JP2006210986A - Sound field design method and sound field composite apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve accuracy in the design and control of a sound field and easily and surely obtain a desired sound field in the case that sound fields by a plurality of speakers are composed and a sound is heard in a way as if to be jumped out by approaching a virtual point sound source to a listening region. <P>SOLUTION: A control line S4 is set to a position close to speakers SP1, SP2, ..., SPm and in front of the speakers SP1, SP2, ..., SPm, a plurality of control points are set on the control line S4, the virtual point sound source 10 is set to a point opposite to the speakers SP1, SP2, ..., SPm with respect to the control line S4, a line including the point of the virtual point sound source 10 and in parallel with the control line S4 is selected for a listening side border S5, and a region at the side of a listener 7 within the listening side border S5 is selected for a region wherein sound fields are composed. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a method for designing a sound field when a sound field is synthesized by a plurality of speakers, and an apparatus for reproducing sound by synthesizing a sound field by a plurality of speakers.

  Apart from multi-channel reproduction such as stereo reproduction, it is considered to reproduce sound by synthesizing a sound field with a plurality of speakers.

  For example, in Patent Document 1 (Japanese Patent Application Laid-Open No. 2004-172786), when a sound field is formed by a speaker array, a plurality of audio signals are supplied to expand a range in which an appropriate sound image localization can be obtained. The output of the digital filter is supplied to a plurality of speakers constituting the speaker array, and a predetermined delay time is set for each digital filter to form a sound field in the closed space and output from the plurality of speakers. It is shown that the reflected sound is reflected by the wall surface of the closed space and then focused on the position of the listener (listener) in the sound field.

  As a method for controlling a sound field in a three-dimensional space, for example, Non-Patent Document 1 (Waseda University, Research Center for Science and Engineering, Acoustic Information Processing Laboratory, Yoshio Yamazaki, “Study on three-dimensional virtual reality based on Kirchoff integral equation”) ), There is a method using Kirchhoff's integral formula as follows.

  That is, assuming a closed curved surface S that does not include a sound source, as shown in FIG. 8, the sound field in the closed curved surface S can be expressed by Kirchhoff's integral formula. In FIG. 8, p (ri) is the sound pressure at the point ri in the closed surface S, p (rj) is the sound pressure at the point rj on the closed surface S, n is the normal at the point rj, and un (rj) is the modulus. The particle velocity in the direction of line n, | ri−rj |, is the distance between point ri and point rj.

  Kirchoff's integral formula is expressed by equation (1) in FIG. 9, and if the sound pressure p (rj) on the closed surface S and the particle velocity un (rj) in the direction of the normal n can be completely controlled, the closed surface S It means that the sound field inside can be completely reproduced.

  In the equation (1), ω is an angular frequency represented by ω = 2πf, where f is an audio frequency, ρ is an air density, and Gij is represented by the equation (2) in FIG. It is what is done.

  Equation (1) is for a steady sound field, but the same can be said for a transient sound field by controlling the instantaneous values of sound pressure p (rj) and particle velocity un (rj).

  Thus, in the sound field design by Kirchhoff's integral formula, it is only necessary to reproduce the sound pressure p (rj) and the particle velocity un (rj) on the virtual closed curved surface S. Since it is impossible to control the sound pressure p (rj) and the particle velocity un (rj) at the continuous points, the sound pressure p (rj) and the particle velocity un ( On the premise that rj) is constant, the closed curved surface S is discretized.

  When the closed surface S is discretized at N points, equation (1) in FIG. 9 is expressed by equation (3) in FIG. 9, and the sound pressure p (rj) at N points on the closed surface S and By reproducing the particle velocity un (rj), the sound field in the closed curved surface S can be completely reproduced.

  As a system for reproducing the sound pressure p (rj) and the particle velocity un (rj) at the N points with M sound sources, a system as shown in FIG. 10 is considered.

  In this system, audio signals from the signal source 1 are supplied to the speakers 3 through the filters 2, respectively, and the sound pressure is measured at N points on the boundary surface of the control region 4. The particle velocity un (rj) in the normal direction is approximately obtained from the sound pressure signal by the two-microphone method.

  At this time, in order to reproduce the sound pressure p (rj) and the particle velocity un (rj) at the N point, the sound pressure at the 2N point may be equal to the original sound field. This results in the problem of obtaining a value such that the sound pressure at point 2N is closest to the original sound field as the transfer function Hi (i = 1 to M) of the filter 2.

  Therefore, Cij is a transfer function between the sound source i (i = 1 to M) and the sound receiving point j (j = 1 to 2N) in the reproduction sound field, and Hi is a transfer function of the filter in the previous stage of the sound source i. An evaluation function for minimizing the difference between the reproduced sound field and the original sound field, as expressed by equation (4) in FIG. 9, where Pj is the transfer function between the sound source and the receiving point j in the original sound field. Think of J.

  In order to obtain the transfer function Hi that minimizes the evaluation function J expressed by the equation (4), the equation (5) in FIG. 9 may be solved.

  Furthermore, as an extension of Kirchhoff's integral formula to a half space, as shown in FIG. 11, the sound source 5 is arranged in a space on one side (left side in the figure) of the boundary surface S1, and in the space on the opposite side (right side in the figure). Assuming a listening area 6 that does not include a sound source, if the sound pressure and particle velocity at all points on the boundary surface S1 or discrete points as described above are controlled by Kirchhoff's integral formula, A desired sound field can be realized in the listening area 6 not included.

  Specifically, as shown in FIG. 12, a plurality of speakers SP1, SP2,... SPm are arranged on the left side (one side) of a certain finite length control line (boundary line) S2, and a plurality of controls are arranged on the control line S2. By setting the points C1, C2... Ck and controlling the sound pressure (amplitude) and phase at each control point C1, C2... Ck, the right side of the control line S2 (speakers SP1, SP2... SPm side) In the listening area on the opposite side, the listener 7 can hear the sound from the speakers SP1, SP2,... SPm as the sound from the virtual point sound source 8 on the left side of the control line S2 (speakers SP1, SP2,... SPm side). Like that.

The prior art documents listed above are as follows.
JP 2004-172786 A Research Center for Science and Engineering, Waseda University, Acoustic Information Processing Laboratory, Yoshio Yamazaki, “Research on 3D Virtual Reality Based on Kirchhoff Integral Equation”, [online], April 1997, [August 27, 2004] ] <URL: http: www.acoust.rise.waseda.ac.jp/publications/happyou/1997-h9.html>

  However, the virtual point sound source is not located on the left side of the control line S2 (speakers SP1, SP2,... SPm side) like the virtual point sound source 8 in FIG. As shown in the diagram, the right side of the control line S2 (on the side opposite to the speakers SP1, SP2... SPm side) is set near the listener 7 so that the sound can be heard as if it jumps out from the vicinity of the listener 7. There is a case.

  However, in that case, the sound field is included in the listening area, and the sound field cannot be expressed by Kirchhoff's integral formula.

  In this case, considering the sounds Ab and Af from the virtual point sound source 9, as shown in FIG. 14, the sound Ab propagating from the virtual point sound source 9 to the control line S2 side is the sound from the speakers SP1, SP2... SPm. Since the propagation direction is different from As, it cannot be expressed and is not a necessary sound. On the other hand, the sound Af propagating from the virtual point sound source 9 to the listener 7 as shown in FIG. 15 is a necessary sound.

  However, as shown in FIG. 15, when the control line S2 is located on the left side of the virtual point sound source 9 (speakers SP1, SP2,... SPm side) and the sound source is included in the listening area, the sound is expressed by the Kirchhoff integration formula as described above. The place cannot be expressed.

  Therefore, as shown in FIG. 16, the control lines S3 for setting the control points C1, C2,... Ck are arranged at positions far away from the speakers SP1, SP2,. It is conceivable that the virtual point sound source 9 is set on the left side of the control line S3 (speakers SP1, SP2,... SPm side) and the right side (listener 7 side) of the control line S3 is set as the listening area.

  According to this, since the virtual point sound source 9 is located close to the listener 7, the sound can be heard as if it jumps out, and since the sound source is not included in the listening area, the sound field is expressed by Kirchhoff's integral formula. It becomes possible.

  However, when the control line S3 is set at a position far away from the speakers SP1, SP2,... SPm as shown in FIG. 16, the control line S2 is set at a position close to the speakers SP1, SP2. Compared with the case of setting, the accuracy of the design and control of the sound field deteriorates, and it becomes difficult to easily and surely realize the desired sound field.

  That is, when the control line S2 is set at a position close to the speakers SP1, SP2,... SPm as shown in FIG. 12, the sound waves reaching the control point C1 from the speakers SP1, SP2. As described above, since there is a large sound pressure difference and phase difference between sound waves that reach the same control point on the control line S2 from each speaker SP1, SP2,... SPm, the accuracy of sound field design and control is improved. A desired sound field can be easily and reliably realized.

  On the other hand, when the control line S3 is set at a position far away from the speakers SP1, SP2,... SPm as shown in FIG. 16, the control point C1 is reached from each speaker SP1, SP2. As indicated by arrows, the sound pressure difference and the phase difference between the sound waves that reach the same control point on the control line S3 from the speakers SP1, SP2,. It becomes difficult to realize a desired sound field easily and reliably.

  Therefore, the present invention improves the accuracy of the sound field design and control, and easily and surely realizes the desired sound field when the virtual point sound source is brought close to the listening area so that sound can be heard as if it pops out. It is something that can be done.

The sound field design method of the present invention is:
When synthesizing a sound field with a plurality of speakers, a control line is set at the front position of each speaker, a plurality of control points are set on the control line, and the control line is on the opposite side of the speaker side. A virtual point sound source is set, and a sound field is designed by setting a region opposite to the control line side of the virtual point sound source as a region for synthesizing a sound field.

  In the sound field design method described above, the virtual point sound source is set not on the speaker side but on the listener side on the opposite side of the control line where each control point is set, and is positioned close to the listener. A sound field that sounds like a popping out can be formed, and the control line where each control point is set is set not on the listener side but on the speaker side opposite to the virtual point sound source. Therefore, the accuracy of sound field design and control is improved, and a desired sound field can be easily and reliably realized.

  As described above, according to the present invention, when a virtual point sound source is brought close to the listening area to make it sound like a sound jumps out, the design and control accuracy of the sound field is improved, and a desired sound field is obtained. It can be easily and reliably realized.

  FIG. 1 shows an example of a sound field synthesizing apparatus according to the present invention, which reproduces sound by forming a sound field designed by a sound field designing method described later.

  The digital audio signal source 100 reads uncompressed PCM (Pulse Code Modulation) audio data from a recording medium such as a compact disk or a hard disk, or ATRAC (registered trademark, abbreviation for Adaptive Transform Acoustic Coding) or MP3. Audio data compressed by (MPEG-1 Audio Layer-3) or the like is read and expanded, or uncompressed PCM audio data is received by a broadcast signal or a network, or ATRAC (registered trademark), MP3, etc. The voice data compressed by the above is received and decompressed.

  The audio data from the digital audio signal source 100 is supplied to a plurality of digital filters 201, 202,. The transfer functions of the digital filters 201, 202,..., 20m are calculated and determined by a sound field design method described later.

  The audio data output from the digital filters 201, 202... 20 m is converted into analog audio signals by the DA converters 301, 302... 30 m, and the converted audio signals are output by the audio amplifier circuits 401, 402. Amplified and supplied to speakers SP1, SP2,. The speakers SP1, SP2,... SPm are configured as a speaker array.

  Although not shown in the figure, the arithmetic processing unit 500 is configured as a computer unit including a CPU, a ROM, and a RAM, and sets filter coefficients of the digital filters 201, 202,.

  Furthermore, in this example, the arithmetic processing unit 500 is also used for sound field design by a sound field design method described later. For this purpose, the storage unit 600 and the key input unit 700 are connected to the arithmetic processing unit 500. The storage unit 600 stores the filter coefficients of the digital filters 201, 202,..., 20m as data of the sound field design result, and the key input unit 700 inputs the sound field design conditions. To perform necessary operations.

  In the sound field designing method according to the present invention, as shown in FIG. 1, the control line S4 is set at a position close to the speakers SP1, SP2... SPm in front (front) of the speakers SP1, SP2. A plurality of control points are set on S4, and a virtual point sound source 10 is set on the opposite side (right side in the figure) of the control line S4 to the speakers SP1, SP2,. Including a line parallel to the control line S4 as a listening side boundary S5, a region on the listener 7 side (right side in the figure) of the listening side boundary S5 is a region for synthesizing a sound field.

  That is, by setting the virtual point sound source 10 between the control line S4 near the speakers SP1, SP2... SPm and the listener 7, as shown in FIG. 2, the right side of the control line S4 (listener 7 side) The region Ec can be a region where the sound field is synthesized.

  However, in this case, as shown in FIG. 3, a region parallel to the control line S4 including the virtual point sound source point x0 is defined as a listening side boundary S5, and an area Ez on the right side (listener 7 side) of the listening side boundary S5. Is the region where the actual sound field is synthesized.

  The synthesized sound field of FIG. 2 and the synthesized sound field of FIG. 3 seem to be the same from the listener 7, but by setting as shown in FIG. 3, the conditions of the Kirchhoff integration formula are satisfied, and the control line is determined by the Kirchhoff integration formula. By controlling the sound pressure and phase at each of the control points x1, x2, x3,... On S4, the sound field generated by the virtual point sound source at the point x0 on the right side (listener 7 side) of the control line S4 is changed to the point x0. It can be realized on the right side (listener 7 side) of the listening-side boundary S5 including.

  In this case, the sound propagating from the virtual point sound source point x0 to the right side (listener 7 side) as indicated by the arrow in FIG. 3 can be described on the left side (speaker side) of the virtual point sound source point x0.

  Specifically, the sound at the point x0 is the sound pressure (amplitude) and phase at each of the control points x1, x2, x3,... On the control line S4, for example, as shown in FIG. The signal S (x2) at the control point x2 is larger than the signals S (x1) and S (x3) at the control points x1 and x3, and the phases of the signals S (x1) and S (at the control points x1 and x3 are as follows. x3) is expressed as being advanced from the signal S (x2) at the control point x2.

  When considered on the frequency axis, in the case of an ideal point sound source, the sound pressure (amplitude) and phase at each control point are the distance r from the virtual point sound source to the control point and the frequency f of the audio signal. As a function of the above, if the virtual point sound source 10 is on the right side (listener side) of the control line S4 as described above, R ( The complex conjugate of r, f) can be taken and the time axis can be reversed. Alternatively, the positive and negative of the time axis may be reversed directly on the time axis to reverse the past and the future.

  Computer simulation of sound field design (sound field formation) by the above sound field design method was performed. FIG. 6 shows the model.

  In this simulation model, the speaker SP is 0.1 m (10 cm) centered on the position of the origin (0, 0) as indicated by a white circle on the x axis of the xy plane (the numerical value of the xy coordinates is meters). ) With a spacing of 14) in the left and right, a total of 28, and the control line S4 at the position of y = −0.1 m before the speaker SP (listener side) is 2.7 m, which is the same as the arrangement of the speakers SP. A large number of control points were set on the control line S4. The virtual point sound source 10 was set at a position where x = 0 m and y = −1 m. The audio signal was a 1 kHz sine wave.

  FIG. 7 shows the simulation results of the model as described above, and shows the ripples of the sound waves in the area surrounded by the oblique lines in FIG. As is apparent from the above, a desired sound field can be formed in a region in front of the virtual point sound source 10 (listener side) by the sound field design method described above.

  1, the control line S4 is set to a position close to the speakers SP1, SP2,... SPm as shown in FIGS. A filter coefficient (transfer function) that forms a sound field when the point sound source 10 is set at a position close to the listener 7, and, as shown in FIG. 12, the control line S2 is set at a position close to the speakers SP1, SP2,. A filter coefficient (transfer function) that forms a sound field when the virtual point sound source 8 is set to the speakers SP1, SP2,... SPm side is prepared and written in the storage device unit 600. By selecting one of the modes by operating the key input unit 700, the filter coefficient (transfer function) corresponding to the selected mode is changed to the digital filter 201. 202 may be configured to be loaded into ‥‥ 20 m.

  According to this, according to circumstances, the user selects either the mode in which the sound can be heard as if it jumps out from the vicinity of the user or the normal mode in which the sound is heard as a sound from a point in the vicinity of the speaker. Can play and listen.

It is a figure which shows an example of the sound case synthesis apparatus of this invention. It is a figure where it uses for description of the sound field design method of this invention. It is a figure where it uses for description of the sound field design method of this invention. It is a figure where it uses for description of the sound field design method of this invention. It is a figure where it uses for description of the sound field design method of this invention. It is a figure which shows the model of simulation. It is a figure which shows the sound pattern of a simulation result. It is a figure which shows the sound field in the virtual closed curved surface which does not contain a sound source. It is a figure which shows the integration formula of Kirchhoff. It is a figure which shows the system which reproduces the sound pressure and particle velocity of N point with M sound sources. It is a figure which shows the principle of expansion to the half space of Kirchhoff's integral formula. It is a figure which shows the specific example of the expansion to the half space of the integration formula of Kirchhoff. It is a figure which shows the case where a virtual point sound source is set on the opposite side to the speaker side of a control line. It is a figure which shows the sound which propagates from the virtual point sound source to the control line side. It is a figure which shows the sound which propagates from the virtual point sound source to the listener side. It is a figure which shows the case where a control point is set in the position close | similar to a listener. It is a figure which shows the sound wave which reaches a control point from each speaker in the case of setting a control point in the position near a speaker. It is a figure which shows the sound wave which reaches a control point from each speaker in the case of setting a control point in the position largely away from the speaker.

Explanation of symbols

  Since all the main parts are described in the figure, they are omitted here.

Claims (4)

  When synthesizing a sound field with a plurality of speakers, a control line is set at the front position of each speaker, a plurality of control points are set on the control line, and the control line is opposite to the speaker side. A sound field design method, wherein a virtual point sound source is set, and a sound field is designed using a region opposite to the control line side of the virtual point sound source as a region for synthesizing a sound field. The sound field design method according to claim 1,
A sound field design method characterized by taking a complex conjugate on the frequency axis for a transfer function from the virtual point sound source to each control point. The sound field design method according to claim 1,
A sound field design method for reversing the positive and negative of the time axis on the time axis for the transfer function from the virtual point sound source to each control point.   A plurality of speakers and a plurality of filters provided at the front stage of each speaker to which the same audio signal is supplied, and a transfer function of each filter is designed by the sound field design method according to claim 1. A sound case generator characterized by having a transfer function that realizes a sound field.

Images