I'm doing some EQ tweaks, and throwing a HPF on the prochannel band of a bus, cutting off at about 80hz. I open up CW Analyzer in the FX chain, and a third-party analyzer (who will remain nameless) and I am seeing movement down below 80 and even 20hz still. What am i missing here??? I want a total cut of all activity below 70-80. Is there any way to do this with eq or is the 24db/octave what gets in the way?

swamp, just try this and see if it makes a difference. Right click on the menu bar of the eq in PC and select "post". See if your analyzer gives you more of what you were hoping for.

-Danny

Thanks Danny.  I think I'm just looking for a deeper cut.  I've tried prochannel eq, using post and running the analyzer in an output bus (because an analyzer won't reflect pceq in post position changes on a track) and also plopping the analyzer into an fx bin in the prochannel. I guess I just need to look around for a filter that gets the results I want.

swamptooth,

Notice how Sonitus HPF @ 80Hz and Q=1.0 slopes differently than FabFilter-Q Low Cut @ 80Hz and Q=1.0.


 

edit to add: I was practicing posting with my android.

The smile is all I could do... I can't type on it.

don't let the graphics confuse you, in other words

Slope is what it is all about. For instance in Studio One the filters that are on the extremes of the spectrum can be set to a whopping 48 dB/octave slope and that is steep and I mean steep.

But even for a slope of say 24 dB /octave if you set that to say even 100 Hz it means at 80 Hz you are still almost flat but at 50 Hz you are 24 dB down in level. Do you have any idea how soft a signal is that is 24 dB below the nominal level and at 25 Hz you are 48 dB down!

I would not be concerned too much about what activity you may be seeing on a spectrum analyser as it is much more important as to how it sounds. I have found from experience that actually less steep slopes can be very useful in taming an otherwise bass heavy track. Don't be in  hurry to necessarily use the steepest slope as it may end up sounding a little too thin. A more gradual slope however can often sound just right to my ears. That is what the slopes are there for.

I have found however where the steep slopes can be very helpful is in a mastering situation where sometimes you do want to take out the bass rather quickly. Danny will agree with this in that the LP64 is also very good for setting up good filter responses. Nothing to stop you either from setting up several nodes all around the same point and choosing a high Q for say 2 or 3 points and together they can make an almost brick wall slope between them. Be prepared to play around with that. You can also do that with the Sonitus too.

Even using the exact same settings, same slope and all, you may hear different results. I've noticed that the first time I was comparing a project I had opened in Studio One and Logic IIRC. The cut wasn't as clean using Logic's EQ and I could hear and see it.

When I inquired, the answer I was given by a Mod on the Logic forum was that Logic's cut off had "Butterworth characteristics, which are the most common type of filter configuration because it provides a good tradeoff between stopband rejection (what you want) and artifacts and phase shift (what you don't want).The response near the cutoff frequency has a much gentler knee,  meaning it doesn't actually meet the dialed in slope until well into the stopband.


Some of the high-end EQ plug-ins like the SSL X-EQ,  UAD Cambridge,  etc.,  let you select the type of filter design,  eg butterworth,  bessel,  elliptical,  gaussian,  chebyshev,  which gives you some flexibility in the rolloff knee. With Logic's EQ,  you can dial in several poles and tweak the Q control of the cutoff to nearly perfectly emulate those filter types,  or in between."


Mike McCue and Bitflipper (IIRC) who are both much more technically versed than I could ever be also elaborated on the subject here when I mentioned it and provide links and further info.

 

I don't know EQ math very well, I would definitely defer to Bitflipper or a handful of other folks that are members here for that sort of info.

I've asked Cakewalk several times to identify the filters they use but I've never gotten an answer.

I think it's really cool to learn that someone at Logic thought it would helpful for you to be able to learn about that.

From what I can gather, very few hi pass filters actually use a "Q" parameter and I have read speculations that the "Q" in Sonnitus may actually be some other parameter sitting in a box that has the "Q" label for use with the other types of filters. How ever, I did once find a vague reference to a hi pass filter, in an article some where, that did have a Q factor but I can't recall its name. I asked Cakewalk about it, but it will not answer the question... I have been told repeatedly to just listen with my ears.

Traditionally "Q" has a specific definition throughout engineering and it is applied to the width of a sinusoidal signal and describes the width of the sine wave at the RMS level, which is 3dB from the peak. It seems like a lot of dsp guis don't match the Q labels with the actual "Q". It can get really confusing because you hear descriptions like modern or vintage etc. which really just refer to the fact that it was very hard to make an analog circuit that didn't deviate from the math a bit and so the labeling on old analog gear wasn't 100% accurate. Now with digital it seems like designers mis label the Q on purpose when they are emulating the old analog gear. It's not easy to keep track of and it would be helpful to know what actual filters are being emulated.




In the mean time here is one way to look at it.

If you use Sonitus or any other EQ to set up a multi-order or multi-pole example of what ever filter it has... you can have some extra control over the slope. Many filter names are associated with multi-order math and it is simply implicit in the naming.

Take a look at this as an example:





I wish I knew more about EQ filters and the math but it requires specific study and a real good teacher.

I'm still looking for that teacher. :-) Hopefully someday I'll get there.


all the best,
mike

swamp, I just tried the PC eq with the slope on the 48 dB setting high passing 80 Hz with 2 good analyzers in the bin on that track...and I'm getting 0 audible activity below where I'm cutting. I see some activity in the 50 Hz range, but it's registering at around -40 dB and doesn't even spike. From 40 Hz on down I'm seeing slight activity at -80 dB and below with nothing spiking at all.

Did you mess with the slope knob at all? I know that's probably a dumb question, but I just thought I'd ask. LOL! :) I'm using both the Waves PAZ analyzer and the Blue Cat multi freq. They are both reading the exact same thing. The Cake analyzer is showing the exact same thing also. Make sure you use "Hi" mode or it will report low end activity due to the lack of resolution and less cpu usage. In high mode, the Cake analyzer should give you numbers like the ones I'm seeing. But if you leave it in lo mode, you see more low end activity. Lo means lo res, not "low end". I know...you knew that already...just mentioning it in case you didn't. :)

If your analyzer has a resolution mode, you want to adjust that. For example, if I leave the Waves PAZ set for its default at 40 Hz, the graph doesn't slope in as tight and makes it look like I have more activity going on in the bottom end than I actually have. I use either the 20 Hz setting or the 10 Hz setting so it gives me a little more resolution. The same with the Blue Cat....it works in numbers from 1 to 10. I like 8 or 9. Anything under 8, and the low end has a little more activity because the resolution isn't as intense. Just some food for thought for you when viewing these analyzers. :)