It’s safe to say that line arrays have been at the forefront of large format sound reinforcement for the past 20 years or so and in some ways still feels new despite its longevity. This, in part, has to do with the refinements and improved designs that have emerged throughout its evolution, but as we take a look back from our vantage point half way through the second decade of the 21st century the question has to be asked: is the line array a tired technology?

Some theory

A line array is defined as a loudspeaker system that is comprised of a number of identical elements, or enclosures, mounted vertically and fed in-phase in an attempt to create a seamless isophasic/cylindrical wave front. In comparison, a traditional point source creates a spherical wave front. The caveat here is that while these definitions all sound great and work in theory neither exists in the physical world because for a line source to exist it must be infinitely narrow and long and for a point source to exist it must be infinitely small. Regardless, clever engineering and a thorough understanding of the limits of the physical world has enabled system designers and manufacturers to produce systems that come very close to acting like line sources through ingenious workarounds.

This, of course, has historically been mostly a function of physical loudspeaker design, with speaker configurations, waveguides and enclosures exhibiting increasing sophistication to combat the limitations of the line source array in the physical world. But what are the limitations?

The main limitation of a line source array is, as stated, that there is no such thing as a true line source. Hence we have the designation “near-line source’.

The way a line array is designed to work is through constructive and destructive interference between the array elements. To illustrate this point, a single conical driver in an enclosure exhibits varying directivity with frequency. Low frequencies are omni-directional and as frequency increases it narrows into a beam that is, for most applications, too narrow to be practically useful. The remedy is to design an enclosure that employs crossovers and two or more elements, possibly with specifically designed waveguides or horns, to achieve more constant coverage across the spectrum.

Now, if you had to stack two identical speakers and feed them an identical signal in phase, a completely different directivity pattern is exhibited. On-axis to this simple array there will be constructive interference where the sound pressure will increase by 6dB. Off-axis, cancellations are measured due to path length differences relative to each driver which results in a lowered sound pressure level (SPL) and in some cases full cancellation. This is often referred to as “combing’.

So we see that maintaining constant directivity across the spectrum is the challenge and, of course, incredibly difficult to achieve. The main goal in constructing a line array is to get all drivers working in unison instead of against each other and, since frequency is the key variable here, the theory goes that all drivers must be closely coupled with the distance between them being less than half the wavelength of the highest frequency you want to reproduce. Of course, with low frequencies this is less of a problem. Most low frequency drivers are arranged well within the “half the wavelength’ limit. Mid and high frequency drivers present more of a challenge where the physical dimensions of the drivers exceed their desired wavelength. This results in individual point sources as opposed to a seamless cylindrical wave front.
To illustrate, if we want to reproduce a frequency of 20 000Hz and we take the formula that wavelength equals the speed of sound divided by frequency, we see that 20 000Hz should be reproduced by a driver no larger than 8.6mm. This is obviously not possible.

How does a traditional line array deal with these problems? What is the future set to bring us now with the advent of inexpensive and widespread DSP? As we will learn, the limits of the physical world are slowly being eroded and researchers are starting to move the line source array into new territory; accomplishing things that were never thought possible.

3dB drop-off per doubling of distance?

Classical line array mathematics proposes a drop-off of 3dB with doubling of distance due to its proposed cylindrical wave-front, which, by way of expansion, doubles in surface area by doubling of distance. By comparison, a point source will, as stated, radiate in a sphere and will, with doubling of distance, expand to four times the surface area which in turn results in 6dB of a loss in level. This is commonly known as the inverse square law. This gives a line array a clear advantage in long throw applications.

However, the 3dB drop in sound transmission per doubling of distance has been shown to only work in the near field. Loudspeaker manufacturer Meyer Sound has concluded in their research that in the physical world – and despite the marketing hype – this effect is true only of a line array with 16 cabinets containing 15-inch drivers at around 350Hz between two and four metres from the array. Further than that, however, and the wave front starts to become spherical again once again loses 6dB per doubling of distance. This is known as the critical distance and varies with distance and frequency.

Therefore, in practice, line array theory is best demonstrated at low frequencies. As the frequency increases, more progressively smaller drivers placed closer together are required to maintain directivity and this is why many line arrays implement a mid-band where eight-inch drivers are crossed over to handle the mid-range. This means that in the physical, practical world line arrays only act as near-line source systems in the low and mid frequencies. It also stands to reason that for the high frequencies other methods must be employed to match the directional characteristics of the low and mid sections. Of course, the most common of these methods is the high frequency wave guide coupled to one or more compression drivers or by implementing ribbon drivers, such as we see in Alcons Audio and SLS loudspeakers systems.

The high frequency wave guide

Instead of using the classical line array model of using constructive and destructive interference (remember, the half wavelength distance is too small); wave guides produce a directional wave front by reflecting sound, by way of the guide’s design, into a specific pattern. The aim should be to closely match the very narrow vertical and very wide horizontal coverage of the low frequency characteristics of the array. By using correct crossover and equalisation, phase plugs and other clever engineering tricks, the high frequency beam and the constructive interference of the low frequencies can be made to phase align, resulting in consistent coverage.

There are many designs for these wave guides or “horns’, as they are colloquially called, and every manufacturer has their own. In fact, some would say that this aspect of line array design is what gives the line array its “sound’. The ribbon driver has been proposed as the perfect line array driver, particularly for short wavelengths, because it is very narrow, extends the entire vertical length of the enclosure, and is spaced very close to adjacent drivers (a natural line source).
Compression drivers have been touted by some manufacturers such as Alcons Audio as insufficient for line array use because they exhibit distortion above 8kHz and are slow. They also surmise that the waveguides needed for compression drivers introduce distortion as well. The combined effect of the “break-up’ effect of the compression driver at 8kHz and the distortion introduced by the wave guide supposedly results in a bad signal to noise ratio.

However, this is neither here nor there because there are manufacturers producing great sounding line arrays using either method so it ultimately comes down to research, development and overall system design.

The future

So we’ve established that while modern line arrays have done much to overcome the limitations of classical line array theory in the practical world, there are still challenges. It’s plain to see that while a line array is a near-line source system, it is far from a true line source. The sidestep implementation of the high frequency wave guide is proof of this as is the impossibility of practically implementing the half wavelength rule across the spectrum.

Other peripheral and helpful systems like delay towers and fills notwithstanding, a line array can still exhibit directivity and coverage problems and start to act like a point source beyond the critical distance.

Because of the widespread availability of DSP, some companies are beginning to do some very clever things with their latest offerings and, according to my research, so far only two companies are getting it right: Martin Audio and EAW with their MLA and ANYA systems, respectively.

Martin Audio MLA

Martin Audio has taken a completely different approach with their Multi-cellular Loudspeaker Array (MLA) technology. The MLA may look like a line array – it’s still a J-shape after all – but, according to Martin Audio, it isn’t. MLA provides independent control of up to 144 individual drivers by using discrete DSP and amplification feeds to each driver and the claim is that they are able to break the constraints of the typical 3dB loss per doubling of distance and deliver exactly the SPL, frequency response and coverage specified by the system designer. According to Martin Audio, the MLA offers the following advantages over traditional line arrays:

• Uniform coverage across the listening area.

• Noise pollution reduction by limiting the audio delivery outside the defined audience area.

• The elimination of delay speakers in venues that would otherwise require them.

• Software-driven changes in the acoustic model without the need to drop the arrays for physical adjustment.

Martin audio addressed these issues by starting from scratch and focussing on what was desirable in the audience area rather than what was happening at the speaker grille. One of the stepping stones to achieving this was developing the Boundary Element Model (BEM) where they investigated the inter-cabinet relationship between elements in an array. This allowed hundreds of virtual array configurations to be assessed in a 3D environment and clarification of the previously ignored effects of adjacent enclosures. The conclusion was that the output of a single element drastically changes when it is inserted into an array, even when the other elements are not even turned on! This is key to the MLA optimisation process.

In contrast to early line arrays that were one zone systems and modern large format systems that may be split up into as many as three or four zones with progressive level and EQ towards the top of the array, MLA provides discrete software controlled phase, level and EQ optimisation for up to 72 different high frequency cells. This is what allows for adjustment of the acoustic model on the fly.

The MLA Display2 optimisation software is the brain of the system. With Display2 you start with what you want to hear and the software works backwards to tell you what combination of position, shape and individual elemental EQ is required to achieve that result. The process starts with the entry of venue dimensions, how many cabinets you want to use, the array position and the coverage area.

You then set the coverage parameters including the audience area; hard avoid areas, reference position, front-to-back SPL, and atmospheric conditions. The software will then calculate the splay angles upon which time you rig the arrays. Optimisation is then the next step where you set optimisation targets for non-audience, audience and hard avoid areas. 3 200 filter coefficients are calculated per enclosure which are then exported and uploaded to the enclosures via VU-NET.

EAW ANYA

EAW has done a similar thing to Martin Audio except the design of the cabinets is radically different and the array is completely straight. This has the obvious advantage of a once-off rigging job without the need of having to adjust any splay angles and also allows out fill arrays to be hung coincidentally. Using EAW Resolution 2 software, total system performance is adapted to produce an asymmetrical output to deliver a consistent coverage and frequency response throughout the coverage area.

EAW calls the technology implemented in the ANYA array ’adaptive performance’, which allows ANYA to produce virtually any 3D wave front surface while simultaneously optimising system frequency response to match the requirements of the venue. Each cabinet contains 22 transducers (2 x 15′ LF, 6 x 5′ MF, 14 x 1′ HF) that are independently fed by 22 class-D amplifier channels (10 000W in total) and 22 channels of DSP, similar to the MLA.

EAW’s Radial Phase Plugs and Concentric Summation Array technology is implemented to ensure the MF drivers enter the horn and sum coherently with the HF wave front. Interestingly the 15-inch LF drivers use what EAW calls an ’off-centre aperture’ loading to increase the spacing of the apparent acoustical centres which extends the horizontal pattern control into the LF range.

Needless to say, given that this is a very new system, the specifics of how it actually creates an asymmetrical wave front are unclear. However, it was debuted at the Coachella festival in California. Dave Rat, who was instrumental in the development of the ANYA, says: “We don’t expect it (ANYA) to compete or replace [for example] the L-Acoustics K1, we expect it to be complementary – something that’s got more control; a different tool for a different job. Something that’s better at getting rid of the sound of the room.’

The wrap

So, indeed strides are being taken in the world of large format sound reinforcement. The traditional line array format doesn’t seem to be going away any time soon; no one would expect it to, but there are definitely some exciting things happening in the DSP-driven world that were previously thought impossible. Systems like the Martin Audio MLA and the EAW ANYA are sure to usher in a whole slew of competing systems that will follow suit and one can only guess where that road will take us.