Microphone array

Abstract

For specific applications, such as e.g. in a sports stadium, a microphone array having particularly high directivity in the vertical direction and a high, but broadly adjustable, directivity in the horizontal direction is provided. The microphone array (100) has a plurality of microphones (110), the output signals of which are combined to produce at least one joint output signal (360). The microphones are directional microphones having a preferred direction of high sensitivity (115) and arranged substantially in one plane on a circle (120) or circle segment, so that each microphone has a different preferred direction of high sensitivity. In this case, the preferred direction of high sensitivity (115) for each of the microphones lies substantially orthogonally in relation to the circle or circle segment. A joint output signal (360) of the microphone array is obtained by beamforming (310,..., 350). The microphone array (100) has an adjustable preferred direction of high sensitivity, wherein the joint output signal (360) contains the sound picked up from this adjustable direction.

Description

[0001] The invention relates to a microphone array. background

[0002] When recording audio in large sports venues, the acoustic events on the field can be particularly interesting for immersive playback, such as the sounds of the ball, the stick, etc., and the conversations of players, referees, coaches, and so on. However, achieving good sound quality and speech intelligibility is difficult due to the numerous ambient noises. This is because microphones often have to be positioned at the edge of the playing field, as a considerable distance from the desired sound sources must be maintained. The background noise essentially consists of the noise from the audience, which is usually located in the stands. Furthermore, the microphones used for recording audio should not obstruct the view of the audience or the cameras that are typically present.

[0003] A typical example is the playing field of a football stadium, where ball sounds, player conversations, the referee's whistle and coach's instructions are to be recorded.

[0004] Similar problems can occur in other sports such as baseball, or in other situations where sound recordings are to be made of sound sources that are widely distributed over a flat surface and may be moving, and which cannot be directly fitted with a microphone despite background noise.

[0005] A solution known as "KICK" from the company LAWO consists of a setup of numerous directional or supercardioid microphones distributed around the perimeter of a soccer field, parallel to the ground (https: / / www.lawo.com / en / products / audio-production-tools / kick.html). To capture the sound of the ball, its position is tracked automatically or semi-automatically using visual tracking. The positional data is fed into an automatic audio mixing unit, which also receives the microphone output signals and processes, weights, and mixes them according to the positional data. The underlying principle is that signals from microphones closest to the current position of the ball are given particular weight. A disadvantage of this well-known solution is the extensive cabling required. The cables and microphones must be set up before each game and dismantled afterward.Additional microphones require additional cables and increase the system's cost. Furthermore, the fixed microphone placement means their optimal pickup area must be relatively wide to cover areas between adjacent microphones. Even so, these areas are captured with poor sound quality and are therefore suboptimal. Moreover, a larger horizontal pickup area (azimuth angle) also increases the vertical pickup area (elevation angle), since the directional characteristics of common microphones are rotationally symmetrical. This results in noise from the upper tiers of the stands being picked up.

[0006] Another possible solution involves manually aligning or tracking directional microphones with particularly high directivity. However, this introduces a time delay. Furthermore, manual alignment requires an operator for each microphone, and handling noise can be transmitted to the microphone during this process. Remote control for tracking the microphones would introduce additional delays as well as motor noise, which would inevitably be picked up by the microphone and become audible as interference. Incorrect alignment of a directional microphone affects different frequencies differently because the directivity of directional microphones is stronger at higher frequencies than at lower ones. This results in a constant change in the timbre of the audio signal.

[0007] Another well-known solution for achieving high directivity is beamforming. This involves combining the output signals of several microphones arranged in an array, for example by means of delay, addition, and filtering.

[0008] The resulting beam, i.e., the area of ​​particularly high sensitivity, has an adjustable direction and is usually rotationally symmetric. The specific shape of the beam depends on the type, number, and arrangement of the microphones, as well as the algorithm used for combining them. Common algorithms include the delay-and-sum (DS) algorithm and the minimum variance distortionless response (MVDR) algorithm, each of which also has its drawbacks. Microphone arrays are typically built from microphones with no or low directivity because they are easy to use and inexpensive. However, achieving high directivity over a wide azimuth angle and comparable directivity with respect to elevation requires a very large number of microphones, resulting in high computational costs.

[0009] It is therefore an object of the present invention to provide a microphone arrangement that solves the problems mentioned above.

[0010] To make multichannel audio recordings, e.g., for 22 channels, an arrangement with circularly arranged shotgun microphones is known (Y. Sasaki, T. Nishiguchi, K. Ono: "Development of multichannel single-unit microphone using shotgun microphone array"). In this setup, adjacent shotgun microphones are used to further narrow the rotationally symmetrical directional characteristic of each individual microphone to its respective direction at low frequencies through filtering. In another known solution (K. Niwa, Y. Koizumi, K. Kobayashi, H. Uematsu: "Binaural sound generation corresponding to omnidirectional video view using angular region-wise source enhancement"), shotgun microphones are used as an alternative to beamforming.

[0011] US 2017 / 076720 A1 describes voice-activated devices which have a microphone array with a plurality of microphones.

[0012] WO 2005 / 013058 A1 describes adaptive beamforming of a microphone array.

[0013] WO 2009 / 009568 A2 describes a microphone array with a plurality of microphones and a beamforming system.

[0014] Meyer et al., "Beamforming for a circular microphone array mounted on spherically shaped objects", The Journal of the Acoustical Society of America, American Institute of Physics, Volume 109, No. 1, January 1, 2001, pages 185 - 193, describes a microphone array with a plurality of microphones. Summary of the invention

[0015] One object of the present invention is to provide a microphone arrangement with particularly high directivity in the vertical direction and a high, but within wide limits adjustable, directivity in the horizontal direction.

[0016] The problem is solved by the microphone array specified in claim 1.

[0017] Thus, a microphone array with a plurality of microphones is provided, the output signals of which are combined to form at least one common output signal. The microphones are directional microphones, each with an interference tube and a microphone capsule exhibiting a preferred direction of high sensitivity towards the interference tube. The microphones are arranged substantially in one plane. The microphones are arranged such that each microphone has a different preferred direction of high sensitivity. The microphones are arranged such that their microphone capsules are located on a circle or segment of a circle, and for each microphone, the preferred direction of high sensitivity is substantially orthogonal to the circle or segment. The common output signal is obtained by modal beamforming.The microphone array has at least one adjustable preferred direction of high sensitivity, with the common output signal containing the sound picked up from this at least one adjustable direction. The interference tubes are arranged radially in a plane.

[0018] The interference tubes have lateral openings on one of their top surfaces to detect sound arriving from the side. The top surfaces with the lateral openings of the interference tubes all point in the same direction.

[0019] Due to the high directivity of the shotgun microphones, both the elevation angle and the azimuth angle of the arrangement's detection range are very small, while the azimuth angle is adjustable over a very large range, up to 360°. The resulting azimuth directionality of the microphone array can be stronger than that of a single shotgun microphone, even if none of the shotgun microphones is pointing in the corresponding direction. In embodiments where the microphones are arranged in a full circle, some shotgun microphones always point in the opposite direction to the actual target direction. This enables a consistent directional characteristic regardless of the orientation of the microphone array. A method for audio recording using shotgun microphones is specified in claim 13. The use of a microphone array is specified in claim 14.

[0020] Further advantageous embodiments are specified in the dependent claims and in the following detailed description. Brief description of the drawings

[0021] Further details and advantageous embodiments are shown in the drawings. These show Fig. 1 a microphone array in a first embodiment; Fig. 2 a shotgun microphone with interference tube; Fig. 3 a block diagram of signal processing for the beamforming algorithm; Fig. 4 a microphone array in a second embodiment; Fig. 5 a microphone array; Fig. 6 a microphone array in one embodiment; Fig. 7 a block diagram of multi-focus signal processing for the beamforming algorithm; Fig. 8 a diagram of the radial components of modal responses of a Sennheiser MKH8070 shotgun microphone; Fig. 9 a microphone array in one embodiment; and Fig. 10 a perspective view of a microphone array in one embodiment. Detailed description of the invention

[0022] Fig. 1Figure 1 shows, by way of example, a circular microphone array 100 with thirty-one directional microphones 110, wherein both the microphone array 100 and each individual directional microphone 110 exhibit very high directivity. Each of the directional microphones 110 contains a microphone capsule, the microphone capsules of all directional microphones 110 being arranged on a circle 120 with radius r around a center point C. Furthermore, each directional microphone 110 contains an interference tube that is orthogonal to the circle 120 and directed radially outwards. The interference tube provides the directional characteristic of the respective directional microphone. The microphones are therefore also referred to as directional microphones. The preferred direction of high sensitivity of each directional microphone is in its respective longitudinal direction, i.e., also orthogonal to the circle 120 or radial to the entire arrangement. Thus, each microphone has a different preferred direction of high sensitivity.The directional microphones are distributed essentially evenly around the circle, so that the angles between each microphone are equal, e.g., 360° / 31 = 11.6°. Furthermore, in one embodiment, all directional microphones can be arranged essentially in a common plane. The entire arrangement is positioned, for example, in a football stadium, essentially horizontally, so that the directional microphones are aligned parallel to the ground.

[0023] Alternatively, the shotgun microphones could be arranged in two or more different planes. These planes should preferably be close together. The microphones can also be arranged in completely different planes, but then the sensitivity of all microphones with respect to a given elevation should be similar. In other words, the "lines of sight" or focal areas of the different microphones should all lie essentially in the same plane at a given distance.

[0024] The radius of the circle (or circular segment) determines the aliasing frequency and the operating frequency range. A larger radius with a constant number of directional microphones improves low-frequency performance by shifting this range to lower frequencies and resulting in a lower aliasing frequency. Increasing the number of microphones leads to a higher aliasing frequency.

[0025] Fig. 2Figure 1 shows an example of a single directional microphone 200, which can be used as a directional microphone 110 in the arrangement 100. The directional microphone 200 contains a tube 210 acting as an interference tube, with a microphone capsule 240 located inside it (not visible in the drawing). The microphone capsule can be electrically connected via an electrical connector 250 at the rear end of the directional microphone. In this example, the directional tube 210 has one or more openings 230 at its front end, which serve as sound inlets. Further openings 220 are distributed along the sides of the tube, allowing sound arriving laterally to enter the tube. This laterally arriving sound also enters the tube through the openings 230, but with a phase shift due to the longer path. Inside the tube, it superimposes with the laterally arriving sound through the side openings 220.Due to interference within the tube, this sound is compensated for, resulting in a lower sensitivity to sound arriving from the side. Only for sound arriving from the front do the components entering the tube through openings 220 and 230 superimpose constructively, leading to a higher sensitivity of the microphone for sound arriving from the front ("endfire shotgun microphone"). The side openings 220 of the interference tube are not normally distributed around its circumference, but are located only on one side, which is referred to below as the top of the shotgun microphone.

[0026] Shotgun microphones offer the advantage of particularly high directivity, which applies to both a very small azimuth angle and a very small elevation angle. The elevation angle is the angle perpendicular to the plane of the image. Fig. 1 . The azimuth angle, i.e., the angle in the plane of the drawing, is also Fig. 1 The directivity of each individual shotgun microphone is very small, but by including neighboring shotgun microphones and by appropriate calculations to combine the different microphone signals, the directivity of the entire arrangement in the plane can be controlled. In particular, the directivity of a rotationally symmetric arrangement as in Fig. 1The microphones can be electronically controlled in any direction on the plane, i.e., to any desired azimuth angle. The elevation angle of the polar pattern of the entire array is the same as the elevation angle of the polar pattern of each individual shotgun microphone, and therefore very small. This eliminates the need to arrange microphones in multiple vertical planes to achieve high vertical directivity. This results in a flat array that, for example, does not obstruct the view of spectators or cameras in a sports stadium when the microphone array is positioned at the edge of the playing field. Furthermore, no calculations are required for a (potentially time-varying) combination of the microphone signals across the vertical axis.

[0027] Another advantage of a rotationally symmetric arrangement as in Fig. 1The advantage lies in the fact that the directivity, as well as the frequency response, is uniform in every direction of the plane, i.e., at any azimuth angle. This prevents any tonal coloration of sounds arriving from the side, such as audience noise, when the direction of high sensitivity of the setup is changed. Furthermore, by processing the microphone signals in multiple parallel and different ways, it is easy to define several directions simultaneously as directions of high sensitivity. This allows the beam to be focused on multiple azimuth angles at once, meaning that multiple sound sources from different directions can be recorded simultaneously with high directivity.

[0028] Various signal processing methods can be used. One possible and particularly advantageous signal processing method for the microphone array is the beamforming algorithm. Here, the beamforming is based on so-called modal beamforming, which is especially suitable for configurations in which all microphones have essentially the same directivity (directional pattern) and are arranged spherically or circularly. For the array's operating frequency range, it is possible to achieve an almost uniform directivity across all frequencies. The number Q of microphones used determines the maximum achievable degree M of the output signal, which corresponds to the spatial resolution of the directional characteristic (beam pattern), according to... M ≤ Q − 1 2 The processing takes place in two steps: (a) frequency-independent mixing (or matrixing) of the microphone signals to generate 2M+1 intermediate or mixed signals, and (b) filtering and subsequent weighting and summing of the intermediate or mixed signals.

[0029] Particularly noteworthy is the option to direct the beam (i.e., the resulting direction of high sensitivity) to a desired azimuth angle ΦT by applying the real-valued weights. g m ϕ T The control (i.e., the specification of information about the desired azimuth angle ΦT) can be achieved either manually or automatically, e.g., by a visual tracking system. Of particular importance is that the actual control of the microphone array is electronic, i.e., contactless, and that the control information varies over time. Furthermore, the filtered signals are weighted accordingly before summing, which facilitates the simultaneous recording of multiple sound sources as targets. An example is given in Fig. 7 shown and explained further below.

[0030] Fig. 3 This shows a block diagram of signal processing for the modal beamforming algorithm for an array of circularly arranged directional microphones. The Q microphone signals X(ω,x 1 ),...,X(ω,x Q ) are processed in a transformation matrix. T ( M )< ( ϕ 1 , ϕ 2 , ..., ϕ Q) 310 frequency-independently mixed. The transformation matrix applies to a desired maximum degree. M ≤ Q − 1 2 and provides (2M+1) output signals. Each output signal is filtered, with one of the (2M+1) filters 320,...,322' appearing once and all others twice as a filter pair 321,321'. For example, filter 321 for the (-M+1)th output of the matrix and filter 321' for the (M-1)th output of the matrix are identical. Each filter or filter pair has its own filter function, corresponding to the order of a specific mode. The output signal of each filter 320,...,322' is weighted in one or more units 330 according to the desired azimuth direction ΦT with a corresponding value (gain). g − M ϕ T , g − M + 1 ϕ T , … , g M ϕ T The 2M+1 weighted filtered mixed signals are summed in a summing unit 340, and the summed signal Y(ω) can then either be output as a signal 360, or optionally filtered in an equalization filter 350 and then output. This enables very flexible time-varying beamforming.

[0031] The following applies to the number of directional microphones and their positions. Generally, the number of microphones determines the spatial resolution of the possible directional characteristic, in particular the maximum directivity index, which indicates the ratio between the output power with respect to a desired target direction and the total output power integrated over all other directions. In the context of modal beamforming, it is advantageous to select the number Q of microphones depending on the required maximum degree M, according to Q = 2M + 1. When using the circular harmonic transformation described below, it is beneficial, given the assumptions made for it, to use a uniform distribution of the microphones in a circle. This ensures a uniform signal quality over all (azimuth) directions, as intended in modal beamforming. Fig. 4Figure 400 shows a microphone array in a second embodiment. In this embodiment, eleven directional microphones 4101, ..., 41011 are arranged radially and evenly distributed over a circle 420. According to the formula Q = 2M + 1, where Q = 11, a signal with a magnitude of at most M = 5 can be generated.

[0032] However, if an algorithm other than modal beam control is used, it can be advantageous to arrange the directional microphones differently, namely not exactly radially, but slightly rotated or offset. This makes the overall arrangement smaller without reducing the length of the individual directional microphones or the diameter of the microphone array.

[0033] Fig. 5 Figure 1 shows a microphone array 500 in which each of the eleven directional microphones 510 1 ,...,510 11 is rotated by an angle α and their microphone capsules are arranged on a circle 520. The algorithm used must take this rotation into account, although very small angles can be neglected.

[0034] Furthermore, for certain applications, it can be advantageous to arrange the directional microphones on a circular segment at a specific angle, for example, when only minimal background noise is expected. However, the disadvantage of a circular segment arrangement compared to a full circle arrangement is that, when positioned near the edge, background noise from directions not pointed by a directional microphone cannot be effectively suppressed. This problem can be partially mitigated by making the circular segment larger than the area being monitored.

[0035] Fig. 6Figure 1 shows a microphone array 600 in an embodiment in which eleven directional microphones 6101, ..., 61011 are evenly distributed over a semicircle 620. This arrangement is well suited for a central orientation close to 0°, corresponding to microphone 6106. An acceptable result can also be achieved for a range of, for example, ±45° around the central orientation. Accordingly, a microphone array of the following is shown in Fig. 6 The shown shape can be used, for example, at the corners of a playing field, where an area of ​​essentially 90° is to be covered.

[0036] However, for a circular segment arrangement of directional microphones, algorithms other than modal beam control are usually more suitable because they are not based on a circularly symmetrical arrangement of the microphones. However, such alternative algorithms have the disadvantage that not only their scalar weights but also their filter functions are direction-dependent. Since calculating the filter functions or filter coefficients is often relatively computationally intensive, these can be calculated in advance. The device then contains a memory in which the respective filter coefficients for specific directions are stored and from which they can be retrieved as needed. In this way, real-time operation is also possible with such alternative algorithms.

[0037] Fig. 7 This shows a block diagram of a multi-focus signal processing system for the beamforming algorithm. As already mentioned in Fig. 3The single-focus signal processing shown includes, in the multi-focus signal processing, a mixing matrix 310 for mixing the microphone signals into (2M+1) mixed signals, where M is the order of the common output signal, and a plurality of (2M+1) filters 320, 321, 321', 322, 322' for filtering the mixed signals, resulting in filtered mixed signals QF -M , QF -M+1 ,..., QF 0 ,..., QF M-1 , QF M . The filtered mixed signals are now passed not only to (2M+1) first weighting units 330 1, but also to (2M+1) second weighting units 330 2 . The first weighting units 330 1 weight each of the filtered mixed signals with a first weight g -M (Φ T1 ),...,g 0 (Φ T1 ),.. ,g M (Φ T1 ), and the second weighting units 330 2 weight each of the filtered mixed signals with a second weight g -M (Φ T2 ),...,g 0 (Φ T2 ),...,g M (Φ T2 ).The weight of each first weighting unit corresponds to the first preferred direction of high sensitivity ΦT1, and the weight of each second weighting unit corresponds to the second preferred direction of high sensitivity ΦT2. The output signals of the first weighting units 3301 and the output signals of the second weighting units 3302 are summed separately in two separate summing units 3401 and 3402, optionally filtered 3501 and 3502, and then output. Thus, the microphone array simultaneously has two preferred directions of high sensitivity ΦT1 and ΦT2. The two output signals 3601 and 3602 contain the audio signals from these two preferred directions of high sensitivity of the microphone array. For example, sounds from the direction of the ball and from the direction of the referee can be extracted and recorded simultaneously.One advantage of this arrangement is that the second weighting units 330 2 process the same filtered mixed signals as the first weighting units 330 1, using only different directional information for the preferred direction of high sensitivity Φ T2. Therefore, the filters 320,...,322' only need to be calculated and implemented once, as they are direction-independent. The weighting units can, for example, be implemented as multipliers. The entire in . Fig. 3 or in Fig. 7 The arrangement shown can be implemented using one or more microprocessors, possibly with appropriate software programs.

[0038] Details of the two-dimensional modal beam control are explained below.

[0039] First, basic assumptions and relationships are explained. In a compact area of ​​interest within three-dimensional space, containing the center of an imaginary coordinate system, free of sound sources, and excited externally by a sound field independent of the z-axis, lies an array of Q acoustic sensors (i.e., microphones) that behave linearly. These are arranged on a circle within the xy-plane of the imaginary coordinate system, with the (two-dimensional) coordinates x q = r 0 ⋅ cos ϕ q sin ϕ q , q ∈ 1 , , Q

[0040] Here, r0 is the radius of the circle and r0 is the azimuth of the q-th microphone, measured counterclockwise in the xy-plane from the x-axis. The representation X ( ω , x q The response of the q-th microphone signal in the frequency domain at an angular frequency ω can be described as a superposition (composition) of responses to individual plane waves arriving from all possible azimuth angles Φ, i.e. X ω x q = ∫ − π π H ω x q ϕ ⋅ C ω ϕ d ϕ

[0041] Here, C(ω,Φ) is the so-called plane wave amplitude density function, which is essentially a frequency domain representation of the sound pressure at the origin of the coordinate system caused by a single, plane wave incident with an azimuth angle Φ. H ( ω , x q , ϕ ) is the directional characteristic of the q-th microphone.

[0042] By series decomposition of the directional characteristic H ( ω, x q , ϕ ) and the amplitude density function of the plane wave C ( ω , ϕ) in series of real-valued orthonormal circular harmonics (a special form of spherical harmonics), defined by trg m ϕ = 1 2 π ⋅ 2 cos mϕ for m > 0 1 for m = 0 2 sin mϕ for m < 0 according to H ω x q ϕ = ∑ m = − ∞ ∞ H m ω x q trg m ϕ C ω ϕ = ∑ m = − ∞ ∞ C m ω trg m ϕ and by exploiting the orthonormality of the circular harmonics, i.e. ∫ − π π trg m ϕ trg m ′ ϕ d ϕ = δ m , m ′ where δ The Kronecker delta function can represent the microphone signal in the frequency domain. X ( ω , x q ) can be formulated as X ω x q = ∫ − π π ∑ m = − ∞ ∞ H m ω x q trg m ϕ ⋅ ∑ m ′ = − ∞ ∞ C m ′ ω trg m ′ ϕ d ϕ = ∑ m = − ∞ ∞ H m ω x q C m ω .

[0043] The individual weights H m (ω, x q ) of the circular harmonic series in (4) are called modal responses of degree m.

[0044] If all microphones have the same directional characteristics and are directed orthogonally to the circle, either outwards or inwards, this can be formally expressed as <menclose notation="box"> H ω x q ϕ = H PROTO ω , r 0 , ϕ − ϕ q < / menclose>

[0045] with H PROTO ( ω , r 0 , ϕ) as a Φ-symmetrical prototype directivity. This can be viewed as a microphone at a position ( r 0 , φ q = 0). Because of this Φ symmetry, the circular harmonic series decomposition of H PROTO ( ω , r 0 , ϕ ) given by H PROTO ω r 0 ϕ = ∑ m = − ∞ ∞ H PROTO , m ω r 0 trg m ϕ with H PROTO , m ω r 0 = 0 for m < 0 .

[0046] In this specific case, the modal responses can be factored into a frequency- and radius-dependent component and another component that depends only on the azimuth angle, according to... <menclose notation="box"> H m ω x q = b m ω r 0 trg m ϕ q < / menclose> with b m ω r 0 = π ⋅ H PROTO , m ω r 0 for m > 0 2 π ⋅ H PROTO , 0 ω r 0 for m = 0 . π ⋅ H PROTO , − m ω r 0 for m < 0 .

[0047] Also noteworthy is the symmetry of the radial components. <menclose notation="box"> b m ω r 0 = b − m ω r 0 ∀ m < / menclose> and the fact that the radial components depend on the product of the angular frequency and the radius: b m ω r 0 = b m ωr 0

[0048] By substituting (12) into (8) the frequency domain representation X ( ω, x q ) of the q-th microphone signal can be expressed as X ω x q = ∑ m = − ∞ ∞ b m ω r 0 C m ω trg m ϕ q

[0049] The following describes the basic principle of modal beamforming. This can be divided into the following two steps: (1) Reconstruction of the underlying superposition of the acting sound field of individual plane waves from the microphone signals X ( ω , x q ), represented by the coefficients C m (ω) the circular harmonic series decomposition of the amplitude density function of the plane waves, and (2) weights of the individual plane waves of the acting sound field according to a desired directional characteristic (target beam-pattern) and subsequently their integration to obtain the output signal of the beamformer.

[0050] A block diagram of a typical modal beam shaper is shown in Fig. 3 and Fig. 7shown as described above. The two steps mentioned are described in more detail below.

[0051] To motivate the reconstruction of the acting sound field, the circular harmonic series decomposition of the microphone signals in the frequency domain is used. X ω x q = ∑ m = − ∞ ∞ X m ω r 0 trg m ϕ q compared to (15). It becomes clear that the series coefficients X m ( ω, r 0) are related to the sought-after circular harmonic series coefficients C m ( ω ) the amplitude density function of the plane wave according to X m ω r 0 = b m ω r 0 C m ω .

[0052] Therefore, two further steps are carried out: (1) The circular harmonic series coefficients of the microphone signals in the frequency domain are estimated by a circular harmonic transformation according to X ^ m ω r 0 = ∑ q = 1 Q w q ⋅ X ω x q ⋅ trg m ϕ q It should be noted here that due to the finite number Q of spatial sampling points xq is the maximum absolute value of degree m that can be reconstructed, is also finite, and depends on the distribution of the spatial sampling points. x q depends on the circle. For example, the weights are all the same for the special case of a uniform distribution, namely 2 π Q , and the maximum absolute value of degree m that can be reconstructed is given by <menclose notation="box"> M = Q − 1 2 < / menclose> By defining the vector X ( ω ), which contains the signals from all microphones, as X ω = X ω x 1 X ω x 2 … X ω x Q T of the vector of all circular harmonic series coefficients as X CH ω r 0 = X ^ − M ω r 0 X ^ − M + 1 ω r 0 … X ^ M ω r 0 T and the discrete circular harmonic transformation matrix as T M ϕ 1 , ϕ 2 , … , ϕ Q = w 1 ⋅ trg − M ϕ 1 w 2 ⋅ trg − M ϕ 2 … w Q ⋅ trg − M ϕ Q w 1 ⋅ trg − M + 1 ϕ 1 w 2 ⋅ trg − M + 1 ϕ Q … w Q ⋅ trg − M + 1 ϕ Q ⋮ ⋮ ⋱ ⋮ w 1 ⋅ trg M ϕ 1 w 2 ⋅ trg M ϕ 2 w Q ⋅ trg M ϕ Q The estimation of the circular harmonic series coefficients can be expressed by the following matrix multiplication: <menclose notation="box"> X CH ω r 0 = T M ϕ 1 , ϕ 2 , … , ϕ Q ⋅ X ω < / menclose> It is particularly important that this matrix is ​​frequency-independent. (2) Taking into account (17) and (14), the circular harmonic series coefficients of the plane wave amplitude density function are estimated in principle as follows: <menclose notation="box"> C ^ m ω = f m ω r 0 ⋅ X ^ m ω r 0 < / menclose> with f m ω r 0 = 1 b m ω r 0 , which is a filtering for each individual estimated circular harmonic series coefficient of the microphone signals X̂ m ( ω, r 0 ) in the frequency domain.

[0053] Using the estimated circular harmonic series coefficients of the amplitude density function of the plane wave, the individual plane waves of the acting sound field are weighted according to a desired directional characteristic, in order to then be integrated or summed.

[0054] The maximum degree M of the circular harmonic series coefficients of the plane wave amplitude density function determines the maximum possible spatial resolution of the desired directional characteristic. Therefore, a prototype of a desired directional characteristic is defined using a truncated circular harmonic series expansion of the same degree M: g ϕ T = 0 ϕ = ∑ m = 0 M g m ϕ T = 0 ⋅ trg m ϕ the for a target azimuth angle ϕ T = 0 is set and Φ -symmetric. Because of the symmetry, the series coefficients are zero for negative degree indices m.

[0055] When the directional characteristic is set to any desired azimuth angle ΦT, its corresponding circular harmonic series coefficients can be calculated from those for ΦT = 0 according to g m ϕ T = cos mϕ T ⋅ g m ϕ T = 0 + sin mϕ T ⋅ g − m ϕ T = 0 ∀ m ∈ − M , … , M

[0056] The current beam shaper output signal Y ( ω) in the frequency domain is calculated as a weighted sum of the circular harmonic series coefficients of the amplitude density function of the plane wave as follows: Y ω = ∑ m = − M M g m ϕ T C ^ m ω

[0057] Because of the equivalence of (28) with Y ω = ∫ − π π g ϕ T ϕ C ω ϕ d ϕ The integration of the weighted contributions of the plane wave to the acting sound field becomes clear.

[0058] For most applications, the frequency-independent directional characteristic used here is advantageous and desirable. However, a frequency-dependent directional characteristic can also be created very easily by making the weighting factors frequency-dependent. This requires one filter for each coefficient of the circular harmonic series coefficients of the plane wave amplitude density function before summation.

[0059] Optionally, an equalizing filter 350,350' can be applied to the output signal. Y ( ω) of the beam shaper can be used to create direction-independent coloration or to compensate for direction-dependent coloration, e.g. to attenuate high-frequency signal components affected by spatial aliasing.

[0060] The radius of the circle on which the microphone capsules of the directional microphones are arranged influences at least two parameters of the array, namely the directivity achievable in practice at low frequencies and the frequency at which spatial aliasing begins.

[0061] The directivity at low frequencies is affected as follows. The radial components bm ( ω, r The modal responses typically have a high-pass characteristic, with the cutoff frequency increasing with the degree index m. To illustrate this, see... Fig. 8An example diagram of the radial components of modal responses for different degrees m of a Sennheiser MKH8070 shotgun microphone, plotted against a product ω · r 0. As can be seen, especially for low spectral frequencies, the contributions of the modes become very small with increasing degree m within the measured microphone signals (16). To reconstruct the corresponding circular harmonic series coefficients of the plane wave amplitude density function, a high gain factor of 1 b m ω r 0 necessary (see (26)), since | b | m | ( ω ,r 0 )| is small. This typically results in low white noise amplification for a high-degree M directional characteristic, meaning that microphone noise is strongly amplified in the beamformer's output signal. Increasing the array's radius r reduces the in Fig. 8The curves shown are essentially shifted to the left, i.e., towards lower frequencies. This leads to a reduction in the high-pass cutoff frequencies and thus reduces the effect of white noise amplification at low frequencies compared to a smaller radius.

[0062] Spatial aliasing is a phenomenon that occurs when, for example, a sound field is sampled at too few sampling points to capture high-frequency spatial oscillations of the sound pressure. Since the relevance of higher-degree circular harmonics (m) within the signature function typically increases with spectral frequency, this also applies to the magnitude of the error caused by spatial aliasing. In particular, the angular frequency at which the contribution of higher-degree circular harmonics to the signature function becomes significant can be considered the frequency at which the aliasing effect becomes disturbing or noteworthy. Essentially, this angular frequency is at ω = M ⋅ c S r 0 where c SThe speed of sound is given by the formula. This means that for a chosen number Q of microphones, the spatial aliasing frequency can be increased by reducing the radius r of the array. Alternatively, for a given array radius, the number of microphones can be increased.

[0063] For microphone arrays used in audible frequencies, the microphone capsules should be positioned on a circle or segment of a circle with a minimum radius of r min = 5 cm. For practical reasons, a maximum radius of approximately r max = 100 cm is recommended. For microphone arrays intended for use in, for example, a sports stadium, it is advantageous if the radius for outward-facing shotgun microphones is between r min = 30 cm and r max = 40 cm, and for inward-facing shotgun microphones, for example, between r min = 40 cm and r max = 60 cm. With the setup described as an example, very high directivity can be achieved, for example, for frequencies from 200 Hz to 3 kHz. For recordings in a sports stadium, frequencies below 3–4 kHz are particularly relevant.

[0064] A smaller microphone array design is possible if the circularly arranged directional microphones point radially inwards. The calculations given above still apply in this case. Fig. 9 Figure 1 schematically shows a microphone array 900 with eleven shotgun microphones in an embodiment where the individual shotgun microphones 9101, ..., 91011 are oriented essentially towards the center C of the array. The respective microphone capsules (not shown) lie on the circle 920 with radius r. With a radius of r = 50 cm and using, for example, Sennheiser MKH8070 shotgun microphones with a length of approximately 46.5 cm (where the microphone capsule is approximately 6 cm from the rear end), the diameter of the entire array is therefore only 2*(50+6) cm = 112 cm instead of 2*(50+40.5) cm = 181 cm.

[0065] Fig. 10Figure 1 shows a perspective view of a similar microphone array 1000 with fifteen directional microphones 1010 1 ,...,1010 15, which are also oriented towards the center C of the array. The microphones can be mounted, for example, on a ring or a plate. It is particularly important to ensure that the lateral openings 220 of the interference tubes of the directional microphones 1010 1 ,...,1010 15 are not obstructed, as they represent the main inlet openings for sound. Thus, the directional microphones 1010 1 ,...,1010 15 are not interfered with by the directional microphones located opposite them, i.e., in the "viewing direction". The directional microphones 1010 1 ,...,1010 15 are therefore arranged such that their upper surfaces with the side openings 220 are freely accessible to sound and preferably all point in the same direction. As in the previously described examples, the directional microphones 1010 1 ,...,1010 15 essentially in one plane, whereby the directivity of the microphone array within this plane can be electronically controlled. Note that the representation in . Fig. 10 It is not necessarily to scale. For example, microphones 1010 1 ,..., 1010 15 should be distributed as evenly as possible over circle 1020.

[0066] A particular advantage of the microphone array according to the invention is that it does not need to be moved but remains stationary, with the direction of highest sensitivity being adjustable by electronic control. In the case of a circular arrangement, this can be set to any direction within the plane of the circle (corresponding to an azimuth angle of 0°–360° with a horizontal setup). In other applications, it may be advantageous to position the circular plane vertically to capture an elevation angle of 0°–360° while keeping the azimuth angle of the detection area very small. Likewise, any intermediate orientation of the microphone plane is possible. As shown in the drawings, no microphone is located in the center of the arrangement. The specified number of directional microphones per array is the minimum required; it is always possible and can be advantageous to increase the number of microphones, as explained above.The number Q can be even or odd.

[0067] In one embodiment, the invention relates to a method for audio recording using a microphone array of directional microphones, wherein at least one common output signal is generated containing the sound in an adjustable preferred direction of high sensitivity of the microphone array, comprising the steps of: mixing several microphone signals in a mixing matrix to produce (2M+1) mixed signals, where M is the order of the common output signal, and wherein the microphone signals originate from the directional microphones and the directional microphones are arranged substantially in a plane and on a circle or circular segment such that for each of the directional microphones a preferred direction of high sensitivity is substantially orthogonal to the outside or inside of the circle or circular segment; filtering the mixed signals in a plurality of (2M+1) filters, resulting in filtered mixed signals.Weighting each of the filtered mixed signals with a weighting in a plurality of (2M+1) weighting units, wherein the weighting of each weighting unit corresponds to the adjustable preferred direction of high sensitivity of the microphone array, and summing the (2M+1) weighted, filtered mixed signals in a summing unit, thereby producing the common output signal.

[0068] The above-described examples are illustrative and can be combined with one another, even if such a combination is not explicitly mentioned. For example, in an array arrangement as in Fig. 5 The individual directional microphones are shown pointing inwards, as shown in Fig. 9 and Fig. 10 .

Claims

1. Microphone array (100) comprising a plurality of microphones (110) whose output signals are combined into at least one common output signal (360), wherein - the microphones are directional microphones, each with an interference tube and a microphone capsule and a preferred direction of high sensitivity (115) in the direction of the interference tube; - the microphones are arranged substantially in a plane; - the microphones are arranged such that each microphone has a different preferred direction of high sensitivity; - the microphones are arranged such that their microphone capsules lie on a circle (120, 920) or circular segment, and for each of the microphones, the preferred direction of high sensitivity (115) points substantially orthogonally to the circle or circular segment; - the common output signal (360) is obtained by modal beamforming (310, ..., 350); and - the microphone array (100) has at least one adjustable preferred direction of high sensitivity, wherein the common output signal (360) contains the sound picked up from this at least one adjustable direction, wherein the interference tubes (210) are arranged radially in a plane, wherein the interference tubes (210) have lateral openings (220) on an upper side of the interference tube (210) for detecting laterally incident sound, wherein the upper sides with the lateral openings (220) of the interference tubes (210) all point in the same direction.

2. Microphone array according to claim 1, wherein the beamforming (310, ..., 350) comprises a directional characteristic of the microphone array defined by a degree M, wherein a higher degree means a more focused directional characteristic, and wherein the plurality of microphones comprises at least 2M+1 microphones.

3. Microphone array according to claim 1 or 2, further comprising an electronic circuit arrangement for processing the output signals of the microphones in order to perform beamforming (310, ..., 350).

4. Microphone array according to claim 3, wherein the electronic circuit arrangement comprises at least the following elements: - a mixing matrix (310) for mixing the microphone signals into 2M+1 mixed signals, where M is the order of the common output signal; - a plurality of 2M+1 filters (320, 321, 321', 322, 322') for filtering the mixed signals, wherein filtered mixed signals (QF-M, QF-M+1, ..., QF0, ..., QFM-1, QFM) are generated; - a plurality of 2M+1 weighting units (330) for weighting each of the filtered mixed signals with a weighting (g-M(ΦT),...,g0(ΦT),..., gM(ΦT)), wherein the weighting of each weighting unit corresponds to the adjustable preferred high sensitivity direction (ΦT) of the microphone array; and - a summation unit (350) for summing the 2M+1 weighted, filtered mixed signals, producing an output signal (360) containing sound from the adjustable high sensitivity preferred direction of the microphone array.

5. Microphone array according to claim 4, wherein the microphone array comprises at least two preferred directions of high sensitivity (ΦT1,ΦT2), and wherein the circuit arrangement contains further weighting units (3302) and at least one further summation unit (3502), wherein the second weighting units (3302) process the same filtered mixed signals as the first weighting units (3301), but with different directional information for the preferred direction of high sensitivity (ΦT2) of the microphone array.

6. Microphone array according to one of claims 1-5, wherein the microphone capsules are arranged on a circle or circular segment with a radius (r) of between rmin = 5 cm and rmax = 100 cm around a center point (c).

7. Microphone array according to claim 6, wherein the radius is between 30 cm and 60 cm.

8. Microphone array according to one of claims 1-7, further comprising a control unit for setting the preferred direction of high sensitivity of the microphone array, or comprising an input for connecting such a control unit.

9. Microphone array according to any one of claims 1-8, wherein for each of the microphones, the preferred direction of high sensitivity (115) points outward relative to the circle or circle segment.

10. Microphone array according to any one of claims 1-9, wherein for each of the microphones, the preferred direction of high sensitivity (115) of each of the microphones points inward relative to the circle or circular segment.

11. Microphone array according to any one of claims 1-10, wherein each of the microphones includes a plurality of linearly arranged microphone capsules, each of which has a fixed directional characteristic.

12. Method for audio recording using a microphone array (100) with a plurality of microphones whose output signals are combined into at least one common output signal (360), wherein - the microphones are directional microphones, each with an interference tube and a microphone capsule and a preferred direction of high sensitivity (115) in direction of the interference tube; - the microphones are arranged substantially in a plane; - the microphones are arranged such that each microphone has a different preferred direction of high sensitivity; - the microphones are arranged such that their microphone capsules lie on a circle (120, 920) or circular segment, and for each of the microphones, the preferred direction of high sensitivity (115) points substantially orthogonally to the circle or circular segment; - the common output signal (360) is obtained by modal beamforming (310, ..., 350); and - the microphone array (100) has at least one adjustable preferred direction of high sensitivity, wherein the common output signal (360) contains the sound picked up from this at least one adjustable direction, wherein the interference tubes (210) are arranged radially in a plane, wherein the interference tubes (210) have lateral openings (220) on an upper side of the interference tubes (210) for detecting sound arriving from the side, wherein the upper surfaces with the lateral openings (220) of the interference tubes (210) all point in the same direction, with the steps - mixing a plurality of microphone signals in a mixing matrix (310) to 2M+1 mixed signals, where M is the order of the common output signal, and where the microphone signals come from the microphones, - filtering the mixed signals in a plurality of 2M+1 filters (320, ..., 322'), wherein filtered mixed signals (QF-M, QF-M+1, ..., QF0, ..., QFM-1, QFM) are generated; - weighting each of the filtered mixed signals with a weighting (g-M(ΦT),..., g0( ΦT) ,..., gM( ΦT)) in a plurality of 2M+1 weighting units (330), wherein the weighting of each weighting unit corresponds to the adjustable preferred high sensitivity direction (ΦT) of the microphone array; and - summing the 2M+1 weighted, filtered mixed signals in a summation unit (350), thereby producing the common output signal (360).

13. The method according to claim 12, wherein the microphone array is placed at a playing field edge of a sports stadium to capture an audio signal in the sports stadium.

14. Use of a microphone array (100) with a plurality of microphones (110) whose output signals are combined into at least one common output signal (360), wherein - the microphones are directional microphones, each having an interference tube and a microphone capsule and a preferred direction of high sensitivity (115) in direction of the interference tube; - the microphones are arranged substantially in a plane; - the microphones are arranged such that each microphone has a different preferred direction of high sensitivity; - the microphones are arranged such that their microphone capsules lie on a circle (120, 920) or circular segment, and for each of the microphones, the preferred direction of high sensitivity (115) points substantially orthogonally to the circle or circular segment; - the common output signal (360) is obtained by modal beamforming (310, ..., 350); and - the microphone array (100) has at least one adjustable preferred direction of high sensitivity, wherein the common output signal (360) contains the sound picked up from this at least one adjustable direction, wherein the interference tubes (210) are arranged radially in a plane, wherein the microphone array (100) is located at a playing field edge of a sports stadium to capture an audio signal in the sports stadium.

Images