Originally posted by Mike Sulzer
View Post
For telephone circuits, phase is not important. Bell Labs tested this extensively, because they wanted to know if they had to design long distance telephone circuits to be phase linear (which preserves wave shapes) or not, with a lot of money riding on the answer. Turns out phase linearity is not required for understandable speech.
For music, it isn't quite as clear, and the better the fidelity the more phase linearity seems to matter. This was the basic point of the stuff about architectural acoustics and the feeling of presence: LARES - Wikipedia, the free encyclopedia. Follow the links to David Griesinger.
There was also a thread on this, and there are some microphone makers that emphasize phase linearity out to 50 KHz.
As for guitars, I have always felt that the handling of the attack transient required phase linearity.
I would point out that given a measured impulse response function, one can compute the complex spectra, from which one can compute the power spectra. So, the impulse response is the more fundamental measurement.
Comment