The two legged stool

Join Our Community Subscribe to Paul's Posts

PS Audio engineer Darren Myers asked me an interesting question yesterday. One I have been pondering for the last twelve hours. Why are speaker measurements made with only one speaker?

This was an intriguing thought. We listen to two speakers yet concern ourselves only with the frequency, amplitude, and phase response of one. Why should this matter? The sound we hear from a stereo speaker is the sum of the left and right speakers. At multiple frequencies the two speakers couple their sound output such that the in-room response is significantly different than just playing one speaker.

This coupling is likely one reason flat speakers don’t sound good. I remember designer John Curl telling me the worst sounding speaker he’d ever heard was his own design that was as flat as possible. But “flat” was measured as a single channel. As the frequency bands dip below about 2kHz, where the wavelength begins to approach the width of a human face, the left and right outputs couple together forming a louder amplitude than what’s measured individually.

Stereo measures differently than mono.

Imagine a system whose amplitude response begins to rise at around 2kHz and gets louder all the way down to maybe 500Hz. If we drew that curve most of us would immediately guess the system’s sound was going to be aggressive. In that critical lower midrange area flat would be nice, even a slight dip might help add depth and musicality. Instead, that’s not what we get when we play stereo on a pair of speakers who individually measure flat. We get this bad bump in loudness.

Arnie Nudell routinely added what became known as the Nudell Dip, a slight trough from 800Hz to about 2kHz. I never questioned why he did this, I just knew it made his designs sound right. But now I think I understand, thanks to Darren’s revelation.

Flat frequency response means something only if it is measured in the way it reaches the ear when played on a stereo system.

We don’t listen in mono anymore.