CNET también está disponible en español.

Ir a español

Don't show this again


Do speakers that measure well in the lab actually sound better?

Can we measure sound quality? The Audiophiliac thinks not.

While measurements provide valuable information about how a speaker performs for designers and engineers, measurements can't predict whether a consumer will prefer one speaker over another. I'd go so far as to claim that choosing a speaker based on how well it performs in a lab would be a mistake, simply because there's a good chance there are better-sounding speakers that don't measure as well.

The KEF LS50 (left) and Falcon LS3/5A (right) speakers measure and sound very different.

Steve Guttenberg/CNET

The problem is simply this: Sound quality is impossible to quantify, and even if it was, all bets are off when you're listening to the speakers in the real world. Your room's size, shape, acoustics, locations of windows, mirrors, furnishings, floor covering and taste are unknowns to the speaker designers, so they can't know how the speakers will perform for each owner. Lab measurements subtract the variables for consistency for the engineers, but that limits their usefulness for consumers.

I'm a subjective reviewer, but my old friend Brent Butterworth relies on measurements and his ears for his speaker tests. We've discussed the pros and cons of subjective versus objective reviewing for years, but last week I interviewed him for this blog.

Butterworth started with this assertion, "Research has shown that people like speakers with a flat measured response on axis, and a gently rolling off response off axis," but then he quickly added, "I liked the Bowers & Wilkins 804 D3 better than my Revel F206 towers, but the 804 D3 doesn't measure as well." Exactly, that's my point, the better measuring speaker isn't necessarily the better sounding speaker.

For the 16 years before I started writing audio reviews, I was a high-end audio salesman. I demonstrated speakers for literally thousands of people and here's what I observed: People like what they like. Some want to totally rock out, some crave purity so their favorite opera singer's voice sounds perfect to them and some want to feel the bass through their feet! It was rare to encounter a customer who prized accuracy, or asked about speakers that measured well.

Butterworth sees it differently, and believes most people, if they compared one speaker against another would choose the more accurate, better measuring speaker. Butterworth continued, "A speaker that measures flat is a safe bet. It isn't going to sound bad. It's going to perform well in a wide variety of rooms, and it's not going to have any overt colorations."

Then I asked Butterworth, "Do you listen to the speakers you're reviewing first, then write about the sound, and measure them later?" He said 80 percent of the time he does, and added "I can tell from listening if a speaker is going to measure flat, but I've been a little surprised from time to time. Almost always, when I hear flaws they show up in the measurements." Right, my ears hear those flaws as well, but Butterworth thinks reviewers are unaware of their biases, and measurements are a check against that.

The very best measuring speakers are among the best sounding he's heard, but according to Butterworth the Bower & Wilkins CM10 wasn't a great measuring speaker, but he still thinks it sounds really good. So even for Butterworth his subjective opinion sometimes trumps his objective measurements. How could it be otherwise? Humans hear differently than microphones because our brains "filter" the sound, and I'm writing reviews for humans, not microphones.

Some reviews with measurements that reveal a technical flaw come off like a "gotcha," as if the speaker designer was unaware of what the measurements would reveal. Assuming the designer was competent they were probably well aware of the "flaw," but saw it as a necessary design compromise. No speaker at any price is perfect, and designers know that people buy speakers to listen to, not to measure. A poorly designed speaker's flaws would be audible to anyone with ears, and the measurements would be superfluous. Of course, folks who put their faith in measurements should buy the best measuring speakers they can afford. As to whether it's a good sounding speaker, that requires listening.

In the final analysis, neither measurement oriented or subjective speaker reviews provide absolute assurance of satisfaction to the readers. All I can do is try my best to describe the sound of a speaker, based on my experiences of hearing a countless number of speakers. Then it's up to the buyers to listen for themselves.

Brent Butterworth reviews speakers for Home Theater Review.