High beam

When the first steered loudspeaker systems turned up at our trade shows, they brought smiles of bemusement to the faces of many onlookers. As recently as 15 years ago, the idea of directing sound of a loudspeaker around a room ‘like a searchlight’ seemed a little far-fetched to the majority of the pro audio community. But like the line arrays that had revolutionised the live sound industry before them, beam steering systems have won over many sceptics in turn.

The purpose of ‘steering’ technology is to focus and direct a beam of sound such that it can be given a required vertical direction. This technology has found ready favour in highly reverberant and architecturally difficult spaces such as railway stations and shopping malls, as well as churches and mosques – for obvious reasons.

At the heart of the technology is control of the relative phase relationships between the sound generated by a relatively large number of individual loudspeaker drivers. The phase and amplitude of sound from each the drivers can be used to determine how the drivers sum (constructively or destructively interfere) to produce a ‘beam’ of sound that can be controlled in the vertical plane. Although there are passive systems – such as those from JBL and Community Professional – these normally combine DSP and amplification within the loudspeaker to simplify set-up and control.

Although the physics behind beam steering can be compared to radar and sonar, steering loudspeakers is more difficult due to the great range of human hearing. Assuming that we require a 20Hz to 20kHz frequency range, the wavelength of a sine wave at 20Hz is approximately 50 feet (15.25m) while the wavelength of 20kHz sine wave is half an inch (13mm). And in order for beam steering to work, the sound radiated from one loudspeaker needs to be influenced by the sound from one or more other loudspeakers in the array. Combined with the requirements of loudspeaker spacing, it is this range that makes life difficult for the designers of steered beam arrays. Generally, it is necessary for the size of the array to be larger than the operating wavelength for beam steering to work. To steer low frequencies with a single size of transducer, very many drivers are theoretically required making their spacing impractical. A further problem sets in above around 2kHz, when effective beam control is compromised by a problem called lobbing, or spatial aliasing, where strong side beams occur. These can be stronger than the main beam, particularly when steering at extreme angles, and are one of the problems highlighted by beam steering’s critics, but used as an asset by its advocates. Additionally, the extent to which beams can be steered restricts the height at which a loudspeaker can be practically installed.

Real world systems

Commercially available steered systems work on a common principle – a set of loudspeaker drivers assembled in an array powered by its own DSP-controlled internal amplifier that selectively delays individual drivers in the array to achieve the required steering and focusing. In most cases, the software control program and onboard DSP also offer comprehensive signal compression and equalisation capabilities.

It was the falling cost of DSP that brought beam steering into the reach of professional audio (having been in military use for some considerable time). The first system was unveiled by Duran Audio in 1996 – the Axys Intellivox array, the successor to the Octarray. The standard Intellivox DC and DS models are available in five lengths – the longer arrays allowing greater control of directivity. Each model uses an array of 4-inch or 6.5-inch drivers of varying length (up to 4.3m using 32 4-inch drivers and offering a frequency response of 130Hz to 10kHz), while the DC/DS115 has two additional tweeters to extend its frequency range (130Hz to 20kHz). With a more accurate high-end response (also 130Hz to 18kHz) the Intellivox DSX is available in four lengths up to 5m, using 28 4-inch drivers and eight tweeters.

The DC and DS series each draw on technologies developed by Duran Audio to allow electronic control of the directivity of loudspeaker arrays – Digital Directivity Control (DDC) and Digital Directivity Synthesis (DDS). DDC is a multichannel loudspeaker array technology providing each loudspeaker channel with its own dedicated audio path through onboard DSP and amplification. Each loudspeaker, or group of loudspeakers, is given a unique set of filters, which electronic manipulation of the vertical dispersion of an array for venues with a flat listening plane. Digital Directivity Synthesis takes this a stage further, and anticipates installations with a more complex listening plane, such as with a raked seating area or split level, DDS is the answer. DDS claims to be able to synthesise any desired 3D radiation pattern from the loudspeaker array within the physical constraints of a pre-defined array (transducer distance, array length and so on) to address the specific configuration of a venue.

‘A lot of people still don’t realise that we have already moved on from beam steering into beam shaping with our Intellivox DDS range,’ international business development manager Nick Screen asserts. ‘This allows us to control both the near-field and far-field directivity of an array, and create a complex dispersion pattern – in short, you can aim the sound where you want it and try to stay away from reflective surfaces in even buildings with complex geometries. OK, it sounds like rocket science until you actually go through the process and then you realise how easy it is to achieve – you simply build a model of the room (in our DDA software) and define your “ideal” result. The algorithm then does all the hard work and, after a few moments, shows you graphically what can be achieved with your chosen array in your chosen position. Once you’re happy, it creates the settings file with the necessary filters for the array. The whole aim of this is to give system designers and engineers a set of tools that allow them to take control and get the most from their arrays. It’s important to realise that this is very different from having a beam steered array with multiple lobes, those lobes will overlap and where they do there will, of course, be undesirable effects.

‘The thing is about our beam shaping technology (DDS) is that the maths behind it can be applied to any array, of any shape, with any combination of drivers,’ he continues. ‘This is why we have also used this technology in our Target touring line array. So we are still a long way ahead of the market, perhaps this is something that’s difficult to understand from a spec sheet. Of course, there are now many manufacturers trying to play catch up with their beam steering products, but we have already been beam shaping for most of the 00s and beam steering for more than 15 years.

‘Moving forward, we plan to expand the portfolio of products,’ he closes. ‘These won’t just be “me too products” or product launches for the sake of product launches they will stay true to our philosophy and core values of creating products that use the benefits of our beam shaping technology to solve real-life problems, both in terms of speech and music reproduction. Our latest development is the Beam Shaping Subwoofers launched at the Pro Light & Sound show. These highly innovative subwoofers will allow engineers to gain control over low frequencies by shaping the directivity of their arrays, the result is low end with definition and punch and banish that flabby undefined bass sound.’

The DSA series of digitally steerable array loudspeakers from EAW offers similar onboard control, using an extension of the Phased Point Source Technology (PPST) developed for EAW’s KF900 line array. The use of phase and frequency ‘shading’ techniques allowed KF900 loudspeaker modules to be operated as a single acoustical element with control over the beam profile, enabling adjustment and steering in the vertical plane.

'The beam forming and beam steering technology first developed for the KF900 series as part of PPST (Phased PointSource Technology) was more than a simple adjust-and-aim process,' says Nathan Butler, EAW principal engineer. 'The complex characteristics of the venue are an important part of the equation and, of course, can't be ignored. This approach was carried into the DSA series and can be seen in the DSAPilot software. In addition to the basic 'specify coverage pattern' method, the user is allowed to match the loudspeaker coverage to the listening area. With this, we have only begun to scratch the surface of possibilities that signal processing offers with regard to steering. With faster computers and more advanced modelling capabilities, much more is, and will be, possible.'

EAW recognised that balancing enclosure size against driver placement represented the first substantial obstacle – while longer lines handle lower frequencies, many applications demanded small enclosures, prompting the creation of two different enclosures – a primary full-range system (DSA250i, with eight 4-inch drivers and eight 1-inch drivers) with ‘meaningful pattern control down to 300Hz’ and a secondary LF-only system that extends this pattern control when needed (the DSA230i, with eight 4-inch drivers). EAW claims ‘radical steering angles and even intentional nulls’ can be achieved, while band limiting the signal to around 8kHz allows for ‘radical shaping with smooth frequency response across a large area.’

Again, high-resolution, design-optimised DSP control is the key to better high-frequency control than could previously be achieved, with the actual performance exceeding the initial design expectations with steering up to 16kHz. Each DSA250i has a fixed 120° horizontal coverage pattern, while properly configured DSP can vary the vertical beam width anywhere from 15° to 120°. EAW claims that the beam can be steered ±30° while still maintaining even frequency response up to 16kHz. As with Duran Audio’s systems, the control software (DSAPilot) accommodates stacking of multiple units of the same model or combinations of the two to achieve more power and greater beam control.

For Renkus-Heinz, it is the Iconyx line, comprising four models using from eight to 32 drivers. As with the EAW two-way system, the Iconyx IC8-R incorporates eight 4-inch and eight 1-inch transducers, with the IC16-R, IC24-R and IC32-R extending size and performance – although, in this case, the two drivers are coaxial units. Multiple sonic beams can be individually shaped and aimed from a single Iconyx array using the system’s software controlled DSP – up to four beams on an IC8-R – and the acoustic centre of the array can be raised or lowered electronically.

‘When we previewed the Iconyx loudspeaker at InfoComm in June 2005, it was billed as the world’s first digitally steerable loudspeaker specifically intended for music as well as speech,’ says senior VP Ralph Heinz, ‘The original Iconyx project was spurred by pent-up demand, particularly from houses of worship and architects in the US and elsewhere, for an affordable product that would simultaneously solve the problem of delivering intelligibility in hugely reverberant spaces like cathedrals and museums, and also meet the aesthetic desire for an minimalist loudspeaker installation.

‘The world had tried simple column loudspeakers, multi-speaker distributed systems, line arrays, yet many influential people – congregations, worship leaders, operators of beautiful public spaces and architects – were still searching for a solution that was both a genuinely innovative acoustic problem-solver and yet architecturally almost unnoticeable. We set to work on this challenge with a combination of unique cabinet design and custom-designed software, DSP and amplification, as well as an intuitive external PC-based software system, BeamWare, to enable fast and accurate system tuning and commissioning. The result was Iconyx: they were unobtrusive and the clarity came as a revelation to most who heard them.

‘Another crucial factor was that we also intended, from the start, that the product would be as good for music reproduction as it would for speech intelligibility. This has been a key factor in its success.’

The heart of each control module is the eight-channel DSP processor/amplifier developed specifically for the Iconyx line. This performs the digital signal processing needed to shape and aim the beams, and also claims to avoid creating the side lobes that trouble steered beams. Broader system control is offered by Rhaon (Renkus-Heinz Audio Operations Network), which manages multichannel audio distribution, user-controlled DSP, user selectable presets and remote system management and control over Cat5 cable using standard Ethernet hardware.

‘The ability to precisely steer beams of sound from a fixed, slender cabinet created huge interest in Iconyx among key consultants and integrators,’ Mr Heinz reports. ‘That momentum led to consultants and customers asking for further developments for larger applications and for A/V rental firms.’ The result was the high-power, portable IC Live range – as specified for President Obama’s Nobel Peace Prize acceptance speech in December – and a new installation-specific version of it. ‘Expect more exciting additions to the range this year and beyond,’ Mr Heinz adds.

While Tannoy readily agrees that its Qflex line isn’t the first commercial application of beam steering, it does claim that the six models in the range represent a ‘notable leap forward’ in the technology, with ‘far greater capability than any previously available beam steering array product’. ‘The first thing you have to get right is the physics, before you get into fancy steering algorithms,’ begins Tannoy’s Graham Hendry. ‘The voice intelligibility side had been well covered by the people who were doing it before us, but we wanted to make it better suited to full-frequency applications.’

That the system is a departure from Tannoy’s long-standing use of dual-concentric drivers is critical in understanding beam steering. ‘Effective steering and beam control requires densely spaced transducers,’ Mr Hendry continues. ‘Tannoy’s hallmark Dual Concentric drivers are not suitable because they prohibit tweeter spacing wider than a woofer diameter. This is not surprising since they are in themselves passive, axis-symmetric “arrays”. The utilisation of coax drivers or vertically arrayed full-range drivers is therefore unsuitable for effective full-range beam steering applications. It’s a simple rule of physics which no amount of corrective DSP can compensate for.’

Certainly, Qflex pushes effective full-range beam control – up to 12kHz, rather than concentrating on the vocal band – with the capability to steer multiple beams at previously angles of ±70°, down to 1° resolution. Unsurprisingly, this is achieved through a proprietary high-frequency array and steering algorithm.

Drawing its name from the ‘the Q factor’ term used to indicate directivity, Qflex uses densely spaced transducers under the control of integrated DSP and the Regularized Least-Squares Multichannel Inversion software algorithm. Qflex 16 features 16 channels of amplification and DSP, and eight 4-inch low/mid and eight 1-inch high-frequency drivers (with the 24, 32, 40 and 48 models following suit). Qflex 8 uses eight 4-inch low-frequency drivers, and is intended as a low-frequency extension for larger array configurations or as a standalone unit for frequencies up to 4kHz. Configuration of filtering and delay for the desired directivity is through BeamEngine software, while Tannoy’s VNet software allows speakers to be configured in a larger system.

‘Contrary to what the others do, we apply a single shaped beam to the required coverage area,’ says Mr Hendry. ‘This beam can be asymmetric in shape; we can also specify quiet areas between a beam. From the defined listening area input by the user, they specify target SPLs over the coverage area. Multiple audience areas can be specified if required. The steering algorithm generated within the beam engine is then saved and loaded to the DSP via the VNet software. This sequence of steps can be carried out in a matter of seconds.’

The MessengerG2 is described by manufacturer Ateïs as a ‘next generation’ design, focused on improved lobe control and overall performance. The five models in the Messenger line-up use between 12 and 48 drivers to offer 94dB SPL at between 25m and 70m. High-power audio DSP provides 24 channels of processing that support precise beam control and even signal pattern that stays within 3dB over a distance of 100m. Each Messenger array delivers a very tightly controlled beam that can be shaped as required for specific applications or environments requiring symmetrical or asymmetrical, single-, dual- or triple-lobe designs, while a variable acoustical centre can compensate for architectural requirements.

The input section is designed to ‘broadcast’ standards and has two fully controllable audio inputs and outputs with override functions and hardware bypass, while remote control by LAN or WAN is possible via a variety of networking and control cards. Additional input-cards can be provided to offer a link between Messenger columns or to interface with an Ateïs network based on M-Net or Ateïs-Net, and can be controlled using Ateïs Studio system designer and controller software. The external interface also accepts the Ateïs Studio remote control software.

The latest beam steering system to appear is Meyer Sound’s CAL (column array loudspeaker), which is available in three versions of column lengths ranging from less than 1m up to 1.8m, and offers a frequency range of 150Hz to 10kHz. The design is based on sound directivity research conducted in collaboration with the University of California Berkeley’s Center for New Music and Audio Technologies, and uses the same digital signal processing as the Meyer Sound D-Mitri digital audio platform.

‘Early this year, we introduced CAL column arrays that deploy beam steering technology that is based on our ongoing collaboration with the University of California’s CNMAT group,’ John McMahon, executive director of digital products, confirms. ‘Using principles of wavefield synthesis, we’ve been able to maximise the density of drivers in a given loudspeaker and, as a result, accomplish a level of precision and control that the industry hasn’t seen before. In addition to packing in the maximum number of drivers – you can’t get them any closer – each driver has its own amplification and processing.’

Drivers and tweeters are powered by dedicated amplifier channels to accurately control vertical behaviour, and the beam can be angled up or down 30° and configured with vertical beam widths from 5° to 60°. Multiple or split beams can be used as needed. A range of control presets offer alternative angle and dispersion options to readily project sound where needed.

The passive approach

While the above systems package amplification and control together, there are alternatives such as those offered by Community Professional and JBL.

Community reckons its Entasys to be ‘the world’s first fully passive three-way true line source column loudspeaker, producing zero vertical lobing up to 16kHz, and user-selectable 6° or 12° vertical coverage pattern configurations. Entasys allows the user to control the low-frequency vertical coverage pattern without affecting the beam width of mid and high frequencies. Passive Entasys Low Frequency Extension Columns can be added to any Entasys array to seamlessly narrow the vertical pattern between 200Hz to1kHz, matching mid and high frequencies’ vertical beam width. Entasys provides consistent horizontal and vertical coverage from 800Hz to 16kHz from a single loudspeaker.

‘Entasys is quite a bit different, in that it is mechanically steered,’ Community’s John Wiggins explains ‘For example, in a house of worship with a high ceiling and a balcony at the rear of a hall with a mid-band RT of 2s, increasing to 3.5s to 4.0s seconds below 500Hz, we used a split-beam array with two sets of two full-range modules, separated by a low-frequency module in the middle of each array. Using the two vertical lobes and careful aiming, we created floor and balcony coverage with five Entasys columns per side. This is the kind of performance that people do not expect from an Entasys column arrangement, but it is proof that it can be done, and done very well.’

The audience areas on the hall floor showed fairly uniform direct SPL, ranging from 107dB in the front corners near the arrays to 102dB at the back of the room underneath the balcony. The SPL in the balcony area is 101dB to102dB, while the potential intelligibility shows C50 at -3 dB (±1dB) and the STI is approximately 0.52.

‘We have – possibly – the only column line array that will exhibit good phase linearity and frequency response from 200Hz to well beyond 16kHz,’ Mr Wiggins adds. ‘Many complex and expensive steerable units begin to lose coherency above 2kHz, limiting their musicality.’

The JBL CBT Series, meanwhile, uses the company’s Constant Beamwidth Technology (CBT) beam forming to deliver constant directivity coverage from a passive column. CBT claims many of the benefits of DSP-controlled powered column speakers but without active steering. CBT does, however, offer greater rejection of off-axis sound, which lowers reverberation and improves speech intelligibility and music quality.

www.duran-audio.com

www.eaw.com

www.renkus-heinz.com

www.tannoy.com

www.ateis-international.com

www.meyersound.com

www.communitypro.com

www.jblpro.com

Published in PAA May-June 2010