Quote:

Originally Posted by

**apple-q**
The bells ring like crazy when used as notches? They ring when you boost stuff but not when you use them as notches.

Without getting too much into the weeds, boosting and cutting are mathematically congruent in an LTI equalizer with an order greater than 2, it's just a question of wether the filter's resonance is in-phase with the input or 180° out-of-phase. If the EQ is ringing, boosting a frequency rings in phase with the input signal and cutting rings out-of-phase (and least this what I hear when I use it).

Edit: Better explanation=

So in a

biquad (just as an example, it's quite possible Avid's ChannelStrip EQs are biquads) your feedforward gain coefficients move further away from unity in as your gain goes up and your Q goes down, but your feedback gain coefficients get closer to unity by the same proportion. If these are the coefficients:

Code:

w0 = 2*pi*f0/Fs
alpha = sin(w0)/(2*Q)
A = 10^(dBgain/40)
Feedforward gains:
b0 = 1 + alpha*A
b1 = -2*cos(w0)
b2 = 1 - alpha*A
Feedback gains:
a0 = 1 + alpha/A
a1 = -2*cos(w0)
a2 = 1 - alpha/A

The feedforward and feedback gains are unity, plus or minus either the product or ratio of alpha and A. A never goes below zero, it's bounded at 0 and is at 1 when dbGain is zero. As gain increases above zero the feedforward coefficients get further away from unity exponentially, while the feedback coefficients get closer to unity. And as gain drops below zero the coefficients move in opposite directions but by the same magnitude. As the gain's sign changes, either the feedforward or feedback network gets further or close from unity, the actual values are the same, they just switch position on the graph.