FIELD OF THE INVENTION

The invention relates to sound reproduction using a speaker array and in particular, but not exclusively to sound reproduction for a surround sound system. BACKGROUND OF THE INVENTION

In recent years, a new class of surround sound systems has been introduced in the market. These systems attempt to produce a convincing surround sound experience with a loudspeaker arrangement situated only in front of the user, rather than having to put five or more speakers around the room as in conventional systems. The impetus for developing such systems is that many consumers do not like to have a large number of speakers in their homes, and are typically especially concerned about the need to have speakers at the back (plus the corresponding wires). There can be many reasons for this: the room geometry or furniture arrangement does not allow for suitable placement, it is aesthetically undesirable to have many speakers and wires in the room, etc.

This has led to the development of products for rendering of multi channel surround sound from a single speaker box. In such approaches, the spatial perception is typically achieved by directional radiation of sound, and the achieved spatial experience is highly dependent on the characteristics of this directional radiation.

Systems have been developed wherein directional radiation of sound is controlled by use of complex and advanced audio beam steering algorithms together with a speaker array comprising a plurality of audio drivers. However, although such approaches may result in advantageous operation in many scenarios, they tend to be complex and costly. In particular, the beam forming algorithms tend to have high computational resource demands and require complex circuitry for controlling the signals fed to each audio driver to achieve the desired beam shape.

Systems have been proposed that provide advantageous results in many embodiments without having the associated cost and complexity implications of traditional beam steering. For example, a single speaker box system has been developed wherein the left front and surround channels are reproduced by a dipole configuration of two loudspeakers. Similarly, the right front and surround channels are reproduced by a dipole configuration of two other loudspeakers. The notch (or specifically the null) direction of the dipole is steered electronically (by delaying the signal to one of the two dipole speakers) towards the position of the listener. This results in the listener not receiving any (or at least very little) sound for the spatial channel directly from the loudspeakers, so that the sound that reaches him consists mainly of sound reflected from the walls. This has proven to give a very wide, spacious, and enveloping sound sensation. Thus, an improved sound sensation is achieved with very low complexity.

The approach may for example also be used in compact stereo systems or surround sound systems based on two separate loudspeaker units, e.g. one for the left front- and surround channels and one for the right front- and surround channels.

However, although the technique is able to create a desirable wide, enveloping, and diffuse sound sensation, it also has some associated drawbacks. In some scenarios, the front channels may be perceived as spatially less well defined, and the sound stage and individual localized sounds sources may be perceived slightly non-focussed and "blurry". This may in some cases degrade the spatial experience.

Thus, in many sound reproduction systems, such as those described above, conflicting requirements exist that a directed sound radiation may seek to address. However, such conflicting requirements often lead to suboptimum trade-offs. For example, it is in many systems desirable to provide a sound stage with both a wide, enveloping and diffuse sensation and well defined sound sources. Furthermore, it is at the same time desirable to have low complexity and thus a system providing an improved trade-off between such requirements would be advantageous. In particular, a sound reproduction system allowing increased flexibility, reduced complexity, an improved spatial experience, reduced cost, reduced computational requirements, improved spatial definition, increased sound stage width, facilitated implementation and/or improved performance would be advantageous.

SUMMARY OF THE INVENTION

Accordingly, the Invention seeks to preferably mitigate, alleviate or eliminate one or more of the above mentioned disadvantages singly or in any combination.

According to an aspect of the invention there is provided a sound reproduction system comprising: a speaker array comprising at least a first driver, a second driver and a third driver; a receiver for receiving an input signal; and a drive circuit for generating from the input signal a first drive signal for the first driver, a second drive signal for the second driver, and a third drive signal for the third driver; wherein the drive circuit is arranged to generate the first drive signal and the third drive signal to drive the first driver and the third driver in a dipole configuration; and to direct a notch of the sound radiation pattern of the speaker array in a first direction by generating the second drive signal with at least a first offset relative to the first drive signal, the first offset comprising at least one offset selected from the group of a phase offset and a gain offset.

The invention may provide an improved sound reproduction system in many scenarios. In particular, an improved trade-off between different requirements may be achieved and in many embodiments an advantageous combination of a wide sound stage with well defined sound sources can be achieved. The system may have low complexity and in particular complex beam steering can be avoided. The approach may in many cases for example allow the direction of the notch to be controlled simply by adjusting a (possibly frequency dependent) gain difference for the driving of the second speaker relative to the first and third speaker.

In many embodiments, a three-loudspeaker configuration may be used to create a directional response where two drivers are used as an unsteered dipole pair with the third driver being used to cancel the sound field of the dipole pair in a selected direction. The cancellation angle can in many scenarios be controlled simply by changing the (possibly frequency-dependent) gain balance between the dipole pair and the second driver.

The approach may for example allow an advantageous rendering of spatial side and/or surround channels in a surround sound system using a single speaker array located in front of the listener. In particular an improved spatial definition may be achieved while maintaining a wide sound image and low complexity. For example, in comparison to a simple dipole arrangement, an improved spatial definition may be achieved without significant complexity increase. Thus, a clearer and spatially more defined sound image can be achieved while maintaining a wide sound stage.

Furthermore, the approach may allow improved power efficiency at low frequencies and/or may provide improved robustness to small variations in system

parameters. In particular, the presence of a clear notch in the desired direction will be more robust and insensitive to variations in various system parameters.

The approach may allow facilitated and/or improved control of the direction of the notch and may in particular allow facilitated dynamic adaptation and variations of the notch direction. The dipole configuration may result in a far field dipole radiation pattern. The dipole configuration may specifically be achieved by the first and third drive signal being substantially identical while having a phase difference of 180°, i.e. the same signal may be fed to the first and third driver but with opposite phases. The notch may specifically be a null. The first offset may specifically be a gain offset and e.g. combined with a phase offset in the frequency domain.

In many embodiments, the first offset may specifically comprise a gain offset combined with a substantially 90° phase offset (often in the interval of [80°; 100°] or even [85°;95°]).

The first direction may be different than a direction of the null (or notch) of the dipole configuration of the first and third drivers. The drive circuit may specifically be arranged to modify a direction of a notch of the sound radiation pattern of the speaker array for the input signal from a direction of a dipole notch for the dipole by generating the second drive signal with the first offset (selected from the group of a phase offset and a gain offset) relative to the first drive signal.

In accordance with an optional feature of the invention, the first offset is frequency dependent.

This may provide improved performance in many embodiments and may in particular compensate for frequency dependent variations in the in-air combinations of sound pressure levels from the individual drivers.

The first offset may be a frequency dependent gain combined with a frequency independent phase offset, e.g. in the interval of [80°; 100°] or even [85°;95°].

In some embodiments, the frequency dependent gain may be substantially proportional to:

Where Δχ/2 is the inter-driver distance, c is the speed of sound, Θ is the desired notch direction, and ω is the frequency.

In accordance with an optional feature of the invention, the drive circuit is arranged to provide a frequency dependent transfer function for generating the first drive signal from the input signal. The drive circuit may similarly be arranged to provide a frequency dependent transfer function for generating the third drive signal from the input signal. Indeed, the frequency response for the first and third drivers may be the same.

This may allow an improved audio quality and may in particular allow compensation for frequency variations and equalization resulting from the interaction between the different drivers of the array.

In particular, the frequency response of the transfer function may provide a compensation for a frequency variation resulting from a frequency variation of the first offset as well as a frequency dependent in-air combining of sound pressure levels from the first driver, the second driver and the third driver in a first direction.

In some embodiments, the generation of the first and/or third drive signals may include a filtering contribution substantially according to:

where Δχ/2 is the inter-driver distance, c is the speed of sound, Θ is the desired notch direction, <¾ is a desired angle of the frequency compensation and ω is the frequency.

In accordance with an optional feature of the invention, a phase difference between a transfer function for generating the first drive signal from the input signal and a transfer function for generating the second drive signal from the input signal is in the interval of 80° to 100° in a frequency interval of 100 to 2 kHz.

This may provide improved performance and/or reduced complexity. In some scenarios a phase difference of the transfer function for generating the first drive signal from the input signal and the transfer function for generating the second drive signal may be in the interval of 85° to 95° in a frequency interval of 100 Hz to 2 kHz.

In particular, the drive signal for the second driver may be phase offset by substantially 90° relative to the drive signals for the first and/or third drivers. The phase offset may specifically be substantially constant over the operating frequency range.

The gain difference between a transfer function for generating the first drive signal from the input signal and a transfer function for generating the second drive signal from the input signal may be frequency dependent. Thus, the transfer function for the second driver may have a varying (in the frequency domain) gain offset and constant phase offset relative to the first and/or third driver.

In accordance with an optional feature of the invention, the drive circuit comprises a first signal processing path for generating the first drive signal, a second signal processing path for generating the second drive signal, and a third signal processing path for generating the third drive signal; and the first offset may be generated by the second signal processing path having a different frequency response than the first signal processing path and the third signal processing path.

This may provide a particularly advantageous implementation.

In accordance with an optional feature of the invention, the first signal processing path and the third signal processing path have substantially identical gain frequency responses.

This may provide a particularly advantageous implementation and may reduce complexity while maintaining high performance. The gain responses may differ by no more than 2dB for a frequency range from 500Hz to 2 kHz. The second signal processing path may have a different gain and may e.g. deviate by more than 6 dB for at least some frequencies.

In accordance with an optional feature of the invention, the first signal processing path, the second signal processing path and the third signal processing path comprise a shared signal processing path segment and separate signal processing path segments for each of the first signal processing path, the second signal processing path and the third signal processing path; and the shared signal processing path segment comprises a filter having a frequency dependent gain and a frequency dependent gain of a separate signal processing path segment for the first signal processing path is different from frequency dependent gain of a separate signal processing path segment for the second signal processing path.

This may provide a particularly advantageous implementation and may reduce complexity while maintaining high performance.

In accordance with an optional feature of the invention, the first signal processing path and the third signal processing path comprise a common filter having a frequency gain variation and a gain of the second signal processing path is substantially frequency independent.

This may provide a particularly advantageous implementation and may reduce complexity while maintaining high performance. The gain of the second signal processing path may vary by no more than 2dB in a frequency range from 500Hz to 2 kHz. The common filter may specifically be a first order low pass filter having a slope of around 6 dB per octave.

In accordance with an optional feature of the invention, the common filter is independent of the first direction; and the sound system further comprises means for modifying the first direction by modifying a frequency independent gain offset of the second signal processing path relative to the first signal processing path.

This may provide a particularly advantageous implementation and may reduce complexity while maintaining high performance. In particular it may allow a low complexity and/or a robust approach for controlling the direction of the notch. The approach may for example allow a particularly advantageous approach for (possibly dynamically) adapting the operation of the system to the specific characteristics and requirements of the particular audio environment.

In accordance with an optional feature of the invention, the sound reproduction system further comprises: a circuit for dividing a received signal into an ambient sound signal and a non-ambient sound signal, and for generating the input signal to comprise the non-ambient sound signal but not the ambient sound signal.

This may provide an improved audio experience in many embodiments.

In accordance with an optional feature of the invention, the sound reproduction system further comprises: a circuit for introducing a time offset for the second drive signal relative to at least one of the first drive signal and the third drive signal, the time offset corresponding to a geometric characteristic of relative positions of at least two drivers of the speaker array.

This may provide an improved audio experience in many embodiments.

In accordance with an optional feature of the invention, the sound system is arranged to render at least one spatial channel, which specifically may be a front side channel.

The invention may provide a particular advantageous rendering of a spatial channel in a surround sound system. The approach may in particular be arranged to provide particularly advantageous performance for a side or surround channel with an improved trade-off between the conflicting requirements for such sound reproduction.

According to an aspect of the invention there is provided a driver for a loudspeaker arrangement comprising a speaker array with at least a first driver, a second driver and a third driver; the driver comprising: a receiver for receiving an input signal; and a drive circuit for generating from the input signal a first drive signal for the first driver, a second drive signal for the second driver, and a third drive signal for the third driver; wherein the drive circuit is arranged to generate the first drive signal and the third drive signal to drive the first driver and the third driver in a dipole configuration; and to direct a notch of the sound radiation pattern of the speaker array in a first direction by generating the second drive signal with at least a first offset relative to the first drive signal, the first offset comprising at least one offset selected from the group of a phase offset and a gain offset.

According to an aspect of the invention there is provided a method of operation for a driver for a loudspeaker arrangement comprising a speaker array with at least a first driver, a second driver and a third driver; the method comprising: receiving an input signal; and generating from the input signal a first drive signal for the first driver, a second drive signal for the second driver, and a third drive signal for the third driver; wherein the generating comprises generating the first drive signal and the third drive signal to drive the first driver and the third driver in a dipole configuration; and directing a notch of the sound radiation pattern of the speaker array in a first direction by generating the second drive signal with at least a first offset relative to the first drive signal, the first offset comprising at least one offset selected from the group of a phase offset and a gain offset.

These and other aspects, features and advantages of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will be described, by way of example only, with reference to the drawings, in which

Fig. 1 illustrates an example of a speaker system for a surround sound system; Fig. 2 illustrates an example of elements of a sound reproduction system in accordance with some embodiments of the invention;

Fig. 3 illustrates an example of a radiation pattern for a dipole configuration; Fig. 4 illustrates an example of a speaker coupling for a dipole configuration; Figs. 5 and 6 illustrate examples of elements of a sound reproduction system in accordance with some embodiments of the invention;

Fig. 7 illustrates an example of frequency responses for an equalization filter for a sound reproduction system in accordance with some embodiments of the invention; and

Figs. 8 to 13 illustrate examples of elements of a sound reproduction system in accordance with some embodiments of the invention. DETAILED DESCRIPTION OF SOME EMBODIMENTS OF THE INVENTION

The following description focuses on embodiments of the invention applicable to a spatial sound reproduction system and specifically to a surround sound reproduction system. However, it will be appreciated that the invention is not limited to this application but may be applied to many other sound reproduction systems including for example a stereo sound reproduction system.

Fig. 1 illustrates an example of a speaker system for a surround sound system in accordance with some embodiments of the invention.

In the example, the speaker system is a single speaker box 101 that comprises two speaker arrays 103, 105. In the example, the speaker box 101 comprises a first speaker array 103 which is used to reproduce the left side front and surround channels and a second speaker array 105 which is used to reproduce the right side front and surround channels. In addition, both speaker arrays 103, 105 are used to reproduce the centre signal.

The following description will focus on the rendering of a spatial channel of a surround sound system and in particular to the rendering of the left front side channel by the left speaker array 103 of Fig. 1. It will be appreciated that the described approach can be applied to other channels including to other spatial channels of a surround sound signal. It will also be appreciated that the system may be symmetric and that the right surround channel(s) may be generated equivalently using the right side array 105 of Fig. 1.

The left speaker array 103 comprises a first driver 107, a second driver 109 and a third driver 111 which are controlled to provide a directional sound radiation. The drivers may for example be loudspeaker units or other sound transducers.

In the example, the left front channel is rendered by radiating the sound in a direction which is not directly towards the assumed listening position 113. Rather, the sound is predominantly directed in a sideways direction 115 that allows the sound to reach the listening position 113 indirectly, such as e.g. via reflections of walls. In addition, a notch is formed in the direction 117 towards the listening position to reduce the direct sound radiation to the listener.

This approach has been found to provide an enhanced audio perception and may in particular provide a perception of more enveloping and wide stereo signal. In particular, the spatial perception of the left front signal may be perceived to originate from the side rather than from the single speaker box 101.

The inventor has realized if such an approach is based on a low complexity two speaker dipole configuration, it tends to result in a suboptimal user experience in some situations. In particular, the inventor has realized that the perception of a potential blurred front image may be caused by such a low complexity approach resulting in sound being radiated in different and possibly undesired directions. In particular, a dipole radiation pattern has two lobes in opposite directions and this results in sound sources that are supposed to be localizable and well-defined (and which are usually mainly positioned in the front channels) possibly becoming somewhat diffuse or blurry in some cases. This is because the sound of, for example, the left-front channel, may actually be radiated to both the left- and right side wall, leading to a "spatial cross-talk" (and indeed with an inverted phase).

However, in the approach described in the following, the three driver speaker arrays are used to provide an improved spatial experience. In particular, the approach may be used to enable a wide sound stage to be produced from a compact device in front of the listener yet with a spatially more defined, and typically less blurry, image.

In particular, a three (or more) driver array is used in a configuration wherein two drivers are configured as an unsteered dipole pair (with the notch in the direction perpendicular to the driver line of the array, 0=0). This may be achieved by feeding the same signal to both drivers but with the phase of one of the drivers being inversed. Furthermore, an additional driver of the array is used to at least partly cancel the field of the dipole in a selected direction. Specifically, the additional driver is used to generate a notch which can be directed towards the listening position. Furthermore, as will be described, the notch direction can be modified by low complexity processing of the signal for the additional driver relative to the two other drivers. Indeed, in many scenarios, a simple gain variation may be used to modify the notch direction.

In more detail, Fig. 2 illustrates a driving arrangement for the speaker array 105 of Fig. 1. In the example, the driver arrangement drives three equally- spaced drivers 107, 109, 111 arranged in line (i.e. a three-speaker line array).

The arrangement of Fig. 2 comprises a receiver 201 which receives an input signal to be rendered by the system. In the example, the input signal is the left front channel of a surround sound multi-channel signal. The receiver 201 is coupled to a drive circuit 203 which is arranged to generate a drive signal for each of the three drivers 107, 109, 111.

Specifically, the drive circuit comprises a first signal processing path 209 which generates a first drive signal for the first driver 107, a second signal processing path 211 which generates a second drive signal for the second driver 109, and a third signal processing path 213 which generates a third drive signal for the third driver. Each of the signal processing paths 209, 211, 213 provides a transfer function for generating the respective drive signal from the input signal provided by the receiver 201. It will be appreciated that although the signal processing paths 209, 211, 213 in Fig. 2 are illustrated as separate and parallel signal processing paths, some processing may be shared or common for all paths. In particular, the signal paths 209, 211, 213 may comprise a shared/ common signal processing path segment as well as separate and individual signal processing path segments for each individual driver.

In the example, the drive circuit 203 is arranged to generate the drive signals for the first driver 107 and the third driver 111 of the speaker array 103 such that these are operated in a dipole configuration. In particular, a first and third drive signal is generated from the input signal such that the two drivers 107, 111 behave as a single dipole in the far field, i.e. at a distance where the drivers 107, 111 are sufficiently close to be considered a single point sound source. The dipole configuration is furthermore achieved by the drivers 107, 111 being sufficiently close for rendering the relevant frequencies as a dipole sound source. Typically, a distance between the centers of the drivers 107, 111 of no more than 50 cm, or in many embodiments advantageously no more than 25 cm, is sufficient to ensure that the two drivers 107, 111 can be operated in a dipole configuration for distances to the listening position of, say, more than lm. An example of a dipole radiation pattern is illustrated in Fig. 3.

In the example, the two drivers 107, 111 operated in a dipole configuration are the outside drivers.

The dipole configuration may specifically be achieved by the first and third drive signal being substantially identical but being phase inverted, i.e. by being substantially 180° out of phase relative to each other. This may for example simply be achieved by feeding the same signal to both drivers with the wires for the driver being reversed for one of the drivers relative to the other as illustrated in Fig. 4. Thus, in the example of Fig. 2, e.g. the third signal processing path 213 may comprise an inverter not included in the first signal processing path 209 (or vice versa). Indeed, in the example, the first and third drive signal for the first and third driver 107, 111 respectively may be the same signal except that one of the signals is phase inverted.

In the example, the drive circuit 203 generates a second drive signal for the second driver 109. In the example, the second driver 109 is located between the two drivers 107, 111 operating as a dipole. Due to the symmetry of the configuration, the resulting configuration can be seen as a coincident dipole-monopole pair with the first and third driver 107, 111 providing the dipole and the second driver 109 providing the monopole. In the example, the second drive signal is generated from the input signal but with at least one of a phase offset and a gain offset relative to the first drive signal. For example, the generation of the second drive signal for the second driver 109 may include a phase offset and/or a gain offset that is not included in the generation of the first drive signal (or the third drive signal). Thus, the second signal processing path 211 comprises a phase shift and/or gain modification function that is not included in the first and third signal processing paths 209, 213.

The second driver 109 is in this way used to modify the radiation pattern of the speaker array 103 from being that of a dipole. In particular, the relative offset applied when generating the second drive signal allows the pattern to be modified such that a notch is steered in a desired direction, and this is in the system used to direct this notch in the direction of the listening position. Thus, a notch (or specifically a null) may be generated in a different direction than the notch (or null) of the dipole configuration. This notch may result in the direct path from the speaker array 103 to the listening position 113 being attenuated such that the sound is predominantly perceived from indirect paths. Furthermore, due to the active use of the three speakers in this way, a more advantageous radiation pattern is achieved and in particular the radiation of sound in the opposite direction (e.g. towards the right in the specific example) is substantially reduced. Thus, the second driver 109 is used to control the radiation pattern such that a notch is provided in the desired direction while at the same time achieving a more desirable radiation pattern.

Furthermore, this modified approach can be achieved with a very low complexity as it merely requires the offsetting of one signal relative to the other. Thus, a much simpler approach than a conventional beam steering approach can be achieved. In particular, as will be demonstrated, it is in many embodiments possible to control the directional sound radiation, and specifically controlling the desired direction of the notch, simply by controlling a gain offset for the second drive signal.

In more detail, the approach uses two drivers out of an array comprising three equally- spaced drivers in a line to provide an unsteered dipole pair with a notch in the direction 0=0 i.e. perpendicular to the speaker line, by simply inverting the phase of one of the two drivers. The remaining driver is then used to (at least partly) cancel the field of the dipole in a selected direction.

Denoting the inter-driver distance Ax/2 (i.e. with the total length of the three- speaker array being Ax), and positioning the origin of the coordinate system at the centre of the second driver 109, the resulting sound pressure level from the three drivers can be determined. It is assumed that all three drivers can be considered to be monopoles.

First, the two drivers forming the dipole (the outer two drivers 107, 1 1 1) are considered to be driven by the same harmonic signal s(t)=d ^m but with the signal of the third driver 1 1 1 being inverted (i.e. a 180-degree phase shift compared to the first driver 107 equivalent to driving it with a signal e^ ⁰ *^).

In the far field, where the sound waves radiated by both speakers can be considered as plane waves, the total sound pressure due to the outer two drivers 107, 1 1 1 in an arbitrary direction a (defined relative to the line through second driver 109 perpendicular to the array) at radial frequency frequency in Hz) is given by:

Pi ₊₃ (ω,α ) = ^β (i)

where c is the speed of sound.

The second driver 109 may then generate a notch, and specifically a null, in a desired direction Θ by modifying this radiation pattern. In order to cancel the sound field in the desired null direction Θ, the second driver 109 consequently specifically has to generate a pressure /¾ = -pi+3 in this direction Θ, i.e. the second driver should generate the sound field:

/? ₂ (co,0 ) = 2 sin ω (2)

This equation thus provides a modification of the second drive signal relative to the first drive signal in order to provide a cancellation in the desired direction.

Fig. 5 illustrates an example of a drive circuit in accordance with such an approach. The input signal s(t) is fed to an equalization filter 501 which is part of a common path segment for the signal processing paths 209, 21 1, 213 for generating the drive signals from the input signal. Initially, this filter will be assumed to simply pass the input signal s(t) through to the three separate signaling processing path segments for generating the three different drive signals (i.e. the system will be considered without the equalization filter 501).

In the example, the first drive signal is generated as identical to the (equalized) input signal and the third drive signal is generated as the inversed (equalized) input signal. Thus, the first and third drivers 107, 111 are coupled in a dipole configuration and generate a sound field having a radiation pattern as indicated in Fig. 4.

The signal processing path for the second driver 109 comprises a filter 503 and a 90° phase shifter 505. This reflects equation 2 which essentially comprises a 90-degree phase shift operation and a frequency dependent gain offset.

Thus, in the system of Fig. 5, the transfer response of the signal processing path generating the second drive signal has a frequency- independent phase offset of 90° relative to the transfer response for generating the first drive signal (and consequently of - 90° relative to the transfer response for generating the third drive signal). Thus, the phase of the second drive signal is halfway between the phases of the first and third drive signals.

The 90° phase shifter 505 may e.g. be implemented as a Hilbert-transform which will be well known to the skilled person. Such a transform may e.g. be implemented by means of a filter. In a digital system, this is typically implemented as an FIR filter. The required length of the FIR filter is determined by the lowest frequency where the processing should still be effective.

It will also be appreciated that a relatively small variation from an exact 90° phase shift will merely result in a reduced attenuation in the desired notch direction and/or in the direction of the notch deviating a bit from the desired direction. It has been found that particularly advantageous performance is typically found if the phase difference between the transfer function for generating the first drive signal from the input signal and the transfer function for generating the second drive signal from the input signal is in the interval of [80°, 100°] or in some cases [85°,95°] for a frequency interval of 100 to 2 kHz.

In addition to the 90° phase shift, equation 2 illustrates that a gain offset should be applied to the second signal processing path 211 relative to the first signal processing path 209. As can be seen, this gain offset depends on the desired direction of the null, or in other words the direction of the null can simply be determined by controlling the gain offset between the different signal processing paths.

Also, as illustrated by equation 2, the gain offset is preferably frequency dependent and is therefore in the example of Fig. 5 implemented as a filter 503 rather than as a simple gain. The filter has the transfer function H? which in the specific example is given by: Thus, in the example, the offset between the first and second signal processing paths 209, 21 1 (and between the transfer function for generating the first drive signal and the second drive signal) thus comprises a substantially frequency independent (constant) phase offset and a frequency dependent (varying) gain offset.

It will be appreciated that the phase shift 505 and the filter 503 can be implemented in the same filter.

The total sound pressure in an arbitrary angle a generated by the system of Fig. 5 can thus be determined as the sum of the individual contributions, i.e. as:

(4)

As can be seen, a null is generated in the desired direction Θ.

However, equation 4 also indicates that a frequency dependent sound pressure level is generated, i.e. that the reproduced frequency response of the system as a whole is not flat but may introduce a coloration of the sound. The frequency response is furthermore dependent on the angle a and thus on the position of the listener.

In some embodiments, the drive circuit 203 is arranged to provide a frequency dependent transfer function for generating the first drive signal from the input signal. Thus, in some embodiments, the generation of the first signal from the input signal introduces a frequency shaping. The same approach may also be applied for the third drive signal.

The frequency response of the transfer function is specifically arranged such that it provides a compensation for the frequency variation/ shaping that result from a frequency variation of the offset of the second drive signal and the frequency dependent in- air combining of sound pressure levels from the first driver 107, the second driver 109 and the third driver 1 1 1 in a first direction. Thus, the frequency shaping may be used to compensate for the frequency shaping that is introduced by the described approach.

In the example of Fig. 5, this compensation is simply achieved by a frequency shaping of the input signal. In particular, the equalization filter 501 is located in a shared or common segment of the signal processing paths 209, 21 1 , 213 such that the input signal is pre-compensated before being applied to the individual signal processing paths. Thus, the shared signal processing path segment comprises an equalization filter 501 that has a frequency dependent gain. This frequency dependent gain is in this example thus common for all drivers 107, 109, 1 1 1 whereas the frequency dependent gain of the separate segments of the signal processing paths are different for the first and second driver 107, 109 (as well as for the second and third driver 109, 1 1 1) by virtue of filter 503.

Thus, in the example of Fig. 5, a frequency varying gain filter in the form of the equalization filter 501 is introduced to pre- filter the input signal so as to compensate for the frequency shaping by the rest of the rendering system.

Specifically, for a reference angle <¾ in which a flat frequency response is desired, the equalization filter 501 may be given by:

As will be described in the following, the desired processing of the rendering system can be achieved precisely or approximately in many different ways.

The signal operations of the exemplary embodiments mainly seek to accurately or approximately introduce:

A 90-degree phase shift for the second driver 109 relative to the first driver

107.

A frequency dependent gain offset for the second driver 109 relative to the first driver 107.

A frequency dependent equalization of the overall signal to be reproduced to compensate for a frequency shaping by the sound rendering.

In the example of Fig. 5 this is achieved by use of the filters H ₂ and H _E Q as defined by equations 3 and 5 as well as by use of a Hilbert-transform FIR filter 505.

It is seen from equations 3 and 5 that both the overall equalization filter HEQ and the centre speaker filter H? depend on both the frequency and the null angle Θ in a non- independent way, since both variables are in the argument of the same sine function.

Consequently, the shape of both filters depend on the null angle. In practice, instead of calculating filter coefficients in real-time using equations 3 and 5, a set of pre-calculated filters for different values of the null angle Θ may be stored in a memory.

Several approximations to the equalization filters of this first embodiment are possible and may enable some convenient simplifications to the more general processing scheme, as will be shown in the following embodiments. For example, in some embodiments, the first and third signal processing paths 209, 213 may advantageously be arranged to have a common frequency gain variation whereas the frequency gain of the second signal processing path 211 is kept substantially constant, e.g. within 2dB in the frequency range from 200 Hz to 3 kHz. Thus, in some embodiments, the gain offset may be introduced by processing in the first and third signal processing paths 209, 213 rather than in the second signal processing path 211.

As an example of such a system, it is first worth noting from equations 3 and 5 that if the reference angle <¾ is chosen to be 0 (so that a flat frequency response is achieved in the on-axis direction, perpendicular to the array), then the overall system equalization filter H _EQ reduces to the inverse of the frequency-dependent gain factor H ₂ for the second driver.

In this case, instead of applying the overall equalization filter of equation 5 to the input signal of the system and applying the frequency-dependent gain filter H _? of equation 3 to the signal fed to the second driver 109, the same effect can be achieved by only applying the equalization filter of equation 5 to the signals fed to the outer two drivers speakers 107, 111. Furthermore, in this case, the generation of the second drive signal need not include any frequency-dependent gain as the equalization filter and the frequency- dependent gain factor H cancel each other.

These considerations lead to the modified and often simplified approach illustrated in Fig. 6. Further Fig. 7 illustrates examples of the shape of the equalization filter H _EQ for the first and second signal processing paths 209, 213 for various values of the null angle Θ and for an inter- speaker spacing of 6 cm (total length of configuration^ 2 cm) and cco=0,.

An inconvenience of the required overall equalization filter H _EQ of equation 5 and the second driver gain factor H _? of equation 3 is that they both depend on the frequency and the null angle Θ in a non-independent way, since both variables are in the argument of the same sine function. This means that the shape of both filters depends on the null angle.

It is often more preferable, e.g. in use cases where the notch direction may be dynamically modified (e.g. to track a user's position), to have fixed equalization filters that are independent of Θ for achieving a flat overall system response, and to have a frequency- independent gain for the second driver that only depends on Θ for controlling the notch angle.

Such an implementation is indeed at least approximately possible. In particular, for sufficiently low frequencies and/or a sufficiently small notch angle, the approximation of sin x ~ x can be used to simplify equations 3 and 5 :

a _OTro* («,e ) (6)

and

(V)

Using such an approximation, H _EQ and H ₂ of Fig. 5 can be replaced by a fixed equalization filter in the path of the first and third drivers 107, 111 with a response that is substantially proportional to l/co (i.e. has -6 dB/octave frequency slope). In addition, the gain for the second driver 109 may be set to g ₂ =sin0 to control the null angle Θ. In addition, the input signal may be scaled by a gain of gi=l/(sin0-sina _o ) which to ensure unity gain of the system.

An example of such a system is shown in Fig. 8.

Thus, in such embodiments, the common filter of the first and third signal processing paths 209, 213 (but not of the second signal processing path 211) is thus independent of the direction of the notch whereas the frequency independent gain of the second signal processing path 211 is used to control the direction of the notch. Thus a very low complexity approach for controlling the notch direction is achieved. The common filter is specifically a first order low pass filter with a slope of 6 dB per octave.

Another example of a system having such characteristics can e.g. be achieved by setting <¾=0, i.e. by specifying a flat response in the on-axis direction. In this case, the gain for second driver 109 is the inverse of the overall input gain, so both can be replaced by a gain g ₃ =l/sin0 in the path of first and third driver 107, 111. An example of such a system is shown in Fig. 9.

In the above examples, the phase offset between the second driver 109 and the first and third drivers 107, 111 is introduced by a phase shift of the second signal processing path 211. However, in some embodiments the phase offset may be introduced by a phase shift in the first signal processing path 209 or the third signal processing path 213.

Figs. 10- 13 illustrates examples of such implementations corresponding to the approaches of Figs. 5, 6, 8 and 9 respectively. An advantage of several of these approaches is that the second signal processing path 211 does not need to actually include any processing (or only a simple gain such as in the system of Fig. 12). Furthermore, by combining the phase shift operation with the other filtering in the appropriate path, an even lower complexity can be achieved.

In the previous examples, it was assumed that the configuration of the drivers 107, 109, 1 1 1 is completely symmetrical and linear, i.e. that the three drivers 107, 109, 1 11 are on the same line, with the second driver 109 being exactly in the centre of the

configuration.

In practice, variations from this assumption may slightly degrade the performance but this performance may be acceptable in many situations. However, in some embodiments, the drive circuit may further be arranged to introduce a time offset for the second drive signal relative to at least one of the first and third drive signal where the time offset is determined in response to a difference between a distance between the first driver 107 and the second driver 109 and a distance between the third driver 1 1 1 and the second driver 109.

For example, if the second driver 109 is not exactly at the center (i.e. at a distance of Ax/2 from both the first and third driver 107, 11 1), but is shifted a distance x ₂ along the line connecting the drivers (x ₂ is negative for a shift towards the first driver 107 and positive towards the third driver 1 1 1).

In this case, equations 1 and 2 are still valid. However, since the position of the second driver 109 has shifted, a frequency- independent delay (At) 2 is added to the second drive signal in order to compensate for this. This delay may specifically have the value: Δ _ί ) ¾ sina _ ₍₈₎

c Note that this time offset (delay or advance dependent on the direction of the displacement and the notch angle a) can e.g. be incorporated in the Hilbert transform filter for the second driver 109 in the system of Fig. 5 and thus an additional delay element is typically not required.

The same concept also applies to configurations in which the drivers 107, 109, 1 1 1 are not positioned on the same line.

The described approach may in particular provide a wide soundstage while maintaining spatially well defined sound source positions. However, in some scenarios and for some sound components the enveloping, diffuse sensation that can be provided by a simple dipole arrangement is actually appreciated by consumers. Therefore, in some embodiments the described approach is used selectively. For example, in some embodiments, the described approach may be used to render the front side channels of a spatial surround sound signal whereas other rendering approaches are used for the surround channels. Thus, the approach may be used to process channels that are intended to be perceived spatially wide yet clearly defined and localized, whereas channels that are advantageously perceived as diffuse and ambient sounds are processed by a rendering approach that provides a more diffuse experience, such as by a simple two driver dipole configuration.

In some embodiments, such a selective approach may be used for differentiated rendering of different parts of a received signal. In particular, in some embodiments, the drive circuit comprises means for dividing the received signal into an ambient sound signal and a non-ambient signal. This approach may for example be applied to a received stereo signal which can be divided into an ambient stereo signal and a non- ambient stereo signal where the ambient signal may tend to comprise background sounds whereas the non-ambient stereo signal may comprise one or more specific dominant sound sources. The received signal is thus divided into signal components that meet an ambiance criterion and signal components that do not meet the ambiance criterion. The criterion may thus provide a suitable test for determining whether the signal components can be considered ambient and thus should be rendered more diffusely and less spatially well defined, or whether the signal components cannot be considered ambient and thus should be rendered spatially more well defined.

Thus, the input signal may be analyzed and separated into localized or spatially defined signal components, and diffuse or ambience components. It will be appreciated that any suitable algorithm for separating a signal into such components may be used and that such algorithms will be known to the skilled person. An example of suitable algorithms may e.g. be found in Harma, Aki; Faller, Christof, "Spatial Decomposition of Time- frequency Regions: Subbands or Sinusoids", AES Convention: 116 (May 2004) Paper Number :6061 ; R. IRWAN AND RONALD M. AARTS, "Two-to-Five Channel Sound Processing", J. Audio Eng. Soc, Vol. 50, No. 11, 2002 November or in US patent publication US20090092258 Al .

The two signal components are then rendered using different approaches. Specifically, the non-ambient signal components are rendered using the described approach whereas the ambient signal components are rendered using another approach, such as e.g. using a simple fixed dipole arrangement. It will be appreciated that at least some of the described embodiments may provide advantages including for example singly or in combination:

Providing a sound rendering with a significantly smaller back lobe resulting in a sound image that is still as wide, but less blurry.

Increased power efficiency. The power efficiency at low frequencies may be twice as high for the same inter-speaker distance as the total distance of the array is twice as high. This means less powerful amplifiers are needed for the same bass output, or higher bass output can be achieved with the same loudspeakers and power resulting in a more full sound. This also means that the cut-off of an additional subwoofer can be lowered, and its power requirements (and hence size) reduced.

The robustness of the system to small variations in system parameters, e.g. sensitivity- or frequency response differences between individual drivers, is better. In particular, the presence of a clear notch in the desired direction is more robust.

The fact that the notch angle is controlled by a simple adjustment of gain balance makes implementation of dynamic use cases in which the notch angle is changed dynamically, e.g. to track listener position, easier to implement (e.g. there is no need for delay interpolation).

Furthermore, a more precise control over the notch angle can be achieved.

It will be appreciated that the above description for clarity has described embodiments of the invention with reference to different functional circuits, units and processors. However, it will be apparent that any suitable distribution of functionality between different functional circuits, units or processors may be used without detracting from the invention. For example, functionality illustrated to be performed by separate processors or controllers may be performed by the same processor or controllers. Hence, references to specific functional units or circuits are only to be seen as references to suitable means for providing the described functionality rather than indicative of a strict logical or physical structure or organization.

The invention can be implemented in any suitable form including hardware, software, firmware or any combination of these. The invention may optionally be

implemented at least partly as computer software running on one or more data processors and/or digital signal processors. The elements and components of an embodiment of the invention may be physically, functionally and logically implemented in any suitable way. Indeed the functionality may be implemented in a single unit, in a plurality of units or as part of other functional units. As such, the invention may be implemented in a single unit or may be physically and functionally distributed between different units, circuits and processors.

Although the present invention has been described in connection with some embodiments, it is not intended to be limited to the specific form set forth herein. Rather, the scope of the present invention is limited only by the accompanying claims. Additionally, although a feature may appear to be described in connection with particular embodiments, one skilled in the art would recognize that various features of the described embodiments may be combined in accordance with the invention. In the claims, the term comprising does not exclude the presence of other elements or steps.

Furthermore, although individually listed, a plurality of means, elements, circuits or method steps may be implemented by e.g. a single circuit, unit or processor. Additionally, although individual features may be included in different claims, these may possibly be advantageously combined, and the inclusion in different claims does not imply that a combination of features is not feasible and/or advantageous. Also the inclusion of a feature in one category of claims does not imply a limitation to this category but rather indicates that the feature is equally applicable to other claim categories as appropriate.

Furthermore, the order of features in the claims do not imply any specific order in which the features must be worked and in particular the order of individual steps in a method claim does not imply that the steps must be performed in this order. Rather, the steps may be performed in any suitable order. In addition, singular references do not exclude a plurality. Thus references to "a", "an", "first", "second" etc do not preclude a plurality. Reference signs in the claims are provided merely as a clarifying example shall not be construed as limiting the scope of the claims in any way.