EP1900250B1 - Electroacoustic method - Google Patents

Abstract

Disclosed is a method for improving the perceived acoustics of a room. Said method comprises the following steps: at least one sound signal is recorded in the vicinity of at least one sound source located in the room via a recording means; the recorded sound signal is replicated once with one respective pulse response function to generate different optimized sound signals, the respective pulse response function being chosen among a selection of predefined pulse response functions for the room; the optimized sound signals are reproduced via sound reproducing means located in the room. A different optimized sound signal is fed to each sound reproducing means while the optimized sound signals are fed to the reproducing means following one respective switching interval such that the same optimized sound signal is fed to no reproducing means in directly succeeding switching intervals.

Description

The present invention relates to a method for improving the perceived acoustics of a room, in particular for improving the room acoustics of an event hall, an open venue or the like by means of electro-acoustic measures.

In event rooms, the room acoustic conditions usually only meet very limited requirements of different types of events. For example, the room acoustic requirements for language events, music theater or music events with different instrumental composition differ significantly. Speech events such as speech theater or lectures require strong early reflections and sufficient signal volume to ensure good intelligibility. A reverberation time for such rooms for mainly language events should be between 0.5 and 1.5 s. In musical events, however, a stronger reverberation of the room depending on volume, style of music and occupation of the orchestra between 1.0 and 2.5 s in classical music and even 2 to 10 s in sacred music is desired. Furthermore, if necessary, lateral reflections to increase the spatiality and the perceived sound source size should be aimed at.

As a rule, the natural room acoustic conditions in function rooms are tailored to a specific type of event, such as a philharmonic orchestra, opera halls or lecture theaters, so that the acoustic results for different event types are often unsatisfactory. Within certain limits, an adaptation of the room acoustic conditions by variable absorber surfaces, adjustable reflectors or additional Hall chambers done, but this requires a high technical and financial effort.

Therefore, electroacoustic systems have been developed with which additional reflections and reverberations can be recorded in different areas of space. A particular difficulty is then a sound discoloring feedback of the rehearsed To prevent signals in recording microphones. Often, this is done by using microphones and loudspeakers with a pronounced directivity and by using dynamic filters to suppress any feedback frequencies. At the same time, however, a homogeneous recording of the sound sources in the entire stage area and of high quality is required for systems for electronic room acoustics influencing. The microphone arrangement should not be changed as far as possible in the case of different types of events. Highly directional microphones for suppressing feedback are therefore rather unsuitable.

In addition, the stage situation is very variable at opera performances or in multi-purpose halls, whereby the natural room-acoustic conditions can also change during the performance. A repositioning of recording microphones is then hardly possible.

Therefore, electroacoustic systems for improving the room acoustics have been proposed recently, which deal with the feedback problem in various ways. From the US 5,109,419 For example, an electroacoustic system is known in which sound signals recorded by means of a microphone are fed to a large number of Hall generators, each of which impress a synthetic reverb on the corresponding signals. These signals are reproduced via a plurality of speakers, which are arranged in a correspondingly equipped hall. The different Hall generators generate a random delay time and each have non-correlated transfer functions. The speakers are controlled so that adjacent speakers are not addressed by the same signal. In that, according to the US 5,109,419 the amplitude and time delays of the individual signals and the transfer functions used for the Hall generators are changed continuously in time and at random, so a feedback can be reduced. However, such synthetically generated additional reflections or reverberation may seem unnatural to a listener. The randomly synthesized reverberation generation also has the disadvantage that the special characteristics of the room in which such an electroacoustic system is used are insufficiently taken into account by the randomization and artificial generation of the reverb.

It is therefore an object of the present invention to provide a method and system for improving a perceived acoustics of a room in which natural acoustic signals of the room are complemented time-synchronously with electroacoustically generated signals as far as possible at every listener seat, without a sound discernible perceptible feedback or the additionally generated electro-acoustic signals can be localized by a listener.

This object is achieved by a method for improving a perceived acoustics of a room with the method steps of claim 1. Furthermore, an inventive sound system with the features of claim 17 solves this problem.

1. a) Aufnehmen mindestens eines Klangsignals über ein Aufnahmemittel in der Nähe mindestens einer Klangquelle in dem Raum;
2. b) Durchführen von jeweils einer Faltung des aufgenommenen Klangsignals mit jeweils einer Impulsantwortfunktion zum Erzeugen von unterschiedlichen optimierten Klangsignalen, wobei die jeweilige Impulsantwortfunktion aus einer Auswahl von vorgegebenen Impulsantwortfunktionen für den Raum ausgewählt werden;
3. c) Wiedergeben der optimierten Klangsignale über in dem Raum angeordnete Klangwiedergabemittel, wobei jedem Klangwiedergabemittel ein unterschiedliches optimiertes Klangsignal zugeführt wird, und wobei jeweils nach einem Umschaltzeitintervall die optimierten Klangsignale den Wiedergabemitteln derart zugeführt werden, dass keinem der Wiedergabemittel in direkt aufeinander folgenden Umschaltzeitintervallen dasselbe optimierte Klangsignal zugeführt wird.

Accordingly, a method is provided with the following method steps:

1. a) recording at least one sound signal via a recording means in the vicinity of at least one sound source in the room;
2. b) performing a convolution of the recorded sound signal, each having an impulse response function, for generating different optimized sound signals, the respective impulse response function being selected from a selection of predetermined impulse response functions for the space;
3. c) reproducing the optimized sound signals via sound reproducing means arranged in the space, wherein a different optimized sound signal is supplied to each sound reproducing means, and wherein after each switching time interval the optimized sound signals are supplied to the reproducing means so that none of the reproducing means receive the same optimized sound signal in directly successive switching time intervals is supplied.

1. a) mindestens ein Mikrofon zum Aufnehmen mindestens eines Klangsignals in der Nähe einer Klangquelle;
2. b) eine Faltungseinrichtung zum Falten des aufgenommenen Klangsignals mit vorgegebenen Impulsantwortfunktionen zum Erzeugen von optimierten Klangsignalen;
3. c) eine an die Faltungseinrichtung gekoppelte Umschalteinrichtung; und
4. d) an die Umschalteinrichtung gekoppelte Lautsprecher zum Wiedergeben der optimierten Klangsignale;
5. e) wobei die Umschalteinrichtung jeweils nach einem Umschaltzeitintervall die optimierten Klangsignale den Lautsprechern derart zuführt, dass keinem der Lautsprecher in direkt aufeinander folgenden Umschaltzeitintervallen dasselbe optimierte Klangsignal zugeführt wird.

Accordingly, the sound system according to the invention provides for a room, in particular for carrying out this method according to the invention:

1. a) at least one microphone for recording at least one sound signal in the vicinity of a sound source;
2. b) convolution means for convolving the recorded sound signal with predetermined impulse response functions to produce optimized sound signals;
3. c) a switching device coupled to the folding device; and
4. d) speakers coupled to the switching means for reproducing the optimized sound signals;
5. e) wherein the switching device in each case after a Umschaltzeitintervall the optimized sound signals to the speakers such that none of the speakers in directly successive Umschaltzeitintervallen the same optimized sound signal is supplied.

According to the invention, a plurality of optimized sound signals are generated from a sound signal of the corresponding sound source in the room, such as music or speech, by convolution with predetermined impulse response functions and reproduced by a plurality of speakers distributed in the room. According to the invention, the feedback security is achieved in that the corresponding emission locations undergo a periodic change in switching time intervals. By switching the reproduced optimized sound signals it is then not possible for a listener to locate the recorded sound. Rather, the inventive method leads to a particularly pleasant spatial sound. Furthermore, additional reflections or reverberation often have to be recorded just in the vicinity of the stage or the sound source in order to achieve the most pleasant possible room acoustics. In prior art systems, this has often been problematic since it can easily cause feedback. By means of the invention, however, the recording of additional reverberation in the vicinity of the stage is also made possible.

Preferably, at least three different optimized sound signals are generated, and at least three reproduction means are provided.

Preferably, the switching time interval is chosen so short that no sound-discoloring feedback is produced between a respective reproducing means and the recording means. As a sound discoloring usually unwanted level increases limited frequency ranges are felt. Already in a preliminary stage of the actual feedback, which can lead to a clearly perceptible whistling, for example, such increases in the signal level are disturbing even by local frequencies. According to the invention, such possible color discolorations are counteracted by the switching.

The switching time interval depends on the distance between the corresponding recording medium or recording microphone and the speakers arranged in the hall for reproducing the optimized sound signals. A preferred switching time interval is approximately in a range between 100 ms and 500 ms.

In one embodiment of the method according to the invention, the switching time interval is changed over time. Different interval durations can be set, for example, depending on the type of sound source and event. The interval duration can also be selected randomly, so that the optimized sound signals are fed completely stochastically to the speakers.

In a preferred embodiment of the method according to the invention, when a change is made from a first to a second optimized sound signal supplied to a playback means, the respective first sound signal optimized in the respective second sound signal is soft-blended. Such a smooth transition results in a particularly homogeneous sound carpet, which is perceived as pleasant in certain types of sound sources.

The optimized sound signals are preferably distributed in a periodic order or randomly stochastich distributed to the reproducing means.

In a further development of the method, a number of optimized sound signals is generated in step b), which is greater than the number of playback means. Then, at least one of the generated optimized sound signals is generated during a respective switching interval not guided to a playback means. In this alternative method, therefore, one of the optimized sound signals is not reproduced, whereby a certain randomization of the reproduction of all optimized sound signals is achieved. Preferably, the given impulse response functions are uncorrelated with each other.

• i) Messen einer Impulsantwortfunktion an einem Ort in dem Raum; und
• ii) Entfernen oder Zufügen von Hallanteilen zu den gemessenen Impulsantwortfunktionen zum Erzeugen einer entsprechenden vorgegebenen Impulsantwortfunktion.
• iii) Abspeichern der gemessenen Impulsantwortfunktion und/oder erzeugten vorgegebenen Impulsantwortfunktion.

In a development of the method, the steps may be provided for determining at least one of the predetermined impulse response functions:

• i) measuring an impulse response function at a location in the room; and
• ii) removing or adding Hall components to the measured impulse response functions to produce a corresponding predetermined impulse response function.
• iii) storing the measured impulse response function and / or generated predetermined impulse response function.

Most preferably, the predetermined impulse response functions are generated for multiple locations in the room. Because the natural impulse response functions are first measured for the adjustment or installation of the method according to the invention in the corresponding room, certain proportions can be added to determine the predetermined impulse response functions in order to achieve the most ideal impulse response function possible. By using impulse response functions measured in the room and by folding the sound signals recorded by the sound source with these modified original impulse response functions, a particularly natural sound impression is achieved according to the invention.

Furthermore, the predetermined impulse response functions can be generated such that early reflection components in a respective impulse response function substantially correspond to an impulse response obtained by a microphone recording with a high direct sound component and strong early reflections. Thus, natural microphone recordings are used to alter the original impulse response function measured in the room such that additional early reflections enhancing the direct sound component are perceived become. Likewise, the predetermined impulse response functions are preferably generated such that reverberation portions substantially correspond to microphone recordings having a high diffuse sound. Thus, it is preferred to use impulse response functions that appear as natural as possible for convolution, which consist of impulse response functions of the room and have additional components for direct sound and diffuse sound. It is equally possible to remove disturbing components in a measured impulse response function. Therefore, according to the invention, the realistic acoustic properties corresponding to the room are first of all determined and reworked or "improved". In this case, therefore, preferably not completely synthetically generated Hall signals are used but "natural" Hall of the room prepared.

In a further development of the method according to the invention, the optimized sound signals during reproduction are subjected to a periodic level change. This periodic level change further reduces the risk of feedback and creates an improved sense of space in a listener. Preferably, the time constant of the periodic level change is then selected as a function of a time constant for a feedback setup in the room.

In a further development of the method according to the invention, the optimized sound signals are delayed such that the reproduced optimized sound signals and the sound signals of the sound source arrive at at least one location of the room substantially simultaneously and preferably the sound signals of the sound source arrive earlier.

In addition, if there are multiple sound sources, the optimized sound signals are delayed such that a direct sound of the sound sources at each location of the room always arrives in time before a reflection component of a reproduced optimized sound signal.

In another preferred embodiment of the method, a plurality of individual sound signals of the sound source are recorded via a plurality of recording means and the corresponding individual sound signals are mixed with one another to form a recorded sound signal for further processing. The individual sound signals are delayed depending on the positions of the recording means. By using several recording media or microphones near the stage it is possible to achieve a particularly homogenous sound, taking into account the delay differences between the microphone and the sound source.

It is also advantageous if the single-tone signals or the recorded sound signals are filtered to reduce higher-frequency components. By filtering, the absorption property of the transmission medium for the sound in the room can be simulated and taken into account.

The predetermined impulse response functions, for example based on a realistic determination, are preferably stored in the sound system according to the invention in a memory coupled to the convolution device. Furthermore, it is preferable for the switching device to be followed by a level adjusting device for periodically changing the levels of the optimized sound signals and / or a delay device for delaying the optimized sound signals.

The speakers in the room are advantageously arranged such that preferably at least three speakers are perceptible at a particular location by a listener in the room. By means of three loudspeakers an improvement of the sound properties is already possible by the electroacoustic measures according to the invention. Perceptible in this case does not necessarily mean that an amplification of the original sound signal of the sound source is made and a listener hears the appropriately rehearsed signals, but that the sound enhancement or the improvement of the acoustics of the room is perceptible.

A particularly preferred embodiment of the sound system according to the invention provides a plurality of signal branches, each with at least one recorded sound signal, a convolution device, a switching device and a group of speakers, wherein the corresponding predetermined impulse response functions are selected in dependence on the positions of the speakers.

The invention further comprises an event hall equipped with a sound system according to the invention.

Fig.1:

eine schematische Darstellung des erfindungsgemäßen Klangsystems in einem Konzertsaal;

Fig. 2:

ein Beispiel einer vorgegebenen Impulsantwortfunktion;

Fig. 3:

eine bevorzugte Ausführungsform des erfindungsgemäßen Klangsystems;

Fig. 4:

eine Darstellung von Gewichtungsfunktionen zur Pegelabsenkung; und

Fig. 5:

eine bevorzugte Weiterbildung des erfindungsgemäßen Klangsystems.

Further advantageous embodiments and modifications of the invention are the subject of the dependent claims and the following description of the embodiments. In the following the invention will be explained in more detail by means of embodiments with reference to the figures. It shows:

Fig.1:

a schematic representation of the sound system according to the invention in a concert hall;

Fig. 2:

an example of a given impulse response function;

3:

a preferred embodiment of the sound system according to the invention;

4:

a representation of weighting functions for level reduction; and

Fig. 5:

a preferred embodiment of the sound system according to the invention.

In the figures, unless otherwise stated, the same or functionally identical elements have been given the same reference numerals.

The Fig. 1 shows an inventive sound system 1 for a room 2, for example, a concert or performance hall. In the hall 2, a stage area 3 is provided, which serves here schematically as a sound source. In the vicinity of the sound source 3, a sound signal E1 (t) is recorded via a microphone 4. The recorded sound signal E1 (t) is fed to a convolver 5, which may be implemented, for example, as a digital signal processor or PC. The folding device 5 is coupled to a memory 6 which holds stored impulse response functions stored. In the example shown here, the convolution device 5 preferably performs three convolutions on the recorded sound signal E1 (t), each having an impulse response function gi (t), thus producing three optimized sound signals L1 (t), L2 (t), L3 (t): $$\mathrm{Li}\left(\mathrm{t}\right)=\int \mathrm{d}\tau \mathrm{egg}\left(\mathrm{t}\right)\mathrm{gi}\left(\mathrm{t}-\tau \right)$$

The three optimized sound signals L1, L2, L3 are supplied to a switching device 7, which is coupled to the folding device 5.

The switching device has a first, second and third input 8, 9 and 10 and three outputs 11, 12 and 13. The switching device 7 switches the optimized sound signals L1, L2, L3 in such a way to their outputs 11, 12 and 13 that, for example during a first Umschaltzeitvalvall the length .DELTA.t the signal L1 can be tapped as a loudspeaker signal N1 at the first output 11, the second optimized sound signal L2 at the second output 12 as the second loudspeaker signal N2 can be tapped off and the third optimized sound signal L3 at the third output 13 as the third Speaker signal N3 can be tapped. In a second or subsequent switching time interval L1 is then fed to the output 12, L2 to the output 13 and L3 to the output 11. In a third switching interval, an assignment L1 to output 13, L2 to output 11 and L3 to output 12 could then follow. The corresponding loudspeaker signals N1, N2, N3 are supplied to loudspeakers 14, 15 and 16, which are arranged in the concert hall 2 and reproduce the corresponding signals.

A listener in the room 2, for example at the place 17, on the one hand perceives the direct sound from the stage 3 and on the other hand reverberation parts which are emitted by the system 1 according to the invention through the loudspeakers 14, 15, 16. The subjective sound impression for a listener results essentially from the original acoustic properties of the room 2 and the predetermined impulse responses with which the recorded signal E1 (t) is folded.

In general, the acoustics of a room can be characterized by its impulse response or its corresponding transfer functions. The corresponding impulse response function designates the sound pressure profile at a specific location of the room as a function of time when a delta-shaped acoustic wave is emitted at another location in the room.

An exemplary impulse response gi (t) is in the Fig. 2 shown. At time 0, a short pulse, such as a shot or clapping, caused and recorded the corresponding sound pressure or the energy density of the sound signal elsewhere. The solid curve A1 can, for example, the impulse response function of in Fig. 1 represented space 2 for a transmitted pulse on the stage 3 and the received sound pressure at the position 20 correspond.

The first early components correspond to a direct sound DS, which propagates without any detours via reflections on any surfaces to the receiver or listener. As early reflections that significantly influence the perceived clarity and distinctness of the sound signals, the parts of the impulse response function gi (t) present in a range up to approximately 80 ms after the direct sound DS are usually referred to. Hall components at longer times are perceived as diffuse DF. As a rule, rooms which have an impulse response function which corresponds approximately to a "fir-tree" and an envelope, such as, for example, the curve A2, are perceived as particularly pleasant. The predetermined and stored impulse response functions which are used in the convolution device 5 can now be generated in such a way that, for a harmonic sound impression, missing reverberation parts, for example the dashed line A3 in FIG Fig. 2 , which are added to the measured impulse response function. This can be done, for example, by recording impulse response functions in rooms with particularly good acoustics, and then inserting such reverberation units into a measured impulse response of the room to be improved. Likewise, disturbing components in a measured impulse response function can be removed and the predetermined impulse response functions generated in this way can be used as the basis for the convolution in the convolver 5.

Generally, by folding a sound signal of a first space with the impulse response of a second space, the sound characteristics of the second space can be impressed on the sound signals. The folding takes place as described in equation (1). However, if such a folded signal is reproduced in the second room, here for example in the hall 2, without further processing, then feedback can occur and a listener would also be able to locate the additionally rehearsed reverb via the speakers, which would normally be annoying.

According to the invention, the folded or optimized signals L1, L2, L3 are therefore replaced by the switching device 7. As predetermined impulse response functions for convolution, for example impulse response functions recorded in the hall at particularly favorable listener positions come into consideration, such as, for example, in the first third of a recitaled concert hall. Then, at adjacent positions 17, 18, 19, 20 in the hall 2, the impulse responses can be recorded and used in modified or Hall added form as a given impulse response function for the convolution. Other improvements are also possible by adding or removing certain reverb portions of these impulse response functions.

Overall, a particularly pleasant perceived impulse response function is used according to the invention. By storing such predetermined impulse response functions, the sound system according to the invention can also be easily adapted to spatial changes in the concert hall 2. Corresponding predetermined impulse response functions for use in speech events, chamber music, orchestral music or lecture events can thus be kept flexible. The sound system according to the invention can also respond to spatial changes by drawing in dividing walls.

The Fig. 3 shows a preferred embodiment of the sound system 100 according to the invention. There is provided a microphone 4 for receiving a sound signal E1, which is fed to a folding device 5. The folding device 5 performs convolutions of impulse response functions stored in the coupled memory with the recorded recording signal E1. The six corresponding convolution signals L1, L2, L3, L4, L5, L6 are fed to the six inputs 21-26 of a switching device 7.

The switching device 7 has four outputs 27, 28, 29, 30, on which four folding signals M1, M2, M3, M4 can be tapped off. The switching device 7 switches, for example, in a stochastic change the folding signals L1-L6 to the outputs 27-30. In this case, two of the optimized sound signals are each over a respective Umschaltzeitintervall .DELTA.t N1-N6 not switched through. In this case, the assignment of the signals L1-L6 present at the inputs of the switching device to the signals M1, M2, M3, M4 which can be tapped off at the outputs 27, 28, 29, 30 of the switching device 7 is preferably repeated periodically with a period duration of approximately 200 ms to 500 ms changed a given algorithm. For example, in the exemplary embodiment considered here, two signals L1-L6 (different signals at successive switching time intervals) are masked out. On adjacent outputs 27, 28, 29, 30 of the switching device 7 are therefore never on the one hand never signals that were generated by means of a convolution with the same impulse response function, and on the other hand, the assignment of the signals differs on the outputs 27, 28, 29, 30 in successive following switching time intervals. Two of the signals L1-L6 are also hidden here, in each case other signals are hidden in successive Umschaltzeitintervallen. By way of example only, a random selection of the signals not to be switched through or to be hidden may be mentioned for each switching time interval.

The corresponding convolution signals M1-M4 are coupled to inputs 31, 32, 33, 34 of a level adjustment device 35. This dynamic level adjustment means 35 provides at its outputs 36-39 weighted convolution signals N1, N2, N3, N4. The level adjuster 35 periodically lowers the respective convolution signals M1-M4 to the level-matched convolution signals N1-N4. This can be done, for example, by a time-dependent weighting factor wi (t): $$\mathrm{Ni}\left(\mathrm{t}\right)=\mathrm{Wed.}\left(\mathrm{t}\right)\mathrm{wi}\left(\mathrm{t}\right)$$

wobei i = 1, ...5 und j = 1, ...4 läuft. Dabei sind die jeweiligen Verzögerungszeiten Δtij an die Position der jeweiligen Lautsprecher 46-55 angepasst, denen das jeweilige Lautsprechersignal zugeführt ist. Ferner können die einzelnen pegelangepassten Faltungssignale N1, N2, N3, N4 auch mit einem Gewichtungsfaktor aij versehen werden. In der Regel genügt es, eine Überlagerung der pegelangepassten Faltungssignale N1-N4 direkt am Zuhörerort durch die Wiedergabe per Lautsprecher 46-55 zu erreichen. In diesem Fall dient die Verzögerungseinrichtung lediglich der Verzögerung der Faltungssignale N1-N4. Eine zusätzliche Superposition gemäß der Gleichung (3) kann jedoch vorteilhaft sein, wenn an bestimmten Zuhörerplätzen wiedergegebene Lautsprechersignale wahrnehmbar sind, die von verschiedenen Zweigen eines erweiterten Klangsystems erzeugt werden. Eine entsprechende Weiterbildung mit drei Zweigen ist im Folgenden in de Figur 5 näher erläutert.In the Fig. 4 For example, weighting factors w1, w2, w3, w4 are shown as a function of time t. In this case, the respective periodic level reduction by the weighting factors wi (t) is set in time such that a corresponding time constant for the reduction of the respective level is selected such that it essentially corresponds to a time constant for a feedback building up in the room. Preferably, the periodicity of the level decreases of Δτ ≠ Δt is thus chosen differently compared to the Umschaltzeitintervalle. Only the orders of magnitude of the time constants for the level reduction and the switching time intervals are similar. Alternatively, stochastic or random level reductions can also be imposed on the convolution signals M1, M2, M3, M4. The dynamic level-lowered Convolution signals N1, N2, N3, N4 are finally fed to a delay device 40, in which the level-matched folding signals N1, N2, N3, N4 are each subjected to delays and as loudspeaker signals 01, 02, 03, 04, 05 to amplifier devices 41, 42, 43 , 44, 45 are guided. To the amplifier device 41-45 finally arranged in the corresponding room or concert hall speakers 46-55 are connected. The loudspeaker signals 01-05 are generated in the delay device 40, for example as superpositions of the level-matched convolution signals N1, N2, N3, N4: $$\mathrm{Oi}={\displaystyle \underset{j}{\Sigma }}\mathrm{nj}\left(\mathrm{t}-\mathrm{\Delta tij}\right)\mathrm{ij}.$$

where i = 1, ... 5 and j = 1, ... 4 is running. The respective delay times Δtij are adapted to the position of the respective loudspeakers 46-55, to which the respective loudspeaker signal is supplied. Furthermore, the individual level-matched convolution signals N1, N2, N3, N4 can also be provided with a weighting factor aij. As a rule, it suffices to achieve a superimposition of the level-matched convolution signals N1-N4 directly at the listener location by the reproduction by loudspeakers 46-55. In this case, the delay means only serves to delay the convolution signals N1-N4. An additional superposition according to equation (3), however, may be advantageous if loudspeaker signals reproduced at certain listener seats are perceivable, which are generated by different branches of an extended sound system. A corresponding development with three branches is described below in de FIG. 5 explained in more detail.

The loudspeaker signals 01-05 are then respectively adapted so that at virtually every listener seat of the hall a direct sound signal which propagates through the transmission medium (air) in the hall from the sound source, always arrives before the additional sound signal generated by the sound signal. The speakers are then preferably arranged so that a listener seat is operated by at least three speakers or recorded echo signals.

The described sound system, which is operated by the method according to the invention, on the one hand has a very high level of feedback reliability due to the time and direction variance the recorded signals via the speakers. On the other hand, it is not possible for a listener to locate the recorded signals directly, since the periodic or stochastic change of the sound signals 01-05 supplied to the loudspeakers takes place locally after a respective switchover time interval which may, for example, be 100-500 ms changed. As a result, the sound appears natural and lively. As a result of this time and direction variance, a listener's psychoacoustically attains an increase in attention, as does, for example, the change in direction of the sound as a result of a movement of a musician when playing the respective instrument. Thus, even without a level increase of the additionally recorded by the sound system according to the invention acoustic signals, the audience sense is improved. Due to the temporal and spatial variance of the additionally recorded sound signals, the perceived reflections generated in this way become more conspicuous. Therefore, only a relatively low level of the recorded signals is required.

By using substantially real impulse responses whose fine structures are "improved" and thus correspond to the sound image or the perceived acoustics of a real room with a particularly pleasant acoustics, in the event room equipped with a system according to the invention an overall sound field arises from the superposition of the natural surround sound and the added convolution signals and a feedback of this recording via the loudspeakers to the microphone inputs of the system, without disturbing feedbacks can occur. Thus, the natural sound of a room with unsatisfactory perceived acoustics can be added by the sound system according to the invention additional natural sounding Hall to create a favorable sound. The perceived acoustics of the room can be flexibly adapted by different impulse response functions, for example for language events, orchestral music, chamber music or solo voices.

In the Figures 5A and 5B a development of the sound system according to the invention is shown. Essentially, the development 101 has three branches each with a convolver 5, 105, 205, a switching device 7, 107, 207, a dynamic level adjustment device 35, 135, 235 and an equalization device 40, 140, 240. Memory 6, 106, 206 are respectively coupled to the folding devices 5, 105, 205 in which Prescribed pulse-response functions are stored. Loudspeaker groups 146, 147, 148 are connected to the equalization devices 40, 140, 240 via amplifiers 141, 142, 143. Essentially, the branches work by the method according to the invention, as it is to the Fig. 3 has been described.

The sound signals G (t), H (t), K (t) supplied to the convolvers 5, 105, 205 are provided by a mixer 102 to the delayed and optionally filtered sound signals F1 recorded by microphones 303, 304, 305, 306, F2, F3 are supplied.

The microphones 303, 304, 305, 306 are arranged in the vicinity of a stage 3, on which one or more sound sources are present. When using the sound system of the invention in an example 5000m ³ large multi-purpose hall, which is to be used for conventional acoustic reinforced events as well as for opening the theater or orchestra concerts, is an arrangement of the microphones 303, 304, 305, 306 at a distance of about 2 m in front of the Stage and a height of 5m above the stage level. For example, these microphones 303, 304, 305 can be designed as high-quality studio condenser microphones with a directional characteristic "wide kidney". Thus, usually all sound sources in the stage area can be detected well.

The thus-detected sound signals E1, E2, E3, E4 are pre-amplified in an amplifier 307 and then supplied to the mixer device 102 as amplified sound signals F1, F2, F3, F4. The mixer 102 delays and mixes the respective pre-amplified sound signals F1, F2, F3, F4. The signal processing is similar, as already shown for example in equation (3), where the Oi (t) corresponds to the signals G (t), H (t) and K (t) and Nj (t) the pre-amplified sound signals Fj (t) corresponds. The sound signals G (t), H (t), K (t) supplied to the convolution devices 5, 105, 205 thus have portions of all the signals E1, E2, E3, E4 received by the four microphones 303, 304, 305, 306, wherein signal components from more distant microphones 303, 304, 305, 306 are taken into account as a function of the sound propagation time between the microphone and the loudspeakers 146, 147, 148 coupled to the respective branch. For example, if a loudspeaker group 146 of the first branch 100 is provided in a rear right area of the room, in the mixer and Retarder 102 signals of the microphones 205, 206 more delayed than the signals of the microphones 203, 204, which are arranged in the vicinity of the right stage area. Accordingly, the recorded sound signals of the microphones 303, 304, 305, 306 for loudspeakers 147 arranged in the middle of the room are then delayed approximately equally.

The mixer device 102 can be implemented, for example, as a digital signal processor, to which the pre-amplified signals F1, F2, F3, F4 are supplied in digitized form (for example via a high-resolution analog-to-digital converter, which is not shown here). The further processing of the branch signals G (t), H (t), K (t) can then take place, for example, by means of a customary personal computer, to which the signals are coupled via a digital input of a sound card. Today's computer technology makes it possible, for example, with several networked computers, practically parallel to fold the three times six different impulse response functions with the signals G (t), H (t), K (t) in real time and output again to a DSP in periodic change , The tasks of the folding device 5, 105, 205 and the switching device 7, 107, 207 can therefore be taken over by a particularly powerful PC or several networked PCs.

The corresponding post-coupled digital signal processor (DSP) performs the tasks of dynamic level adjustment in the level adjusters 35, 135, 235 and the tasks of the delay and equalizer devices 40, 140, 240. The latter further performs signal quality enhancing filtering or equalization to perform the transfer function to compensate for the speaker. The corresponding output signals 01-05, T1-T5 and X1-X5 are digitally converted to analog (for example, with a conventional DA converter, which are not shown here) and the amplifiers 141, 142, 143 supplied.

The amplifiers each feed two to three loudspeakers at their outputs, which are regularly fitted distributed in the ceiling and wall area of the hall. In some cases, speakers are also placed below podiums or grandstands. The density of the speaker network should be chosen so that each listener seat is powered by 3-4 speakers.

The present invention thus provides a method and a sound system for improving a perceived acoustics of a room, which enables a natural and lively sound image based on real determined spatial properties by means of electro-acoustic interference. By folding with predetermined impulse response functions, which are constructed from measured impulse response functions of ideally ideal rooms, a particularly pleasant surround sound is achieved. By changing the Einkopplungsrichtung the rehearsed additional optimized sound signals by switching in the switching devices unwanted localization of the additional signals is avoided by a listener and achieved a particularly high level of feedback security. By the method and system according to the invention, the room acoustics of event rooms can be significantly improved.

Although the present invention has been described in terms of preferred embodiments, it is not limited thereto, but modifiable in a variety of ways. The number of parallel convolution signals can be changed as desired. The use of the invention is not limited to closed rooms, but a corresponding sound system is also suitable for improving the room acoustics perception of open or semi-open venues. Furthermore, synthesized impulse response functions can also be used as predetermined impulse response functions. The waveforms shown in the figures are to be understood as exemplary only for explanation. Size specifications, for example, the volume of the rooms to be sounded or microphone distances, too, are also to be understood only as an example and virtually any change or scalable.

LIST OF REFERENCE NUMBERS

1

sound system

2

event hall

3

stage

4

microphone

5

convolver

6

Storage

7

switchover

8, 9, 10

entrance

11, 12, 13

output

14, 15, 16

speaker

17 - 20

listening position

21 - 26

entrance

28 - 30

output

31 - 34

entrance

35 - 39

output

40

equalizer

41 - 45

amplifier

46 - 55

speaker

100

sound system

101

sound system

102

mixer

105

convolver

106

Storage

107

switchover

135

Level matching device

140

delay means

141, 142, 143.

amplifier

146, 147, 148

Speaker group

205

convolver

206

Storage

207

switchover

235

Level matching device

240

delay means

303-306

microphone

307

amplifier

A1

Impulse response function

A2

envelope

A3

sound component

DS

direct sound

DF

diffuse sound

E1-E4

microphone signal

F1-F4

pre-amplified signal

G, H, K

recorded sound signal

L1-L5, P1-P6

convolution signal

U1-U6

convolution signal

M1-M4, Q1-Q4

switched convolution signal

V1-V4

switched convolution signal

N1-N4, R1-R4

Level-adjusted signal

Y1-Y4

Level-adjusted signal

O1-O5, T1-T5

Loudspeaker signal

X1-X5

Loudspeaker signal

.delta.t

switching time

w1-w4

weighting function

Claims (29)

1. Method of improving the audible acoustics of a room (1), comprising the method steps of:

   a) recording at least one sound signal (E1 (t)) with a recording means (4) in the vicinity of at least one sound source (3) in the room (1);

   b) convolving each recorded sound signal (E1(t)) with a respective impulse response function (gi(t)) for generating different optimised sound signals (L1 (t), L2(t), L3(t)), each impulse response function being selected from a selection of predetermined impulse response functions for the room (1);

   c) playing back the optimised sound signals (N1, N2, N3) with the sound playback means (14, 15, 16) arranged in the room (1), a different optimised sound signal (N1, N2, N3) being fed to each sound playback means (14, 15, 16), and the optimised sound signals (N1, N2, N3) being fed in each case to the playback means (14, 14, 16) after a predetermined switching time interval (Δt) in such a way that the same optimised sound signal (N1, N2, N3) is not fed to any of the playback means (14, 15, 16) in consecutive switching time intervals (Δt).
2. Method according to claim 1, characterised in that at least three different optimised sound signals (L1, L2, L3) are generated and at least three playback means (14, 15, 16) are provided.
3. Method according to either claim 1 or 2, characterised in that the switching time interval (Δt) in step c) is selected to be short enough, preferably in a range of from 100 to 500 ms, to prevent feedback which would colour the sound between a playback means (14, 15, 16) and the recording means (4).
4. Method according to any one of the preceding claims, characterised in that, in step c), during a change from a first to a second optimised sound signal (N1, N2, N3) fed to a playback means (14, 15, 16), each first sound signal is cross-faded into each second optimised sound signal (N1, N2, N3).
5. Method according to any one of the preceding claims, characterised in that the optimised sound signals (N1, N2, N3) are fed to the playback means (14, 15, 16) in a periodic sequence or in a stochastically distributed manner.
6. Method according to any one of the preceding claims, characterised in that the duration of the switching time interval (Δt) is changed.
7. Method according to claim 1, characterised in that a number of optimised sound signals (L1 to L6) are generated in step b), this number being greater than the number of playback means (14, 15, 16), and in that at least one of the optimised sound signals (L1 to L6) is not fed to a playback means (14, 15, 16) in step c) during a switching interval (Δt).
8. Method according to any one of the preceding claims, characterised in that the predetermined impulse response functions (gi(t)) are not correlated with one another.
9. Method according to any one of the preceding claims, characterised in that the method of determining at least one of the predetermined impulse response functions (gi(t)) comprises the step of:

   i) measuring an impulse response function (gi(t)) at a location (17) in the room (2).
10. Method according to claim 9, characterised in that the impulse response function (gi(t)) measured is stored.
11. Method according to either claim 9 or claim 10, characterised in that a further step is provided:

    ii) removing or adding reverberation components (A3) to the impulse response function (gi(t)) measured to generate a corresponding predetermined impulse response function.
12. Method according to any one of claims 9 to 11, characterised in that the predetermined impulse response functions (gi(t)) are generated by measuring impulse response functions at different locations (17 to 20) in the room (2).
13. Method according to any one of the preceding claims, characterised in that the predetermined impulse response functions (gi(t)) are generated in such a way that early reflection components basically correspond to microphone recordings with a high direct sound component (DS)
14. Method according to any one of the preceding claims, characterised in that the predetermined impulse response functions (gi(t)) are generated in such a way that reverberation components basically correspond to microphone recordings with high diffuse sound (DF).
15. Method according to any one of the preceding claims, characterised in that the optimised sound signals (M1 to M4) are subjected to a periodic change in level during playback.
16. Method according to claim 15, characterised in that a time constant for the periodic change in level is selected as a function of a time constant for a build up of feedback in the room.
17. Method according to any one of the preceding claims, characterised in that the optimised sound signals (N1, N2, N3, N4) are delayed in such a way that the played-back optimised sound signals (O1 to 05) and the sound signals of the sound source (3) reach at least one location in the room (2) substantially simultaneously.
18. Method according to any one of the preceding claims, characterised in that, in the case of a plurality of sound sources, the optimised sound signals (N1 to N4) are delayed in such a way that direct sound from the sound sources reaches each location in the room before a reflection component of a played-back optimised sound signal (O1 to 05).
19. Method according to any one of the preceding claims, characterised in that a plurality of individual sound signals (E1 to E4) from the sound source (3) are recorded in step a) by a plurality of recording means (303, 304, 305, 306) and the corresponding individual sound signals (E1 to E4) are mixed together to form a sound signal (G(t), H(t), K(t)) for further processing, the individual sound signals (E1 to E4) being delayed as a function of the positions of the recording means (303, 304, 305, 306).
20. Method according to any one of the preceding claims, characterised in that the individual sound signals (E1 to E4) or the recorded sound signals are filtered to reduce highfrequency components.
21. Sound system (1) for a room (2), in particular for carrying out the method according to at least one of claims 1 to 20, comprising:

    a) at least one microphone (3) for recording at least one sound signal (E1(t)) in the vicinity of the sound source (3);

    b) a convolution device (5) for convolving the recorded sound signal (E1 (t)) with predetermined impulse response functions (gi(t)) for generating at least three optimised sound signals (L1, L2, L3);

    c) a switching device (7) coupled to the convolution device (5); and

    d) loudspeakers (14, 15, 16), coupled to the switching device (7), for playing back the optimised sound signals (L1, L2, L3);

    e) the switching device (7) in each case feeding the optimised sound signals (L1, L2, L3) to the loudspeakers (14, 15, 16) after a switching time interval (Δt) in such a way that the same optimised sound signal (L1, L2, L3) is not fed to any of the loudspeakers (14, 15, 16) in consecutive switching time intervals (Δt).
22. Sound system (1) according to claim 21, characterised in that at least three optimised sound signals (L1, L2, L3) are generated by convolving the recorded sound signal (E1) with three different impulse response functions (gi(t)) in each case and at least three loudspeakers (14, 15, 16) are provided.
23. Sound system (1) according to either claim 21 or claim 22, characterised in that the predetermined impulse response functions (gi(t)) are stored in a storage means (6) coupled to the convolution device (5).
24. Sound system (1) according to any one of claims 21 to 23, characterised in that a level adjustment device (35) for periodically changing the level of the optimised sound signals (M1, M2, M3, M4) is connected downstream of the switching device (7).
25. Sound system according to any one of claims 21 to 24, characterised in that a delay device (40) is provided for delaying the optimised sound signals (M1, M2, M3, M4).
26. Sound system (1) according to any one of claims 21 to 25, characterised in that the loudspeakers are arranged in the room (2) in such a way that at least three loudspeakers (14, 15, 16) can be heard by a listener at any location (17) in the room.
27. Sound system (1, 100, 101) according to any one of claims 21 to 26, characterised in that the following are further provided:

    - a plurality of microphones (303, 304, 305, 306) arranged in the vicinity of a stage (3) in the room (2) to record a plurality of individual sound signals (E1 to E4); and

    - a mixing device (307, 102) for mixing the individual sound signals (E1 to E4) to form the recorded sound signal (G(t), H(t), K(t)).
28. Sound system (101) according to any one of claims 21 to 27, characterised in that the sound system (1) has a plurality of signal paths, each having at least one recorded sound signal (G(t), H(t), K(t)), a convolution device (5, 105, 205), a switching device (7, 107, 207) and a set of loudspeakers (146, 147, 148), the corresponding predetermined impulse response functions (gi(t)) being selected as a function of the position of the loudspeakers.
29. Function hall (2) comprising a sound system (1) according to any one of claims 21 to 28.

Images