A few EQ questions...

nebulae · Post by **nebulae** » Tue Oct 16, 2007 7:13 pm

for the nooBs, I believe the "learn macro" parts he's referring to is if you had an EQ8 in a rack.

Bassic Dave · Post by **Bassic Dave** » Tue Oct 16, 2007 7:35 pm

Thats interesting, but why would you want all 8 bands doing the same cut? Does this mean if i only use one or a few bands some low freqs will get through the eq? And is this a defect in EQ8 or common to all EQs?
Thanks.
D.

Kent_in_CO · Post by **Kent_in_CO** » Tue Oct 16, 2007 7:49 pm

I noticed some of that bleed-through when I looked at EQ'd parts with FreqAnalyst. Is this common with a lot of EQ's? Can this bleed-through have a noticeable impact on the mix?

Oh, and thanks Robert...that trick worked great to get a steeper curve.

Robert Henke · Post by **Robert Henke** » Tue Oct 16, 2007 8:27 pm

yes .this is true for almost all EQs.

the key word here is "slope". it defines how much amplitude change you'll get per octave. if a filter has a slope of 12dB/Octave it will cut the signal by 12dB per octave, just as the name implies. now if you cascade those filers, you will of course also cascade the loss, this is why the curve gets steeper.
However, most filters work a bit like resonators. they take a bit of the output signal and feed it back to its input. by doing so one can create pretty useful filters like bandpass, lowpass, notch,.... with only a few calculations. Most digital filters and most analog filters work like this. However, this feedack changes the behaviour in time. you can try this:

create a very short sound, a sound that is just a little "tic!" and apply EQ 8 with a lowpass filter with a very high Q. now tune this filter. you will hear a harmonic sound. if you re-record this sound and look at the waveform you will see that it is much longer then the impulse you put in. The more dramatic a filter filters, the longer usually the impulse response gets. So, with feedback filters you allways have a trade off between how much they cut out and how much they smear the signal in time.

There are other types of filters, so called FIR filters ( finite impulse repsonse ) that can create very dramatic filtering but they need a lot more CPU and introduce latency.

For the purpose of mixing often too much filtering is worse then filterin g only a bit.

R.

nebulae · Post by **nebulae** » Tue Oct 16, 2007 8:42 pm

Bassic Dave wrote:Thats interesting, but why would you want all 8 bands doing the same cut? Does this mean if i only use one or a few bands some low freqs will get through the eq? And is this a defect in EQ8 or common to all EQs?
Thanks.
D.

It's not a defect...as I said earlier, most EQs are like this because they have a set amount of slope. Thus, in order to increase the slope, with every filter you add on, the slope gets steeper. That's about the easiest way to explain it without getting technical...for which you'll have to turn to Henke or Tone Deft, or some other loveable geek.

nebulae · Post by **nebulae** » Tue Oct 16, 2007 8:46 pm

Robert Henke wrote: For the purpose of mixing often too much filtering is worse then filterin g only a bit.

Yeah, this is definitely true. In a mix, sometimes you just need a gentle slope to take down frequencies.

However, if you really need a very sharp filter, as in 384db/octave (yeah, you heard me!), check out Rubberfilter - it's free. http://www.kvraudio.com/get/2842.html

Nogi · Post by **Nogi** » Wed Oct 17, 2007 5:25 am

nebulae wrote: However, if you really need a very sharp filter, as in 384db/octave (yeah, you heard me!), check out Rubberfilter - it's free. http://www.kvraudio.com/get/2842.html

I tried this when it was first released into the wild and it crashed me 7000 times in 5 seconds. Is it better now? (I can't risk my currently stellar system stability - that said, I still want in the beta BAAAD!)