Why Crank It Up?

Related link: http://www.pcmag.com/article2/0,1759,1822559,00.asp

PC magazine just reviewed the new Altec Lansing inMotion iM7, comparing it to several other portable, powered speakers for digital music players. Midway through the review, I came across this curious line:

The sound quality of the iM7 is quite impressive. We cranked it up all the way and found minimal distortion—significantly less than when we maxed out the volume on our [Bose] SoundDock.

I’ve seen this kind of “testing” in many other magazine and Web reviews. And so I wondered again: What does it prove? Should the manufacturer have clamped a governor on the volume knob so that it couldn’t distort? I don’t think so; there are cases where volume is more important than fidelity.

Dan Brown once told a funny story about taking revenge on a neighbor who was pounding on his wall: Dan aimed his amp at the wall, leaned a guitar against it, and then turned up the gain until the amp started shrieking with feedback. He left it going until the police came.

Which reminds me of another powered speaker review I read recently: The reviewer wrote that he had cranked the system so loud that his deaf neighbor complained! She couldn’t hear the music, but the force of the sound had knocked the hanging pictures off her wall.

Anyway, my point is that (A) loud-and-distorted is sometimes worthwhile, and (B) most people who buy powered computer speakers probably want to know how they sound at normal listening levels, because that’s how they’ll be using them.

Specs Offenders

Last year at the notorious Project Bar-B-Q computer music conference, I moderated a brainstorming group that wanted to devise a meaningful way to measure sound quality. Somehow, every year BBQ tries to determine what quality audio is, and every year it reaches an impasse. Indeed, I began the two-day session by telling the group to take a new approach, because I didn’t want them to waste my time. That was pretty arrogant of me, considering that the group was composed of bigwigs from Audio Precision, Analog Devices, Dolby, DTS, Intel, Sigmatel, Waves, and more, but they rose to the challenge.

Here is their report and recommendation, “A Whole-system Testing Framework for PC Audio.” It starts by acknowledging that audio specifications are so misused that they’re often meaningless, and then suggests a daring solution.

Incidentally, Project Bar-B-Q is hosted by the Fat Man, who writes an O’Reilly blog as well. Drop by and tell him what computer audio problems you’d like the BBQ Brain to attack this October.

How do you test speaker quality?