EP2485504B1 - Generation of quiet areas within the listener zone of multi-channel playback systems - Google Patents

Description

The present invention relates to the reproduction of an acoustic wave field in a listener zone by means of several loudspeakers, and more particularly to the formation of a wave-free section within the listener zone.

In addition to the known stereophonic methods for sound reproduction, methods aiming at a physical reconstruction of any wave field are gaining in importance in research and practice. Of particular interest are methods that allow analytic calculation of the drive signal for the speakers. This is the case, among others, with the Wellenfeldsyntlzese and Higher-Order Ambisonics. These two methods have the common goal of synthesizing a desired field in a broad area, so as to create a spatial impression in a relatively large homogeneous listening area compared to conventional stereophonic methods. Ideally, the listener area covers the entire area surrounded by the speakers.

The physical basis of both methods is based on the assumption of a continuous distribution of sound sources around the audience area to be sounded. The practical realization by a finite number of spatially discrete loudspeakers must be explicitly considered. In the following, the theory of Higher-Order Ambisonics and Wave Field Synthesis will be briefly discussed, with the reproduced sources of the auditory scene as virtual sources and the speakers used for reproduction as secondary sources.

The derivation of the control functions in Higher-Order Ambisonics is done by explicitly solving the underlying physical problem. For this purpose, the speakers are characterized by a circular or spherical continuous distribution of typically models acoustic monopoles that are assigned a weighting function. The integral equation following from this approach can be solved by the representation of the relevant acoustic fields in the form of spherical harmonics. Higher-order Ambisonics allows, in principle, the exact reconstruction of almost any sound field within the secondary source distribution. In practice, however, the continuous distribution of monopoles is realized by a spatially discrete distribution of a finite number of loudspeakers. An overview of Higher-Order Ambisonics, for example, the work of J. Daniel, "Representation of Acoustics Acoustics, Application for the Transmission and Playback of the Sound Effects of Multimedia", PhD thesis, Université Paris 6, 2001 be removed.

The method of spectral division (Spectral Division Method SDM, see for example J. Ahrens and S. Spors. Sound field reproduction using planar and linear arrays of loudspeakers. IEEE Transactions on Audio, Speech and Language Processing, 18 (8): 2038 - 2050, November 2010 , doi: 10.1109 / TASL.2010.2041106) represents in principle a transfer of the higher-order Ambisonics approach to planar and linear geometries. In SDM, the underlying physical problem, formulated as an integral equation, is explicitly solved. The solution is achieved by the transition into the spatial spectral range by means of a spatial Fourier transformation. As with Higher-Order Ambisonics, the speaker drive functions in this range are calculated by spectral division. In addition to the analytical expression for a desired sound field of a prototypical source, the propagation function (Green's function) of the loudspeakers is needed. After back transformation of the control function from the spectral range to the local area, this is then available for implementation.

The physical basis of wave field synthesis is qualitatively based on the Huygens-Fresnel principle and quantitatively on the Rayleigh integrals (see for example S. Spors, R. Rabenstein, and J. Ahrens, "The theory of wave field synthesis revisited," Audio Eng. Soc. Conv. Paper, Vol. 124 ). These state that an almost arbitrary sound field within a half-space or a half-level can be resynthesized by a distribution of secondary sources on the edge. The secondary sources must be controlled with the directional gradient of the desired sound field on the edge. This forms the physical basis for the use of linear or planar secondary source distributions. In practice, however, are usually piecewise used linear loudspeaker arrays. For this purpose, appropriate extension is made in theory (see for example S. Spors, R. Rabenstein, and J. Ahrens, "The theory of wave field synthesis revisited," Audio Eng. Soc. Conv. Paper, Vol. 124 ).

As with Higher-Order Ambisonics and SDM, the theory of wave field synthesis assumes a spatially continuous distribution of secondary sources. In practice, however, spatially discrete distributions of loudspeakers are used. This spatial discretization leads to a number of artifacts in the audience.

1. 1 Wiedergabe in einem begrenzten Bereich umgeben von einer Ruhezonen. Zum Beispiel in einem Büro, Wohnzimner, etc.
2. 2 Konstruktion von akustisch getrennten Bereichen, um verschiedene Inhalte in einem physikalisch offenen Gebiet so wiederzugeben, dass sich die Inhalte nicht vermischen. Zum Beispiel für die Wiedergabe von Stereosignalen ohne die Verwendung von Kopfhörern.
3. 3 Minimieren des Einflusses des Raumes, in dem die Wiedergabe stattfinden soll. Eine Möglichkeit dies zu erzielen ist, die Reflektion an den Wänden zu verhindern, indem die Energie des Schallfeldes nahe den Wänden minimiert wird.
4. 4 Unterdrückung des Schallfeldes in einem Ort, in dem sich ein Aufnahmesystem befindet, um Rückkopplungseffekte zu unterdrücken.

For some applications, it may be interesting to provide a quiet zone within a target PA, ie the listener zone, or to play in a limited area surrounded by a quiet zone. By way of example, the following possible applications may be mentioned:

1. 1 Play in a limited area surrounded by quiet areas. For example in an office, living room, etc.
2. 2 Construction of acoustically separated areas to render different contents in a physically open area so that the contents do not mix. For example, for the playback of stereo signals without the use of headphones.
3. 3 Minimizing the influence of the room in which the playback is to take place. One way to do this is to prevent reflection on the walls by minimizing the energy of the sound field near the walls.
4. 4 Suppression of the sound field in a location where a recording system is located to suppress feedback effects.

Approaches for generating quiet zones (acoustic contrast control) are, for example, from the document of J.-W. Choi and Y.-H. Kim, "Generatian of an acoustically bright zone with an illuminated region using multiple sources," Journal of the Acoustic Society of America, Vol. 111, No. 4, pp. 1695-1700, April 2002 known. These methods are based on the approach of a series of control points in the audience area and quiet areas. By minimizing the energy in the quiet zones while optimizing synthesis in the listener area, the loudspeaker drive signals are calculated.

There are also higher order Ambisonics based procedures. So will in TW Abhayapala and YJ Wu, Spatial Soundfield Reproduction with Zones of Quiet, Audio Engineering Society Convention 127, 2009 a method proposed for suppressing the wave field in a circular area in an ambisonics-based sound reproduction using a circular array. This is exemplary in Fig. 5 illustrated. In the playback room, zeros occur next to the desired wave field. This method is numerically motivated and can not be transferred to linear arrays. Furthermore, this procedure provides implicit specifications for the location of the rest area. The quiet zones may not overlap with the rendering area, so a quiet zone may not be within the propagation path of the desired rendering area, as in Fig. 5 shown. Here the thick black ring indicates the arrangement of the loudspeakers, which in the example shown are thus arranged in a circle around the listener zone. The desired wave field is a plane wave that is displayed on the left side of the listener zone. On the right is a small ring in which the sound field is suppressed.

It is an object of the present invention to overcome the limitations of the prior art and an improved method and apparatus for suppressing the sound field in certain areas of the listener zone, also referred to as quiet zones, while synthesizing a field to be displayed in a target area - ie outside the quiet zones - provide. The location of the quiet zones in the listener zone or the arrangement of the loudspeakers should be subject to as few restrictions as possible. The object of the invention is achieved by in the method and a device according to claims 1 and 8, respectively.

The invention provides a method for reproducing an acoustic wave field in a listener zone by means of several loudspeakers. Applying a spatial window function, such as a rectangular function, to the wave field to be rendered defines a wave-free section in the listener zone in which the wave field is suppressed. This can be done for example by a multiplication of the target sound field with the window function, this corresponds to a spatial convolution in the spatial frequency range.

To be able to reproduce the resulting desired wave field with the speakers, corresponding loudspeaker drive signals are determined and these

Ansteueruxagssignale supplied to the speakers. The determination of the drive signals may be performed by a higher order Ambisonics method or a spectral division method. Preferably, the desired wave field is bandlimited to increase the quality of the suppression of the sound field in the corresponding area of the listener zone.

Furthermore, the invention provides a corresponding device for reproducing an acoustic wave field in the listener zone, wherein the wave field is suppressed in one or more predetermined sections of the listener zone. In this case, the loudspeakers may be arranged on an open contour, in particular a line, or else on a contour surrounding the listener zone, that is to say for example on a circle or a rectangle.

• Fig. 1 eine Simulation der Wiedergabe einer ebenen Welle umgeben von einer ruhigen Zone;
• Fig. 2 die Energieverteilung des in Fig. 1. gezeigten Schallfeldes (in dB);
• Fig. 3 die Energieverteilung des in Fig. 1 gezeigten Schallfeldes, wenn das Wellenfeld nur in einem bestimmten Gebiet unterdrückt werden soll (in dB);
• Fig. 4 eine Simulation der Wiedergabe einer ebenen Welle umgeben von einer ruhigen Zone bei rechteckiger Arraygeometrie; und
• Fig. 5 ein mit dem im Stand der Technik bekannten Ambisonics- Verfahren wieder gegebenes Schallfeld.

The invention will be described in more detail below with reference to the attached figures. Show it:

• Fig. 1 a simulation of the reproduction of a plane wave surrounded by a quiet zone;
• Fig. 2 the energy distribution of in Fig. 1 , shown sound field (in dB);
• Fig. 3 the energy distribution of in Fig. 1 shown sound field, if the wave field is to be suppressed only in a certain area (in dB);
• Fig. 4 a simulation of the reproduction of a plane wave surrounded by a quiet zone with rectangular array geometry; and
• Fig. 5 a sound field reproduced with the Ambisonics method known in the art.

All methods of multi-channel acoustic reproduction systems use a pre-defined mixing matrix which maps the source signals to particular loudspeaker signals to produce a desired auditory impression. The derivation of the control functions in Higher-Order Ambisonics and SDM is done by explicitly solving the underlying physical problem.

The suppression of the wave field in certain zones can be interpreted as a multiplication of the original desired field with certain spatial window functions that suppress the sound field in certain areas. Considered in the spatial spectral range, this spatial multiplication is a spectral convolution and the spatial suppression is therefore done by a suitable spectral spread of the spatial spectrum of the reproduced spatial signal.

To clarify the principle of operation of the invention, the following is an example of the synthesis of acoustic wave fields using linear arrays underlying equations. In the discussion, the reproduction of the wave field is described in one plane, so z = 0. The extent of the array runs in the direction of the x-axis.

dabei stellt D die Ansteuerungsfunktion der Sekundärquellen dar und G die Green'sche Funktion, die den Ausbreitungspfad eines akustischen Monopols an der Position (x₀=(x ₀,y ₀,z ₀)) zum Punkt (x=(x,y,z))charakterisiert. Im Idealfall entspricht das synthetisiert Schallfeld dem gewünschten Schallfeld.The synthesized sound field can be understood as the result of a spatial convolution integral. $$P\left(\mathbf{x}\omega \right)={\displaystyle \underset{-\infty }{\overset{\infty }{\int }}}D\left({\mathbf{x}}_0\omega \right)G\left(\mathbf{x}-{\mathbf{<figure-callout\; id="0"\; label="x"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">x</figure-callout>}}_0.\omega \right)d{\mathbf{x}}_0.$$

where D represents the drive function of the secondary sources and G the Green's function, which determines the propagation path of an acoustic monopole at the position (x ₀ = ( x ₀ , y ₀ , z ₀ )) to the point (x = ( x , y , z )). Ideally, the synthesized sound field corresponds to the desired sound field.

eine zeitliche Fouriertransformation obiger Gleichung gefolgt von zwei räumlichen Fouriertransformationen entlang der x- und y-Achse liefern $$\bar{P}\left(k_x{k}_{y}z\omega \right)=8{\pi }^3\delta \left(k_x-k_{\mathit{pw},x}\right)\delta \left(k_y-k_{\mathit{pw},y}\right)\delta \left(\omega -{\omega }_{\mathit{pw}}\right)\mathrm{.}$$

For example, the desired field is a monochromatic plane wave, which is given by the equation: $$p\left(xyt\right)=e^{-i\left(k_{\mathit{pw}.x}x+k_{\mathit{pw}.y}-{\omega }_{\mathit{pw}}t\right)}.$$

provide a temporal Fourier transform of the above equation followed by two spatial Fourier transforms along the x and y axes $$\tilde{P}\left(k_x{k}_{y}z\omega \right)=\mathrm{8th}{\mathrm{<figure-callout\; id="3"\; label="\pi "\; filenames="imgf0004.png,imgf0005.png"\; state="\{\{state\}\}">\pi </figure-callout>}}^3\delta \left(k_x-k_{\mathit{pw}.x}\right)\delta \left(k_y-k_{\mathit{pw}.y}\right)\delta \left(\omega -{\omega }_{\mathit{pw}}\right)\mathrm{,}$$

What is desired now is a boundary of the wave field on a stripe perpendicular to the array. This can be understood as a spatial fenestration. One possible window function is the rectangular function. The spatial Fourier transformation of the rectangular function of width (1 / a) is. $${\mathrm{\Pi }}_x\left(a\mathbf{x}\right)\circ -•\frac{1}{\sqrt{2{\mathit{<figure-callout\; id="2"\; label="\pi a"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\pi a</figure-callout>}}^2}}\mathrm{sinc}\left(\frac{k_x}{2\mathit{\pi a}}\right)\mathrm{,}$$

Overall, the desired spatially limited wave field is given by the following convolution: $$\begin{array}{l}{\tilde{P}}_{\mathrm{\Pi }}\left(k_x{k}_y\omega \right)={\mathrm{\Pi }}_x\left(a\mathbf{x}\right)*\tilde{P}\left(k_x{k}_y\omega \right).\\ =\frac{1}{\sqrt{2{\mathit{<figure-callout\; id="2"\; label="\pi a"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\pi a</figure-callout>}}^2}}\mathrm{sinc}\left(\frac{k_x-k_{\mathit{pw}.x}}{2\mathit{\pi a}}\right)\delta \left(k_y-k_{\mathit{pw}.y}\right)\delta \left(\omega -{\omega }_{\mathit{pw}}\right)\mathrm{,}\end{array}$$

In order to calculate the driving functions for the speakers, the method of spectral division can be used. This is based on the fact that the convolution integral in (G1.1) can be transformed into a multiplication in the spatial Fourier domain. Thus, (G1.1) can be transformed as follows $$\tilde{D}\left(k_x{k}_y\omega \right)=\frac{{\tilde{P}}_{\mathrm{\Pi }}\left(k_x{k}_y\omega \right)}{G\left(k_x{k}_y\omega \right)}\mathrm{,}$$

The quality of the suppression of the sound field in a certain area depends on the spatial aliasing frequency of the playback system and the cut-off frequency of the reproduced material, because only the spectral components of the reproduced material, which are below the aliasing frequency of the system can be selectively suppressed. In addition, a band limitation of the material to be reproduced can be used to avoid spatial aliasing.

In a preferred embodiment of the method according to the invention, the mixing matrix on the reproduction side is determined frequency-selectively on the basis of the desired wave field, the position of the recording system and the geometry of the reproduction system. As an application example, a linear loudspeaker array is discussed, wherein the interception zone with two virtual lines, which are perpendicular to the array, should be divided into three areas. In one area, conventional flat wave reproduction should be based on the spectral division method. In the two other areas in which the recording system can be located, the wave field should be suppressed.

The desired wave field is a multiplication of the plane waves with the spatial rectangular function or another suitable window function. For the determination of Driving functions for the speakers according to the embodiment, the method of spectral division is used. This method requires an expression for both the desired wave field and Green's function in the spatially transformed region for linear array geometries. by means of the Fourier transformation along a reference line on which the reproduction should be exact. The driving functions in the spatial spectral range are the result of element-wise division of the spectrum of the desired field by the spectrum of Green's function. The thus determined spectra of the drive functions can be transformed back into the spatial domain by means of an inverse Fourier transformation.

In Fig. 1 is the result of a simulation of the presented method and shown in Fig. 2 you can see the energy distribution of the wave field. In Fig. 3 the energy distribution is reproduced when the quiet zone, ie the wave-free section, divides the display area into two areas.

Array geometries that enclose the audience space in two dimensions, such as rectangular arrays, allow two-dimensional suppression of the field. An example of this is in Fig. 4 shown. The procedure for carrying out the method according to the invention is similar to that described above. The main difference is that the window function here is two-dimensional. The same applies to the three-dimensional rendering.

The spatially limited wave field is in the spectral range a two-dimensional convolution of the desired field with the spatially transformed window function. For example, the in Fig. 4 Wave field shown a superposition of plane waves, wherein in the synthesis of the individual plane waves, the desired resting area was explicitly considered. The driving functions of the speakers for the in. Fig. 4 The plane waves shown were determined by means of the aforementioned method of spectral division.

Claims (11)

1. A method of reproducing an acoustic wave field in a listener zone comprising the suppression of the wave field in one or more predetermined portions of the listener zone by means of a plurality of loudspeakers, the method comprising the steps of:

   - determining a desired wave field in the listener zone by applying a spatial window function for forming wave-free portions on a wave field to be reproduced;

   - determining gating signals for the loudspeakers for generating the desired wave field in the listener zone; and

   - supplying the gating signals to the loudspeakers.
2. The method of claim 1, wherein the wave field to be reproduced is multiplied by the spatial window function for determining the desired wave field.
3. The method of claim 1, wherein a spatial convolution of the spectrum of the wave field to be reproduced with the window function in the spatial frequency area is performed for determining the desired wave field.
4. The method of claim 1, wherein the wave field to be reproduced is spectrally spread for forming the wave-free portion.
5. The method of any of the preceding claims, wherein the window function is a rectangular function.
6. The method of any of the preceding claims, wherein the gating signals are determined by a higher-order ambisonics method or a spectral division method.
7. The method of any of the preceding claims, wherein the desired wave field is bandlimited.
8. A device for reproducing an acoustic wave field in a listener zone comprising the suppression of the wave field in one or more predetermined portions of the listener zone by a plurality of loudspeakers, the device comprising:

   - a plurality of loudspeakers;

   - a device for supplying gating signals to the loudspeaker for generating a desired wave field comprising a wave-free portion in the listener zone, wherein the desired wave field is formed by applying a window function to a wave field to be reproduced.
9. The device of claim 8, wherein the loudspeakers are arranged on an open contour, in particular a line.
10. The device of claim 8, wherein the loudspeakers are arranged on a contour surrounding the listener zone.
11. The device of claim 10, wherein the loudspeakers are arranged on a circle or a rectangle.

Images