Line Arrays: A Buyer's Guide- Part I

In Uncategorizedby tfwm

Line arrays have been readily available to the installation market for the past two to three years, and have enjoyed “favored child” status with most of the major audio manufacturers during this period. The question remains, however, whether they represent the appropriate solution when upgrading your existing system or designing for your new facility. The answer, as always, depends on your application. In this issue, we’ll look at what’s currently available on the market, how arrays operate, and some key purchasing issues to consider. Next issue, we’ll dive into some additional concepts to consider, as well as take a glimpse at the next generation.

Let’s look first at the goals, and the variety of solutions available in today’s market. To set a baseline, let’s presume that the objective of any quality system should be to provide clear communication and accurate music reinforcement, and that system performance should be as consistent as possible from seat to seat.

The development of the modern line array has yielded a new set of tools that potentially represent a substantial improvement over traditional systems, as we’ll explore further. Their performance, however, hinges greatly on the design of the array elements, and how they are arranged and installed. Line arrays can definitely be created poorly, creating extremely uneven frequency and SPL response, and unwanted power lobes at various frequencies. We’ll explore some of these issues, and how to anticipate in advance what a given array will generate. While there are a number of line arrays designed specifically for voice reinforcement, in this article we’ll focus on arrays that provide full-bandwidth performance.

In the church market, it’s clear that effective communication of the message is the reason the system is there to begin with. For better or worse, the overall image projected by the church is also a key component to consider. Today’s audiences expect a fairly high level of audio performance, and a system with poor performance communicates its own message. The bottom line is that systems have to be good enough to keep people satisfied (and returning), while being, in the truest sense of the phrase, cost-effective. With church budgets being typically tight, there is little room for getting it right next time.

The introduction of full-frequency line-arrays in the 1990’s caused a certain amount of confusion because line array performance didn’t fit into the neat descriptive boxes we’d used in the past to define system performance. New terms were introduced, many of which attempted to articulate in simple terms the complex performance of the waveforms developed by an array. Thus, terms like “cylindrical wavefront” and “line-source” were used to educate consumers and users. While these terms may have oversimplified many of the concepts, keep in mind that they were all introduced in an attempt to explain new performance parameters. True array performance was different enough to justify the new terms.

Line arrays can be an elegant and efficient solution to a number of problems that have plagued sound reinforcement systems since the second speaker was placed next to the first. While line arrays are capable of a substantial increase in performance over a traditional system, they must be designed and installed with careful regard to the physics involved with array behavior. Bolting cabinets into a linear arrangement will not create the performance of a true full-bandwidth line array, unless the array elements are specifically designed to create a line-source.

Just about any audio manufacturer can park gear on a stage and make it sound good. The challenge is to make the total system, designed for the correct coverage area, respond so that the performance is consistent no matter where you are in the facility. This, by the way, is where the consultants and design-build contractors earn their money.

To understand the difference between a demo and a well-engineered and installed system, it’s extremely important to understand the root physics issues. Remember that when two radiating acoustic sources reproducing the same signal have overlapping coverage patterns, the result includes phase-induced interference. This interference is exhibited in a number of forms, most notably cancellations (comb-filtering) within the primary coverage area, leading to variations in SPL and frequency response and off-axis side-lobes (uncontrolled energy outside of the primary coverage pattern). The amount of interference is directly related to the distance separating the two sources, and wavelengths related to this distance (frequency). If, however, the distance between acoustic sources can be reduced, the negative, or destructive, cancellations within the primary coverage area are minimized. If the acoustic center spacing can be reduced far enough, the result is constructive interaction, where the output of the acoustic sources sum coherently, thereby increasing the total output. Note that these are universal principles and apply to any array or cluster of speakers.

The optimum solution, therefore, would be to create a system that effectively reduces the distance between array elements to zero, thereby creating a phase-coherent single acoustic source. This theoretical result would eliminate phase-induced cancellations, while minimizing off-axis lobing and destructive inter-element interaction. Creating a single acoustic source capable of generating a configurable coverage pattern with sufficient output is the goal of modern array designers.

Line array performance
Please note that a line array, by definition, is an assembly of devices in a line, no more, no less. Since single acoustic devices radiate energy as a point source, expanding both horizontally and vertically, it’s not typically a great idea to create an array of devices whose overlapping patterns will interact. Properly constructed, however, purpose-built elements in a line array will have coverage patterns that can couple coherently and combine to form a single acoustic source in the form of a line (line-source), creating phase-coherent summation in the primary coverage area.

A well-designed line array forms a radiating plane or ribbon of energy that expands primarily in one dimension (as opposed to a point source), with destructive interference creating null zones, or areas of cancellations (typically above and below an array). These null areas can be 12 to 15dB below the level of the primary coverage area. While this is also frequency and array-length dependent, it’s a huge potential benefit for situations where gain-before-feedback is problematic, and where poor mic technique and extensive lavaliere mic use is prevalent (sound familiar?) The primary advantages of a well-constructed line array remains it’s ability to create very even frequency and SPL coverage throughout the seating areas (since the array can act as a single acoustic source), and better stereo imaging, since you’ve reduced the amount of interference within the primary sources.

Line-arrays are also by nature very efficient. Since array elements do not have to overcome cancellations created by their interaction with neighboring elements, the physical size of the system is frequently smaller than a traditional system. The primary benefits here are better sightlines, and in many instances, lower system costs (fewer cabinets, plus fewer amp channels). Arrays also provide the ability to transmit energy efficiently over quite a distance, particularly in the higher frequencies. While this is again dependent on the size and shape of the array, this can help offset HF air-loss absorption, and minimize or eliminate the need for delay fill speakers (again, fewer speakers and their related amp channels needed).

An additional benefit is the apparent extension of the near field, exhibited as 3dB attenuation per doubling of distance (as opposed to far field and point source’s 6dB attenuation per doubling of distance). Please note, however, that near-field extension is highly dependent on the size and shape of the array, frequency, and listener position. For example, an array may transition from near field to far field behavior in a matter of a few feet at the lower end of the frequency spectrum, while high frequency transition from near field to far field may take place hundreds of feet from the array. The immediate impact of extended near field though, is an extended level of immediacy or presence from the mid-frequencies on up.

How it works
What you hear as broad bandwidth line-source array behavior is primarily dictated by three factors: the size (length) of the array, the spacing between the individual components, and listener position. Simply stated, devices will couple coherently if their acoustic sources are within a half-wavelength of each other at their highest operating frequency. Alternately, if the acoustic centers cannot be placed close enough together (due to the size of the wavelength and the size of the driver), devices will also couple coherently if their output’s waveform is relatively flat (curvature less than a 1/4 wavelength) and their total radiating area occupies at least 80% of the face of the array1. Research has shown that the flatter the waveform; first, the closer the performance will be to approximating a perfect line source, and second, the lower the side lobes will be, keeping spurious energy sources at bay. This second set of criteria tightly defines how much an array can be articulated (shaped) before the inter-element coupling breaks, resulting in a series of independent point-sources (with the inherent comb-filtering issues).

To put this in physical terms, if two devices operate up to 200 Hz, their acoustic centers can be placed up to 34″ apart (1132 [speed of sound] divided by 200 [frequency], divided by 2) and still generate a coherent wavefront with relatively well-behaved side lobes (at least 12-14dB below the level of the primary lobe). On the other hand, compression drivers operating up to 17kHz would have to be small enough to allow their acoustic centers to be placed within 0.4″ in order to maintain the first coupling criteria. Thus, it is relatively easy to build a coupling array in the lower frequencies, up to the point where driver size forces the second criteria to come into play. This dictates both the size of the driver, and the amount of separation available between array elements. It also means that any driver operating above approximately 1kHz will probably need to adhere to the second set of coupling criteria.

Since acoustic sources radiate spherical wavefronts, some manipulation must take place in order to adhere to the second set of criteria, making the wavefront transition from one that is spherical, to one that is relatively flat. The first device to address this aspect of high-frequency waveform manipulation was the DOSC waveguide, introduced by L-ACOUSTICS in the early 1990’s. Later product development included other forms of complex manifold technology as utilized by manufacturers like Adamson, Meyer Sound, Nexo, etc. This is a highly proprietary field, since the success of this device will dictate the success of the coupling performance in the higher frequency range. The effects of a curved wave-front’s negative interaction within an array is well documented, so make sure the devices you’re interested in can support their claims.

One notable exception to maintaining these coupling criteria is EAW, whose arrays incorporate the approach that various technologies should be used for the various frequency bands. In other words, line array behavior should be maintained where inter-device spacing allows, and other relevant technologies used where appropriate for the mid-range and up. Thus, high-frequency coverage is handled via traditional horns. This approach takes advantage of low-frequency coupling and its inherent efficiency gains. The use of horns in the mid-high area allows specific coverage angles to be used for specific coverage areas, which typically increases the total vertical coverage of a given array. This approach, however, should not be considered as an example of line-source behavior in the upper frequency ranges.

It should be noted here that high-frequency manifold developments reflect working with the physical waveform to create seamless coverage over a specific coverage area. Tailored pattern control has also been realized by manipulating phase and SPL between various array elements, though care must be exercised in order to maintain the “single source” performance benefits of the array. This DSP-centric approach represents the next wave of array development, and the new products by EAW, Meyer Sound and others reflect this approach. More on this in the following article.

An interesting side note is that the HF waveform manifolds from Adamson, L-ACOUSTICS, Nexo, Meyer Sound, etc. all seek to generate a seamless ribbon of energy along the face of the array. The question to be raised then, is whether a ribbon driver (with a waveform that inherently meets the coupling criteria) is viable as a driver in line arrays. At this point only three professional audio manufacturers, SLS Loudspeakers, Stage Accompany, and Alcons Audio are actually using HF ribbon driver arrays. While early ribbon drivers could not handle sufficient power levels or generate sufficient SPL, recently developed ribbon drivers take advantage of new materials and technologies to allow greater short-term power handling than was previously available. While generally still a few dB shy of matching the output of systems using traditional HF horns and compression drivers, potential ribbon driver benefits include coherent coupling between HF components (given the nature of their inherently flat waveform), greater transient response, and much lower distortion levels than that which is typical of a compression driver. However, spacing between driver assemblies still must adhere to the 80% rule in order to create coherent summation.

System Design
One of the beauties of line array design is that it can be molded to fit fairly specific performance requirements. Since the elements (theoretically) couple coherently, an array’s shape can be straightened to generate additional SPL for long-throw situations. Essentially, the straight portion of an array increases the total number of components covering a given area, thus effectively decreasing the total area being covered by these components, providing denser coverage (and thus higher SPLs). Similarly, if you want to reduce the SPL (to the near-fill areas for example), you can curve the array (within limits) to dissipate the energy. The key issue here is that the array can be articulated as long as the coupling criteria are maintained. If the coupling criteria are not maintained, the array will generate variances in both SPL and frequency response within the primary lobe, and unacceptable off-axis lobes. Thus, the primary consideration is to what degree they can be separated before the cylindrical waveform breaks.

Articulated vs. constant-curvature arrays
Another approach taken by a number of manufacturers is to move the apparent acoustic centers behind the physical device, where they merge to effectively create a single acoustic source. Most arrays using this approach create a constant curvature array (think of the merged acoustic centers being the reference point, and the front of the array elements creating a section of an expanding sphere). This approach, used by L-Acoustics in their ARCS series and by Renkus-Heinz in their RPA arrays for example, yields a theoretical single acoustic source whose output has a curvature that remains constant (hence the name). Nexo also uses a variation of this approach in their Geo-series array elements.

While solving the issue of multiple acoustic sources, systems of this type typically cannot be effectively articulated into a straight line in order to drive energy further into the house (doing so would move the acoustic centers apart). Thus, they generate consistent SPL across the length of the array as opposed to the variable output available from an articulated array. The exception to the model is the Nexo Geo system, which uses a series of internal acoustic reflectors to generate a configurable pattern. Within these limits, constant-coverage arrays do couple effectively, and should be considered in the appropriate application.

What does all this mean to a system designer/user/buyer? Bottom line, do your homework and pay attention to the physics. There is hard science behind the development of many of these products. Unfortunately, the array market has become congested, and some marketing claims may exceed what can be realistically accomplished. Looks can definitely be deceiving, but the legitimate contenders will be able to back up their claims with research, and will support independent predictive software packages such as EASE and CATT.

System Cost Issues
Beyond the individual array element costs, there are three key considerations. First is ancillary hardware costs. With many systems, a given amp channel can drive multiple components. DSP-centric systems may need a DSP channel and individual amp channel per band, per element. Thus, for a simple three-way, six element array, you may need up to 18 DSP and amp channels per side. Pricing of self-powered systems such as those available from EAW, JBL, Meyer, and Renkus-Heinz include DSP and amplification. Each non-powered system can use different DSP processing and amplification schemes, so you’ll have to verify each design and its DSP and amp requirements to achieve a true “apples to apples” comparison.

Another hotly debated topic has been the various array rigging systems. While primarily the concern of the touring production companies, the type of rigging used can also have an effect on the cost of an installation if it significantly increases (or decreases) assembly time. Early arrays used rigging systems composed of a variety of specific components to dictate the angles between cabinets. Most recent arrays, however, utilize rigging components that can be used at a variety of angles, and which remain attached to the individual array element. These captive rigging systems are generally included in the cost of the array element, while non-captive specific-value rigging components are not (which effectively increases the cost of the system). Additionally, manufacturers are addressing system costs by offering installation versions that eliminate expensive rigging components. For example, the McCauley IN.LINE’ array elements are purpose-built for permanent installations, and JBL has introduced installation versions of their arrays.

What to watch
Vertical coverage capabilities are probably the biggest issue that affects consistent system performance and cost. Vertical coverage is traditionally calculated as being a sum of the angles between the array elements. Thus, if a manufacturer allows their arrays to be constructed with greater inter-element angles, they will be at a competitive advantage since fewer elements are required to cover the vertical dimension. Remember however that the amount an array can be articulated is tightly defined by the coupling criteria, and exceeding the theoretical limits results in the system acting, again, as an assembly of inter-reacting point sources. Maximum inter-element angles are based on math (and as such, are absolutely definable), taking into account such things as the size of the components, size of the array elements, and listener position. Many manufacturers provide rigging hardware that will accommodate angles that exceed the theoretical maximum allowable values, so pay attention. When in doubt, have the arrays modeled in an independent predictive program, with the aim of keeping frequency and SPL coverage as consistent as possible.

Predictive software
Virtually every manufacturer has created design support software to calculate the number of elements required for a given system. Some are more extensive than others, providing information such as aim angles, center of gravity, horizontal coverage, weights, heights etc. Some of the newer tools also indicate SPL variation over distance at various frequencies, which is extremely useful in determining the effects of various array configurations, transition to far-field effects, etc. One of the most cost-effective methods to explore array behavior before you buy is to examine the array’s performance in modeling software that includes phase information (since this is the basis for line array performance.)

Both the CATT and EASE 4.0 programs now utilize phase information to calculate array coverage in a room and inter-element interference, shown as coverage balloons at a given distance from the array. Using information provided by the manufacturers (EAW, Electro-Voice, JBL, L-Acoustics, Renkus-Heinz, and others support EASE array models), these predictions illustrate the interaction between elements at various frequencies and inter-element splay angles.

Since the response of any array changes with distance (again, based on the transition from near field to far field and contingent on the size and shape of the array), these balloons should be examined at various distances. However, they do show what happens to system response as the arrays are articulated. Request these predictions (based on the specific array you’ll need) since they represent a relatively objective perspective on different arrays and their behavior at various inter-element angles and frequencies. Again, it’s easier to get an array to couple nicely at lower frequencies, so make sure you see plots reflecting the higher frequencies as well.

The next arena of battle is likely to be steerable patterns and horizontal coverage. Over the years, the industry developed an extensive set of tools in the form of a variety of constant-directivity horns to tightly control coverage within specific portions of a coverage pattern. Most line array offerings are not yet as extensive, offering a limited palette from which to attempt to construct exact coverage patterns as the horizontal pattern transitions from the front of the house to the back. To date, only the McCauley offers array elements in a full range of horizontal coverage angles (60o, 90 o, and 120 o versions), which can be used as needed within an array. Other remaining issues include the aspect of smooth interaction from adjacent arrays, as would be found in systems employing left-left/left/right/right-right, LCR, and matrix-fed configurations.

Line arrays present the capability of providing significant performance increases over traditional designs, if constructed accurately. It remains the buyer’s responsibility to examine particular manufacturer’s claims in the light of the available research and objective modeling programs, especially if the claims yield suspiciously advantageous bids. In the next issue, we’ll explore the potential of DSP-based systems, as well as some other alternate solutions available on the market today.