First, I have a problem with using Split mode. I’ve been using it exclusively for years, and never questioned it, as I’ve seen everyone else use it too. But here’s the thing: When using it, you have to set your crossover frequency very carefully. If it’s too low, then the de-esser will kick in whenever a singer sings a loud note, because there will be enough content above the crossover frequency to trigger the effect during regular non-sibilance sounds. Then you have to back off the threshold, and now you’re not catching the sibilance as much as you would like. On the other hand, when you set the threshold too high, then you’re only de-essing some of the sibilance, and that doesn’t sound so good.
This is fine if your singer doesn’t have a big dynamic range, or if you don’t need heavy de-essing. In this case, you can easily find a nice balance of settings that gets the job done. But when I’m needing heavy de-essing, or if the singer has a big dynamic range, I find that I end up compromising. So, a couple of years ago, I started de-essing in “wide” mode, and I found that it works really well much of the time, especially if I break the other rule of de-essing at the same time. Next time you’re having trouble getting just the right settings for your de-esser, try putting it in wide mode. Here, gain reduction will be applied evenly to the whole spectrum, just like a regular compressor, which will free you up to set the frequency wherever you want. Now you can set that parameter very high to make sure than only the sibilance trigers the effect. I usually go somewhere above 9k. The common school of thought is that you don’t want to apply gain reduction to the whole spectrum when de-essing because sibilance has tons of high frequencies, and it’s only those frequencies that you want to turn down. But that’s exactly why wide mode works. Sibilance is all high frequencies; There is really no mid or low frequency content when a singer sings an “s” sound. So why worry about turning down the mids and lows when they just aren’t there to begin with? OK, I know what you might be saying: Yes, that’s cool in theory, but when you apply heavy de-essing in wide mode, sometimes you can hear it pumping. And that can be true unless you put your de-esser after the compression and equalizer.
Like I said earlier, most people put the de-esser before the compressor and EQ, because they don’t want to re-adjust the de-esser everytime they adjust the other two effects. I’m willing to put up with this inconvenience because there’s a bit of magic that happens when you put the de-esser last in the chain. I don’t know about you, but when I’m mixing agressive pop, dance, or rock music, I do two things to the vocal most of the time: The first thing I almost always do is compress the vocal heavily. When that happens, most of the dynamics get squashed: The vocal level becomes very even, except for the sibilance. The sibilance doesn’t have much mid and low frequency content, so even when it sounds uncomfortably loud, its average level is pretty low, so it passes the compressor with much less gain reduction than other sounds. So now what you’re left with is sound that’s really even regardless of whether the singer sings quiet or loud notes, except for the the sibilance, which is really popping through. Because of this, you can now really crank the threshold on your de-esser without worrying about it trigger non-sibilant vocals. This is a great way to apply massive amounts of de-essing when needed. You can’t apply nearly as much de-essing to a vocal when you put it first in the chain.
The second bit of magic that happens when working this way is this: It’s true that when you apply lots of de-essing to a vocal in wide mode, it sometimes doesn’t sound very transparent. But here’s what happens when you put the de-esser after the eq: For some reason, the more compression and high frequency boost you apply, the more transparent the de-esser starts sounding when placed last in the chain. I really don’t know why this happens, but it really just works much of the time. I just go with it.
There’s a couple of things to keep in mind when using these techniques: First, they’re great on aggressive music, but totally not necessary on softer, more organic mixes like orchestral and jazz. Here, you might as well just put the de-esser first so that you don’t have to worry about re-balancing the threshold every time you tweak the compressor or EQ. Secondly, in theory, you would think that setting the frequency so high might prevent the de-esser from kicking in when the singer is singing a “sh” or “tsh” type of sibilance instead of an “ess”. I find that in practice, this is never a problem for me. But I guess it’s possible, in which case, you could always use more than one de-esser to catch different types of sibilance. The very talented Robert L. Smith uses multiple de-essers. Check out his description of the technique here.
Well, I hope this explanation will appease the Audio Engineering gods. Check out the video below to see and hear all this in action and to find out which de-esser I think blows away the competition.
