JP2018137531A - Gain setting device, loudspeaker system, gain setting method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a gain setting device, gain setting method and program, performing gain setting for transmitting audio into a room with a sound volume as uniform as possible, while preventing howling.SOLUTION: A gain setting device includes a transfer function acquisition unit, a gain calculation unit, and a gain setting unit. The transfer function acquisition unit acquires a first transfer function of a first system from a speaker position to a microphone, a second transfer function of a second system from the speaker position to multiple hearing positions, a third transfer function of a third system from multiple speakers to the multiple hearing positions, and a fourth transfer function of a fourth system from multiple speakers to the microphone. By using the first through fourth transfer functions obtained in the transfer function acquisition unit, energy above a certain level is obtained at all hearing positions, and the gain of a fifth system from the microphone to the multiple speakers is calculated so that the loop gain in a feedback system from the multiple speakers to the microphone will be less than 1, and is set in each of the five systems.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a gain setting device, a gain setting method, and a program for setting a system gain from a microphone to a speaker.

  Patent Document 1 discloses a temple acoustic system that adjusts the volume of a speaker so that the sound of the speaker reaches all attendees at the same volume in the main hall of the temple. In the temple acoustic system of Patent Document 1, a microphone is installed on the ceiling near the speaker position, and a plurality of speakers are installed in a matrix on the ceiling at the listening position. And the sound system for temples of patent documents 1 sets up the volume of each speaker so that the speaker near a microphone is small.

JP-A-9-65470

  The system of Patent Document 1 does not consider howling. If a uniform volume is set at all listening positions, howling is likely to occur in a speaker system close to the microphone.

  On the other hand, if a low gain is set so as not to generate howling, it is not possible to sufficiently squeeze out, and a listening position where it is difficult to hear the speaker's voice occurs.

  On the other hand, an object of the present invention is to provide a gain setting device, a gain setting method, and a program for performing gain setting for transmitting sound at a volume as uniform as possible in a room so that no howling occurs. .

  The gain setting device according to the present invention includes a transfer function acquisition unit, a gain calculation unit, and a gain setting unit. The transfer function acquisition unit includes a first transfer function of a first system from a speaker position to a microphone, a second transfer function of a second system from the speaker position to a plurality of listening positions, and the plurality of listening from a plurality of speakers. A third transfer function of the third system reaching the position and a fourth transfer function of the fourth system extending from the plurality of speakers to the microphone are acquired. The gain calculation unit obtains energy of a certain level or more at all of the plurality of listening positions using the first to fourth transfer functions obtained by the transfer function acquisition unit, and the plurality of speakers. The fifth system gain setting value from the microphone to the plurality of speakers is calculated so that the loop gain in the feedback system from the microphone to the microphone is less than 1. The gain setting unit sets a gain setting value based on the gain calculated by the gain calculation unit in the fifth system.

The sound energy at each listening position is the sum of the sound energy output from the speaker and the direct sound energy from the speaker. The energy of the direct sound from the speaker is expressed by the second transfer function: H _{T1, Ln} . The energy of the sound output from the speaker is expressed by the first transfer function H_T1 _{, M1} , the third transfer function H _{Sn, Ln,} and the gain Gn of the fifth system. Then, the condition for obtaining constant energy at all of the plurality of listening positions is as follows:
H _{T1, L1} (ω) ² + (H _{T1, M1} (ω) ・ H _{S1, L1} (ω) ・ G1 (ω)) ²・ ・ ・ + (H _{T1, M1} (ω) ・ H _{Sn, L1} (ω) ・ Gn (ω)) ²
= H _{T1, L2} (ω) ² + (H _{T1, M1} (ω) · H _{S1, L2} (ω) · G1 (ω)) ² ... + (H _{T1, M1} (ω) · _{HSn, L2} (ω) · Gn (ω)) ²
...
= H _{T1, Ln} (ω) ² + (H _{T1, M1} (ω) · H _{S1, Ln} (ω) · G1 (ω)) ² ... + (H _{T1, M1} (ω) · H _{Sn, Ln} (ω) ・ Gn (ω)) ²
It becomes. The condition for the loop gain p of the feedback system to be less than 1 is to use the fourth transfer function H _{Sn, M1} ,
(G1 (ω) · H _{S1, M1} (ω)) ² + (G2 (ω) · H _{S2, M1} (ω)) ² ··· + (Gn (ω) · H _{Sn, M1} (ω)) ² = p ² <1
It is represented by The loop gain p is a value that does not provide sufficient feedback, for example, 10 log ₁₀ (p ² ) = − 6 dB.

  The gain calculation unit calculates the gain Gn (ω) of the fifth system from the microphone to the speaker by the above relational expression. As a result, the gain setting device can transmit the sound at a volume as uniform as possible in the room without howling.

  The gain setting device of the present invention can perform gain setting that can transmit sound at a volume as uniform as possible in a room without howling.

It is a top view of the room where a loudspeaker system was installed. It is sectional drawing of the room where the loudspeaker system was installed. It is a block diagram which shows the structure of a loudspeaker system. It is a flowchart which shows operation | movement of information processing apparatus. It is a figure which shows the relationship between the distance from a speaker, and energy. It is a block diagram which shows the structure of 4 A of signal processing apparatuses which concern on the application example 3. FIG. It is a top view of the room | chamber interior in which the loudspeaker system 1A which concerns on the application example 5 was installed. It is a block diagram which shows the structure of the signal processing apparatus 4B which concerns on the application example 5. FIG. 9A is a plan view of a room where the loudspeaker system 1B according to Application Example 7 is installed, and FIG. 9B is a cross-sectional view of the room where the loudspeaker system 1B according to Application Example 7 is installed. is there.

  FIG. 1 is a plan view of a room where the loudspeaker system 1 is installed, and FIG. 2 is a cross-sectional view (a cross-sectional view taken along the plane AA shown in FIG. 1). FIG. 3 is a block diagram showing the configuration of the loudspeaker system 1.

  The loudspeaker system 1 includes a microphone M1, a plurality of speakers (in this example, nine speakers S1 to S9), a signal processing device 4, and an information processing device 5.

  In the example of FIG. 2, the loudspeaker system 1 is configured such that the microphone M1 is suspended from the ceiling in the room and directed downward in the room. However, the installation mode of the microphone M1 is not limited to the example of FIG. The microphone M1 may be installed at a position away from the speaker T. For example, the loudspeaker system 1 may be configured such that the microphone M1 is embedded in the ceiling.

  The loudspeaker system 1 has a plurality of speakers embedded in a ceiling and installed facing downward in the room. However, the installation mode of a plurality of speakers is not limited to the example of FIG. For example, the plurality of speakers may also be suspended from the ceiling in the same manner as the microphone M1.

  The loudspeaker system 1 has a plurality of speakers arranged in a matrix in plan view as shown in FIG. The loudspeaker system 1 has the speaker S1 installed at a position closest to the microphone M1. In FIG. 1, the direction in which the microphone M1 is installed when viewed from the speaker S1 is defined as the Y direction, and the direction orthogonal to the Y direction is defined as the X direction. The loudspeaker system 1 has the speakers S1 as the center row, and the speakers S4 and S7 are arranged in the same direction as the speakers S1 along the X direction. In the loudspeaker system 1, the speaker S2 is installed in the center row in the same manner as the speaker S1, and is installed at a position farther from the microphone M1 than the speaker S1. In the loudspeaker system 1, the speaker S5 and the speaker S8 are installed along the X direction in the same column as the speaker S2, with the speaker S2 as a central row. In the loudspeaker system 1, the speaker S3 is installed in the center row in the same manner as the speaker S1, and is installed at a position farther from the microphone M1 than the speaker S2. In the loudspeaker system 1, the speaker S6 and the speaker S9 are installed along the X direction in the same column as the speaker S3, with the speaker S3 as a central row. In addition, the number of speakers and the installation mode in plan view are not limited to the examples shown in FIGS.

  In the example of FIG. 2, the speaker T is located directly below the microphone M1. The microphone M1 picks up the sound spoken by the speaker T.

  As shown in FIG. 3, the signal processing device 4 includes a matrix mixer 42 and a gain adjuster 43.

  The microphone M1 inputs a sound signal related to the acquired sound to the microphone amplifier 40. The microphone amplifier 40 amplifies the input sound signal and then inputs the amplified sound signal to the matrix mixer 42 in the signal processing device 4. The matrix mixer 42 distributes the input sound signal to a plurality of systems (in this example, the same nine systems as the number of speakers).

  The gain adjuster 43 adjusts the gain of the input sound signal in the gain adjusters 431 to 439 of each system and outputs the gain to the speaker amplifiers 451 to 459.

  The speaker amplifiers 451 to 459 amplify the input sound signals and output the amplified sound signals to the speakers S1 to S9.

  In this way, the loudspeaker system 1 loudspeaks the sound acquired by the microphone M1 and delivers it to the listeners L1 to L9.

  The information processing device 5 corresponds to the gain setting device of the present invention, and is composed of, for example, a personal computer. The information processing apparatus 5 includes a CPU and a storage medium (not shown). The CPU of the information processing device 5 operates as the transfer function acquisition unit 51, the gain calculation unit 52, and the gain setting unit 53 illustrated in FIG. 3 by executing the program stored in the storage medium.

FIG. 4 is a flowchart showing the operation of the information processing apparatus 5. First, the transfer function acquisition unit 51 acquires each transfer function in the room (s11). As shown in FIG. 2, the indoor transfer function includes a first transfer function H _{T1, M1 of} the first system from the speaker position T to the microphone M1, and a plurality of listening positions Ln (in this example, from the speaker position T). second transfer function H _{T1, Ln of} the second system up to n = 1 to 9) and third transfer function H of the third system from the plurality of speakers Sn (n = 1 to 9) to the plurality of listening positions Ln. _{Sn, Ln} and a fourth transfer function H _{Sn, M1 of} the fourth system from the plurality of speakers Sn to the microphone M1.

Each transfer function may be actually measured by outputting a sample sound such as white noise from a speaker position or a speaker and acquiring the sample sound at a microphone or listening position. However, the transfer function acquisition unit 51 accepts inputs such as the shape of the room, the position of the microphone, the position of the speaker, the listening position, the speaker position, and the like, and estimates each transfer function from the frequency characteristics of the voice, Each transfer function may be acquired. At this time, the transfer function acquisition unit 51 includes the microphone conversion efficiency in the first transfer function H _{T1, M1} and includes the speaker conversion efficiency in the third transfer function H _{Sn, Ln} and the fourth transfer function H _{Sn, M1.} Thus, each transfer function is estimated.

  The gain calculation unit 52 calculates the gain Gn of the fifth system from the microphone M1 to the plurality of speakers Sn using the transfer function acquired by the transfer function acquisition unit 51 (s12).

  The gain Gn of the fifth system corresponds to the gain setting value to be set in the gain adjusters 431 to 439 of each system in the gain adjuster 43 shown in FIG.

  In the present embodiment, the gain calculation unit 52 obtains a gain setting value that does not cause howling under the condition that the sound energy is a certain level or higher at all listening positions.

  First, the sound energy at each listening position Ln corresponds to the sum of the sound energy output from the plurality of speakers Sn and the direct sound energy from the speaker. Therefore, the gain calculation unit 52 defines the following conditional expression, assuming that this added value is equal at all listening positions Ln.

H _{T1, L1} (ω) ² + (H _{T1, M1} (ω) ・ H _{S1, L1} (ω) ・ G1 (ω)) ²・ ・ ・ + (H _{T1, M1} (ω) ・ H _{Sn, L1} (ω) ・ Gn (ω)) ²
= H _{T1, L2} (ω) ² + (H _{T1, M1} (ω) · H _{S1, L2} (ω) · G1 (ω)) ² ... + (H _{T1, M1} (ω) · _{HSn, L2} (ω) · Gn (ω)) ²
...
= H _{T1, Ln} (ω) ² + (H _{T1, M1} (ω) · H _{S1, Ln} (ω) · G1 (ω)) ² ... + (H _{T1, M1} (ω) · H _{Sn, Ln} (ω) ・ Gn (ω)) ²
Further, the condition that howling does not occur is when the loop gain p of the feedback system from the plurality of speakers Sn to the microphone M1 is less than 1. Therefore, the gain calculation unit 52 uses the fourth transfer function H _{Sn, M1} ,
(G1 (ω) · H _{S1, M1} (ω)) ² + (G2 (ω) · H _{S2, M1} (ω)) ² ··· + (Gn (ω) · H _{Sn, M1} (ω)) ² = p ² <1
It is represented by The loop gain p is a value that does not provide sufficient feedback, for example, 10 log ₁₀ (p ² ) = − 6 dB.

Gj (ω) ² = xj, H _S j _{, M1} (ω) ² = aj, H _{T1, L} j (ω) ² = bj, H _{T1, M1} (ω) ² · H _{S1, L} j (ω) ² = Cj (j = 1 to n), the above relational expression can be expressed by the following determinant 1.

ゲイン算出部５２は、当該行列式をｘjについて解くことにより、ゲインＧｎ(ω)を算出する。

The gain calculation unit 52 calculates the gain Gn (ω) by solving the determinant for xj.

  Then, the gain setting unit 53 sets a gain setting value based on the gain Gn (ω) calculated by the gain calculation unit 52 in the gain adjusters 431 to 439 (s13).

  Here, the gain Gn (ω) calculated by the gain calculation unit 52 has frequency characteristics. When the signal processing device 4 includes an equalizer, the gain setting unit 53 divides the gain Gn (ω) into the gain setting value G0 and the equalizer setting value (frequency characteristic) Leq0 (ω) to set the gain. The value G0 is set in the gain adjuster 431 to the gain adjuster 439, and the equalizer set value (frequency characteristic) Leq0 (ω) is set in the equalizer.

  Further, for each transfer function, the overall value gain Gn can be obtained by using, for example, an overall value obtained by performing A characteristic correction and frequency correction in accordance with audio frequency characteristics, instead of a value for each frequency. In this case, since the gain Gn (ω) optimized for each frequency cannot be obtained, it is preferable to perform in advance a process such as equalizing to reduce the gain of the frequency that is easy to howling.

  Thereby, the information processing apparatus 5 which is a gain setting apparatus can perform the gain setting which can transmit an audio | voice with the sound volume as uniform as possible indoors so that howling may not arise.

(Application 1)
At the listening position Ln near the microphone M1, the energy of the direct sound from the speaker increases. Therefore, in the above calculation formula, the gain Gn (ω) may be a negative value so as to reduce the energy of the direct sound from the speaker at the listening position. However, it is difficult for the speaker Sn to absorb sound energy at the listening position. Therefore, when the gain Gn (ω) becomes a negative value, the gain calculation unit 52 according to the application example offsets the negative value to 0 and calculates the gain Gn (ω) again. Thus, as shown in FIG. 5, when the gain Gn (ω) is a negative value, the energy from the speaker at a position close to the microphone M1 becomes 0, and the direct sound from the speaker becomes dominant. . On the other hand, at a position far from the speaker and where the energy of the speaker's direct sound is small, the sound from the speaker is dominant, and the sound energy is uniform regardless of the listening position. Thereby, a sound energy larger than a certain value (energy of a direct sound from the speaker) is obtained at a listening position close to the speaker, and a certain energy is obtained at a listening position far from the speaker. That is, the energy of sound is a certain level or more at all of the plurality of listening positions.

(Application example 2)
A speaker near the microphone M1 is likely to affect howling. Therefore, the gain calculation unit 52 according to the application example 2 multiplies the above-described added value by a predetermined coefficient wl on the assumption that the gain of the speaker located near the microphone M1 is reduced.

[H _{T1, L1} (ω) ² + (H _{T1, M1} (ω) ・ H _{S1, L1} (ω) ・ G1 (ω)) ²・ ・ ・ + (H _{T1, M1} (ω) ・ H _{Sn, L1} (ω) ・ Gn (ω)) ² ] w1
= [H _{T1, L2} (ω) ² + (H _{T1, M1} (ω) · H _{S1, L2} (ω) · G1 (ω)) ² ... + (H _{T1, M1} (ω) · _{HSn, L2} (ω) · Gn (ω)) ² ] w2
...
= [H _{T1, Ln} (ω) ² + (H _{T1, M1} (ω) • H _{S1, Ln} (ω) • G1 (ω)) ² ... + (H _{T1, M1} (ω) • H _{Sn, Ln} (ω) ・ Gn (ω)) ² ] wn
Gj (ω) ² = xj, H _S j _{, M1} (ω) ² = aj, H _{T1, L} j (ω) ² = bj, H _{T1, M1} (ω) ² · H _{S1, L} j (ω) ² = Cj (j = 1 to n), the above relational expression can be expressed by the following determinant 2.

係数Ｗnを大きくすると、対応するスピーカにおけるゲインＧｎ(ω)の算出結果は、小さくなる。したがって、ゲイン算出部５２は、話者に近い受聴位置（例えば受聴位置Ｌ１）における係数（例えばＷ１）は大きい値を設定し、話者に遠い受聴位置（例えば受聴位置Ｌ９）における係数（例えばＷ９）は小さい値を設定して、ゲインＧｎ(ω)を算出する。

When the coefficient Wn is increased, the calculation result of the gain Gn (ω) in the corresponding speaker is decreased. Therefore, the gain calculation unit 52 sets a coefficient (for example, W1) at a listening position (for example, listening position L1) close to the speaker to a large value, and a coefficient (for example, W9) at the listening position (for example, listening position L9) far from the speaker. ) Sets a small value and calculates the gain Gn (ω).

(Application 3)
FIG. 6 is a block diagram illustrating a configuration of a signal processing device 4A according to Application Example 3. The signal processing device 4 </ b> A further includes an equalizer (EQ) 41 and an EQ 44 with respect to the signal processing device 4. Other configurations are the same as those of the signal processing device 4.

  The EQ 41 adjusts the frequency characteristic on the input side. The EQ 44 has frequency characteristics in EQ 441 to EQ 449 provided for each system on the output side.

  EQ41 and EQ44 are, for example, notch filters that reduce the gain of a specific frequency. For EQ41 and EQ44, for example, the frequency and gain (notch depth) are set in advance by the user. Alternatively, when detecting howling, EQ41 and EQ44 automatically set a notch having a predetermined depth with respect to the detected howling frequency.

These EQ41 and EQ44 are signal processing in the fifth system. Therefore, the transfer function acquisition unit 51 according to the application example 3 acquires the transfer functions of EQ 41 and EQ 44 (that is, setting of the notch filter) as the fifth transfer function of the fifth system excluding the gain Gn (ω). Assuming that the transfer function of _EQ41 is H _EQM1 (ω) and the transfer function of _EQ44 is H _EQSn (ω), the gain calculation unit 52 obtains the gain Gn (ω by the following conditional expression that makes the added value of sound energy uniform. )

H _{T1, L1} (ω) ² + H _{T1, M1} (ω) ²・ H _EQM1 (ω) ²・ [(H _EQS1 (ω) ・ H _{S1, L1} (ω) ・ G1 (ω) ・ ・ ・ + H _EQSn ( ω) ・ H _{T1, M1} (ω) ・ H _{Sn, L1} (ω) ・ Gn (ω)) ² ]
＝ H _{T1, L2} (ω) ² ＋ H _{T1, M1} (ω) ²・ H _EQM1 (ω) ²・ [(H _EQS1 (ω) ・ H _{S1, L2} (ω) ・ G1 (ω) ・ ・ ・ + H _EQSn (ω) ・ H _{T1, M1} (ω) ・ H _{Sn, L2} (ω) ・ Gn (ω)) ² ]
...
＝ H _{T1, Ln} (ω) ² ＋ H _{T1, M1} (ω) ²・ H _EQM1 (ω) ²・ [(H _EQS1 (ω) ・ H _{S1, Ln} (ω) ・ G1 (ω) ・ ・ ・ + H _EQSn (ω) ・ H _{T1, M1} (ω) ・ H _{Sn, Ln} (ω) ・ Gn (ω)) ² ]
Thereby, the gain setting apparatus can perform gain setting including the influence of signal processing such as a notch filter.

(Application 4)
In the above example, the conditional expression is set on the assumption that the number of listening positions and the number of speakers match. However, in practice, the number of listening positions may be different from the number of speakers. Therefore, when the number of listening positions is large, the gain calculation unit 52 according to Application Example 4 considers that the number of listening positions is the same as the number of speakers (reducing or increasing the number of listening positions). ) Calculate the gain.

  However, when the number of listening positions is reduced, the accuracy of the calculated gain decreases. However, when the number of listening positions is larger than the number of speakers, the number of equations is larger than the number of unknowns (gains), so the gain is not uniquely determined. Therefore, the gain calculating unit 52 calculates the gain a plurality of times by changing the listening position, and estimates the most appropriate gain of each system from the calculated plurality of gains. For example, the gain setting device optimizes the least squares method under the condition that the uniformity of energy at the listening position is minimized (that is, the condition that the howling is not satisfied and the difference between each side of the conditional expression is minimized). The gain which is an unknown is obtained by the method.

  Accordingly, the gain setting device can set an appropriate gain setting value even when the number of listening positions is different from the number of speakers.

(Application example 5)
FIG. 7 is a plan view of a room in which the loudspeaker system 1A according to the application example 5 is installed. FIG. 8 is a block diagram illustrating a configuration of a signal processing device 4B according to Application Example 5. The loudspeaker system 1A according to the application example 5 includes a plurality of microphones (a microphone M1, a microphone M2, and a microphone M3).

  In the application example 5, the speaker T is located under the microphone M1, the microphone M2, or the microphone M3. Since the speaker T has a conversation while moving, the distance to each microphone changes.

  The microphone M1, the microphone M2, and the microphone M3 pick up sounds uttered by the speaker T. The microphone amplifier 491, the microphone amplifier 492, and the microphone amplifier 493 amplify sound signals related to sounds acquired by the microphone M1, the microphone M2, and the microphone M3, respectively, and input the amplified signal to the signal processing device 4B.

  The signal processing device 4B further includes an automixer 420 with respect to the signal processing device 4. The automixer 420 mixes the sound signals input from the microphone M1, the microphone M2, and the microphone M3 and outputs the mixed sound signals to the matrix mixer 42 at the subsequent stage. The automixer 420 automatically adjusts the gain of each microphone according to the level of the sound signal input from the microphone M1, the microphone M2, and the microphone M3.

  For example, the automixer 420 selects one of the microphones having the highest level from the sound signals input from the microphone M1, the microphone M2, and the microphone M3, and outputs the sound signal of the selected microphone to the subsequent stage. . Alternatively, the automixer 420 sets the gain of a system with a high level high, and sets the gain of a system with a low level low.

  The gain calculation unit 52 according to the application example 5 calculates the gain in consideration of the change in the setting of the automixer 420. Specifically, the gain is calculated as follows.

(1) A gain is calculated for each of the plurality of microphones.
In this case, assuming that one of the microphones is turned on, the gain is calculated for each microphone. The gain calculation unit 52 calculates the gain for each route from the microphone M1, the microphone M2, and the microphone M3 to each speaker.

The condition for howling is as follows when the number of microphones is m.
Microphone M1: (G1, 1 (ω) • H _{S1, M1} (ω)) ² + (G1, ² (ω) • H _{S2, M1} (ω)) ² ... + (G1, n (ω) • H _{Sn, M1} (ω)) ² = p <1
Microphone M2: (G2, 1 (ω) • H _{S1, M2} (ω)) ² + (G2, ² (ω) • H _{S2, M2} (ω)) ² ... + (G2, n (ω) • H _{Sn, M2} (ω)) ² = p <1
Microphone Mm: (Gm, 1 (ω), H _{S1, Mm} (ω)) ² + (Gm, 2 (ω), H _{S2, Mm} (ω)) ² ... + (Gm, n (ω) H _{Sn, Mm} (ω)) ² = p <1
Further, the condition that the added value of the sound energy is uniform at all listening positions is as follows.
Microphone M1: H _{T1, L1} (ω) ² + (H _{T1, M1} (ω) • H _{S1, L1} (ω) • G1 _{, 1} (ω)) ² … + (H _{T1, M1} (ω) • H _{Sn, L1} (ω) ・ G1, n (ω)) ²
＝ H _{T1, L2} (ω) ² + (H _{T1, M1} (ω) ・ H _{S1, L2} (ω) ・ G1,1 (ω)) ²・ ・ ・ + (H _{T1, M1} (ω) ・ H _{Sn , L2} (ω) ・ G1, n (ω)) ²
...
= H _{T1, Ln} (ω) ² + (H _{T1, M1} (ω) · H _{S1, Ln} (ω) · G1,1 (ω)) ² ... + (H _{T1, M1} (ω) · H _{Sn , Ln} (ω) ・ G1, n (ω)) ²
Microphone M2: H _{T1, L1} (ω) ² + (H _{T1, M2} (ω), H _{S1, L1} (ω), G2,1 (ω)) ² ... + (H _{T1, M2} (ω), H _{Sn, L1} (ω) ・ G2, n (ω)) ²
＝ H _{T1, L2} (ω) ² + (H _{T1, M2} (ω) ・ H _{S1, L2} (ω) ・ G2,1 (ω)) ²・ ・ ・ + (H _{T1, M2} (ω) ・ H _{Sn , L2} (ω) ・ G2, n (ω)) ²
...
= H _{T1, Ln} (ω) ² + (H _{T1, M2} (ω) · H _{S1, Ln} (ω) · G2,1 (ω)) ² ... + (H _{T1, M2} (ω) · H _{Sn , Ln} (ω) ・ G2, n (ω)) ²
Microphone Mm: H _{T1, L1} (ω) ² + (H _{T1, Mm} (ω), H _{S1, L1} (ω), G _m , 1 (ω)) ² ... + (H _{T1, Mm} (ω)・ H _{Sn, L1} (ω) ・ G _m , n (ω)) ²
= H _{T1, L2} (ω) ² + (H _{T1, Mm} (ω) · H _{S1, L2} (ω) · G _m , 1 (ω)) ² ... + (H _{T1, Mm} (ω) · H _{Sn, L2} (ω) ・ G _m , n (ω)) ²
...
= H _{T1, Ln} (ω) ² + (H _{T1, Mm} (ω) · H _{S1, Ln} (ω) · G _m , 1 (ω)) ² ... + (H _{T1, Mm} (ω) · H _{Sn, Ln} (ω) ・ G _m , n (ω)) ²
The gain calculation unit 52 calculates the gain Gm, n (ω) from the above conditional expression.

  In this case, automixer 420 selects one of the sound signals input from microphone M1, microphone M2, and microphone M3, and outputs the sound signal of the selected microphone to the subsequent stage.

  When gain is obtained for each microphone as described above, the value of the gain Gm, n (ω) of each system for m microphones and n speakers is uniquely determined. Therefore, when the gain of each system can be set individually, it is not necessary to change the gain regardless of the microphone selected by the automixer 420. However, when the gain setting unit 53 does not set the gain of all the microphones × speakers but simply sets the gain for each speaker, the calculation is performed for the selected microphone. The gain Gn of each speaker thus set is set in the gain adjuster 43.

(2) Calculate gains for all systems at once.
In this case, it is assumed that the gain of the automixer 420 changes according to the level of the input signal (including the case where all the input signals are mixed), and the equation is solved collectively including the gain of the automixer 420. That is, the gain calculation unit 52 collectively calculates the gains of paths from all microphones to all speakers.

For example, at the listening position L1, the added value of sound energy is as follows. When the automixer 420 is used, the transfer function to each microphone such as H _{T1, M1} (ω) includes the gain of the automixer 420 for each listening position.
H _{T1, L1} (ω) ² + (H _{T1, M1} (ω) · H _{S1, L1} (ω) · G1,1 (ω)) ² ... + (H _{T1, M1} (ω) · H _{Sn, L1} (ω) ・ G1, n (ω)) ²
+ H _{T1, L1} (ω) ² + (H _{T1, M2} (ω), H _{S1, L1} (ω), G2,1 (ω)) ² ... + (H _{T1, M2} (ω), H _{Sn, L1} (ω) ・ G2, n (ω)) ²
+ H _{T1, L1} (ω) ² + (H _{T1, Mm} (ω) · H _{S1, L1} (ω) · G _m , 1 (ω)) ² ... + (H _{T1, Mm} (ω) · H _{Sn , L1} (ω) · G _m , n (ω)) ²
Similarly, if the sum of energy at all listening positions is found and a conditional expression is set when the sum of these energy is uniform, the gain of the path from all microphones to all speakers is calculated in a lump. can do.

  However, when the speaker position changes, the gain of the automixer 420 also changes, so the calculated gain value also changes. Therefore, the gain setting unit 53 selects the one with the lowest risk of howling (for example, the one with the lowest gain) and sets it in the gain adjuster 43. Alternatively, the gain calculation unit 52 sets a plurality of speaker positions and obtains a gain for each speaker position. The gain setting unit 53 associates the speaker position with the setting of the automixer 420 and selects one of the gains according to the setting of the automixer 420.

(3) The gain is calculated by integrating a plurality of microphones into one microphone.
In this case, the gain calculation unit 52 calculates the gain using the same conditional expression as the above-described conditional expression (determinant 1) when there is one microphone. However, since the setting of the automixer 420 changes when the speaker position changes, the gain calculation unit 52 sets a plurality of speaker positions and obtains a gain for each speaker position.

  The gain setting unit 53 selects a gain that hardly causes howling (for example, a gain having the lowest gain) from among the calculated gains, and sets the selected gain in the gain adjuster 43. Alternatively, the gain setting unit 53 associates the speaker position with the setting of the automixer 420 and selects one of the gains according to the setting of the automixer 420.

(Application example 6)
Speakers in the same row (for example, listening position L3, listening position L6, and listening position L9) do not change greatly from the speaker position, and therefore the same gain may be set for the speakers in the same row. In this case, the gain calculation unit 52 may calculate the gain assuming that there is one speaker in the same column (for example, assuming that there is one speaker in the center row).

  When calculating gains for all the speakers, the gain calculation unit 52 sets any one calculation result. For example, the speaker having the lowest gain among the speakers in the same column is selected and set in the gain adjuster 43.

(Application example 7)
9A is a plan view of a room where the loudspeaker system 1B according to Application Example 7 is installed, and FIG. 9B is a cross-sectional view of the room where the loudspeaker system 1B according to Application Example 7 is installed. is there.

  The loudspeaker system 1B of the application example 7 includes a microphone and a speaker for each speaker. Other points are the same as the example shown in FIGS.

  In Application Example 7, the number of microphones, the number of speaker positions, the number of speakers, and the number of listening positions all match. Therefore, in this case, the gain calculation unit 52 calculates the gain for each of the plurality of microphones, for example. That is, the gain calculation unit 52 can obtain the gain in the same manner as the example (1) shown in the application example 5 described above.

52 ... Gain calculation unit M1, M2, M3 ... Microphone S1 ... Speaker S9 ... Speakers 1, 1A, 1B ... Loudspeaker system 4, 4A, 4B ... Signal processing device 5 ... Information processing device 40 ... Microphone amplifier 41 ... EQ
42 ... Matrix mixer 43 ... Gain adjuster 44 ... EQ
49, 491, 492, 493 ... microphone amplifier 51 ... transfer function acquisition unit 52 ... gain calculation unit 53 ... gain setting unit 420 ... auto mixer 431 ... gain adjuster 439 ... gain adjuster 441 ... EQ
451 ... Speaker amplifier 459 ... Speaker amplifier

Claims (11)

A first transfer function of the first system from the speaker position to the microphone, a second transfer function of the second system from the speaker position to the plurality of listening positions, and a third system from the plurality of speakers to the plurality of listening positions A transfer function acquisition unit that acquires the third transfer function of the fourth system and the fourth transfer function of the fourth system from the plurality of speakers to the microphone;
Using the first to fourth transfer functions obtained by the transfer function acquisition unit, energy of a certain level or more is obtained at all of the plurality of listening positions, and reaches the microphone from the plurality of speakers. A gain calculation unit for calculating a gain of a fifth system from the microphone to the plurality of speakers so that a loop gain in the feedback system is less than 1;
A gain setting unit that sets a gain setting value based on the gain calculated by the gain calculation unit in the fifth system;
A gain setting device. The transfer function acquisition unit further acquires a fifth transfer function related to signal processing excluding the gain in the fifth system,
The gain calculation unit further calculates the gain using the fifth transfer function.
The gain setting device according to claim 1. The gain calculation unit, when there is a system in which the gain has a negative value, offsets the negative value to 0 and calculates the gain again.
The gain setting device according to claim 1 or 2. When the number of listening positions is different from the number of speakers, the gain calculating unit regards the number of listening positions as the same as the number of speakers, calculates the gain, and calculates the gain from the calculated gain. Estimating the gain of each system of the fifth system,
The gain setting unit sets the gain setting value based on the gain estimated by the gain calculation unit;
The gain setting device according to any one of claims 1 to 3. The transfer function acquisition unit acquires the first transfer function to the fourth transfer function based on a frequency characteristic of sound.
The gain setting device according to any one of claims 1 to 4. The gain calculation unit calculates the gain for each microphone of the plurality of microphones when there are a plurality of the microphones;
The gain setting unit sets at least one calculation result among the gains calculated for each microphone.
The gain setting device according to any one of claims 1 to 5. The gain calculation unit calculates the gain by integrating the plurality of microphones into one microphone when there are the plurality of microphones.
The gain setting device according to any one of claims 1 to 5. The gain calculating unit calculates the gain for each speaker position among the plurality of speaker positions when there are a plurality of speaker positions;
The gain setting unit sets at least one calculation result among the gains calculated for each speaker position.
The gain setting device according to any one of claims 1 to 7. A gain setting device according to any one of claims 1 to 8,
With a microphone,
Multiple speakers,
A gain adjuster for adjusting each gain of the fifth system to a value according to a gain setting value set by the gain setting unit;
Loudspeaker system with A first transfer function of the first system from the speaker position to the microphone, a second transfer function of the second system from the speaker position to the plurality of listening positions, and a third system from the plurality of speakers to the plurality of listening positions Transfer function acquisition processing for acquiring the third transfer function of the fourth system and the fourth transfer function of the fourth system from the plurality of speakers to the microphone;
Using the first to fourth transfer functions obtained in the transfer function acquisition process, a certain level of energy is obtained at all of the plurality of listening positions, and the plurality of speakers reach the microphone. A gain calculation process for calculating a gain of a fifth system from the microphone to the plurality of speakers so that a loop gain in the feedback system is less than 1;
A gain setting process for setting a gain setting value based on the gain calculated in the gain calculation process to the fifth system;
How to set the gain. A first transfer function of the first system from the speaker position to the microphone, a second transfer function of the second system from the speaker position to the plurality of listening positions, and a third system from the plurality of speakers to the plurality of listening positions Transfer function acquisition processing for acquiring the third transfer function of the fourth system and the fourth transfer function of the fourth system from the plurality of speakers to the microphone;
Using the first to fourth transfer functions obtained in the transfer function acquisition process, a certain level of energy is obtained at all of the plurality of listening positions, and the plurality of speakers reach the microphone. A gain calculation process for calculating a gain of a fifth system from the microphone to the plurality of speakers so that a loop gain in the feedback system is less than 1;
A gain setting process for setting a gain setting value based on the gain calculated in the gain calculation process to the fifth system;
For causing an information processing apparatus to execute the program.

Images