DE102008004674A1 - Signal recording with variable directional characteristics - Google Patents

Abstract

A signal processor is used to generate a substitution signal (10) having a predetermined spatial directional characteristic using a first signal (8a) having a known spatial directional characteristic and a second signal (8b) having a known spatial directional characteristic. The first signal and the second signal are converted into a spectral representation. In the signal processor, the spectral representations of the first signal and the second signal are combined according to a combination rule to obtain amplitude parameters of a spectral representation of the substitution signal having a predetermined directivity (10). According to the combination rule, the absolute amounts of amplitude parameters of the spectral representations of the first signal and the second signal are combined so that the predetermined directivity differs from the directivity of the first (8a) and second (8b) signals.

Description

The The present invention is concerned with the recording of signals, such as audio signals, and more particularly how, out With known directional characteristics recorded signals a substitution signal can be derived, which deviates from a directional characteristic has a spatial characteristic of the space to be able to describe the filling signal.

at Multi-channel audio reproduction systems is a listener of surrounded by a plurality of speakers. The simplest multi-channel playback system is a stereo setup with two speakers. Without additional artificial influences on the sound to be reproduced a stereo system can only provide such sound sources with proper Reproduce locators that are on the line that the connects two speakers of the stereo setup. Sound or a Signal coming from other directions may not be correct provided that it is the spatial orientation As. Are multiple speakers used around the listener are arranged, several spatial directions can be correct be reproduced, resulting in a more natural spatial Sound impression for the listener may result. The best known of such multi-channel speaker systems or layouts the standardized 5.1 playback setup (ITU-R 7754-1) that out consists of five loudspeakers operating at angles of 0 °, + 30 ° and + 110 ° relative to the listening position are arranged. In addition, a number of others Playback setups with different numbers of speakers proposed or used at different positions.

Around To be able to reproduce sound spatially correctly distributed, It is first necessary to record the sound so that the spatial direction of the individual sound sources is maintained or can be reproduced. This can be done by that already after or during the recording the recorded Signals are mixed so that for each of the five Playback speaker of the ITU system an audio channel is generated assigned to the corresponding speaker during playback becomes.

In the recent past, another efficient way of recording spatial audio signals has been proposed, which makes it possible to reproduce the spatial impression prevailing in the recording with different loudspeaker systems whose relative geometric orientation need not be known a priori. This procedure is called DirAC ( Directional Audio Coding, Pulkki, V., "Directional audio coding in spatial sound reproduction and stereo upmixing", in Proceedings of The AES 28th International Conference, pp. 251-258, Pitea, Sweden, June 30-July 2, 2006 ) designated. As already mentioned, a signal recorded in this way can be played back with any speaker setup. The aim is to record the signal recorded by DirAC so that it can be reproduced as precisely as possible by means of any multi-channel speaker system, the spatial sound impression of the place where the recording was performed, should be reproduced as precisely as possible.

In The DirAC scenario is the audio signal with an omnidirectional Microphone (W) and recorded with a set of microphones that allow both an intensity vector to determine the sound field as well as the diffuse ("diffuseness") thereof.

The intensity vector indicates the direction from which the recorded sound contributes maximum energy to the recorded signal. In the form of a parameter, the diffusibility impressively describes the uniformity of the spatial sound. If the sound is perceived with identical intensity from all directions, a case of maximum diffusivity is achieved. On the other hand, there is minimal diffuseness when the sound is perceived or recorded only by a single sound source from a well-defined direction. After the DirAC analysis, the intensity vector also points in this direction. One way to record a signal is to use three microphones (X, Y, Z) with a "dumbbell" polar pattern that are oriented so that their maximum directivity is parallel to the axes of a Cartesian coordinate system (see, for example: Craven G. and Gerzon M. "Coincident microphone simulation covering three dimensional space and yielding various directional outputs". United States Patent 4042779 ). Due to the dumbbell-shaped directional characteristic, such a microphone is often referred to as a dipole (see, for example: GW Elko: "Superdirectional microphone arrays" in SG Gay, J. Benesty (eds.): "Acoustic Signal Processing for Telecommunication", Chapter 10, Kluwer Academic Press, 2000, ISBN: 978-3-540-41953-2 ; and Merimaa, J., Applications of a 3-D Microphone Array, in Proceedings of the AES 112th Convention, Munich, Germany, May 2002 ).

One way to obtain the signals W, X, Y and Z is to use a so-called "soundfield" microphone use that directly generates all of these signals. The signals W, X, Y and Z may alternatively be determined with a plurality of omnidirectional microphones located at different locations in 3-dimensional space, such as at the corners of an octahedron. From two spatially separated omnidirectional microphones, a virtual microphone with a directional characteristic, the maximum points in the direction of the connecting line between the two omnidirectional microphones, are formed by the signals of the two microphones are subtracted from each other.

It However, there are a variety of applications where it is not possible is to arrange the microphones in three dimensions in space. For example it may only be possible, the microphones in a predetermined Orientation on the surface of a plane to arrange. Furthermore, it may be necessary to reduce costs, the number of the microphones used to record a spatial signal used to reduce. This can cause that the number of microphones is lower than that actually for the inclusion of a signal with the components W, X, Y and Z required would. Should be in such a scenario, a virtual sound signal be generated, which has a predetermined directional characteristic, the is not generated by physically existing microphones must the signal by appropriate combination of existing signals Microphones are generated.

Become in this way for all relevant missing spatial directions signals with one in the corresponding Directional direction showing directionally generated, for example The DirAC algorithm continues to be successfully applied to to record a spatial audio signal. As an an example for the sound recording like a normally relative flat table microphone arrangement serve. Generalizing this can be be equated with a case in which the arrangement of microphones is limited to one level. Become omnidirectional Microphones used in such an arrangement can only Directional microphone signals in the x and y directions generated within the plane by measurement. Another microphone For example, in the center of a rectangular array of omnidirectional Signals are arranged to record an omnidirectional signal W. For example, from the recorded signals a speaker, which sits in the vertical direction in front of the display or in front of the plane is located to be able to locate or the signal to amplify the speaker appropriately is it is necessary to produce a signal that has a directivity in Z direction has its maximum signal energy thus from the z direction comes. To generate such a signal, only the signals of the remaining four or five microphones be used. The subtraction of two microphone signals, analogous to Generation of the x- or y-directed virtual microphone signals comes not considered, since all microphones used with respect their Z coordinate does not differ.

By means of known beam-forming techniques, when using a planar microphone array or a microphone matrix arranged in a plane, a signal having a directional characteristic which has a maximum in the z-direction can be generated In the case of WLAN antennas, for example, so-called "filter-and-sum-beam-formers" (see, for example: GW Elko: "Superdirectional microphone arrays" in SG Gay, J. Benesty (eds.): "Acoustic Signal Processing for Telecommunication", Chapter 10, Kluwer Academic Press, 2000, ISBN: 978-0792378143 ; and J. Bitzer, KU Simmer: "Superdirective microphone array" in M. Brandstein, D. Ward (eds.): "Microphone Arrays - Signal Processing Techniques and Applications", Chapter 2, Springer Berlin, 2001, ISBN: 978-540- 41953-2 ) or differential microphone arrays. In this case, predominantly signals of individual microphones or antennas are phase-shifted and added in a weighted manner, so that a structural interference of the individual antennas or microphones results for a phase shift assigned spatial direction, a signal recorded from this spatial direction is thereby amplified.

through However, this method does not allow a signal with a directional characteristic of a dipole in the z-direction, if as a signaling microphones or detectors a microphone array in the x and y plane is available. For Methods such as the DirAC algorithm that rely on that the input signals have dipole-like directional characteristics, So these known beam-forming methods are not applicable.

Furthermore, simple beam-forming methods in which signals of different detectors or microphones are added with a spatially directional phase shift, not suitable to produce a directional characteristic for signals whose bandwidth is so large that the phase difference for different frequencies of the male or signal to be detected is more than 360 ° only due to the different wavelength and the geometric distance of the detectors or microphones used within the bandwidth to be detected. If a method of constant phase-shifted superimposition were applied to such a signal, it would inevitably be within increasing bandwidth to reinforcements and also to extinctions. From a uniform directional characteristic for the detector can then no longer be spoken. For WLAN frequencies of, for example, 5.15 GHz to 5.725 GHz, this problem does not arise because due to the comparatively small bandwidth of the individual channels (for WLAN 802.11a, for example 0.03 GHz) with constant phase shift for all frequencies to be detected reaches a constructive interference can be. However, when recording broadband RF signals or audio signals, applying a constant phase shift with summation of the signals is no longer feasible. If audio signals from 20 Hz to 20 kHz are to be recorded, the wavelengths change from about 15 m to 1.5 cm, so that with usual geometric dimensions (for example, with a distance of adjacent microphones of a few cm) from a constant phase shift between the single microphones for all detected frequencies can no longer be talked about.

Similar considerations can be done in a two-dimensional case, as he, for example can occur in telecommunications applications. It may be sufficient there the acoustic environment only within a 2-dimensional Level to model or capture. This may be enough if only a level speaker configuration used for playback shall be. Assuming that the plane is through the X and Y axes it is sufficient, an omnidirectional signal W, the X and the Y signal record, which for example by means of a microphone array of four microphones, each at the corners a square arranged can be achieved.

however In practice it is often not even possible to use microphones in two dimensions, for example when the microphones to be placed on the top frame of a laptop display. In this case, only linear microphone arrays can be used be used, d. H. All microphones are arranged on a straight line. Without restriction of generality, this line is called Consequently, X-direction defined within an XY plane can only the omnidirectional signal W and the dipole signal X directly to be recorded. A typical setup could be three Microphones consist of two spaced apart omnidirectional Microphones can be used to control the X signal (dipole polar pattern) to measure and being an additional, in the center of the two X-microphones arranged omnidirectional microphone used is to measure the omnidirectional signal W. Even if it is sufficient should be to perform the sound analysis only in the XY plane, In such a scenario, it is necessary to have a signal with directional characteristics in Y-direction to determine or calculate. In the case of a following In addition, DirAC analysis must provide this signal the presupposed Have dipole directional characteristic. As in the previous 3-dimensional case, a signal with a directional characteristic in V-direction using conventional beam-forming techniques be generated. By means of the three discussed above However, microphones can have such a directional characteristic with dipole shape can not be achieved.

Also, a recently proposed extension of the Filter-and-Sum-Beam-Forming ( H. Kamiyanagida, H. Saruwatari, K. Takeda, F. Itakura: "Direction of arrival estimation based on nonlinear microphone array" in Proceedings of the IEEE Int Conf. On Acoustics, Speech and Signal Processing (ICASSP), pp. 3033 -3036, Salt Lake City, May 2001 ) does not lead to a directional characteristic, which has a maximum in one spatial direction for all frequency components of the signal to be detected. Rather, in the proposed extension, the direction in which a plurality of audio sources are located with respect to the pickup position is estimated. For this purpose, the amplitude square of a signal generated by means of a beam-forming microphone array is subtracted from the amplitude square of another output signal of a beam-forming microphone array. By evaluating the signal thus generated, the directivity of the microphone array can be permanently adjusted so that the direction of minimum sensitivity of the array corresponds to the direction of the active sound sources.

By this type of adjustment of the directivity (beam-forming) will not about a predetermined directivity with one in a large Room area does not vanish, but the directivity is merely changed or tracked that the direction of minimum sensitivity of the microphone array corresponds to the direction of the active sound sources. The extension of the beam-forming algorithm therefore allows the Estimate the direction of a plurality of active sound sources, whereas no statement about the origin of the maximum Signal energy of the audio signal, that is the intensity of the signal or the diffusivity of the signal can be derived.

It is therefore required signals of predetermined directivity to be able to generate as substitution signals.

This object is achieved by a signal processor according to claim 1, a method according to Pa tentanspruch 19 and a computer program according to claim 20 solved.

at In some embodiments of the invention, a signal is generated predetermined spatial directional characteristic of a first signal of known spatial directional characteristic and a second signal of known spatial directional characteristic generated by first the first and the second signal be converted from a temporal into a spectral representation. The spectral representations of the first signal and the second Signals are processed according to a combination rule, that of the known directional characteristics of the first and second Signal depends, so combined that a spectral representation a signal having the predetermined spatial directional characteristic is obtained, which differ both from the directional characteristic of the first signal as well as the second signal. In some embodiments, the directional characteristic is identical for all spectral ranges of the generated signal.

at In some embodiments, in particular, the amplitude amounts the spectral representations of the input signals formed before these are combined to give amplitude values for the substitution signal generated to create.

The Spectral representation of the signal obtained by the combination can thus have a directional characteristic, which differs from the Directional characteristic of the recorded first and second signals differs, in particular a directional characteristic whose maximum Amplification factor points in a spatial direction, which is from the spatial direction of the maximum sensitivity of the signals known directivity differs. By the combination the signals in their spectral representation can be achieved be that the directional characteristic for the entire resulting Sig nal in an identical space area is the same. This means in particular, that is different for different frequency components the signal to be detected or generated is the maximum the sensitivity or the maximum amplification factor the directional characteristic is in the same spatial direction. This is even the case when the bandwidth of the recorded Signals or the signal to be generated is so large that a superposition of the signals in the temporal representation with a constant phase shift would cause that is the maximum directivity or the maximum the directional characteristic spatially within the spectral range or the bandwidth of the signal to be generated significantly would change. By combining the spectral Representation can be achieved, however, that a directional characteristic is generated for all signal components a broadband signal in the same space area or is similar or identical in the same spatial direction.

at In some embodiments of the invention, the spectral Representation of the generated signal directly from a signal processor issued or available for further processing posed. In further embodiments of the invention the spectral representation of the generated signal is returned to the temporal representation implemented to a temporal representation of the signal having the predetermined directivity.

at Some embodiments of the invention will be audio signals or audio recording of pieces of music or Environmental sounds or speakers processed by means of won a 2-dimensional or 1-dimensional microphone array become. Consequently, also information about the location or the position of the sound sources in a respect the arrangement of the microphones orthogonal direction gained. at In other embodiments, high-frequency signals become larger Bandwidth or any other signals processed that a signal is generated whose maximum contribution to the signal amplitude comes from a predetermined spatial direction, that is in other Words a predetermined spatial directional characteristic having. In this sense, a signal with directional characteristics as a signal with directionally weighted signal components to understand so that there are one or more spatial directions, from which signal components with maximum gain or amplitude recorded or reconstructed, whereas others Spaces exist where the signal components are damped or completely suppressed.

In some embodiments of the invention, a short-term frequency transformation is used for conversion into the spectral representation, in which the signals to be transformed or converted are processed in blocks. For this purpose, for example, a filter bank or a frequency transformation can be used, in which each signal part of predetermined length, which may for example consist of a sequence of a predetermined number of signal samples, is assigned a plurality of amplitude and phase values, as described, for example, in the short-term Fourier Transformation (SFT) is the case. A continuous signal in time representation is transformed into a sequence of amplitude and phase factors or represented as a sequence of these factors, each signal portion, ie each independently processed time interval is assigned a plurality of amplitude values P (k, n) (the index k indicates the analyzed frequency band). In the combination of the spectral representations of the transformed signals, only the magnitudes of the amplitudes are combined to obtain the signal having a predetermined directivity. This ensures that the resulting directional characteristic can be identical for all frequencies of the frequency band to be distinguished. This may be the case even if the locations where the signals of known directional characteristics were recorded are not known exactly. The maximum of the sensitivity can thus be obtained in a simple and mathematically uncomplicated manner for all frequencies in the same spatial direction. In some embodiments of the present invention, the signal of predetermined spatial directional characteristics thus obtained may be immediately further used in its spectral representation to estimate from which spatial direction in the observed frequency band the maximum sound or signal energy is coming.

at In some embodiments, the spectral representation the generated signal having the predetermined directivity used by means of a DirAC analyzer or a DirAC algorithm the DirAC parameterization in two dimensions or three dimensions to receive directly. In doing so substitutes or replaces the generated one Signal a signal that is inaccessible to direct measurement is. In this sense, this is generated by the signal processor Signal also to be called a substitution signal. Hereinafter Therefore, the terms "substitution signal" and signal used synonymously with a predetermined directional characteristic. Using of embodiments of signal processors with DirAc analyzers Thus, the great advantage that a complete DirAC analysis and thus a parametrization of the spatial Sound impression is possible without a complex recording of three independent signals, each in pairs vertical directional characteristic to perform have to.

at Presenting the information, in what intensity the signal can be recorded from the three spatial directions, about it a statement is made as to how diffuse the recorded Signal is. The intensity vector gives the energy flux density at. In a three-dimensional space has the intensity vector three orthogonal components (eg x, y, z), which together form the direction of the energy flow.

are in all three spatial directions the components of the intensity vector equal or almost zero, it can be assumed that the Signal or the sound or sound the measuring space evenly fills, since in the frequency interval studied out all spatial directions small or zero Components of the intensity vector are present.

at Further embodiments of the invention, the spectral Representation of the generated signal predetermined directional characteristic converted into a temporal representation, so that approximately a signal is received, as seen from a virtual microphone predetermined directivity would have been recorded. In this case, the phase factors of the implementation in the frequency domain from any of (input) signals of predetermined directivity used to distinguish between the individual frequency ranges To obtain as realistic as possible a phase relationship. This can cause that, though in the implementation the directional dependence takes into account only the amplitudes were generated, a signal whose audible artifacts due to the partially ignored phase information are barely perceptible.

preferred Embodiments of the present invention will be with reference to the attached figures, explained in detail. Show it:

1 An embodiment of a signal processor;

2 an embodiment of a signal combiner of the signal processor;

3 an example of an arrangement of microphones for recording signals for signal processors;

4 an example of a directional characteristic of recorded signals;

5 an example of a predetermined directivity of a generated signal;

6A and 6B Embodiments of a device for generating a signal of predetermined directivity;

7 Exemplary embodiment for deriving a DirAC parameterization; and

8th An embodiment of a method for generating a signal of predetermined directivity.

1 shows an embodiment of a signal processor 2 , which is a signal converter 4 and a signal combiner 6 includes. The in 1 shown signal processor 2 is used to generate a signal having a predetermined spatial directional characteristic (a substitution signal) using a first signal 8a known spatial directional characteristic and a second signal 8b known spatial directional characteristic. The signals 8a (P ₁ ) and 8b (P ₂ ) can for example be recorded by a microphone or received by an antenna and are in a temporal representation. As in 1 indicated, further embodiments of the invention may use more than two signals as input signals, in particular, the number of signals used as input signals is in principle not limited to the top. The signal converter 4 serves to get the first signal 8a and the second signal 8b in each case in a spectral representation of the signals P _i (k, n) to implement, where i = 1, 2.

The signal combiner 6 combines the spectral representations of the first signal 8a and the second signal 8b according to a combination rule to the signal 10 to produce with a predetermined spatial directional characteristic. In this case, the signal generated 10 have a directional characteristic, which differs from the directional characteristic of the signals 8a and 8b different. In some embodiments of the invention, the combination rule that the signal combiner pertains to 6 used to the signal 10 to produce with a predetermined directional characteristic exclusively of the known directivity of the first and the second signal 8a and 8b from.

In the embodiments which use more than two signals of known directivity as input signals, it is also necessary that the directivity for each input signal is known to produce a signal whose directivity corresponds to the predetermined or desired directivity. The in 1 shown signal processor 2 thus generates the signal 10 with a predetermined spatial directional characteristic in a spectral representation of the same.

in the The following briefly becomes the general case of generating the signal predetermined directional characteristic of any number L> 2 of signals known Directional described before the invention of the underlying Concept based on two specific embodiments is shown again in more detail.

there become the embodiments of the invention discussed on the basis of audio signals, although the concept too any other signals, such as high frequency signals or radio signals is applicable. Starting point for the Processing of the signals by embodiments of Signal processors according to the invention are known L> 2 signals Directivity. These signals can either be direct Measured with the known directional characteristic (omnidirectional or directed) microphone signals or be signals, taken from a directional characteristic output of a microphone array become. How the directional characteristic was generated is immaterial, as long as this is known.

wobei j die imaginäre Einheit bezeichnet. Von dem Signalkombinierer des Signalprozessors werden die Beträge P₁(k, n) dieser Eingangssignale unter Verwendung einer Kombinationsregel derart kombiniert, dass ein durch die Kombination resultierender Betrag |D(k, n)| einer Spektraldarstellung des zu erzeugenden Signals zugeordnet ist. Dies korrespondiert zu einem Mikrofonsignal, das die vorbestimmte räumliche Richtcharakteristik aufweist. Dabei ist es insbesondere möglich, die Richtcharakteristik innerhalb eines breiten Raumbereichs und eines großen Frequenzbereichs konstant oder ähnlich zu halten. Allgemein kann die Kombination der Beträge durch folgende Formel beschrieben werden: D(k, n) = g(|P1(k, n)|, |P2(k, n)|, ..., |PL(k, n)|), wobei D(k, n) das erzeugte Signal (das Substitutionssignal) beschreibt. Die Funktion g(.) beschreibt die Kombinationsregel gemäß derer die Beträge der Eingangssignale kombiniert werden, welche prinzipiell aus beliebigen linearen und nichtlinearen Funktionen gebildet beziehungsweise zusammengesetzt sein kann.Each of the input signals is divided into a series of discrete time sections or signal sections (signal blocks). The signal blocks are converted into a spectral representation, for example in the short-term frequency domain. In the following, P _i (k, n) denotes the first input signal in the frequency band. The index n denotes the number or number of the signal block currently under consideration from the series of signal blocks into which the input signal has been split. After the frequency conversion, the signal can be represented in each frequency band of interest as an amplitude or magnitude P _i (k, n) and as a phase Φ (k, n), so that the following applies:

where j denotes the imaginary unit. From the signal combiner of the signal processor, the amounts P ₁ (k, n) of these input signals are combined using a combination rule such that an amount | D (k, n) | a spectral representation of the signal to be generated is assigned. This corresponds to a microphone signal having the predetermined spatial directional characteristic. In particular, it is possible to keep the directivity constant or similar within a wide spatial range and a large frequency range. Generally, the combination amounts are described by the following formula: D (k, n) = g (| P 1 (k, n) |, | P 2 (k, n) |, ..., | P L (k, n) |), where D (k, n) describes the generated signal (the substitution signal). The function g (.) Describes the combination rule according to which the amounts of the input signals are combined, which in principle can be formed or composed of arbitrary linear and nonlinear functions.

In some embodiments of the invention, one of the phases of the input signals is used as the phase information of the generated signal D (k, n). For example, if the first signal from the signal combiner is selected as the signal whose phase is used, the combination rule can be defined as follows:

A signal combiner 6 which implements the above-described concept is schematically illustrated in FIG 2 shown. The input signals are thereby 8a . 8b and 8c already in their spectral representation. As described above, the phase information may be from the first input signal 8a can be extracted, wherein in further embodiments, the phase information of any other input signals can be used. Further embodiments may completely dispense with the phase information. The phase information 12 becomes. from the first input signal 8a extracted, taking from an amount 14a the amount 14b of the first input signal 8a is formed. In an equivalent way are used by the contributors 16a respectively 18a the amount values 16b and 18b the input signals 8b and 8c educated. In a combination block 20 will be the sums 14b . 16b and 18b the input signals 8a . 8b and 8c using a combination rule (function g) combined to the magnitude value 22 of the generated signal. An optional multiplier 24 serves to adjust the signal D (k, n) by multiplying the magnitude value 22 and the phase factor 12 to form, as described in the previous sections.

By applying a suitable combination law g which depends on the directional characteristic of the input signals, it is possible to generate a signal having a predetermined spatial directional characteristic, which in particular has a directional characteristic which differs from that of the input signals. From the in 2 sketched signal combiner 6 Thus, a signal with a predetermined spatial directional characteristic is generated in a spectral representation. This signal can be used immediately in conjunction with the input signals 8a to 8c Derive parameters that describe the basic characteristics of the received signal or audio signal in the recording environment. These can be, for example, the DirAC parameters, ie the direction of the instantaneous intensity per frequency range and the diffusivity of the signal in each of the frequency ranges considered. In further embodiments, the generated signal in the spectral representation of predetermined directivity can also be transformed back into a temporal representation.

there In particular, it is possible to have a predetermined directional characteristic in a large space and frequency range, without assumptions about the statistics of the input signals used to have to do. Such assumptions about the signal statistics, such as stationary behavior, the different spectral properties or the spatial Coherence between multiple signals typically becomes taken in non-linear filter techniques as they reduce background noise used in speech enhancement systems (also known as spectral subtraction or Wiener filtering).

embodiments By contrast, signal combiners require only the Knowledge of the directional characteristic of the recording of the input signals used microphones or microphone arrays and meet a priori, no assumptions about the statistics of the input signals or via the statistics of their spectral compositions. This makes it efficient, based on simple algorithms Way possible, with input signals of known directional characteristics a signal with a predetermined spatial directional characteristic to create.

3 schematically shows a possible microphone set up, by means of which the signals W can be received with omnidirectional directional characteristic, X with dipole-shaped directional characteristic in the X direction and Y with dipole-shaped directional characteristic in the Y direction. 3 shows a scenario in which five microphones 30a - 30e are mounted in one plane. It is assumed that it is due to geometric edge conditions is impossible, other microphones outside of the 3 to install in the plan shown in the supervision. By means of such an arrangement can be achieved by combinations of the microphones 30a - 30d recorded signals generate signals having a directional characteristic in the X direction and in the Y direction. The central microphone 30e with omnidirectional directivity can be used, for example, to record an omnidirectional signal W and to provide it as an input signal with omnidirectional directional characteristic.

3 is only one of any number of possible examples that make it possible to record signals with directional characteristic in the X direction, in the Y direction and without specific directional characteristic, ie omnidirectional signals W.

The description of the following embodiment, in which a signal with a predetermined spatial directional characteristic is generated, on the basis of which, for example, a DirAC parametrization of the spatial audio signal can be performed, assumes that the input signal is an omnidirectional signal W, a signal having a dipole-like directional characteristic X in the X direction and a signal Y with directional characteristic in the Y direction are available, as for example by means of in 3 shown 2-dimensional microphone arrays can be obtained.

As already mentioned, it is assumed that the direction maximum sensitivity of the signal X (k, n) the X direction and the signal Y (k, n) the Y direction of a Cartesian coordinate system is. Furthermore, the signals W, X and Y are already in a spectral Present representation, that is, for each frequency range and time block or signal part of the signals exists Amplitude parameter and a phase parameter, as in the preceding Sections has been described. To the sound field with sufficient accuracy 3-dimensional description is required that a signal with a directional characteristic is generated, whose Maximum has a component in the Z direction. Becomes a DirAC parameterization in particular, it is necessary to generate a signal which has a dipole-like directional characteristic with a maximum or with maximum sensitivity in the Z direction.

4 illustrates the directional characteristic of a signal which is formed according to the following combination rule from the amplitudes X (k, n) and Y (k, n) of the input signals X and Y:

4 shows a 3-dimensional representation of the directivity of the signal formed according to the above combination or combination rule. Shown is the gain factor (the directional weighting factor), which contains the signal from the relevant spatial direction in the combination signal against the position of the source in the X direction 40 and in the Y direction 42 , An amplification factor of 1 means that the signal is recorded unattenuated, that is to say with the amplitude or intensity not reduced by the combination of the two individual signals X and Y. As simple geometric considerations show, the gain factor along the Z-axis is constantly zero, since in this direction neither the X signal nor the Y signal or the microphones assigned to this signal are sensitive. From the above-mentioned combination of the X-signals and the Y-signals thus results in a directional effect, the in 4 shown torus whose axis of rotation is the Z-axis. If it is further considered that the directional characteristic of an omnidirectional signal W has no maximum, ie has a spherical shape, it becomes evident that a signal of dipole-shaped directional characteristic in the Z-direction can be obtained if the signals W (k, n), X (k , n) and Y (k, n) are combined according to the following combination rule:

By The above relationship thus becomes a signal of predetermined spatial Directional characteristic (Z-directional dipole) provided the directional characteristics of the input signals X, Y and W are known.

Out The combination rule thus results in the amount of generated Directly in the relevant frequency range.

In one embodiment of the invention, the phase φ _w (k, n) of the phase of the omnidirectional signal W (k, n) can be chosen accordingly, so that the signal generated by a phase information has the form:

5 illustrates in a one-dimensional representation the amplification factor as a function of the angle α ( 52 ) between the Z-axis and the direction of the signal source. The solid line 54 describes the directional characteristic of a conventional dipole. The dotted curve 56 illustrates the directional characteristic of the signal Z (k, n) which was combined by means of the above combination rule from the signals W, X and Y or derived from these signals.

As it is in 5 can be seen, the phase information is no longer complete due to the combination, which is characterized by the fact that the directivity of the so-called homophasic dipole 56 is equivalent to that of the "classical" dipole only for the angles 0 to 90 ° and 270 to 360 ° The absolute magnitude of the directional gain factors in the range between 90 ° and 270 ° is identical to the absolute value of the classical dipole, but no negative ones occur If, for example, in the case of the DirAC parameterization, only information is required from which spatial direction the maximum intensity is recorded, this loss of information is acceptable.Therefore, if by means of the above combination rule the signal Z with the directivity just described has been generated subsequent and signal-based DirAC analysis will yield a correct determination of the intensity vector within the half-space of the Cartesian Z-axis coordinate system.

The 6A and 6B illustrate again schematically embodiments of an apparatus for generating a signal of predetermined directivity or a system for generating a DirAC parameterization based on a first signal X with directional characteristic in the X direction, a second signal Y with directional characteristic in the Y direction and an omnidirectional signal W. , 6A schematically shows an embodiment of a signal processor 2 to which a first signal X, a second signal Y and a third signal W are supplied. The directional characteristics of the signals correspond to the embodiment described above, so that by means of the signal processor 2 a signal Z can be generated with a predetermined spatial directional characteristic, in which the maximum direction-dependent weighting factor occurs in the Z direction. As described with reference to the preceding embodiments, the signals X, Y and W initially recorded in a temporal representation by microphones are converted into a spectral representation, whereupon a signal Z having the predetermined directional characteristic is formed by the above-described combination of the magnitude values.

As it is in 6B Alternatively, the spectral representation of the X, Y and W signals together with the spectral representation of the particular Z signal may be a DirAC analyzer 60 supplied, as described above, the characteristic of the DirAC parameterization of an acoustic spatial signal parameters, namely a direction vector 62 and a diffusivity parameter (diffuseness parameter) 64 generated. Together with the omnidirectional signal W in its temporal representation or in its spectral representation, these parameters make it possible to prefer any spatial directions in a subsequent reproduction or to reproduce signals only from the arbitrary spatial directions, or to reproduce the sound field true to the original.

Based on 7 In the following, it will be illustrated how a DirAC parameterization can be performed by means of a substitution signal of predetermined directivity generated by an embodiment of a signal processor, or how the DirAC parameters (a direction vector and a diffuse value or diffuse value ψ) are formed can be. 7 shows in a Cartesian coordinate system with the axes X, Y and Z at the corners of an octahedron and in the center seven omnidirectional microphones, by means of which the complete information required for the DirAC parameterization or analysis, ie signals with directivity in X, Y and Z and an omnidirectional signal W can be obtained. At the in 7 shown, only to be understood as an example arrangement, the signals can be obtained with directional characteristic of the difference between two spaced apart in the relevant spatial direction omnidirectional microphones. At the in 7 Thus, a signal with the directional characteristic in the X direction (signal X (k, n)) can thus be obtained by subtracting the optionally frequency-dependent normalized signals of the microphones 70a and 70b be won. Furthermore, a signal with directional characteristic in the Y direction (signal Y (k, n)) can be obtained by subtracting the possibly frequency-dependent normalized signals of the microphones 72a and 72b be won. In addition, a signal with directional characteristic in the Z direction (signal Z (k, n)) by subtracting possibly suitable frequency-dependent normalized signals of the microphones 74a and 74b be won. The center-mounted omnidirectional microphone 76 (W) is used to record a signal W (k, n) with omnidirectional characteristics, as required for a DirAC analysis. The signal with omnidirectional characteristic can alternatively also be calculated as mean value of all microphone signals or as mean value of the external microphone signals.

in the For the sake of simplicity, it is assumed below that a Transformation into the spectral range for the concerned Signals already occurred, so that indices k each considered the Specify the spectral range, whereas the index n the considered Signal part or signal block of the blocks used here by way of example Frequency analysis (eg short-time Fourier transform SET) describes.

For the DirAC analysis, an intensity vector is first calculated whose X-component results from the following relationship: I x (k, n) = Re {E {W * (k, n) X (k, n)}}, where the operator * designates the generation of the complex conjugate.

It goes without saying that the equivalent relationships apply to the Y and Z components of the intensity vector, although the above formula describes only the X component. Would, as in the 7 Case shown, all information, so the signal X, Y, Z and W are available, could be the coordinates of in 7 exemplified direction of maximum volume 78 determine as follows:

In the above equations, the angle φ denotes the azimuth 79a and the angle θ the elevation 79b So those angles that indicate the direction of the sound source 78 describe clearly relative to the origin of the coordinate system. The diffuseness ψ is determined in the DirAC analysis as follows: E (k, n) = 1 2 (E {| X (k, n) | 2 } + E {| Y (k, n) | 2 } + E {| Z (k, n) | 2 } + E {| W (k, n) | 2 }),

If, as discussed in detail in the preceding embodiment, a microphone signal with directional characteristic in the Z direction is not available, the signal Z ~ (k, n) can be generated by means of an embodiment of a signal processor as described above. This signal has the basis of the 5 illustrated directional characteristic. Therefore, the signal thus generated in the frequency domain can be used directly for the derivation of the DirAC parameters. According to the above considerations, the intensity vector for the DirAC parameterization in the Z direction consequently results in: I ~ z (k, n) = Re {E {W * (k, n) Z ~ (k, n)}}

Consequently, even in the case where there is no generically picked up signal in the Z direction, an elevation angle can be used 79b Derive the position of the sound source 78 in Z-direction indicates. This is possible by using the intensity I ~ _z (k, n) to calculate the angle according to the following formula:

As already mentioned above, by the fact that the phase information in which the combination rule is not taken into account, the directivity of the generated signal may correspond to that of a homophasic dipole. Therefore, a unique association is the position of the sound source 78 only possible in the half-plane with positive Z-values. The diffuseness parameter results in the case of a 2-dimensional microphone array which generates X, Y and W signals, equivalent to the case of full DirAC microphone signals, by replacing the value I _z with the above-derived intensity I ~ _z becomes.

A DirAC analysis is also possible if one is only interested in the 2-dimensional analysis of the sound field. In this case, for example, the omnidirectional microphone 76 can be omitted, as will be shown below, from the X and Y signals, a substitution signal can be generated with torus-shaped directional characteristic, which allows the interesting 2-dimensional DirAC parameters, namely azimuth 79a and to determine the diffuseness parameter ψ _2D .

First, as a substitution signal of predetermined directivity, a signal W _T (k, n) is formed, which has a toroidal directivity, as already described above. The signal W _T (k, n) can be represented as follows from the signals X and Y:

The phase parameter is optional. The 2-dimensional intensity can be formed as follows: I 2D, x (k, n) = Re {E {W T * (k, n) X (k, n)}} where the Y-direction "2-dimensional intensity" is equivalent to the above formula for the Y signal.

Durch das erzeugte Signal vorbestimmter Richtcharakteristik (Signal W_T mit Torus-förmiger Richtcharakteristik) lässt sich der „2-dimensionale" Gesamtenergiewert gemäß folgender Formel errechnen: E2D(k, n) = 12 (E{|X(k, n)|2} + E{|Y(k, n)|2} + E{(WT(k, n)|2}). From the X and Y intensities, the azimuth angle 79a be determined directly according to the following formula:

By the generated signal of predetermined directional characteristic (signal W _T with toroidal directional characteristic), the "2-dimensional" total energy value can be calculated according to the following formula: e 2D (k, n) = 1 2 (E {| X (k, n) | 2 } + E {| Y (k, n) | 2 } + E {(W T (k, n) | 2 }).

Using the substitution signal W _T , which was generated by means of an embodiment of a signal processor, the diffuseness parameter ψ can thus also be calculated in the 2-dimensional case according to the following formula:

As above embodiments have given, it is by the use of an embodiment of an inventive Signal processor possible to save microphones or geometric Boundary conditions that are not an ideal placement of microphones allow to compensate and still complete Perform DirAC analysis of a spatial signal.

In the further embodiment of the present invention discussed below, it is assumed that only one linear microphone array is available as signal suppliers, ie as signal sources for the input signals of the signal processor. This has the immediate result that a signal with a directional characteristic only in the direction that is defined by the connecting line between the microphones, is available. Without limiting generality, this direction is referred to below as the X direction. Consequently, only one omnidirectional signal W (k, n) and one signal X (k, n) can be provided as an input to the signal processor. Nevertheless, an acoustic analysis can at least take place within the XY plane, for example with the DirAC parameterization, since, as will be described below, a signal Y (k, n) with a predetermined, dipole-like directional characteristic can be generated from the signals W and X. Analogous to the above considerations, the following combination rule can be modified to determine the values of the signal to be generated according to the following relationship:

The directional characteristic of the signal thus generated corresponds to the above considerations following a torus whose axis of rotation is the X-axis. This means especially for a 2-dimensional view tion within the XY plane that the directional characteristic of the signal Y ^ within the XY plane corresponds to a dipole whose maximum weighting factor lies on the Y-axis and whose directional characteristic is symmetrical to the Y-axis. The directional characteristic thus generated fulfills the conditions for a subsequent DirAC analysis.

An optional phase factor may correspond to the phase factor of the omnidirectional signal W (k, n), so that if phase information is desired, the combination rule may be extended to form the generated signal such that it can be described according to the following formula:

The thus generated signal Y ^ (k, n) has a directional characteristic accordingly a homophasic dipole within the X-Y plane. Thus, will a subsequent DirAC analysis on a correct determination of the Intensity vector within the half-plane of the X-Y plane lead, which includes the positive Y-axis. Consequently, can also in the case of a 1-dimensional microphone array a DirAC analysis be carried out, which is a highly flexible processing of the analyzed signal, if embodiments used according to the invention signal processors can be used to provide an additional signal to generate predetermined spatial directional characteristic.

The following consideration, as generated by means of the above Signal Y ^ (k, n) also in the case of a linear microphone array a subsequent DirAC analysis (2-dimensional case) can. The intensity vector for the DirAC analysis is calculated for the X component as in the normal case, because this component together with an omnidirectional signal immediately available.

For the determination of the Y-component, the generated substitution signal of predetermined directivity Y ^ (k, n) is used, so that the following applies: I ~ 2D, y (k, n) = Re {E {W * (k, n) Y ~ (k, n)}}

To determine the azimuth angle, the intensity I ~ _{2D, Y} thus generated and the measured intensity I _{2D, x} can then be used. In the case of a linear microphone array, too, a diffuseness parameter ψ _2D can be derived from the following relationships, even if only microphones can be used which are arranged linearly one behind the other in one spatial direction: E ~ 2D (k, n) = 1 2 (E {| X (k, n) | 2 } + E {| Y ~ (k, n) | 2 } + E {| W (k, n) | 2 }),

As described above, so can also make a statement about the spatial composition of the signal are taken if only with a 1-dimensional receiver array can be recorded.

• – Bereitstellen einer Mehrzahl von Signalen mit bekannter Richtcharakteristik (entweder direkt mit dieser Richtcharakteristik gemessen oder durch Beam-Forming mittels eines Mikrofon-Arrays errechnet).
• – Transformieren jedes Eingangssignals in die Kurzzeitfrequenzdomäne.
• – Für jeden zu berücksichtigenden Frequenzbereich: Bestimmen des Betrags und einer Phaseninformation jedes Eingangssignals.
• – Für jeden interessierenden Frequenzbereich: Bestimmen des Betrags des zu erzeugenden Signals, indem die Beträge der Eingangssignale gemäß einer Kombinationsregel miteinander kombiniert werden.
• – Sofern eine Phaseninformation gewünscht wird, Kombinieren des Betrags des erzeugten Signals mit einem der Phasenfaktoren der spektralen Darstellungen einer der Eingangssignale.

Some embodiments of the present invention thus carry out the following steps:

• - Providing a plurality of signals having a known directional characteristic (either measured directly with this directional characteristic or calculated by beam-forming using a microphone array).
• - Transform each input signal into the short-time frequency domain.
• For each frequency range to be considered, determine the magnitude and phase information of each input signal.
• For each frequency range of interest: determining the magnitude of the signal to be generated by combining the magnitudes of the input signals according to a combination rule.
• If phase information is desired, combining the magnitude of the generated signal with one of the phase factors of the spectral representations of one of the input signals.

As already mentioned above, allows the application Signal processors according to the invention the more flexible Application of microphones or microphone arrays with regard to on the determination of parameters for spatial reconstruction a recorded soundscape or a recorded signal.

It is allowed to receive signals of a predetermined spatial Directivity to produce (such as a dipole characteristic).

About that In addition, it is possible using inventive Signal processors the spatial selectivity of Improve microphone arrays by, for example, the spatial Range by the directional weighting factor is big, can be limited. This can be, for example be achieved by the output signals of the microphone arrays known Directional characteristic by means of embodiments according to the invention Signal processors are processed.

8th shows an embodiment of a method for generating a signal with a predetermined spatial directional characteristic. In a signal generation step 100 becomes a first signal of known spatial directional characteristic 8a and a second signal of known spatial directional characteristic 8b provided. Furthermore, the temporal representation of the first and the second signal in a step of the transformation into a spectral representation of the first signal 8a and a spectral representation of the second signal 8b implemented. In a combination step, the spectral representation of the first signal 8a and the spectral representation of the second signal 8b combined according to a combination rule such that a spectral representation of the signal to be generated resulting from the combination has the predetermined spatial directivity, wherein the predetermined directivity differs from the directivity of the first and the second signal.

In an optional inverse transformation step 104 the spectral representation of the generated signal is converted into a time representation in order to obtain a signal with a predetermined spatial directional characteristic, which can be reproduced for example by means of a loudspeaker.

In an alternative analysis step 106 becomes from the generated signal predetermined spatial directional characteristic 10 and the input signals 8a and 8b derived a parametrization or their spectral representation of the spatial properties of the recorded audio signal.

Dependent from the circumstances, the inventive Method for generating a signal with predetermined spatial Directional characteristic can be implemented in hardware or in software. The implementation may be on a digital storage medium, in particular a floppy disk or CD with electronically readable control signals done so interact with a programmable computer system can that process of the invention for generating a signal having a predetermined spatial Directional characteristic is executed. Generally exists the invention thus also in a computer program product with a stored on a machine-readable carrier Program code for carrying out the inventive Procedure if the computer program product on a machine expires. In other words, the Invention thus as a computer program with a program code to implement the method, if the computer program runs on a computer.

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• US 4042779 [0006]

Cited non-patent literature

• - Directional Audio Coding, Pulkki, V., "Directional audio coding in spatial sound reproduction and stereo upmixing", in Proceedings of The AES 28th International Conference, pp. 251-258, Pitea, Sweden, June 30-July 2, 2006 [ 0004]
• - GW Elko: "Superdirectional microphone arrays" in SG Gay, J. Benesty (eds.): "Acoustic Signal Processing for Telecommunication", Chapter 10, Kluwer Academic Press, 2000, ISBN: 978-3-540-41953-2 [ 0006]
• Merimaa, J., "Applications of a 3-D Microphone Array," in Proceedings of the AES 112th Convention, Munich, Germany, May 2002 [0006]
• GW Elko: "Superdirectional microphone arrays" in SG Gay, J. Benesty (eds.): "Acoustic Signal Processing for Telecommunication", Chapter 10, Kluwer Academic Press, 2000, ISBN: 978-0792378143 [0010]
• - J. Bitzer, KU Simmer: "Superdirective microphone array" in M. Brandstein, D. Ward (eds.): "Microphone Arrays - Signal Processing Techniques and Applications", Chapter 2, Springer Berlin, 2001, ISBN: 978-540 -41953-2 [0010]
• H. Kamiyanagida, H. Saruwatari, K. Takeda, F. Itakura: "Direction of arrival estimation based on nonlinear microphone array" in Proceedings of the IEEE Int Conf. On Acoustics, Speech and Signal Processing (ICASSP), pp. 3033-3036, Salt Lake City, May 2001 [0015]

Claims (25)

Signal processor ( 2 ) for generating a substitution signal ( 10 ) having a predetermined directivity using a first signal ( 8a ) having a known directional characteristic and a second signal ( 8b ) having a known directional characteristic, comprising: a signal converter ( 4 ), the first signal ( 8a ) and the second signal ( 8b ) to implement in a spectral representation with a plurality of frequency bands, for each of which at least one amplitude parameter is determined; and a signal combiner ( 6 ) to the spectral representation of the first signal ( 8a ) and the second signal ( 8b ) according to a combination rule to obtain a spectral representation of the substitution signal ( 10 ) with the predetermined, from the directional characteristics of the first ( 8a ) and second ( 8b ) Signals of differing directivity, wherein, according to the combination rule, the absolute values of the amplitude parameters of the first signal ( 8a ) and the second signal ( 8b ) and combined to produce an amplitude parameter of a spectral representation of the substitution signal ( 10 ) to obtain. Signal processor ( 2 ) according to claim 1, further comprising a second signal converter for generating the spectral representation of the substitution signal ( 10 ) into a temporal representation. Signal processor according to one of Claims 1 or 2, in which signal combiners ( 6 ) is formed for different frequency bands of the spectral Darstellun conditions of the first ( 8a ) and the second (b) signal to use the same combination rule. Signal processor ( 2 ) according to one of the preceding claims, in which the signal combiner ( 6 ) is designed to use a combination rule which only depends on the known directional characteristic of the first ( 8a ) and the second ( 8b ) Signal depends. Signal processor ( 2 ) according to one of the preceding claims, in which the signal converter ( 4 ) has a block processor to the first ( 8a ) and the second signal ( 8b ) into discrete signal parts of constant length, the signal converter ( 4 ) is adapted to convert each of the signal parts into a spectral representation with a plurality of frequency bands. Signal processor ( 2 ) according to one of the preceding claims, in which the signal combiner is designed ( 4 ), to use such a combination rule that at a first signal ( 8a ) with omnidirectional directivity and a second signal ( 8b ) is generated with dipole-shaped directional characteristic in a dipole direction, a substitution signal having a dipole-shaped directional characteristic in a predetermined direction, which is perpendicular to the dipole direction. Signal processor ( 2 ) according to one of the preceding claims, in which the signal combiner ( 6 ) is adapted to use a combination rule according to which the magnitude of an amplitude parameter D (k, n) of the substitution signal is determined by the following combination of the amplitude parameters P (k, n) of the first signal ( 8a ) and P (k, n) of the second signal ( 8b ) can be described:

Signal processor ( 2 ) according to one of the preceding claims, in which the signal converter ( 4 ) is designed to additionally determine a phase parameter for each frequency band as part of the spectral representation of the signals. Signal processor according to Claim 8, in which the signal combiner ( 6 ) is adapted to use a combination rule according to which one of the spectral representation of the substitution signal ( 10 ) associated phase factor φ using a phase factor φ ₁ (k, n) of the first ( 8a ) and / or a phase factor φ ₂ (k, n) of the second ( 8b ) Signal is obtained. Signal processor according to Claim 9, in which the signal combiner ( 6 ) is formed, as a phase factor φ either the phase factor φ ₁ (k, n) of the first ( 8a ) or the phase factor φ ₂ (k, n) of the second ( 8b ) To use signals. Signal processor according to one of the preceding claims, which is designed to receive the substitution signal ( 10 ) using a third signal ( 8c ) predetermined directional characteristic, wherein the signal converter ( 4 ), the third signal ( 8c ) in the spectral representation with several Fre implement bands; and the signal combiner ( 6 ) is adapted to provide a spectral representation of the substitution signal ( 10 ) with the predetermined, from the directional characteristics of the first ( 8a ), second ( 8b ) and third ( 8c ) Signals of differing directivity, wherein, according to the combination rule, in addition the absolute value of an amplitude parameter of the third signal ( 8c ) is combined to obtain the amplitude parameter of the spectral representation of the substitution signal ( 10 ) to obtain. Signal processor ( 2 ) according to claim 11, wherein the signal combiner is formed ( 4 ), to use such a combination rule that at a first signal ( 8a ) with omnidirectional directivity, a second signal ( 8b ) with a dipole-shaped directional characteristic in a dipole direction and a third signal ( 8c ) is generated with a dipole-shaped directional characteristic in a direction perpendicular to the dipole direction of the second dipole a substitution signal having a dipole directional characteristic in a predetermined direction, which is perpendicular to the dipole direction and the second dipole direction. Signal processor ( 2 ) according to claim 11 or 12, wherein the signal combiner ( 6 ) is adapted to use a combination rule according to which the magnitude of an amplitude parameter D (k, n) of the substitution signal is determined by the following combination of the amplitude parameters P ₁ (k, n) of the first signal ( 8a ), P (k, n) of the second signal ( 8b ) and P ₃ (k, n) of the third signal ( 8c ) can be described:

Signal processor ( 2 ) according to one of claims 11 to 13, in which the signal converter ( 4 ) is designed to additionally determine a phase parameter for each frequency band as part of the spectral representation of the signals. Signal processor according to claim 14, wherein the signal combiner ( 6 ) is adapted to use a combination rule according to which one of the spectral representation of the substitution signal ( 10 ) associated phase factor φ using a phase factor φ ₁ (k, n) of the first ( 8a ) and / or a phase factor φ ₂ (k, n) of the second ( 8b ) and / or a phase factor φ ₃ (k, n) of the third ( 8c ) Signal is obtained. Signal processor ( 2 ) according to claim 15, wherein the signal combiner ( 6 ) is formed, as a phase factor φ either the phase factor φ ₁ (k, n) of the first ( 8a ), the phase factor φ ₂ (k, n) of the second ( 8b ) or the phase factor φ ₃ (k, n) of the third ( 8c ) To use signals. Signal processor ( 2 ) according to claim 9 or 15, wherein the signal combiner ( 6 ) is adapted to use a combination rule according to which the substitution signal ( 10 ) from the magnitude of the amplitude parameter | D (k, n) | and the phase factor φ is combined as follows: D (k, n) = | D (k, n) | e jφ , Signal processor according to one of preceding claims, which is adapted to audio signals to use and generate an audio signal. A method for generating a signal having a predetermined spatial directional characteristic using a first signal having a known spatial directional characteristic and a second signal having a known spatial directional characteristic, comprising: converting the first signal ( 8a ) and the second signal ( 8b ) in a spectral representation with a plurality of frequency bands, for each of which at least one amplitude parameter is determined; and combine the spectral representation of the first signal ( 8a ) and the second signal ( 8b ) to obtain a spectral representation of the substitution signal ( 10 ) with the predetermined, from the directional characteristics of the first ( 8a ) and second ( 8b ) Signals of differing directivity, wherein the absolute values of the amplitude parameters of the first signal ( 8a ) and the second signal ( 8b ) and combined to produce an amplitude parameter of a spectral representation of the substitution signal ( 10 ) to obtain. Computer program with a program code for execution The method of claim 19 when the program is on a computer expires. Signal analysis system for generating a parameterization which describes a spatial characteristic of a signal, having the following features: a signal processor according to any one of claims 1 to 18; and a spatial signal analyzer to use the spectral representations of the first signal ( 8a ), the second signal ( 8b ) and the substitution signal ( 10 ) to generate the parameters of the parameterization Signal analysis system according to claim 21, in which the spatial signal analyzer is designed to a DirAC parameterization to create. Signal analysis system according to one of the claims 21 or 22, in which the signal combiner ( 6 ) is formed from a first signal ( 8a ) with a dipole-shaped directional characteristic in an X-direction and a second signal ( 8b ) with a dipole-shaped directional characteristic in a Y direction orthogonal thereto, a substitution signal ( 10 ) with a toroidal directional characteristic; and the sound field analyzer is configured to determine a directional parameter indicative of an incident direction of the sound field in the XY plane and a diffusivity parameter indicative of a diffusibility of the signal energy in the XY plane. Signal analysis system according to one of the claims 21 or 22, in which the signal combiner ( 6 ) is formed from a first signal ( 8a ) with omnidirectional directivity and a second signal ( 8b ) with a dipole-shaped directional characteristic in an X-direction a substitution signal ( 10 ) with a toroidal directional characteristic; and the sound field analyzer is configured to determine a direction parameter indicating an incident direction of the sound field in an XY plane and a diffusivity parameter indicating a diffusibility of the signal energy in the XY plane. Signal analysis system according to one of the claims 21 or 22, in which the signal combiner ( 6 ) is formed from a first signal ( 8a ) with omnidirectional directivity, a second signal ( 8b ) with a dipole-shaped directional characteristic in an X-direction and a third signal ( 8c ) having a dipole-shaped directional characteristic in a Y direction orthogonal to the X direction, a substitution signal ( 10 ) with a dipole-shaped directional characteristic in a Z-direction perpendicular to the X-direction and Y-direction; and the sound field analyzer is configured to determine a first and a second direction parameter indicating an incident direction of the sound field in the 3-dimensional space and a diffusivity parameter indicating a diffusibility of the signal energy in the 3-dimensional space.