(19)
(11)EP 2 930 954 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
22.07.2020 Bulletin 2020/30

(21)Application number: 14163702.5

(22)Date of filing:  07.04.2014
(51)International Patent Classification (IPC): 
H04S 7/00(2006.01)

(54)

Adaptive filtering

Adaptive Filterung

Filtrage adaptatif


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(43)Date of publication of application:
14.10.2015 Bulletin 2015/42

(73)Proprietor: Harman Becker Automotive Systems GmbH
76307 Karlsbad (DE)

(72)Inventor:
  • Christoph, Markus
    94315 Straubing (DE)

(74)Representative: Westphal, Mussgnug & Partner Patentanwälte mbB 
Werinherstrasse 79
81541 München
81541 München (DE)


(56)References cited: : 
EP-A1- 1 843 635
US-A1- 2007 019 826
US-B1- 6 760 451
EP-A1- 1 986 466
US-A1- 2010 305 725
  
  • NELSON P A ET AL: "ADAPTIVE INVERSE FILTERS FOR STEREOPHONIC SOUND REPRODUCTION", IEEE TRANSACTIONS ON SIGNAL PROCESSING, IEEE SERVICE CENTER, NEW YORK, NY, US, vol. 40, no. 7, 1 July 1992 (1992-07-01), pages 1621-1632, XP000307653, ISSN: 1053-587X, DOI: 10.1109/78.143434
  
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description

TECHNICAL FIELD



[0001] The disclosure relates to an adaptive filter system and method.

BACKGROUND



[0002] Spatial sound field reproduction techniques utilize a multiplicity of loudspeakers to create a virtual auditory scene over a large listening area. Several sound field reproduction techniques, e.g., wave field synthesis (WFS) or Ambisonics, make use of a loudspeaker array equipped with a plurality of loudspeakers to provide a highly detailed spatial reproduction of an acoustic scene. In particular, wave field synthesis is used to achieve a highly detailed spatial reproduction of an acoustic scene to overcome limitations by using an array of, e.g., several tens to hundreds of loudspeakers.

[0003] EP 1986466 A1 discloses a method for automated tuning of a sound system. The sound system includes delay lines, equalizing filters, and at least two loudspeakers. The method includes reproducing a useful sound signal through the loudspeakers, measuring sound pressure values at at least one location, and providing a target transfer function for tuning the delay lines and the equalizing filters of the sound system, wherein the target transfer function represents a desired transfer characteristics of the sound system. The method further includes adjusting the delay of the delay lines, and adjusting amplitude responses of the equalizing filters such, that the actual transfer characteristics of the sound system approximates the target function.

[0004] US 2010/0305725 A1 discloses a method for designing an audio precompensation controller for a multichannel audio system, with a prescribed number N of loudspeakers in prescribed positions so that listeners positioned in any of P>1 spatially extended listening regions should be given the illusion of being in another acoustic environment that has L sound sources located at prescribed positions in a prescribed room acoustics. The method provides a unified joint solution to the problems of equalizer design, crossover design, delay and level calibration, sum-response optimization and up-mixing. A multi-input multi-output audio precompensation controller is designed for an associated sound generating system including a limited number of loudspeaker inputs for emulating a number of virtual sound sources. The method includes estimating, for each loudspeaker input signals, an impulse response at each of a set of measurement positions that cover the P listening regions, specifying a target impulse response (target stages) for each virtual sound source at each measurement position, and determining adjustable filter parameters of the audio precompensation controller so that a criterion function is optimized.

[0005] US 6760451 B1 discloses a method of processing an audio signal. The method includes executing at least one spectral transform operation to derive a signal spectrum, processing the spectrum to effect smoothing thereof, said smoothing being variable with frequency across the signal spectrum, and conditioning the response of a filter for use in a subsequent audio reproduction system with the smoothed spectrum.

[0006] US 2007/0019826 A1 discloses a frequency equalization system that substantially equalizes the room frequency responses generated by at least one loudspeaker within a listening area so that the frequency responses in the listening area are substantially constant.

[0007] The publication "Adaptive Inverse Filters for Stereophonic Sound Reproduction", Nelson et al., IEEE Transactions on Signal Processing, Vol. 40, No. 7, July 1992, discloses another system for equalizing the response of a multichannel sound reproduction system using least square filters and modeling delays.

[0008] Spatial sound field reproduction techniques overcome some of the limitations of stereophonic reproduction techniques. However, technical constraints prohibit the employment of a high number of loudspeakers for sound reproduction. Wave field synthesis (WFS) and Ambisonics are two similar types of sound field reproduction. Though they are based on different representations of the sound field (the Kirchhoff-Helmholtz integral for WFS and the spherical harmonic expansion for Ambisonics), their aim is congruent and their properties are alike. Analysis of the existing artifacts of both principles for a circular setup of a loudspeaker array came to the conclusion that HOA (Higher-Order Ambisonics), or more exactly near-field-corrected HOA, and WFS meet similar limitations. Both WFS and HOA and their unavoidable imperfections cause some differences in terms of the process and quality of the perception. In HOA, with a decreasing order of the reproduction, the impaired reconstruction of the sound field will probably result in a blur of the localization focus and a certain reduction in the size of the listening area.

[0009] For audio reproduction techniques such as wave field synthesis (WFS) or Ambisonics, the loudspeaker signals are typically determined according to an underlying theory, so that the superposition of sound fields emitted by the loudspeakers at their known positions describes a certain desired sound field. Typically, the loudspeaker signals are determined assuming free-field conditions. Therefore, the listening room should not exhibit significant wall reflections, because the reflected portions of the reflected wave field would distort the reproduced wave field. In many scenarios such as the interior of a car, the necessary acoustic treatment to achieve such room properties may be too expensive or impractical.

SUMMARY



[0010] An acoustic multi-input multi-output system with M microphones and K loudspeakers includes K ≥ 1 equalizing filter modules that are arranged in signal paths downstream of an input signal path, the input signal path being supplied with an input signal, wherein the equalizing filter modules have controllable transfer functions, and K equalizing filter control modules that are configured to control the transfer functions of the K equalizing filter modules according to an adaptive control algorithm, based on M>1 error signals and the input signal on the input signal path. The adaptive control algorithm includes a windowed magnitude constraint with an integrated post-ringing constraint, the magnitude constraint and post-ringing constraint being psychoacoustic constraints that are configured to model psychoacoustic properties of a human ear. The filter control modules are configured to iteratively adapt the transfer functions of each of the K equalizing filter modules. The windowed magnitude constraint with the integrated post-ringing constraint is configured to input a set of cosine signals with equidistant frequencies and equal amplitudes into the K equalizing filter modules upon adaption, weight a signal output by the filter modules with a frequency-dependent windowing function, sum up the filtered and windowed cosine signals to provide a sum signal, and scale the sum signal to provide an updated impulse response of the K equalizing filter modules for controlling the transfer function of the K equalizing filter modules.

[0011] A method for equalizing an acoustic multiple-input multiple-output system with M microphones and K loudspeakers includes equalizing filtering with K ≥ 1 controllable transfer functions in a signal path downstream of an input signal path, the input signal path being supplied with an input signal, and controlling with K filter control signals of the controllable transfer function for filtering according to an adaptive control algorithm, based on M>1 error signals and the input signal on the input signal path. The adaptive control algorithm includes a windowed magnitude constraint with an integrated post-ringing constraint, the magnitude constraint and post-ringing constraint being psychoacoustic constraints that are configured to model psychoacoustic properties of a human ear. The controlling comprises iteratively adapting the transfer function of the equalizing filtering. The windowed magnitude constraint with the integrated post-ringing constraint comprises inputting a set of cosine signals with equidistant frequencies and equal amplitudes into the K equalizing filter modules upon adaption, weighting signals output by the equalizing filtering with a frequency-dependent windowing function, summing up the filtered and windowed cosine signals to provide a sum signal, and scaling the sum signal to provide an updated impulse response of the equalizing filtering for controlling the K transfer functions of the equalizing filtering.

BRIEF DESCRIPTION OF THE DRAWINGS



[0012] The system and methods may be better understood with reference to the following drawings and description. The components in the figures are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.
Moreover, in the figures, like referenced numerals designate corresponding parts throughout the different views.

Figure 1 is a flow chart illustrating a simple acoustic Multiple-Input Multiple-Output (MIMO) system with M recording channels (microphones) and K output channels (loudspeakers), including a multiple error least mean square (MELMS) system or method.

Figure 2 is a flowchart illustrating a 1 × 2 × 2 MELMS system or method applicable in the MIMO system shown in Figure 1.

Figure 3 is a diagram illustrating a pre-ringing constraint curve in the form of a limiting group delay function (group delay differences over frequency).

Figure 4 is a diagram illustrating the curve of a limiting phase function (phase difference curve over frequency) derived from the curve shown in Figure 3.

Figure 5 is an amplitude time diagram illustrating the impulse response of an all-pass filter designed according to the curve shown in Figure 4.

Figure 6 is a Bode diagram illustrating the magnitude and phase behavior of the all-pass filter shown in Figure 5.

Figure 7 is a block diagram illustrating a setup for generating individual sound zones in a vehicle.

Figure 8 is a magnitude frequency diagram illustrating the magnitude frequency responses at each of the four zones (positions) in the setup shown in Figure 7 using a MIMO system solely based on more distant loudspeakers.

Figure 9 is an amplitude time diagram (time in samples) illustrating the corresponding impulse responses of the equalizer filters of the MIMO system that forms the basis of the diagram shown in Figure 8.

Figure 10 is a schematic diagram of a headrest with integrated close-distance loudspeakers applicable in the setup shown in Figure 7.

Figure 11 is a schematic diagram of an alternative arrangement of close-distance loudspeakers in the setup shown in Figure 7.

Figure 12 is a schematic diagram illustrating the alternative arrangement shown in Figure 11 in more detail.

Figure 13 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7 when a modeling delay of half the filter length and only close-distance loudspeakers are used.

Figure 14 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 13.

Figure 15 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7 when a length-reduced modeling delay and only close-distance loudspeakers are used.

Figure 16 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 15.

Figure 17 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7 when a length-reduced modeling delay and only system, i.e., far-distance, loudspeakers are used.

Figure 18 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 17.

Figure 19 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7 when an all-pass filter implementing the pre-ringing constraint instead of a modeling delay and only close-distance loudspeakers are used.

Figure 20 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results to the frequency characteristics at the four desired positions shown in Figure 19.

Figure 21 is an amplitude frequency diagram illustrating the upper and lower thresholds of an exemplary magnitude constraint in the logarithmic domain.

Figure 22 is a flow chart of a MELMS system or method with a magnitude constraint that is based on the system and method described above in connection with Figure 2.

Figure 23 is a Bode diagram (magnitude frequency responses, phase frequency responses) of the system or method using a magnitude constraint, as shown in Figure 22.

Figure 24 is a Bode diagram (magnitude frequency responses, phase frequency responses) of a system or method using no magnitude constraint.

Figure 25 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7 when only the eight more distant loudspeakers in combination with a magnitude and pre-ringing constraint are used.

Figure 26 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 25.

Figure 27 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7 when only more distant loudspeakers in combination with a pre-ringing constraint and a magnitude constraint based on windowing with a Gauss window are used.

Figure 28 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 27.

Figure 29 is an amplitude time diagram illustrating an exemplary Gauss window.

Figure 30 is a flow chart of a MELMS system or method with a windowing magnitude constraint that is based on the system and method described above in connection with Figure 2.

Figure 31 is a Bode diagram (magnitude frequency responses, phase frequency responses) of a system or method when only more distant loudspeakers in combination with a pre-ringing constraint and a magnitude constraint based on windowing with the modified Gauss window are used.

Figure 32 is an amplitude time diagram illustrating an exemplary modified Gauss window.

Figure 33 is a flow chart of a MELMS system or method with a spatial constraint that is based on the system and method described above in connection with Figure 22.

Figure 34 is a flow chart of a MELMS system or method with an alternative spatial constraint that is based on the system and method described above in connection with Figure 22.

Figure 35 is a flow chart of a MELMS system or method with a frequency-dependent gain constraint LMS, which is based on the system and method described above in connection with Figure 34.

Figure 36 is a magnitude frequency diagram illustrating the frequency-dependent gain constraints corresponding to four more distant loudspeakers when using crossover filters.

Figure 37 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7 when only more distant loudspeakers in combination with a pre-ringing constraint, a windowed magnitude constraint and an adaptive frequency (dependent gain) constraint are used.

Figure 38 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 37.

Figure 39 is a Bode diagram of a system or method when only more distant loudspeakers in combination with a pre-ringing constraint, a windowed magnitude constraint and an adaptive frequency (dependent gain) constraint are used.

Figure 40 is a flow chart of a MELMS system or method that is based on the system and method described above in connection with Figure 34, with an alternative frequency (dependent gain) constraint.

Figure 41 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7, with applied equalizing filters when only more distant loudspeakers in combination with a pre-ringing constraint, a windowed magnitude constraint and the alternative frequency (dependent gain) constraint in the room impulse responses are used.

Figure 42 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 41.

Figure 43 is a Bode diagram of the equalizing filters applied to the setup shown in Figure 7 when only more distant loudspeakers in combination with a pre-ringing constraint, a windowed magnitude constraint and the alternative frequency (dependent gain) constraints in the room impulse responses are used.

Figure 44 is a schematic diagram illustrating the sound pressure levels over time for pre-masking, simultaneous masking and post-masking.

Figure 45 is a diagram illustrating a post-ringing constraint curve in the form of a limiting group delay function as group delay differences over frequency.

Figure 46 is a diagram illustrating the curve of a limiting phase function as phase difference curve over frequency derived from the curve shown in Figure 45.

Figure 47 is a level time diagram illustrating the curve of an exemplary temporal limiting function.

Figure 48 is a flow chart of a MELMS system or method that is based on the system and method described above in connection with Figure 40, with a combined magnitude post-ringing constraint.

Figure 49 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7, with applied equalizing filters when only more distant loudspeakers in combination with a pre-ringing constraint, a magnitude constraint-based non-linear smoothing, a frequency (dependent gain) constraint and a post-ringing constraint are used.

Figure 50 is an amplitude time diagram illustrating the impulse responses corresponding to the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 49.

Figure 51 is a Bode diagram of the equalizing filters applied to the setup shown in Figure 7 when only more distant loudspeakers in combination with a pre-ringing constraint, a magnitude constraint-based non-linear smoothing, a frequency (dependent gain) constraint and a post-ringing constraint are used.

Figure 52 is a magnitude time diagram illustrating the curve of an exemplary level limiting function.

Figure 53 is an amplitude time diagram corresponding to the magnitude time curve shown in Figure 52.

Figure 54 is a magnitude time diagram illustrating the curve of exemplary window functions with exponential windows at three different frequencies.

Figure 55 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7, with applied equalizing filters when only more distant loudspeakers in combination with a pre-ringing constraint, a magnitude constraint, a frequency (dependent gain) constraint and a windowed post-ringing constraint are used.

Figure 56 is an amplitude time diagram illustrating the impulse responses of the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 55.

Figure 57 is a Bode diagram of the equalizing filters applied to the setup shown in Figure 7, with applied equalizing filters when only more distant loudspeakers in combination with a pre-ringing constraint, a magnitude constraint, a frequency (dependent gain) constraint and a windowed post-ringing constraint are used.

Figure 58 is a magnitude frequency diagram illustrating an exemplary target function for the tonality of a bright zone.

Figure 59 is an amplitude time diagram illustrating the impulse responses in the linear domain of an exemplary equalizing filter with and without applied windowing.

Figure 60 is a magnitude time diagram illustrating the impulse responses in the logarithmic domain of an exemplary equalizing filter with and without applied windowing.

Figure 61 is a magnitude frequency diagram illustrating the frequency characteristics at the four positions in the setup shown in Figure 7, with applied equalizing filters when all loudspeakers in combination with a pre-ringing constraint, a magnitude constraint, a frequency (dependent gain) constraint and a windowed post-ringing constraint are used and the response at the bright zone is adjusted to the target function depicted in figure 58.

Figure 62 is an amplitude time diagram illustrating the impulse responses of the equalization filter of the MIMO system, which results in the frequency characteristics at the four desired positions shown in Figure 61.

Figure 63 is a flow chart of a system and method for reproducing wave fields or virtual sources using a modified MELMS algorithm.

Figure 64 is a flow chart of a system and method for reproducing virtual sources corresponding to a 5.1 loudspeaker setup using a modified MELMS algorithm.

Figure 65 is a flow chart of an equalizing filter module arrangement for reproducing virtual sources corresponding to a 5.1 loudspeaker setup at the driver position of a vehicle.

Figure 66 is a flow chart of a system and method that uses a modified MELMS algorithm to generate virtual sound sources corresponding to a 5.1 loudspeaker setup at all four positions of a vehicle.

Figure 67 is a diagram illustrating spherical harmonics up to fourth order.

Figure 68 is a flow chart of a system and method for generating spherical harmonics in a target room at a distinct position using a modified MELMS algorithm.

Figure 69 is a schematic diagram illustrating a two-dimensional measuring microphone array disposed on a headband.

Figure 70 is a schematic diagram illustrating a three-dimensional measuring microphone array disposed on a rigid sphere.

Figure 71 is a schematic diagram illustrating a three-dimensional measuring microphone array disposed on two ear cups.

Figure 72 is a process chart illustrating an exemplary process for providing a magnitude constraint with integrated post-ringing constraint..


DETAILED DESCRIPTION



[0013] Figure 1 is a signal flow chart of a system and method for equalizing a multiple-input multiple-output (MIMO) system, which may have a multiplicity of outputs (e.g., output channels for supplying output signals to K ≥ 1 groups of loudspeakers) and a multiplicity of (error) inputs (e.g., recording channels for receiving input signals from M ≥ 1 groups of microphones). A group includes one or more loudspeakers or microphones that are connected to a single channel, i.e., one output channel or one recording channel. It is assumed that the corresponding room or loudspeaker-room-microphone system (a room in which at least one loudspeaker and at least one microphone is arranged) is linear and time-invariant and can be described by, e.g., its room acoustic impulse responses. Furthermore, Q original input signals such as a mono input signal x(n) may be fed into (original signal) inputs of the MIMO system. The MIMO system may use a multiple error least mean square (MELMS) algorithm for equalization, but may employ any other adaptive control algorithm such as a (modified) least mean square (LMS), recursive least square (RLS), etc. Input signal x(n) is filtered by M primary paths 101, which are represented by primary path filter matrix P(z) on its way from one loudspeaker to M microphones at different positions, and provides M desired signals d(n) at the end of primary paths 101, i.e., at the M microphones.

[0014] By way of the MELMS algorithm, which may be implemented in a MELMS processing module 106, a filter matrix W(z), which is implemented by an equalizing filter module 103, is controlled to change the original input signal x(n) such that the resulting K output signals, which are supplied to K loudspeakers and which are filtered by a filter module 104 with a secondary path filter matrix S(z), match the desired signals d(n). Accordingly, the MELMS algorithm evaluates the input signal x(n) filtered with a secondary pass filter matrix S(z), which is implemented in a filter module 102 and outputs K × M filtered input signals, and M error signals e(n). The error signals e(n) are provided by a subtractor module 105, which subtracts M microphone signals y'(n) from the M desired signals d(n). The M recording channels with M microphone signals y'(n) are the K output channels with K loudspeaker signals y(n) filtered with the secondary path filter matrix S(z), which is implemented in filter module 104, representing the acoustical scene. Modules and paths are understood to be at least one of hardware, software and/or acoustical paths.

[0015] The MELMS algorithm is an iterative algorithm to obtain the optimum least mean square (LMS) solution. The adaptive approach of the MELMS algorithm allows for in situ design of filters and also enables a convenient method to readjust the filters whenever a change occurs in the electro-acoustic transfer functions. The MELMS algorithm employs the steepest descent approach to search for the minimum of the performance index. This is achieved by successively updating filters' coefficients by an amount proportional to the negative of gradient (n), according to which w(n + 1) = w(n) + µ(-(n)), where µ is the step size that controls the convergence speed and the final misadjustment. An approximation may be in such LMS algorithms to update the vector w using the instantaneous value of the gradient (n) instead of its expected value, leading to the LMS algorithm.

[0016] Figure 2 is a signal flow chart of an exemplary Q × K × M MELMS system or method, wherein Q is 1, K is 2 and M is 2 and which is adjusted to create a bright zone at microphone 215 and a dark zone at microphone 216; i.e., it is adjusted for individual sound zone purposes. A "bright zone" represents an area where a sound field is generated in contrast to an almost silent "dark zone". Input signal x(n) is supplied to four filter modules 201-204, which form a 2 x 2 secondary path filter matrix with transfer functions Ŝ11(z), Ŝ12(z), Ŝ21(z) and Ŝ22(z), and to two filter modules 205 and 206, which form a filter matrix with transfer functions W1(z) and W2(z). Filter modules 205 and 206 are controlled by least mean square (LMS) modules 207 and 208, whereby module 207 receives signals from modules 201 and 202 and error signals e1(n) and e2(n), and module 208 receives signals from modules 203 and 204 and error signals e1(n) and e2(n). Modules 205 and 206 provide signals y1(n) and y2(n) for loudspeakers 209 and 210. Signal y1(n) is radiated by loudspeaker 209 via secondary paths 211 and 212 to microphones 215 and 216, respectively. Signal y2(n) is radiated by loudspeaker 210 via secondary paths 213 and 214 to microphones 215 and 216, respectively. Microphone 215 generates error signals e1(n) and e2(n) from received signals y1(n), y2(n) and desired signal d1(n). Modules 201-204 with transfer functions Ŝ11(z), Ŝ12(z), Ŝ21(z) and Ŝ22(z) model the various secondary paths 211-214, which have transfer functions S11(z), S12(z), S21(z) and S22(z).

[0017] Furthermore, a pre-ringing constraint module 217 may supply to microphone 215 an electrical or acoustic desired signal d1(n), which is generated from input signal x(n) and is added to the summed signals picked up at the end of the secondary paths 211 and 213 by microphone 215, eventually resulting in the creation of a bright zone there, whereas such a desired signal is missing in the case of the generation of error signal e2(n), hence resulting in the creation of a dark zone at microphone 216. In contrast to a modeling delay, whose phase delay is linear over frequency, the pre-ringing constraint is based on a non-linear phase over frequency in order to model a psychoacoustic property of the human ear known as pre-masking. An exemplary graph depicting the inverse exponential function of the group delay difference over frequency is and the corresponding inverse exponential function of the phase difference over frequency as a pre-masking threshold is shown in Figure 4. "Pre-masking" threshold is understood herein as a constraint to avoid pre-ringing in equalizing filters.

[0018] As can be seen from Figure 3, which shows a constraint in the form of a limiting group delay function (group delay differences over frequency), the pre-masking threshold decreases when the frequency increases. While at a frequency of approximately 100 Hz, a pre-ringing represented by a group delay difference of about 20 ms is acceptable for a listener, at a frequency of approximately 1,500 Hz, the threshold is around 1.5 ms and may reach higher frequencies with an asymptotic end-value of approximately 1 ms. The curve shown in Figure 3 can be easily transformed into a limiting phase function, which is shown in Figure 4 as phase difference curve over frequency. By integrating the limiting phase difference function, a corresponding phase frequency characteristic can be derived. This phase frequency characteristic may then form the basis for the design of an all-pass filter with a phase frequency characteristic that is the integral of the curve shown in Figure 4. The impulse response of an accordingly designed all-pass filter is depicted in Figure 5, and its corresponding Bode diagram is depicted in Figure 6.

[0019] Referring now to Figure 7, a setup for generating individual sound zones in a vehicle 705 using the MELMS algorithm may include four sound zones 701-704 corresponding to listening positions (e.g., the seat positions in the vehicle) arranged front left FLPos, front right FRPos, rear left RLPos and rear right RRPos. In the setup, eight system loudspeakers are arranged more distant from sound zones 701-704. For example, two loudspeakers, a tweeter/midrange loudspeaker FLSpkrH and a woofer FLSpkrL, are arranged closest to front left position FLPos and, correspondingly, a tweeter/midrange loudspeaker FRSpkrH and a woofer FRSpkrL are arranged closest to front right position FRPos. Furthermore, broadband loudspeakers SLSpkr and SRSpkr may be arranged next to sound zones corresponding to positions RLPos and RRPos, respectively. Subwoofers RLSpkr and RRSpkr may be disposed on the rear shelf of the vehicle interior, which, due to the nature of the low-frequency sound generated by subwoofers RLSpkr and RRSpkr, impact all four listening positions front left FLPos, front right FRPos, rear left RLPos and rear right RRPos. Additionally, vehicle 705 may be equipped with yet other loudspeakers, arranged close to sound zones 701-704, e.g., in the headrests of the vehicle. The additional loudspeakers are loudspeakers FLLSpkr and FLRSpkr for zone 701; loudspeakers FRLSpkr and FRRSpkr for zone 702; loudspeakers RLLSpkr and RLRSpkr for zone 703; and loudspeakers RRLSpkr and RRRSpkr for zone 704. All loudspeakers in the setup shown in Figure 7 form respective groups (groups with one loudspeaker) except loudspeaker SLSpkr, which forms a group of passively coupled bass and tweeter speakers, and loudspeaker SRSpkr, which forms a group of passively coupled bass and tweeter speakers (groups with two loudspeakers). Alternatively or additionally, woofer FLSpkrL may form a group together with tweeter/midrange loudspeaker FLSpkrH and woofer FRSpkrL may form a group together with tweeter/midrange loudspeaker FRSpkrH (groups with two loudspeakers).

[0020] Figure 8 is a diagram illustrating the magnitude frequency responses at each of the four zones 701-704 (positions) in the setup shown in Figure 7 using equalizer filters, a psychoacoustically motivated pre-ringing constraint module and the system loudspeakers, i.e., FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr. Figure 9 is an amplitude time diagram (time in samples) illustrating the corresponding impulse responses of the equalizer filters for generating a desired crosstalk cancellation in the respective loudspeaker paths. In contrast to the simple use of a modeling delay, the use of a psychoacoustically motivated pre-ringing constraint provides sufficient attenuation of the pre-ringing. In acoustics, pre-ringing designates the appearance of noise before the actual sound impulse occurs. As can be seen from Figure 9, the filter coefficients of the equalizing filters, and thus the impulse responses of the equalizing filters, exhibit only little pre-ringing. It can additionally be seen from Figure 8 that the resulting magnitude frequency responses at all desired sound zones tend to deteriorate at higher frequencies, e.g., above 400 Hz.

[0021] As shown in Figure 10, loudspeakers 1004 and 1005 may be arranged in a close distance d to listener's ears 1002, e.g., below 0.5 m, or even 0.4 or 0.3 m, in order to generate the desired individual sound zones. One exemplary way to arrange loudspeakers 1004 and 1005 so close is to integrate loudspeakers 1004 and 1005 into headrest 1003 on which listener's head 1001 may rest. Another exemplary way is to dispose (directive) loudspeakers 1101 and 1102 in ceiling 1103, as shown in Figures 11 and 12. Other positions for the loudspeakers may be the B-pillar or C-pillar of the vehicle in combination with loudspeakers in the headrest or the ceiling. Alternatively or additionally, directional loudspeakers may be used instead of loudspeakers 1004 and 1005 or combined with loudspeakers 1004 and 1005 at the same position as or another position than loudspeakers 1004 and 1005.

[0022] Referring again to the setup shown in Figure 7, additional loudspeakers FLLSpkr, FLRSpkr, FRLSpkr, FRRSpkr, RLLSpkr, RLRSpkr, RRLSpkr and RRRSpkr may be disposed in the headrests of the seats in positions FLPos, FRPos, RLPos and RRPos. As can be seen from Figure 13, only loudspeakers that are arranged in close distance to a listener's ears, such as additional loudspeakers FLLSpkr, FLRSpkr, FRLSpkr, FRRSpkr, RLLSpkr, RLRSpkr, RRLSpkr and RRRSpkr, exhibit an improved magnitude frequency behavior at higher frequencies. The crosstalk cancellation is the difference between the upper curve and the three lower curves in Figure 13. However, due to the short distance between the loudspeaker and the ears such as a distance less than 0.5 m, or even less than 0.3 or 0.2 m, pre-ringing is relatively low, as shown in Figure 14, which illustrates the filter coefficients and thus the impulse responses of all equalizing filters, for providing crosstalk cancellation when using only headrest loudspeakers FLLSpkr, FLRSpkr, FRLSpkr, FRRSpkr, RLLSpkr, RLRSpkr, RRLSpkr and RRRSpkr, and, instead of the pre-ringing constraint, a modeling delay whose delay time may correspond to half of the filter length. Pre-ringing can be seen in Figure 14 as noise on the left side of the main impulse. Arranging loudspeakers in close distance to a listener's ears may in some applications already provide sufficient pre-ringing suppression and sufficient crosstalk cancellation if the modeling delay is sufficiently shortened in psychoacoustic terms, as can be seen in Figures 15 and 16.

[0023] When combining less distant loudspeakers FLLSpkr, FLRSpkr, FRLSpkr, FRRSpkr, RLLSpkr, RLRSpkr, RRLSpkr and RRRSpkr with a pre-ringing constraint instead of a modeling delay, the pre-ringing can be further decreased without deteriorating the crosstalk cancellation at positions FLPos, FRPos, RLPos and RRPos (i.e., the inter-position magnitude difference) at higher frequencies. Using more distant loudspeakers FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr instead of less distant loudspeakers FLLSpkr, FLRSpkr, FRLSpkr, FRRSpkr, RLLSpkr, RLRSpkr, RRLSpkr and RRRSpkr and a shortened modeling delay (the same delay as in the example described above in connection with Figures 15 and 16) instead of a pre-ringing constraint exhibits worse crosstalk cancellation, as can be seen in Figures 17 and 18. Figure 17 is a diagram illustrating the magnitude frequency responses at all four sound zones 701-704 using only loudspeakers FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr disposed at a distance of more than 0.5 m from positions FLPos, FRPos, RLPos and RRPos in combination with equalizing filters and the same modeling delay as in the example described in connection with Figures 15 and 16.

[0024] However, combining loudspeakers FLLSpkr, FLRSpkr, FRLSpkr, FRRSpkr, RLLSpkr, RLRSpkr, RRLSpkr and RRRSpkr, which are arranged in the headrests with the more distant loudspeakers of the setup shown in Figure 7, i.e., loudspeakers FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr, and, as shown in Figures 19 and 20, using a pre-ringing constraint instead of a modeling delay with reduced length can further decrease (compare Figures 18 and 20) the pre-ringing and increase (compare Figures 17 and 19) the crosstalk cancellation at positions FLPos, FRPos, RLPos and RRPos.

[0025] Alternative to a continuous curve, as shown in Figures 3-5, a stepped curve may also be employed in which, for example, the step width may be chosen to be frequency-dependent according to psychoacoustic aspects such as the Bark scale or the mel scale. The Bark scale is a psychoacoustic scale that ranges from one to 24 and corresponds to the first 24 critical bands of hearing. It is related to but somewhat less popular than the mel scale. It is perceived as noise by a listener when spectral drops or narrow-band peaks, known as temporal diffusion, occur within the magnitude frequency characteristic of a transfer function. Equalizing filters may therefore be smoothed during control operations or certain parameters of the filters such as the quality factor may be restricted in order to reduce unwanted noise. In case of smoothing, nonlinear smoothing that approximates the critical bands of human hearing may be employed. A nonlinear smoothing filter may be described by the following equation:


wherein n = [0, ..., N-1] relates to the discrete frequency index of the smoothed signal; N relates to the length of the fast Fourier transformation (FFT); ┌x-1/2┐ relates to rounding up to the next integer; α relates to a smoothing coefficient, e.g., (octave/3-smoothing) results in α = 21/3, in which A(jω) is the smoothed value of A(jω); and k is a discrete frequency index of the non-smoothed value A(jω), k ∈ [0, ..., N-1].

[0026] As can be seen from the above equation, nonlinear smoothing is basically frequency-dependent arithmetic averaging whose spectral limits change dependent on the chosen nonlinear smoothing coefficient α over frequency. To apply this principle to a MELMS algorithm, the algorithm is modified so that a certain maximum and minimum level threshold over frequency is maintained per bin (spectral unit of an FFT), respectively, according to the following equation in the logarithmic domain:




wherein f = [0, ..., fs/2] is the discrete frequency vector of length (N/2+1), N is the length of the FFT, fs is the sampling frequency, MaxGaindB is the maximum valid increase in [dB] and MinGaindB is the minimum valid decrease in [dB].

[0027] In the linear domain, the above equation reads as:





[0028] From the above equations, a magnitude constraint can be derived that is applicable to the MELMS algorithm in order to generate nonlinear smoothed equalizing filters that suppress spectral peaks and drops in a psychoacoustically acceptable manner. An exemplary magnitude frequency constraint of an equalizing filter is shown in Figure 21, wherein upper limit U corresponds to the maximum valid increase MaxGainLimdB(f) and lower limit L corresponds to the minimum allowable decrease MinGainLimdB(f). The diagrams shown in Figure 21 depict upper threshold U and lower threshold L of an exemplary magnitude constraint in the logarithmic domain, which is based on the parameters fs = 5,512 Hz, α = 21/24, MaxGaindB = 9 dB and MinGaindB = - 18 dB. As can be seen, the maximum allowable increase (e.g., MaxGaindB = 9 dB) and the minimum allowable decrease (e.g., MinGaindB = -18 dB) is achieved only at lower frequencies (e.g., below 35 Hz). This means that lower frequencies have the maximum dynamics that decrease with increasing frequencies according to the nonlinear smoothing coefficient (e.g., α = 21/24), whereby according to the frequency sensitivity of the human ear, the increase of upper threshold U and the decrease of lower threshold L are exponential over frequency.

[0029] In each iteration step, the equalizing filters based on the MELMS algorithm are subject to nonlinear smoothing, as described by the equations below.

Smoothing:



[0030] 






Double sideband spectrum:



[0031] 


with ASS(N-n)* = complex conjugate of ASS(N-n).

Complex spectrum:



[0032] 



[0033] Impulse response of the inverse fast Fourier transformation (IFFT):



[0034] A flow chart of an accordingly modified MELMS algorithm is shown in Figure 22, which is based on the system and method described above in connection with Figure 2. Magnitude constraint module 2201 is arranged between LMS module 207 and equalizing filter module 205. Another magnitude constraint module 2202 is arranged between LMS module 208 and equalizing filter module 206. The magnitude constraint may be used in connection with the pre-ringing constraint (as shown in Figure 22), but may be also used in standalone applications, in connection with other psychoacoustically motivated constraints or in connection with a modeling delay.

[0035] However, when combining the magnitude constraint with the pre-ringing constraint, the improvements illustrated by way of the Bode diagrams (magnitude frequency responses, phase frequency responses) shown in Figure 23 may be achieved in contrast to systems and methods without magnitude constraints, as illustrated by the corresponding resulting Bode diagrams shown in Figure 24. It is clear that only the magnitude frequency responses of systems and methods with magnitude constraints are subject to nonlinear smoothing, while the phase frequency responses are not essentially altered. Furthermore, systems and methods with magnitude constraints and pre-ringing constraints exert no negative influence on the crosstalk cancellation performance, as can be seen from Figure 25 (compared to Figure 8), but post-ringing may deteriorate, as shown in Figure 26, compared to Figure 9. In acoustics, post-ringing designates the appearance of noise after the actual sound impulse has occurred and can be seen in Figure 26 as noise on the right side of the main impulse.

[0036] An alternative way to smooth the spectral characteristic of the equalizing filters may be to window the equalizing filter coefficients directly in the time domain. With windowing, smoothing cannot be controlled according to psychoacoustic standards to the same extent as in the system and methods described above, but windowing of the equalizing filter coefficients allows for controlling the filter behavior in the time domain to a greater extent. Figure 27 is a diagram illustrating the magnitude frequency responses at sound zones 701-704 when using equalizing filters and only the more distant loudspeakers, i.e., loudspeakers FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr, in combination with a pre-ringing constraint and a magnitude constraint based on windowing with a Gauss window of 0.75. The corresponding impulse responses of all equalizing filters are depicted in Figure 28.

[0037] If windowing is based on a parameterizable Gauss window, the following equation applies:


wherein

and α is a parameter that is indirect proportional to the standard deviation σ and that is, for example, 0.75. Parameter α may be seen as a smoothing parameter that has a Gaussian shape (amplitude over time in samples), as shown in Figure 29.

[0038] The signal flow chart of the resulting system and method shown in Figure 30 is based on the system and method described above in connection with Figure 2. A windowing module 3001 (magnitude constraint) is arranged between LMS module 207 and equalizing filter module 205. Another windowing module 3002 is arranged between LMS module 208 and equalizing filter module 206. Windowing may be used in connection with the pre-ringing constraint (as shown in Figure 22), but may be also used in standalone applications, in connection with other psychoacoustically motivated constraints or in connection with a modeling delay.

[0039] Windowing results in no significant changes in the crosstalk cancellation performance, as can be seen in Figure 27, but the temporal behavior of the equalizing filters is improved, as can be seen from a comparison of Figures 26 and 28. Using a window as a magnitude constraint, however, does not result in such a huge smoothing of the magnitude frequency curve as with the other version, as will be apparent when comparing Figure 31 with Figures 23 and 24. Instead, the phase time characteristic is smoothed since smoothing is performed in the time domain, as will also be apparent when comparing Figure 31 with Figures 23 and 24. Figure 31 is a Bode diagram (magnitude frequency responses, phase frequency responses) of a system or method when only more distant loudspeakers in combination with a pre-ringing constraint and a magnitude constraint based on windowing with the modified Gauss window are used.

[0040] As windowing is performed after applying the constraint in the MELMS algorithm, the window (e.g., the window shown in Figure 29) is shifted and modified periodically, which can be expressed as follows:



[0041] The Gauss window shown in Figure 29 tends to level out when parameter α gets smaller and thus provides less smoothing at smaller values of parameter α. Parameter α may be chosen dependent on different aspects such as the update rate (i.e., how often windowing is applied within a certain number of iteration steps), the total number of iterations, etc. In the present example, windowing was performed in each iteration step, which was the reason for choosing a relatively small parameter α, since repeated multiplications of the filter coefficients with the window are performed in each iteration step and the filter coefficients successively decrease. An accordingly modified window is shown in Figure 32.

[0042] Windowing allows not only for a certain smoothing in the spectral domain in terms of magnitude and phase, but also for adjusting the desired temporal confinement of the equalizing filter coefficients. These effects can be freely chosen by way of a smoothing parameter such as a configurable window (see parameter α in the exemplary Gauss window described above) so that the maximum attenuation and the acoustic quality of the equalizing filters in the time domain can be adjusted.

[0043] Yet another alternative way to smooth the spectral characteristic of the equalizing filters may be to provide, in addition to the magnitude, the phase within the magnitude constraint. Instead of an unprocessed phase, a previously adequately smoothed phase is applied, whereby smoothing may again be nonlinear. However, any other smoothing characteristic is applicable as well. Smoothing may be applied only to the unwrapped phase, which is the continuous phase frequency characteristic, and not to the (repeatedly) wrapped phase, which is within a valid range of -π ≤ φ < π.

[0044] In order also to take the topology into account, a spatial constraint may be employed, which can be achieved by adapting the MELMS algorithm as follows:

wherein

and Gm(e) is the weighting function for the mth error signal in the spectral domain.

[0045] A flow chart of an accordingly modified MELMS algorithm, which is based on the system and method described above in connection with Figure 22 and in which a spatial constraint LMS module 3301 substitutes LMS module 207 and a spatial constraint LMS module 3302 substitutes LMS module 208, is shown in Figure 33. The spatial constraint may be used in connection with the pre-ringing constraint (as shown in Figure 33), but may also be used in standalone applications, in connection with psychoacoustically motivated constraints or in connection with a modeling delay.

[0046] A flow chart of an alternatively modified MELMS algorithm, which is also based on the system and method described above in connection with Figure 22, is shown in Figure 34. A spatial constraint module 3403 is arranged to control a gain control filter module 3401 and a gain control filter module 3402. Gain control filter module 3401 is arranged downstream of microphone 215 and provides a modified error signal e'1(n). Gain control filter module 3402 is arranged downstream of microphone 216 and provides a modified error signal e'2(n).

[0047] In the system and method shown in Figure 34, (error) signals e1(n) and e2(n) from microphones 215 and 216 are modified in the time domain rather than in the spectral domain. The modification in the time domain can nevertheless be performed such that the spectral composition of the signals is also modified, e.g., by way of the filter that provides a frequency-dependent gain. However, the gain may also simply be frequency independent.

[0048] In the example shown in Figure 34, no spatial constraint is applied, i.e., all error microphones (all positions, all sound zones) are weighted equally so that no special emphasis or insignificance is applied to particular microphones (positions, sound zones). However, a position-dependent weighting can be applied as well. Alternatively, sub-areas may be defined so that, for example, areas around the listener's ears may be amplified and areas at the back part of the head may be damped.

[0049] It may be desirable to modify the spectral application field of the signals supplied to the loudspeakers since the loudspeakers may exhibit differing electrical and acoustic characteristics. But even if all characteristics are identical, it may be desirable to control the bandwidth of each loudspeaker independently from the other loudspeakers since the usable bandwidths of identical loudspeakers with identical characteristics may differ when disposed at different locations (positions, vented boxes with different volume). Such differences may be compensated by way of crossover filters. In the exemplary system and method shown in Figure 35, a frequency-dependent gain constraint, herein also referred to as a frequency constraint, may be used instead of crossover filters to make sure that all loudspeakers are operated in an identical or at least similar fashion, e.g., such that none of the loudspeakers are overloaded, which leads to unwanted nonlinear distortions. Frequency constraints can be realized in a multiplicity of ways, two of which are discussed below.

[0050] A flow chart of an accordingly modified MELMS algorithm, which is based on the system and method described above in connection with Figure 34, but may be based on any other system and method described herein, with or without particular constraints, is shown in Figure 35. In the exemplary system shown in Figure 35, LMS modules 207 and 208 are substituted by frequency-dependent gain constraint LMS modules 3501 and 3502 to provide a specific adaptation behavior, which can be described as follows:


wherein k = 1, ... , K, K being the number of loudspeakers; m = 1, ..., M, M being the number of microphones; k,m(ejΩ,n) is the model of the secondary path between the kth loudspeaker and the mth (error) microphone at time n (in samples); and |Fk(ejΩ)| is the magnitude of the crossover filter for the spectral restriction of the signal supplied to the kth loudspeaker, the signal being essentially constant over time n.

[0051] As can be seen, the modified MELMS algorithm is essentially only a modification with which filtered input signals are generated, wherein the filtered input signals are spectrally restricted by way of K crossover filter modules with a transfer function Fk(ejΩ). The crossover filter modules may have complex transfer functions, but in most applications, it is sufficient to use only the magnitudes of transfer functions |Fk(ejΩ)| in order to achieve the desired spectral restrictions since the phase is not required for the spectral restriction and may even disturb the adaptation process. The magnitude of exemplary frequency characteristics of applicable crossover filters are depicted in Figure 36.

[0052] The corresponding magnitude frequency responses at all four positions and the filter coefficients of the equalizing filters (representing the impulse responses thereof) over time (in samples), are shown in Figures 37 and 38, respectively. The magnitude responses shown in Figure 37 and the impulse responses of the equalizing filters for establishing crosstalk cancellation shown in Figure 38 relate to four positions when applying equalizing filters in connection with exclusively more distant loudspeakers such as loudspeakers FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr in the setup shown in Figure 7 in combination with a frequency constraint, a pre-ringing constraint and a magnitude constraint, including windowing with a Gauss window of 0.25.

[0053] Figures 37 and 38 illustrate the results of the spectral restriction of the output signals by way of the crossover filter modules below 400 Hz, which is the minor influence of the front woofers FLSpkrL and FRSpkrL in the setup shown in Figure 7, and the absence of any significant influence on the crosstalk cancellation, as can be seen from a comparison of Figures 37 and 27. These results are also supported when comparing the Bode diagrams shown in Figures 39 and 31, in which the diagrams shown in Figure 39 are based on the same setup that forms the basis of Figures 37 and 38 and shows a significant change of the signal supplied to woofers FLSpkrL and FRSpkrL when they are next to front positions FLPos and FRPos. Systems and methods with frequency constraints as set forth above may tend to exhibit a certain weakness (magnitude drops) at low frequencies in some applications. Therefore, the frequency constraint may be alternatively implemented, e.g., as discussed below in connection with Figure 40.

[0054] A flow chart of an accordingly modified MELMS algorithm, as shown in Figure 40, is based on the system and method described above in connection with Figure 34, but may be alternatively based on any other system and method described herein, with or without particular constraints. In the exemplary system shown in Figure 40, a frequency constraint module 4001 may be arranged downstream of equalizing filter 205, and a frequency constraint module 4002 may be arranged downstream of equalizing filter 206. The alternative arrangement of the frequency constraint allows for reducing the complex influence (magnitude and phase) of the crossover filters in the room transfer characteristics, i.e., in the actual occurring transfer functions Sk,m(ejΩ,n) by way of prefiltering the signals supplied to the loudspeakers, and in the transfer functions of their models k,m(ejΩ,n), which is indicated in Figure 40 by Ŝ'k,m(ejΩ,n). This modification to the MELMS algorithm can be described with the following equations:




wherein Ŝ'k,m(ejΩ,n) is an approximation of S'k,m(ejΩ,n).

[0055] Figure 41 is a diagram illustrating the magnitude frequency responses at the four positions described above in connection with Figure 7 when equalizing filters are applied and only the more distant loudspeakers, i.e., FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr in the setup shown in Figure 7, are used in connection with a pre-ringing constraint, a magnitude constraint (windowing with a Gauss window of 0.25) and a frequency constraint that is included in the room transfer functions. The corresponding impulse responses are shown in Figure 42, and the corresponding Bode diagrams are shown in Figure 43. As can be seen in Figures 41-43, the crossover filters have a significant impact on woofers FLSpkrL and FRSpkrL next to front positions FLPos and FRPos. Particularly when comparing Figures 41 and 37, it can be seen that the frequency constraint on which the diagram of Figure 41 is based allows for a more distinct filtering effect at lower frequencies and that the crosstalk cancellation performance deteriorates a little bit at frequencies above 50 Hz.

[0056] Depending on the application, at least one (other) psychoacoustically motivated constraint may be employed, either alone or in combination with other psychoacoustically motivated or not psychoacoustically motivated constraints such as a loudspeaker-room-microphone constraint. For example, the temporal behavior of the equalizing filters when using only a magnitude constraint, i.e., non-linear smoothing of the magnitude frequency characteristic when maintaining the original phase (compare the impulse responses depicted in Figure 26), is perceived by the listener as annoying tonal post-ringing. This post-ringing may be suppressed by way of a post-ringing constraint, which can be described based on an energy time curve (ETC) as follows:

Zero padding:



[0057] 


wherein wk is the final set of filter coefficients for the kth equalizing filter in a MELMS algorithm with length N/2, and 0 is the zero column vector with length N.

FFT conversion:



[0058] 


ETC calculation:



[0059] 








wherein Wk,t(ejΩ) is the real part of the spectrum of the kth equalizing filter at the tth iteration step (rectangular window) and

represents the waterfall diagram of the kth equalizing filter, which includes all N/2 magnitude frequency responses of the single sideband spectra with a length of N/2 in the logarithmic domain.

[0060] When calculating the ETC of the room impulse response of a typical vehicle and comparing the resulting ETC with the ETC of the signal supplied to front left high-frequency loudspeaker FLSpkrH in a MELMS system or method described above, it turns out that the decay time exhibited in certain frequency ranges is significant longer, which can be seen as the underlying cause of post-ringing. Furthermore, it turns out that the energy contained in the room impulse response of the MELMS system and method described above might be too much at a later time in the decay process. Similar to how pre-ringing is suppressed, post-ringing may be suppressed by way of a post-ringing constraint, which is based on the psychoacoustic property of the human ear called (auditory) post-masking.

[0061] Auditory masking occurs when the perception of one sound is affected by the presence of another sound. Auditory masking in the frequency domain is known as simultaneous masking, frequency masking or spectral masking. Auditory masking in the time domain is known as temporal masking or non-simultaneous masking. The unmasked threshold is the quietest level of the signal that can be perceived without a present masking signal. The masked threshold is the quietest level of the signal perceived when combined with a specific masking noise. The amount of masking is the difference between the masked and unmasked thresholds. The amount of masking will vary depending on the characteristics of both the target signal and the masker, and will also be specific to an individual listener. Simultaneous masking occurs when a sound is made inaudible by a noise or unwanted sound of the same duration as the original sound. Temporal masking or non-simultaneous masking occurs when a sudden stimulus sound makes other sounds that are present immediately preceding or following the stimulus inaudible. Masking that obscures a sound immediately preceding the masker is called backward masking or pre-masking, and masking that obscures a sound immediately following the masker is called forward masking or post-masking. Temporal masking's effectiveness attenuates exponentially from the onset and offset of the masker, with the onset attenuation lasting approximately 20 ms and the offset attenuation lasting approximately 100 ms, as shown in Figure 44.

[0062] An exemplary graph depicting the inverse exponential function of the group delay difference over frequency is shown in Figure 45, and the corresponding inverse exponential function of the phase difference over frequency as the post-masking threshold is shown in Figure 46. "Post-masking" threshold is understood herein as a constraint to avoid post-ringing in equalizing filters. As can be seen from Figure 45, which shows a constraint in the form of a limiting group delay function (group delay differences over frequency), the post-masking threshold decreases when the frequency increases. While at a frequency of approximately 1 Hz, a post-ringing with a duration of around 250 ms may be acceptable for a listener, at a frequency of approximately 500 Hz, the threshold is already at around 50 ms and may reach higher frequencies with an approximate asymptotic end-value of 5 ms. The curve shown in Figure 45 can easily be transformed into a limiting phase function, which is shown in Figure 46 as phase difference curve over frequency. As the shapes of the curves of post-ringing (Figures 45 and 46) and pre-ringing (Figures 3 and 4) are quite similar, the same curve may be used for both post-ringing and pre-ringing but with different scaling. The post-ringing constraint may be described as follows:

[0063] Specifications:

is the time vector with a length of N/2 (in samples),
t0 = 0 is the starting point in time,
a0db = 0 dB is the starting level and
a1db = -60 dB is the end level.

Gradient:



[0064] 

is the gradient of the limiting function (in dB/s),
τGroupDelay(n) is the difference function of the group delay for suppressing post-ringing (in s) at frequency n (in FFT bin).

Limiting function:



[0065] LimFctdB(n,t) = m(n)tS is the temporal limiting function for the nth frequency bin (in dB), and


is the frequency index representing the bin number of the single sideband spectrum (in FFT bin).

Time compensation/scaling:



[0066] 



0 is the zero vector with length tMax, and

tMax is the time index in which the nth limiting function has its maximum.


Linearization:



[0067] 


Limitation of ETC:



[0068] 


Calculation of the room impulse response:



[0069] 

is the modified room impulse response of the kth channel (signal supplied to loudspeaker) that includes the post-ringing constraint.

[0070] As can be seen in the equations above, the post-ringing constraint is based here on a temporal restriction of the ETC, which is frequency dependent and whose frequency dependence is based on group delay difference function τGroupDelay(n). An exemplary curve representing group delay difference function τGroupDelay(n) is shown in Figure 45. Within a given time period τGroupDelay(n)fS, the level of a limiting function LimFctdB(n,t) shall decrease according to thresholds a0dB and a1db, as shown in Figure 47.

[0071] For each frequency n, a temporal limiting function such as the one shown in Figure 47 is calculated and applied to the ETC matrix. If the value of the corresponding ETC time vector exceeds the corresponding threshold given by LimFctdB(n, t) at frequency n, the ETC time vector is scaled according to its distance from the threshold. In this way, it is assured that the equalizing filters exhibit in their spectra a frequency-dependent temporal drop, as required by group delay difference function τGroupDelay(n). As group delay difference function τGroupDelay(n) is designed according to psychoacoustic requirements (see Figure 44), post-ringing, which is annoying to a listener, can be avoided or at least reduced to an acceptable degree.

[0072] Referring now to Figure 48, the post-ringing constraint can be implemented, for example, in the system and method described above in connection with Figure 40 (or in any other system and method described herein). In the exemplary system shown in Figure 48, combined magnitude and post-ringing constraint modules 4801 and 4802 are used instead of magnitude constraint modules 2201 and 2202. Figure 49 is a diagram illustrating the magnitude frequency responses at the four positions described above in connection with Figure 7 when equalizing filters are applied and only the more distant loudspeakers, i.e., FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr in the setup shown in Figure 7, are used in connection with a pre-ringing constraint, a magnitude constraint (windowing with a Gauss window of 0.25), a frequency constraint that is included in the room transfer functions and a post-ringing constraint.

[0073] The corresponding impulse responses are shown in Figure 50, and the corresponding Bode diagrams are shown in Figure 51. When comparing the diagram shown in Figure 49 with the diagram shown in Figure 41, it can be seen that the post-ringing constraint slightly deteriorates the crosstalk cancellation performance. On the other hand, the diagram shown in Figure 50 shows that post-ringing is less than in the diagram shown in Figure 42, which relates to the system and method shown in Figure 40. As is apparent from the Bode diagrams shown in Figure 51, the post-ringing constraint has some effect on the phase characteristics, e.g., the phase curves are smoothed.

[0074] Another way to implement the post-ringing constraint is to integrate it in the windowing procedure described above in connection with the windowed magnitude constraint. The post-ringing constraint in the time domain, as previously described, is spectrally windowed in a similar manner as the windowed magnitude constraint so that both constraints can be merged into one constraint. To achieve this, each equalizing filter is filtered exclusively at the end of the iteration process, beginning with a set of cosine signals with equidistant frequency points similar to an FFT analysis. Afterwards, the accordingly calculated time signals are weighted with a frequency-dependent window function. The window function may shorten with increasing frequency so that filtering is enhanced for higher frequencies and thus nonlinear smoothing is established. Again, an exponentially sloping window function can be used whose temporal structure is determined by the group delay, similar to the group delay difference function depicted in Figure 45.

[0075] The implemented window function, which is freely parameterizable and whose length is frequency dependent, may be of an exponential, linear, Hamming, Hanning, Gauss or any other appropriate type. For the sake of simplicity, the window functions used in the present examples are of the exponential type. Endpoint a1dB of the limiting function may be frequency dependent (e.g., a frequency-dependent limiting function a1dB(n) in which a1dB(n) may decrease when n increases) in order to improve the crosstalk cancellation performance.

[0076] The windowing function may be further configured such that within a time period defined by group delay function τGroupDelay(n), the level drops to a value specified by frequency-dependent endpoint a1dB(n), which may be modified by way of a cosine function. All accordingly windowed cosine signals are subsequently summed up, and the sum is scaled to provide an impulse response of the equalizing filter whose magnitude frequency characteristic appears to be smoothed (magnitude constraint) and whose decay behavior is modified according to a predetermined group delay difference function (post-ringing constraint). Since windowing is performed in the time domain, it affects not only the magnitude frequency characteristic, but also the phase frequency characteristic so that frequency-dependent nonlinear complex smoothing is achieved. The windowing technique can be described by the equations set forth below.

Specifications:



[0077] 

is the time vector with a length of N/2 (in samples),
t0 = 0 is the starting point in time,
a0db = 0 dB is the starting level and
a1dBMin= -120 dB is the lower threshold.

Level limiting:



[0078] 

n is a level limit,


is a level modification function,
a1dB(n) = LimLevdB(n)LevModFctdB(n), wherein


is the frequency index representing the bin number of the single sideband spectrum.

Cosine signal matrix:



[0079] CosMat(n,t) = cos(2πntS) is the cosine signal matrix.

Window function matrix:



[0080] 

is the gradient of the limiting function in dB/s,
τGroupDelay(n) is the group delay difference function for suppressing post-ringing at the nth frequency bin,
LimFctdB(n,t) = m(n)tS is the temporal limiting function for the nth frequency bin,

is the matrix that includes all frequency-dependent window functions.

Filtering (application):



[0081] 

is the cosine matrix filter, wherein wk is the kth equalizing filter with length N/2.

Windowing and scaling (application):



[0082] 

is a smoothed equalizing filter of the kth channel derived by means of the previously described method.

[0083] The magnitude time curves of an exemplary frequency-dependent level limiting function a1dB(n) and an exemplary level limit LimLevdB(n) are depicted in Figure 52. Level limiting function a1dB(n) has been amended according to level modification function LevModFctdB(n), shown as the amplitude frequency curve in Figure 53, to the effect that the lower frequencies have been less limited than the upper frequencies. The windowing functions WinMat(n,t), based on exponential windows, are illustrated in Figure 54 at frequencies 200 Hz (a), 2,000 Hz (b) and 20,000 Hz (c). Magnitude and post-ringing constraints can thus be combined with each other without any significant performance drops, as can further be seen in Figures 55-57.

[0084] Figure 55 is a diagram illustrating the magnitude frequency responses at the four positions described above in connection with Figure 7 when equalizing filters are applied and only the more distant loudspeakers, i.e., FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr in the setup shown in Figure 7, are used in connection with a pre-ringing constraint, a frequency constraint, a windowed magnitude and a post-ringing constraint. The corresponding impulse responses (amplitude time diagram) are shown in Figure 56, and the corresponding Bode diagrams are shown in Figure 57. The previously described windowing technique allows for a significant reduction of spectral components at higher frequencies, which is perceived by the listener as more convenient. It has to be noted that this special windowing technique is not only applicable in MIMO systems, but can also be applied to any other system and method that use constraints such as general equalizing systems or measurement systems.

[0085] In most of the aforementioned examples, only the more distant loudspeakers, i.e., FLSpkrH, FLSpkrL, FRSpkrH, FRSpkrL, SLSpkr, SRSpkr, RLSpkr and RRSpkr in the setup shown in Figure 7, were used. However, employing more closely arranged loudspeakers such as loudspeakers FLLSpkr, FLRSpkr, FRLSpkr, FRRSpkr, RLLSpkr, RLRSpkr, RRLSpkr and RRRSpkr may provide additional performance enhancement. Accordingly, in the setup shown in Figure 7, all loudspeakers, including the eight loudspeakers disposed in the headrests, are employed to assess the performance of a windowed post-ringing constraint in view of the crosstalk cancellation performance. It is assumed that a bright zone is established at the front left position and three dark zones are generated at the three remaining positions.

[0086] Figure 58 illustrates, by way of a magnitude frequency curve, a target function that is the reference for tonality in the bright zone and may be simultaneously applied to the pre-ringing constraint. The impulse responses of an exemplary equalizer filter based on the target function shown in Figure 58 with and without applied windowing (windowed post-ringing constraint) are depicted in Figure 59 as amplitude time curves in the linear domain and in Figure 60 as magnitude time curves in the logarithmic domain. It is apparent from Figure 60 that the windowed post-ringing constraint is capable of significantly reducing the decay time of the equalizing filter coefficients and thus of the impulse responses of the equalizing filters based on the MELMS algorithm.

[0087] From Figure 60, it can be seen that the decay is in accordance with psychoacoustic requirements, which means that the effectiveness of the temporal reduction increases successively when frequency increases without deteriorating the crosstalk cancellation performance. Furthermore, Figure 61 proves that the target function illustrated in Figure 58 is met almost perfectly. Figure 61 is a diagram illustrating the magnitude frequency responses at the four positions described above in connection with Figure 7 when using all loudspeakers (including the loudspeakers in the headrests) in the setup shown in Figure 7 and equalizing filters in combination with a pre-ringing constraint, a frequency constraint, a windowed magnitude and a windowed post-ringing constraint. The corresponding impulse responses are shown in Figure 62. In general, all types of psychoacoustic constraints such as pre-ringing constraints, magnitude constraints, post-ringing constraints and all types of loudspeaker-room-microphone constraints such as frequency constraints and spatial constraints may be combined as required.

[0088] Referring to Figure 63, the system and method described above in connection with Figure 1 may be modified not only to generate individual sound zones, but also to generate any desired wave fields (known as auralization). To achieve this, the system and method shown in Figure 1 has been modified in view of primary path 101, which has been substituted by controllable primary path 6301. Primary path 6301 is controlled according to source room 6302, e.g., a desired listening room. The secondary path may be implemented as a target room such as the interior of vehicle 6303. The exemplary system and method shown in Figure 63 is based on a simple setup in which the acoustics of desired listening room 6302 (e.g., a concert hall) are established (modeled) within a sound zone around one particular actual listening position with the same setup as shown in Figure 7 (e.g., the front left position in vehicle interior 6303). A listening position may be the position of a listener's ear, a point between a listener's two ears or the area around the head at a certain position in the target room 6303.

[0089] Acoustic measurements in the source room and in the target room may be made with the same microphone constellation, i.e., the same number of microphones with the same acoustic properties, and disposed at the same positions relative to each other. As the MELMS algorithm generates coefficients for K equalizing filters that have transfer function W(z), the same acoustic conditions may be present at the microphone positions in the target room as at the corresponding positions in the source room. In the present example, this means that a virtual center speaker may be created at the front left position of target room 6303 that has the same properties as measured in source room 6302. The system and method described above may thus also be used for generating several virtual sources, as can be seen in the setup shown in Figure 64. It should be noted that front left loudspeaker FL and front right loudspeaker FR correspond to loudspeaker arrays with high-frequency loudspeakers FLSpkrH and FRSpkrH and low-frequency loudspeakers FLSpkrL and FRSpkrL, respectively. In the present example, both source room 6401 and target room 6303 may be 5.1 audio setups.

[0090] However, not only may a single virtual source be modeled in the target room, but a multiplicity I of virtual sources may also be modeled simultaneously, wherein for each of the I virtual sources, a corresponding equalizing filter coefficient set Wi(z), I being 0, ..., I-1, is calculated. For example, when modeling a virtual 5.1 system at the front left position, as shown in Figure 64, I = 6 virtual sources are generated that are disposed according to the ITU standard for 5.1 systems. The approach for systems with a multiplicity of virtual sources is similar to the approach for systems with only one virtual source, which is that I primary path matrixes Pi(z) are determined in the source room and applied to the loudspeaker set up in the target room. Subsequently, a set of equalizing filter coefficients Wi(z) for K equalizing filters is adaptively determined for each matrix Pi(z) by way of the modified MELMS algorithm. The I × K equalizing filters are then superimposed and applied, as shown in Figure 65.

[0091] Figure 65 is a flow chart of an application of accordingly generated I × K equalizing filters that form I filter matrixes 6501-6506 to provide I = 6 virtual sound sources for the approximate sound reproduction according to the 5.1 standard at the driver's position. According to the 5.1 standard, six input signals relating to loudspeaker positions C, FL, FR, SL, SR and Sub are supplied to the six filter matrixes 6501-6506. Equalizing filter matrixes 6501-6506 provide I = 6 sets of equalizing filter coefficients W1(z)-W6(z) in which each set includes K equalizing filters and thus provides K output signals. Corresponding output signals of the filter matrixes are summed up by way of adders 6507-6521 and are then supplied to the respective loudspeakers arranged in target room 6303. For example, the output signals with k = 1 are summed up and supplied to front right loudspeaker (array) 6523, the output signals with k = 2 are summed up and supplied to front left loudspeaker (array) 6522, the output signals with k = 6 are summed up and supplied to subwoofer 6524 and so forth.

[0092] A wave field can be established in any number of positions, e.g., microphone arrays 6603-6606 at four positions in a target room 6601, as shown in Figure 66. The microphone arrays providing 4 × M are summed up in a summing module 6602 to provide M signals y(n) to subtractor 105. The modified MELMS algorithm allows not only for control of the position of the virtual sound source, but also for the horizontal angle of incidence (azimuth), the vertical angle of incidence (elevation) and the distance between the virtual sound source and the listener.

[0093] Furthermore, the field may be coded into its eigenmodes, i.e., spherical harmonics, which are subsequently decoded again to provide a field that is identical or at least very similar to the original wave field. During decoding, the wave field may be dynamically modified, e.g., rotated, zoomed in or out, clinched, stretched, shifted back and forth, etc. By coding the wave field of a source in a source room into its eigenmodes and coding the eigenmodes by way of a MIMO system or method in the target room, the virtual sound source can thus be dynamically modified in view of its three-dimensional position in the target room. Figure 67 depicts exemplary eigenmodes up to an order of M = 4. These eigenmodes, e.g., wave fields that have the frequency-independent shapes shown in Figure 67, may be modeled by way of specific sets of equalizing filter coefficients to a certain degree (order). The order basically depends on the sound system present in the target room such as the sound system's upper cutoff frequency. The higher the cutoff frequency is, the higher the order should be.

[0094] For loudspeakers in the target room that are more distant from the listener and that thus exhibit a cutoff frequency of fLim = 400 ... 600 Hz, a sufficient order is M = 1, which are the first N = (M + 1)2 = 4 spherical harmonics in three dimensions and N = (2M + 1) = 3 in two dimensions.


wherein c is the speed of sound (343 m/s at 20° C), M is the order of the eigenmodes, N is the number of eigenmodes and R is the radius of the listening surface of the zones.

[0095] By contrast, when additional loudspeakers are disposed much closer to the listener (e.g., headrest loudspeakers), order M may increase dependent on the maximum cutoff frequency to M = 2 or M = 3. Assuming that the distant field conditions are predominant, i.e., that the wave field can be split into plane waves, the wave field can be described by way of a Fourier Bessel series, as follows:


wherein

are the Ambisonic coefficients (weighting coefficients of the Nth spherical harmonic),

is a complex spherical harmonic of mth order, nth grade (real part σ = 1, imaginary part σ = -1), P(r,ω) is the spectrum of the sound pressure at a position r = (r, θ, ϕ), S(jω) is the input signal in the spectral domain, j is the imaginary unit of complex numbers and jm(kr) is the spherical Bessel function of the first species of mth order.

[0096] The complex spherical harmonics

may then be modeled by the MIMO system and method in the target room, i.e., by the corresponding equalizing filter coefficients, as depicted in Figure 68. By contrast, the Ambisonic coefficients

are derived from an analysis of the wave field in the source room or a room simulation. Figure 68 is a flow chart of an application in which the first N = 3 spherical harmonics are generated in the target room by way of a MIMO system or method. Three equalizing filter matrixes 6801-6803 provide the first three spherical harmonics (W,X and Y) of a virtual sound source for the approximate sound reproduction at the driver's position from input signal x[n]. Equalizing filter matrixes 6801-6803 provide three sets of equalizing filter coefficients W1(z)-W3(z) in which each set includes K equalizing filters and thus provides K output signals. Corresponding output signals of the filter matrixes are summed up by way of adders 6804-6809 and then supplied to the respective loudspeakers arranged in target room 6814. For example, the output signals with k = 1 are summed up and supplied to front right loudspeaker (array) 6811, the output signals with k = 2 are summed up and supplied to front left loudspeaker (array) 6810 and the last output signals with k = K are summed up and supplied to subwoofer 6812. At listening position 6813 then, the first three eigenmodes X, Y and Z are generated that together form the desired wave field of one virtual source.

[0097] Modifications can be made in a simple manner, as can be seen from the following example in which a rotational element is introduced while decoding:

wherein

are modal weighting coefficients that turn the spherical harmonics in the desired direction (θDes,ϕDes).

[0098] Referring to Figure 69, an arrangement for measuring the acoustics of the source room may include microphone array 6901 in which a multiplicity of microphones 6903-6906 are disposed on a headband 6902. Headband 6902 may be worn by a listener 6907 when in the source room and positioned slightly above the listener's ears. Instead of a single microphone microphone arrays may be used to measure the acoustics of the source room. The microphone arrays include at least two microphones arranged on a circle with a diameter corresponding to the diameter of an average listener's head and in a position that corresponds to an average listener's ears. Two of the array's microphones may be disposed at or at least close to the position of the average listener's ears.

[0099] Instead of a listener's head, any artificial head or rigid sphere with properties similar to a human head may also be used. Furthermore, additional microphones may be arranged in positions other than on the circle, e.g., on further circles or according to any other pattern on a rigid sphere. Figure 70 depicts a microphone array including a multiplicity of microphones 7002 on rigid sphere 7001 in which some of microphones 7002 may be arranged on at least one circle 7003. Circle 7003 may be arranged such that it corresponds to a circle that includes the positions of a listener's ears.

[0100] Alternatively, a multiplicity of microphones may be arranged on a multiplicity of circles that include the positions of the ears but that the multiplicity of microphones concentrates to the areas around where the human ears are or would be in case of an artificial head or other rigid sphere. An example of an arrangement in which microphones 7102 are arranged on ear cups 7103 worn by listener 7101 is shown in Figure 71. Microphones 7102 may be disposed in a regular pattern on a hemisphere around the positions of the human ears.

[0101] Other alternative microphone arrangements for measuring the acoustics in the source room may include artificial heads with two microphones at the ears' positions, microphones arranged in planar patterns or microphones placed in a (quasi-)regular fashion on a rigid sphere, able to directly measure the Ambisonic coefficients.

[0102] Referring again to the description above in connection with Figures 52-54, an exemplary process for providing a magnitude constraint with integrated post-ringing constraint as shown in Figure 72 may include iteratively adapting the transfer function of the filter module (7201), inputting a set of cosine signals with equidistant frequencies and equal amplitudes into the filter module upon adaption (7202), weighting signals output by the filter module with a frequency-dependent windowing function (7203), summing up the filtered and windowed cosine signals to provide a sum signal (7204), and scaling the sum signal to provide an updated impulse response of the filter module for controlling the transfer functions of the K equalizing filter modules (7205).

[0103] It is to be noted that in the system and methods described above that both the filter modules and the filter control modules may be implemented in a vehicle but alternatively only the filter modules may be implemented in the vehicle and the filter control modules may be outside the vehicle. As another alternative both the filter modules and the filter control modules may be implemented outside vehicle, e.g., in a computer and the filter coefficients of the filter module may be copied into a shadow filter disposed in the vehicle. Furthermore, the adaption may be a one-time process or a consecutive process as the case may be.


Claims

1. An acoustic multiple-input multiple-output system with M microphones and K loudspeakers comprising:

K ≥ 1 equalizing filter modules (103, 205, 206) that are arranged in signal paths downstream of an input signal path, the input signal path being supplied with an input signal, wherein the equalizing filter modules have controllable transfer functions, and

K filter control modules (106, 207, 208) that are configured to control the transfer functions of the equalizing filter modules (103, 205, 206) according to an adaptive control algorithm based on M>1 error signals and the input signal on the input signal path, characterized in that

the adaptive control algorithm comprises a windowed magnitude constraint with an integrated post-ringing constraint, the magnitude constraint and post-ringing constraint being psychoacoustic constraints that are configured to model psychoacoustic properties of a human ear, wherein the filter control modules (106, 207, 208) are configured to iteratively adapt the transfer functions of each of the K equalizing filter modules (103, 205, 206);

further characterized in that the windowed magnitude constraint with the integrated post-ringing constraint is configured to

input a set of cosine signals with equidistant frequencies and equal amplitudes into the K equalizing filter modules (103, 205, 206) upon adaption,

weight a signal output by the filter modules (103, 205, 206) with a frequency-dependent windowing function,

sum up the filtered and windowed cosine signals to provide a sum signal, and

scale the sum signal to provide an updated impulse response of the K equalizing filter modules (103, 205, 206) for controlling the transfer function of the K equalizing filter modules (103, 205, 206).


 
2. The system of claim 1, wherein the window function has a temporal length that decreases with increasing frequency.
 
3. The system of claim 1 or 2, wherein the window function is of exponential, linear, Hamming, Hanning or Gauss type.
 
4. The system of any of claims 1-3, wherein the K filter control modules are configured to apply a temporal limiting function with a start value and an end value to the signals output by the K equalizing filter modules (103, 205, 206).
 
5. The system of claim 4, wherein the end value of the limiting function decreases with increasing frequency.
 
6. The system of claim 4, wherein the end value of the limiting function decreases according to a cosine function.
 
7. A method for equalizing an acoustic multiple-input multiple-output system with M microphones and K loudspeakers comprising:

equalizing filtering with K ≥ 1 controllable transfer functions (103, 205, 206) in signal paths downstream of an input signal path, the input signal path being supplied with an input signal, and

controlling with K filter control signals (106, 207, 208) of the K controllable transfer functions (103, 205, 206) for filtering according to an adaptive control algorithm based on M>1 error signals and the input signal on the input signal path, characterized in that
the adaptive control algorithm comprises a windowed magnitude constraint with an integrated post-ringing constraint, the magnitude constraint and post-ringing constraint being psychoacoustic constraints that are configured to model psychoacoustic properties of a human ear, wherein controlling comprises iteratively adapting the transfer functions of the equalizing filtering;

further characterized in that the windowed magnitude constraint with the integrated post-ringing constraint comprises:

inputting a set of cosine signals with equidistant frequencies and equal amplitudes into the equalizing filtering upon adaption,

weighting signals output by the equalizing filtering with a frequency-dependent windowing function,

summing up the filtered and windowed cosine signals to provide a sum signal, and

scaling the sum signal to provide an updated impulse response of the equalizing filtering for controlling the K transfer functions of the equalizing filtering.


 
8. The method of claim 7, wherein the window function has a temporal length that decreases with increasing frequency.
 
9. The method of claim 7 or 8, wherein the window function is of exponential, linear, Hamming, Hanning or Gauss type.
 
10. The method of any of claims 7-9, wherein controlling further comprises applying a temporal limiting function with a start value and an end value to the signals output by the filter module.
 
11. The method of claim 10, wherein the end value of the limiting function decreases with increasing frequency.
 
12. The method of claim 10, wherein the end value of the limiting function decreases according to a cosine function.
 
13. A computer program product configured to cause a processor to execute the method of any of claims 7-12.
 


Ansprüche

1. Akustisches System mit mehreren Eingängen, mehreren Ausgängen, mit M Mikrofonen und K Lautsprechern, umfassend:

K ≥ 1 Entzerrerfiltermodule (103, 205, 206), in Signalwegen angeordnet sind, die einem Eingangssignalweg nachgeschaltet sind, wobei dem Eingangssignalweg ein Eingangssignal zugeführt wird, wobei die Entzerrerfiltermodule über steuerbare Übertragungsfunktionen verfügen, und

K Filtersteuermodule (106, 207, 208), die dazu ausgelegt sind, die Übertragungsfunktionen der Entzerrerfiltermodule (103, 205, 206) gemäß einem adaptiven Steueralgorithmus auf Grundlage von M > 1 Fehlersignalen und dem Eingangssignal auf dem Eingangssignalweg zu steuern, dadurch gekennzeichnet, dass

der adaptive Steueralgorithmus eine gefensterte Größenbeschränkung mit einer integrierten Post-Ringing-Beschränkung umfasst, wobei die Größenbeschränkung und die Post-Ringing-Beschränkung psychoakustische Beschränkungen sind, die dazu ausgelegt sind, psychoakustische Eigenschaften eines menschlichen Ohrs nachzubilden, wobei die Filtersteuermodule (106, 207, 208) dazu ausgelegt sind, die Übertragungsfunktionen jedes der K Entzerrerfiltermodule (103, 205, 206) iterativ anzupassen;

ferner dadurch gekennzeichnet, dass die gefensterte Größenbeschränkung mit der integrierten Post-Ringing-Beschränkung zu Folgendem ausgelegt ist

Eingeben eines Satzes von Kosinussignalen mit äquidistanten Frequenzen und gleichen Amplituden in die K Entzerrerfiltermodule (103, 205, 206) nach der Anpassung,

Gewichten eines von den Filtermodulen (103, 205, 206) ausgegebenen Signals mit einer frequenzabhängigen Fensterungsfunktion,

Summieren der gefilterten und gefensterten Kosinussignale, um ein Summensignal bereitzustellen, und

Skalieren des Summensignals, um eine aktualisierte Impulsantwort der K Entzerrerfiltermodule (103, 205, 206) zum Steuern der Übertragungsfunktion der K Entzerrerfiltermodule (103, 205, 206) bereitzustellen.


 
2. System nach Anspruch 1, wobei die Fensterfunktion eine zeitliche Länge aufweist, die mit steigender Frequenz abnimmt.
 
3. System nach Anspruch 1 oder 2, wobei die Fensterfunktion zum Typ exponentiell, linear, Hamming, Hanning oder Gauß gehört.
 
4. System nach einem der Ansprüche 1-3, wobei die K Filtersteuermodule dazu ausgelegt sind, eine zeitliche Begrenzungsfunktion mit einem Startwert und einem Endwert auf die Signale, die von den K Entzerrerfiltermodulen (103, 205, 206) ausgegeben werden, anzuwenden.
 
5. System nach Anspruch 4, wobei der Endwert der Begrenzungsfunktion mit steigender Frequenz abnimmt.
 
6. System nach Anspruch 4, wobei der Endwert der Begrenzungsfunktion gemäß einer Kosinusfunktion abnimmt.
 
7. Verfahren zum Entzerren eines akustischen Systems mit mehreren Eingängen, mehreren Ausgängen, mit M Mikrofonen und K Lautsprechern, umfassend:

Enzerrungsfiltern mit K ≥ 1 steuerbaren Übertragungsfunktionen (103, 205, 206) in Signalwegen, die einem Eingangssignalweg nachgeschaltet sind, wobei dem Eingangssignalweg ein Eingangssignal zugeführt wird, und

Steuern mit K Filtersteuersignalen (106, 207, 208) der K steuerbaren Übertragungsfunktionen (103, 205, 206) zum Filtern gemäß einem adaptiven Steueralgorithmus auf Grundlage von M > 1 Fehlersignalen und dem Eingangssignal auf dem Eingangssignalweg, dadurch gekennzeichnet, dass

der adaptive Steueralgorithmus eine gefensterte Größenbeschränkung mit einer integrierten Post-Ringing-Beschränkung umfasst, wobei die Größenbeschränkung und die Post-Ringing-Beschränkung psychoakustische Beschränkungen sind, die dazu ausgelegt sind, psychoakustische Eigenschaften eines menschlichen Ohrs nachzubilden, wobei das Steuern das iterative Anpassen der Übertragungsfunktionen des Enzerrungsfilterns umfasst;

ferner dadurch gekennzeichnet, dass die gefensterte Größenbeschränkung mit der integrierten Post-Ringing-Beschränkung Folgendes umfasst:

Eingeben eines Satzes von Kosinussignalen mit äquidistanten Frequenzen und gleichen Amplituden in das Entzerrungsfiltern nach der Anpassung,

Gewichten von Signalen, die von dem Entzerrungsfiltern ausgegeben werden, mit einer frequenzabhängigen Fensterungsfunktion,

Summieren der gefilterten und gefensterten Kosinussignale, um ein Summensignal bereitzustellen, und

Skalieren des Summensignals, um eine aktualisierte Impulsantwort des Entzerrungsfilterns zum Steuern der K Übertragungsfunktionen des Entzerrungsfilterns bereitzustellen.


 
8. Verfahren nach Anspruch 7, wobei die Fensterfunktion eine zeitliche Länge aufweist, die mit steigender Frequenz abnimmt.
 
9. Verfahren nach Anspruch 7 oder 8, wobei die Fensterfunktion zum Typ exponentiell, linear, Hamming, Hanning oder Gauß gehört.
 
10. Verfahren nach einem der Ansprüche 7-9, wobei das Steuern ferner das Anwenden einer zeitlichen Begrenzungsfunktion mit einem Startwert und einem Endwert auf die durch das Filtermodul ausgegebenen Signale umfasst.
 
11. Verfahren nach Anspruch 10, wobei der Endwert der Begrenzungsfunktion mit steigender Frequenz abnimmt.
 
12. Verfahren nach Anspruch 10, wobei der Endwert der Begrenzungsfunktion gemäß einer Kosinusfunktion abnimmt.
 
13. Computerprogrammprodukt, das dazu ausgelegt ist, einen Prozessor zu veranlassen, das Verfahren nach einem der Ansprüche 7-12 auszuführen.
 


Revendications

1. Système acoustique à entrées multiples et sorties multiples avec M microphones et K haut-parleurs comprenant :

K≥1 modules de filtrage d'égalisation (103, 205, 206) qui sont agencés dans des chemins de signaux en aval d'un chemin de signal d'entrée, le chemin de signal d'entrée étant alimenté par un signal d'entrée, dans lequel les modules de filtrage d'égalisation ont des fonctions de transfert pouvant être commandées, et

K modules de commande de filtre (106, 207, 208) qui sont configurés pour commander les fonctions de transfert des modules de filtrage d'égalisation (103, 205, 206) selon un algorithme de commande adaptatif basé sur M>1 signaux d'erreur et le signal d'entrée sur le chemin de signal d'entrée, caractérisé en ce que

l'algorithme de commande adaptatif comprend une contrainte d'amplitude fenêtrée avec une contrainte post-sonnerie intégrée, la contrainte d'amplitude et la contrainte post-sonnerie étant des contraintes psychoacoustiques qui sont configurées pour modéliser les propriétés psychoacoustiques d'une oreille humaine, dans lequel les modules de commande de filtre (106, 207, 208) sont configurés pour adapter de manière itérative les fonctions de transfert de chacun des K modules de filtrage d'égalisation (103, 205, 206) ;

caractérisé en outre en ce que la contrainte d'amplitude fenêtrée avec la contrainte post-sonnerie intégrée est configurée pour

entrer un ensemble de signaux cosinus avec des fréquences équidistantes et des amplitudes égales dans les K modules de filtrage d'égalisation (103, 205, 206) lors de l'adaptation,

pondérer un signal émis par les modules de filtrage (103, 205, 206) avec une fonction de fenêtrage dépendant de la fréquence,

additionner des signaux cosinus filtrés et fenêtrés pour fournir un signal de somme, et

mettre à l'échelle le signal de somme pour fournir une réponse impulsionnelle mise à jour des K modules de filtrage d'égalisation (103, 205, 206) afin de commander la fonction de transfert des K modules de filtrage d'égalisation (103, 205, 206) .


 
2. Système selon la revendication 1, dans lequel la fonction de fenêtre a une longueur temporelle qui diminue avec l'augmentation de la fréquence.
 
3. Système selon la revendication 1 ou 2, dans lequel la fonction de fenêtre est de type exponentiel, linéaire, Hamming, Hanning ou Gauss.
 
4. Système selon l'une quelconque des revendications 1 à 3, dans lequel les K modules de commande de filtre sont configurés pour appliquer une fonction de limitation temporelle avec une valeur de début et une valeur de fin aux signaux émis par les K modules de filtrage d'égalisation (103, 205, 206).
 
5. Système selon la revendication 4, dans lequel la valeur de fin de la fonction de limitation diminue avec l'augmentation de la fréquence.
 
6. Système selon la revendication 4, dans lequel la valeur de fin de la fonction de limitation diminue selon une fonction cosinus.
 
7. Procédé pour égaliser un système acoustique à entrées multiples et sorties multiples avec M microphones et K haut-parleurs comprenant :

le filtrage d'égalisation avec K≥1 fonctions de transfert pouvant être commandées (103, 205, 206) dans des chemins de signaux en aval d'un chemin de signal d'entrée, le chemin de signal d'entrée étant alimenté par un signal d'entrée, et

la commande avec K signaux de commande de filtre (106, 207, 208) des K fonctions de transfert pouvant être commandées (103, 205, 206) pour filtrer selon un algorithme de commande adaptatif basé sur M>1 signaux d'erreur et le signal d'entrée sur le chemin de signal d'entrée, caractérisé en ce que

l'algorithme de commande adaptatif comprend une contrainte d'amplitude fenêtrée avec une contrainte post-sonnerie intégrée, la contrainte d'amplitude et la contrainte post-sonnerie étant des contraintes psychoacoustiques qui sont configurées pour modéliser les propriétés psychoacoustiques d'une oreille humaine, dans lequel la commande comprend l'adaptation itérative des fonctions de transfert du filtrage d'égalisation ;

caractérisé en outre en ce que la contrainte d'amplitude fenêtrée avec la contrainte post-sonnerie intégrée comprend :

l'entrée d'un ensemble de signaux cosinus avec des fréquences équidistantes et des amplitudes égales dans le filtrage d'égalisation lors de l'adaptation,

la pondération de signaux émis par le filtrage d'égalisation avec une fonction de fenêtrage dépendant de la fréquence,

l'addition des signaux cosinus filtrés et fenêtrés pour fournir un signal de somme, et

la mise à l'échelle du signal de somme pour fournir une réponse impulsionnelle mise à jour du filtrage d'égalisation pour commander les K fonctions de transfert du filtrage d'égalisation.


 
8. Procédé selon la revendication 7, dans lequel la fonction de fenêtre a une longueur temporelle qui diminue avec l'augmentation de la fréquence.
 
9. Procédé selon la revendication 7 ou 8, dans lequel la fonction de fenêtre est de type exponentiel, linéaire, Hamming, Hanning ou Gauss.
 
10. Procédé selon l'une quelconque des revendications 7 à 9, dans lequel la commande comprend en outre l'application d'une fonction de limitation temporelle avec une valeur de début et une valeur de fin aux signaux émis par le module de filtrage.
 
11. Procédé selon la revendication 10, dans lequel la valeur de fin de la fonction de limitation diminue avec l'augmentation de la fréquence.
 
12. Procédé selon la revendication 10, dans lequel la valeur de fin de la fonction de limitation diminue selon une fonction cosinus.
 
13. Produit de programme informatique configuré pour amener un processeur à exécuter le procédé selon l'une quelconque des revendications 7 à 12.
 




Drawing























































































































Cited references

REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description




Non-patent literature cited in the description