JP5447199B2 - Sound processor - Google Patents

Description

  The present invention relates to a technique for generating a plurality of acoustic signals corresponding to different directivities.

  A technique for generating a plurality of surround sound signals (surround sound collection) has been proposed. For example, Patent Document 1 surrounds four sound signals using four unidirectional sound collecting devices directed in different directions and one omnidirectional sound collecting device. Technology is disclosed. It is also possible to generate multiple surround sound signals by sound source separation using known techniques such as independent component analysis (ICA) and computational auditory scene analysis (CASA). Is possible.

JP-A-5-191886

  However, in the technique of Patent Document 1, since five sound collecting devices are required to generate four types of surround sound signals, there is a problem that the apparatus necessary for surround sound collection is enlarged. In addition, in sound source separation using independent component analysis or computational auditory scene analysis, there is a problem that processing time (latency) and power consumption increase due to the complexity of arithmetic processing. In view of the above circumstances, an object of the present invention is to generate a plurality of surround sound signals by a simple process while reducing the total number of sound collecting devices.

  Means employed by the present invention to solve the above problems will be described below. In order to facilitate understanding of the present invention, in the following description, the correspondence between each element of the present invention and the element of each of the embodiments described later is indicated in parentheses, but the scope of the present invention is not limited to the embodiment. It is not intended to limit the configuration.

The acoustic processing apparatus of the present invention includes a unidirectional first acoustic signal (for example, an acoustic signal XC) directed forward and a bi-directional second acoustic signal (for example, an acoustic signal XR) directed to the right front and left rear. ) And a third bi-directional acoustic signal (e.g., an acoustic signal XL) directed to the left front and right rear, the right front signal in which the incoming sound from the right front is emphasized ( For example, the right front directing means (for example, the right front directing unit 32) that synthesizes the sound signal YR from the first sound signal and the second sound signal, and the left front signal (for example, the sound signal) in which the incoming sound from the left front is emphasized. YL) is synthesized from the first acoustic signal and the third acoustic signal, the left front directing means (for example, the left front directing unit 34), the incoming sound from the left front and the front are suppressed, and the incoming sound from the right rear is Enhanced right rear signal (eg acoustic signal YRS ) From the first acoustic signal and the third acoustic signal, the right rear directing means (for example, the right rear directing unit 42), and the incoming sound from the right front and front are suppressed and the incoming sound from the left rear is emphasized. Left rear directing means (for example, left rear directing unit 44) for synthesizing the left rear signal (for example, acoustic signal YLS) from the first acoustic signal and the second acoustic signal.
In the above configuration, the right front signal is synthesized from the first acoustic signal and the second acoustic signal, the left front signal is synthesized from the first acoustic signal and the third acoustic signal, and the first acoustic signal and A right rear signal is synthesized from the third acoustic signal and a left rear signal is synthesized from the first acoustic signal and the second acoustic signal. Therefore, there is an advantage that a surround type acoustic signal is generated by a simple process while reducing the total number of sound collecting devices.

  An acoustic signal having a specific directivity (unidirectional or bi-directional) α means an acoustic signal substantially equivalent to a sound collection signal in a sound collection device having directivity α, and has directivity α. In addition to the acoustic signal actually generated by the sound collection device, the sound signal to which the directivity α is given by the signal processing on the sound signal generated by the sound collection device having directivity other than the directivity α (including omnidirectionality) Is a concept that also includes A directional acoustic signal directed in a specific direction is also grasped as an acoustic signal in which an incoming sound from the direction is emphasized compared to an incoming sound from another direction. For example, the first acoustic signal corresponds to a signal that emphasizes the incoming sound from the front. Similarly, the second acoustic signal corresponds to a signal in which the incoming sound from the right front and the left rear is emphasized, and the third acoustic signal corresponds to a signal in which the incoming sound from the left front and the right rear is emphasized.

  In addition, when the polarity of the 1st acoustic signal with respect to the incoming sound from the front and the polarity of the 2nd acoustic signal with respect to the incoming sound from the right front are the same polarity, the right front directing means has the first acoustic signal and the second acoustic signal. The right front signal can be generated by summing with the signal. On the other hand, when the polarity of the first acoustic signal with respect to the incoming sound from the front and the polarity of the second acoustic signal with respect to the incoming sound from the right front are opposite, the right front directing means has the first acoustic signal and the second acoustic signal. A right forward signal may be generated by subtraction between the signals. The same applies to the generation of the left front signal. In addition, the process of synthesizing the right rear signal from the first acoustic signal and the third acoustic signal and the process of synthesizing the left rear signal from the first acoustic signal and the second acoustic signal are also performed in the direction of sound arrival and the acoustic signal. It can be appropriately selected according to the polarity.

  The sound processing apparatus according to a preferred aspect of the present invention includes coefficient setting means (for example, a coefficient setting unit 70) that generates a processing coefficient sequence in which coefficient values for each frequency are set so that incoming sound from the front is suppressed. The right rear directing means includes a first acoustic signal and a third acoustic signal that synthesizes a fourth acoustic signal (for example, an acoustic signal RA) in which the incoming sound from the left front is suppressed among the third acoustic signals. Combining means (for example, the combining unit 52R) and first suppression processing unit (for example, the suppressing processing unit 64R) for generating the right rear signal by applying each coefficient value of the processing coefficient sequence to each frequency component of the fourth acoustic signal. And the left rear directing means synthesizes the fifth acoustic signal (for example, the acoustic signal LA) in which the incoming sound from the right front of the second acoustic signal is suppressed from the first acoustic signal and the second acoustic signal. Means (for example, the combining unit 52L) and each of the fifth acoustic signals By the action of each coefficient value of the processing coefficient sequence of wavenumber component and a second suppression processing means for generating a left rear signal (e.g. suppression processing unit 64L). In the above aspect, the component of the incoming sound from the front (residual component) remaining in the fourth acoustic signal and the fifth acoustic signal is suppressed by the first suppression processing means and the second suppression processing means. Compared to a configuration that does not suppress the component (a configuration in which the component of the incoming sound from the front is mixed in the right rear signal or the left rear signal), the listener can clearly perceive the sense of localization of the sound image behind the listener. Is possible.

  In a preferred aspect of the present invention, the coefficient setting means includes sum component generation means (for example, a sum component generation unit 72) for generating a sum component by adding the spectra of the fourth acoustic signal and the fifth acoustic signal, Difference component generation means (for example, difference component generation unit 74) that generates a difference component in which the incoming sound is suppressed by subtracting the spectrum of the fourth acoustic signal and the fifth acoustic signal, and a processing coefficient sequence from the sum component and the difference component Coefficient sequence generation means (for example, coefficient sequence generation unit 76) to be generated. In the above aspect, since the processing coefficient sequence is generated by addition and subtraction of the spectrum in the frequency domain, there is an advantage that the residual components in the fourth acoustic signal and the fifth acoustic signal can be effectively suppressed by simple processing.

  In a preferred aspect of the present invention, the right rear directing means includes spectral subtraction for subtracting the spectrum of the first acoustic signal from the spectrum of the third acoustic signal, and flooring for setting a component value that becomes a negative number in the spectral subtraction to a predetermined value. The left rear direction means performs spectral subtraction for subtracting the spectrum of the first acoustic signal from the spectrum of the second acoustic signal, and a component value that becomes a negative number in the spectral subtraction is a predetermined value. The flooring process set to is executed to generate a left rear signal. In the above aspect, since the residual component that can be generated by spectrum subtraction is suppressed by the flooring process, the localization of the sound image behind the listener is clearly shown to the listener as compared with the configuration that does not suppress the residual component. It is possible to perceive. In addition, the specific example of the above aspect is explained in full detail as 3rd Embodiment, for example.

  The sound processing apparatus according to a preferred aspect of the present invention includes a first sound signal based on a first original signal, a second original signal, and a third original signal generated by omnidirectional sound pickup devices installed at different positions. And directivity providing means (for example, directivity providing unit 80) for generating the second acoustic signal and the third acoustic signal. In the above aspect, the surround-type sound signal is generated from the sound signals (first original signal, second original signal, and third original signal) generated by the non-directional sound collecting device. There is an advantage that the cost required for realizing the sound processing device is reduced as compared with a configuration using a sound collecting device of one directivity or bi-directional property.

  The acoustic processing device according to each of the above aspects is realized by hardware (electronic circuit) such as a DSP (Digital Signal Processor) dedicated to processing of an acoustic signal, or a general-purpose calculation such as a CPU (Central Processing Unit). It is also realized by cooperation between the processing device and a program (software). The program of the present invention is a program that emphasizes the incoming sound from the right front from the unidirectional first acoustic signal directed forward and the bi-directional second acoustic signal directed right front and rear left. A left front signal in which the incoming sound from the left front is emphasized is synthesized from the right front directing process for combining the front signal and the first acoustic signal and the bi-directional third acoustic signal directed to the left front and right rear. Left front directing processing, and right rear pointing that combines the first and third acoustic signals with a right rear signal in which incoming sounds from the left front and front are suppressed and an incoming sound from the right rear is emphasized. Processing and left rear pointing processing for synthesizing a left rear signal in which the incoming sound from the front right and the front is suppressed and the incoming sound from the left rear is emphasized from the first acoustic signal and the second acoustic signal. To run. According to the above program, the same operation and effect as the sound processing apparatus of the present invention are realized. The program of the present invention is provided to a user in a form stored in a computer-readable recording medium and installed in the computer, or provided from a server device in a form of distribution via a communication network and installed in the computer. Is done.

1 is a block diagram of a sound processing apparatus according to a first embodiment. It is explanatory drawing of the position and directional characteristic of each sound collection apparatus. It is a schematic diagram of a reproduction system. It is a block diagram of a front signal generation part. It is explanatory drawing of operation | movement of a front signal generation part. It is a block diagram of a back signal generation part. It is explanatory drawing of operation | movement of a back signal production | generation part. It is a block diagram of a coefficient setting part. It is a block diagram of the sound processing apparatus which concerns on 2nd Embodiment. It is explanatory drawing of operation | movement of the position of each sound collection apparatus, and a directivity provision part. It is a block diagram of a directivity provision part. It is explanatory drawing of operation | movement of a directivity provision part. It is a block diagram of the back signal generation part in a 3rd embodiment.

<A: First Embodiment>
FIG. 1 is a block diagram of a sound processing apparatus 100A according to the first embodiment of the present invention. As shown in FIG. 1, three sound collection devices 12 (12C, 12R, 12L) are connected to the sound processing apparatus 100A. The sound collecting device 12C generates a time domain acoustic signal XC representing the waveform of surrounding sounds (speech and music). Similarly, the sound collection device 12R generates an acoustic signal XR, and the sound collection device 12L generates an acoustic signal XL.

  FIG. 2 is a schematic diagram showing the arrangement and directivity characteristics of each sound collecting device 12. As shown in FIG. 2, one side along a predetermined reference line DC viewed from an arbitrary position in the horizontal plane is expressed as “front” and the opposite side is expressed as “rear”. Further, when viewed from an arbitrary position in the horizontal plane, one side along the perpendicular of the reference line DC is denoted as “left” and the opposite side is denoted as “right”.

  The sound collection device 12C is a unidirectional microphone that is installed in front of the reference point P0 on the reference line DC and is directed forward. That is, the sound collection device 12C generates an acoustic signal (unidirectional acoustic signal) XC in which the incoming sound from the front is emphasized compared to the incoming sound from the other direction. In the following, the polarity of the acoustic signal XC with respect to the incoming sound from the front is defined as positive polarity (positive phase) for convenience.

  The sound collection device 12R is installed in front of the reference point P0 (between the front and the right), and the sound collection device 12L is installed in front of the reference point P0 (between the front and the left). Specifically, the sound collecting device 12R is installed on a straight line DR that forms an angle θR (acute angle) clockwise with respect to the reference line DC, and the angle θL (acute angle) is counterclockwise with respect to the reference line DC. A sound collecting device 12L is installed on the straight line DL formed. The sound collecting device 12R and the sound collecting device 12L are in positions symmetrical with respect to the reference line DC.

  The sound collection device 12R is a bidirectional microphone whose directivity axis is parallel to the straight line DR. That is, the sound collection device 12R generates an acoustic signal (bidirectional acoustic signal) XR in which incoming sounds from the right front and left rear are emphasized compared to incoming sounds from other directions. The polarity of the acoustic signal XR is the same polarity (positive polarity) as that of the acoustic signal XC with respect to the incoming sound from the right front, and the opposite polarity (negative polarity) of the acoustic signal XC with respect to the incoming sound from the left rear. .

  The sound collection device 12L is a bidirectional microphone whose directional axis is parallel to the straight line DL. That is, the sound collection device 12L generates the acoustic signal XL in which the incoming sounds from the left front and the right rear are emphasized as compared with the incoming sounds from the other directions. The polarity of the acoustic signal XL is the same polarity (positive polarity) as the acoustic signal XC with respect to the incoming sound from the left front, and the opposite polarity (negative polarity) with respect to the incoming sound from the right rear. .

  The acoustic processing apparatus 100A in FIG. 1 is a signal process for three systems of acoustic signals X (XC, XR, XL) generated in parallel by three sound collection devices 12 (12C, 12R, 12L), and has different directivities. 5 (ie, surround format) acoustic signals Y (YC, YR, YL, YRS, YLS) are generated. The acoustic signal YC is a signal in which the incoming sound from the center with respect to the reference point P0 is emphasized as compared with the incoming sound from the other direction. In the first embodiment, the acoustic signal XC generated by the sound collection device 12C is output as the acoustic signal YC. However, a configuration in which the acoustic signal YC is generated by a predetermined process (for example, amplification) on the acoustic signal XC may be employed.

  The acoustic signal YR is a signal in which the incoming sound from the right front with respect to the reference point P0 is emphasized, and the acoustic signal YL is a signal in which the incoming sound from the left front with respect to the reference point P0 is emphasized. The acoustic signal YRS is a signal in which the incoming sound from the right rear (Right-Surround) with respect to the reference point P0 is emphasized, and the acoustic signal YLS is the incoming sound from the left rear (Left-Surround) with respect to the reference point P0. Is an emphasized signal.

  FIG. 3 is an explanatory diagram of an acoustic system that reproduces the acoustic signal Y (YC, YR, YL, YRS, YLS) generated by the acoustic processing apparatus 100A. As shown in FIG. 3, the acoustic system includes five sound emitting devices 16 (16C, 16R, 16L, 16RS, and 16LS) arranged so as to surround the listener A. Each acoustic signal Y generated by the acoustic processing apparatus 100A is supplied in parallel to the five sound emitting devices 16 and reproduced as sound waves. The sound emitting device 16C disposed in front of the listener A reproduces a sound wave corresponding to the acoustic signal YC. Similarly, the sound emitting device 16R on the right front of the listener A reproduces the acoustic signal YR, and the sound emitting device 16L on the left front of the listener A reproduces the acoustic signal YL. The sound emitting device 16RS on the right rear side of the listener A reproduces the acoustic signal YRS, and the sound emitting device 16LS on the left rear side of the listener A reproduces the acoustic signal YLS. Therefore, it is possible to make the listener A perceive a feeling of localization such that sound comes from a virtual sound source (sound image) at an arbitrary position around the listener A.

  As shown in FIG. 1, the sound processing device 100 </ b> A includes an arithmetic processing device 22 and a storage device 24. The storage device 24 stores a program PGM executed by the arithmetic processing device 22 and data used by the arithmetic processing device 22. A known recording medium such as a semiconductor recording medium or a magnetic recording medium or a combination of a plurality of types of recording media is arbitrarily adopted as the storage device 24. The arithmetic processing unit 22 executes the program PGM stored in the storage device 24 to thereby convert the three acoustic signals X (XC, XR, XL) to the five acoustic signals Y (YC, YR, YL, YRS, A plurality of functions (forward signal generation unit 30 and rear signal generation unit 40A) for generating (YLS) are realized. A configuration in which each function of the arithmetic processing unit 22 is distributed over a plurality of integrated circuits, or a configuration in which a dedicated electronic circuit (DSP) realizes each function may be employed.

  The front signal generation unit 30 generates an acoustic signal YR and an acoustic signal YL from the acoustic signal X (XC, XR, XL). As shown in FIG. 1, the front signal generation unit 30 includes a right front directing unit 32 and a left front directing unit 34. The right front pointing unit 32 combines the acoustic signal YR from the acoustic signal XC and the acoustic signal XR, and the left front pointing unit 34 combines the acoustic signal YL from the acoustic signal XC and the acoustic signal XL.

  FIG. 4 is a block diagram of the front signal generation unit 30, and FIG. 5 is an explanatory diagram of the operation of the front signal generation unit 30. As shown in FIG. 4, the right front directing unit 32 includes a combining unit 36R and an adjusting unit 38R. The synthesizer 36R adds the acoustic signal XC and the acoustic signal XR. The adjustment unit 38R adjusts the amplitude of the output signal of the synthesis unit 36R to generate the acoustic signal YR. A multiplier that multiplies a predetermined coefficient (for example, 0.5) is preferably used as the adjustment unit 38R. As understood from FIG. 5, the component corresponding to the incoming sound from the right front of the acoustic signal XR and the acoustic signal XC have the same polarity. Therefore, the calculation includes addition of the acoustic signal XC and the acoustic signal XR. In the generated acoustic signal YR, the incoming sound from the right front with respect to the reference point P0 is emphasized compared to the incoming sound from the other direction.

  The left front directing unit 34 in FIG. 4, similarly to the right front directing unit 32, includes a combining unit 36L that adds the acoustic signal XC and the acoustic signal XL, and an adjusting unit 38L (adjusting the amplitude of the output signal of the combining unit 36L). For example, a multiplier for multiplying a predetermined coefficient). Therefore, as can be understood from FIG. 5, the acoustic signal YL generated by the left front directing unit 34 is a signal in which the incoming sound from the left front relative to the reference point P0 is emphasized as compared with the incoming sound from the other direction. Become. In the right front directing unit 32 and the left front directing unit 34, a configuration in which the front and rear of the signal addition by the combining unit (36R, 36L) and the amplitude adjustment by the adjusting unit (38R, 38L) are reversed is also employed. obtain.

  The rear signal generator 40A in FIG. 1 generates the acoustic signal YRS and the acoustic signal YLS from the acoustic signal X (XC, XR, XL). FIG. 6 is a block diagram of the rear signal generation unit 40A. As shown in FIG. 6, the backward signal generation unit 40A includes a combining unit 52R, a combining unit 52L, a component suppressing unit 54, an adjusting unit 56R, and an adjusting unit 56L.

  FIG. 7 is an explanatory diagram of the operations of the combining unit 52R and the combining unit 52L. The synthesizer 52R in FIG. 6 generates the acoustic signal RA by synthesizing the acoustic signal XC and the acoustic signal XL. Specifically, the synthesizer 52R is an adder that generates an acoustic signal RA (RA = XC−XL) by reverse phase addition (subtraction) between the acoustic signal XC and the acoustic signal XL. The component of the incoming sound from the left front of the acoustic signal XL has the same polarity (positive polarity) as the acoustic signal XC, and the component of the incoming sound from the right rear of the acoustic signal XL has the opposite polarity (negative polarity) to the acoustic signal XC. ). Therefore, as can be understood from FIG. 7, in the acoustic signal RA generated by the reverse phase addition of the acoustic signal XC and the acoustic signal XL, the component of the incoming sound from the left front (acoustic signal XC) of the acoustic signal XL. And the component of the incoming sound from the right rear is emphasized.

  On the other hand, the synthesizer 52L in FIG. 6 generates the acoustic signal LA by synthesizing the acoustic signal XC and the acoustic signal XR. Specifically, the synthesizing unit 52L generates an acoustic signal LA (LA = XC−XR) by reverse phase addition (subtraction) between the acoustic signal XC and the acoustic signal XR. Therefore, as can be understood from FIG. 7, in the acoustic signal LA, the incoming sound component from the right front (the overlapping component with the acoustic signal XC) of the acoustic signal XR is suppressed and the incoming sound from the left rear is suppressed. The components of are emphasized.

  By the way, as understood from FIG. 7, the acoustic signal RA generated by the synthesis unit 52R includes a component of the incoming sound from the front which is not included in the acoustic signal XL among the acoustic signal XC. Similarly, a component of the incoming sound from the front that is not included in the acoustic signal XR of the acoustic signal XC remains in the acoustic signal LA generated by the synthesis unit 52L. The component suppression unit 54 in FIG. 6 generates the acoustic signal RD and the acoustic signal LD by suppressing the components of the incoming sound from the front (hereinafter collectively referred to as “residual components”) in each of the acoustic signal RA and the acoustic signal LA. To do.

  As shown in FIG. 6, the component suppression unit 54 includes a frequency analysis unit 62, a signal processing unit 64, a waveform synthesis unit 66, and a coefficient setting unit 70. The frequency analysis unit 62 sequentially generates a frequency spectrum (complex spectrum) RB of the acoustic signal RA and a frequency spectrum (complex spectrum) LB of the acoustic signal LA for each unit period (frame) on the time axis. The frequency spectrum RB is a series of K frequency components RB [1] to RB [K] corresponding to different frequencies. Similarly, the frequency spectrum LB is a series of K frequency components LB [1] to LB [K]. For the generation of the frequency spectrum RB and the frequency spectrum LB, a known frequency analysis such as a short-time Fourier transform can be arbitrarily employed.

  The coefficient setting unit 70 generates a processing coefficient sequence GS for suppressing the residual components of the acoustic signal RA and the acoustic signal LA. The processing coefficient sequence GS is a sequence of K coefficient values gS [1] to gS [K] corresponding to different frequencies, and is sequentially generated for each unit period using the frequency spectrum RB and the frequency spectrum LB. Is done. The coefficient value gS [k] (k = 1 to K) corresponds to a gain (spectrum gain) with respect to the frequency component RB [k] of the acoustic signal RA and the frequency component LB [k] of the acoustic signal LA, and is 0 or more and 1 It is set within the following range. Specifically, the coefficient values gS [1] to gS [K] of the processing coefficient sequence GS are closer to zero as the coefficient value gS [k] of the frequency having a larger residual component power (amplitude) becomes closer to zero. The coefficient value gS [k] at a frequency with a lower power is a value closer to 1. A specific method for generating the processing coefficient sequence GS will be described later.

  The signal processing unit 64 causes the processing coefficient sequence GS generated by the coefficient setting unit 70 to sequentially act on the frequency spectrum RB and the frequency spectrum LB in each unit period. As shown in FIG. 6, the signal processing unit 64 generates a frequency spectrum RC (frequency components RC [1] to RC [K]) by applying a processing coefficient sequence GS to the frequency spectrum RB; And a suppression processing unit 64L that generates a frequency spectrum LC (frequency components LC [1] to LC [K]) by applying a processing coefficient sequence GS to the frequency spectrum LB.

Specifically, as shown in the following formula (1A), the suppression processing unit 64R multiplies the frequency component RB [k] of the frequency spectrum RB and the coefficient value gS [k] of the processing coefficient sequence GS to obtain the frequency spectrum. The frequency component RC [k] of RC is calculated. Similarly, as shown in the following formula (1B), the suppression processing unit 64L multiplies the frequency component LB [k] of the frequency spectrum LB and the coefficient value gS [k] of the processing coefficient sequence GS to generate the frequency spectrum LC. The frequency component LC [k] is calculated.

  6 generates a time-domain acoustic signal RD and acoustic signal LD from the frequency spectrum RC and frequency spectrum LC generated by the signal processing unit 64. Specifically, the waveform synthesizer 66 converts the frequency spectrum RC for each unit period into a time domain signal by inverse Fourier transform, and generates an acoustic signal RD by connecting the unit periods before and after. Similarly, the waveform synthesizer 66 generates an acoustic signal LD from each frequency spectrum LC.

  The adjusting unit 56R in FIG. 6 generates the acoustic signal YRS by adjusting the amplitude of the acoustic signal RD generated by the component suppressing unit 54. For example, a multiplier that amplifies the acoustic signal RD by multiplication of a predetermined value can be employed as the adjustment unit 56R. Similarly, the adjustment unit 56L (for example, a multiplier) in FIG. 6 generates the acoustic signal YLS by adjusting the amplitude of the acoustic signal LD generated by the component suppression unit 54.

  As can be understood from the above description, the incoming sound from the right rear is emphasized in the acoustic signal YRS, and the incoming sound from the left rear is emphasized in the acoustic signal YLS. That is, the synthesis unit 52R, the frequency analysis unit 62, the suppression processing unit 64R, the waveform synthesis unit 66, and the adjustment unit 56R function as the right rear pointing unit 42 that synthesizes the acoustic signal YRS from the acoustic signal XC and the acoustic signal XL. The synthesizing unit 52L, the frequency analyzing unit 62, the suppression processing unit 64L, the waveform synthesizing unit 66, and the adjusting unit 56L function as the left rear directing unit 44 that synthesizes the acoustic signal YLS from the acoustic signal XC and the acoustic signal XR.

Next, a specific configuration of the coefficient setting unit 70 will be described with reference to FIG. As shown in FIG. 8, the coefficient setting unit 70 includes a sum component generation unit 72, a difference component generation unit 74, and a coefficient sequence generation unit 76. The sum component generator 72 sequentially adds the sum component M (K) for each unit period by adding the frequency spectrum RB of the acoustic signal RA and the frequency spectrum LB of the acoustic signal RA as expressed by the following formula (2). A complex spectrum represented by frequency components M [1] to M [K].

The difference component generation unit 74 in FIG. 8 sequentially subtracts the difference component S (K frequency components S [1]) for each unit period by subtraction between the frequency spectrum RB of the acoustic signal RA and the frequency spectrum LB of the acoustic signal LA. ~ Complex spectrum represented by S [K]. Specifically, the difference component generation unit 74 calculates the frequency components S [1] to S [K] by calculation (weighted subtraction) of the following formula (3).

  Symbol α in Equation (3) is a predetermined coefficient (0 ≦ α ≦ 1). As understood from the equation (3), the frequency component S [k] of the difference component S is subtracted from the frequency component RB [k] from the frequency component LB [k] by a ratio (weighted value) according to the coefficient α. Is calculated. Therefore, the difference component S is a signal obtained by suppressing the component in the direction corresponding to the coefficient α relative to the component in the other direction in the acoustic signal RA and the acoustic signal LA. In the present embodiment, the numerical value of the coefficient α is set so that the residual component (arrival sound from the front) that is commonly included in the acoustic signal RA and the acoustic signal LA is suppressed by the difference component S (for example, α = 0.5). The reason why the expression (3) includes division by the maximum value max (α, 1-α) is that the intensity of the frequency component S [k] is equal to the frequency component RB [k] and the frequency component LB [k]. It is for maintaining.

  The coefficient sequence generation unit 76 in FIG. 8 sequentially generates a processing coefficient sequence GS for each unit period from the sum component M generated by the sum component generation unit 72 and the difference component S generated by the difference component generation unit 74. As shown in FIG. 8, the coefficient sequence generation unit 76 includes a first generation unit 762 that generates a processing coefficient sequence GE (a sequence of K coefficient values gE [1] to gE [K]), and a processing coefficient sequence GE. And a second generation unit 764 that generates a processing coefficient sequence GS (coefficient values gS [1] to gS [K]).

数式(4)の変数（パワー）Ｐ[k]は、例えば以下の数式(5A)および数式(5B)で算定される。

The first generation unit 762 calculates each coefficient value gE [k] of the processing coefficient sequence GE by the following equation (4). The reason why the equation (4) includes division by the amplitude | M [k] | is to normalize the coefficient value gE [k] to a numerical value of 1 or less (0 ≦ gE [k] ≦ 1). .

The variable (power) P [k] in Expression (4) is calculated by, for example, Expression (5A) and Expression (5B) below.

As understood from the equation (5A), the numerical value of the variable P [k] in the equation (4) is the power of the sum component M | M [k] | ² to the power of the difference component S | S [k] | ² Is set to the value obtained by subtracting (spectrum subtraction). However, the value of the variable P [k] at a frequency at which the power | M [k] | ² of the sum component M is equal to or less than the power | S [k] | ^{2 of the} difference component S is the power | M [ k] | ² is set to a product of a predetermined coefficient (flooring coefficient) β (flooring process).

  Since the frequency component S [k] of the difference component S is a component in which the residual component is suppressed, the series of variables P [1] to P [K] calculated by the equations (5A) and (5B) is a sum component. This corresponds to a power spectrum in which residual components included in M are emphasized. Therefore, the coefficient values gE [1] to gE [K] of the processing coefficient sequence GE are closer to 1 as the coefficient value gE [k] of the frequency having a larger residual component power (amplitude), and the residual component power is smaller. The frequency coefficient value gE [k] is closer to zero.

数式(6)から理解されるように、第２生成部７６４が生成する処理係数列ＧSの係数値ｇS[1]〜ｇS[K]は、残留成分の強度（パワー）が大きい周波数の係数値ｇS[k]ほどゼロに近い数値となり、残留成分の強度（パワー）が小さい周波数の係数値ｇS[k]ほど１に近い数値となる。したがって、周波数スペクトルＲBおよび周波数スペクトルＬBの各々に信号処理部６４が処理係数列ＧSを乗算することで、前方からの残留成分を抑圧して右後方からの到来音を強調した音響信号ＹRS（ＲD）と、前方からの残留成分を抑圧して左後方からの到来音を強調した音響信号ＹLS（ＬD）とが生成される。 The second generation unit 764 in FIG. 8 generates a processing coefficient sequence GS for suppressing residual components using the processing coefficient sequence GE generated by the first generation unit 762. Specifically, the second generation unit 764 subtracts each coefficient value gE [k] of the processing coefficient sequence GE from a predetermined value (1 in this embodiment) as defined by the following formula (6). Thus, each coefficient value gS [k] of the processing coefficient sequence GS is calculated.

As understood from the equation (6), the coefficient values gS [1] to gS [K] of the processing coefficient sequence GS generated by the second generation unit 764 are the coefficient values of the frequencies having a large residual component strength (power). The value gS [k] is closer to zero and the frequency coefficient value gS [k] having a smaller residual component strength (power) is closer to 1. Accordingly, the signal processor 64 multiplies each of the frequency spectrum RB and the frequency spectrum LB by the processing coefficient sequence GS, thereby suppressing the residual component from the front and enhancing the incoming sound from the right rear by the acoustic signal YRS (RD ), And an acoustic signal YLS (LD) that suppresses the residual component from the front and emphasizes the incoming sound from the left rear.

  According to the first embodiment described above, five surround sound signals Y (YC, YR, YL, YRS, YLS) are processed by signal processing for three sound signals X (XC, XR, XL). Since it is generated, for example, it is possible to reduce the size of the acoustic processing apparatus 100A as compared with the technique of Patent Document 1. In addition, a plurality of acoustic signals Y are generated by a very simple process compared to sound source separation using independent component analysis (ICA) or computational auditory scene analysis (CASA). Therefore, there is an advantage that processing time (latency) and power consumption are reduced. In the computational auditory scene analysis using the acoustic harmonic structure, the target of sound source separation is limited to the acoustic signal having the harmonic structure, but the first embodiment does not assume the harmonic structure of the acoustic signal X. According to the above, there is an advantage that the acoustic signal X can be processed regardless of the presence or absence of the harmonic structure.

  In the first embodiment, the acoustic signal YRS is generated by suppressing the residual component coming from the front in the acoustic signal RA synthesized from the acoustic signal XC and the acoustic signal XL by the suppression processing unit 64R. The acoustic signal YLS is generated by suppressing the residual component coming from the front in the acoustic signal LA synthesized from the acoustic signal XC and the acoustic signal XR by the suppression processing unit 64L. Therefore, it is possible to make the listener A clearly perceive the sense of localization of the sound image of the sound source located behind the reference point P0 (left rear, right rear) as compared with the configuration in which the residual component is not suppressed.

<B: Second Embodiment>
A second embodiment of the present invention will be described. In the first embodiment, the directivity (unidirectional, bi-directional) sound collecting device 12 is used, but in the second embodiment, five non-directional sound collecting devices 12 are used. An acoustic signal Y (YC, YR, YL, YRS, YLS) is generated. In addition, about the element which an effect | action and function are equivalent to 1st Embodiment in each following illustration, the code | symbol referred by the above description is diverted and each detailed description is abbreviate | omitted suitably.

  FIG. 9 is a block diagram of the sound processing apparatus 100B of the second embodiment. As shown in FIG. 9, three omnidirectional sound collecting devices 12 (12C, 12R, 12L) are connected to the sound processing device 100B. The sound collecting device 12C generates an acoustic signal VC, the sound collecting device 12R generates an acoustic signal VR, and the sound collecting device 12L generates an acoustic signal VL.

  FIG. 10 is a schematic diagram showing the position of each sound collecting device 12. As shown in FIG. 10, the sound collecting device 12C is installed at a reference point P0 on the reference line DC. The sound collection device 12R is installed on a straight line DR that forms an angle θR (acute angle) with respect to the reference line DC, and the sound collection device 12L has an angle θL (| θL | = | θR with respect to the reference line DC. |) On the straight line DL. The sound collecting device 12R and the sound collecting device 12L are in positions symmetrical with respect to the reference line DC.

  As shown in FIG. 9, the sound processing device 100B of the second embodiment has a configuration in which a directivity imparting unit 80 is added to the sound processing device 100A of the first embodiment. The directivity imparting unit 80 generates an acoustic signal XC, an acoustic signal XR, and an acoustic signal XL by signal processing (directivity imparting) on the acoustic signal VC, the acoustic signal VR, and the acoustic signal VL. As in the first embodiment, the acoustic signal XC is a signal (unidirectional signal) in which the incoming sound from the front is emphasized. The acoustic signal XR is a bidirectional signal in which incoming sounds from the right front and left rear are emphasized, and the acoustic signal XL is a bidirectional signal in which incoming sounds from the left front and right rear are emphasized. Signal.

  FIG. 11 is a block diagram of the directivity providing unit 80. As shown in FIG. 11, the directivity imparting unit 80 includes a bidirectional directivity imparting unit 82R, a bidirectional directivity imparting unit 82L, and a single directivity imparting unit 84. The bi-directional imparting unit 82R generates a bi-directional acoustic signal XR from the acoustic signal VC and the acoustic signal VR. Specifically, the bidirectional directing unit 82R generates the acoustic signal XR by subtracting the acoustic signal VC from the acoustic signal VR. When the acoustic signal VC is subtracted from the acoustic signal VR, as shown by the solid line in FIG. 10, the incoming sound from the direction of the perpendicular bisector DVR between the sound collecting device 12R and the sound collecting device 12C is the incoming sound from the other direction. It is suppressed compared with. That is, a dead angle of sound collection is formed in the direction of the vertical bisector DVR. Therefore, the acoustic signal XR generated by the bidirectional directing unit 82R is a bidirectional signal (ie, from the right front and the left rear) in which the directivity axis forms an angle θR with respect to the reference line DC, as in the first embodiment. Signal). Since the acoustic signal VC is subtracted from the acoustic signal VR, the acoustic signal XR becomes positive with respect to the incoming sound from the right front as in the first embodiment.

  Similar to the bidirectional directing unit 82R, the bidirectional directing unit 82L in FIG. 11 generates the acoustic signal XL by subtracting the acoustic signal VC from the acoustic signal VL. By subtracting the sound signal VC from the sound signal VL, a sound collection dead angle is formed in the direction of the perpendicular bisector DVL between the sound collection device 12L and the sound collection device 12C. Therefore, the acoustic signal XL formed by the bidirectional directing unit 82L is a signal in which incoming sounds from the left front and right rear are emphasized, as in the first embodiment.

  The unidirectional imparting unit 84 generates a unidirectional acoustic signal XC directed forward. As shown in FIG. 11, the unidirectional assignment unit 84 includes a combining unit 842, a delay unit 844, and a combining unit 846. FIG. 12 is an explanatory diagram of the operation of the unidirectional imparting unit 84. The synthesizer 842 generates an acoustic signal VRL by adding the acoustic signal VR and the acoustic signal VL. Since the sound collection device 12R and the sound collection device 12L are omnidirectional, the acoustic signal VRL generated by the synthesis unit 842 is generated between the sound collection device 12R and the sound collection device 12L as shown by the chain line in FIG. It approximates to a signal generated by a virtual omnidirectional sound pickup device 12RL located at a point.

  The delay unit 844 in FIG. 11 generates an acoustic signal VC_t obtained by delaying the acoustic signal VC by the delay amount t. The delay amount t is the time length (t = d / c) required for the incoming sound (speed c) from the rear to propagate the distance d from the sound collection device 12C to the virtual sound collection device 12RL in FIG. Set to The synthesis unit 846 in FIG. 11 generates the acoustic signal XC by subtracting the acoustic signal VC_t delayed by the delay unit 844 from the acoustic signal VRL generated by the synthesis unit 842. Since the components of the incoming sound from the rear are in phase with the acoustic signal VRL and the acoustic signal VC_t, they are canceled out by subtraction by the synthesis unit 846. That is, a blind spot for sound collection is formed. Therefore, as indicated by a solid line in FIG. 12, the acoustic signal XC is a unidirectional signal in which the incoming sound from the front is emphasized.

  The front signal generation unit 30 and the rear signal generation unit 40A in FIG. 9 process the three systems of acoustic signals X (XC, XR, XL) generated by the directivity providing unit 80 by the above method. The contents of the processing by the front signal generation unit 30 and the rear signal generation unit 40A are the same as those in the first embodiment. Therefore, the same effects as those of the first embodiment are realized in the second embodiment. In the second embodiment, since the omnidirectional sound collecting device 12 (12C, 12R, 12L) can be used, the unidirectional sound collecting device 12C, the bi-directional sound collecting device 12R, and the sound collecting device are used. Compared to the first embodiment that requires the device 12L, there is an advantage that the cost required to realize the sound processing apparatus 100B can be reduced.

<C: Third Embodiment>
The acoustic processing device 100C of the third embodiment has a configuration in which the rear signal generation unit 40A of the first embodiment is replaced with the rear signal generation unit 40B of FIG. As shown in FIG. 13, the backward signal generation unit 40B includes a frequency analysis unit 92, a subtraction processing unit 94R, a subtraction processing unit 94L, and a waveform synthesis unit 96. The frequency analysis unit 92 includes a frequency spectrum QC (frequency components QC [1] to QC [K]) of the acoustic signal XC, a frequency spectrum QR (frequency components QR [1] to QR [K]) of the acoustic signal XR, and the acoustic signal. The XL frequency spectrum QL (frequency components QL [1] to QL [K]) is sequentially generated for each unit period by the same processing as the frequency analysis unit 62.

  The subtraction processing unit 94R generates the frequency spectrum QRS (frequency components QRS [1] to QRS [K]) of the acoustic signal YRS from the frequency spectrum QC of the acoustic signal XC and the frequency spectrum QL of the acoustic signal XL. The subtraction processing unit 94L generates a frequency spectrum QLS (frequency components QLS [1] to QLS [K]) of the acoustic signal YLS from the frequency spectrum QC of the acoustic signal XC and the frequency spectrum QR of the acoustic signal XR. The waveform synthesis unit 96 synthesizes the time domain acoustic signal YRS from the frequency spectrum QRS sequentially generated by the subtraction processing unit 94R, and generates the time domain acoustic signal YLS from the frequency spectrum QLS sequentially generated by the subtraction processing unit 94L. To do. The processing of the subtraction processing unit 94R and the subtraction processing unit 94L will be described in detail below.

The frequency component QRS [k] of the frequency spectrum QRS generated by the subtraction processing unit 94R is expressed by the following formula (7).

The symbol j in Equation (7) means an imaginary unit, and the symbol θR [k] means the phase angle of the acoustic signal XL. The subtraction processing unit 94R calculates the power PRS [k] (PRS [k] = | QRS [k] | ² ) of the equation (7) by the calculation of the following equations (8A) and (8B).

As understood from the equation (8A), the power PRS [k] of the acoustic signal YRS subtracts the power | QC [k] | ² of the acoustic signal XC from the power | QL [k] | ² of the acoustic signal XL. Calculated by processing (spectrum subtraction). However, the power PRS [k] of a frequency that becomes a negative number in spectral subtraction is set to a product of the power | QL [k] | ² of the acoustic signal XL and a predetermined coefficient (for example, a positive number of 1 or less) β. (Flooring process).

The component of the incoming sound from the left front of the acoustic signal XL (the overlapping component with the acoustic signal XC) is suppressed by the spectral subtraction of Equation (8A). Assuming that the frequency of the residual component (arrival sound from the front) of the acoustic signal XC is different from that of the acoustic signal XL, the power | QC [k] | ² of the residual component is the power | QL [k] | ^Since it exceeds ² (| QL [k] | ² <| QC [k] | ² ), the power PRS [k] at the frequency where the residual component (arrival sound from the front) is present in the frequency spectrum QRS The predetermined value (β | QL [k] | ² ) is set in the flooring process (8B). Therefore, the residual component of the acoustic signal XC is suppressed in the frequency spectrum QRS. As can be understood from the above description, the frequency analysis unit 92, the subtraction processing unit 94R, and the waveform synthesis unit 96 suppress the incoming sound from the left front and the front (residual component) and also the incoming sound from the right rear. Functions as the right rear directing unit 42 for generating the acoustic signal YRS in which is emphasized.

Similar to the subtraction processing unit 94R, the subtraction processing unit 94L subtracts the power | QC [k] | ² of the acoustic signal XC from the power | QR [k] | ² of the acoustic signal XR to thereby reduce the power PLS of the acoustic signal YLS. By executing spectral subtraction for calculating [k] and flooring processing for setting power PLS [k] of a frequency that becomes a negative number by spectral subtraction to a predetermined value (β | QR [k] | ² ) A frequency spectrum QLS of the signal YLS is generated. That is, the frequency analysis unit 92, the subtraction processing unit 94L, and the waveform synthesis unit 96 generate the acoustic signal YLS in which the incoming sound from the front right and the front (residual component) is suppressed and the incoming sound from the left rear is emphasized. It functions as the left rear directing unit 44 to be generated.

  As described above, in the third embodiment, the five surround sound signals Y (YC, YR, YL, YRS, YLS) are signals corresponding to the three sound signals X (XC, XR, XL). Since it is generated by processing, it is possible to reduce the size of the acoustic processing device 100C as in the first embodiment. In the above, the third embodiment has been described based on the first embodiment. However, the rear signal generation unit 40A (FIG. 9) of the second embodiment is replaced with the rear signal generation unit 40B of the third embodiment. Can also be employed.

<D: Modification>
Each of the above forms can be variously modified. Specific modifications are exemplified below. Two or more aspects arbitrarily selected from the following examples can be appropriately combined.

(1) Modification 1
In each of the above embodiments, five systems of acoustic signals Y (YC, YR, YL, YRS, YLS) are generated, but a configuration for generating four surround-type acoustic signals Y (YC, YR, YL, YS). Is also adopted. The acoustic signal YS is a signal in which the incoming sound from the rear (left rear or right rear) is emphasized. For example, the acoustic signal YRS generated by the rear signal generation unit 40 (40A, 40B) is combined with the acoustic signal YLS. (For example, addition or subtraction).

(2) Modification 2
In the first and second embodiments, the frequency spectrum RB of the acoustic signal RA and the frequency spectrum LB of the acoustic signal LA are multiplied by each coefficient value gS [k] of the processing coefficient sequence GS (formula (1A), formula (1B)), the content of the processing by the signal processing unit 64 (the suppression processing unit 64R, the suppression processing unit 64L) is appropriately changed. For example, each coefficient value gE [k] of the processing coefficient sequence GE calculated by the equation (4) becomes a numerical value closer to 1 as the coefficient value gE [k] of the frequency having a larger power (amplitude) of the residual component becomes a residual component. The coefficient value gE [k] at a frequency with a small power becomes a value closer to zero. Therefore, a configuration in which the suppression processing unit 64R generates the frequency spectrum RC (RC [k] = RB [k] −gE [k]) by subtracting the processing coefficient sequence GE from the frequency spectrum RB of the acoustic signal RA, or suppression. A configuration in which the processing unit 64L generates the frequency spectrum LC (LC [k] = LB [k] −gE [k]) by subtracting the processing coefficient sequence GE from the frequency spectrum LB of the acoustic signal LA may be employed. In the configuration using the processing coefficient sequence GE for suppressing residual components, the generation of the processing coefficient sequence GS is omitted.

(3) Modification 3
In the first and second embodiments, the processing coefficient sequence GS generated by spectral subtraction of the sum component M and the difference component S is applied to the acoustic signal RA (frequency spectrum RB) and the acoustic signal LA (frequency spectrum LB). However, a known technique (so-called minus one process) for selectively suppressing the incoming sound from the front can be arbitrarily adopted for suppressing the residual component.

(4) Modification 4
In the third embodiment, the set value (β | QL [k] | ² ) obtained by the flooring process of Expression (8B) is arbitrarily changed. For example, a configuration may be employed in which the power PRS [k] of a frequency that becomes a negative number in the spectral subtraction of Expression (8A) is set to a predetermined value (eg, zero) that is unrelated to the acoustic signal XL or the acoustic signal XC. The same applies to the calculation of the power PLS [k] of the acoustic signal XLS.

(5) Modification 5
A configuration in which the processing described in the above embodiments is applied only to components within a predetermined frequency band may be employed. Further, it is possible to adopt a configuration in which each acoustic signal X (XC, XR, XL) is divided into a plurality of frequency bands, and the processes in the above embodiments are performed in parallel for each frequency band and then synthesized.

100A, 100B, 100C .... Sound processing device, 12 (12C, 12R, 12L) ... Sound collecting device, 16 .... Sound emitting device, 22 .... Calculation processing device, 24 ... Storage device, 30 ... Front signal Generating unit, 32... Right front directing unit, 36R, 36L... Combining unit, 38R, 38L... Adjusting unit, 34 ... Left front directing unit, 40A, 40B. Directing unit 44... Left rear directing unit 52 R and 52 L Combining unit 54. Component suppressing unit 56 R and 56 L Adjusting unit 62 Frequency analyzing unit 64 Signal processing unit 64 R 64L: Suppression processing unit, 66: Waveform synthesis unit, 70: Coefficient setting unit, 72: Sum component generation unit, 74: Difference component generation unit, 76: Coefficient sequence generation unit, 762: First Generation unit, 764... Second generation unit, 80... Directivity providing unit, 82R, 82L. 84 ...... unidirectional imparting unit, 842,846 ...... combining unit, 844 ...... delay unit.

Claims (5)

A unidirectional first acoustic signal directed forward, a bi-directional second acoustic signal directed right front and left rear, and a bi-directional third acoustic signal directed left front and right rear; An acoustic processing device for processing
A right front directing means for synthesizing a right front signal in which an incoming sound from the right front is emphasized from the first acoustic signal and the second acoustic signal;
Left front directing means for synthesizing a left front signal in which an incoming sound from the left front is emphasized from the first acoustic signal and the third acoustic signal;
A right rear directing means for synthesizing a right rear signal in which the incoming sound from the left front and the front is suppressed and the incoming sound from the right rear is emphasized from the first acoustic signal and the third acoustic signal;
Left rear directing means for synthesizing a left rear signal in which the incoming sound from the right front and the front is suppressed and the incoming sound from the left rear is emphasized from the first acoustic signal and the second acoustic signal. Sound processing device. Coefficient setting means for generating a processing coefficient sequence in which coefficient values for each frequency are set so that incoming sound from the front is suppressed,
The right rear pointing means is
First synthesizing means for synthesizing from the first acoustic signal and the third acoustic signal a fourth acoustic signal in which the incoming sound from the left front of the third acoustic signal is suppressed;
First suppression processing means for generating the right rear signal by causing each coefficient value of the processing coefficient sequence to act on each frequency component of the fourth acoustic signal,
The left rear pointing means is
Second synthesizing means for synthesizing from the first acoustic signal and the second acoustic signal a fifth acoustic signal in which the incoming sound from the right front of the second acoustic signal is suppressed;
The acoustic processing apparatus according to claim 1, further comprising: second suppression processing means for generating the left rear signal by causing each coefficient value of the processing coefficient sequence to act on each frequency component of the fifth acoustic signal. The coefficient setting means includes
Sum component generating means for generating a sum component by adding the spectrum of the fourth acoustic signal and the fifth acoustic signal;
Difference component generation means for generating a difference component in which the incoming sound from the front is suppressed by subtracting the spectrum of the fourth acoustic signal and the fifth acoustic signal;
The sound processing apparatus according to claim 2, further comprising coefficient sequence generation means for generating the processing coefficient sequence from the sum component and the difference component. The right rear pointing means performs spectral subtraction for subtracting the spectrum of the first acoustic signal from the spectrum of the third acoustic signal and flooring processing for setting a component value that becomes a negative number in the spectral subtraction to a predetermined value. To generate the right rear signal,
The left rear pointing means performs spectral subtraction for subtracting the spectrum of the first acoustic signal from the spectrum of the second acoustic signal, and flooring processing for setting a component value that becomes a negative number in the spectral subtraction to a predetermined value. The sound processing apparatus according to claim 1, wherein the left rear signal is generated. The first acoustic signal, the second acoustic signal, and the third acoustic signal from the first original signal, the second original signal, and the third original signal generated by the omnidirectional sound pickup devices installed at different positions. The sound processing apparatus according to claim 1, further comprising directivity providing means for generating

Images