JPH0749694A - Device for generating reverberation sound - Google Patents

Abstract

PURPOSE:To follow up the movement of a sound source and a listening position in real time by ignoring the angle of incidence on a vertual reflection wall (wall) and all adding incident sounds on the wall. CONSTITUTION:The sounds of the sound source 11-1, etc., are inputted to a reverberation generating device 15, and are delayed according to distances between the sound sources 11-1, etc., and the walls 1-4 to be made incident on the walls 1-4. The sounds reflected by the walls 1-4 are made incident on the walls l-4 also, and are inputted to adder 152-1, etc., to be added regardless of an incident direction. The total sounds are revised by a frequency characteristic of the reflction by the walls through a damping filter 153-1, etc., and generate a time difference through an allpass filter 154-1, etc., to be inputted to a delay line 155-1, etc. Then, the total sounds are inputted to the amplifier 156-1-1, etc., and are received with the attenuation according to the direction of the walls 1-4 and a reflection direction to be transmitted to respective listening positions.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a reverberation sound generating apparatus for performing a calculation for imparting reverberation to a sound emitted from one or a plurality of sound sources and transmitting the reverberation to one or a plurality of listening positions.

[0002]

2. Description of the Related Art A reverberant sound generator has been proposed in the past, which gives a reverberant sound to an input sound signal by calculation, for example, as if playing in a concert hall (eg Japanese Patent Publication No. 4-32600). See the bulletin). In the proposed reverberant sound generator, for example, assuming a room having a reflective wall such as a concert hall, and assuming each position of the sound source and the listener in the room, in addition to the direct sound directly from the sound source to the listener , The sound emitted from the sound source to the listener after being reflected by each reflective wall, the sound reflected by each reflective wall to the listener after being reflected by another reflective wall, etc. to each reflective wall. Reverberation sound is generated by performing calculations that take into account the incident angle, reflection angle, and other factors.

[0003]

In recent years, there has been proposed a technique of moving a sound image in the space with time in addition to making the sound image stand still in the space. For example, in this technique, the reverberation sound as described above is used. Problems arise when combining generators. That is, when the position of the sound source moves in the above proposed reverberant sound generator, the sound radiated from the sound source is reflected by the reflection wall, or is reflected by another reflection wall and the locus of the sound until reaching the listener. Changes and the calculation is performed assuming a finite number of reflection walls, it is necessary to move the finite number of reflection walls as the sound trajectory changes, and all changes in the sound trajectory are calculated. To do so, the amount of calculation is enormous, and it is almost impossible to sequentially move the sound source online and follow the reflected sound. The same applies when the listener (listening position) moves.

In view of the above circumstances, it is an object of the present invention to provide a reverberant sound generator capable of following a movement of a sound source or a listening position in real time.

[0005]

Means for Solving the Problems A reverberation sound generating apparatus of the present invention that achieves the above object is: (1) First calculation for transmitting sounds radiated from one or a plurality of sound sources to a plurality of virtual reflection walls. Means (2) Among the plurality of virtual reflection walls, the sounds incident on the virtual reflection wall are added together, and the reflection sound reflected from the virtual reflection wall is weighted by the reflection direction. (3) The reflected sound reflected from the virtual reflective wall is weighted in the reflection direction to one or a plurality of listening positions. It is characterized in that a third arithmetic means for transmitting is provided.

[0006]

In the reverberant sound generator of the present invention, the reverberation sound is incident on the virtual reflection wall (hereinafter, sometimes abbreviated as "reflection wall" or simply "wall") regardless of the incident angle. Since the configuration is such that all incident sounds are added, there is no need to move the virtual reflection wall even if the sound source moves, and it is possible to perform separate calculations before and after entering the virtual reflection wall. The calculation becomes extremely simple. Further, since the position of the reflection wall may be fixed, the calculation of the route in which the reflection sound reflected from the reflection wall is transmitted to another reflection wall and the route in which the reflection sound reflected from the reflection wall is transmitted to the listening position is extremely performed. It will be easy. Therefore, the reverberant sound can be followed in real time as the sound source or the listening position moves.

It should be noted that the process of adding a plurality of incident sounds incident on the reflecting wall regardless of the incident direction to collectively obtain the reflected sound is, for example, a process in the case of assuming a perfect diffusion wall. Even if such a process of ignoring the incident direction is performed, reverberant sound that is not realistic is not generated.

[0008]

EXAMPLES Examples of the present invention will be described below. Here, a three-dimensional room having L wall surfaces is used as a basic model, and one or a plurality of (K) point sound sources and one or a plurality of (M) listening positions are arranged in the basic model. For each listening position, the state in which the sound emitted from each of the K sound sources is reflected between each of the L wall surfaces and the state in which the reflected sound is detected from the listening position including the direction localization of each wall surface is simulated. By reproducing them separately, a high-quality stereoscopic reverberation sound is generated.

The room having L wall surfaces is virtual and does not have to have the shape of an L-face body that actually exists. The playback means may be headphones (two outputs),
There may be two or more speakers (the number of outputs is N).
The same reproducing means need not be used for each listening position, and the value of N may be different.

By arranging a plurality of listening positions, it is possible to simultaneously detect reflected sounds in the same space at different listening positions. Reverberation with a high quality of reality must be able to recognize the sense of spatial spread, the sense of the sound source, the position and distance of the reflecting wall, and the presence of space (air). In order to give the reverberant sound a sense of three-dimensional expanse, the reflected sound from each wall direction is virtually localized in the direction of the wall by sound image localization means that localizes the direction of the sound image viewed from the listening position. It can be carried out. This sound image localization means is a device for adding localization in the direction of the sound image by arithmetic processing.

In order to more clearly recognize the position (distance) and presence of each wall and sound source, the distance between the sound source and each wall, the distance between the walls, and the distance between each wall and the listening position are defined. It is necessary to generate a reflected sound and a direct sound with a delay time according to the distance between the sound source and the distance between the sound source and the listening position.
It becomes possible to recognize the position of the wall by the relationship between the reflected sounds and the position of the sound source by the relationship between the reflected sound and the direct sound.

Further, in order to be able to recognize the existence of space (air), it is necessary to create reflections of complicated paths in various directions so as to completely fill the interior of the room with respect to reflections between the walls. To obtain each delay time, the coordinates of the center position of each wall, the position of the sound source, and the listening position (x, y,
z) is decided. As a result, the distance r between them is obtained. From this, t = r / v (v: sound velocity) is obtained.

The conventional sound image localization device localizes only direct sound,
In other words, it was localization only in the direction of the sound image, but if combined with such a reverberation sound generator, sounds in all directions can be heard at the same time, so there is less discomfort than in the case of stereoscopic localization with only direct sound,
A natural stereoscopic space and sound image localization can be obtained. FIG. 1 is a block diagram of a sound image localization apparatus including the reverberation sound generation apparatus of one embodiment of the present invention, which is configured according to the above concept.

In this sound image localization apparatus, all processing is performed on a signal. However, for convenience of explanation, here, for example, a signal representing a sound is simply expressed as "sound", which is a term representing a physical phenomenon that the signal means. It may be replaced with and explained.
Sounds radiated from the K sound sources 11_1, ..., 11_K are input to the direct sound processing means 13 and the reverberation sound processing means 15. The direct sound processing means 13 includes the sound sources 11_1, ..., 1
It calculates the attenuation and delay of each sound emitted from 1_K until it is directly transmitted to each of the M listening positions. The reverberation sound generation means 15 corresponds to an embodiment of the reverberation sound generation device of the present invention, and each sound source 11_
The attenuation and delay of the sound emitted from 1, ..., 11_K until it is transmitted to each listening position after being reflected once by each of the L reflection walls or multiple reflections between the reflection walls. It is something that is calculated.

The sound attenuated and delayed by the direct sound processing means 13 and the reverberation sound generating means 15 is input to the sound image localization means 17. In the sound image localization means 17, the sound entering the right ear and the sound entering the left ear of the human being located at each of the M listening positions have different phases, and have different frequency characteristics of the sound due to the influence of the head or auricle. , 11_K are localized for the direct sound, and the reverberant sound is localized for the reverberant sound in the direction of each of the L reflection walls viewed from each listening position. Let As a result, a signal to be sent to each of the reproduction means 19_1, ..., 19_M having N speakers is generated for each listening position.

FIG. 2 is a block diagram of the sound image localization means 17 shown in FIG. Here, the part of the listening position 1 is shown. For the listening position 1 output from the direct sound processing means 13 and the reverberation generating means 15 shown in FIG. 1, respectively.
The number of sounds corresponding to the number K of the sound sources 11_1, ..., 11_K and the number of sounds corresponding to the number L of the reflection walls are respectively set in the sound image localization blocks 170_1, ..., 170_K, 170_K + 1,
..., 170_K + L, respectively. Each sound image localization block 170_1, ..., 170_K, 170_K +
1, ..., 170_K + L, the sound sources 1 to K and the reflection walls 1 to 1 for the listener located at the listening position 1 when sound is emitted from the N speakers arranged at the predetermined positions, respectively.
Sound localization was performed so that sound could be heard from the direction of L.
A number of signals corresponding to the number N of speakers are generated and output. Each sound image localization block 170_1, ..., 170_
The signals output from K, 170_K + 1, ..., 170_K + L are input to and added by each adder 171_1, ..., 171_N for each signal sent to each of the N speakers, whereby the listening position 1 An output signal is produced.

If any of the sound sources 1 to K moves in the left-right and up-down directions when viewed from the listening position 1, it is necessary to change the direction of the sound image localization for the sound source, and therefore the corresponding sound image localization block is set. The set filter coefficient and the like are changed. However, since it is assumed here that the reflection wall does not move even if the sound source moves, the sound image localization direction of the reverberation sound does not move due to the movement of the sound source. On the other hand, when the listening position 1 is moved with time, considering the listening position 1 as the center, both the position of the sound source and the position of the reflection wall move relatively. Therefore, the direct sound reverberates according to the movement. The sound image localization direction of the sound is also changed.

The sound image localization blocks 170_1, ..., 1
The configuration of 70_K, 170_K + 1, ..., 170_K + L is known and is not directly related to the present invention, and thus detailed description of the sound image localization block is omitted. FIG. 3 is a block diagram of the direct sound processing means 13 shown in FIG. Sounds radiated from the sound sources 1 to K are input to the delay lines 130_1, ..., 130_K, and the sound sources 11_
, ..., 11_K (see FIG. 1) and each listening position are delayed by an amount corresponding to each distance, and then each amplifier 131_
1, ..., 131_K, ..., 131_K (M-1) +1,
…, 131_K · M, and this time, each sound source 11
Since _1, ..., 11_K are assumed to be point sound sources, attenuation corresponding to each distance between each of the sound sources 11_1, ..., 11_K and each listening position is calculated, whereby each listening position 1
A direct sound for each of ~ M is generated.

Next, referring to FIG. 1, which corresponds to an embodiment of the present invention.
The reverberation sound generating means 15 shown in FIG. 1 will be described. First, the concept for constructing the reverberation sound generating means in this embodiment will be described. The position of the wall surface of the reflecting wall is represented by the center position of the wall surface, and the reflection method between the wall surfaces is considered. Assuming regular reflection as shown in Fig. 4, the wall surface is flat, and when considering a reflection wall arbitrarily arranged, the sound emitted from the sound source or the sound reflected by another reflection wall is incident on the reflection wall. Although the angle and the angle of the sound reflected from the reflection wall toward the other reflection wall or the listening position do not necessarily match, it is assumed that those angles approximate to the closest direction. Since the number of walls is limited, if only one reflected sound can be made for one incident sound, only a simple path can be reflected. If the reflected sound is distributed not only in regular reflection but also in other directions as shown in FIG. 5, one reflection is distributed over a plurality of wall surfaces, and the reflection is distributed over a plurality of wall surfaces. Although the reflection of the path is possible, if the walls in the incident direction are different, it is necessary to make a plurality of reflections of different levels separately.
A complicated structure such as (only the reflection wall 1 is described) will be formed. Incidentally, FIG. 6 and FIG.
Indicates only the signal flow on the wall surface, and the illustration of the delay is omitted. Further, in each of these figures, a triangle in the signal path represents an amplifier (including an attenuator), and a symbol having an addition sign inside a white circle represents an adder.

However, as shown by the broken line in FIG. 7, assuming that the wall surface has a shape of a perfect diffusing surface in which the intensity of the reflected sound is determined only by the angle from the normal of the reflecting wall surface regardless of the incident angle. , All the incident sounds can be added and then distributed to the reflected sounds in each direction, so that the reflection of a complicated path can be made with a simple configuration as shown in FIG. 8 (only the wall 1 is described). Like

Since the wall surface has a complicated shape to have such a property and it is considered that reflection occurs even in one wall surface, reflection in the same wall surface is also added to FIG. By adding the reflection on the same wall, it is possible to increase the feeling of space expanse and the presence of walls even if the number of walls is small. The delay time of each reflection path is obtained by using the wall as a point, but since the reflection wall is actually a surface, a large number of reflections with different delay times are overlapped. To simulate it, insert an all-pass filter in the reflection path.
The all-pass filter causes various time differences. As a result, although the direction localization by the sound image localization means is a point, it becomes possible to feel the localization also in the space between the points, and as a result, it can be recognized as a wall surface.

In the reflection path, the reflection characteristics of the wall are as follows:
A dump filter for simulating the frequency characteristics of reflection (a simple filter such as a high-pass filter or low-shelving filter, a low-pass filter or a high-shelving filter, or a measurement of the impulse response of an actual reflected sound from a wall) May be used). Figure 9
FIG. 2 is a block diagram showing an algorithm of the reverberation sound generating means 15 shown in FIG. 1, which is constructed based on the above idea. Here, as an example, the number of sound sources K = 2, the number of reflection walls L = 4, and the number of listening positions M = 2.

The sounds radiated from the sound sources 11_1 and 11_2 are input to the reverberation generating means 15, and first, the delay line 15
0_1, 150_2 causes a delay according to the distance between each sound source 11_1, 11_2 and each wall 1-4, and each sound source 1
Since 1_1 and 11_2 are assumed to be point sound sources, each amplifier 151_1_1, 151_1_2; 151_2_1,
151_2_2; ...; 151_4_1, 151_2-4_2
Is attenuated according to each distance between each sound source 11_1, 11_2 and each wall 1-4, and is incident on each wall 1-4.
The delay lines 150_1 and 150_2 for calculating the delays from the sound sources 11_1 and 11_2 to the walls 1 to 4 are delay lines 130_1 and 130_1 in the direct sound processing unit 13 shown in FIG.
..., 130_K may also be used.

The sounds reflected by the four walls 1 to 4 including themselves are also incident on each of the walls 1 to 4. Each of these walls 1
The sound incident on 4 is added to each adder 152_1, ..., 152_4
To the respective walls 1 to 4 and added regardless of the incident direction of the incident sound. Each adder 152_1, ..., 152_4
, The total sound of each of the walls 1 to 4 is passed through each of the dump filters 153_1, ..., 153_4, whereby the total sound is changed by the frequency characteristic of the reflection of the wall, and each of the all-pass filters 154_1 ,. Via 154_4, various time differences are generated thereby, so that the delay lines 155_1, ..., 155_4 are provided with a spread as walls (planes) instead of points. In each of the delay lines 155_1, ..., 155_4, a delay of sound transmission between one wall and another wall, a delay of sound transmission due to multiple reflection in one wall, and each wall 1 to 4 and each listening Subjected to the delay of sound transmission between positions, each amplifier 156_1_1,
..., 156_1_6; ..., 156_4_1; ..., 156
It is input to _4_6 to be attenuated (see FIG. 7) according to the direction of each of the walls 1 to 4 and the reflection direction, and re-enters each of the walls 1 to 4 or is transmitted to each listening position.

Here, each amplifier 15 for calculating the attenuation coefficient
1_1_1, 151_1_2; ...; 151_4_1, 1
51_4_2, each dump filter 153_1, ..., 15
3_4, all-pass filters 154_1, ..., 154
_4, the connection order of the delay lines 155_1, ..., 155_4 for calculating the delay due to the distance between the wall and the listening position may be different. However, the delay lines 155_1, ..., 155
The effect changes depending on whether another element is connected to _4 before or after. Also, a plurality of different walls may be arranged in the same direction. Further, if the delay time of each part is equivalent, it can be distributed to other delay parts.

Since the sound source is assumed to be a point sound source, distance attenuation occurs. In the path from the sound source to the listening position (direct sound) and the path from the sound source to each wall, when the sound source moves, the attenuation coefficient of each changes with a value proportional to the reciprocal 1 / r of the distance r of the path. The path from the wall can be a surface sound source, so no distance attenuation occurs. However, there is attenuation due to the wall direction (reflection direction).

In order to express the movement of the position of the sound source by the reflected sound, the delay line 150_, which corresponds to the distance from the sound source to each wall,
1,150_2 each delay time and the coefficient which gives the input distance attenuation (amplifiers 151_1_1, 151_1_2;
... 151_4_1, 151_4_2 only), and other parameters may be fixed. For the direct sound, delay, attenuation of distance, and localization of direction are controlled.

The expression of the movement of the listening position by the reflected sound is as follows:
Each delay line 1 corresponding to each distance from each wall to its listening position
Amplifiers 156_1_1, 55_1, ..., 155_4 due to respective delay times and changes in the reflection direction of the sound reflected from the wall
156_1_2; ... Controls only the amplification factors of 156_4_1 and 156_4_2, and other parameters may be fixed.
In addition, when the listening position moves, the wall moves relative to the listening position, so that the sound image localization means 17 (see FIG. 1) also controls the direction localization of the reverberant sound. For direct sound, delay, distance attenuation, and direction localization are controlled.

In order to move the sound source position and the listening position, it is preferable to change the above delay time while interpolating so that noise is not generated. However, with simple things such as linear interpolation, the sound quality when in the interpolation state Deteriorates. In order to avoid this, it is preferable that interpolation is not performed when the sound source is in a stationary state, and interpolation is performed only when the sound source is in a moving state. Interpolation of the delay time in the moving state not only prevents noise but also produces a Doppler effect on the reflected sound, resulting in a more realistic reverberant sound.

The values of the parameters of the attenuation coefficient in the reflection path between the walls, the dump filter, the all-pass filter, and the delay due to the distance between the walls are preferably different values for each wall. When the same value is present, it is difficult to recognize as a different sound in each direction, and the reverberant sound spreads less.
The reflected sound on the perfect diffusion surface has a directionality as shown in FIG. Since this direction changes depending on the angle of the wall surface, by adjusting the angle of the wall surface, that is, the direction of the reflection distribution, it is possible to adjust the degree of reverberation and the degree of spread.

In FIG. 9, the frequency characteristics of the reflected sound in each direction are the same for one wall, but they may be different. FIG. 10 is a block diagram showing an algorithm of the wall 1 configured so that the frequency characteristic of reflected sound can be adjusted for each reflection direction. The reference numerals are the same as those in FIG.

The dump filter 153_1_1, ..., 15
3_1_6 is arranged after the delay line 155_1.
In FIG. 11, each dump filter 153_1_1, ..., 1
53_1_6 is arranged immediately before being input to the adder 152_1. If the frequency characteristics of the reflected sound to each listening position are the same, FIG. 10 is equivalent even if it is modified as shown in FIG.

If the dump filter has a direct structure as shown in FIG. 12, and the frequency characteristics of each reflected sound have the same pole positions and different zero points, the pole portions can be made common as shown in FIG. Can be easy. Incidentally, FIG.
2. In FIG. 13, square blocks without reference numerals represent unit delays. FIG. 13 shows an example in which the dump filter is of the first order, but it may be configured by a second or higher order filter.

If the sound emitted from the sound source has directionality, a more realistic simulation can be performed. Not only a single direction but also a plurality of directions may be used, and each direction may be arbitrarily weighted. Of course, sound may be emitted evenly in all directions. In the case of emitting sound in all directions in FIG. 14, in the case of a stereo speaker in FIG. 15 (an example in which the sound is fixed in the listening position direction), in the case of a rotating speaker in FIG. 16 (the sound source moves while the sound emitting direction changes). Example)

By arranging the respective walls so that the following conditions are satisfied, an efficient and good spread feeling can be obtained even with a small number of walls. Here, it is assumed that only the direction of the wall around the listening position is considered and the distance is not considered. (1) The relations from the center to each other are arranged so that they are equal on all walls (make the sense of all directions equal). (2) Do not place other walls on a straight line that passes through the center and one wall (It has been experimentally confirmed that it is better if the reflected sound does not pass through the center in order to create a feeling of spaciousness. (3) The walls on the leftmost and rightmost sides should be left-right symmetric (because it feels most sensitive in the left-right direction, the sense of breadth on the left and right should be the same). (4) Even if all walls are mapped horizontally, they do not overlap and are evenly arranged (the left and right directions can be clearly recognized, but the up and down directions cannot be clearly recognized). . (5) Even if all walls are mapped horizontally and the rear ones are mapped forward, they do not overlap each other and are evenly arranged (the front and rear directions cannot be clearly recognized).

When used for music production, the following consideration is preferable. (1) When it is not desired to change the pitch even if the sound source and the listening position move and the delay time changes, the crossfade is used instead of interpolation. (2) When it is desired not to cause a rhythm shift, the delay time of the direct sound is set to 0 and the delay time is subtracted from the entire reverberation generating means. At this time, the change amount of the delay time of the reverberation sound due to the movement of the sound source and the listening position is different from the original change amount, and the pitch change is also different. Therefore, in this case as well, it is preferable to connect by the crossfade. This makes it possible to prevent the harmony and rhythm from being deteriorated while leaving the acoustic image of the room and the acoustic image of the position of the sound source, which is partially different from the actual simulation.

The above is an example of simulating a realistic space, but even if the same algorithm is used, it is possible to create a special space if there is a bias such as partial reflection between walls.

[0038]

As described above, according to the present invention,
Since the sound incident on the virtual reflection wall is added regardless of the incident direction, and the power of the reflected sound is distributed according to the direction and the reflection direction of the virtual reflection wall, the movement of the sound source,
It is only necessary to adjust some parameters when moving the listening position, and reverberation can be made to follow those movements in real time.

[Brief description of drawings]

FIG. 1 is a block diagram of a sound image localization device including a reverberation sound generation device according to an embodiment of the present invention.

FIG. 2 is a block diagram of a sound image localization means.

FIG. 3 is a block diagram of direct sound processing means.

FIG. 4 is a diagram showing regular reflection of sound.

FIG. 5 is a diagram showing sound dispersion.

FIG. 6 is a diagram showing an example of an algorithm simulating sound dispersion.

FIG. 7 is a diagram showing complete diffusion of sound.

FIG. 8 is a diagram showing an example of an algorithm simulating complete diffusion.

FIG. 9 is a block diagram showing an algorithm of a reverberation sound generating means.

FIG. 10 is a block diagram showing an algorithm of wall 1.

FIG. 11 is a block diagram showing an algorithm of wall 1.

FIG. 12 is a block diagram showing a configuration example of a dump filter.

FIG. 13 is a block diagram showing an algorithm of wall 1.

FIG. 14 is a diagram showing an example of sound emission from a sound source.

FIG. 15 is a diagram showing an example of sound emission from a sound source.

FIG. 16 is a diagram showing an example of sound emission from a sound source.

[Explanation of symbols]

11_1, ..., 11_K Sound source 13 Direct sound processing means 15 Reverberation generating means 17 Sound image localization means 19_1, ..., 19_M Reproducing means 150_1, 150_2, 155_1, ..., 155_4
Delay line 151_1_1, 151_1_2; ...; 151_4_
1, 151_4_2, 156_1_1, ..., 156_1
_6; ...; 156_4_1, ..., 156_4_6 amplifier 152_1, ..., 152_4 adder 153_1, ..., 153_4 dump filter 154_1, ..., 154_4 all-pass filter

Claims (1)

[Claims] 1. A first calculation means for transmitting sounds radiated from one or a plurality of sound sources to a plurality of virtual reflection walls, adding the incident sounds incident on the virtual reflection walls to each other, and the virtual reflection walls. From the virtual reflection wall, the second calculation means for transmitting the reflection sound reflected by the reflection sound to the reflection sound by weighting the reflection sound according to the reflection direction, and transmitting the reflection sound to at least one virtual reflection wall among the plurality of virtual reflection walls. The reflected sound reflected
A reverberation sound generation apparatus comprising: a third calculation means for weighting the reflected sound in the reflection direction and transmitting the weighted sound to one or a plurality of listening positions.