied to the speaker while discretely changing over the measurement range, the sound pressure of the sound reproduced by the speaker is converted into an electric signal by the microphone, and then converted into a digital signal by A / D conversion. After cutting out one cycle of data, the sampling time Δt is changed at each frequency by interpolation, and Δt * N
The harmonic distortion is obtained by Fourier series approximation with N points of data satisfying = 1 /.

Function The present invention can measure the harmonic distortion of the speaker with a small amount of data in a short time by the method described above. That is, a sine wave signal of a frequency is input to a speaker while discretely changing a frequency within a measurement range from an oscillator, and a generated sound pressure signal is converted into an electric signal by a microphone. This signal is converted into a digital signal by the A / D converter and stored. Data that is an integral multiple of at least one cycle of each frequency is extracted, and interpolation is performed by changing the sampling time so that the number of data satisfies the following equation. That is, if the frequency and period are t, the sampling time at the time of capturing is Δt, the number of data is M, the sampling time after interpolation is Δt ′, and the number of data is N. Δt ′ is determined so that the relationship of N and M are integers. At this time, the interpolation is performed so that the number of data points N is the same even if the oscillation frequency is different. Data continuity can be realized by this interpolation. Then, the fundamental wave component, the second-order harmonic wave, and the third-order harmonic wave component are obtained by Fourier series approximation to obtain harmonic distortion.

Embodiments Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows a method for measuring harmonic distortion of a speaker that realizes the method for measuring harmonic distortion of a speaker according to the present invention. In FIG. 1, 1 is an oscillator that generates a sine wave signal of a frequency, 2 is an amplifier that drives a speaker, and 4 is a microphone. The sine wave signal of the frequency is used as the input signal, and the speaker 3 is driven while discretely changing the frequency within the measurement range. The reproduced sound from the speaker 3 is converted into an electric signal by the microphone 4, and is continuously taken into the memory 7 as a digital signal through the A / D converter 6. The sampling time for capturing should be selected so that the number of data in one cycle is 10 or more. That is, when the measurement upper limit frequency is 20 KHz and the number of data points is 10, 5 [μsec] = Δt. The more the number of data points of one frequency, the higher the accuracy of interpolation, but the amount of data increases. Therefore, the sampling time Δt may be changed with the measurement frequency to prevent the increase of the number of data. Then, waveform processing is performed for each frequency to obtain harmonic distortion.

The waveform processing in which the frequency is specified will be described below. Second
The figure shows the data stored in the memory 7 after being converted from an analog signal to a digital signal by the A / D converter 6. In FIG. 2, * indicates original signal data stored as sound pressure for each sampling time Δt. Next, as shown in FIG. 3, the waveform cut-out processing circuit 8 takes out data for at least one cycle t time. Based on this data The sampling time is changed from Δt to Δ so that the following equation is satisfied.
Change to t '. Data is interpolated by the interpolation processing circuit 10 and 1
Interpolation data having a cycle of N points is obtained. FIG. 4 shows the data after the interpolation.

Note that M is the number of data points in one cycle before interpolation and is an integer value, and N is the number of data points in one cycle after interpolation and is an integer value. If N is usually 50 to 100 points, it is a sufficient number to obtain the distortion characteristic of the speaker. The data of FIG. 4 is subjected to Fourier series approximation processing by the Fourier series approximation processing circuit 10 to obtain the fundamental wave, second order, third order, ... Nth order harmonic component.
This value is displayed on the display circuit 11. The interpolation processing circuit 9 is based on the assumption that the Fourier series approximation is based on a continuous periodic function, and when the data shown in FIG. This is done in order to generate an error that harmonic distortion becomes larger than it actually is due to the discontinuity. FIG. 6 shows the data after the interpolation, and the discontinuity point as in FIG. 5 does not occur.

In this embodiment, the Fourier series approximation process is performed for one cycle, but data for one cycle or more may be used.

Further, the interpolation data of one cycle may be repeated several times, and the pseudo continuous wave arranged in a continuous manner may be subjected to FFT processing.

As described above, according to the present invention, the sine wave of the frequency is applied as an input signal to the speaker while being discretely changed within the measurement range, and the sound pressure output from the speaker is converted into an electric signal by the microphone. , A / D conversion is used as a digital signal and one cycle of data for each frequency is cut out, and then the sampling time Δt is changed at each frequency by interpolation. Since the harmonic distortion is obtained by Fourier series approximation with the data of N points satisfying the above, it is sufficient to capture at least one period of data, and therefore the capture time is FFT by the FFT analyzer.
It is possible to obtain the effect of being extremely short as compared with the case of using the algorithm.

For example, when capturing frequency data of 301 points selected so that the audible frequency band 20 Hz to 20 KHz is evenly spaced on the logarithmic axis F
FT requires 188.7 seconds as shown above, while the present invention requires However ₁ Becomes Further, since interpolation is performed so that the number of data points in one cycle of each frequency is the same, all the coefficients for obtaining the inner product of the Fourier series approximation may be the same, and a simple memory and a multiplier can be used.

[Brief description of drawings]

FIG. 1 is a block diagram of a speaker harmonic distortion measuring apparatus for realizing a speaker harmonic distortion measuring method according to an embodiment of the present invention, and FIG. 2 is a state diagram showing data after A / D conversion, and FIG. FIG. 4 is a state diagram showing cut-out one waveform data, FIG. 4 is a state diagram showing data after interpolation, FIGS. 5 and 6 are state diagrams showing continuity of data by interpolation, and FIG. 7 is a conventional state diagram. It is a block diagram of an apparatus which realizes a harmonic distortion measuring method. 1 ... Oscillator, 2 ... Amplifier, 3 ... Speaker, 4 ...
Microphone, 6 ... A / D converter, 7 ... memory circuit,
8 ... Waveform cutout processing circuit, 9 ... Interpolation processing circuit, 10 ...
… Fourier class approximation processing circuit, 11 …… Display circuit.

Claims (1)

(57) [Claims] 1. A sine wave of a frequency f is applied as an input signal to a speaker while f is discretely changed in a measurement range, and the sound pressure of the sound reproduced by the speaker is converted into an electric signal by a microphone, and then the The electrical signal is A / D converted into a digital signal, and at least one cycle of data for each frequency is cut out from this digital signal and then interpolated,
The sampling time Δt is changed at each frequency, A method for measuring harmonic distortion of a speaker, which obtains harmonic distortion by Fourier series expansion with N points of data satisfying