De-essing: Automatic v's Manual

Ssssss...

When it comes to recording vocals one of the most unwanted artefacts that can protrude from the vocalist or processing on the vocal  via the mic, compressor or eq is the dreaded "Sss" "Shh" known as sibilance. Sometimes the sibilance can be so predominant it can take your ears off with the sheer shrill frequencies like nails down a blackboard. A vocal can sound reasonably fine but slap on a compressor to help level things out and then you can find the sibilance is like a jet aeroplane taking off.

Automatic

The way to combat sibilance is to use a de-esser. A unit or plugin that compresses just the 'S'

frequencies that usually fall somewhere between 3 kHz and 10 kHz, depending upon the singer or processing in use. It's like a mini one band compressor.

The most common classic hardware de-esser dbx 902

The hardware units like the classic dbx 902 shown above were limited with only two controls threshold and frequency. The Threshold control sets the level when the de-essing process begins. The Frequency control allows you to fine tune the frequency to exactly where the S’s occur. Now when it comes to plugin De-essers we are presented with far more precise control and they are able to do more complex tasks than their hardware counterparts could ever do.

Some of these other features include;-

Split-Band - Here, the signal is split into two frequency ranges high and low bands, a range that contains the sibilant frequencies, and a range that does not. Allowing for only the 'Sss' in the selected frequency to be compressed and the rest of the signal to go untouched.

Wideband - As it's name suggests lets you target a wider area. This can be useful on instruments rather than specifically used on vocals. For instance for the removal of finger squeak on an acoustic guitar track. However on a vocal it tends to leave the leading edge of the 'S' unalterterd and you'll tend to catch some other non-sibilant words and reduce them too.

Sidechain - A little confusing terminology found on some plugins as it refers to simply the ability to solo out the de-essing and help you listen and target the 'S' sound more effectively.

One other automatic way of de-essing is to use  pre-triggering of the de-esser. This can be achieved by making a duplicate copy of the vocal track and move the copy 50ms ahead of the main vocal. Now you insert a de-esser on the original vocal track and key it's side chain to the copy. If your de-essing lacks a side chain facility then try using a multiband compressor that has that function. It will act the same way.

Through this method everytime an 'S' comes through on the copied track it will activate the de-esser on the main vocal track but 50ms earlier than when it will sound on the original track. This allows the de-esser to reach peak gain reduction before the 'S' actually sounds. The benefit of this method is a more natural transparent sound.

Manual

De-essing by manually seeking out those annoying 'S's can be painstakingly long but more rewarding in my view as it sounds more natural. This is best done by soloing the vocal in question and listening for any 'S' 'Shh' sounds that really stand out too much. Once found you have two methods of reducing their volume, much like an automatic de-esser hardware/plugin does.

Remember to first duplicate your vocal track. 

Never do this amount of editing on your main vocal track or you could get lost or really mess things up and find you have too many undo's or worse find no way of getting back to your original.

First is by using the clip gain reduction. This is were you would find the troublesome 'S' , cut the audio each side of it and pull down the event/clip gain by -5 to -6db. On playing back it should still be heard but not be so much in your face so to speak.

Second method of manually de-essing is by using automation. This is slightly more easier to do as you don't need to slice or split up the audio. Like the method above locate the 'S' and using volume automation draw in a -5 to -6db volume decrease. This means as the 'S' begins to sound the volume dips lessening its audio impact.

Conclusion

Which ever method you choose one thing is for sure, you want to lessen the impact of any 'S' sounds but not take anything else away. Most automatic de-essers no matter how well setup and tweaked will always reduce some of the frequencies that your not wanting to target. This is where manual de-essing has the upper hand. It's far more specific and precise honing in solely on the annoying 'S's and leaving the adjacent frequencies untouched.

Note it's very easy to over DeEss. If your singer sounds like someone took out his/her front teeth you
are probably overdoing it. If you hear the “DeEssing effect” that too is usually a sign of over
DeEssing. If the audio passage sounds rather natural and free from annoying Sizzle and distorted
Esses then you got it right. 

Watch the video below as I demonstrate the pros and cons of automatic and manual de-essing.