The Gold Standard: Anechoic Chamber

Pros: Easy, accurate, dependable
Cons: Even if you had the space for it, you couldn't afford it

Speaker measurement is complicated because you have to isolate the sound of the speaker from the acoustical effects and environmental noises of the surroundings. For example, if your listening room tends to boost bass at 80 Hz, and that shows up in your speaker measurements, you'll think the speaker has a boost at 80 Hz when it's really the room. Sort of like if you were auditioning that really good singer next to another singer who's out of tune.

The best way to eliminate the acoustical effects of the surroundings, and to eliminate the influence of environmental noise, is by measuring speakers in an anechoic chamber, a space filled with giant fiberglass wedges that absorb all sound. Put the speaker in the chamber, put the microphone in front of it, and it's almost as if you'd hoisted the speaker and mic up a mile high over the Pacific Ocean on a windless day when no ships or planes are passing within 100 miles. It's just speaker and microphone.

So why doesn't everyone measure speakers this way? The shell and the wedges for an anechoic chamber cost typically about $500,000. You also need a minimum of about a 30- by 30- by 30-foot space in your building to install them. Once the chamber's installed, you'll still need to invest in measurement gear, including microphones, a computer audio interface of some sort, and a computer with appropriate software.

Allan and I measured the F208 in Harman's full-space anechoic chamber (shown above). The on-axis response curve we measured can be seen in the graph at right, with the blue trace.

But even in an anechoic chamber, doing just an on-axis measurement-say, with the mic directly in front of the tweeter-only tells you so much. It's like if all you asked the really good singer you were auditioning to perform was "Mary Had a Little Lamb," because it gives you only a limited perspective of what the product does. Yes, a speaker radiates sound forward, but it also radiates sound in every other direction, too. These off-axis sound waves eventually reach your ears by reflecting off the walls, ceiling, and floor. So if the sound of those off-axis waves doesn't reasonably match the on-axis waves, you've got a problem.

Speaker companies, and most of the pro reviewers who do measurements, measure the response of speakers both on-axis and off-axis. Harman takes it to the extreme by measuring speakers on-axis, then off-axis in 10-degree increments in a 360-degree circle both horizontally and vertically. Often some of these measurements are combined to create an averaged response across a "listening window." Past practice at S&V was to average the results at 0°, ±10°, ±20°, and ±30° horizontally. With the merger of S&V and Home Theater, this practice will be changing to the standard employed by Home Theater and sister publication Stereophile: an average of 0°, ±15° horizontally, and ±15° vertically.