Peak overshoots after subtractive EQ

[Eurocide]

Today I loaded a ready-made drumloop which is already normalized to -0.1db.
It has an annoying emphasis at 160 Hz and I reduced it with the Cambridge by -9db.
Then I noticed, that the overall peak goes up to +3.6 db!
Just to be curious, I put the precision limiter afterwards and it also reacts to this new peaks.
But they are not audible though.
BTW, the standard C4 EQ produces the same effect.
What is the reason of this behaviour? EQ phase shiftings? Intersample/phantom peaks?
:-k

 

[Fundy]

It's do with the way filters work and harmonics. I don't know the exact technical reasons other than the perceived levels are increased in another spectral region when you turn another down (less is more).

The Q/Bandwidth affects how much is turned up and down, conversely this will change the perceived volume of related harmonic values. So when you dropped the volume -9dB at 160Hz, depending on the bandwidth setting used this would cascade onto 320Hz, 640Hz and so on.

Like I said before a lot of it is science and to get a fuller understanding you would have to do some proper research. Although I do know, a standard parametric equaliser uses a 3dB/Oct setting.

 

[boody]

Today I loaded a ready-made drumloop which is already normalized to -0.1db.
It has an annoying emphasis at 160 Hz and I reduced it with the Cambridge by -9db.
Then I noticed, that the overall peak goes up to +3.6 db!
Just to be curious, I put the precision limiter afterwards and it also reacts to this new peaks.
But they are not audible though.
BTW, the standard C4 EQ produces the same effect.
What is the reason of this behaviour? EQ phase shiftings? Intersample/phantom peaks?
:-k

try linear phase. It should cut 'clean'.
Budy