Skip to main content

Sound And Sensibility

  • Since the last time this column took up valuable advertising space in this fine periodical, a lot has transpired. I've been evicted after 28 years of living at the same address, hired attorneys to negotiate with attorneys, had my rent triple, and moved to the end of the world. I'm now about as far west as one can go in San Francisco without falling into the Pacific Ocean; well okay, you can go one block farther, and okay, you don't technically fall into the ocean, actually onto the road just above the ocean, but I can look out my living room window and see Ocean Beach.
  • Lots of changes, and actually they're turning out to be for the better, with the exception of the guy running into the parked car behind our car (parked in front of our house) and turning that car into totalled sandwich meat with his car and our car as the bread...smashing in the rear of our car in the process, or the tree which blew down and crunched the car parked behind our car (again parked in front of our house) and also hit our car (minor scratches). All true. So if you come to visit me don't park in the space behind our may not be safe.
  • The other thing that seems unsafe these days is sufficient awareness of the many challenges inherent to a successful sound system design. Those of you with long memories may recall FDR's definition of personal liberties via the Four Freedoms (that would be "school memories" not "were there" memories). Well, in sound system design we have, courtesy of Chris Foreman, our own definitions of inherent liberties for system performance, but they're expressed as questions via the Three Will-its. Will it be loud enough? Will it go into feedback? Will it be intelligible?
  • What's cause for concern is the apparent simplification of system design afforded by some of the new series of software and components. Here's where Chris' three Will-its are so great, because in three simple queries he's encapsulated years of investigation into audio systems performance and the psychoacoustics of the listener. Embedded within those queries are the engineering drivers such as polar response, directivity, wavelength, sensitivity, power handling, distance to listeners, EAD, PAG-NAG, critical distance, T60, Hopkins-Stryker, ambient noise level, Alcons, STI (plus some I've probably forgotten); which will, if correctly juggled, allow the response to these three queries to be answered affirmative, negative, and affirmative, respectively.
  • Note that in spite of the proliferation of processing power and associated software wizards, the 3-Ws are still the factors our end-users/clients consider the benchmark of a successful system. Yet at the design end of things a certain laissez-faire attitude has crept in, encouraged by the new tools at hand.
  • Now don't misunderstand me, I've nothing against new technology or tools, it's great to be able to solve intelligibility/coverage problems in a five-second room with $25,000 of steered arrays versus the $100,000-plus pewback system I'd have needed to use not much more than a decade ago. What is a concern is where the new tools are being used without really considering what they're doing or whether in fact it needs to be done.
  • For example, consider one type of component, currently available from a variety of manufacturers either as a physical device or a simulacrum via software, the ubiquitous "feedback assassin." There's no question of the usefulness of such a device, but one has to question its necessity. Whenever a particular component starts to become generic in a system design you have to consider whether that device is truly generally applicable, whether it's the current flavor of the month, or whether a "CFOTM" has simply crept into common usage.
  • Back in the day, the efforts of a significant number of practitioners and researchers established the mechanisms and components of feedback in audio systems. From their work the PAG-NAG series of equations evolved which allowed a designer to define the limits (distance of loudspeaker to talker's mic, distance of loudspeaker to farthest listener, distance of talker to mic, etc.) for a particular system to be stable and nominally (there can always be the one-off exception) feedback-free. Although a system designed in this manner doesn't really need a feedback assassin, it may still be there as insurance to catch the odd event, particularly in situations where even the risk of feedback is unacceptable. Seems sensible enough.
  • However, often in my experience, the quickest way to get an existing system to sound better is to simply bypass any existing feedback assassin rather than reconfigure it. Does that mean that the technology is a problem? Not at all, rather it means it's being used without much thought, either being "pushed" into ring mode at levels that are atypical of normal use or being configured to "hold" every ring filter rather than letting them run dynamic.
  • As system design becomes more generic and "automated", whether via manufacturer's tools or Wiki-knowledge, and as this approach to design becomes more pervasive, there's the danger system design will become a commodity like so many other things, and its value will fall to the lowest common denominator. In that market the only viable way to compete is with the added value you can bring to the table. In my mind that value includes understanding the laws of physics behind the tools you use, rather than merely letting the set-up wizards run the show. Might be worth your consideration.