As sound reproduction has transitioned to the digital realm, audio professionals have a wealth of capabilities that were unthinkable in the analog days.
Today, loudspeakers with companion digital processors are increasingly a tool of choice, going well beyond just the higher end of the application spectrum.
Virtually all types and brands of digital loudspeaker processors offer a rich suite of features, often overwhelmingly so. Some have capabilities that are unique, while others seek to capture the most important functions at the lowest possible price point. Nearly all provide a comprehensive set of functions that, if compared in a vacuum to the analog crossovers of yesteryear, would actually seem to be science fiction rather than reality.
The advanced capabilities - now often taken for granted – have unquestionably raised the performance potential of sound systems by at least an order of magnitude. But they also open a Pandora’s Box.
For example, how do we discern the choice of a Bessel curve over that of a Linkwitz-Riley? Why select a 12 dB/octave crossover slope instead of a 48 dB/octave slope? What’s actually going on with asymmetrical crossovers, extensive EQ capabilities, the use of digitally adjustable all-pass filters, incremental driver delay, phase filter controls, and the many other features that have become commonplace in the modern DSP loudspeaker controller?
Product user manuals do a great job of telling us how to access the various features of their products. But do we actually know why we might want to utilize these features? The answers are not simple and come from a deeper understanding of electro-acoustics, loudspeakers, and processing.
This is the first in a series of articles, to be continued regularly here on PSW in a step-by-step manner, to provide those answers. Here, we’ll start by looking at how to view and understand the on-axis and off-axis response of a typical single 2-way loudspeaker.
The Bruel & Kjaer 4007 Series microphone positioned to measure the Apogee AE-5 2-way loudspeaker. (click to enlarge)
Importance Of Instrumentation
While many aspects of sound system calibration can be set by ear, it’s physiologically impossible to set up incremental time delays, all-pass filters, and phase filters relying solely on listening. Yes, you can select recommended crossover points and crossover rates, and even woofer delays from a loudspeaker manufacturer’s list of recommendations, and that might get you pretty far down the road.
But there is no substitution for taking the measurements yourself and making the optimal adjustments accordingly. Manufacturers are as overworked as the rest of us, and while most of them do their best to provide optimal guidelines, they don’t have access to all possible loudspeaker controllers, or in many cases, to optimal measurement instrumentation. For 17 years I was one of these folks. (By the very nature of the business, their suggestions are generic at best.)
Listening Vs Measuring
A common argument goes that “if whatever I do to adjust the equipment doesn’t sound better than before I adjusted it, then it’s meaningless.” On the surface, it’s a good argument, and it certainly applies to on-site, real-time mixing.
But dig one level deeper and you have to consider that whatever music you’re using as your guideline is written in a certain key and does not contain all possible fundamentals and harmonics.
For example, change the key from E to E-flat, and that nice, even bass line might suddenly reveal some holes or peaks in the response. Or, when the drummer shows up for sound check and re-tunes the kick drum to his preference, the tonality could change from tight and punchy to dull and mushy.
Moreover, trying to make micro-second alterations to mid-high crossover points - without the aid of instrumentation - is like trying to guess what bacteria looks like without using a microscope. It’s a sub-optimal condition at best. The only way to achieve the best possible result is through high-resolution measurement.
Let’s start with the basic process for optimizing a loudspeaker by looking at how to view the on-axis and off-axis response of a single 12-inch cone driver found in a typical 2-way biamplified loudspeaker.
Where pertinent, this will also include details on how the selection of each DSP signal conditioning choice was arrived at, and why each is important to the final result: making the loudspeaker as flat as possible within its operating range.
The Hewlett-Packard 35665A dual-channel FFT-based spectrum analyzer with Jensen Twin-Servo mic preamp and XTA DP 548 digital loudspeaker controller. (click to enlarge)
The measurements utilized for this article were conducted using a Hewlett-Packard 35665A dual channel FFT-based dynamic signal analyzer, a Jensen Twin-Servo 990 mic preamp, and a Bruel & Kjaer 4007 microphone. Equalization, where it appears, was performed with an XTA DP548 digital loudspeaker processor/controller.
All measurements were made in a near-field environment, within 1 meter of the loudspeaker, in a 45-foot and 25-foot carpeted room. The DUT (Device Under Test) was an Apogee AE-5 loudspeaker powered by a Hafler D-220 amplifier.
All signal level devices and the amplifier were tested with the HP analyzer to verify that they were flat within +/- .25 dB.
The B&K 4007 microphone was tested in relation to two other identical mics, none of which displayed a deviation greater than 0.5 dB from 25 Hz to 18 kHz.