Apparatus and method for generating sound in a spatial domain

Abstract

The invention relates to a device for acoustically irradiating a spatial region (200), comprising at least one sound transducer array (100) with a plurality of flatly arranged individual sound transducers (11). The sound transducer array (100) is designed and can be operated so as to emit wavefronts (30) which correspond to a model with at least two virtual sound sources (1, 2) that lie behind the at least one sound transducer array (100) when viewed from the direction of the acoustically irradiated spatial region (200). The invention is characterized by a computing means (50) for actuating the individual sound transducers (11) in the at least one sound transducer array (100), wherein emission times for the same audio content are determined on the basis of at least one first virtual sound source (1) for sound transducers (11) in a horizontal arrangement (10) in the sound transducer array (100) and on the basis of at least one second virtual sound source (2) for sound transducers (11) in a vertical arrangement (20) in the sound transducer array (100). Emission times (t1) for the same audio content for the sound transducers (11) can be determined by the computing means (50) using the distance of the sound transducers to at least one first virtual sound source (1), regardless of a vertical position of the at least one virtual sound source (1) and the respective sound transducer (11), and the propagation time (t2) to each sound transducer (11) is added to the calculated values, said propagation time resulting from the distance of each sound transducer (11) to the at least one second virtual sound source (2), regardless of the respective horizontal position of the at least one first virtual sound source (1) and the respective sound transducer (11). The invention also relates to a method.

Description

[0001] The invention relates to a device for sound reinforcement of a spatial area with the features of claim 1, a method for sound reinforcement with the features of claim 9 and a computer program product with the features of claim 17.

[0002] For sound reinforcement in spatial areas, such as event halls or outdoor areas, it is known to use systems that operate on the principle of wave field synthesis. Such systems are, for example, known from WO 2015 022579 A2 and WO 2001 023104 A2.

[0003] Further information on wave field synthesis can be found in Makarski, Goertz, Weinzierl, Moldrzyk: On the development of loudspeakers for wave field synthesis. Tonmeistertagung 2008, S. Spors, R. Rabenstein, J. Ahrens: The theory of wave field synthesis revisited. In: 124th AES Convention, Amsterdam, May 2008.

[0004] The principle of wave field synthesis in the audio domain can be realized by convolving the signal into an impulse response specific to the position of each individual transducer. This allows for the synthesis of both direct sound waves and reflections from the recording space during playback. The process is mathematically complex. Therefore, in practice, it is typically reduced to a single horizontal array of transducers arranged around the listener in a heavily acoustically damped playback room. This method is unsuitable for live sound reinforcement applications due to the signal redundancy resulting from the convolution into the impulse response.

[0005] In the model-based approach of wave field synthesis, virtual sound sources can be generated that produce their elementary waves with very low latency relative to the nearest transducers. This makes the method fundamentally suitable for use in public address systems. Transducer fields based on the principle of wave field synthesis are also known.

[0006] A sound transducer array radiates sound over a surface, which can be flat or curved. A sound transducer array can also have both flat and curved sections. A sound transducer array typically comprises an array of transducers, such as loudspeakers, typically dynamic loudspeakers, arranged across the surface. However, the use of other transducer principles, such as electrostatic or piezoelectric transducers, or even microelectromechanical systems (MEMS), is also possible.

[0007] The task, therefore, is to create improved sound transducer fields and methods for sound reinforcement of spatial areas.

[0008] The invention is defined in the attached independent claims. The dependent claims define preferred embodiments of the invention.

[0009] This involves using a device for sound reinforcement of a spatial area with at least one transducer array comprising a plurality of individual transducers (e.g., loudspeakers) arranged in a planar configuration. The transducer array can, for example, have a flat or curved shape. It is also possible for the transducer array to have both curved and flat sections. The transducers can be arranged in a grid, but this is not mandatory as long as their geometric position remains precisely defined. The transducers can be identical, but this is not required.

[0010] The sound transducer array is configured and operational to emit wavefronts corresponding to a model with at least two virtual sound sources located behind the at least one sound transducer array, as seen from the area being covered. The area being covered can be, for example, an enclosed space, the audience area within an enclosed space, a part of the outdoor area, or an audience area outdoors.

[0011] A computational tool is used to control individual transducers in at least one transducer array, whereby the computational tool determines the radiation times for the same audio content for the transducers in the at least one transducer array based on their distance to at least one first virtual sound source, whereby the different vertical positions of the individual transducers in a horizontal arrangement with respect to the first virtual sound source are not taken into account, so that the radiation times can be determined independently of the vertical position of the at least one virtual sound source and the respective transducer, and that the travel time to the respective transducer, which can be determined from the distance of the respective transducer to at least one second virtual sound source, is added to these calculated values.where the different horizontal position of the individual sound transducers in a vertical arrangement with respect to the second virtual sound source is not taken into account, so that the time of flight is independent of the respective horizontal position of the at least one second virtual sound source and the respective sound transducer.

[0012] The additional latency of the audio signal caused by adding the values ​​of the transit times can be eliminated, for example, by subtracting a common value from each transit time calculated for the individual transducers, which is smaller than the smallest of all the values ​​calculated in the result of the addition.

[0013] In one embodiment, the computing device is designed such that several wavefronts with the same audio content can be generated by the sound transducer field using a common virtual sound source for the horizontal plane (azimuth plane) and at least two vertically staggered virtual sound sources for the vertical plane (elevation plane).

[0014] Such an arrangement allows for the creation of a combined wavefront, enabling effective sound coverage even in spatial areas with complex geometries. For example, each of the staggered levels can be assigned its own input channel, allowing the sound pressure levels and frequency responses of individual beams with the same audio content to be adjusted independently. This means that multiple wavefronts can be assigned their own input channel with the same signal content, while maintaining separate access to level and equalization settings.

[0015] In another embodiment, the acoustic beams in the elevation plane have different vertical opening angles, whereby the size of individual supply areas in the spatial area is adjusted to each other in order to be able to use the maximum available diaphragm deflection of the individual sound transducers in all areas.

[0016] Furthermore, in one embodiment, the computing device is designed such that the vertical curvature of the wavefronts is divided into different frequency ranges, wherein, in particular, only an inner area in the vertical center of the transducer field is equipped with transducers that radiate the upper frequency range of the audio content, the curvature of the wavefront in this frequency range can be determined by a separate, secondary virtual sound source such that its wavefronts in at least a part of the sound-covered spatial area, e.g., the audience area, have the same vertical boundary as the wavefront of the outer transducers for the same beam. This prevents linear distortion of the audio content.In particular, a time compensation between the wavefronts of the different frequency ranges can be designed such that the time errors caused by the different distances of the second virtual sound sources to the transducer field for the different frequency ranges between the wavefronts of the different frequency ranges of the same audio content are halved at the boundaries of the supply area of ​​the respective beam by shifting the virtual sound source for the lower frequency band, after correcting the time difference in the middle of the wavefront, again by half the difference between the two radii of curvature at the boundary of the supply area in the direction of the two-dimensional transducer arrangement.

[0017] For flexible use of a sound reinforcement system, in one embodiment the delay times and levels of the audio content determined in the model-based approach of wave field synthesis are converted into impulse responses in the computing tool, with which the sound transducers of a wave field synthesis system based on the convolution of the audio content into impulse responses can be controlled.

[0018] Furthermore, the computational tool can be designed such that the curvatures of the wavefronts in the elevation plane and the curvatures of the same wavefronts in the azimuth plane are different, with virtual sound sources located at different distances from the transducer field being used as reference points for calculating the delay times and the level of the same audio content for each individual transducer in the transducer field, the localization of the nearby first virtual sound source is decisive for the localization of the sound event in the azimuth plane, and the more distant second virtual sound source determines the curvature and orientation of the same wavefront in the elevation plane.

[0019] The problem is also solved by a method having the features of claim 9 and a computer program product having the features of claim 17.

[0020] In the following, embodiments of the invention are also described with reference to drawings, showing Fig. 1 a schematic representation of an embodiment of a device for sound reinforcement of a spatial area; Fig. 2 a schematic perspective representation of an embodiment of a device for sound reinforcement of a sub-area of ​​an arena-shaped area; Fig. 3 a schematic perspective representation of an embodiment of a device for sound reinforcement of an arena-shaped area with four second virtual sound sources in the elevation plane; Fig. 4 a schematic sectional view of an embodiment of a device with a determination of sound propagation in the overlap area; Fig. 5 a schematic perspective representation of an embodiment for sound reinforcement of a spatial area with different distance zones; Fig. 6 a perspective view of an embodiment with a subdivided transducer array for sound reinforcement of a spatial area; Fig. 7 a sectional view through an embodiment according toFig. 6 ; Fig. 7A a sectional view through an embodiment according to Fig. 7 with time off in lieu.

[0021] Based on Fig. 1 The basic principle of an embodiment of a device for sound reinforcement of a spatial area 200 by means of a sound transducer field 100 is presented.

[0022] The transducer field 100, which operates on the principle of wave field synthesis, generates an "acoustic curtain." From a single mono signal, and by calculating sound propagation times and levels based on the distance between a virtual sound source 1 and the respective transducer 11 in the transducer field 100, the signals that a transducer would receive from a real sound source at the position of the virtual sound source 1 from a microphone located directly behind it in a partition are derived using a model-based approach. The wavefront of a real sound source is thus reconstructed, as if by a "curtain."

[0023] In this process, each virtual sound source 1 behind this arrangement is physically reconstructed from a multitude of individual sound transducers according to Huygens' principle. The curvature of the wavefront 30 corresponds to that of a wavefront emanating from a real sound source at the position of the virtual sound source 1. Therefore, unlike phantom sound sources in psychoacoustically based methods, the virtual sound source 1 does not change its point of origin with the listener's position.

[0024] Therefore, apart from diffraction effects due to the finite area of ​​the transducer field 100, it is only audible in the area in which the virtual sound source 1, 2 is located within the transducer field 100 as seen from the listener.

[0025] Using transducer fields 100 based on the principle of wave field synthesis, the curvature and orientation of the wavefronts 30 in the azimuth direction (i.e., in the horizontal) and elevation direction (i.e., in the vertical) can be determined.

[0026] Transducer arrays 10 with individually operated transducers 11 generate cylindrical waves when a parallel wavefront is to be produced. The cylindrical waves are reflected off the ceiling and floor of a room, which is why the room must be heavily damped in order to at least partially meet the condition of a source-free volume theoretically required for wave field synthesis.

[0027] In transducer fields 100, the size of the transducer field and the position of the virtual sound source 1 determine the area in which the audience is served by this virtual sound source 1. Outside the area where transducers 11 are located between the virtual sound source 1 and the sound receiver, the sound pressure decreases very rapidly. Figuratively speaking, the virtual sound source 1 is only audible in the area where it would also be visible from the audience's position within the area of ​​the transducer field 11. In this way, a wavefront 30 can be directed precisely, for example, towards the audience area, without unintentionally hitting the room's reflective surfaces.

[0028] Due to the limited coverage area, the virtual sound source 1 in the model-based approach must be positioned close behind the transducer array to cover a wide audience area within the spatial area 200. While a distant virtual sound source 1 would have the advantage that the surface area of ​​the wavefront 30 in the audience area would increase only slowly with distance, and the level would decrease correspondingly slowly with distance from the generating transducer array 100, it would lose the key advantage of virtual sound sources 1 over phantom sound sources: their localization is independent of the listener's position. Although it would retain its position, this would then deviate significantly depending on the viewer's perspective, for example, from the azimuth of a performer on stage. Furthermore, the coverage area would be almost entirely limited to the size of the transducer array 100.

[0029] The vertical height of the transducer field 100 cannot be too low relative to the wavelength of the wavefront 30 to be synthesized. Otherwise, sufficient directivity will not be achieved in the elevation plane. This directivity in both planes is, however, a prerequisite for abandoning the requirement of a source-free volume for the playback room, which arises from the theoretical derivation of the principle of wave field synthesis. Only if the excitation of unwanted reflections can be largely avoided by selectively supplying the desired playback area will the volume of the playback area remain largely source-free. Only then is the principle applicable even in acoustically untreated rooms.

[0030] The resulting necessary vertical extension of the transducer field 100 leads to a large opening angle in the elevation plane when the virtual sound source 1 is positioned close behind it. This, in turn, is a prerequisite for covering a wide audience area in the azimuth plane and for the localization in the azimuth plane to correspond, for example, to the position of the performer on stage.

[0031] Overall, such a wavefront 30, synthesized from a single virtual sound source 1, forms a spherical section whose size and shape are determined by the dimension of the sound transducer field and the position of the virtual sound source 1.

[0032] This spherical segment is characterized by having the same radius of curvature in the azimuth and elevation planes. A spherical surface increases quadratically with its radius. Accordingly, the sound intensity decreases rapidly with distance from the virtual sound source 1 because it is distributed over the quadratically increasing spherical surface.

[0033] The advantage of wavefield synthesis, namely that wavefronts can be generated with a transducer field 100 whose surfaces grow only slowly with distance or, as in the case of a parallel wavefront, even remain constant with distance, is not realized if the virtual sound source 1 is positioned close to the transducer field 100. Conversely, the coverage area in the azimuth plane is reduced or even limited to the width of the transducer field 100 if the virtual sound source 1 is positioned far behind the two-dimensional transducer field.

[0034] This can be avoided if a curvature KE of the wavefront 30 of a transducer field 100 in the elevation plane E is decoupled from a curvature KA of the same wavefront 30 in the azimuth plane A. This is discussed in connection with the Fig. 1 depicted.

[0035] With decoupling, it is possible, for example, for a wavefront to emanate from a virtual sound source 1 located near the transducer field 100, exhibiting only slight curvature in the elevation plane E. This results in the wavefront level decreasing less sharply with distance from the transducer field 100 than would be expected from the spherical wave segment originating from this position, due to the quadratic increase in its surface area with distance. However, the localization of the primary virtual sound source 1 must be maintained from all listener positions within the audience area of ​​the spatial region 200.

[0036] The decoupling of the radii of curvature KE, KA can be achieved by having a computing device 50 control the individual transducers 11 in the at least one transducer array 100 as follows: The emission times of the same audio content for the individual transducers 11 of the transducer array 100 are determined as a function of their distance to the first virtual sound source 1, whereby the different vertical position of the individual transducers 11 with respect to the first virtual sound source 1 is not taken into account. The calculated values ​​are temporarily stored.

[0037] In the second step, the emission times of the same audio content for the individual transducers 11 in the transducer field 100 are determined as a function of their distance to at least one second virtual sound source 2, whereby the different horizontal position of the individual transducers 11 with respect to the second virtual sound source 2 is not taken into account. These calculated values ​​are also temporarily stored.

[0038] After adding the temporarily stored values, the shortest radiation time can then be subtracted to avoid unnecessary redundancy in the playback, which would arise in the virtual part of the wave field synthesis due to the travel time from the respective virtual sound source to the transducer field 100.

[0039] Then the sound propagation times from the first step are added to the sound propagation times from the second step and the audio signal, which is assigned to the beam of the second virtual sound source 2, is emitted by the individual sound transducers 11 with these delay times.

[0040] The result is a wavefront that can have a different curvature in the elevation plane than in the azimuth plane. Several such wavefronts can be generated simultaneously in a common transducer field 100, whereby the first virtual sound source 1 determines the common curvature of all these wavefronts in the azimuth plane, while independent control can be achieved in the elevation plane with a number of differently positioned additional virtual sound sources.

[0041] This radius of curvature KE in the elevation plane is independent of the position of the primary virtual sound source 1 to be localized and depends only on the distance of the secondary virtual sound source position 2, which is decisive for the curvature of the wavefronts 30 in the elevation plane E, to the transducer field 100. If this distance approaches infinity, a cylindrical wave segment is created around the position of the primary virtual sound source 1, which is curved only in the azimuth plane A.

[0042] The surface area of ​​a cylindrical wave increases linearly with its radius, not quadratically like the surface area of ​​a spherical wave segment. Therefore, the level drop of such a wavefront, apart from losses due to airborne sound insulation and diffraction, is only -3 dB per doubling of distance instead of -6 dB, as with a spherical wave segment. Depending on the position of the secondary virtual sound source, which determines the opening angle and orientation of the wavefront 30 in the elevation plane E, the theoretical level drop with distance then lies between these values.

[0043] In the embodiment of the Fig. 1 The transducer array 100 is designed as a flat surface with a rectangular shape, meaning that the individual transducers 11 lie geometrically in one plane and form a kind of array. This is not strictly necessary. In other embodiments, the transducers 11 can lie on a surface with a curvature. The surface curvature can be identical or different within the array. The transducer array 100 can also be composed of several arrays with different curvatures.

[0044] Further embodiments of the device and details of the method are described with reference to the additional figures.

[0045] In Fig. 2 is an arena-shaped audience area in an event hall, the spatial area to be covered by sound is 200.

[0046] This shows Fig. 2 The separation of the radii of curvature KA, KE of the wavefront 30 into the azimuth and elevation planes. The primary one, the localization of the sound event in the azimuth plane A (not shown separately here, see Fig. 1 The first virtual sound source 1, which determines the sound source, is positioned close to the transducer field 100, so that the entire listening area within the spatial region 200 in the azimuth plane A is supplied with sound from the wavefront emanating from it, according to the principle of wave field synthesis. In the azimuth plane A, this position determines the curvature KA of the wavefront 30, which is why it is localized at this position regardless of the listener's position. This preserves the significant advantage of virtual sound sources over phantom sound sources based on psychoacoustic effects.

[0047] A reduced curvature of the wavefronts 30, which, due to the slower increase in surface area with distance, offers the advantage of improved level stability, can only be achieved in the elevation plane E. Given a transducer field size 100 determined by the wavelength of the radiated wavefronts 30 and the geometry of the listening area 200, it becomes possible to avoid exciting unwanted reflections from the spatial area 200 and thus, in principle, fulfill the theoretical requirement of a source-free volume in the playback area. Therefore, a second virtual sound source 2 is defined for the same wavefront 30, which determines the curvature KE and orientation of the wavefront 30 in the elevation plane E. The distance of the second virtual sound source from the transducer field 100 then determines the opening angle of the wavefront 30 in the elevation plane E, its vertical position, and its vertical orientation.

[0048] Both virtual sound sources 1 and 2 have the same audio content. Therefore, the second virtual sound source 2 of wavefront 30 does not need to be assigned a separate audio channel.

[0049] In the model-based approach of wave field synthesis, the value for the respective delay and the level of the signal are determined in a first step for each individual transducer 11 in the azimuth plane, independently of the vertical axis. For each individual transducer 11 in the elevation plane, the respective delay and the level of the signal are determined independently of the horizontal axis.

[0050] In a second step, the delay times related to the second virtual sound source 2 are added to these values, which depend on the azimuth position of the first virtual sound source 1. These are also calculated only on a single line, this time vertically at the azimuth position of the second virtual sound source 2.

[0051] Because of the greater distance between the second virtual sound source 2 and the transducer field 100, adding both values ​​would introduce additional latency to the audio signal. This would be unacceptable for live events. Therefore, before adding the delay values ​​of the elevation plane E, the smallest of these delay values ​​is subtracted from all calculated elevation delay values. The reduced values ​​then represent the curve curvature in the elevation plane E without unnecessarily increasing the propagation time and thus the latency of the system.

[0052] The calculated time-of-flight values ​​of the second virtual sound source 2 are then adjusted according to the vertical position of the respective sound transducer 11 (i.e., in the vertical arrangement 20 in Fig. 2 ) is added to the runtime of the first virtual sound source 1 on the azimuth plane A. When the audio signal for the common wavefront 30 is fed to the power amplifiers of the respective transducers 11 with the delay times assigned to the position of the respective transducers 1, 2, the wavefront 30 with its different radii of curvature KA, KA is created.

[0053] High level stability with distance is achieved when the second virtual sound source 2 is positioned far away from the transducer field 100. This reduces the curvature KA of the wavefront 30 in the elevation plane E, but also the vertical opening angle of the wavefront 30. This will usually result in the audio signal not reaching the entire audience area within the spatial region 200 in this plane. Therefore, in one embodiment, several wavefronts with different radii of curvature are arranged one above the other.

[0054] In Fig. 3 This is illustrated using an arena sound system similar to the forms described above. The first virtual sound source 1, which determines the localization, can be assigned further independent second virtual sound sources 2a, 2b, 2c, 2d for the elevation plane E.

[0055] In the listening area 200, several planes of wavefronts 30a, 30b, 30c, 30d with slight curvature but different orientations in the elevation plane E are generated. The delay times for the audio signal are calculated in the same way as described for a single wavefront 30. The curvature of all wavefronts 30a, 30b, 30c, 30d in the azimuth plane A is further calculated by determining the horizontal distance of the first virtual sound source 1 to the respective transducer 11 in the vertical arrangement 20 (see Fig. 1 ) calculated. According to their vertical position in the transducer field 100, the corresponding delay time, which results from the position of the second virtual sound source 2a, 2b, 2c, 2d, is added to the respective value.

[0056] The audio signal itself can be provided by a single input channel. However, an advantageous embodiment of the invention is to assign a separate input channel to each of the wavefronts 30a, 30b, 30c, 30d thus generated. The signal content should be the same, but the separate access to the level and equalization of each individual wavefront 30a, 30b, 30c, 30d offers significant advantages. The sound pressure can be very effectively balanced between the front and rear seats of the audience area 200. At long distances, the loss due to airborne sound insulation in the upper transmission range can also be compensated for by appropriate equalization of the input signal of the respective wavefront 30a, 30b, 30c, 30d.

[0057] Despite this separation of the input signals, wavefronts 30a, 30b, 30c, and 30d are supplied with coherent signals. Due to the inherent principle, overlapping areas occur at the boundaries of the individual wavefronts 30a, 30b, 30c, and 30d. The height of the transducer field 100 in the vertical direction, relative to the wavelength of the audio signal, determines the sharpness of the transition region.

[0058] The superposition of two coherent signals only avoids interference-induced cancellations and peaks in the signal's frequency response if both wavefronts 30a, 30b, 30c, 30d arrive at the listener simultaneously. This issue must be considered when positioning the virtual sound sources 2a, 2b, 2c, 2d that determine the wavefront elevation.

[0059] In Fig. 4 The spectator arena has 200 seats according to regulations. Fig. 3 The diagram is shown in cross-section. Behind the transducer array 100, the first virtual sound source 1 determines the localization in the azimuth plane A. The distance between a second virtual sound source 2a, 2d for the elevation plane E and the transducer array 100 then determines the vertical opening angle of the beam, i.e., the directional sound.

[0060] To ensure that the unavoidable overlap with the vertically adjacent beam in the spectator area 200 remains largely free of comb filter effects in the frequency response caused by phase-shifted addition of the coherent signals, a circular arc around the intersection point B of the second virtual sound source 2a is drawn to determine the position of the adjacent virtual sound source. The extended line L from the intersection point B through the lower edge of the transducer array 100, with the circular arc around B, then represents the position of a further second virtual sound source 2d for the beam above. From the spectators in the overlap area of ​​the two beams, the two second virtual sound sources 2a and 2d are then equidistant, so that the sound waves emanating from them arrive simultaneously. Therefore, no time-of-flight interference of the coherent signals occurs at the intersection point B of the two beams.

[0061] In the same way, the starting points of further levels are determined until the entire audience area 200 is supplied with audio signals. It is important to note that the subtraction of the latency times in the virtual part must not be performed independently for each beam. This would cause the arrival times of the individual beams to shift relative to each other. First, all arrival times of all virtual sound sources 2a, 2b, which determine the curvature of the wavefronts 30a, 30d in the elevation plane E, are determined for all transducers 11 along their vertical line in the transducer field 100. Then, the common minimum value from all these delay times is subtracted from each calculated delay time for each individual transducer 11. This prevents unnecessary latency while preserving the temporal relationships between the individual beams.

[0062] According to the described method, the ratio of the beam opening angles to each other is fixed by the geometric relationship between the audience area (200) and the vertical height of the sound transducer field (100). This also determines the sound pressure level achievable in each individual beam.

[0063] In practice, this leads to the problem that the individual beams, with the same vertical opening angle, cover audience areas of very different sizes. For example, on a flat audience area, the beam for the rear audience area must cover a much larger area than the beam for the front audience area. To achieve a balanced sound pressure level, the diaphragm excursion of the individual transducers 11 in the transducer array 100 is then significantly greater for the more distant areas, which in turn limits the maximum achievable sound pressure level in those areas. For a uniform level distribution, the beams covering smaller audience areas must then be attenuated to this sound pressure level. This reduces the maximum possible sound power of the overall system.

[0064] In the Fig. 5 The illustration shows, as an example, the sound coverage of a flat audience area 200 from a transducer array 100 positioned above the stage area. To utilize the full power of the transducer array 100 across the entire audience area 200, the audience area is divided into three approximately equal sections A, B, and C. Each section is assigned a virtual sound source 2a, 2b, 2c, which determines the curvature of the wavefront in the elevation plane E. To ensure that their coverage areas are of equal size, a different distance to the transducer array 100 is defined for each section due to the different beam angles.

[0065] To determine the positions of the virtual sound sources 2a, 2b, 2c, the audience area 200 is first divided into approximately equal areas A, B, and C, which are located at different distances from the position of the sound transducer field 100. The intersection of the extended lines from the lower and upper boundaries of area AB, BC to the lower and upper edges of the sound transducer field 100 is then the position of the corresponding virtual sound source 2a, 2b, 2c.

[0066] However, the latency of the individual wavefronts resulting from their differing distances to the transducer array cannot be corrected in the same way as if they had the same opening angles. In the overlapping regions, they would arrive at different times. This is because the wavefronts of the two virtual sound sources originate from the bottom edge of the transducer array, one from its top edge, and the other from its bottom edge. Therefore, the path of each wavefront to the listener is of a different length in the respective overlapping region. A time compensation referenced to the center of the transducer array would thus lead to comb filtering effects in the frequency response, associated with these timing errors. The radii Ra, Rb, and Rc around the boundaries AB and BC of the overlapping regions therefore determine the time correction for the individual wavefronts.

[0067] To determine the delay times, the distances between the acoustic centers of the individual sound transducers are calculated in a first step, and the resulting sound travel times are calculated in the virtual part of the wave field synthesis. From the buffered values ​​assigned to the virtual sound source A, the difference in radii through points A and B is then subtracted by the overlap region AB. This ensures that there is no time difference between these wavefronts in the elevation plane within the overlap region AB.

[0068] In the second step, the transit time, which corresponds to the difference between radii B and C around the overlap area BC, is subtracted from these temporarily stored values ​​so that their wavefronts arrive simultaneously in this area. The same procedure is followed if further virtual sound sources are needed to further divide the playback area.

[0069] After the final time equalization between the virtual sound sources for the elevation plane, the travel time between the virtual sound source closest to the transducer array and the nearest transducer is subtracted from all calculated and buffered values. This ensures that no unnecessary latency remains in the elevation plane values.

[0070] In the azimuth plane, the propagation time is calculated and temporarily stored for each transducer, regardless of its vertical position. By subtracting the shortest calculated propagation time from all other propagation times, the unnecessary signal latency that would result from the distance of the virtual sound source to the transducer array is eliminated.

[0071] Even after adding the calculated values ​​for the azimuth plane and the values ​​for the different beams in the elevation plane, wavefronts with the shortest possible latency times result.

[0072] In a further embodiment, the vertical curvature KE of the wavefronts 30 within a transducer field 100 is divided into different frequency ranges. The high complexity of a transducer field 100 based on the principle of wave field synthesis arises primarily from the need to arrange the transducers 11 in close proximity to avoid unwanted aliasing effects and side lobes of the directional characteristic. The matching of the transducer geometry to the wavelength of the signal is known from other audio applications. The long-wavelength frequency range is radiated by a widely extended transducer arrangement, while the upper frequency range is radiated only in the center.

[0073] In wave field synthesis, it is possible to significantly reduce the spacing between individual transducers 11 in the upper frequency range by using different beam opening angles in the elevation plane E from transducers 11 of varying heights and divided into different frequency ranges. This significantly improves the controllability of the individual wavefronts. The aliasing effects caused by the finite spacing of the elementary waves in wave field synthesis are significantly reduced with the same hardware effort.

[0074] Figur 6 This shows such a radiation pattern, divided into two frequency ranges, with two superimposed wavefronts of different vertical curvature. Above the stage area, at the front of Arena 201, a full-range section 102 with closely spaced high-frequency transducers is arranged between the low- and mid-frequency section 101 of the transducer array 100.

[0075] If only the second virtual sound sources 2a and 2b determined the curvature of the entire wavefronts in the elevation plane E, large audience areas would not be supplied with the upper frequency components. Therefore, this frequency range must be processed separately. The curvature of the wavefronts in this frequency range is then determined by the virtual sound sources 6 and 7.

[0076] Fig. 7 represents the situation for a single beam in the side view according to the embodiment of the Fig. 6 dar.

[0077] The size of the audience area 200 to be covered by the respective beam is determined by the position of the second virtual sound source 2a, which determines the vertical opening angle of the beam. An extended straight line from the lower and upper boundary lines of the beam in the audience area 200 through the lower and upper boundary of the area 102 of the transducer field 100, which is also equipped with high-frequency transducers, then determines, at its intersection point, the position of the respective virtual sound source 2b for the high-frequency range.

[0078] If the sound propagation time corresponding to the distance difference S is subtracted from all calculated travel times for the virtual sound source 2a from the sound source to the respective transducers 11, both wavefronts arrive simultaneously in the center of the coverage area within the spatial area 200, i.e., without phase shift. The selected audience area is supplied with all frequency components. Again, the smallest of all calculated delay times can then be subtracted from all calculated values ​​to reduce signal redundancy.

[0079] However, it should be noted that even after this, the wavefront of virtual sound source 2b has a smaller radius of curvature for the high-frequency range than the wavefront emanating from virtual sound source 2a. This results in the two signals no longer being in phase at the edges of the supply area. Comb filtering effects in the crossover region would then be the consequence.

[0080] For this reason, the vertical opening angle, determined by the position of the virtual sound source 2a, must not be too large. At small angles, the time difference is initially very small, but it increases rapidly with increasing opening angle. Depending on the given geometric conditions, the opening angle must then be chosen so that the slight time shift V occurring at the edges of the coverage area in the crossover frequency range leads to compensable level errors, but not to mutual cancellation of the signals. In addition to limiting the vertical coverage angle, a low crossover frequency between the virtual sound sources 2a and 2b also helps to prevent cancellation of the signal components under given geometric conditions.

[0081] Another solution to reduce the physically caused problem is, after compensating the travel times between the second virtual sound sources 2a, 2b, to additionally increase the signal of virtual sound source 2a by half the sound travel time difference (see Fig. 7A ) at the boundaries of the coverage area. This gives the wavefront from virtual sound source 2b in the center of the coverage area a small lead W ahead of the wavefront of virtual sound source 2b. As a result, the time difference is halved at the edges of the coverage area. An equally large difference then arises in the center of the beam in the opposite direction, creating two points in phase between. The halved time differences shift the frequency of the cancellation upwards by one octave.

[0082] The division into frequency ranges is not fundamentally limited to two areas. The transducer array 100 can also be supplemented by additional transducers 11, which radiate exclusively the low-frequency audio range from their upper and lower edges. The time alignment correction then occurs analogously to the procedure described for the mid- and high-frequency ranges. However, it is essential to ensure that the entire frequency range is always radiated in the vertical center of the two-dimensional transducer arrangement. Otherwise, the upper and lower rows of transducers would radiate cylindrical waves that would propagate uncontrollably into the adjacent supply areas.

[0083] It is also possible to combine the beams, divided into frequency ranges, with different radiation angles. All the described methods have in common that the different wavefronts arrive simultaneously in the audience area, but are generated at slightly staggered times in the transducer field 100. At this point, the signals of the individual beams are shifted relative to each other in the upper frequency range. This has the advantage that not all diaphragm excursions add up linearly, as they would with coherent signals.

[0084] In a transducer field 100 based on the principle of wave field synthesis, the inherent advantage arises that the matching of the individual transducers 11 to the air resistance improves with increasing signal wavelength. Because adjacent transducers 11 operate nearly synchronously in this frequency range, their diaphragms no longer operate with a completely mismatched, almost load-free response, as the air in front of the diaphragm can move freely in all directions. The adjacent transducers 11 generate the same sound pressure simultaneously, so that the weight of the air column in front of the diaphragm forms a better-matched load resistance.

[0085] Unfortunately, this effect is not noticeable in the high-frequency range because the driver diaphragms are inherently better adapted to the short wavelength. Furthermore, the movement of adjacent diaphragms in this range is no longer in phase. Therefore, the inherent improvement in the efficiency of the transducers 11 within a transducer field 100 does not materialize in the upper frequency range. Consequently, a comparatively high level of acoustic power must be generated here to achieve the same level as in the mid- and low-frequency ranges. For this reason, a non-coherent superposition of the diaphragm excursions in this range is a significant advantage of the method described here. Doubling the acoustic power then requires only 3 dB more power, not 6 dB as with coherent superposition.

[0086] Furthermore, the acoustic performance of the system is directed only towards the audience area, which fundamentally increases the efficiency of the system and at the same time helps to ensure that the requirement for the source-free volume of the playback room in wave field synthesis is almost fully met.

[0087] The described method refers to the model-based approach of wave field synthesis. However, the delay times and level values ​​determined using this method can also easily be generated in the form of an impulse response. Therefore, the described method is not limited to the model-based approach of wave field synthesis, as the values ​​can be converted to an impulse response method.

[0088] Other possible forms of training are described in the following requirements. Bezugszeichenliste

[0089] 1. First virtual sound source 2. Second virtual sound source 2. Virtual sound source for determining the curvature of the wavefront in the upper plane 2b. Virtual sound source for determining the curvature of the wavefront in the second plane 2c. Virtual sound source for determining the curvature of the wavefront in the third plane 2d. Virtual sound source for determining the curvature of the wavefront in the lower plane 10 Horizontal arrangement of transducers in a transducer field 11 Transducers in the transducer field 20 Vertical arrangement (in elevation direction) of transducers in the transducer field 30 Wavefront 30 a resultant wavefront of the upper plane 30 b resultant wavefront of the second plane 30 c resultant wavefront of the third plane 30 d resultant wavefront of the lower plane 50 Calculating tools 100 Transducer field 101 Low-midrange area of ​​a transducer field 102 Full-range area of ​​a transducer field 200 Spatial area 201 Front of a spatial area AAzimuth plane BIntersection point EElevation plane KAKcurvature of the wave field in the azimuth plane KEKcurvature of the wave field in the elevation plane LLine for determining the overlap Ra, Rb, RcRadies of the boundaries between planar areas SDistance difference between virtual sound sources VTime shift WTime shift

Claims

1. Apparatus for generating sound in a spatial area (200) comprising at least one sound transducer field (100) with a plurality of individual acoustic transducers (11) arranged in a plane, wherein the sound transducer field (100) is configured and operable to emit wave fronts (30) that correspond to a model having at least two virtual acoustic sources (1, 2) which, as seen from the spatial area (200) supplied with sound, are located behind the at least one sound transducer field (100), such that a curvature (KE) of the wave front (30) of the sound transducer field (100) in an elevation plane (E) can be decoupled from a curvature (KA) of the same wave front (30) in an azimuthal plane (A), characterized by computing means (50) for driving the individual acoustic transducers (11) in the at least one sound transducer field (100), wherein radiation times for the same audio content for acoustic transducers (11) in the sound transducer field (100) can be defined on the basis of at least a first virtual acoustic source (1) and can also be defined for acoustic transducers (11) in the sound transducer field (100) on the basis of at least a second virtual acoustic source (2), and wherein the computing means (50) enables determining radiation times (t1) for the same audio content for the acoustic transducers (11) in the at least one sound transducer field (100) by means of their distance from the at least one first virtual acoustic source (1), wherein the different vertical positions of the individual acoustic transducers (11) in a horizontal arrangement (10) relative to the first virtual acoustic source (1) are not taken into account, such that the radiation times (t1) can be determined independently of a vertical position of the at least one virtual acoustic source (1) and the respective acoustic transducer (11), and such that the propagation time (t2) to the respective acoustic transducer (11), which can be determined from the distance of the respective acoustic transducer (11) to the at least one second virtual acoustic source (2), is added to said calculated values, wherein the different horizontal positions of the individual acoustic transducers (11) in a vertical arrangement (20) relative to the second virtual acoustic source (2) are not taken into account, such that the propagation time (t2) results irrespective of the respective horizontal position of the at least one second virtual acoustic source (2) and of the respective acoustic transducer (11).

2. Apparatus according to claim 1, characterized in that the computing means (50) is configured such that the sound transducer field (100) can generate multiple wave fronts (30) with identical audio content by means of a common virtual acoustic source for the horizontal plane and at least two vertically staggered virtual acoustic sources for the vertical plane.

3. Apparatus according to claim 2, characterized in that a dedicated input channel can be assigned to multiple wave fronts (30a, 30b, 30c, 30d) with identical signal content, wherein separate access to level and equalization is provided.

4. Apparatus according to any of the preceding claims, characterized in that beams in the elevation plane (E) have different vertical opening angles, enabling the size of individual coverage zones (A, B, C) in the spatial area (200) to be adjusted relative to each other in order to utilize, in all zones, a maximum available diaphragm excursion of the individual acoustic transducers (11).

5. Apparatus according to any of the preceding claims, characterized in that the computing means (50) is configured such that the vertical curvature of the wave fronts (30a, 30b, 30c, 30d) is divided into different frequency ranges, wherein in particular only an inner zone (102) in the vertical center of the sound transducer field (100) is equipped with acoustic transducers (11) that radiate an upper frequency range of the audio signal, wherein the curvature of the wave front (30a, 30b, 30c, 30d) can be determined in said frequency range by a dedicated secondary virtual acoustic source (2a, 2b, 2c, 2d), such that its wave fronts (30a, 30b, 30c, 30d)) exhibit, in at least a partial region of the spatial area (200) supplied with sound, particularly in the listener area, the same vertical boundary as the wave front of the acoustic transducers (11) located at the outer region for the same beam.

6. Apparatus according to claim 5, characterized in that a time compensation between the wave fronts (30a, 30b, 30c, 30d)) of the different frequency ranges is configured such that the timing errors arising from the different distances between the second virtual acoustic sources (2a, 2b, 2c, 2d) and the sound transducer field (100), at the boundary of the coverage zone of the respective beam, are halved by shifting the virtual acoustic source for the lower frequency band toward the boundary of the coverage zone in the direction of the two-dimensional sound transducer arrangement, after correcting the timing deviation at the center of the wave front by half of the difference between the two radii of curvature.

7. Apparatus according to any of the preceding claims, characterized in that the temporal delays defined in the model-based wave field synthesis approach and the levels of the audio content are converted in the computing means (50) into impulse responses with which the acoustic transducers (11) of a wave field synthesis system based on convolution of the audio signal with impulse responses can be driven.

8. Apparatus according to any of the preceding claims, characterized in that the computing means (50) is configured to realize the shapes of the wave fronts (30a, 30b, 30c, 30d)) in the elevation plane and the curvatures of the same wave fronts in the azimuthal plane, wherein virtual acoustic sources (1, 2) having different distances from the sound transducer field (100) are taken into account for calculating the delay times and / or levels of the same audio content for each individual acoustic transducer (11) of the sound transducer field (100) as reference points for calculating the delays in both the azimuthal and the elevation plane, wherein the location of the first, closer virtual acoustic source (1) determines the localization of the acoustic event in the azimuthal plane, and the further-away second virtual acoustic source (2, 2a, 2b, 2c, 2d) defines the curvature and orientation of the same wave front in the elevation plane.

9. Method for generating sound in a spatial area (200) with at least one sound transducer field (100) comprising a plurality of individual acoustic transducers (11) arranged in a plane, which is configured and operable to emit wave fronts (30) corresponding to a model having at least two virtual acoustic sources (1, 2) that, when viewed from the spatial area (200) supplied with sound, are located behind the at least one sound transducer field (100), characterized in that a) a computing means (50) determines radiation times for the same audio content for acoustic transducers (11) in the sound transducer field (100) as a function of a first virtual acoustic source (1), and determines such times for acoustic transducers (11) in the sound transducer field (100) as a function of at least one second virtual acoustic source (2), and b) the computing means (50) determines propagation times (t1) for the same audio content for the acoustic transducers (11) based on their distance from the at least one first virtual acoustic source (1), wherein the different vertical positions of the individual acoustic transducers (11) in a horizontal arrangement (10) relative to the first virtual acoustic source (1) are not taken into account, such that the radiation times (t1) can be determined independently of a vertical position of the at least one virtual acoustic source (1) and the respective acoustic transducer (11), and c) the propagation time (t2) to the acoustic transducer (11) is added to said calculated values, the propagation time being derived from the distance of the respective acoustic transducer (11) from the at least one second virtual acoustic source (2), wherein the different horizontal position of the individual acoustic transducers (11) in a vertical arrangement (20) relative to the second virtual acoustic source (2) are not taken into account, such that the propagation time (t2) results irrespective of the respective horizontal position of the at least one first virtual acoustic source (1) and the respective acoustic transducer (11).

10. Method according to claim 9, characterized in that the computing means (50) simultaneously generates multiple wave fronts (30) with identical signal content from two virtual acoustic sources (1, 2), which are separated for the azimuthal and elevation plane and vertically staggered in the sound transducer field (100).

11. Method according to claim 9, characterized in that a dedicated input channel is assigned to multiple wave fronts (30a, 30b, 30c, 30d) with identical signal content, with separate access to level and equalization.

12. Method according to any of claims 9 to 11, characterized in that beams in the elevation plane (E) have different vertical opening angles, enabling individual coverage zones (A, B, C) to be dimensioned relative to each other such that in all zones the maximum diaphragm excursion of the individual acoustic transducers (11) can be utilized.

13. Method according to any of claims 9 to 12, characterized in that the computing means (50) is configured such that the vertical curvature of the wave fronts (30a, 30b, 30c, 30d) is divided into different frequency ranges, wherein in particular only the inner zone (102) in the vertical center of the sound transducer field (100) is equipped with acoustic transducers (11) that radiate the upper frequency range of the audio signal, and wherein the curvature of the wave front (30a, 30b, 30c, 30d) in said frequency range is determined by a respective secondary virtual acoustic source (2a, 2b, 2c, 2d) such that its wave fronts (30a, 30b, 30c, 30d) have, in the spatial area (200) supplied with sound, the same vertical boundary as the wave front of the outer acoustic transducers (11) for the same beam.

14. Method according to claim 13, characterized in that a time compensation between the wave fronts (30a, 30b, 30c, 30d) of the different frequency ranges is configured such that the timing errors resulting from the different distances between the second virtual acoustic sources (2a, 2b, 2c, 2d) and the sound transducer field (100), at the boundaries of the coverage zone of the respective beam, are halved by shifting the virtual acoustic source for the lower frequency range at the boundary of the coverage zone in the direction of the two-dimensional transducer arrangement after correcting the timing difference at the center of the wave front by half of the difference between the two radii of curvature.

15. Method according to any of claims 9 to 14, characterized in that the delay times defined in the model-based wave field synthesis approach and the levels of the audio content are converted in the computing means (50) into impulse responses with which the acoustic transducers (11) of a wave field synthesis system based on convolution of the audio signal with impulse responses can be driven.

16. Method according to any of claims 9 to 15, characterized in that the computing means (50) implements the shapes of the wave fronts (30a-30d) in the elevation plane and the curvatures of the same wave fronts in the azimuthal plane differently, wherein virtual acoustic sources (1, 2) with different distances from the sound transducer field (100) are used as reference points for determining delay times and / or levels of the same audio content for each acoustic transducer (11), wherein the localization of the first, closer virtual acoustic source (1) determines the localization of the acoustic event in the azimuthal plane, and the further-away second virtual acoustic source (2, 2a, 2b, 2c, 2d) determines the curvature and orientation of the same wave front in the elevation plane.

17. Computer program product comprising instructions which, when executed by a computer, cause the computer to carry out a method according to any of claims 9 to 16.

Images