Sweeping for “problematic frequencies”

frank1985

Active member
I don’t know if this should be in the newbie subforum, but one piece of advice I’m always coming across with regards to eq is sweeping for “problematic frequencies” with a high Q bell curve, boosted at around 10db or so. How do I know exactly that a certain frequency is problematic, as when boosted to that extent, all frequencies sound harsh and problematic to me :p I don’t quite know what I’m looking for...
 
The thing they either never mention or don't stress enough in these explanations is that you aren't actually looking for problems, you should already hear the problem and be looking for "where" in the frequency spectrum to make your cuts. In other words , say you are working on a complete song/mix and the kick drum sounds "boxy" which is a weird term that i dont really feel describes what most people mean. Anyway, something in the lower mids of the kick is sticking out and bothering you or making it sound muffled/muddy/flat or what have you. You can now take a wide Q boost and sweep it until you find the general area of the timbre that is bothersome(which will often be a "resonant peak" then gradually narrow and sweep until you locate the problem area. There is also a practice of using sweeping to find the resonant peak/s of a track( these frequencies will usually be obviously louder than non resonant freq when sweeping) and cutting these frequencies to even out a performance. This can be especially useful on bass guitar tracks since different notes tend to sound louder than others even at the same relative rms db. The easiest way to know if you have a problem is to use reference tracks in the genre that you are working in and comparing the tonality/timbre. This will help you decide what you like/dislike about your recording and what to mess with.
 
you aren't actually looking for problems, you should already hear the problem and be looking for "where" in the frequency spectrum to make your cuts.

This ^^

You have already discovered that a pre-emptive approach reveals nothing all that helpful.

Frequency sweeping is a technique to pinpoint and identify a frequency that you have already found to be a problem.
 
Unless your room is very well tuned, you may be hearing a room problem not a mix problem (check with headphones?)
 
I never boost, I do cut sweep. In fact I rarely, if ever, boost any freq, period. And while there was a time I would use steep cut sweeps, I discovered that just because I needed my ears to have a drastic comparison but my ears are better now so My depth/width isnt so dramatic.
 
I think that if you are a beginner, you should start by trying to fix only problems that you can clearly hear, like they're so obvious that you don't need to train yourself to hear them.

In the meanwhile it's a good idea to watch videos and read articles and educate yourself on more complex stuff.

And about the boosting thing. I know it's much easier to hear something if you boost it, but I'd suggest to try learning to fix problems by cutting. You identify the problem, then you take a wide Q cut and when you can't hear the problem anymore, it's likely that you've fixed it.

It'd actually be interesting what people have to say about how to educate yourself on this.
 
First I'd think for searching a moderately narrow Q rather than wide. The search' by cutting is an interesting idea that may have some merit! It goes directly to the end step perhaps.
I tend to think 'boost and sweep 'highlights better -perhaps.
I think the process can also temporarily skew our perception of tonal balance. Therefor; having found the target, leave the filter out for a bit, then begin the cut.
 
The recent EQ installment in the iZotope mastering "are you listening" series briefly covered this. The presenter was not a fan of sweeping to look for problems because he felt that everything would sound like a problem when you stick a narrow boost on it. (I have to agree - never found it helpful.) He suggests doing a sweep with a cut, and listen for when the problem goes away, though he suggests (IIRC) developing an idea on where the problem is, so you search area is relatively small.
 
one piece of advice I’m always coming across with regards to eq is sweeping for “problematic frequencies” .

What do you mean by "problematic" if something is that far off, then re-record it.

Its good to learn what frequencies are friendly to which instruments before you record or mix them. All instruments have specific placements.
 

Attachments

  • Spectrum.png
    Spectrum.png
    448.6 KB · Views: 40
What do you mean by "problematic" if something is that far off, then re-record it..

Re-recording may not be the answer. For example, a guitar may have a particular note that is standing out. For example, a bottom G might, for some reason, be boomy. So you can cut a bit of EQ on that note.
 
Re-recording may not be the answer. For example, a guitar may have a particular note that is standing out. For example, a bottom G might, for some reason, be boomy. So you can cut a bit of EQ on that note.

True, most instruments naturally double their sound pressure on certain low notes. There are several ways it can be minimized.
 
I never sweep, it's much easier and quicker to call up a spectrum analyser. You can immediately see the offending frequency and then deal with it using your preferred method of EQ.
 
I never sweep, it's much easier and quicker to call up a spectrum analyser. You can immediately see the offending frequency and then deal with it using your preferred method of EQ.

Assuming you are familiar with what particular instruments look like on the analyzer. Every instrument looks different. I work in Cubase, every EQ channel is tied to an analyzer. It is a great tool.
 
Assuming you are familiar with what particular instruments look like on the analyzer. Every instrument looks different. I work in Cubase, every EQ channel is tied to an analyzer. It is a great tool.
This assumes that the thing you hear in the mix that's bothering you is actually caused by a peak in a particular track. It's not always a single track, or even an observable peak IME. It's just something that catches your ear that you feel is out of place. Then, having some idea of where to start your testing, and using the the [cut] sweep method will be more useful, especially if it's really kind of a transient that can be very hard to pinpoint visually (again IMO/IME).
 
This assumes that the thing you hear in the mix that's bothering you is actually caused by a peak in a particular track. It's not always a single track, or even an observable peak IME. It's just something that catches your ear that you feel is out of place. Then, having some idea of where to start your testing, and using the the [cut] sweep method will be more useful, especially if it's really kind of a transient that can be very hard to pinpoint visually (again IMO/IME).

And to add..once more :>) Sometimes the 'problem isn't necessarily a problem of level -but one of dwell time. This more likely with instruments (or room but that's another discussion) in the lower to low mids.
 
Assuming you are familiar with what particular instruments look like on the analyzer. Every instrument looks different.
Not sure what you mean. If there is a peak at a certain frequency, an SA will show it, regardless of the instrument that is playing it.
 
Not sure what you mean. If there is a peak at a certain frequency, an SA will show it, regardless of the instrument that is playing it.

Um, you still have to have an idea of what you're trying to get rid of. The biggest peak on your analyzer should be the fundamental of the instrument you are playing and cutting that "by eye", as it were, would just cut the note you are playing. If you are eq'ing, say, a bass track of whatever type, your fundamental is going to be the biggest peak but the second, third and fourth harmonics will be the next highest unless you have some mistake you are trying to get rid of. If you don't know where your tone and harmonics are supposed to be for that instrument and then go cutting stuff that is visually higher, you won't be doing your self a favor. So , by all means , use an analyzer, but know how what you see relates to what you hear in order to make the changes that you wish. IOW an offending frequency is often NOT the highest peak on analyzer.
 
IOW an offending frequency is often NOT the highest peak on analyzer.
No, but it's the highest amongst others around it. If you cut the highest one and you don't hear an improvement, I think we would all assume at that point that the wrong frequency has been cut. In that case, try a different peak. Having said that, nine times out of ten the frequency I first go to is the right one. I mean, nobody is seriously going to be looking at 5kHz (for example) when they can hear some boxy, honky mid tones.
 
simply dropping freqs that peak,
is not the answer.
it may be THAT peak, is really one you need,
or possibly is creating a masking issue,
which may need the tweak of a fundamental and not a harmonic.
 
Back
Top