Factory Direct: Digital To Steer, Aim, Optimize
The development of EAW DSA Series loudspeakers

By David Gunness

Editor's note: This Factory Direct was submitted by EAW. Live Sound makes every effort to eliminate any use of marketing inspired hyperbole.

Steering the vertical beamwidth.

EAW’s new Digitally Steerable Array technology, implemented in the DSA Series of loudspeakers, essentially simplifies the company’s KF900/PPST technology that permits digital steering and aiming of an array’s output and adapts it to applications where column loudspeakers would typically be specified. Like a KF900 array, each driver in a DSA Series loudspeaker is applied its own individual amplification and digital signal processing (DSP).

Through use of DSA Pilot software program, the vertical coverage pattern can be adjusted from 15 degrees to 120 degrees, and coverage can be aimed at plus/minus 30 degrees.

DSA loudspeakers are housed in compact, column-speaker-type enclosures that require no external amplification or processing. Users connect AC power, audio signal, and network communications cables, with Pilot then recognizing each loudspeaker connected to the network, allowing control of all of them from a single interface.

A look at the component set of the DSA250.

EARLIER TO NOW Developed and refined through the later 1990s, the KF900 Series showed that high-resolution measurements of each individual driver in an array could serve as the basis for an advanced modeling program that accurately predicts total array performance at a variety of locations. Further, by manipulating various DSP parameters, this modeling program could optimize performance and even steer the array’s overall output.

By 2001, focused work began on what would become the DSA Series. In a nutshell, we defined that these loudspeakers would need to deliver high output, high definition performance within the constraints of its given application. The given application, in this case, was defined as the small to mid-sized installation in spaces with problematic acoustics such as houses of worship, transportation terminals, and museums.

Peak output was defined to be in the neighborhood of 120 dB SPL, to provide adequate levels. Because speech intelligibility is a critical issue for these facilities, the system needed to provide maximum clarity through the vocal range. Broadband frequency response would also be needed to deliver full-range music reproduction.

Almost from the beginning, we decided to model the enclosures on traditional column speakers. In addition to minimizing visual impact, this configuration would enable the creation of a line array that was central to steering output in the vertical plane.

To this point, balancing enclosure size against driver placement represented the first substantial obstacle. While longer lines extend line array benefits to lower frequencies, a very dense array of sources was required to obtain the degree of beam-shaping precision required. Further, we determined that a staggered array of low frequency cones provided the dense vertical spacing required while also providing a bit more horizontal control. In the high frequency section, a dense array of shallow horns, reminiscent of a miniature KF910, provided the necessary density.

With the basic concepts of the physical design in place, we turned attention to the problems of developing DSP parameters and a control surface that could shape and aim loudspeaker and array output in the vertical plane. The accumulated knowledge of the KF900 project had to be captured in the form of C++ source code as the basis of Pilot software.

ARRAY DENSITY

The design-engineering phase of the DSA project began with the assumption that a single DSA250 full-range loudspeaker would meet the minimum performance requirements. To achieve the necessary broadband pattern control, the low-frequency (LF) section’s line length was specified to provide meaningful pattern control to around 300 Hz.

The high-frequency section is made up of eight 1-inch soft dome tweeters on a single, shallow multi-cell horn.

A line comprised of eight 4-inch LF cones, slightly offset, created the array density that precise DSP control requires. Output and driver spacing drove the high-frequency (HF) section design where eight 1-inch soft-dome tweeters were loaded on a single, very shallow multi-cell horn. The goal was to keep the drivers’ acoustic centers as close as possible. Again, this high-resolution design optimized DSP control to substantially higher frequencies, and indeed, actual performance exceeded expectations with steerability to 16 kHz. Each DSA250 loudspeaker offers a fixed 120-degree horizontal coverage pattern. This wide-angled coverage allows for fairly wide horizontal spacing between DSA250 modules in an installation, with vertical steerability helping to eliminate the dead spots in between.

All signal processing and amplification are included as internal components of the loudspeaker system. Each DSA250 loudspeaker is driven with 480 watts of power (8 by 40 watts LF, 8 by 20 watts HF) and has 16 channels of DSP that includes conventional high- and low-pass filtering, parametric equalization (EQ) and limiting as well as proprietary filters developed specifically to facilitate steering. The internal power and DSP modules are outfitted with a robust driver and electronic protection system.

Network communications cable/ connectors are standard EIA-485 (RS-485), with an optional CobraNet interface available. When connected, all loudspeakers in a system comprise a network in which Pilot software recognizes each individual loudspeaker as well as arrays of loudspeakers acting as a single unit.

Single DSA250 with 30-degree beam (left), and with 15-degree steering (right). Both showing octave-band polars, 500 Hz to 8 kHz.

Several factors led enclosures to be fabricated of extruded aluminum. In addition to strength and rigidity, the use of aluminum allowed the entire enclosure to act as a heat sink, an important consideration for the 16 internal amplifiers. Heat sink ribs on the rear of also act as wire channels for a clean installation. Molded plastic end caps house all connectors with AC power on one end and audio signal and network communications on the other.

Actual measured performance of prototype and pilot run systems met expectations and prove out the concepts of vertical beamwidth variability and steerability. Peak output varies from 120 to 130 dB (SPL), depending on the DSP-controlled vertical beamwidth and steering angle. As a rule, a narrow beam focused straight ahead delivers maximum output with wide beams and sharp coverage angles slightly attenuating peak SPL. Frequency range response is flat from 100 Hz to 16 kHz.

GETTING IT VERTICAL

Of course, most of the testing efforts were focused on the variability and steerability of the vertical beamwidth. Indeed, properly configured DSP can vary a DSA250 loudspeaker’s vertical beamwidth anywhere from 15 degrees to 120 degrees. Further, the beam can be steered plus/minus 30 degrees while maintaining even frequency response up to 16 kHz.

Internal power and DSP modules designed to work in tandem with the transducers.

Beyond those angles, a complex set of variables comes into play. With proper thought and, occasionally, a modest performance compromise, radical steering angles can be achieved. In particular, band-limiting the signal to around 8 kHz allows for radical shaping with smooth frequency response across a large area. The DSA230 loudspeaker is simply the LF section of a DSA250, housed in a separate, smaller enclosure. It includes eight 4-inch LF cones as well as eight channels of power and processing.

The addition of a DSA230 can effectively double the line length of the overall LF section, providing directivity to 150 Hz with nominal pattern control maintained to 300 Hz.

DIFFERENT PATH

Pilot software was created on a separate development path from DSA loudspeakers, with its roots lying in the KF900 project’s FChart optimization program. FChart uses high-resolution measurements of each individual driver in an array to predict and optimize total array performance. Pilot uses this same approach to control multiple loudspeakers in an installation environment.

DSA Pilot, developed to assist in design and installation.

Pilot assists in both the design and the installation of a sound reinforcement system yet requires no advanced acoustical knowledge. Once Pilot knows the dimensions of the audience area to be covered, and the location of the loudspeakers, it controls the DSP settings to form a vertical beam matched to the space. While the control computer can be disconnected after installation, many users will prefer to leave it connected for monitoring purposes. (It’s multi-level password protected to prevent tampering.)

Ultimately, Pilot allows users to create multiple DSP configurations to meet a variety of needs such as multiple zones or voice-only/music options. However, it does not attempt to replace the system designer; rather, it is intended only to be helpful to develop a detailed “sketch” showing loudspeaker placement and horizontal coverage.

When the basic system design is set, the user begins by building the total system on Pilot’s “Network” screen. Select pre-configured clusters ranging from one to several loudspeakers to suit the cluster’s output and control requirements. Name each cluster and module appropriately to speed the installation process.

Where the audio signal and network communications feed the loudspeaker.

Next, a coverage area profile can be created for each cluster. Define the size and slope of the vertical coverage area’s floor and input the cluster’s mounting height. Pilot displays the ideal vertical coverage beam as well as the cluster’s current vertical beam. The user can adjust the cluster’s beam manually or Pilot can optimize the beam to match the ideal. Results can be exported to EASE 3.0 for display or further study. (EASE 4.0 users can use the “DSACluster” loudspeaker DLL, which allows them to adjust the cluster’s steering parameters within the EASE environment.) Finally, the user addresses settings for the clusters as a whole such as parametric EQ or gain and delay settings for secondary clusters in a larger system. Once the system is installed and connected, Pilot then seeks out and identifies all clusters and loudspeaker modules on the network.

The user selects an identified loudspeaker, activating an LED on the enclosure. Each loudspeaker is identified and associated with its mate in the design, confirming that the system is properly installed. Pilot then determines if any groups in the system should be treated as clusters or individual modules, and then adjusts each loudspeaker’s DSP setting in conformity to the design, and the system is ready for use. The user is left to adjust front-end equalization to taste.

LEAD WITH THE HIGHS

Even though we’re relatively early in the process of fully defining in-depth performance factors with the DSA Series, here’s some recommended guidelines and application notes:

• With 120-degree horizontal coverage, horizontal arraying of multiple clusters is rarely required. When it is, splay angles between clusters should be at least 120 degrees.

• DSA250 loudspeakers are designed to “lead with the highs,” so the HF section should be closest to the listeners whenever possible, typically putting the HF on the bottom. However, if the loudspeaker were covering a balcony at or above its level, then the HF should be on top.

• Users can add as many DSA230’s as they wish to a single DSA250. Pilot will recognize and control as many modules as needed to reach the desired output or directivity goal.

• When desired, directly adjacent DSA modules can be treated separately so the lower one can, for example, cover a main listening area while the upper covers a balcony.

FLYING OPTIONS

Each DSA250 and DSA230 includes a standard mounting kit with two 2-part brackets and a mounting template. One part of each bracket mounts to the wall and the other part to the loudspeaker. Then lift the 51-inch, 100-pound loudspeakers and set them in place. The brackets allow swivel of up to plus/minus 55 degrees, which provides convenient access to connectors. After being aimed in the horizontal plane, tightening the connecting bolts then secures the module in position.

Mounting brackets that can swivel up to +55 degrees.

In situations where installers must suspend DSA loudspeakers, an optional suspension bracket is available. This permits the physical angling of DSA enclosures to provide maximum output at a down angle. However, note that Pilot does support applications in which loudspeakers are not flush mounted. Work on the DSA Series, its concepts and principles, continues, as frankly should be the case with any promising technology. At the same time, the current iteration hits the mark in terms of a smaller, “smarter” option offering customizable tools that, applied correctly, can help solve both standard and difficult application challenges.

David Gunness is director of research and development for EAW, and holds several patents related to loudspeaker design. He can be reached at david.gunness@eaw.com

October 2003 Live Sound International