A word on aggressive HPF/LPF/EQ

[Note: This is taken from an email conversation with a customer, with their permission]

Ok so.

Going by everything you have said, you are trying to chop the bottom/top off your sound like an absolute knife, to create the sharpest possible cutoff, to remove as much ‘excess’ as possible. I’ll explain why this is a bad idea.

Simply put, eq and filters work by using phase shift to get things to add up/cancel out, so you add or subtract gain at a given frequency.

What’s phase shift? The simplest way to think about it is shifting a sine wave in time, or, left to right/right to left on your screen. Right? If you have a sine wave that fills your screen and you change the phase, you are effectively moving it left or right, so it starts at a different point in the wavecycle.

Ok so. Go here, and look at graphs 5 and 6.

The ‘power vs freq’ graph shows the gain vs frequency – this is what your graphic display on the eq shows you.

The ‘phase vs freq’ shows you what happens to the phase when using the filters shows in the power graph. This is what your graphic eq display doesn’t show you.

(This is not a fault with Renaissance eq, this is how eq works.)

So WTF does this mean? Well, when the phase graph is above or below the line, you’re getting phase shift. You can see that different frequencies are phase-shifted by different amounts, less so as you move up above the cutoff frequency (100Hz in this graph). But the key things here are that

1) things are shifted a long way away from the cutoff frequency

2) different frequencies are shifted by different amounts.

Now, the crazy stuff below the cutoff is where the phase ‘wraps’, so it goes one whole cycle out of phase and therefore ‘comes back around’, so to speak – think about moving the wave on your screen again. You can see, hopefully, that the green line – the softest curve on the power graph – ‘loops’ much more smoothly on the phase graph than the white line, which is the sharpest filter and looks… well, pretty messed up on the phase graph. This demonstrates point 3;

3) the sharper the filter/eq band, the more twisted the phase response

So, now you put all three of those together and look at a real-world application. Ok, a hypothetical real-world application. Let’s say a nice sharp, tight snare drum. Let’s say the fundamental of the snare is at 250Hz (or whatever, not important)

It should be clear that even high-passing at 100Hz – which is ‘pointless’ because it isn’t touching the sound, right? – is going to mess with the sound, because of point 1). That graph shows that there’s phase shift happening up to and beyond 1kHz, miles away from 100Hz. What is this phase shift doing to the snare drum?

Remember the wave ‘sliding’ left and right. In our nice tight theoretical snare drum, all the frequencies present start at exactly the same time, when the stick hits it. ‘POP!’ But what happens when all the constituent frequencies are moved in time by different amounts? They are smeared out in time; they don’t all start at the same time any more. ‘Fffffooopp’. This is often called phase smear, which should hopefully make sense now.

What happens when you use a sharper filter, or add another filter to make it even sharper? More drastic phase distortion/smear (because of point 3), meaning more smear, meaning fffffoooopppp. Our nice tight snare drum – despite its frequency response being essentially unchanged (remember we are high passing below the snare’s fundamental) – is now smeared, and will sound flatter than it did before, because all the constituent parts aren’t hitting at the same time any more. This is what I hear in your drums. This is why they sound lifeless.

All of this applies to low pass filters too, and to normal eq bands (where it occurs above and below the selected freq). Add aggressive HPF to aggressive LPF, along with several heavy bands of sharp bell eq, and it’s a recipe for flat, lifeless drums.

There are more things to consider – ringing and so on, but for now that should explain a lot.

What to do instead? Use gentler cuts. 12dB/oct, 18dB/oct, 24db/oct are plenty. Remember that because they are gentler, you can go further because they aren’t cutting so hard. Linear phase is another option, but that has a whole different set of side effects to consider. Linear phase isn’t a bad choice for high pass filtering though, but as always, it depends.

Bottom line: The more aggressively you do stuff, the more damage you do. Be mindful of mixing with your eyes; ‘sharper must be better’ etc.

________________

  1. #1 by woe84 on October 25, 2014 - 7:55 pm

    hey bob

    thank you, this is very useful for me…greetz from hamburg northgermany, woe

  2. #2 by jj soderling on October 26, 2014 - 3:20 am

    So the more f**kery, the crappier the end result. Got it 🙂

    • #3 by bobmaccsblog on October 26, 2014 - 7:11 am

      You basically just saved me writing any more blog posts, ever again 😉

  3. #4 by Profane on October 26, 2014 - 1:27 pm

    Learned something new. Thanks. I’ll keep an eye out for that.

  4. #5 by duffah on October 29, 2014 - 2:04 pm

    Out of interest, what was the slope being applied to cause the smear?

    • #6 by bobmaccsblog on October 29, 2014 - 6:09 pm

      All eq’s cause smear, and the more aggressive the slope/sharper the filter, the more it smears it. This was the most aggressive slope you can get with REQ, at the top and the bottom, applied several times over (multiple eqs all doing the same aggressive filtering). So you can imagine. You almost end up with an all-pass filter.

      I’ll do a follow up to this article at some point, including pix, as we had another episode recently…

  5. #7 by robin allinson on January 24, 2015 - 1:59 am

    Hi Bob, great bit of info that often gets overlooked as people add more and more on to the chain to get “that sound” 🙂 keep them coming. Am enjoying reading it all so far .

  6. #8 by jw on February 9, 2015 - 2:21 am

    great read…..is there an au plugin that i can use in logic to keep an eye on this phase response whilst i’m working in logic?

    • #9 by bobmaccsblog on February 9, 2015 - 8:04 am

      Please don’t take this the wrong way, but…. Are you joking? Genuine question. Even if there is such a plugin, I have no idea what you’d do with the information. Just remember that the phase response is tied to the amplitude/freq response, and don’t do loads of sharp/aggressive eq all over the place. It’s as simple as that. You don’t need a plugin to tell you not to do daft stuff.

  7. #10 by jw on February 9, 2015 - 10:48 am

    no the reason i ask is because there is such a tool on pc as i’ve seen the guy who works for pro tools expert/fab filter use it and i was wondering if there was a mac version. i didn’t mean literally for using it in mixing situations.

    • #11 by bobmaccsblog on February 9, 2015 - 11:05 am

      Ah right. I think spectrafoo is the standard, but have no idea if it’s available in plugin form. Sorry for the misunderstanding!

  1. Aggressive and Over EQ’ing | DdashFENS
  2. Monoing everything below whateverHz on your master buss | Bob Macc's Blog

Leave a reply to jj soderling Cancel reply