Filter slopes question

Have an urge to learn, or a calling to teach? Want to share some useful Youtube videos? Do it here!
User avatar
Jagwah
Posts: 1514
Joined: 16 Jan 2015

Post 29 Aug 2019

I'm just curious about aggressive or steep slopes and why they are less common in filters.

Mixing wise, to have a super steep slope like Omega Trimmer's 192 dB/oct seems like it could be quite useful, like it could more easily create a tight mix with elements right next to each other with super steep slopes. Why do I rarely see it being used, and rarely see it in typical filters?

User avatar
Loque
Posts: 5783
Joined: 28 Dec 2015

Post 29 Aug 2019

I am not the uber pro here, but soft slopes sound more musical, higher slopes change the sound more dramatic and make them often very weak. Very high slopes make only sense, if you really want to cut everything out and accept the sound change they introduce and the ultra high slope cuts does not make much sense, since they are below the noise floor which are audible or could be represented in CD quality. I use the very high slopes, if i really do not want any influence of a noise floor in the final sound, mainly if i have further treatment with the level, like distortion, wave shapers and other stuff.

My final conclusion today is, you do not really need them, but its good to have the choice which satisfies your feeling that you can completely remove something from a sound.
:reason: 10, Win10 64Bit.

PhillipOrdonez
Posts: 509
Joined: 20 Oct 2017
Location: Colombia

Post 29 Aug 2019

Steep slopes can cause issues, affect phase and sound strange. Many EQs aren't capable of applying clean processing when the slope is too high. When things sound too technical they appear sterile and lifeless.

RobC
Posts: 1118
Joined: 10 Mar 2018

Post 29 Aug 2019

IIR filters, as far as I've heard, and if understood correctly, can't make a perfect cut. It has a crappy artificial sound to it, like you can hear in many modern Deep House songs. Bolier & Trobi - In The Water is a "good" example for that awful side effect.

I suspect it's caused by the fact, that while it does a somewhat sharp cut at some point, it unfortunately boosts close to it. One video showed on a graph, that it kind of looks like a sine wave.

Another solution is using Linear Phase filters/EQ. Drawback is pre-echoing, and latency. The best use I found for that, is splitting and rendering multiple bands, then creating a brick wall graphical EQ - only, you can solo, or mute any bands. On the same level, there are no audible artifacts. Once you start moving the faders big time, you may notice more and more ringing, chirping, pre-echoing, etc. Wonderful for subtle changes, though - or as long as you find it tolerable. Still more clean than IIR filters.

If you design one shot sounds, you can take care of the pre echoing, but forget about busy loops and songs.

For low cut and high frequency roll off, it's more than perfect. Some light lower sub bass pre-echoing won't be bothersome, or really noticeable, and even a 15 kHz roll off isn't noticeable at all in my experience.

User avatar
Jagwah
Posts: 1514
Joined: 16 Jan 2015

Post 31 Aug 2019

Thanks for sharing your wisdom guys, much appreciated. I will avoid super steep slopes in the future and might look in to Linear phase EQ, cheers!!

User avatar
selig
Moderator
Posts: 7902
Joined: 15 Jan 2015

Post 31 Aug 2019

One issue not mentioned is that practically no sound has frequencies limited to a narrow range, except for a sine wave (and that's assuming the sine wave doesn't play a melody that moves around!). Splitting a mix into narrow bands will give you a mix of everything that falls in that range, which will often be a little bit of every instrument in the mix, and this will change as instrument play different notes etc.

One instrument may dominate, so you might find this approach useful to extract a trigger signal, but nothing useful for a mix situation in most cases.
Selig Audio, LLC

RobC
Posts: 1118
Joined: 10 Mar 2018

Post 31 Aug 2019

selig wrote:
31 Aug 2019
One issue not mentioned is that practically no sound has frequencies limited to a narrow range, except for a sine wave (and that's assuming the sine wave doesn't play a melody that moves around!). Splitting a mix into narrow bands will give you a mix of everything that falls in that range, which will often be a little bit of every instrument in the mix, and this will change as instrument play different notes etc.

One instrument may dominate, so you might find this approach useful to extract a trigger signal, but nothing useful for a mix situation in most cases.
For not-one-shot-sounds, the only way would be perhaps with multi-band (maybe parallel) compression, and then nudging the faders of each band.

Edit: ...also, multi-band gating for reducing background sounds.

User avatar
Jagwah
Posts: 1514
Joined: 16 Jan 2015

Post 01 Sep 2019

selig wrote:
31 Aug 2019
One issue not mentioned is that practically no sound has frequencies limited to a narrow range, except for a sine wave (and that's assuming the sine wave doesn't play a melody that moves around!). Splitting a mix into narrow bands will give you a mix of everything that falls in that range, which will often be a little bit of every instrument in the mix, and this will change as instrument play different notes etc.

One instrument may dominate, so you might find this approach useful to extract a trigger signal, but nothing useful for a mix situation in most cases.
Thanks Giles I was hoping you might chime in. It is a weird one for sure, I saw a producer using it on a pad that wasn't intented to move up and down the pitch register too much. I take it it's best to typically avoid, but I will play around with it anyway just to undersatand it a bit better.

RobC
Posts: 1118
Joined: 10 Mar 2018

Post 01 Sep 2019

In case of a good, expressive, dynamic rap performance (not the mumbling crap category - I can't stand most raps), I would perhaps want to keep most frequencies on the spectrum.

You can't really avoid at least some kind of background sound (I intentionally don't say 'noise' this time) leaking, such as sub bass that mostly vehicles create. But since I do want to keep some of those sub bass punches that a good rapper has during performance - and I don't want to just do an 80 Hz low cut <- for that, the easiest solution is to gate such frequency bands individually. Anywhere from 1 band per octave, or per 1/3rd octave. So, it's good for reducing/removing low level frequencies that are still loud enough to be bothersome.

It's not about avoiding tools, but using them when appropriate, as well as appropriately.

User avatar
Jagwah
Posts: 1514
Joined: 16 Jan 2015

Post 02 Sep 2019

RobC wrote:
01 Sep 2019
In case of a good, expressive, dynamic rap performance (not the mumbling crap category - I can't stand most raps), I would perhaps want to keep most frequencies on the spectrum.

You can't really avoid at least some kind of background sound (I intentionally don't say 'noise' this time) leaking, such as sub bass that mostly vehicles create. But since I do want to keep some of those sub bass punches that a good rapper has during performance - and I don't want to just do an 80 Hz low cut <- for that, the easiest solution is to gate such frequency bands individually. Anywhere from 1 band per octave, or per 1/3rd octave. So, it's good for reducing/removing low level frequencies that are still loud enough to be bothersome.

It's not about avoiding tools, but using them when appropriate, as well as appropriately.
Thanks for your input, much appreciated.

  • Information
  • Who is online

    Users browsing this forum: CommonCrawl [Bot] and 0 guests