Sweep technique to cut nasty frequencies. Should I be cutting the "whistle" frequencies?

[BillMurray]

Hi there,

It seems most mixing tutorials I watch, at some point, they will talk about sweeping a given frequency range with a narrow band eq to identify harsh frequencies and then perform a narrow band cut where the undesirable frequency was found.

All I hear when I do this is a whistling type sound. Is it that whistle sound that I should be listening out for? And should I cut those frequencies? If so, then I'd cut the entire frequency spectrum! Because when I boost a narrow eq to do a sweep, I hear that whistling sound constantly, regardless of where I am on the frequency spectrum.

What am I not understanding here?

Thanks for reading.

 

[Lorcan Ward]

That whistling sound is cause you've pushed the volume of that particular frequency way above the mix so it cuts through like that. I've watched guys do it in a studio and its more down to experience and knowing what little undesirable frequencies can be cut rather than specifically searching, as well as making room for other instruments. I prefer to work the other way and sweep the opposite direction until I notice a fizz or hiss disappears like around 4k. Try A/Bing your cuts in a mix and you'll find you end up adding a lot back in.

 

[DudeManBrother]

You should only do it if you hear a particular frequency standing out. Ex. if you hear an obnoxious ping after every snare hit: you’ll memorize the pitch of that ping, and sweep until you find the correct pitch, then play the mix to determine how much needs to be cut. Or setting a wider Q to find the resonant buildup in the bass and kick, so you can reduce one, and leave room for the other.

Don’t just sweep arbitrarily on every instrument; only when your hearing something in the mix that is standing out. Then be sure to bypass the change to make sure you’re happy with it. Too much surgical EQing can really neuter your sound in a hurry. Some of that nastiness is what we love about certain tones.

 

[NickLAudio]

I look at the VU meter while sweeping a narrow boost on an EQ for cutting purposes. Yes, the whole frequency range will "whistle" but more often than not, the "harsh" frequencies will either spike the meter or visually get significantly louder. Once I find the visually loudest spiked frequency, I then make the cut there, of course checking it against the full mix as I go along. This surgical EQ technique I reserve mostly for sounds with a lot of complex frequency data like distorted guitars and the like. It also works wonders for video editing where you're trying to take out specific background noise such as heating/ac units, fans, or even the natural hiss from gain/headroom on shotgun mics.

Not my pic below but minus the slight boosts here and there, a general surgical EQ looks relatively close to something like this in the end...
.

 

[TedEH]

I definitely do that kind of thing to cut out snare ringing and stuff like that. The whistle is what you're listening for, but it's about searching for the places where the whistle gets harsh, or matches the pitch of what you want to remove, etc. A lot of times you know you hit it when the whistle suddenly gets louder or more obnoxious. I find that a lot of times I have to put those cuts on the ringing frequency of a snare at different octaves and take a little bit away from each. And, as was said, you can usually visually see it in the tool you're using - it'll be where there's a big spike or something.

 

[Seybsnilksz]

I do it mostly in drum overhead and room tracks. As you say, you could do it forever as every frequency makes a whistle when you sweep with a narrow Q. I try to only do it when I hear it without searching for it, and I try not to do too many. Sometimes ring can make a track stand out more in the context of the whole mix, even if it sounds ugly in solo. Very common with snare drums.

 

[GunpointMetal]

The biggest thing is don't go looking for bad frequencies until you hear them, or they pop out in the mix. I've seen a lot of people starting out try and "fix" everything before they even have a basic balance set up, and that's just gonna lead to more frustration later. If your guitar sounds harsh, find the frequency that's harsh and pull it down, if it sounds fine, don't worry about it.

 

[DudeManBrother]

Being that this question looks pretty well answered, I’ll share a cool little trick I recently learned. It works like a champ with Fabfilter Pro-Q3; but I bet it could be achieved with any EQ plugin that has an EQ match function:

Insert Pro-Q3 onto Guitar 1 and side chain to Guitar 2, or insert Pro-Q3 in there as well. EQ match 2 onto 1 (or vise versa).

Choose a number of EQ points (it can range from 0-24). Select them all and use the Invert option.

Next, reduce the level so it’s not too drastic.

You can copy this EQ onto the original track, and invert it again to subtly emphasize the impact.

Finally, check each EQ point on/off in the mix and adjust/remove whatever’s necessary.

What’s cool about this is it sculpts out a little bit of room the other guitar occupies, and fills in the spots that guitar lacks. It’s a cool way to create some separation and extra stereo width.

 

[pablometal]

Subscribed

 

[newamerikangospel]

I actually use a narrow Q and pull it down -6db and sweep it around that way. It will sound like a phaser when solo'ed but when it gets to a point of sounding "natural" or the sound settles in, I generally find a node is clouding something right there. If I have something I cant find, I will boost and sweep, but I generally always keep it at +3db max. That sound be enough to pop the sound out without making everything sound like an issue (because everything sounds annoying with a narrow Q and boosted 20db).

 

[nightlight]

Great tips here. I'm guilty of always following a formula when it comes to
EQing my guitars, especially looking for fizz at around 3.6-4khz. Didn't realise that was a silly way to approach it.

 

[KingAenarion]

I teach a lot of Uni students audio, and the rules I've developed are such:

1) The first priorities you should be removing are non-musical/non-content related sounds.

For example, in a guitar, things below 80-150Hz aren't really musical pitch material on the instrument unless tuned low, and even then the bass guitar occupies that space. Things about 8-10kHz are also not really naturally present in a guitar amplifier's speakers as they don't have tweeters. There are usually weird resonances and whatnot due to distortion that aren't pitch material at all, so they can go, same with whistling sounds.

2) Always listen to something in context first, and always reference your changes. In isolation and boosted or soloed, that whistling sound might not sound good, but in context it might be what gives your guitar bite or presence.

3) Always cut over boosting, your gain structure will thank you.

4) There is rarely, if ever, need to make dozens of cuts. If you're having to do this to something to make it fit, rethink the sounds you're working with.

 

[pablometal]

what i do is to scoop al the whistling sounds and make the guitar sound like shit, then i start to decrease the gain scale/mix/wet idk (in fabfilter) until it sounds good and then i finish removing stuff but removing too much makes the guitar sound really bad, is important to keep a natural sound

 

[Drew]

I teach a lot of Uni students audio, and the rules I've developed are such:

1) The first priorities you should be removing are non-musical/non-content related sounds.

For example, in a guitar, things below 80-150Hz aren't really musical pitch material on the instrument unless tuned low, and even then the bass guitar occupies that space. Things about 8-10kHz are also not really naturally present in a guitar amplifier's speakers as they don't have tweeters. There are usually weird resonances and whatnot due to distortion that aren't pitch material at all, so they can go, same with whistling sounds.

2) Always listen to something in context first, and always reference your changes. In isolation and boosted or soloed, that whistling sound might not sound good, but in context it might be what gives your guitar bite or presence.

3) Always cut over boosting, your gain structure will thank you.

4) There is rarely, if ever, need to make dozens of cuts. If you're having to do this to something to make it fit, rethink the sounds you're working with.

Excellent post.

 

[isotropy]

Pro Q looks so pretty.

...that is all.

 

[thrashinbatman]

Pro Q looks so pretty.

...that is all.

I love ProQ. The "isolate frequency" feature or whatever it's called is great. It makes sweeping pretty much unnecessary. The way they do their analyzer is also really good; it makes it really easy to find potential problems. It still requires you to have a good ear and know what all to cut out, but it takes a lot of the leg work out of the process.

 

[isotropy]

I love ProQ. The "isolate frequency" feature or whatever it's called is great. It makes sweeping pretty much unnecessary. The way they do their analyzer is also really good; it makes it really easy to find potential problems. It still requires you to have a good ear and know what all to cut out, but it takes a lot of the leg work out of the process.

It's the only plugin I want to get at this point.

 

[thrashinbatman]

It's the only plugin I want to get at this point.

It's become my "desert island" plugin. I use other EQs for my big, sound sculpting moves, but the ProQ is on almost every track of mine nowadays. It's really changed the game for me, particularly by making me realize how much ugliness I was letting slip by in my mixes prior.


I'm really selling this thing, but it's genuine excitement. I really like what it's capable of.

 

[hensh!n]

One thing I’ve learned as a guitar player...is it’s often the “nasty frequencies” that actually allow your tone to cut through a mix. Too smooth of a tone may sound great on its own, but muddy in context. At least as far as rock/metal is concerned.

 

[Drew]

I guess the only other thing I can contribute here aside from saying yet again to go reread @KingAenarion 's post...

...is that the "sweep a narrow Q boost, find something that sounds bad, and turn it down" is a process that it's VERY easy to overdo. The problem here is with a narrow enough band and a strong enough boost, pretty much anything sounds bad. If you're going to go down this road, I guess my suggestion would be to use two EQs - one for the broader tonal shaping stuff (honestly, I tend to want more of a "color" EQ for this anyway, I really like the Satson Burnley73 for the saturation it can impart and because it's tough to make it NOT sound musical, but even if you're using something pretty transparent like ReaEQ definitely use two instances) where you're getting rid of stff outside of the active frequency range of an instrument (i.e, the 80-150hz and 8k+ stuff on guitars he mentions). Then, use a separate instance (or a very transparent EQ if you're using different plugins) to sweep and cut bands...

...And then, when everthing's said and done, if you have three or four notches, pull up the rest of the mix and play the track back in the context of the full mix, and flip your second EQ in and out of bypass, alternately listening to the instrument with the bands bypassed and active. Unless you're sure the mix is better with those notches in, then strongly consider leaving them bypassed. It's not hard to quickly neuter a mix by going too heavy on this stuff.