Please ensure Javascript is enabled for purposes of website accessibility Jump to content

Simple question about EQ/Cut parameter priority order


themetallikid
 Share

Recommended Posts

When you have an EQ that is paired High/Lo cuts....which of those is processed first...I'd assume the cuts are, then the frequency boost/cuts. 

 

Its a trivial question, but my mind needs to know how it works. 

 

 

This stems from me liking a sharp spike at 125hz for palm mutes, but adding a low cut up to 150hz.  If the cut is processed afterwards, is it defeating the boost at 125hz?

Link to comment
Share on other sites

When you have an EQ that is paired High/Lo cuts....which of those is processed first...I'd assume the cuts are, then the frequency boost/cuts. 

 

Its a trivial question, but my mind needs to know how it works. 

 

 

This stems from me liking a sharp spike at 125hz for palm mutes, but adding a low cut up to 150hz.  If the cut is processed afterwards, is it defeating the boost at 125hz?

It's all dependent on where you are doing it in your signal chain. The first thing in your signal chain is what happens first. 

Link to comment
Share on other sites

Well....thanks, I guess......but that's not what I was asking. 


 


I was asking that within the EQ block only (not related to signal chain placement), there are frequency parameters (Parametric block is what I'm picturing in my head) for Low/Mid/High Frequencies, and then also LOW/HI Cut.  I assume the LOW/Hi Cut is happening first, and then the LOW/MID/HIGH boost/cuts are happening??  As opposed to me boosting the 125hz I mentioned originally, and then it going through a 150hz Low Cut and defeating the boost at 125hz. 


Link to comment
Share on other sites

When you have an EQ that is paired High/Lo cuts....which of those is processed first...I'd assume the cuts are, then the frequency boost/cuts.

 

Its a trivial question, but my mind needs to know how it works.

 

 

This stems from me liking a sharp spike at 125hz for palm mutes, but adding a low cut up to 150hz. If the cut is processed afterwards, is it defeating the boost at 125hz?

Can't say I know the answer, but it seems to me that having the info won't help you one way or the other, in this case. Modeling is weird. The "why's" get rather nebulous.

 

You've been very specific about settings that you like...you've already won. What difference does it make what's happening first if the desired result is achieved? Even if you've guessed wrong, are you gonna mess with settings that are working?

Link to comment
Share on other sites

A cut one place results effectively in boosts around the cut. But they’re not the same. Generally you cut narrow (higher Q) and boost wide as boosts trend to have more perceived impact than cuts. Cutting what you don’t want instead of boosting what you do want will both sound more natural, and improve downstream headroom.

 

If you have any distortion blocks in the signal path, then generally cutting bass and increasing treble before the distortion, while boosting bass and cutting treble after distortion will provide a better overall tone. Balancing these cuts and boosts can help keep a natural tone when you turn the guitar volume down and the distortion cleans up. Cutting bass before distortion reduces mud, cutting treble after distortion reduces ice-pick and fizz.

Link to comment
Share on other sites

Hmmm, while I get that theoretically you want to Know How It Works, me too, and in general the order of processing matters a ton, in this case, within a single EQ block, I doubt very much it matters, and here's why I think that...

 

Say you have EQ > drive > EQ. Cutting lows before the drive and boosting them after is hugely different than the other way around. Less lows into the drive is tighter, with some compensation for the missing lows in the post-drive bass boost, where they don't make a mess. Other way round, you get flub in the drive, then pull out the lows that might possibly have made that make some emotional sense. Yuck in general, with the caveat that whatever works for you in your context, does.

 

The big thing that's really different about that than 2 bands in a single EQ block is that there are no nonlinearities like the drive block going on. IMO that basically makes this a demo of the transitive property of non-overlapping linear EQ bands -- doesn't matter what order they're in.

 

Of course if you're running hot enough to clip inside the EQ, then that's not the case, but most likely Just Don't, it won't be pretty.

Link to comment
Share on other sites

 

Well....thanks, I guess......but that's not what I was asking. 

 

I was asking that within the EQ block only (not related to signal chain placement), there are frequency parameters (Parametric block is what I'm picturing in my head) for Low/Mid/High Frequencies, and then also LOW/HI Cut.  I assume the LOW/Hi Cut is happening first, and then the LOW/MID/HIGH boost/cuts are happening??  As opposed to me boosting the 125hz I mentioned originally, and then it going through a 150hz Low Cut and defeating the boost at 125hz. 

 

I see what you are asking now...lol. 

 

What you can do is play around with that idea with the global EQ open in editor. It has a very nice visual representation of the real time frequency curve. Play around with the settings to see how the hi/lo "ramps" and the mid frequency curves interact with each other. 

 

My SWAG is that the signal is being processed in a parallel fashion so that none of the frequencies are necessarily being processed in a chronological order...

Link to comment
Share on other sites

I can't imagine how the sequence of processing comes into play any more than it does on any hardware effect.  The signal enters the block from the signal chain and the same electronic circuit transformations take place to that signal but done mathematically before it gets passed back to the signal chain.  The only thing the settings affect are the variables in the math which have been determined ahead of time by the configuration.  The same as it would be were you to turn the adjustment knob on any hardware effect.  If you're not concerned about the sequence of processing in a hardware circuit, you shouldn't need to be concerned with it here.  Especially since we're talking about EQ which is the simplest of all effects processing-wise.

Link to comment
Share on other sites

Hmmm, while I get that theoretically you want to Know How It Works, me too, and in general the order of processing matters a ton, in this case, within a single EQ block, I doubt very much it matters, and here's why I think that...

 

Say you have EQ > drive > EQ. Cutting lows before the drive and boosting them after is hugely different than the other way around. Less lows into the drive is tighter, with some compensation for the missing lows in the post-drive bass boost, where they don't make a mess. Other way round, you get flub in the drive, then pull out the lows that might possibly have made that make some emotional sense. Yuck in general, with the caveat that whatever works for you in your context, does.

 

The big thing that's really different about that than 2 bands in a single EQ block is that there are no nonlinearities like the drive block going on. IMO that basically makes this a demo of the transitive property of non-overlapping linear EQ bands -- doesn't matter what order they're in.

 

Of course if you're running hot enough to clip inside the EQ, then that's not the case, but most likely Just Don't, it won't be pretty.

This. If you are staying within the linear limits (not too loud that you're clipping, neither too soft that there is signal degradation or gating), there should be no difference.

 

Since the OP is asking about the EQ processes within the same block, assuming the above prereq is true, there should be negligible difference.

 

You can test it out with this signal chain:

EQ1>EQ2>EQ1-duplicate

Set footswitch to toggle EQ1 off when EQ1-duplicate is toggled on.

 

In the case where there's a non-linear block in between, e.g. amp, distortion pedal, noise gate, etc, then see zooey's reply.

Link to comment
Share on other sites

A general rule of thumb is to cut first, add later.

I, for instance, cut the lows either in amp or right after and bring it up again later. This is how it usually is done in studios.

Cut first, add later.

 

That's not what he was asking. He's asking which parameter gets processed first in a block when a signal goes through a single EQ block, the high parameters or the low parameters.

Link to comment
Share on other sites

I understand the answers provided, lol....I guess it was just a random question...in my head....I would tend to think the HI/LO cuts get processed first, and then any shaping of the frequencies happens, it would make less sense in my head to shape frequencies, and then put the LO/HI Filters on after and affect the previous changes. 

Link to comment
Share on other sites

It's an interesting question. There are enough potential variables in the following test that the results would almost certainly be inaccurate but all other things being equal and just as a "thought experiment" this would be kind of interesting. I just threw up in my mouth a little saying the phrase "thought experiment".

 

I suppose one way you could test cut precedence within a single parametric block would be to put a single parametric with both your selected parametric band cut and the high/low cut on Path 1. On Path 2 place two parametric blocks next to each other with one having only the exact same high/low cut as the parametric band cut you put on Path 1 and the other having exactly the same parametric high/low cut. Then solo the outputs from Path 1 and Path 2 and compare them. Now swap the positions of the two parametric blocks on Path 2 and compare them with Path 1 again. If the signal on Path 2 on only one of the two swapped positions in Path 2 matched the signal on Path 1 you could theoretically infer the precedence of the signal path on Path 1 through a single parametric block. If neither swapped position of the two parametric block on Path 2 was the same you could infer that the single parametric on Path 1 was processing the signal to both the band cuts and the high/low cuts simultaneously. If both swapped positions on Path 2 were identical you could infer that either you did not have enough of a cut dialed up to make an audible difference or it just didn't matter.  ;)

 

However, have to stress this as only a thought experiment under ideal conditions due to real world variables that would probably invalidate the results such as for example the two paths on the Helix having inherently slightly different sounds. I suppose you could get around that problem by only using Path 1 for the test with three parametric blocks - one parametric for the single block test and the other two parametrics for your swap test comparison. And there is the fact that even if there is a precedence order to which cut is processed first within the single parametric block on Path 1 the signal might take longer to pass between the two separate parametric blocks on Path 2 and be subject to some modification in the signal path as it was passed from one block to the other, rendering your comparison invalid. I am sure there are probably a ton of other potential variables that I haven't considered that would render this test null and void. I'm sure someone else is thinking right now, they could improve this test with phase inversion and see if the two approaches cancel each other out. I can't believe I just typed this inane idea out, I should delete it immediately. Or maybe Line6 could just tell us how parametric precedence or lack thereof works. Sheesh, I need a hobby, oh that's right, I already have one, I mean a different hobby.  :)

Link to comment
Share on other sites

For the curious check out http://mysite.mweb.co.za/residents/cyb00746/audiodocs/Digital_EQ.htm

If we assume the filter transforms happen in a floating point or high resolution fixed point domain it probably does not matter much in what sequence they are executed and the outcome is practically the same - just guessing ;-)

Link to comment
Share on other sites

Theoretical analysis is nice of you're into that...but I have a very simple checklist:

 

1. Sounds good--->Done.

2. Sounds like crap ---> Not done. Turn some knobs.

Yeah, ain't nobody got time for Fourier Transforms in the digital realm. 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...