Fixing Room Resonances—After The Fact

After all that’s been said about acoustical treatment, you’d think that anyone who was halfway serious about recording would take some steps to improve their room’s acoustical properties. However, recordings are still being made where unwanted room resonances are part of the tracks. In a situation where a mostly acoustic recording is done with multiple tracks recorded in the same room, and there are resonances, these frequencies may build up over the course of several tracks to add some serious peaks in the final mix.You can try to fix this during the mastering process; you’ll need a very precise digital filter to do so, and I use the Har-Bal program (basically a 8,192 stage FIR filter optimized for mastering) for this type of application. However, if you’re handed tracks like this to mix, sometimes you can get rid of the resonances during the mixing process itself, which can yield better results. This involves finding where the resonances lie, then processing each track to ge
Author:
Publish date:

SO . . . HOW BAD IS IT?

As you’re setting up tracks to mix, some resonances might jump out at you if the acoustics are really bad. What’s more, you may hear this on all tracks that were recorded in the room, unless they were close-miked or recorded in significantly different ways.

If possible, it can be helpful to record a sweep tone in the tracking room to give you a very rough idea of where the resonances might lie. Figure 1 shows one such sweep, which was then brought into Har-Bal for analysis. The green line shows average signal level, while the yellow line shows peaks — which are basically identical, as we’re simply looking at a sine wave, not a complex signal.

The major resonance areas are outlined in red. Granted, there are plenty of other peaks and dips, but these aren’t as severe.

THE BIG FIX

Har-Bal is a stand-alone application, not a plug-in. Therefore, it’s easiest to have linear tracks to work on, as opposed to lots of smaller segments. This may require exporting portions of the mix to create tracks that last the duration of the song. That way, when you re-import the tracks back into the song, you can insert them at the same start time and know they’ll sync up. (Some hosts will let you call up an editor to edit a particular track, in which case you can use whatever EQs you normally use to do surgical-type EQ.)

Then, before even thinking of doing any processing, look/listen for resonances the tracks have in common. But be careful, because lots of resonances are part of a characteristic instrument sound; you don’t want to mess with these

Once you’ve identified the main peaks, work on each track individually. Har-Bal has a function where if you come up with a particular compensating EQ curve, you can save it and apply it to other tracks (Figure 2). If the curve is gentle enough, this can provide an excellent point of departure for doing finer tweaks, but always listen carefully to double-check — a track may have been recorded differently, and have different resonance characteristics. With other programs (I’ve also used this technique with iZotope’s Ozone 3), you can almost certainly save any EQ preset you come up with and apply that as well.

TIME TO MIX

Now you can re-import the tracks into your project, and begin your mix. You may be surprised how much more easily a mix comes together if the tracks have been properly treated for resonances; the more tracks that were recorded in the same room, the more of a difference this will make. A side benefit of this technique is that by reducing built-up peaks, you’ll have more headroom available for the mix, and will be able to get a “louder” sound without having to resort to as much compression or limiting.

So the next time you need to mix a project that was recorded under less-than-ideal acoustic conditions, try fixing the tracks before you mix — it might make the difference between “okay” and “yeah!”