Compression (AGAIN!) Before EQ or After EQ???

[Massive Master]

If there's a "rule of thumb" it would be corrective EQ first, compression after, shaping EQ after. Not "rules" -- Just guidelines that tend to work most of the time.

[Ethan Winer]

If there's a "rule of thumb" it would be corrective EQ first, compression after, shaping EQ after. Not "rules" -- Just guidelines that tend to work most of the time.

I'll go farther and say that is a useful rule. The quote below from my book explains the principles in more detail.

--Ethan

One of the most common questions I see asked in audio forums is if it’s better to equalize before compressing or vice versa. In many cases you’ll have one EQ before compressing and a second EQ after. Let’s take a closer look.

If a track has excessive bass content that needs to be filtered, you should do that before the compressor. Otherwise rumbles and footsteps, or just excessive low-frequency energy, will trigger the compressor to lower the volume unnecessarily. If you compress an unfiltered track, the compressor lowers and raises the volume as it attempts to keep the levels even, but those volume changes are not appropriate and will likely detract from the sound. The same applies for other frequencies that you know will be removed with EQ, such as excess sibilance on a vocal track or a drum resonance you plan to notch out. Therefore, you should do any such corrective EQ before compressing.

However, if you boost desirable frequencies before compressing, the compression tends to counter that boost. As you apply more and more EQ boost, the compressor keeps lowering the volume, reducing that boost. In fact, this is how many de-essers work: They sense the amount of high-frequency content in the sibilance range, then reduce either the overall volume or just the high frequencies, depending on the particular de-esser’s design. So when you’re using EQ to change the basic tone of a track, that’s best done after compressing. Again, there are few rules with art, and I encourage you to experiment. My intent is merely to explain the logic and theory behind mixing decisions that have a basis in science or that are sensible most of the time.

Figure 7.4 shows the Track Controls for a bass track from one of my projects, with three plug-ins inserted: an EQ, a compressor, then another EQ. The first EQ in the chain applies a gentle 6 dB per octave roll-off below 60 Hz to reduce the overall low-frequency content. This is followed by a compressor having a fairly aggressive 10:1 ratio, which is then followed by the EQ that actually alters the tone of the bass with a slight boost at 175 Hz.

One situation where compressing first usually makes sense is when the signal chain includes a severe distortion effect. Distortion tends to bring up the noise floor quite a bit because of the high gain it applies, so compressing after distortion raises the noise even further. Another time you’ll want to compress first is with an echo effect whose repeating echoes decay over time. When a compressor follows an echo effect, it can raise the level of the echoes instead of letting them fade away evenly as intended.

[bouldersoundguy]

I don't think in terms of desirable or undesirable frequencies. I think in terms of spectral balance. That is, I think of the ratio of x frequency to y frequency. Compressors do not change the spectral balance on the time scale of perception of the present, so the "un-eqing" argument doesn't stand.

[Mo Facta]

Essentially. That too. I just think that blanket processes are solutions applied before the real problems have been identified, if any at all.

Cheers

[mixsit]

..Compressors do not change the spectral balance on the time scale of perception of the present, so the "un-eqing" argument doesn't stand.

Pretty deep way to say it
Just based on day to day experience I'd tend to agree with you. I think if you compared the dynamic variation a compressor is seeing vs the amount of eq that's being done, the compressor's difference in response to that eq could be roughly proportional.
I.e let's say you have 5 -- 10db dynamic swings, and that is the full bandwidth changing. Now throw in a few of your typical tone trims- pre or post comp- and maybe some of those are cuts, some boosts.
With the exception of major pre shaping / clean up' eq, I'd expect we'd be looking at pretty small differences.

[RecordingMaster]

If there's a "rule of thumb" it would be corrective EQ first, compression after, shaping EQ after. Not "rules" -- Just guidelines that tend to work most of the time.

I like this theory and have tried it a few times. However since I am mostly itb, it just made a case for too many darn plugins per track!

To the OP:
Not to say I'm the mastermind genius mixer who has the answer, but I find myself eqing before (mostly cuts and the least amounts of boosting needed). IF the compressor is doing something odd the signal due to my eq settings, then it's time for further investigation, but usually is never a problem.

[Omega757]

I always EQ after compressing,it's not a rule of thumb just how i work.Find what works best for any given situation for you.

[RAMI]

I always EQ after compressing,it's not a rule of thumb.

Well, just for the sake of nit-picking; Isn't"always" doing something pretty much the definition of "rule of thumb"?

[CrowsofFritz]

Well, just for the sake of nit-picking; Isn't"always" doing something pretty much the definition of "rule of thumb"?

I think he might have meant it as being non-universal as opposed to personal.

[ashcat_lt]

Also, I generally think of a "rule of thumb" as like a starting point - a usually safe first guess. Something that one "always" does is more of a hard and fast rule, if not a rut!