JP2007300404A - Speaker array apparatus and sound beam set method for speaker array apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To more precisely obtain a beam angle to be set to a sound beam of a multi-channel audio signal to be outputted to a speaker array apparatus. <P>SOLUTION: The speaker array apparatus is provided with: beam formation parts 111 to 115 for inputting a signal of each channel of multi-channel audio signal to generate each sound beam directed towards a beam angle set every channel; adders 121 to 12n for composing them; a speaker array 13; a beam control processing part 142 for outputting test sound beams from the beam formation part 111 at every prescribed beam angle; and two microphone sets 2 installed back and forth. A sound pickup signal of the test sound beam is obtained every beam angle. Front and behind of an arrival direction of the test sound is determined by obtaining a cross-correlation function between two sound pickup signals by the two microphones and a beam angle of the sound beam of each channel is set according to a plurality of peaks of correlation functions and the arrival direction of the test sound. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a speaker array device that emits a plurality of sound beams and reproduces a multi-channel audio signal using reflection of a wall.

  Conventionally, by using a speaker array device in which a plurality of speakers are arranged and reflection of sound by a wall in a listening room, multi-channel sound that gives a feeling that sound is output from around the listener is reproduced in a pseudo manner. A speaker array device has been put into practical use (see, for example, Patent Document 1).

  This speaker array apparatus simultaneously outputs sound beams directed in different directions corresponding to each of the multi-channel audio signals from an array-shaped speaker installed in front of the listener. These sound beams reach the listener through the reflection of the wall from the side or the rear where there is no speaker. Therefore, by installing this speaker array device in the front, it is possible to give a feeling that different sounds are being output from around the listener (see FIG. 1 described later).

In such a speaker array device, an indirect element such as wall reflection is used to form a sound field, so that the setting of the sound field is difficult depending on the environment of the room.
Patent Document 1 describes an automatic setting method for the speaker array device. In this apparatus, there is a description that a sound beam having a test sound as an input is emitted from the speaker array and the test sound is picked up from a microphone installed at the listening position. Also, a peak with a high signal level is calculated while turning the direction in which the sound beam of the test sound is output, and the beam angle corresponding to this peak is automatically set as the direction in which the sound beam of the multi-channel audio signal is output. There is a statement to that effect.
JP 2006-13711 A

  However, the method of Patent Document 1 still has a problem that the accuracy of setting the sound beam emitting direction of the speaker array device is not good. That is, in the method of Patent Document 1, since the beam angle of the test sound beam having a high volume is only selected as the direction of outputting the sound beam of the multi-channel audio signal, the beam angle to be set may be wrong. For example, there is a possibility that the sound beam of the rear channel assumed by the input of the multi-channel audio signal is erroneously set to the front channel. Further, when setting the speaker array device, there is a problem that it is vulnerable to disturbance noise caused by other sounds, such as when a car passes around the listening room.

  Therefore, an object of the present invention is to provide a speaker array device and a sound beam setting method for the speaker array device that can more accurately determine the beam angle at which the sound beam emitting direction of the speaker array device should be set. .

  The present invention has the following configuration as means for solving the above problems.

(1) The present invention
A speaker array having a plurality of speakers arranged in a matrix or a line;
The audio signal of each channel of the multi-channel audio signal is input, the timing to output to each speaker of the speaker array is controlled, and the sound beam directed in the direction set for each signal of each channel is generated. Audio signal beam output means for outputting a signal obtained by synthesizing these sound beams;
A test sound beam output means for inputting a test sound and controlling a timing of outputting the sound to each speaker of the speaker array and outputting a test sound beam in a predetermined direction;
Test audio beam direction indicating means for instructing the test audio beam output means to turn a horizontal beam angle of the test audio beam within a predetermined angle range;
A sound collecting means that includes two microphones arranged side by side at the listening position, and that collects sound including the sound of the test sound beam and records a sound collecting signal;
The arrival time of the two sound pickup signals picked up by the two microphones is determined for each beam angle of the test sound beam, and whether the direction of arrival of the sound beam corresponding to the beam angle is the front of the microphone. Before and after judging means for judging whether it is backward,
A plurality of peak values in the picked-up signal recorded for each beam angle of the test sound beam are selected, and a beam angle corresponding to the selected peak is determined as a horizontal beam of the sound beam of the multi-channel audio signal. Candidate selection means for selecting as angle candidates;
The front-rear direction determined by the front-rear determination unit for the collected sound signal corresponding to the beam angle selected as the candidate is matched with the front-rear direction in which the input of the multi-channel audio signal assumes speaker output. And a setting means for setting a beam angle candidate as a beam angle of an audio beam of the multi-channel audio signal.

  According to this configuration, since the two microphones are installed in the front-rear direction, the front-rear determination unit determines the arrival time of the collected sound signal, and the direction in which the test sound beam arrives is determined by the microphone. It can be judged whether it is forward or backward.

  Further, the candidate selection means can select the peak of the sound pickup signal from the two sound pickup signals as a horizontal angle candidate of the sound beam of the multi-channel audio signal, and the signal of the beam angle selected as the candidate is forward. Therefore, the multi-channel audio signal matches the front-rear direction assuming the speaker output, so that each channel of the multi-channel audio signal is matched. The sound beam can be set accurately.

  Here, “front” refers to the direction of viewing the speaker array from the listener, and “rear” refers to the opposite direction.

(2) The present invention
The front-rear determination unit determines the front and rear of the arrival time by calculating a cross-correlation function for the collected sound signals collected by the two microphones for each of the beam angles.

  If comprised in this way, the front and back of the said arrival time can be judged more accurately than the simple comparison of the head part of a sound-collection signal by the cross correlation function of the said two sound-collection signals. Here, the cross-correlation function can be expressed by, for example, [Equation 1] and [Equation 3] described later.

  Note that the peak can be substituted with a cross-correlation function for the arrival time difference. When the two sound pickup signals are a (t) and b (t), if the waveforms of these two signals completely match with the time difference τ, the cross-correlation function of a (t) and b (t) is There is a peak at τ, which is the integral value of the square of the peak value. When the peak is obtained with the cross-correlation function in this way, determination is made based on a value based on the correlation of each waveform over a certain period of time, so that robust measurement can be performed for noise unrelated to the waveform.

(3) The present invention
A speaker array having a plurality of speakers arranged in a matrix or a line;
An audio beam of a multi-channel audio signal that is input to each channel of the multi-channel audio signal and is output in a direction set for each signal of each channel by controlling the timing of output to each speaker of the speaker array. And a multi-channel audio beam output means for outputting a signal obtained by synthesizing the audio beams of these multi-channel audio signals.
A test sound beam output step of inputting a test sound and controlling a timing of outputting to each speaker of the speaker array and outputting a test sound beam in a predetermined direction;
A test sound beam direction indicating step for instructing the test sound beam output means to turn a beam angle in a horizontal direction of the test sound beam within a predetermined angle range;
A sound collecting step of collecting sound of the test sound beam and recording a sound collecting signal for each beam angle of the test sound beam, using two microphones installed side by side at the listening position; and
The arrival time of the two sound pickup signals picked up by the two microphones is determined for each beam angle of the test sound beam, and the direction in which the test sound beam corresponding to the beam angle arrives is the microphone. Before and after determining whether the front or rear of the
A plurality of peak values in the picked-up signal recorded for each beam angle of the test sound beam are selected, and a beam angle corresponding to the selected peak is determined as a horizontal beam of the sound beam of the multi-channel audio signal. A candidate selection step for selecting as an angle candidate;
The front-rear direction determined by the front-rear determination unit for the two collected sound signals corresponding to the beam angle selected as the candidate is matched with the front-rear direction in which the input of the multi-channel audio signal assumes a speaker output, And a setting step of setting a selected beam angle candidate as a beam angle of a sound beam of the multi-channel audio signal.

  If this invention is implemented, there exists an effect | action similar to the effect | action and effect of (1), and an effect.

  According to the present invention, a speaker array that artificially reproduces multi-channel sound that gives a feeling as if sound is being output from the surroundings of a listener by using sound reflection by the speaker array device and a wall in the listening room. For the device, the sound beam of the multi-channel audio signal can be set accurately.

<Overview of speaker array system>
The outline of the speaker array system will be described with reference to FIG. FIG. 1 is a plan view of a state in which a sound beam of a 5-channel multi-channel audio signal is emitted from the speaker array device as viewed from above. The speaker array system 1 is used for setting the speaker array apparatus main body 10, a video apparatus 109 that projects an image on a screen, and a speaker array apparatus main body 10 in front of a listener L who is a listener inside a rectangular parallelepiped listening room 100. A microphone set 2 is provided. Here, “front” refers to the direction of viewing the speaker array apparatus body 10 from the listener L who is the listener, and “rear” refers to the opposite direction.

  The speaker array system 1 includes a plurality of speakers arranged in an array, and multi-channel sound is transmitted from the listener L using the speaker array 11 provided only in front of the listener L and the reflection of the walls 101 to 104. It can be heard from the surroundings. In order to localize the multi-channel sound independently, the speaker array device body 10 generates and outputs a sound beam of a multi-channel audio signal having directivity in predetermined directions different from each other for each channel. To do.

  Specifically, the Cch (channel) sound beam is emitted toward the front (listener L), and the Lch and Rch sound beams are reflected at the positions of the left and right virtual Lch 105 and virtual Rch 106 as viewed from the listener. To the listener L. Further, the sound beams of SLch and SRch are reflected twice by the left and right wall surfaces 102 and 104 and the rear wall surface 103 when viewed from the listener L, and are emitted from the virtual SLch 107 and the virtual SRch 108 toward the listener L. By controlling the direction in which the sound beam is emitted in this way, it is possible to give the listener L the feeling as if a speaker was placed at the position of the virtual Lch 105, virtual Rch 106, virtual SLch 107, and virtual SRch 108.

  Further, the speaker array system 1 of the present embodiment has a function of automatically setting the beam angle. When the test sound is output from the speaker array apparatus main body 10, the two microphones are installed at the listening position so as to be front and rear. The horizontal beam angle of the sound beam of the multi-channel audio signal is set based on the sound collection signal from which the microphone picks up the test sound and the arrival time difference between the sound collection signals.

<Configuration of Speaker Array System of the Present Embodiment>
The internal configuration of the speaker array system according to the embodiment of the present invention will be described with reference to FIG. FIG. 2 is a block diagram showing this configuration. In the following description, an example for a 5ch surround system will be described as an example of a speaker array device. In the following description, in a 5ch surround system, the front left channel is L (Left) ch, the front right channel is R (Right) ch, the center channel is C (Center) ch, and the rear left channel is SL ( SurroundLeft) ch and the rear right channel are called SR (Surround Right) ch.

  The speaker array system 1 includes a speaker array device body 10 and a microphone set 2 used for setting the speaker array device body 10. The speaker array apparatus body 10 includes an Lch terminal, an Rch terminal, an SLch terminal, an SRch terminal, and a Cch terminal as external input terminals for 5ch multichannel audio signals. The speaker array device main body 10 receives the 5ch multi-channel audio signal as an input, and outputs a “multi-channel audio signal audio beam” which is a bundle of five audio beams in which the signals of each channel are respectively directed in a predetermined direction. Beam forming units 111 to 115 to be generated, adders 121 to 12n for combining outputs of these beam forming units, a speaker array 13, a system control unit 14, a storage unit 15, an operation unit 16, a display unit 17, and an A / A D converter 18 is provided.

  Each of the beam forming units 111 to 115 shown in FIG. 1 has a delay unit 31 having n delay taps corresponding to the speakers (301 to 30n) of the speaker array 12, and a delay tap that is an output of the delay unit 31. Gain adjusters 331 to 33n connected to each other and adders 121 to 12n for adding the signals output from the respective gain adjusters 331 to 335 are provided, and one multi-channel audio signal is input as a predetermined direction. An audio beam of the directed multi-channel audio signal is generated. For example, the beam forming unit 111 inputs a Cch multi-channel audio signal.

  The sound beam of the multi-channel audio signal is generated by adjusting the timing (delay amount) that the delay unit 31 outputs to each speaker of the speaker array based on the delay amount set by the beam control processing unit 142. When this sound beam is output, the sound is propagated greatly because the timing to be output to each speaker is aligned in a predetermined direction, and the sound propagated in the other direction does not match this timing, so the amplitude is attenuated and the sound is reduced. Can be. Therefore, according to the sound beam technology, it is possible to emit sound having directivity in this direction.

The delay unit 31 can be composed of a ring buffer or the like, and adjusts the timing (delay amount) output to each speaker. Each of the delay units 31 of the beam forming units 121 to 125 inputs one audio signal from among the multi-channel audio signal inputs, and the input audio signal is recorded in a ring buffer constituting the delay unit 31, In each delay tap, the data recorded in the ring buffer is read with a delay by an amount instructed by the beam control processing unit 142. As a result, an output of the audio signal obtained by delaying one audio signal input is generated corresponding to each of the speakers 301 to 30n of the speaker array 13, and is output to the gain adjusters 331 to 33n.
The gain adjusters 331 to 33n adjust the gain of the output signal of the delay unit 31.
The adders 121 to 12n combine the outputs of the beam forming units 111 to 115 for each of the speakers 301 to 30n of the speaker array 13. Further, the outputs of the adders 121 to 12n are amplified by an amplifier (not shown) and output to the speaker array 13.

  The speaker array 13 includes speakers 301 to 30n arranged in a matrix or in a row and a power amplifier (not shown) (see FIG. 2 described later), and the audio signals output from the adders 121 to 125 are supplied to this power. The sound is converted into sound through an amplifier and emitted. When the synthesized sound of the sound beam of the multi-channel audio signal is emitted by the speaker array 11, the sound beam directed in the direction described with reference to FIG. 1 is simultaneously emitted corresponding to the sound signal input of each channel. Sounded.

The system control unit 14 includes a user I / F processing unit 141, a beam control processing unit 142, and a measurement data analysis processing unit 143.
The user I / F processing unit 141 outputs a control signal to each unit of the speaker array system 1 according to the operation received by the operation unit 16. Further, the user I / F processing unit 141 outputs contents to be notified to the listener to the display unit 16 according to the status of the apparatus.

  The beam control processing unit 142 sets the beam angle of the sound beam of each multi-channel audio signal. Specifically, the beam control processing unit 142 refers to the delay amount data set in the delay unit 31 from the storage unit 15 and outputs the data to the delay unit 31.

  The system control unit 14 including the beam control processing unit 142 has a function of setting the beam angle of the sound beam of each multi-channel audio signal. When executing the sound beam setting mode for executing this function, the beam control processing unit 142 inputs test sound to one of the beam forming units 111 to 115 and outputs it from the speaker array 11. Turn the beam direction horizontally. The beam control processing unit 142 sets the beam angle of the test audio beam, and sets the beam angle of the audio beam of each multi-channel audio signal based on the analysis result of the measurement data analysis processing unit 143.

  The measurement data analysis processing unit 143 determines an optimum beam angle for the sound beam of the multi-channel audio signal based on the test sound signal collected by the microphone set 2 when the sound beam setting mode described above is executed. In order to set the beam angle, the measurement data analysis processing unit 143 instructs the beam control processing unit 142 to rotate the beam angle of the test audio beam, and emits the test audio beam at a predetermined angle step. The sounded test sound beam is picked up by the microphone set 2 and stored in the storage unit 15. When the collection of the audio signal is completed, the measurement data analysis processing unit 143 reads out the collected sound signals of the two microphones stored in the storage unit 15 for each beam angle of the test audio beam, and calculates their cross correlation. Based on the magnitude of the cross-correlation function, the arrival direction of the sound beam that has reached the microphone is estimated, and an appropriate beam angle is set for the sound beam of each channel based on the estimation of the arrival direction. Details will be described later.

  The storage unit 15 includes a ROM and a RAM, stores various control data, and stores a digital audio signal output from the A / D converter 18 via the system control unit 14. For example, the storage unit 15 stores delay amount data to be set in the delay unit 31, and the beam forming units 111 to 115 generate a sound beam directed in a predetermined direction based on the delay amount data. The storage unit 15 accepts various setting inputs from the listener and outputs signals corresponding to the inputs to the system control unit 14.

  The operation unit 16 instructs the system control unit 14 to perform various operations. For example, the operation unit 16 processes multi-channel audio signal input to instruct to emit sound from the speaker array 13, instructs to adjust the volume of the sound, and the sound beam of the multi-channel audio signal is transmitted. It is possible to instruct execution of a sound beam setting mode for adjusting the beam angle.

  The display unit 17 displays the content transmitted to the listener L based on the control signal output from the system control unit 14.

  The A / D converter 18 converts the two analog sound pickup signals picked up by the microphone set 2 into digital sound signals (sound pickup signal data), and outputs them to the system control unit 14.

  The microphone set 2 includes two omnidirectional microphones and is connected to the A / D converter 18. The microphone set 2 collects the test sound emitted from the speaker array 11 when the sound beam setting mode is executed. Is converted into an electrical signal. The microphone set 2 is used when the sound beam setting mode is executed. At this time, the microphone set 2 is installed by the user U in the front-rear direction of the listener L, that is, in the direction of the straight line connecting the listener L and the speaker array apparatus body 10. It is used in the state that was done. The microphone set 2 picks up the test sound and converts it into an electrical signal when the sound beam setting mode is executed, and obtains a sound collection signal of the test sound.

  Next, an installation position of the microphone set 2 and a configuration example of the microphone set 2 will be described with reference to FIG. FIG. 3A shows the installation position of the microphone set 2 and corresponds to the positional relationship of FIG. FIG. 3B shows an AA arrow view (90-degree rotation diagram) of the configuration example of the microphone set 2 of FIG. As shown in FIG. 3A, the microphone set 2 has microphones 21 and 22 arranged side by side on a straight line connecting the listener L and the speaker array apparatus body 10 at the position of the listener L in FIG. At this time, the height of the microphone set 2 is preferably adjusted to the listening position of the listener. As shown in FIG. 3B, the microphone set 2 includes a front microphone 21, a rear microphone 22, a support base 23 that supports the microphone 21, and a leg 24 that stands the support base 23.

  By configuring the microphone set 2 in this way and installing it at the listening position, the direction in which the sound beam arrives from the arrival time difference, which is the time difference in which the test sound beam was picked up by the microphones 21 and 22, is forward. It can be determined whether it is backward. Specifically, the arrival time difference can be obtained from the time difference when the cross-correlation function of the collected sound signals of the two test voices obtained by the microphones 21 and 22 becomes the maximum value.

  Next, an arrangement example of the speaker array will be described with reference to FIG. 4A and 4B are diagrams showing this arrangement. FIG. 4A shows a case where speakers are arranged in a matrix, FIG. 4B shows a case where three rows of speakers are arranged in a line (matrix arrangement), and FIG. 4C shows a speaker in a line. Is arranged in three rows and the speakers in the second row are shifted (honeycomb-like arrangement).

  Here, as shown in FIG. 4, the speaker array 11 includes a plurality of (n) speakers 30 arranged in a predetermined arrangement in a line on one panel, and the timing of outputting surround sound from each speaker is set. It is adjusted for each channel and radiated in the form of a beam, and delay control is performed so that the sound beam of the multi-channel audio signal is focused at an arbitrary position such as a wall surface. Then, the sound of each channel is reflected on the wall of the room where the speaker array system 1 is installed, thereby creating a sound source at an arbitrary point and forming a sound field of multi-channel sound.

<Description of setting method of multi-channel audio signal sound beam>
Next, a method for setting an audio beam of a multi-channel audio signal will be described with reference to FIGS.
First, the suitability of the beam angle set for the sound beam will be exemplified as a premise for obtaining the beam angle set for the sound beam of the multi-channel audio signal with reference to FIG. FIG. 5 is a diagram for explaining the suitability of the direction in which the sound beam arrives at the microphone set 2, and the reference numerals and positional relationships correspond to those in FIG. 3.

  When the beam angle θ of the test sound beam in FIG. 5 is θ = θ1, the sound beam 42 is reflected by the left wall 102 and then reaches the microphone set 2, which is appropriate as the beam angle of the Lch sound beam. On the other hand, when the beam angle θ = θ6 of the test sound beam, the sound beam 42 is reflected by the left wall 102 and the right wall 104 and then reaches the microphone set 2, so that the sound image is localized from the opposite direction, L It is inappropriate as the beam angle of the sound beam of the channel.

  When the test audio beam angle θ = θ2 in FIG. 5, the test audio beam is reflected by the left wall 102 and the rear wall 103 and then reaches the microphone set 2. Therefore, the beam angle for outputting the SLch audio beam is Appropriate and can be set to the beam angle of the SLch audio beam.

  In the case where the beam angle θ of the test sound beam in FIG. 5 is θ = θ3, the sound beam 41 directly reaches the microphone set 2 and is therefore suitable as the beam angle of the Cch sound beam.

  Further, the beam angles θ = θ4 and θ5 of the test sound beam are appropriate as the beam angles of the SRch and Rch sound beams, respectively, similarly to θ3 and θ2 described above.

FIG. 6 is an explanatory diagram of an audio beam setting method for multi-channel audio signals of the speaker array device.
A method in which the beam control processing unit 142 turns the test sound beam will be specifically described with reference to FIG. FIG. 6A is a diagram for explaining microphone arrangement and beam turning. When the installation of the speaker array apparatus body 10 and the microphone set 2 is completed and the sound beam setting mode is set, the beam control processing unit 142 moves the speaker array 11 to the listening room 100 as shown in FIG. From one direction parallel to the front surface of the speaker array 11 (hereinafter referred to as 0 degree direction) to the other direction (hereinafter referred to as 180 degree direction). Is turned in predetermined angle increments (for example, in increments of 2 degrees).

  At this time, the beam control processing unit 142 sends the test sound to the beam forming unit 111 that is one of the beam forming units 111 to 115 for each beam angle between 0 degrees and 180 degrees on the front surface of the speaker array 11. Enter and emit a test sound beam. The system control unit 14 collects the collected sound signals for each beam angle by the two microphones 21 and 22 and stores the two collected sound signal data for each beam angle. The measurement data analysis processing unit 143 reads the stored data from the storage unit 15 and obtains a cross-correlation function between the two collected sound signal data for each angle θ of the test sound beam.

  Note that the test sound output when executing the sound beam setting mode is selected to have no autocorrelation. For example, sound waves having no autocorrelation such as white noise are suitable. TSP (time stretched pulse) may be used.

  Next, a method for obtaining this cross-correlation function will be described with reference to FIG. FIG. 6B is a graph in which the horizontal axis plots the mutual time difference τ [milliseconds] of the two microphones and the cross-correlation function of the collected sound signals of the two microphones. As a result of obtaining the cross-correlation function, when there is mutual correlation, a peak 51 is obtained, and the time difference τ at which this peak is observed is the arrival time difference between the two microphones 21 and 22 of the microphone set 2. Although the time difference τ is positive in the figure, there are cases where it is both positive and negative, and the direction of arrival can be determined by this positive / negative.

First, the cross-correlation function for the continuous time t will be described. When the function of the continuous time t is a (t) and b (t), the cross-correlation function R _ab (τ) shifted by the time difference τ is obtained by integrating the product of a (t) and b (t + τ) with time. It can be expressed as follows.

The cross-correlation function R _ab (τ) is maximized if a (t) and b (t + τ) obtained by shifting the time difference τ by b (t) temporarily match. If the time difference τ at which the cross-correlation function R _ab (τ) becomes the maximum value is calculated by changing τ, the arrival time difference can be obtained if a (t) and b (t) are sound pickup signals. The absolute value of the arrival time difference is the maximum value when the voice comes from the front or from the back, and this value is the time obtained by dividing the distance L between the microphones 21 and 22 by the speed of sound 340 m / sec ( This is referred to as τ _max .). Accordingly, assuming that the arrival time difference is T, T is obtained in the range of −τ _max ≦ τ ≦ τ _max , and is obtained as τ when the value of this function is maximized. That is, it can be obtained by the following equation.

  Actually, since the collected sound signal of the test sound is actually collected signal data of discrete values as described above, the collected sound signals of the microphone 21 (front) and the microphone 22 (rear) are each expressed as az function. If (z) and b (z) are represented and the sampling time is sp, the result is as follows.

  Also, when the discrete functions a (z) and b (z) are used, the arrival time difference T can be calculated using the above-described [Equation 2]. However, since the response data has a finite length, it may be obtained within that range.

The arrival time difference T calculated in this way can be positive or negative as described above. Based on the sign of T, it can be determined whether the direction in which the sound that has reached the microphones 21 and 22 has arrived is from the front or the rear. This will be described with reference to FIG. For example, if the sound collection signal a (z) of the front microphone 21 reaches T ₀ (T ₀ > 0) faster than the sound collection signal b (z) of the rear microphone 22, a (z) And b (z + T ₀ / sp) are substantially the same signal, the cross-correlation function R _ab (τ) is the maximum value at this T ₀ , and T is a positive value like the peak 51 in FIG. T ₀ become. On the other hand, if b (z) arrives later by T ₁ (T ₁ > 0), a (z) and b (z + (− T ₁ / sp)) are substantially the same signal, and FIG. Like the peak 52 in B), T has a negative value -T ₁ .

Further, the measurement data analysis processing unit 143 obtains these cross-correlation functions R _ab (τ) for each beam angle θ of the test sound beam. Since the function of R _ab (τ) is obtained for each beam angle θ of the test sound beam, the cross-correlation function of the beam angle is expressed as R _abθ (τ). The arrival time difference T described above is also T (θ) by adding θ to the argument.

Here, R _abθ (T (θ)) is the peak value of the sound signal obtained by measuring the test sound with one microphone around the microphone set 2 in an ideal case where there is no ambient noise or disturbance. You are calculating the square. That is, R _abθ (T (θ)) is an alternative value for the volume in the microphone set 2. When the peak is obtained by the cross-correlation function in this way, determination is made based on the correlation value of each waveform over a certain period of time, so that robustness is obtained by squaring the peak value of the audio signal against noise unrelated to the waveform. Can be calculated.

The measurement data analysis processing unit 143 calculates R (T (θ)) obtained by dividing R _abθ (T (θ)) by the measurement time Tn of a (t) as an alternative value of the volume in the microphone set 2, A set with the positive and negative of T (θ) and T (θ) is stored in the storage unit 15.

  Based on the above description, a method for determining the beam angle from R (T (θ)) will be described with reference to FIG. In FIG. 6C, the horizontal axis represents the beam angle θ of the test sound beam, and the vertical axis represents the value R (T (θ)) obtained by dividing the maximum value of the cross-correlation function at each beam angle by the measurement time. It is a graph. Since the volume of the sound collected by the microphone set 2 is reduced according to the number of reflections by the wall before the sound is collected, if the listening room 100 has a rectangular parallelepiped shape and the wall is uniform, FIG. The result should be as shown in Here, when the test sound beams 41 to 46 are set as shown in FIG. 5A, the volume of the test sound beam 41 that directly reaches the listener L is large, and then the test sound beams 42 and 45 that are reflected once and arrive. Then, the sound volume decreases in the order of the test sound beams 43, 44, and 46 (including those that are symmetrical with 46) that are reflected twice. Accordingly, it is assumed that R (T (θ)) also becomes smaller, so it can be estimated that the peaks 410 to 460 in FIG. 6C correspond to the test sound beams 41 to 46 in FIG.

  Therefore, as shown in FIG. 6C, one of the beam angles θa1 to θa7 of the test sound beam corresponding to the peaks 410 to 460 is set as the beam angle set to the sound beam of the multichannel audio signal. Here, in the speaker array system 1 of the present embodiment, as described above, it is possible to determine whether the direction in which the test sound has arrived at least with respect to the listener L is the front or back, so that the test beam corresponding to the direction in which the test sound has arrived. The angle ranges 53 and 54 in which the test sound arrives from the rear wall 103 are calculated, and these peaks are divided.

  When the peaks are divided in this way, the beam angle set for the sound beam is as follows. Θa3 corresponding to the largest peak 410 measuring the direct sound in the front middle is set to Cch. Also, θa2 and θa4 corresponding to the peaks 430 and 440 included in the angle ranges 53 and 54 where the test sound comes from the rear wall 103 are set to SLch and SRch. Further, Lch is set to any one of θa1 and θa6 corresponding to the left peaks 420 and 460 other than the angle range 53, and La and θa5 corresponding to the right peaks 450 and 461 other than the angle range 54 are set. Any one of θa7 is set to Rch.

For Lch and Rch, the peaks are further narrowed down from θa1, θa6, θa5, and θa7 corresponding to the peaks 420, 450, 460, and 461. When the Lch is set like the test sound beam 46 in FIG. 5, it is reflected twice and is heard from the wall on the opposite side of the left and right, and this impairs the sound image localization. This needs to be prevented. Therefore, the measurement data analysis processing unit 143 has a beam angle 0 ≦ θ ≦ min (θ _L− of the test audio beam outside the angle ranges 53 and 54 that are the beam angle ranges of the test audio beam from which the test audio has arrived from the rear. ), Max (θ _R− ) ≦ θ ≦ 180), θa 1 and θa 5 corresponding to the peaks 420 and 450 closest to min (θ _L− ) and max (θ _R− ) are respectively Lch and Rch. Is determined as the beam angle.

  Although FIG. 6B shows data from which noise has been removed, the waveform of measurement data is actually distorted or finely changed due to noise or the like. Also, the horizontal axis of the graph shown in FIG. 6C is set to the beam angle, and the vertical width is set to the gain of the audio data collected by the microphone set 2. In order to easily detect a plurality of peaks from the audio data, a threshold is set to a level at which only an audio beam that has been reflected by the wall up to two times can be detected.

  If the peak does not appear sufficiently, for SL and SRch, the maximum value of R (T (θ)) in the angle ranges 53 and 54, or the angle θ of the test audio beam within this angle range. The center is taken, and Lch and Rch are outside the range of the angular ranges 53 and 54 and take the values closest to the angular ranges 53 and 54 among the maximum values of R (T (θ)). Even when no peak appears in this way, it is desirable that the angle of the sound beam be set so that the volume of the sound arriving at the listener L is increased. This is because, when the beam angle θ of the test sound beam having a small volume is set, unlike the case where it is appropriately set as described in FIG. 5, the sound is reflected many times in the listening room 100 and the sound reaches the listener L. This is because the sound image localization may be impaired. Also, considering the efficiency of sound propagation, it is desirable in terms of efficiency to select the beam angle θ of the test sound beam that increases the sound volume at the position of the listener L.

  Next, the sound beam setting mode described above will be described using the flowchart of FIG. In the following flow, it is assumed that the microphone set 2 is set and the mode is set automatically. This flow is broadly divided into a measurement flow 61 for obtaining collected signal data by outputting test sound for each beam angle, a data processing flow 62 for processing the measured data, and a multi-channel based on the data processing flow. It can be divided into an angle setting flow 63 for setting the beam angle of the audio beam of the audio signal.

The measurement flow 61 in FIG. 7 will be described. In this flow, the operations from ST1 to ST5 are performed for each angle θ of the test sound beam described in FIG. Here, the angle of the test sound beam is set in increments of 2 degrees from 2 degrees to 178 degrees.
In ST1, the angle θ of the test sound beam is set.
In ST2, the delay amount necessary for outputting the sound beam having the angle θ of the test sound beam is read from the storage unit 15.
In ST3, the delay amount read in ST2 is set in the beam forming unit 111, which is one of the beam forming units 111 to 115.
In ST4, a test voice is input to Cch which is an input of the beam forming unit 111. Thereby, a test sound beam having directivity in the direction of the angle θ can be generated.
In ST5, the sound collection signal data obtained by digitizing the two sound collection signals collected by the microphone set 2 via the A / D converter 18 are stored in the storage unit 15, respectively.

The data processing flow 62 in FIG. 7 will be described. In this flow, the operations from ST11 to ST13 are respectively performed on the data for each angle θ of the test sound beam stored in the storage unit 15.
In ST11, a cross-correlation function R _ab (τ) in a range of −τ _max ≦ τ ≦ τ _max is obtained according to [Equation 3].
In ST12, τ that maximizes R _ab (τ) is obtained from the cross-correlation function R _ab (τ), and this value τ is set as T (θ) as a function of the beam angle θ of the test speech beam. Further, the sign of T (θ) is obtained. Further, R (T (θ)) normalized by dividing the cross-correlation function R _ab (T (θ)) at τ = T (θ) by the measurement time Tn of a (t) is calculated. A set with (θ) is stored in the storage unit 15.
In ST13, the value obtained in ST12 is stored.

The data processing flow 63 in FIG. 7 will be described. In this flow, the operations from ST11 to ST13 are performed for each beam angle of the test sound beam.
In ST21, among the values of R (θ), the angle θ of the test sound beam that is the maximum value of the peak near the center (θ = 90 degrees) is set as the angle of the sound beam of Cch.
In ST22, angle ranges 53 and 54 (see FIG. 6) in which T (θ) is negative in the left side (0 ≦ θ ≦ 90 degrees) and the right side (90 degrees ≦ θ ≦ 180 degrees) are obtained (this region). Are θ _L− and θR ₋ , respectively). The angle θ of the test sound beam at which R (θ) has a peak value in θ _L− is set to the angle of the LSch sound beam. On the right side, θ at which R (θ) has a peak value is set as the SRch sound beam angle.

In ST23, out of the regions θ _L− and θ _R− where T (θ) is a negative value in the region where T (θ) is a positive value, 0 ≦ θ ≦ min (θ _L− ), max The value of the beam angle θ of the test sound beam in which R (θ) has a peak value in the region of (θ _R− ) ≦ θ ≦ 180 is obtained. Of beam angle theta of the obtained test sound beam, most theta is in the range of _{0 ≦ θ ≦ min (θ L-} ) to set the close to min (θ _L-) the beam angle of the Lch. Of the obtained beam angles θ of the test sound beam, θ on the right side (90 degrees ≦ θ ≦ 180 degrees), ie, max (θ _R− ) ≦ θ ≦ 180, is the closest to max (θ _R− ). Set the beam angle of the Rch sound beam.

A plan view showing a state where a 5-channel sound beam is emitted from a speaker array system 1 is a block diagram showing a schematic configuration of a speaker array system according to an embodiment of the present invention. The figure which shows the installation position and structural example of the microphone set of the speaker array system which concerns on embodiment of this invention Speaker array layout The figure for demonstrating the appropriateness | suitability of the beam angle set to the audio | voice beam of a multichannel audio signal Explanatory drawing of audio beam setting method of speaker array device Flow chart of sound beam setting mode of speaker array device

Explanation of symbols

1-speaker array system, 10-speaker array device main body 111-115-beam forming unit, 121-12n-adder, 13-speaker array 14-system control unit, 141-user I / F processing unit 142-beam control processing Unit, 143-measurement data analysis processing unit 15-storage unit, 16-operation unit, 17-display unit, 18-A / D converter 2-microphone set, 21-microphone, 22-microphone, 23-support base, 24- Leg 30-speaker, 31-delay unit, 331-33n-gain adjuster 41-46-sound beam 51-peak, 53-angle range, 54-angle range 61-measurement flow, 62-data processing flow, 63-angle Setting flow 100-listening room 101-front wall, 102-left wall, 103-rear wall, 104-right wall 105-virtual Lc , 106- virtual Rch, 107- virtual SLch
108-virtual SRch, 109-video device, L-listener

Claims (3)

A speaker array having a plurality of speakers arranged in a matrix or a line;
The audio signal of each channel of the multi-channel audio signal is input, the timing to output to each speaker of the speaker array is controlled, and the sound beam directed in the direction set for each signal of each channel is generated. Audio signal beam output means for outputting a signal obtained by synthesizing these sound beams;
A test sound beam output means for inputting a test sound and controlling a timing of outputting the sound to each speaker of the speaker array to output a test sound beam in a predetermined direction;
Test audio beam direction indicating means for instructing the test audio beam output means to turn a horizontal beam angle of the test audio beam within a predetermined angle range;
A sound collecting means that includes two microphones arranged side by side at the listening position, and that collects sound including the sound of the test sound beam and records a sound collecting signal;
The arrival time of the two sound pickup signals picked up by the two microphones is determined for each beam angle of the test sound beam, and whether the direction of arrival of the sound beam corresponding to the beam angle is the front of the microphone. Before and after judging means for judging whether it is backward,
A plurality of peak values in the picked-up signal recorded for each beam angle of the test sound beam are selected, and a beam angle corresponding to the selected peak is determined as a horizontal beam of the sound beam of the multi-channel audio signal. Candidate selection means for selecting as angle candidates;
The front-rear direction determined by the front-rear determination unit for the collected sound signal corresponding to the beam angle selected as the candidate is matched with the front-rear direction in which the input of the multi-channel audio signal assumes speaker output. A speaker array device comprising: setting means for setting a beam angle candidate as a beam angle of an audio beam of the multi-channel audio signal.   The speaker array device according to claim 1, wherein the front-rear determination unit determines the front and rear of the arrival time by calculating a cross-correlation function for the collected sound signals collected by the two microphones for each of the beam angles. A speaker array having a plurality of speakers arranged in a matrix or a line;
An audio beam of a multi-channel audio signal that is input to each channel of the multi-channel audio signal and is output in a direction set for each signal of each channel by controlling the timing of output to each speaker of the speaker array. And a multi-channel audio beam output means for outputting a signal obtained by synthesizing the audio beams of these multi-channel audio signals.
A test sound beam output step of inputting a test sound and controlling a timing of outputting to each speaker of the speaker array and outputting a test sound beam in a predetermined direction;
A test sound beam direction indicating step for instructing the test sound beam output means to turn a beam angle in a horizontal direction of the test sound beam within a predetermined angle range;
A sound collecting step of collecting sound of the test sound beam and recording a sound collecting signal for each beam angle of the test sound beam, using two microphones installed side by side at the listening position; and
The arrival time of the two sound pickup signals picked up by the two microphones is determined for each beam angle of the test sound beam, and the direction in which the test sound beam corresponding to the beam angle arrives is the microphone. Before and after determining whether the front or rear of the
A plurality of peak values in the picked-up signal recorded for each beam angle of the test sound beam are selected, and a beam angle corresponding to the selected peak is determined as a horizontal beam of the sound beam of the multi-channel audio signal. A candidate selection step for selecting as an angle candidate;
The front-rear direction determined by the front-rear determination unit for the two sound pickup signals corresponding to the beam angle selected as the candidate is matched with the front-rear direction in which the input of the multi-channel audio signal assumes a speaker output, A setting step of setting a selected beam angle candidate as a beam angle of an audio beam of the multi-channel audio signal, and an audio beam setting method for a speaker array device.

Images