[cerberus]

is it correct that phase shift = delay short enough that it is perceived in the frequency domain? i tend to assume from observation that using too much eq can accumulate these delays, eventually it shows up in the time domain as smearing.

an aspect of eq's sound that was not discussed which is somewhat mysterious to me are side lobe resonances... are these the same for every curve? my observation is perhaps there are ways of minimizing them, or perhaps we like pultecs becuase they tend to be so dirty this way, but it works well on music? jay didn't raise whether side band resonance is important to the individual color of an eq.

phase rotators are said to be made with banks of all-pass filters...but i have a hard time understanding how a filter that passes "all frequencies" is a filter at all. are transformers considered all-pass filters? i've seen transfer functions made from line transformers. they look like filters, or combinations of filters...more or less flattish in the middle but not exactly...and with roll off at both ends. isn't that a band-pass filter? in a digital system, by the time we get to up the nyquist freq: nothing there. so every curve has to roll off before that, how could any be "all-pass"?


jeff dinces