Vocal sibilance is a tonal harshness that arise from syllables like S,T and Z. And often have a occur in the frequency range of 5-9kHz. This problem is often caused by the artist vocal formant but can be exaggerated when using “wrong” microphone or a bad microphone placement.
How To Prevent Sibilance
The best option in my view would be for the vocalist to learn how to sing without too much sibilance. (But it could be hard for you as a recording engineer/producer to make the singer learn a new micro vocal technique during a session). So here are a few tips that I like to use.
Choose The Right Mic
If you notice that the singer has a lot of sibilance you may want to try out a more “darker/deeper sounding” microphone. Look for something that has a smooth top end. The most expensive microphone is not necessary always the best for your situation. Play around.
Find The Right Spot
Even if you just have one microphone you can still get the right sound. Play around with the placement of the microphone. Personally I like to aim for the corner of the mouth if the sibilance is too harsh or exaggerated. This technique can interfere with the frequency response, so be careful.
If Those Don't Work, Then What?
If you have tried out the techniques above and still have a lot of sibilance, you probably can fix it in the mix. I always advocate to “fix” things in the recording stage so as not to end up with a 24 hour edit session. But sometimes you just can’t get rid of the problems.
You can automate the gain of all the S:s, T:s and Z:ts. A bit time consuming but worth it in the end.
This compression is reacting only to the problem-frequencies. It is fast and easy to tweak in and find the problem but it can make your S:s sounds “spikey”. You also might catch some of the treble of non sibilant words and pull down some of the “glitter” of the vocal.
This is a trick I learned recently that works really well. Duplicate the vocal-track and nudge the copy about 50ms earlier in the sessions than the main vocal-track. Now grab a De-esser on your main track and key it to the copy. Every time a “s” hits on the copied signal it will activate the de-esser on the main signal 50ms before the actual "s" from the main signal. The De-esser will reach peak gain reduction before the “s” and will remove the "s" sound very transparently.