JPS6321363B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a signal processing device, and is particularly suitable for use when generating and processing signals of various frequencies.

[Summary of the invention]

The present invention performs D/A conversion on the digital code read from a storage device in which an analog signal is digitally coded and written in advance, and at this time, changes the playback cycle of the D/A converted analog signal. In addition to controlling the readout clock pulse of the storage device, there is also a filter constructed of, for example, a multiplication type D/A converter, for removing unnecessary frequency components and extracting a desired band signal after the D/A converter. The filter is characterized in that a cutoff frequency of the filter is changed in accordance with a control signal for setting the reproduction period of the analog signal.

As a result, the unnecessary frequency components that change according to changes in the reproduction period of the analog signal can be removed without using many filters.

[Conventional technology]

When measuring the characteristics of various electronic circuits such as amplifiers, room acoustic characteristics, etc., a reference signal with a predetermined band spectrum is added to the input side of the measurement target, and the phase and frequency of the reference signal obtained from the output side are measured. , a method of determining various characteristics of the measurement target based on changes in the envelope or the like may be used.

For example, the method announced by Mr. Schroeder has been known as a method for measuring room acoustic characteristics. In this method, an impulse reference signal such as a tone burst signal having the required band spectrum is emitted from a speaker, the emitted sound is captured by a microphone, and the signal is added to a signal processing device, and the reverberation is generated based on the above signal. I am trying to obtain a curve (envelope). According to this method,
The ensemble average of the reverberation curve can be determined with just one measurement.

When the above measurement method is actually applied, conventionally, data analysis has been performed by digitally processing the analog input signal to be measured. For this reason, it takes a lot of time and effort to obtain the measurement results, and the apparatus has become large-scale. A method has been proposed that improves this point and obtains a reverberation curve in real time by processing the input signal under measurement as an analog signal. In this method, a reverberation curve is obtained by synchronously detecting an input signal to be measured in a signal processing device. In this case, as will be described later, since unnecessary frequency components are included in the multiplication output of the synchronous detection circuit, a low-pass filter for removing them is provided in the signal processing device.

Therefore, in such a reverberation curve measuring device, when the frequency of the reference signal is changed to change the frequency to be measured, the frequency and band of unnecessary frequency components in the multiplication output also change. Therefore, it is necessary to change the cutoff frequency of the low-pass filter according to this change.

[Problem that the invention seeks to solve]

In order to change the cut-off frequency of the low-pass filter, the usual method requires providing a plurality of low-pass filters each having a different cut-off frequency, and switching between these in accordance with the frequency to be measured. This not only complicates the device but also increases the number of circuit elements that must be accurate, making it extremely difficult to obtain highly accurate measurement data.

Also, when measuring the characteristics of various other electronic circuits, unnecessary frequency components may appear when processing the signals obtained on the output side, and the frequency and band of these components may differ from the frequency of the reference signal, that is, the frequency under test. Since it varies with frequency, the cutoff frequency of the low-pass filter must be changed to remove it.

[Means for solving problems]

In the present invention, the reading speed of the storage means in which digitally coded analog signals are stored can be changed by the clock pulse control means (comprised of the frequency divider 18 and the code generator 18a in the embodiment), and The cutoff means of the filter is changed in accordance with the change in the reading speed.

[Effect]

Unnecessary frequency components corresponding to the cycle of the reproduced analog signal read from the storage means can be removed without using many filters.

〔Example〕

FIG. 1 shows a circuit system of the reverberation curve measuring device 1 to which the present invention is applied.

This reverberation curve measuring device 1 mainly includes a tone burst signal input circuit 2 and a signal processing device 3. Furthermore, a speaker 5 and a microphone 6 are placed at predetermined positions in the room 4 where the reverberation curve is measured.

The tone burst signal input circuit 2 includes an address counter 7, a ROM 8, a D/A converter 9, and an amplifier 10. Further, the signal processing device 3 includes multipliers 11a and 11b, and a low-pass filter 12.
a, 12b, squarer 13a, 13b and adder 1
4 and an amplifier 1.
6, clock generator 17, frequency divider 18, frequency division ratio specification code generator 18a, address counter 19,
ROM20a, 20b, D/A converter 21a, 2
1b, squarer 22, divider 23, calibration coefficient unit 24
and an output terminal 25.

Then, the tone burst signal output from the tone burst signal input circuit 2 is applied to the speaker 5,
A signal obtained by collecting the radiated sound by the microphone 6 (hereinafter referred to as an input signal) is synchronously detected by the synchronous detection circuit 15, and the envelope of the reverberation curve obtained based on this detection output is output to the output terminal 25.
I try to take it out more often.

Next, a circuit operation for obtaining a reverberation curve using the above configuration will be explained.

First, a method for obtaining the tone burst signal e ₀ will be explained with reference to FIGS. 2 and 3.

Figure 2 A has constant amplitude and a single frequency.
₀ ( _{e.g. 0} = 50Hz _to 10KHz). FIG. 3A shows the frequency spectrum of the sinusoidal signal _e1 . Figure 2B
represents a modulated signal e ₂ whose level changes over a predetermined time d. For example, e ₂ =0.54− as this modulation signal e ₂
A signal having a waveform expressed by 0.46 cos (2πt/d) can be used. Figure 3B shows the above modulated signal.
This shows the frequency spectrum of e ₂ , and this modulated signal e ₂ has signal energy concentrated near zero frequency. Next, the above modulation signal e ₂
When the above sinusoidal signal e ₁ is amplitude modulated by
A tone burst signal e ₀ =e ₁ ×e ₂ having a waveform close to an impulse waveform as shown in FIG. C can be obtained. FIG. 3C shows the frequency spectrum of the tone burst signal e ₀ , and the energy of the signal is concentrated around the center frequency ₀ , so that measurement is hardly affected by external noise.

Tone burst signal obtained as described above
e ₀ is digitally encoded and stored in ROM8 in Figure 1.
is written in advance.

Further, the waveform of the signal e ₃ =cos (2π ₀ t), which is a reference subcarrier for synchronous detection, is digitally coded and written in advance in the ROM 20a.
The waveform of a signal e ₄ =sin (2π ₀ t), which is a reference subcarrier for synchronous detection, is digitally coded and written in advance in the ROM 20b. Furthermore, ROM
One of 20a and 20b is omitted, and ROM2
Add part of the output of 0a or 20b to a 90° shift register, and convert the output of this shift register to the D/A.
It may also be added to converter 21a or 20b.

On the other hand, the clock generator 17 always has a constant frequency.
The clock pulse of _c is output. This clock pulse is divided into the input clock pulse p by a frequency divider 18 whose frequency division ratio is set based on the code obtained from the code generator 18a to _c1 = (1/m ⁿ ) _c . This clock pulse p is sent to address counters 7 and 19.

The address counter 7 is sequentially advanced by the input clock pulse p, whereby the digital tone burst signal e ₀ written in the ROM 8 is read out. The read signal is converted into an analog tone burst signal _e0 by the D/A converter 9, and after being amplified to an appropriate level by the amplifier 10,
It is added to the speaker 5 and radiated as sound. The radiated sound is a combination of direct sound and reverberant sound reflected from the walls of the room 4, etc., and is collected by the microphone 6.

At this time, the next input signal from microphone 6 is
e _i is obtained.

e _i = m(t)・sin(2π ₀ t+φ) − In the formula, m(t) is the tone burst signal
The amplitude change of e ₀ depending on the state of room 4,
That is, it represents the envelope of the desired reverberation curve. Further, φ in the second term of the equation represents the amount of phase change in the tone burst signal e ₀ depending on the state of the room 4.

The input signal e _i obtained as described above is amplified to an appropriate level by the amplifier 16 and then subjected to envelope detection by the synchronous detection circuit 15.

On the other hand, as the address counter 19 is sequentially advanced by the input clock pulse p,
The digital signal _e3 is read out from the ROM20a, and the digital signal e3 is read out from the ROM20b.
e ₄ is read. These signals e ₃ and e ₄ are D/A
Analog signals e ₃ ,
e ₄ and applied to one input terminal of each of the multipliers 11a and 11b. Multiplier 11a, 1
The input signal e _i is applied to the other input terminal of 1b, and the multiplier 11a multiplies the signals e _i and _e3 , and the multiplier 11b multiplies the signals e _i and _e4 . .

Therefore, the multiplication output F ₁ from the multiplier 11a is F ₁ =e _i ×e ₃ = m(t)・sin(2π ₀ t+φ)・cos(2π ₀ t)
=m(t)/2{sin(4π ₀ t+φ)+sinφ}− is obtained. Also, the multiplier output from the multiplier 11b
As F ₂ , F ₂ = e _i ×e ₄ = m(t)・sin(2π ₀ t)・sin(2π ₀ t)=
m(t)/2{-cos(4π ₀ t+φ)+cosφ}− is obtained.

Next, these multiplication outputs F ₁ and F ₂ are passed through the low-pass filter 1
2a and 12b, respectively, and remove the unnecessary frequency component represented by the first term in { } in the above equation, the outputs G ₁ and G ₂ of the low-pass filters 12a and 12b are G ₁ = m(t)/2sinφ − G ₂ =m(t)/2cosφ − is obtained.

Note that the low-pass filters 12a and 12b will be explained in detail later, but in the present invention, these low-pass filters 12a and 12b are connected to the multiplier D/
It is constructed using an A converter. Then, a code signal for setting the frequency division ratio of the frequency divider 18 obtained from the code generator 18a is used, and the cutoff frequency is controlled according to the frequency to be measured based on this code signal. ing.

Next, the above outputs G ₁ and G ₂ are squared by squarers 13a and 1, respectively.
3b, square them respectively to form G ₁ ² and G ₂ ² , and then add G ₁ ² and G ₂ ² to the adder 14 to obtain the addition output X as X=G ₁ ² +G ₂ ² = {m(t)/2} ²・{sin ² (φ)
+cos ² (φ)}={m(t)/2} ² − is obtained.

Next, when the above output X is added to the squarer 22, the output Y is obtained as follows: Y=m(t)/2 -.

This output Y is applied to one input terminal of the divider 23, and a signal representing the calibration coefficient "2" from the calibration coefficient unit 24, that is, a signal for doubling the output Y of the above equation, is applied to the other input terminal. By adding, Z=m(t) − is obtained as the output Z of this divider 23.

This output Z is m of the input signal e _i shown in the above formula.
(t) and indicates the envelope of the reverberation curve.

If it is desired to extract this output Z as an output that changes on a logarithmic scale, the desired output can be obtained by further passing the output Z through a logarithmic amplifier.

Next, a case where the frequency to be measured is changed will be explained.

The frequencies of tone burst signal e ₀ and signals e ₃ and e ₄ written in ROM8, 20a, 20b are
₀ , and one wavelength of these signals is made up of a series of digital codes consisting of q discrete quantities. Since the signal read from each ROM 8, 20a, 20b becomes ₀ Hz, is the reading speed at which the above digital code is read out sequentially,
That is, the frequency _c1 of the input clock pulse p that must be sent to the address counters 7 and 19 is:
_c1 = q・₀ .

This input clock pulse p is supplied to the clock generator 1.
7 oscillation frequency _c is divided by frequency divider 18 by 1/(m ⁿ )
(m and n are integers), so _c = q・m ⁿ・₀ ∴ ₀ = _c / q・m ⁿ −.

Therefore, in the reverberation curve measuring device 1 shown in FIG.
The frequency ₀ of the tone burst signal e ₀ applied to the speaker 5, that is, the input signal from the microphone 6
To change the frequency to be measured by changing _the frequency ₀ of e _i , all ROMs ^are
8, 20a, and 20b can be changed in frequency. That is, by keeping m of the frequency divider 18 constant and changing n, the frequency to be measured can be changed logarithmically, and n=1
By changing m as , the frequency to be measured can be changed linearly. In this case, measurements can be performed under the same conditions for each frequency to be measured. The frequency division ratio 1/m ⁿ of the frequency divider 18 is set to the values of m and n by a code signal from the code generator 18a.

Based on the envelope signal Z=m(t) of the reverberation curve expressed by the formula obtained as above,
For example, various room acoustic characteristics such as D value, time center of gravity, reverberation time, etc. can be determined.

Next, the low-pass filters 12a and 12b will be explained. As described above, the low-pass filters 12a and 12b are configured using multiplication type D/A converters (hereinafter simply referred to as MDACs), and the frequency dividers 18
The cut-off frequency is controlled using a code signal for setting the frequency division ratio.

The frequency spectra of the multiplication outputs F ₁ and F ₂ expressed by the above equations are shown in FIG. In Figure 4, the frequency component of center frequency ₂₀ is
This is an unnecessary frequency component represented by the first term in { } in the equation. The center frequency ₂₀ and its band of this frequency component change when the frequency to be measured is changed by changing the frequency division ratio 1/ ^mn using the method described above. This low pass filter 12
a and 12b are provided to remove the above-mentioned unnecessary frequency components.

FIG. 5 shows an embodiment of the low-pass filters 12a and 12b.

These low-pass filters 12a and 12b have resistors R ₁ , R ₂ ,
It is constructed as shown in the figure by an integrating circuit 34 comprising a capacitor Cs and an operational amplifier 30, and an MDAC 31. Then, the multiplication output F ₁ or F ₂ of the above formula is applied to the input terminal 32, and the output G ₁ or G ₂ expressed by the above formula from which the frequency component of the center frequency ₂₀ has been removed is output from the output terminal 33. It is designed to provide the following.

The MDAC 31 receives a D/A conversion output M _o 〓 ⁱ⁼¹ b _i /2 ⁱ determined by the digital code b _i of the code signal given by a predetermined number of bits from the code generator 18 a, and an input added from the operational amplifier 30. The product obtained by multiplying the signal e _r is obtained as the output e _put . Therefore, this e _put is expressed by the following formula.

e _put = e _r ·M _o 〓 ^{i = 1} b _i /2 ⁱ − Note that, as this MDAC 31, for example, a known one constructed of a ladder network can be used.

In FIG. 5, the negative input point of the operational amplifier 30 is virtual ground, and the currents I ₁ , I ₂ , Is flowing through the resistors R ₁ , R ₂ and the capacitor Cs, respectively, are expressed by the following equations.

However, e _io : input F ₁ , F ₂ e _put : output G ₁ , G ₂ e _r : input of MDAC 31 Further, when Kirchhoff's law is applied to the above point, the following relationship is established.

I ₁ +I ₂ +Is=0 − The relationship between e _io and e _put is determined from the above equation as shown in the following equation.

According to this formula, the apparent size of the feedback resistor R ₂ of the integrating circuit 34 changes depending on the set value of the digital code b _i , and this causes the difference between the low-pass filters 12a and 12b. It turns out that the cutoff frequency can be changed.

According to the above, in the reverberation curve measuring device 1 shown in FIG. 1, when the frequency to be measured is changed by changing the frequency _c1 of the input clock pulse p, the cutoff frequencies of the low-pass filters 12a and 12b are changed accordingly. Can be controlled automatically. Therefore, it is not necessary to provide many low-pass filters, and the number of circuit elements that require accuracy can be reduced, so that highly accurate measurement data can be obtained.

The above has described the case where the present invention is applied to a reverberation curve measuring device, but the present invention is not limited to this, and the present invention can be applied to various measurement targets such as electronic circuits such as amplifiers or acoustic circuits. It is.

Next, a general application example of the present invention will be explained with reference to FIG.

In FIG. 6, a reference signal e ₀ having a predetermined band spectrum is digitally encoded and written in the ROM 40 in advance. This reference signal
e ₀ is read out by a clock pulse obtained by dividing the clock pulse from the clock generator 41 into 1/m ⁿ by a frequency divider. The read digital reference signal e ₀ is converted into an analog reference signal e ₀ by the D/A converter 43 and then applied to the input terminal of the measurement object 44 . From the output terminal of this measurement object 44, a reference signal e ₀ is output according to the characteristics of this measurement object 44.
A signal e _i whose frequency, phase, amplitude, etc. have changed is obtained. This signal e _i is processed by a signal processing circuit 45, and required measurement data is obtained from its output terminal 46. Unnecessary frequency components appearing during this signal processing process are removed by a filter 47. This filter 47 is, for example, a low-pass filter shown in FIG. 5, and its cutoff frequency is controlled by a code signal from a code generator 48. The code signal is also applied to the frequency divider 42 to change its frequency division ratio,
The frequency of the reference signal e ₀ read from the ROM 40 is changed to change the frequency to be measured.

According to the above, when changing the frequency of the clock pulse for reading out the ROM 40 to change the frequency to be measured, the cutoff frequency of the filter 47 can be changed accordingly. Note that the filter 47 may be a low-pass filter, a high-pass filter, or a bandpass filter.

〔Effect of the invention〕

According to the present invention, digital codes such as waveform data stored in a storage device are read out at different readout speeds, and this is D/A converted, thereby generating analog signals with different periods and converting the analog signals into is supplied to a filter whose cutoff frequency is controlled according to the reading speed, that is, the cutoff frequency of the filter is changed according to the cycle of waveform data read from the storage device.
Even if waveform data with different periods are read from the storage device, unnecessary frequency components can be removed using the same filter.

[Brief explanation of the drawing]

Fig. 1 is a circuit diagram showing an embodiment in which the present invention is applied to a reverberation curve measuring device, Fig. 2 A, B,
C is a waveform diagram showing the formation process of the tone burst signal, Figure 3 A, B, and C are the frequency spectra of Figure 2 A, B, and C, Figure 4 is the frequency spectrum of the multiplication output, and Figure 5 is the low frequency spectrum. FIG. 6 is a circuit diagram showing an embodiment of the filter. FIG. 6 is a circuit diagram showing a general application example of the present invention. In addition, in the symbols used in the drawings, 1... Reverberation curve measuring device, 4... Room, 5... Speaker, 6...
...microphone, 8...ROM, 9...D/A
Converter, 11a, 11b... Multiplier, 12a, 1
2b...Low pass filter, 17, 41...Clock generator, 18, 42...Frequency divider, 25...Output terminal, 30...Operation amplifier, 31...Multiplication type D/A
Converter, 34... Integrating circuit, 44... Measurement object,
45...signal processing circuit, 47...filter.

Claims (1)

[Scope of Claims] 1. Storage means in which analog signals are digitally encoded and written in advance; D/A conversion means to which digital signals read from the storage means are supplied; and from the D/A conversion means. clock pulse control means for controlling clock pulses supplied to the storage means to set the period of the output analog signal; and clock pulse control means for removing unnecessary frequency components from the analog signal output from the D/A conversion means. filtering means, the filtering means is configured such that its cutoff frequency is controlled in accordance with the clock pulse controlled by the clock pulse control means, and the analog signal output from the D/A conversion means. A signal processing device characterized in that unnecessary frequency components contained in the analog signal are removed according to the period of the analog signal.