JPH05232979A - Reverberation adding device - Google Patents

Abstract

PURPOSE:To smoothly change reverberation in the case of changing reverberation time based upon a feedback type or cascade type FIR filter. CONSTITUTION:A reverberation adding device having a delay memory 2 for sequentially delaying a digital signal and an adder 4 for adjusting the amplitude of the sequentially delayed signal and convoluting the signal to form a reverberation sound is provided with an amplitude adjusting variable multiplier 51 for feeding back the signal delayed for a prescribed time by the memory 2 to its input and adjusting the amplitude of a feedback signal in order to adjust reverberation time and plural amplitude adjusting variable multipliers 31 for adjusting the amplitude of sequentially delayed signals in order to smoothly change the amplitude of a reverberation sound before and after the feedback signal correspondingly to the amplitude adjustment of the feedback signal.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a reverberation adding device for simulating the feeling of presence and expanse of a music venue in a room of a general household or in a vehicle. Particularly, it is an object of the present invention to smoothly change the reverberation when the reverberation time by the FIR (Finite Impulse Response) filter is changed in the reverberation adding apparatus.

[0002]

2. Description of the Related Art Conventionally, as a technique in the field of such a reverberation adding device, there is one described in Japanese Patent Publication No. 2-62875. In this publication, IIR (Infi
(nite Impulse Response) type filter, multiple delay circuits with different delay times are provided in parallel, input signals are commonly applied to these delay circuits, and the combined output of these delayed output signals or a partial output thereof is delayed. Feedback is given to the common input side of the circuit, and the combined output is taken out as a reverberation signal. This is intended to reduce coloration and flutter echo.

As an FIR (Finite Impulse Response) filter type reverberation adding device which does not cause the occurrence of such coloration, there is one described in JP-A-60-51895. In this publication, a delay memory that stores input signal data at a constant sampling period, a first coefficient calculation unit that adds a coefficient to a plurality of data stored in the delay memory, and a first coefficient calculation unit A plurality of distribution addition units that add the calculation results to a plurality of groups for convolution and convolution, a second coefficient calculation unit that adds a coefficient to the output of each distribution addition unit, and a calculation of the second coefficient calculation unit A final addition unit that adds outputs and outputs as a reverberation signal is provided. Further each second
The coefficient calculator of 1 has one delay circuit, two amplitude adjusters and an adder. This makes it easy to set and change the coefficient parameter and to easily set the reverberation characteristic.

The latter is advantageous over the former in that coloration and the like can be surely removed. However, the latter FI
The R-filter type reverberation adding device has a problem in that, when trying to form a reverberation sound for a long time, the number of calculation steps for adding coefficients increases, and therefore the following feedback system is proposed. FIG. 8 is a diagram showing a feedback type FIR filter type reverberation adding apparatus which is a premise of the present invention. The reverberation adding device shown in this figure has a digital signal v
of the adder 1 for inputting i, the delay memory 2 including a RAM (Random Access Memory) having a multi-tap for delaying the signal from the adder 1 at each sampling cycle, and the delay memory 2 A plurality of multipliers 3 connected to the multi-tap, an adder 4 that adds the outputs of the plurality of multipliers 3 to convolve it to form an output signal vo, and a signal from the last tap of the delay memory 2. The signal adjusted by adjusting the amplitude is added to the adder 1
And a multiplier 5 that feeds back to the other input of the.

FIG. 9 is a diagram showing a reverberation pattern obtained by the reverberation adding apparatus of FIG. As shown in FIG. 3A, the signals input to the delay memory 2 are delayed by the delay memory 2, the amplitudes are adjusted by the plurality of multipliers 3, and the signals are convolved by the adder 4 to be formed. Although a reverberation pattern is shown, as shown in FIG. 7B, the delay of an arbitrary point (the final point supplied to the plurality of multipliers 3 in the example of this figure) in the signal delayed in the delay memory 2 is delayed. The amplitude of the signal is adjusted by the multiplier 5 and fed back to the input of the delay memory 2 via the adder 1 to form the reverberation pattern of the reverberation time t1. Such a reverberation pattern is
For example, it is formed by setting a plurality of multipliers 3, coefficients of the multipliers 5, or delay times of the delay memory 2 for taking out signals of the multipliers 3 for each characteristic of a music hall such as a concert hall, a church, and a live house. It In this way, the number of calculation steps is reduced while suppressing coloration.

[0006]

By the way, in the feedback FIR filter type reverberation adding apparatus which is the premise of the present invention, it may be required to shorten or lengthen the reverberation time depending on the preference of the user. Here, the reverberation time is usually defined as the time until the amplitude attenuates from the steady state to -60 dB. When adjusting the reverberation time, the reverberation time t2 is shortened by decreasing the coefficient of the multiplier 5, as shown in FIG. Also, as shown in FIG.
When the coefficient of the multiplier 5 is increased, the reverberation time t3 becomes longer.

However, as shown in FIGS. 3C and 3D, when a reverberation pattern is formed by the feedback signal,
There is a problem that the reverberation pattern has a step. The step generated in such a reverberation pattern reproduces an unnatural sound, which is not preferable. Therefore, an object of the present invention is to provide a reverberation adding device that can easily prevent a step from being generated in a reverberation pattern when the reverberation time is adjusted in view of the above problems.

[0008]

In order to solve the above problems, the present invention forms a reverberation sound by delaying a delay memory for sequentially delaying a digital signal and adjusting the amplitude of the sequentially delayed signal. A reverberation adding device having an adder for adjusting the amplitude is provided with a variable multiplier for amplitude adjustment and a plurality of variable multipliers for amplitude adjustment. The variable multiplier for amplitude adjustment feeds back the signal delayed by the delay memory for a predetermined time to its input, and adjusts the amplitude of the feedback signal in order to adjust the reverberation time. The plurality of variable multipliers for amplitude adjustment adjust the amplitude of the sequentially delayed signal in order to smoothly change the amplitude of the reverberation sound before and after the feedback signal in response to the amplitude adjustment of the feedback signal. ..

The reverberation adding device having a first delay memory for sequentially delaying a digital signal and an adder for adjusting and revolving the amplitude of the sequentially delayed signal to form a reverberant sound is provided as a second reverberation device. A delay memory, a plurality of multipliers for reverberation time adjustment, and a plurality of multipliers for amplitude adjustment are provided. The second delay memory sequentially delays the output signal of the adder. The plurality of multipliers for adjusting the reverberation time fetch the delay signal from the delay memory at arbitrary intervals and adjust the amplitude. The plurality of multipliers for amplitude adjustment are sequentially delayed by the delay memory in order to smoothly change the amplitude of the reverberation sound before and after each of the plurality of multipliers corresponding to the amplitude adjustment of the plurality of multipliers. Adjust the signal amplitude. The output signals of the plurality of multipliers are convolved to form a reverberation sound.

[0010]

According to the reverberation applying apparatus of the present invention, the amplitude-adjustable variable multiplier feeds back the signal delayed by the delay memory for a predetermined time to its input, and adjusts the reverberation time by the feedback signal. The amplitude is adjusted. The plurality of variable multipliers for adjusting the amplitude adjust the amplitude of the sequentially delayed signal in order to smoothly change the amplitude of the reverberation sound before and after the feedback signal in response to the amplitude adjustment of the feedback signal. It This adjustment is obtained by multiplying the coefficient of the plurality of basic variable multipliers by, for example, the proportional constant corresponding to the signal delay time of each variable multiplier in relation to the amplitude adjustment of the feedback signal. It is performed by a relatively simple process.

The output signal of the adder is sequentially delayed by the second delay memory. The delay signals are taken out from the delay memory at arbitrary intervals by the plurality of multipliers for adjusting the reverberation time, and the amplitude is adjusted.
The plurality of multipliers for amplitude adjustment are sequentially delayed by the delay memory in order to smoothly change the amplitude of the reverberation sound before and after each of the plurality of multipliers corresponding to the amplitude adjustment of the plurality of multipliers. Adjust the signal amplitude. The output signals of the plurality of multipliers are convolved to form a reverberation sound.
Such adjustment can be obtained by multiplying the coefficients of the plurality of basic variable multipliers as described above by, for example, a proportional constant corresponding to the signal delay time of each variable multiplier in relation to the amplitude adjustment of the feedback signal. Therefore, it is performed by a simple process without changing the basic characteristics.

[0012]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a reverberation applying apparatus according to a first embodiment of the present invention. The reverberation adding apparatus shown in the figure has an adder 1 for inputting a digital signal vi to one side thereof, and a RAM (Random Accumulation) having a multi-tap for delaying the signal from the adder 1 at every sampling cycle, for example.
ESS Memory), a plurality of variable multipliers 31 connected to the multi-tap of the delay memory 2 and whose coefficients are changed by a method described later, and outputs of the plurality of variable multipliers 31. An adder 4 for adding for convolution and forming an output signal vo, and the variable multipliers 3
The amplitude of the signal from any delay tap of the delay memory 2 to 1 is adjusted in conjunction with the coefficient change of the plurality of variable multipliers 31, and the signal thus adjusted is added to the adder 1
And a variable multiplier 51 which feeds back to the other input of the.

Assuming that the coefficient of the variable multiplier 51 for changing the reverberation time is set to a · G, its input signal considering the feedback signal to the delay memory 2 is as follows.
Here, if the initial value of the signal vi input to the delay memory 2 is vi (0) and the sampling period is τ, the multi-tap output signal of the delay memory 2 is vi (τ), vi
(2.tau.), ..., Vi (n.tau.), G is a reference coefficient, and a is a coefficient that is corrected each time the reverberation time is changed.
The input signal considering the feedback signal to the delay memory 2 is as follows, where nτ is the delay time for extracting the feedback signal.

Vif = vi (τ) + vi (nτ)  (aG) + vi (2nτ)  (aG) ² + ... + vi (knτ)  (aG) ^k + ... (1) Next The relationship between the coefficient G and the correction coefficient a will be described. FIG. 2 is a diagram for explaining the relationship between the coefficient G of the variable multiplier 51 and its correction coefficient a. As shown in the figure, the total delay time T of the delay memory 2 is T = nτ, and the reference reverberation time is TR.
1 and assume that the reverberation time has changed to TR2 from this state. In this case, the coefficient of the variable multiplier 51 is changed from G to a · G. Reverberation time is defined as:

20 LOG (G) ^k1 = -60 dB, therefore LOG (G) =-3 / K1, where K1 = TR
1 / T, LOG (a · G) = LOG (a) + LO
G (G) = − 3 / K2, where K2 = TR2 / T, and thus LOG (a) = 3 (1 / K1-1 / K2). Also, LOG (a ・ G) / LOG (G) = TR1 / T
R2 is established. Thus, a = 10 ^{3 (T / TR1-T / TR2)} is obtained.

Next, the FIR filter is the delay memory 2,
It is composed of a plurality of variable multipliers 31 and an adder 4, but the coefficient of the plurality of variable multipliers 31 in one music venue is g.
(1), g (2), ..., G (n).
FIG. 3 is a diagram showing the windowing process, and FIG. 4 is a diagram for explaining the relationship between the coefficient G of the variable multiplier 51 and the coefficient g (m11) of the reverberation group having the variable multiplier 31. Figure 3
As shown in, the reverberation signals formed by the coefficients of the plurality of variable multipliers 31 are not regularly arranged. However, the reverberation signal has a certain reverberation time in a certain mass. Therefore, the reverberation signals having the same reverberation time are grouped into the above g (1), g (2), ...
(N) is replaced by [g (m1) · b (m1), {g (m1) · b (m
1)} ^m12-m11 , {g (m1) ・ b (m
1)} ^m13-m11 , ...,], [g (m2) · b (m2),
{G (m2) ・ b (m2)} ^m22-m21 , {g (m2) ・
b (m2)} ^m23-m21 , ...,], [g (m3) ・ b (m
3), {g (m3) · b (m3)} ^m32-m31 , {g (m
3) ・ b (m2)} ^m33-m31 ...,], ...
(M1), b (m2), b (m3), ... Are correction coefficients for the reverberation group, and the reverberation time of each is Tm1,
Tm2, Tm3, ... In this case, the reverberation pattern v1 of the FIR filter is the adder 4 in which feedback is ignored.
The output signal of can be expressed by the following equation.

V1 = [vi (m11τ) · g (m1) · b (m1) + vi (m12τ) · {g (m1) · b (m1)} ^m12-m11 + vi (m13τ) · {g (m1) ) ・ B (m1)} ^m13-m11 + ...] + [vi (m21τ) ・ g (m2) ・ b (m2) + vi (m22τ) ・ {g (m2) ・ b (m2)} ^m22-m21 + vi (m23τ) ・ {g (m2) ・ b (m2)} ^m23- m21 + ...] + [vi (m31τ) ・ g (m31) ・ b (m3) + vi (m32τ) ・ {g (m32) ・^{b (m3)} m32-m31} + vi (m33τ) · {g (m33) · b (m3)} m33-m31 + ... ] + ... ... (2) where taken from Day Rey memory 2, for example, every 1 sampling period Is supplied to the multiplier 31. As shown in FIG. 4, m11 <m12 <m13, ... Are delay memos in the reverberation group m1. 2 shows the extraction order of the multiplier 31 from. For example, if m12 leads m11 by one sampling period, m12-m11 = 1, and if m13 leads m11 by two sampling periods, m12-m11 = 2.

Here, for example, when the reverberation time of FIG. 2 is set to TR1, in this case, b (m
1) = b (m2) = b (m3) = ... = 1.0 is set. From FIG. 4, n1 and nn1 are determined so that TR1 = n1 · T = nn1 · Tm1. Then 20 LOG
(G) ⁿ¹ = −60 dB, therefore LOG (G) = −
3 / n1, similarly LOG (g (m11)) = − 3 / nn
1 is established. From these, n1 = −3 / LOG (G),
nn1 = −3 / LOG (g (m11)), and from the above equation, −3T / LOG (G) = − 3Tm1 / LOG (g
(M11)) is established. Therefore, LOG (g (m1
1)) = (Tm1 / T) LOG (G) holds. When this is transformed, g (m11) = 10 ^{(Tm1 / T) LOG (G)} . Therefore, in each reverberation group, g (m11) and G have the relationship of the above equation, and the average of the entire reverberation group is G.
Matches Therefore, even if G changes, if the coefficient of each reverberation group can maintain the above relationship, it is possible to maintain the reverberation characteristic as it is, without the step difference being generated for each feedback as in the conventional case. That is, when the coefficient of the variable multiplier 51 is changed to a · G to adjust the reverberation time, g (m11) · b (m1) = 10 from the above equation as well.
^{It is (Tm1 / T) LOG (aG)} . Therefore, when the reverberation time is changed from TR1 to TR2, g (m1) · b (m1) / g (m1) = 10
^{(Tm1 / T) LOG (aG)} / 10 ^{(Tm1 / T) LOG (G)} , that is, b (m
1) = 10 ^{(Tm1 / T) (LOG (aG) -LOG (G))} = 10
^{(Tm1 / T) (TR1 / TR2-1) LOG (G)} is obtained. Similarly, other coefficients b (m2), b (m3) ... Are obtained.

From the above, the output vo of the adder 4 is obtained by the following calculation. vo = [vi (m11τ) · g (m1) · b (m1) + vi (m12τ) · {g (m1) · b (m1)} ^m12-m11 + vi (m13τ) · {g (m1) · b (M1)} ^m13-m11 + ...] + [vi (m21τ) ・ g (m2) ・ b (m2) + vi (m22τ) ・ {g (m2) ・ b (m2)} ^m22-m21 + vi (m23τ) ) ・ {G (m2) ・ b (m2)} ^m23- m21 + ...] + [vi (m31τ) ・ g (m31) ・ b (m3) + vi (m32τ) ・ {g (m32) ・ b (m3 ^{)} m32-m31 + vi (} m33τ) · {g (m33) · b (m3)} m33-m31 + ... ] + ...] + [[vi ((n + m11) τ ) · g (m1) · b (m1 ) + Vi ((n + m12) τ) ・ {g (m1) ・ b (m1)} ^m12-m11 + vi ((n + m13) τ) ・ {g (m1) ・ b (m1)} ^m13-m11 +…] + [Vi ((n + 21) τ) · g (m2 ) · b (m2) + vi ((n + m22) τ) · {g (m2) · b (m2)} m22-m21 + vi ((n + m23) τ) · {g (m2 ) ・ B (m2)} ^m23-m21 +…] + [vi ((n + m31) τ) ・ g (m31) ・ b (m3) + vi ((n + m32) τ) ・ {g (m32) ・ b (m3 )} ^m32-m31 + [vi ((n + m33) τ ) · {g (m33) · b (m3)} m33-m31 + ... ] + ...] · a · G + [[vi ((2n + m11) τ ) · g (M1) ・ b (m1) + vi ((2n + m12) τ) ・ {g (m1) ・ b (m1)} ^m12-m11 + vi ((2n + m13) τ) ・ {g (m1) ・ b (m1) } ^M13-m11 + ...] + [vi ((2n + m21) τ) ・ g (m2) ・ b (m2) + vi ((2n + m22) τ) ・ {g (m2) ・ b (m2)} ^m22-m21 + vi ( 2n + m23) τ) · { g (m2) · b (m2)} m23-m21 + ... ] + [vi ((2n + m31) τ ) · g (m31) · b (m3) + vi ((2n + m32) τ) · {g (m32) · b ( m3)} m32-m31 + vi ((2n + m33) τ) · {g (m33) · b (m3)} m33-m31 + ... ] + ...] · (a · G) ² + (3) FIG. 5 is a diagram showing the configuration of the reverberation pattern correction unit. The reverberation pattern correction unit shown in the figure receives a plurality of variable multipliers 31 and variable multipliers 51 in response to a reverberation time change instruction from the user.
, And various basic coefficients g (1), g (2), ..., G to be set in the variable multiplier 51.
An FIR filter coefficient storage unit 101 for storing (n),
When there is an instruction to change the reverberation time, g (1), g
(2), ..., G (n) are pre-classified for each reverberation sound group and stored, and a coefficient a · G of the feedback signal of the variable multiplier 51 from the reverberation time change instruction.
And a coefficient for forming each group reverberation sound for a new reverberation time from the reverberation time of each group reverberation sound and the delay time of the entire filter based on the coefficient a · G of the feedback signal. The reverberation correction unit 104 for correction is included.

FIG. 6 is a diagram showing a reverberation pattern corrected by the first embodiment. As shown in this figure (a),
When the reverberation adding device operating with a predetermined basic coefficient is instructed to change the reverberation time from the outside, according to the present embodiment, the coefficient of the variable multiplier 51 is changed as shown in FIGS. Is not only changed, but also the coefficients of the plurality of variable multipliers 31 are changed so that there is no step difference in each feedback, unlike the conventional case.
Moreover, since the characteristics of the coefficients of the plurality of variable multipliers 31 can be maintained and changed easily, natural sound can be easily reproduced and the initial purpose can be achieved.

In the above, the reverberant sound formed by the multiplier 31 is divided into a plurality of groups and the respective correction coefficients are obtained, but it is also possible to perform the correction with approximately one coefficient. FIG. 7 is a diagram showing a reverberation adding apparatus according to the second embodiment of the present invention. The reverberation adding device shown in the figure has a multi-tap RAM (Random Access Memo) for delaying the digital signal vi, for example, every sampling period.
ry) etc., a plurality of variable multipliers 31 which are connected to the multi-tap of the delay memory 2 and whose coefficients are changed by a method described later, and the output of each of the plurality of variable multipliers 31 is convolved. A first FIR filter including an adder 4 for adding in order to eliminate the delay, and a delay memory 12 including a RAM (Random Access Memory) having a multi-tap for delaying a signal from the adder 4 for each sampling period, for example. , A plurality of variable multipliers 13 that are connected to the multi-tap of the delay memory 2 and whose coefficients are changed by a method described later, and an adder 14 that adds the outputs of the plurality of variable multipliers 31 for convolution. ,
It includes a second FIR filter that is cascaded with the first FIR filter to determine the reverberation time.

In the first embodiment, since the FIR filter is fed back, some coloration is unavoidable. Therefore, as in the second embodiment, the interval between the multipliers 13 is slightly shifted. As will be described below, the calculation scale can be made smaller than that using only a conventional FIR filter without coloration. When only the coefficient of the multiplier of the second FIR filter is changed in order to change the reverberation time, the reverberation pattern has a step difference before and after the multiplier as in the above case. However, the FIR filter described later can smoothly and easily change the reverberation time. First, the fact that the calculation scale can be reduced will be described. In the delay memory 2, the interval between each of the plurality of variable multipliers 31 is a sampling cycle τ,
Let the predelay time be nτ. In the delay memory 12, the time intervals of the multipliers 13 are m1τ, m2τ,
..., mkτ. Here, the following approximation is established.

N.apprxeq.m1.apprxeq.m2.apprxeq..apprxeq.mk According to the condition of the above equation, about n ² reverberant sounds are formed by this cascade type FIR filter. The number of multipliers for forming this reverberation is about 2n.
When the conventional cascade type is not used, about n ² arithmetic units are simply required, so that the ratio can be reduced by about 2n / n ² = 2 / n. Therefore, this cascade type F
According to the IR filter, the calculation scale can be reduced while preventing coloration.

Next, it will be explained that the reverberation time in such a cascade type FIR filter can be changed smoothly and easily. Let g be the coefficient of the plurality of variable multipliers 31 in FIG.
(1), g (2), ..., G (n), and these coefficients are classified into groups having the same reverberation time as in the first embodiment. The coefficients of the plurality of multipliers 13 are 1, a · G,
(A · G) ² , ..., (a · G) ^k−1 are set in this order. This multiplier 13 is a variable multiplier 51 for feedback of the first embodiment.
It corresponds to.

Similarly to the first embodiment, the correction coefficient a of the multiplier 13 is obtained by the reverberation time change instruction, and the coefficients g (1), g (2), ..., G (of the variable multiplier 31 are obtained. The correction of n) is performed. The reverberation pattern obtained by synthesizing the first and second FIR filters is smooth without a step before and after the multiplier 13, as shown in FIG. Further, the processing of the above signals can be similarly performed by the simple reverberation pattern correction unit shown in FIG.

[0026]

As described above, according to the present invention, F
When a reverberation sound obtained by an IR filter is fed back or cascaded to form a reverberation sound for a long time, when the reverberation time is changed, there is no discontinuity in the reverberation pattern and it becomes smooth, so that a natural sound can be formed. ..

[Brief description of drawings]

FIG. 1 is a diagram showing a reverberation applying apparatus according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating a relationship between a coefficient G of a variable multiplier 51 and a correction coefficient a thereof.

FIG. 3 is a diagram showing a state of windowing processing.

FIG. 4 is a diagram showing a coefficient G of a variable multiplier 51 and a variable multiplier 3;
It is a figure explaining the relationship with coefficient g (m11) of a certain reverberation group.

FIG. 5 is a diagram showing a configuration of a reverberation pattern correction unit.

FIG. 6 is a diagram showing a reverberation pattern corrected by the first embodiment.

FIG. 7 is a diagram showing a reverberation adding device according to a second embodiment of the present invention.

FIG. 8 is a diagram showing a feedback FIR filter type reverberation adding apparatus that is a premise of the present invention.

9 is a diagram showing a reverberation pattern obtained by the reverberation adding device of FIG.

[Explanation of symbols]

 1, 4, 14 ... Adder 2, 12 ... Delay memory 31, 13 ... Plural variable multipliers 51 ... Variable multipliers

Claims (3)

[Claims] 1. A reverberation adding apparatus having a delay memory (2) for sequentially delaying a digital signal, and an adder (4) for adjusting and convolving the amplitude of the sequentially delayed signal to form a reverberation sound. A variable multiplier (5) for amplitude adjustment, which feeds back a signal delayed by a predetermined time by the delay memory (2) to its input and adjusts the amplitude of the feedback signal to adjust the reverberation time.
1) and a plurality of variable multiplications for amplitude adjustment for adjusting the amplitude of the sequentially delayed signal in order to smoothly change the amplitude of the reverberation sound before and after the feedback signal in response to the amplitude adjustment of the feedback signal. And a reverberation adding device (31). 2. A reverberation having a first delay memory (2) for sequentially delaying a digital signal, and an adder (4) for adjusting and convolving the amplitude of the sequentially delayed signal to form a reverberation sound. A second delay memory (12) for sequentially delaying the output signal of the adder (4), and a delay signal taken out from the delay memory (12),
Multiple multipliers for adjusting reverberation time for adjusting amplitude (13)
Corresponding to the amplitude adjustment of the plurality of multipliers (13), the delay memory (2) is sequentially delayed to smoothly change the amplitude of the reverberation sound before and after each of the plurality of multipliers (13). And a plurality of amplitude adjusting multipliers (31) for adjusting the amplitudes of the generated signals, and convolution of the output signals of the plurality of multipliers (13) to form a reverberation sound. .. 3. The reverberation adding apparatus according to claim 1 or 2, wherein the multipliers (31) commonly correct the amplitudes of the signals that have a constant change in amplitude among the sequentially delayed signals.