Article
Mode-Matching-Based Sound Field Recording and Synthesis with Circular Double-Layer Arrays Takuma Okamoto
ID
National Institute of Information and Communications Technology, 3-5, Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-0289, Japan;
[email protected]; Tel.: +81-774-98-6409 Received: 30 May 2018; Accepted: 3 July 2018; Published: 4 July 2018
Abstract: A sound field control approach is investigated for recording a primary sound field and synthesizing it at a secondary field without exterior radiation using circular double-layer arrays of microphones and loudspeakers. Although the conventional least-squares (LS) and generalized singular value decomposition (GSVD) approaches are based on numerical solutions and control the discretized interior and exterior sound pressures, this paper provides a mode-matching-based analytical method with circular double-layer receiver and source. The primary sound field cylindrical harmonic spectrum is analytically estimated from the recorded sound pressures without forbidden frequencies, and the driving signals of the loudspeakers for synthesizing it are analytically derived without interior and exterior control points. Computer simulations demonstrate the effectiveness of the proposed analytical formulation with circular double-layer arrays of microphones and loudspeakers. Compared to the conventional numerical LS and GSVD approaches, the interior sound field synthesis is more accurate and the exterior sound propagation is more effective in the proposed method under both free-field and reverberant conditions. Keywords: sound field synthesis; sound field recording; loudspeaker array; cylindrical harmonic expansion; Ambisonics
1. Introduction Generating acoustically bright and dark zones is an important and attractive problem for acoustic communication technology, and many approaches are being investigated [1–15]. Although these methods only control the acoustic contrast or energy between the bright and dark zones, multizone sound field synthesis methods also simultaneously control not only the sound pressures but also multiple sound fields in multiple regions [16–25]. Most sound field synthesis approaches assume that the synthesis environment is a free-field and the synthesis target area is inside the loudspeaker array [26–30]. Figure 1a shows an example of a two-dimensional interior sound field synthesis based on higher-order Ambisonics [29] with a circular loudspeaker array in a free-field. Although a plane wave is synthesized inside the array, undesired sound pressures propagate outside the array. In addition, the actual synthesis environment is not a free-field, and reverberation occurs by undesired sound pressures propagating outside the array. Consequently, reverberation degrades the synthesis accuracy in the target area (Figure 1b). Therefore, considering the sound pressures propagating outside the array is important both to reduce the undesired sound propagation and to maintain the synthesis accuracy in the target area in the actual reverberant conditions. Considering the reverberant conditions, exterior propagation-free sound field synthesis approaches have been proposed [31–34]. These approaches can synthesize a sound field inside an array without propagating undesired sound pressures outside the array (Figure 1c). As a result, reverberations do not occur, and the synthesis accuracy inside the array remains high in actual Appl. Sci. 2018, 8, 1084; doi:10.3390/app8071084
www.mdpi.com/journal/applsci
Appl. Sci. 2018, 8, 1084
2 of 17
reverberant conditions. These methods are categorized as multizone sound field control approaches where the synthesized and acoustically dark zones are simultaneously set inside and outside the array. In these methods, the synthesized sound field is assumed to be a simple plane wave. To extend exterior propagation-free sound field synthesis for reproducing a primary sound field recorded by a microphone array [35], this paper focuses on sound field recording and synthesis methods without exterior propagation using arrays of microphones and loudspeakers.
(a) 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
(b) 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
(c) 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
Figure 1. Comparison of two-dimensional sound field syntheses with a circular array: (a) typical interior sound field synthesis in a free-field; (b) typical interior sound field synthesis in a reverberant environment; and (c) undesired propagation-free sound field synthesis in a reverberant environment.
To achieve such a sound field recording and synthesis approach, pressure matching-based least-squares (LS) [36,37] and generalized singular value decomposition (GSVD) [38] methods [39] are directly introduced. The LS and GSVD methods are based on the Kirchhoff–Helmholtz (KH) integral [40] and introduce double-layer arrays of microphones and loudspeakers. The LS method is, however, based on numerical inversion and requires repeated calculations to select the optimal regularization parameters. In the GSVD approach, iterative calculations are also required for finding the optimal threshold to automatically decide the singular value truncation parameters [40]. In previous works [14,34,41,42], the spatial Fourier transform-based analytical approaches to sound field control without repeated calculations outperformed the conventional LS methods in terms of control accuracy. For more accurate interior sound field synthesis without exterior sound propagation than the conventional LS and GSVD approaches without iterative computations, this paper proposes an analytical approach with continuous circular double-layer source and receiver. The continuous circular double-layer source and receiver are finally discretized to circular double-layer arrays of microphones and loudspeakers. The proposed method is based on cylindrical harmonic expansion, which is a spatial Fourier transformation in cylindrical coordinates that is categorized as a mode-matching scheme. Although the conventional LS and GSVD methods require interior and exterior discretized control points, the proposed method does not, since a mode matching-based method controls the cylindrical harmonic spectrums of the interior and exterior sound fields as control points instead of the sound pressures. As in the GSVD method [39], the proposed method is also formulated in a two-dimensional height-invariant sound field with a two-dimensional Green’s function, circular double-layer arrays, and cylindrical harmonic expansion. The proposed analytical method can also be easily extended for three-dimensional sound fields by introducing the three-dimensional Green’s function, spherical double-layer arrays, and spherical harmonic expansion. The rest of this paper is organized as follows. Section 2 briefly introduces the conventional LS and GSVD methods. The proposed formulation for circular double-layer source and receiver is analytically derived in Section 3. In Section 4, computer simulations compare the proposal with the conventional
Appl. Sci. 2018, 8, 1084
3 of 17
LS and GSCD methods in both the free-field and reverberant conditions. Section 5 finally concludes this paper. 2. Pressure Matching-Based Sound Field Recording and Synthesis without Exterior Propagation with Double-Layer Arrays The conventional LS- and GSVD-based sound field recording and synthesis methods are briefly introduced here for comparison with the proposed approach. 2.1. Kirchhoff–Helmholtz Integral Let S be a two-dimensional volume without sound sources, bounded by surface δS. Sound pressure P(r, k) at point r for wavenumber k = 2π f /c is uniquely defined from sound pressures P(r 0 , k) and particle velocities ∂P(r 0 , k)/∂n at boundary δS by the KH integral [40]: αP(r, k) =
Z S
∂G (r, r 0 , k) ∂P(r 0 , k ) G (r, r 0 , k) − P(r 0 , k) ∂n ∂n
dr 0 ,
(1)
where α = 1 if r is inside S, α = 1/2 if r is on δS and α = 0 if r is outside S. f and c are the temporal frequency and the speed of sound, respectively. The two-dimensional free-field Green’s function is given: G (r, r 0 , k) =
j H0 (k|r − r 0 |), 4
(2)
√ where j = −1 and Hm is the m-th order Hankel function of the first kind. In the conventional LS and GSVD methods for recording a primary sound field and synthesizing it without exterior propagation, double-layer receivers and secondary sources are, respectively, introduced to achieve monopole and approximated dipole distributions on δS and simultaneously control both the sound pressures and their derivatives based on the KH integral. 2.2. Least-Squares Approach In the recording stage, a primary sound field is first recorded by using a double-layer array of Iint microphones (Figure 2a). In the synthesis stage, two double-layer arrays of Iint and Iext control points are arranged for simultaneously controlling both the interior and exterior sound fields. In pressure matching approaches, the arrangement of the interior control points is just the same as that of the microphone array in the primary sound field. A secondary sound field is then controlled by a double-layer array of L loudspeakers located between the interior and exterior control points with driving signals of L × 1 vector D (k ) (Figure 2b). The secondary sound pressures at the control points are defined as I (=Iint + Iext ) × 1 vectors P (k) and given: P ( k ) = G ( k ) D ( k ),
(3)
where G (k) is the I × L transfer function matrix between the receivers and the sources whose coefficients are given: Gil (k) =
j H0 (k|r i − r l |). 4
(4)
Appl. Sci. 2018, 8, 1084
4 of 17
To achieve interior sound field synthesis without exterior sound propagation, the target secondary sound pressures are set: P (k) =
(
0, outside the loudspeaker array , Porg (k), inside the loudspeaker array
(5)
where Porg (k ) is the Iint × 1 vector of the primary sound pressures captured by the microphone array. The optimal driving signals are then derived based on the least-squares solution: DLS (k) = ( G H (k) G (k ))−1 G H (k)P (k),
(6)
where DLS (k) is the driving signal for the LS method and G H (k) is the Hermitian transpose of G (k ). With the derived driving signals, both the interior and exterior sound fields are simultaneously controlled. Exterior double-layer control points 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 sha1_base64="bs8QlLljyYde7WbNFuyzJJm+ZMk=">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 sha1_base64="MFJp9d/tfL1BF9EA1oPBTbZUME0=">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 sha1_base64="pCdPZAKO691oL9LOhiKUyxSpnWI=">AAACiXichVExTxsxGH1caUtDaUK7VOpyIqViIfqOpRETEqrUkQRCkNIoujsMWHHOJ58TkUb9A/yBDkxUqhDq0rVdu/AHGPgJiDFIXTr0y+WkChD0s2w/P3/v87MdxEomluh8wnkw+fDR46knuemnM8/yhdnnm4numlDUQq202Qr8RCgZiZqVVomt2Ai/EyhRD9qro/16T5hE6mjD9mPR7Pi7kdyRoW+ZahVev9u3wkht3G3dZdGi8vvCuKGOrNHKjbWMbNIqFKlEabi3gZeBIrJY04VjfMA2NEJ00YFABMtYwUfCrQEPhJi5JgbMGUYy3Rf4hBxru5wlOMNnts3jLq8aGRvxelQzSdUhn6K4G1a6mKczOqEhndI3uqA/d9YapDVGXvo8B2OtiFv5g5frv/+r6vBssfdPda9nix2UU6+SvccpM7pFONb3Pn4eri9X5wdv6Atdsv8jOqdffIOodxV+rYjqIXL8Ad7N574NNpdKHpW8ChVXytlXTOEV5rDA7/0WK3iPNdT43AN8xw/8dKYdzyk7y+NUZyLTvMC1cFb/AqVNmBc=
Double-layer loudspeaker array
Double-layer microphone array
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
Interior double-layer control points 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
(a) Recording stage
(b) Synthesis stage
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
Figure 2. Arrangements: (a) double-layer microphone array for recording a primary sound field; and (b) double-layer arrays of loudspeakers and control points for synthesizing recorded primary sound field at a secondary field.
2.3. Generalized Singular Value Decomposition Approach In the GSVD approach [39], G (k ) is also decomposed into interior component Gint (k ) and exterior component Gext (k): G (k) =
"
Gext (k) Gint (k)
#
.
(7)
GSVD is applied to Equation (7). Gint (k) and Gext (k ) are then decomposed: Gext (k) = U (k)C (k) X H (k), H
Gint (k) = V (k)S(k) X (k),
(8) (9)
where U and V are complex unitary matrices, X is a complex square invertible matrix, and C and S are pseudo-diagonal matrices containing singular values: C H (k)C (k) + S H (k )S(k) = I,
(10)
Appl. Sci. 2018, 8, 1084
5 of 17
where I is the L × L identity matrix. The optimal driving signals of the GSVD method are obtained: H DGSVD (k) = ( X H )−1 (k)S+ K ( k )V ( k ) Porg ( k ),
(11)
where SK (k) retains the first K largest coefficients of S(k) along the diagonal and S+ K ( k ) is the generalized inverse of SK (k). The tradeoff between the interior field synthesis accuracy and the exterior field propagation reduction can be controlled by singular value truncation parameter K. As in a previous work [39], K was automatically chosen to satisfy that all diagonal terms of C K (k) are smaller than 0.01 where C K (k) also retains the first K smallest coefficients of C (k) along the diagonal. When K = L, the GSVD method just corresponds to the LS approach. The detailed derivation of the GSVD approach was previously described [39]. 3. Proposed Analytical Formulation In the LS and GSVD methods, the discretized primary interior sound pressures recorded by a double-layer microphone array are directly controlled. The proposed approach, on the other hand, controls the cylindrical harmonic spectrum of the estimated primary interior sound field rather than the discretized sound pressures. As a result, the proposal does not require interior and exterior control points. 3.1. Cylindrical Harmonic Expansion of Two-Dimensional Sound Field The cylindrical harmonic expansion of a two-dimensional sound field is briefly introduced here. Spherical coordinates relative to Cartesian coordinates are defined in Figure 3. In two-dimensional sound field, z = 0 and θ = π/2. z
]T p r = x2 + y 2 + z 2 p x2 + y 2 1 ✓ = tan z 1 y = tan x
r = [r, ✓, ✓ r
y
x
Figure 3. Definition of spherical coordinates relative to Cartesian coordinates.
Its interior expansion in a region, which is homogeneous and free of sources, is given: ∞
P(r, φ, k) =
∑
m=−∞
A˚ m (k ) Jm (kr )e jmφ ,
(12)
where A˚ m (k ) and Jm are the interior sound field cylindrical harmonic spectrum and the m-th order Bessel function. The corresponding expansion for regions exterior to any sound sources is given: ∞
P(r, φ, k) =
∑
m=−∞
B˚ m (k) Hm (kr )e jmφ ,
(13)
where B˚ m (k) is the exterior sound field cylindrical harmonic spectrum. 3.2. Interior and Exterior Sound Field Control with Circular Double-Layer Monopole Source and Receiver A sound field recorded by a continuous circular monopole receiver distribution that is centered at the origin with radius R0 is converted into the cylindrical harmonic domain [40]:
Appl. Sci. 2018, 8, 1084
6 of 17
1 P˚m ( R0 , k ) = 2π
Z 2π 0
P( R0 , φ, k)e− jmφ dφ.
(14)
The cylindrical harmonic spectrum of the interior sound field is then obtained: P˚m ( R0 ) A˚ m (k ) = . Jm (kR0 )
(15)
Equation (15), however, has a forbidden frequency (Bessel-zero) problem where Jm (kR0 ) = 0 and is dependent on receiver radius R0 and wavenumber k [40]. To avoid the forbidden frequency problem for open circular monopole receivers, a circular double-layer monopole receiver centered on the origin with radii R1 and R2 is introduced (Figure 4a). Sound fields recorded by both layers are converted into the cylindrical harmonic domain by Equation (14) as P˚m ( R1 , k) and P˚m ( R2 , k). By applying a method with spherical double-layer monopole receivers [43] into circular double-layer monopole receivers in the proposed method, the cylindrical harmonic spectrum is calculated:
(1 − β m ( R1 , R2 , k)) P˚m ( R1 , k) + β m ( R1 , R2 , k) P˚m ( R2 , k) A˚ m (k ) = , (1 − β m ( R1 , R2 , k)) Jm (kR1 ) + β m ( R1 , R2 , k) Jm (kR2 )
(16)
where β m ( R1 , R2 , k ) =
(
0, 1,
| Jm (kR1 )| ≥ | Jm (kR2 )| . | Jm (kR1 )| < | Jm (kR2 )|
(17)
Equation (16) indicates that A˚ m (k) is obtained from a single layer of Jm (kR1 ) or Jm (kR2 ) with a large absolute value based on the assumption that Jm (kR1 ) and Jm (kR2 ) do not simultaneously become zero. In the proposal, therefore, the absence of mutual zeros needs to hold for all cylindrical harmonic orders. Circular double-layer monopole source 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
y
y
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
AAACY3ichVFNSwJBGH7cvswszToEEUhidJLZCLJOQpeOfmQKJrK7jba47i67q2DSH6hr0aFTQUT0M7r0Bzr4B4LoaNClQ6/rQpRU7zAzzzzzPu88MyObmmo7jHV9wsjo2PiEfzIwFZyeCYVnI3u20bQUnlcMzbCKsmRzTdV53lEdjRdNi0sNWeMFub7d3y+0uGWrhr7rtE1ebkg1Xa2qiuQQlWlXwjGWYG5Eh4HogRi8SBvhW+zjAAYUNNEAhw6HsAYJNrUSRDCYxJXRIc4ipLr7HMcIkLZJWZwyJGLrNNZoVfJYndb9mrarVugUjbpFyiji7IndsR57ZPfshX38Wqvj1uh7adMsD7TcrIROFnLv/6oaNDs4/FL96dlBFUnXq0reTZfp30IZ6FtHF73cVjbeWWHX7JX8X7Eue6Ab6K035SbDs5cI0AeIP597GOTXEpsJMbMeSyW9n/BjEctYpefeQAo7SCNPx3Kc4gznvmchKESE+UGq4PM0c/gWwtInZSOJww==
R1
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
r1
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
x
x
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 , r1 , k) + D P˚ (r> , r2 , k)= 2π D
(23)
j = 2π Jm (kr ) D˚ m (r1 , k) Hm (kr1 ) + D˚ m (r2 , k) Hm (kr2 ) , 4
j = 2π Hm (kr ) D˚ m (r1 , k) Jm (kr1 ) + D˚ m (r2 , k) Jm (kr2 ) , 4
(22)
(24)
where r< , r1,2 and r> , r1,2 denote P˚ (r, r1,2 , k) and G˚ (r, r1,2 , k) for r < r1,2 and r > r1,2 employed in [40] and j G˚ (r< , r1,2 , k ) = Jm (kr ) Hm (kr1,2 ), 4 j G˚ (r> , r1,2 , k ) = Jm (kr1,2 ) Hm (kr ). 4
(25) (26)
Appl. Sci. 2018, 8, 1084
8 of 17
To synthesize the interior sound field recorded by the circular double-layer monopole receiver without exterior propagation, P˚ (r< , r1 , k) and P˚ (r> , r2 , k ) are set to P˚ (r< , r1 , k) = A˚ m (k ) Jm (kr ), P˚ (r> , r2 , k ) = 0.
(27) (28)
Equations (27) and (28) are then represented in matrix form as in [31,34]: j 2π · 4
"
Hm (kr1 ) Jm (kr1 )
Hm (kr2 ) Jm (kr2 )
#"
˚ m (r1 , k ) D ˚ Dm ( r 2 , k )
#
=
"
A˚ m (k) 0
#
,
(29)
and the driving function of the circular double-layer monopole source is analytically derived:
−2jJm (kr2 ) A˚ m (k) , π { Hm (kr1 ) Jm (kr2 ) − Hm (kr2 ) Jm (kr1 )} 2jJm (kr1 ) A˚ m (k) ˚ m (r2 , k ) = D . π { Hm (kr1 ) Jm (kr2 ) − Hm (kr2 ) Jm (kr1 )}
˚ m (r1 , k ) = D
(30) (31)
Although the driving signals of the LS and GSVD methods derived in Equations (6) and (11) include the control point locations, those of the proposed method derived in Equations (30) and (31) are independent of the control point locations. The driving signals only depend on estimated cylindrical harmonic spectrum A˚ m (k) and circular source radii r1 and r2 since Jm (kr ) and Hm (kr ) are completely cancelled in Equations (30) and (31). Unlike the LS and GSVD methods, the proposed method, therefore, does not require interior and exterior control points. Continuous circular double-layer receiver and source are finally discretized into circular double-layer arrays of loudspeakers and microphones. When the number of microphones of each layer is Iint /2 with equiangular sampling, the recorded sound fields for radii R1 and R2 in the cylindrical harmonic domain are calculated: P˚m ( R1,2 , k) ≈
=
1 2π · Iint /2 2π 2 Iint
Iint /2
∑
Iint /2
∑
P( R1,2 , φi , k )e− jmφi
i =1
P( R1,2 , φi , k)e− jmφi .
(32)
i =1
When the number of loudspeakers of each layer is L/2 with equiangular sampling, order m of the cylindrical harmonic spectrum in Equations (30) and (31) can be calculated up to M = b( L/2 − 1)/2c, where b·c is the floor function. However, the cylindrical harmonic spectrum at |n| > |kr1,2 | is the evanescent component and is quite large because of the inverse propagation when r > r1,2 [40]. To calculate a stable filter as in [42], m is up to M = b|kr1,2 |c if b|kr1,2 |c < b( L/2 − 1)/2c since using only the propagation component for |n| ≤ |kr1,2 | is sufficient to control the target interior sound field. The driving signal of each loudspeaker at φl for r1 and r2 in the frequency domain is obtained: M
D (r1,2 , φl , k) =
∑
˚ m (r1,2 , k )e jmφl , D
l = 1, 2, · · · L/2.
(33)
m=− M
The purpose of introducing circular double-layer arrays of microphones and loudspeakers in the LS and GSVD methods differs from that in the proposed method. In the LS and GSVD methods, the circular double-layer arrays are introduced to satisfy the KH integral. Hence, an approximated dipole distribution is required. On the other hand, the approximated dipole distribution is not required because the proposed method is not based on the KH integral. Circular double-layer arrays are introduced in the proposed method to avoid the forbidden frequencies in the recording stage and to
Appl. Sci. 2018, 8, 1084
9 of 17
simultaneously control both the interior and exterior sound field cylindrical harmonic spectrums in the synthesis stage. In addition, the proposal can also be used when the radii of a circular double-layer array of microphones are larger than those of a circular double-layer array of loudspeakers. The proposed analytical method can also be easily extended for three-dimensional sound fields by introducing the three-dimensional Green’s function, spherical double-layer arrays and spherical harmonic expansion with the spherical convolution theorem [29] instead of the two-dimensional one in Equation (20), circular arrays shown in Figure 4, and circular harmonic expansion in Equations (12) and (13) with the circular convolution theorem for Equations (21) and (23). 4. Computer Simulations To evaluate the proposed analytical approach with circular double-layer arrays of microphones and loudspeakers, computer simulations based on the proposed method were performed and compared with the conventional numerical LS and GSVD methods. 4.1. Simulation Conditions In all the simulations, two-dimensional sound fields were assumed and both free-field and reverberant conditions were evaluated. The speed of sound c was 343.36 m/s. In the reverberant condition, the two-dimensional room size was 15 m × 12 m and the room center was centered at the origin. The transfer functions in the reverberant condition were calculated by the image method [47] in the frequency domain with reflection coefficient 0.3 and reflection order 10. The primary sound field was calculated as a cylindrical wave propagating in a free-field from a point at x = [5, 3]T . Both the amounts of microphones, interior and exterior control points for the LS and GSVD methods, and loudspeakers for each layer were Iint /2 = L/2 = 24, and their total numbers were Iint = L = 48. Then, each layer had 24 elements with equiangular sampling. The central radii of the circular double-layer arrays of microphones and loudspeakers were ( R1 + R2 )/2 = 1.0 m and (r1 + r2 )/2 = 2.0 m. Two circular double-layer arrays of control points with radii R1 , R2 , R3 and R4 were introduced for the LS and GSVD methods and their central radii were ( R1 + R2 )/2 = 1.0 m and ( R3 + R4 )/2 = 3.0 m. Three types of layer intervals ∆ = R2 − R1 = R4 − R3 = r2 − r1 = 0.05, 0.25, and 0.5 m were evaluated. The arrangements of microphones, loudspeakers and control points are illustrated in Figure 5. The case with ∆ = 0.05 m is the same condition with previous simulations for the LS and GSVD [38]. The spatial Nyquist frequency of the circular secondary source was approximately derived as [31,39]: f Nyq ≈
c( L/2) . 4πr2
(34)
In the simulations, the f Nyq values were about 324, 309, and 291 Hz for ∆ = 0.05, 0.25, and 0.50 m, respectively. To evaluate the synthesized sound field, the synthesis error and exterior acoustic contrast at position r were defined as E(r ) = 10 log10
| Porg (r ) − Psyn (r )|2 , | Porg (r )|2
(35)
C (r ) = 10 log10
| Psyn (r )|2 , | Porg (r )|2
(36)
where Porg (r ) and Psyn (r ) are, respectively, the primary and synthesized sound pressures at position r. To evaluate the interior sound field synthesis accuracy and exterior acoustic contrast, the spatially-averaged synthesis error within r ≤ 1.5 m and exterior acoustic contrast with 2.5 m ≤ r ≤ 4.5 m were, respectively, calculated for 10 Hz ≤ f ≤ 500 Hz.
Appl. Sci. 2018, 8, 1084
10 of 17
In the GSVD method, singular value truncation parameter K in Equation (11) was also automatically chosen to satisfy that all the diagonal terms of C K (k) are smaller than 0.01 [39]. The maximum cylindrical harmonic order was M = b( L/2 − 1)/2c = 11 in the proposed approach. Exterior double-layer control points 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 sha1_base64="bs8QlLljyYde7WbNFuyzJJm+ZMk=">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 sha1_base64="MFJp9d/tfL1BF9EA1oPBTbZUME0=">AAACfnichVHLThsxFD1MSwkhbQIrpG5G5aFuGt1hA2KFhJC6DI+QSAFFMxNDLZzxyONEhIgfyA900RVIqKq6YQtbNv2BLviEqksqdcOCm0kkRFHhWraPj32uj32DWMnEEl2POC9ejr4ay4xnJ3Kv3+QLk7ntRLdMKMqhVtpUAz8RSkaibKVVohob4TcDJSrBwWp/v9IWJpE62rKdWOw2/f1I7snQt0zVC7Nrh1YYqY3b0C0WfVB+Rxg31JE1WrmxlpFN6oUZKlIa7mPgDcEMhlHSha/YQQMaIVpoQiCCZazgI+FWgwdCzNwuuswZRjLdFzhGlrUtPiX4hM/sAY/7vKoN2YjX/ZxJqg75FsXdsNLFHP2kb3RDP+g7/aLb/+bqpjn6Xjo8BwOtiOv53vTm32dVTZ4tPt2rnvRssYel1Ktk73HK9F8RDvTto883m8sbc915OqXf7P+ErumKXxC1/4Rn62LjC7JcAO/f734MtheKHhW9dUIGb/EO7/mbF7GCjyihzNf1cI4LXDoTjucsDUrljAxrNoUH4SzfARWzlzA= sha1_base64="pCdPZAKO691oL9LOhiKUyxSpnWI=">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
Double-layer loudspeaker array
Double-layer microphone array
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
r2 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
R1 R2
r1
R1 R2
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
R3
AAACaXichVFNLwNRFD0d3/XRYiNsGg2xmtwpqrWSsLCk1ZIgMjMeRqczk5nXJjT+gJWdYEUiIn6GjT9g4SeIJYmNh
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
R4 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
Interior double-layer control points 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
(a) Recording stage
(b) Synthesis stage
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
AAACeHichVHLSiNBFD1pR0fjK+pGmE1j8LUJt7MZcSW4cZkYo4EoobstY2Gnu+mqBGLwB/wBFy5kBkTFz3DjD7jwE4ZZOjAgLrzpNIiKeouqOnXqnlunqpzQk0oT3aeMvm/9A98Hh9LDI6Nj45mJyU0VNCNXlN3AC6KKYyvhSV+UtdSeqISRsBuOJ7acg9Xu/lZLREoG/oZuh2KnYdd9uSddWzNVy0wtOItmqe3rfaGkMpW266KWyVKO4jDfAysBWSRRCDIX2MYuArhoogEBH5qxBxuKWxUWCCFzO+gwFzGS8b7AEdKsbXKW4Ayb2QMe67yqJqzP625NFatdPsXjHrHSxCzd0RU90C1d0x96+rBWJ67R9dLm2elpRVgbP54u/f9S1eBZY/9F9alnjT0sxV4lew9jpnsLt6dvHZ48lJbXZztz9Jv+sv9fdE83fAO/9c89L4r1U6T5A6y3z/0ebOZzFuWsYj67spR8xSB+YAYL/N4/sYI1FFDmc9s4wyWuUo+Gacwbi71UI5VopvAqjPwzBxyRKg==
(r1 + r2 )/2 = 2.0 m
(R1 + R2 )/2 = 1.0 m
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
(R1 + R2 )/2 = 1.0 m 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
(R3 + R4 )/2 = 3.0 m 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
= 0 05 0 25 and 0 5 m
F gure 5 Compu er s mu a ons cond ons a doub e ayer m crophone array and b doub e ayer arrays o oudspeakers and con ro po n s Each ayer has 24 e emen s w h equ angu ar samp ng
4 2 Inter or Synthes s Accuracy and Exter or Acoust c Contrast F gures 6 and 7 respect ve y show the resu ts of the pr mary sound fie d synthes zed sound fie d nter or sound fie d synthes s accuracy and exter or acoust c contrast by the LS GSVD and proposed mode-match ng methods w th ∆ = 0 25 m n both the free-fie d and reverberant cond t ons for = 200 The resu ts n F gure 6 nd cate that the LS and proposed methods rea ze a h gher nter or contro accuracy than the GSVD method Th s s because the convent ona GSVD method reduces the exter or propagat on above the spat a Nyqu st frequency at the expense of the nter or synthes s accuracy [39] The resu ts n F gure 6 a so suggest that the proposed method can ach eve the h ghest exter or acoust c contrast As a resu t the nter or synthes s accurac es of the LS and GSVD methods are degraded by the reverberat ons caused by exter or propagat ons (F gure 7) On the other hand on y the proposed method can successfu y contro both the nter or and exter or sound fie ds n the reverberant cond t on (F gure 7) F gures 8 and 9 respect ve y p ot the resu ts of the spat a y-averaged nter or synthes s error and exter or acoust c contrast defined n Equat ons (35) and (36) As descr bed above the convent ona GSVD approach reduces the exter or propagat on beyond the spat a Nyqu st frequency at the expense of the nter or synthes s accuracy [39] The resu ts n F gure 9 show that the GSVD method successfu y reduces the exter or propagat on over the spat a Nyqu st frequency compared w th the LS and proposed methods However the nter or fie d s no onger synthes zed by the GSVD method at h gher frequenc es (F gure 8) The resu ts n F gures 6–9 demonstrate the ma n mpact of th s paper compared to convent ona numer ca approaches the proposed method can more accurate y synthes ze the nter or fie d wh e reduc ng the exter or propagat on be ow the spat a Nyqu st frequency n both the free-fie d and reverberant cond t ons As descr bed above th s s because the proposed method rea zes the h ghest exter or acoust c contrast and successfu y contro s both the nter or and exter or sound fie ds n the reverberant cond t on These resu ts a so suggest the effect veness of the spat a Four er transform-based approach compared w th the LS method [14 34 41 42]
Appl. Sci. 2018, 8, 1084
11 of 17
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
Acoustic contrast
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
Synthesis error
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
Synthesized field
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
Primary field
Unlike the LS and GSVD methods, which depend on layer intervals ∆, the proposed method’s interior field synthesis accuracy is independent of ∆ (Figure 8) because the proposed method can successfully estimate the interior sound field cylindrical harmonic spectrum from the circular double-layer microphone array using Equation (16) without the forbidden frequencies. The proposed method does not control the discretized sound pressures. Instead, it controls the estimated sound field in the cylindrical harmonic domain, as derived in Equations (30) and (31). However, the driving signals of the LS and GSVD methods obviously include the control point and loudspeaker locations, as described in Equations (6) and (11).
(a) LS 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
(b) GSVD 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
(c) Proposed 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
Figure 6. Results of primary sound field, synthesized sound field, interior sound field synthesis accuracy and exterior acoustic contrast by LS, GSDV, and proposed mode-matching methods with ∆ = 0.25 m in free-field condition for f = 200 Hz. Blue ◦, 4 and + are microphones, loudspeakers and control points, respectively.
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
Acoustic contrast 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
Synthesis error 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
Synthesized field
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
Primary field
Appl. Sci. 2018, 8, 1084
(a) LS
AAACaHichVG7SgNBFD1Z3/GVaKFiEwyKNmE2CD4qwcbCIjHGBFRkdx11zL7YnQRi8Aes7EStFETEz7DxByzyCZJSwcbCm82CqKh3mJkzZ+65c2ZGd03hS8bqEaWtvaOzq7sn2tvXPzAYiw9t+E7ZM3jecEzHK+qaz01h87wU0uRF1+OapZu8oJeWm/uFCvd84djrsurybUvbt8WeMDRJ1Ma0NpNYze3EkizFgkj8BGoIkggj48RusYVdODBQhgUOG5KwCQ0+tU2oYHCJ20aNOI+QCPY5jhElbZmyOGVoxJZo3KfVZsjatG7W9AO1QaeY1D1SJjDJntgde2GP7J49s/dfa9WCGk0vVZr1lpa7O4Mno7m3f1UWzRIHn6o/PUvsYT7wKsi7GzDNWxgtfeXo7CW3uDZZm2LXrEH+r1idPdAN7MqrcZPla5eI0geo35/7J8inUwspNZtOLs2GP9GNcUxgmp57DktYQQZ5OvYQpzjHRaShxJURZayVqkRCzTC+hDLxARDGiuc=
12 of 17
(b) GSVD
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
(c) Proposed
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
Figure 7. Results of primary sound field, synthesized sound field, interior sound field synthesis accuracy and exterior acoustic contrast by LS, GSDV, and proposed mode-matching methods with ∆ = 0.25 m in reverberant condition for f = 200 Hz. Blue ◦, 4 and + are microphones, loudspeakers and control points, respectively.
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
= 0.50 m
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
= 0.25 m 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
= 0.05 m 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
= 0.50 m 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
= 0.25 m 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
= 0.05 m
Appl. Sci. 2018, 8, 1084 13 of 17
Free-field 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
Free-field
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
Reverberant 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
Figure 8. Results of spatially-averaged synthesis error for interior sound field within r ≤ 1.5 m.
Reverberant
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
Figure 9. Results of spatially-averaged acoustic contrast for exterior sound field with 2.5 m ≤ r ≤ 4.5 m.
Appl. Sci. 2018, 8, 1084
14 of 17
4.3. Driving Signal Stability
= 0.25 m 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
= 0.50 m
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
= 0.05 m
To evaluate the driving signal stability, Figure 10 plots maximum and minimum driving signal powers 10 log10 max(| D (r1,2 , φl )|2 ) and 10 log10 min(| D (r1,2 , φl )|2 ) . The driving signals of the proposed method for all the layer intervals below the spatial Nyquist frequency are stable, similar to those of the LS method. These results suggest that the forbidden frequencies for Jm (kr1 ), Jm (kr2 ), Jm (kR1 ) and Jm (kR2 ) in Equations (16), (30) and (31) were effectively avoided by the proposed method in these array configurations. The driving signal dynamic ranges of the GSVD method for all layer intervals were larger than the others, although the driving signal stability was not evaluated in a previous work [38].
Figure 10. Results of maximum and minimum driving signal power.
4.4. Limitations The proposed method has two limitations. One is that it cannot control the exterior field above the spatial Nyquist frequency because it is based on the spatial Fourier transformation in cylindrical coordinates [40] (Figure 9). The other is that circular and spherical double-layer array configurations with equiangular sampling are required for the proposed method to derive analytical formulations. In contrast, the conventional LS and GSVD methods can manage other configurations such as the rectangular and quadrate double-layer array employed in [39]. Consequently, the effectiveness of the proposed analytical approach using circular double-layer arrays of microphones and loudspeakers was confirmed in terms of its interior sound field synthesis accuracy and exterior propagation reduction within the spatial Nyquist frequency compared with the conventional numerical LS and GSVD methods in both the free-field and reverberant conditions. 5. Conclusions This study proposed a sound field control approach to recording a primary sound field and synthesizing it at a secondary field without exterior radiation using circular double-layer arrays of microphones and loudspeakers. Although the conventional LS and GSVD methods are based on a numerical approach and control the discretized sound pressures at the interior and exterior control points, a mode-matching-based analytical method with circular double-layer receiver and source was proposed. The primary sound field cylindrical harmonic spectrum was analytically obtained from the recorded sound pressures without forbidden frequencies, and the driving signals of the loudspeakers
Appl. Sci. 2018, 8, 1084
15 of 17
for synthesizing it were analytically derived without interior and exterior control points. The proposed analytical approach’s better effectiveness with circular double-layer arrays of microphones and loudspeakers was confirmed by computer simulations in terms of the interior sound field synthesis accuracy and exterior propagation reduction within the spatial Nyquist frequency compared with the conventional numerical LS and GSVD methods in both the free-field and reverberant conditions. The proposal can also be easily extended for three-dimensional sound fields by introducing the three-dimensional Green’s function, spherical double-layer arrays, and spherical harmonic expansion. Funding: This study was partly supported by JSPS KAKENHI Grant Numbers JP15K21674 and JP18K11387. Conflicts of Interest: The author declare no conflict of interest.
References 1. 2. 3. 4. 5. 6. 7.
8. 9. 10. 11. 12. 13. 14. 15.
16. 17. 18. 19.
Choi, J.W.; Kim, Y.H. Generation of an acoustically bright zone with an illuminated region using multiple sources. J. Acoust. Soc. Am. 2002, 111, 1695–1700. [CrossRef] [PubMed] Jones, M.; Elliot, S.J. Personal audio with multiple dark zones. J. Acoust. Soc. Am. 2008, 124, 3497–3506. [CrossRef] [PubMed] Chang, J.H.; Lee, C.H.; Park, J.Y.; Kim, Y.H. A realization of sound focused personal audio system using acoustic contrast control. J. Acoust. Soc. Am. 2009, 125, 2091–2097. [CrossRef] [PubMed] Elliot, S.J.; Cheer, J.; Murfet, H.; Holland, K.R. Minimally radiating sources for personal audio. J. Acoust. Soc. Am. 2010, 128, 1721–1728. [CrossRef] [PubMed] Shin, M.; Lee, S.Q.; Fazi, F.M.; Nelson, P.A.; Kim, D.; Wang, S.; Park, K.H.; Seo, J. Maximization of acoustic energy difference between two spaces. J. Acoust. Soc. Am. 2010, 128, 121–131. [CrossRef] [PubMed] Elliott, S.J.; Cheer, J.; Choi, J.W.; Kim, Y. Robustness and regularization of personal audio systems. IEEE Trans. Audio Speech Lang. Process. 2012, 20, 2123–2133. [CrossRef] Coleman, P.; Jackson, P.J.B.; Olik, M.; Møller, M.; Olsen, M.; Pedersen, J.A. Acoustic contrast, planarity and robustness of sound zone methods using a circular loudspeaker array. J. Acoust. Soc. Am. 2014, 135, 1929–1940. [CrossRef] [PubMed] Cai, Y.; Wu, M.; Yang, J. Sound reproduction in personal audio systems using the least-squares approach with acoustic contrast control constraint. J. Acoust. Soc. Am. 2014, 135, 734–741. [CrossRef] [PubMed] Shin, M.; Fazi, F.M.; Nelson, P.A.; Hirono, F.C. Controlled sound field with a dual layer loudspeaker array. J. Sound Vib. 2014, 333, 3794–3817. [CrossRef] Helwani, K.; Spors, S.; Buchner, H. The synthesis of sound figures. Multidimens. Syst. Signal Process. 2014, 25, 379–403. [CrossRef] Coleman, P.; Jackson, P.J.B.; Olik, M.; Pedersen, J.A. Personal audio with a planar bright zone. J. Acoust. Soc. Am. 2014, 136, 1725–1735. [CrossRef] [PubMed] Olik, M.; Jackson, P.J.B.; Coleman, P. Optimal source placement for sound zone reproduction with first order reflections. J. Acoust. Soc. Am. 2014, 136, 3085–3096. [CrossRef] [PubMed] Gálvez, M.F.S.; Elliott, S.J.; Cheer, J. Time domain optimization of filters used in a loudspeaker array for personal audio. IEEE/ACM Trans. Audio Speech Lang. Process. 2015, 23, 1869–1878. [CrossRef] Okamoto, T.; Sakaguchi, A. Experimental validation of spatial Fourier transform-based multiple sound zone generation with a linear loudspeaker array. J. Acoust. Soc. Am. 2017, 141, 1769–1780. [CrossRef] [PubMed] Olivieri, F.; Fazi, F.M.; Fontana, S.; Menzies, D.; Nelson, P.A. Generation of private sound with a circular loudspeaker array and the weighted pressure matching method. IEEE/ACM Trans. Audio Speech Lang. Process. 2017, 25, 1579–1591. [CrossRef] Wu, Y.J.; Abhayapala, T.D. Spatial multizone soundfield reproduction: Theory and design. IEEE Trans. Audio Speech Lang. Process. 2011, 19, 1711–1720. [CrossRef] Radmanesh, N.; Burnett, I.S. Generation of isolated wideband sound fields using a combined two-stage Lasso-LS algorithm. IEEE Trans. Audio Speech Lang. Process. 2013, 21, 378–387. [CrossRef] Poletti, M.A.; Fazi, F.M. An approach to generating two zones of silence with application to personal sound systems. J. Acoust. Soc. Am. 2015, 137, 598–605. [CrossRef] [PubMed] Betlehem, T.; Zhang, W.; Poletti, M.; Abhayapala, T. Personal sound zones: Delivering interface-free audio to multiple listeners. IEEE Signal Process. Mag. 2015, 32, 81–91. [CrossRef]
Appl. Sci. 2018, 8, 1084
20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.
33. 34.
35. 36. 37. 38. 39. 40. 41.
42.
43. 44.
16 of 17
Jin, W.; Kleijn, W.B. Theory and design of multizone soundfield reproduction using sparse methods. IEEE/ACM Trans. Audio Speech Lang. Process. 2015, 23, 2343–2355. Radmanesh, N.; Burnett, I.S.; Rao, B.D. A Lasso-LS optimization with a frequency variable dictionary in a multizone sound system. IEEE/ACM Trans. Audio Speech Lang. Process. 2016, 24, 583–593. [CrossRef] Poletti, M.A.; Fazi, F.M. Generation of half-space sound fields with application to personal sound systems. J. Acoust. Soc. Am. 2016, 139, 1294–1302. [CrossRef] [PubMed] Zhang, W.; Abhayapala, T.D.; Betlehem, T.; Fazi, F.M. Analysis and control of multi-zone sound field reproduction using modal-domain approach. J. Acoust. Soc. Am. 2016, 140, 2134–2144. [CrossRef] [PubMed] Buerger, M.; Hofmann, C.; Kellermann, W. Broadband multizone sound rendering by jointly optimizing the sound pressure and particle velocity. J. Acoust. Soc. Am. 2018, 143, 1477–1490. [CrossRef] [PubMed] Donley, J.; Ritz, C.H.; Kleijn, W.B. Multizone soundfield reproduction with privacy and quality based speech masking filters. IEEE/ACM Trans. Audio Speech Lang. Process. 2018, 26, 1041–1055. [CrossRef] Berkhout, A.J.; de Vries, D.; Vogel, P. Acoustic control by wave field synthesis. J. Acoust. Soc. Am. 1993, 93, 2764–2778. [CrossRef] Spors, S.; Rabenstein, R.; Ahrens, J. The theory of wave field synthesis revisited. In Proceedings of the 124th Audio Engineering Society Convention 2008, Amsterdam, The Netherlands, 17–20 May 2008. Poletti, M.A. Three-dimensional surround sound systems based on spherical harmonics. J. Audio Eng. Soc. 2005, 53, 1004–1025. Ahrens, J.; Spors, S. An analytical approach to sound field reproduction using circular and spherical loudspeaker distributions. Acta Acust. Acust. 2008, 94, 988–999. [CrossRef] Ahrens, J.; Spors, S. Sound field reproduction using planar and linear arrays of loudspeakers. IEEE Trans. Audio Speech Lang. Process. 2010, 18, 2038–2050. [CrossRef] Poletti, M.A.; Abhayapala, T.D. Interior and exterior sound field control using general two-dimensional first-order sources. J. Acoust. Soc. Am. 2011, 129, 234–244. [CrossRef] [PubMed] Poletti, M.A.; Abhayapala, T.D.; Samarasinghe, P. Interior and exterior sound field control using two dimensional higher-order variable-directivity sources. J. Acoust. Soc. Am. 2012, 131, 3814–3823. [CrossRef] [PubMed] Chang, J.H.; Jacobsen, F. Sound field control with a circular double-layer array of loudspeaker. J. Acoust. Soc. Am. 2012, 131, 4518–4525. [CrossRef] [PubMed] Okamoto, T. Analytical approach to 2.5D sound field control using a circular double-layer array of fixed-directivity loudspeakers. In Proceedings of the 42nd IEEE International Conference on Acoustics, Speech and Signal ICASSP 2017, New Orleans, LA, USA, 5–9 May 2017; pp. 91–95. Zhang, W.; Samarasinghe, P.N.; Chen, H.; Abhayapala, T.D. Surround by sound: A review of spatial audio recording and reproduction. Appl. Sci. 2017, 7, 532. [CrossRef] Kirkeby, O.; Nelson, P. Reproduction of plane wave sound fields. J. Acoust. Soc. Am. 1993, 94, 2992–3000. [CrossRef] Betlehem, T.; Abhayapala, T.D. Theory and design of sound field reproduction in reverberant rooms. J. Acoust. Soc. Am. 2005, 117, 2100–2111. [CrossRef] [PubMed] Bai, Z.; Demmel, J.W. Computing the generalized singular value decomposition. SIAM J. Sci. Comput. 1993, 14, 1464–1438. [CrossRef] Pasco, Y.; Gauthier, P.A.; Berry, A.; Moreau, S. Interior sound field control using generalized singular value decomposition in the frequency domain. J. Acoust. Soc. Am. 2017, 141, 334–345. [CrossRef] [PubMed] Williams, E.G. Fourier Acoustics: Sound Radiation and Nearfield Acoustic Holography; Academic Press: London, UK, 1999. Okamoto, T. Generation of multiple sound zones by spatial filtering in wavenumber domain using a linear array of loudspeakers. In Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing ICASSP 2014, Florence, Italy, 4–9 May 2014; pp. 4733–4737. Okamoto, T. Analytical methods of generating multiple sound zones for open and baffled circular loudspeaker arrays. In Proceedings of the IEEE Workshop on Applications of Signal Processing to Audio and Acoustics WASPAA 2015, New Paltz, NY, USA, 18–21 October 2015. Balmages, I.; Rafaely, B. Open-sphere designs for spherical microphone arrays. IEEE Trans. Audio Speech Lang. Process. 2007, 15, 727–732. [CrossRef] Poletti, M.A. A unified theory of horizontal holographic sound systems. J. Audio Eng. Soc. 2000, 48, 1155–1182.
Appl. Sci. 2018, 8, 1084
45. 46.
47.
17 of 17
Tiana-Roig, E.; Jacobsen, F.; Grande, E.F. Beamforming with a circular microphone array for localization of environmental noise sources. J. Acoust. Soc. Am. 2010, 128, 3535–3542. [CrossRef] [PubMed] Fahim, A.; Samarasinghe, P.; Abhayapala, T.D. Extraction of exterior field from a mixed sound field for 2D height-invariant sound propagation. In Proceedings of the 2016 IEEE International Workshop on Acoustic Signal Enhancement (IWAENC), Xi’an, China, 13–16 September 2016. Allen, J.B.; Berkley, D.A. Image method for efficiently simulating small-room acoustics. J. Acoust. Soc. Am. 1979, 65, 943–950. [CrossRef] c 2018 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
