JP2511253B2 - Digital graphic equalizer - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital graphic equalizer, and more particularly to a graphic equalizer which extracts a digital audio signal such as a PCM audio signal by varying its level for each frequency band.

2. Description of the Related Art Digital filters used in conventional digital graphic equalizers generally use individual band enhancement /
It has an attenuating filter and is configured to combine them to obtain an overall characteristic, and is not configured by one filter.

Problems to be Solved by the Invention In the conventional FIR digital filter, the band emphasis characteristic K _b
Since it is necessary to provide a filter for each band in order to obtain (boost amount L _b ) and band attenuation characteristic K _c (cut amount L _c ), there is a problem that the configuration becomes complicated.

On the other hand, for example, a technique using an IIR digital filter described in Japanese Patent Laid-Open No. 58-182315 is known. However, when this is used, the phase is non-linear because it is an IIR digital filter, and Due to the error, the SN ratio in the high frequency range is deteriorated, and further, there is a problem that it is easy to oscillate and it is difficult to obtain a desired overall characteristic with one filter.

The present invention consists of one FIR digital filter,
An object of the present invention is to provide a digital graphic equalizer that can obtain desired band enhancement and attenuation characteristics at high speed with a small amount of coefficient data.

Means for Solving the Problems In FIG. 1, the control device 9 and the DSP (digital signal processor) 4 change the filter coefficient according to the designation from the characteristic input section 13 and add the coefficient on the band side. The present invention is an embodiment of a means for obtaining and holding a new coefficient once, and a means for changing the frequency characteristic with the new coefficient.

Function FIR digital bandpass filter virtualized by multiple bands 1
7 _k (k = 1 to n) (FIG. 4) A plurality of coefficients h _ki (i
= 1 to 1) (FIG. 8), the coefficient of the filter of the corresponding band corresponding to the specification from the characteristic input unit 13 is changed according to the specification, and a plurality of bands (1 to 1) are divided into a plurality of bands (1 to l). 1 to n) The same number l of coefficients h _{1 to} h _n (FIG. 5) as the number of coefficients are obtained, and the frequency characteristic of the FIR digital filter 40 (FIG. 4) is changed based on the obtained coefficients. At this time, the coefficient changing means switches the coefficient held during the operation of the FIR digital filter to a new coefficient. Therefore, the characteristics of the FIR digital filter can be switched in real time.

Embodiment FIG. 1 shows a block system diagram of an embodiment of the equalizer of the present invention. In the figure, 13 is a characteristic input section, and 14 is a display section. As shown in FIG. 2, characteristic changing switches SW ₁ , SW ₂ , ...
W _n and display element groups 14 ₁ , 14 ₂ , ... 14 _n are provided. Four
Is a DSP for performing digital filter operation to realize the n-band equalizer characteristics K _{1 to} K _n set in the characteristic input unit 13 as shown in FIG.
It has 4 ₂ , program ROM 4 _1, etc.

The digital audio signal supplied to the input terminal 1 through an optical fiber (not shown) is electrically converted by the optical interface 2 and is also serial / parallel converted, supplied to the latch circuit 3 and latched therein. . The output of the latch circuit 3 has its equalizer characteristic changed by the DSP 4 described later according to the designation of the characteristic input unit 13, supplied to the latch circuit 5, converted into parallel / serial by the optical interface 6, and converted into an optical signal. , Taken out from terminal 15,
It is supplied to the optical fiber.

DSP4 is as shown substantially in the Figure 4, the desired equalizer characteristic by varying the coefficients of a digital bandpass filter 17 ₁ to 17 _n of the n bands by rewriting the coefficient memory 4 ₂ K ₁
.About.K _n , and the digital / bandpass filter shown in FIGS. 6 (A) and 6 (B) can be considered.

6 (A) and 6 (B) respectively show a circuit of the digital bandpass filter and its schematic block system diagram. In FIG. 6A, 1-sample delay units 28 _{1 to} 28 _l and multiplication unit 29
_{1 to} 29 _l , adder units 30 ₂ to 30 _l to obtain the characteristics of the solid line in FIG. 7 (A), low-pass filter (22), 1-sample delay unit 33 _{1 to} 33 _l , multiplication unit 34 ₁ to 34C _l, and solid high-pass filter to obtain the characteristics (high pass filter) 24 of the same at the addition unit 35 ₂ to 35 _l FIG (B) is respectively configured, whereas, 1 sample delay part 32 ₁ Up to 32 _m , all-pass filter (substantially delay circuit and attenuator) 23 in multiplication units (level adjustment units) 31 and 36
Is configured.

The coefficients a _{1 to} a _l of the multipliers 29 _{1 to} 29 _l of the low-pass filter 22 are the tenth
Figure is a function window (A), the other hand, the coefficient b ₁ ~b _l multiplier 34 ₁ to 34C _l of high-pass filter 24 is a function window shown in FIG. (B), the output of the all-pass filter 23 Is supplied to the addition units 30 _m and 35 _m of the intermediate tap m of the low-pass filter 22 and the high-pass filter 24.

In this case, as shown in FIG. 7 (A), the low-pass filter (solid line) and the all-pass filter (dotted line) are combined to obtain the characteristic of the broken line, and these are combined to show in FIG. 7 (C). The band attenuation characteristic K _c is obtained.

On the other hand, in the case of obtaining the band emphasis characteristic, the coefficients a _{1 to} a _l of the multiplication units 29 _{1 to} 29 _l of the band filter 22 are variably set, and FIG.
While being configured to obtain the characteristics shown in
The coefficients b ₁ to b _l of the multiplication units 34 _{1 to} 34 _l are variably set to obtain the characteristic shown in FIG. Similar to the case of obtaining the above band attenuation characteristic, as shown in FIG. 7 (D), the high-pass filter (solid line) and the all-pass filter (dashed line) are combined to obtain the characteristic of the broken line, and FIG. ), Low-pass filter (solid line) and all-pass filter (dashed line)
Are combined to obtain the characteristic of the broken line, and these are combined to obtain the band emphasis characteristic K _b shown in FIG.

FIG. 8 is a simplified version of the bandpass filter shown in FIG. 6 (A).
22 coefficients a _i (a _{1 to} a _l ) and high-pass filter 24 coefficients b _i (b ₁ to a _l ).
b _l ) and the coefficient added by the adder 31 (31 _{1 to} 31 _l ) h _ki (h ₁₁
To h 1) _(and which was the FIG. 11) is set to the coefficient multiplier 29 (29 ₁ ~29 _l), is equivalent to the circuit diagram shown in substantially FIG. 6 (A).

The reason why these are equivalent is as follows.

Referring to FIGS. 6 and 8, the delay circuit 28 (FIGS. 6 (A) and 8) and the delay circuit 33 (FIG. 6 (A)) are
The sample delay signals X (n-1), X (n-2), ..., X (n-1) are generated from the input signal X (n), and these are connected in cascade. The delay circuits 28 and 33 are composed of a plurality of 1-sample delay sections (Z ⁻¹ ) and have the same function.

Further, the delay circuit 32 of FIG. 6 (A) is provided for the purpose of producing the m-sample delay signal X (n-m) from the signal X (n), and a plurality of 1-sample delay sections (Z ^{− 1} )
And has the same function as (a part of) the delay circuit 28. Therefore, the delay circuit 28 multiplies the k-th signal X (n−k) in these delay circuits by the coefficient ak, the delay circuit 33 multiplies the coefficient bk, and the delay circuit 32
Now multiply the coefficients at and bt respectively, and then adder 30l and 35l
And sum of each [Where 1 ≦ k ≦ l] and the sum of at · X (n−m),
as well as [However, the sum of 1 ≦ k ≦ l] and bt · X (n−m))
The obtained, is adding them at the adder 37, advance coefficient ak as shown in FIG. 8, bk, and at, the coefficient h _ki keep by adding bt, by multiplying with them Equivalent to obtaining the sum (equal to the mathematical formula), FIG. 6 (A) and FIG.
Is equivalent to the figure.

The reason why the composite characteristic KC of FIG. 7 (C) is obtained from the composite of FIG. 7 (A) and FIG. 7 (B) is as follows.

In general, when considering the original filter, it is known that this mirror image filter is in a mirror image relationship, that is, the pass band of the original filter becomes the stop band of the mirror image filter, and the stop band of the original filter becomes the pass band of the mirror image filter. However, when such two filters are combined, they interfere with each other to form a flat filter, that is, an all-pass filter, and the characteristics of each other cancel each other out.

However, in reality, the level at the point (frequency) at which the low-pass filter in FIG. 7 (A) intersects the flat characteristic and at the point (frequency) at which the high-pass filter in FIG. 7 (B) intersects the flat characteristic is It does not match the level-Lc in Fig. 7 (C),
Since it is about 6 dB (decibel) lower than that, it is possible to match the level-Lc in the combined characteristic obtained by adding these low-pass filter and high-pass filter, and the characteristic of FIG. 7 (C) is possible. is there. That is, although "interference" is actually caused by the cancellation of these two characteristics, the level of the minimum part of the cut curve (attenuation curve) is set within -Lc as shown in FIG. 7 (C). It turns out that it is possible.

The level at the point (frequency) at which the low-pass filter in FIG. 7 (E) intersects the flat characteristic and at the point (frequency) at which the high-pass filter in FIG. 7 (D) intersects the flat characteristic is shown in FIG. 7 (C). ) Level Lb does not match the level −Lb and is set about 6 dB lower than that, these low-pass filter, high-pass filter, and all-pass filter are added. With the obtained combined characteristic, it is possible to obtain a boost characteristic (enhancement characteristic) having a peak of level + Lb as shown in FIG. 7 (C).

In the case of one band calculated individually, the coefficient does not need to be normalized, but when processing a plurality of bands, it is necessary to normalize the coefficient before simply adding,
The coefficient data to be added in a plurality of bands is normalized in advance.

The band-pass filters shown in FIGS. 6 (A) and 8 are applied to the band-pass filters 17 _{1 to} 17 _l shown in FIG. 4, respectively. The band-pass filter shown in FIG. 5 is shown in FIG. It is equivalent to the band-pass filters 17 _{1 to} 17 _n and the entire adder 19 (indicated by 40), and in practice, the band-pass filter shown in FIG. 5 is used in the present invention. In FIG. 5, h ₁₁ to h ₁ obtained by the addition method shown in FIG. 8 are the coefficients of the first band, h ₂₁ to h _2l obtained by the same method below are the coefficients of the second band, h the ₃₁ to h _3l coefficients of the third band, ..., a h _n1 to h _nl the coefficient of the n band adding section (operation section) 4 _41, 4 _42, ..., respectively by 4 _4l s Is calculated and supplied to the multipliers 29 ₁ , 29 ₂ , ... 29 _{l, respectively} , to obtain a result equivalent to the individual calculation of the bandpass filters 17 _{1 to} 17 _n shown in FIG. 4 by one bandpass filter. I will get it.

The overall characteristics of parallel connection as shown in Fig. 4 are
It is known that it is not a simple superposition of the figures, but an interfering characteristic. That is, a limited condition in which such interference can be sufficiently ignored, that is, the characteristics of FIG. 3 are not crossed, for example, the characteristics K ₁ , K ₄ in FIG.
The interference can be neglected by making the characteristics such that K _K and K _n are separated. That is, the center frequency of the characteristic K ₁ is 75 Hz
The gain of the boost / cut characteristic is almost zero dB at 180 Hz, the center frequency of the characteristic K ₄ is 600 Hz, and the gain of the boost or cut characteristic is almost zero dB at 220 Hz, and the characteristics are not crossed.

Therefore, in the present embodiment, when there is interference, for example, -K ₁ (center frequency 75 Hz) and K ₂ (center frequency 150 Hz)
When the close characteristics such as are combined, the frequencies at which the cut characteristics and the boost characteristics become 0 dB are crossed, so the peaks of each are reduced and the level between the peaks becomes almost flat. . In this embodiment, the strict equalization is limited to the fact that they do not cross each other.

Next, switching of the equalizer characteristic will be described. When the equalizer characteristic is switched, the coefficient data corresponding to the coefficient of the digital filter which substantially constitutes the DSP 4 is switched. It is the CPU (control means) 10 in the control device 9 that controls this switching operation. Is ROM11, RAM1
It is configured to operate according to the flow chart shown in FIG. 9 based on the control signal from 2.

The switch position control data is input by operating the switch of the predetermined band of the characteristic input unit 13 (step 101 in FIG. 9A), and the predetermined segment of the display element group corresponding to this band is displayed ( Step 102),
A filter coefficient (for example, h ₃₁ , h ₃₂ , ... H _3l for the third band) corresponding to the designated equalizer characteristic is retrieved from the ROM (storage unit) 11 (step 103). The end of the input operation does not mean that the filter coefficient is searched and changed after waiting for the final input operation, but the input result causes the change of the filter coefficient at every minimum movement operation of the input operation, and the above operation is completed. Time is the true end.

When the filter coefficients are retrieved from the ROM 11, they are transferred to the RAM 12, and these are sequentially sent to the latch circuit (first holding means) 8 at the interrupt timing (step 111 in FIG. 9B).

The latch circuit 8 latches data (for example, h31, h32, ... h3l) of a predetermined one band (for example, the third band), and each time, a predetermined one of the data 11 to nl is latched. These data for one band (ie eg
The addresses of h ₃₁ , h ₃₂ , ... h3l) are specified.

At the same time, the address data is supplied to the address memory 7, and the address corresponding to each latch data is designated (step 112). At this time, CPU 10 is 1 _m out of 1 _{1 to} n _l.
Up to n _l are sequentially counted up, and it is judged by the counter set that all the data have been input (steps 113 to 115).

The DSP 4, the output data of the latch circuit 8 is written into the coefficient memory 4 ₂ at the specified address (the second holding portion) by the address memory 7. When the filter coefficients are written in all the addresses, the adding units 4 ₄₁ , 4 ₄₂ , ...
4 _4l Is calculated and stored in the memory units 4 ₃₁ to _{43 1} (FIG. 5). In this case, the filter coefficients of the coefficient memory 4 ₂ (1
2) constituting the memory unit (third holding unit) 4 ₃₁ to ₄₃ 1 of 1)
It is stored in the memory part of the non-operating page of the two memory parts and is switched immediately, so that the new coefficient
h ₁ , h ₂ , ..., H _l are obtained, a filter operation is performed for each sampler, and a new equalizer characteristic is obtained.

In this way, each time the equalizer characteristic is specified in the characteristic input unit 13, Is calculated to obtain new coefficients h ₁ , h ₂ , ..., H _l , and the equalizer characteristic is changed in real time.

In the above embodiment, Although the operation of is performed by the DSP 4, it may be performed by the CPU 10.

Further, the band emphasis characteristic K _b and the band attenuation characteristic K _c are not limited to those shown in FIG. 3, and may be of a system in which the gain of the band pass filter is changed.

EFFECTS OF THE INVENTION According to the digital graphic equalizer of the present invention, the configuration is simple because one FIR digital filter is used, and since the FIR digital filter is used, the IIR digital filter is used. Since it is used, compared to the one using IIR digital filter, there is no deterioration in the SN ratio in the high frequency range, problems such as oscillation do not occur, and the set of coefficients is converted into one system set. Therefore, it is possible to combine various characteristics with a small amount of coefficient data,
As a result, the degree of freedom in characteristics is large, and since it can be realized with one filter configuration, the configuration is simple and it is possible to perform calculations in a short time. For example, a high-speed processor such as DSP can be made compact and low cost. In addition, when the desired equalizer characteristic is designated from the characteristic input section, the coefficient held in the holding section is switched while the operation of the FIR digital filter and the operation of the coefficient changing unit are performed in series. Since the frequency characteristics have been changed, the filter characteristics can be changed in real time.

[Brief description of drawings]

1 and 2 are a block system diagram of an embodiment of the equalizer of the present invention and a schematic diagram of a part thereof, FIGS. 3 and 4, respectively.
FIG. 5 is a block diagram of an equalizer characteristic diagram and a band filter for obtaining an equalizer characteristic, FIG. 5 is a circuit diagram of an FIR digital filter used in the equalizer of the present invention, and FIG. 6 is a circuit diagram of an FIR digital band filter. FIG. 7 is a frequency characteristic diagram of the filter, FIG. 8 is a circuit diagram of the FIR digital band filter, FIG. 9 is a flowchart for explaining the operation of the CPU, and FIGS. 10 and 11 are diagrams showing the coefficient values of the filter. . 1 ... digital audio signal input terminal, 4 ... DSP, 4 ₂
... coefficient memory, 7 ... address memory, 8 ... latch circuit,
9 ... Control device, 10 ... CPU, 11 ... Program ROM, 12 ... CP
U, 12 ... RAM for CPU work, 13 ... Characteristic input section, 14 ... Display section, 1
5 ... Output terminal block, 28 _{1 to} 29 _l Multiply block, 30 ₂ to 30 _l , 32 _{1 to} 32
_l … Adder, 40… FIR digital filter.

Claims (1)

(57) [Claims] 1. A characteristic input unit (13) for designating a desired equalizer characteristic for each of a plurality of bands, and a storage unit (11) for storing each coefficient of a virtual FIR digital bandpass filter for each of the plurality of bands. A control means (10) for extracting a coefficient corresponding to the designation from the storage section (11) and supplying the coefficient to the first holding section (8), and from the first holding section (8) A second holding unit (4 ₂ ) in which the supplied coefficient is written, and a plurality of coefficient for each band is added for each of the plurality of coefficients based on the coefficient written in the _second holding unit (4 ₂ ). a calculation unit 4 ₄₁ to 4 _4l) for made from the coefficient changing means comprises a third holding unit for temporarily storing the result of the addition _{(4 31 ~4 3l) (4} ), changes the coefficient Means (4), during the filter operation, based on the coefficient obtained by the addition, frequency characteristics Is changed to configure one digital filter, a digital graphic equalizer.