For professional results when recording vocals, you’re going to need a de-esser to tame sibilance. Maybe you’ve already checked out our roundup of the best de-esser plugins for vocals but want to learn more about de-essing and sibilance in general? Well, you’re in the right place.
In this article, we’ll unravel sibilance with explosive insss-ight – and some maddening wordplay. Let’s take a look at the fundamental concepts of de-essing alongside a step-by-step guide to de-essing and some general tips to improve your vocal tracks and mixing skills.
What is a De-Esser?
De-essers are specialized tools that reduce the audio signal level either in part or whole to mitigate sibilance in vocals or other instruments. They do so without considerably altering the overall character of the voice. In simple words, if you use a sidechain compressor that uses sibilance as a key (instead of a kick drum), you’ve got yourself a traditional de-esser.
That is just a basic overview. Many modern plugins and racks have their own refinements and will work in different ways. Moreover, each de-esser plugin has its own proprietary settings and presets that cater to specific situations like dialog processing instead of singing.
What is Sibilance?
Sibilance generally refers to a ‘sizzling’ harshness in vocal recordings with ‘s’, ‘sh’, ‘f’, ‘x’, or soft ‘c’ sounds. It arises from a combination of factors such as the vocalist’s voice type, the distance from and type of microphone used, and other factors like heavy compression in the post-processing stage.
These ‘sibilant events’ are usually found in the higher end of the frequency spectrum and tend to be more irksome when listening to music on earbuds and smaller speakers. Before you get dismissive because you have a nice pair of studio monitors to listen to music on, just remember that millions of people listen to music on AirPods and – yes – tiny phone and laptop speakers.
Sibilance itself is not bad – it would sound weird if we removed it entirely. But it can be distracting if it’s too present. Our ears are especially sensitive to frequencies in this range, so they can really take over in a mix if not attenuated properly.
A Brief Hisss-tory
De-essing debuted in the 1940s, established by Warner Brother Films. Though still primitive in form, it performed the vital role of taming sibilance in speech and dialogue for the next two decades. In the 1960s, it took roots in the music recording industry, but only at the mastering stage.
Unlike today, that was an era when they recorded the entire band or ensemble simultaneously. The technology was not ready for a ‘dynamic compressor’ with a ‘set and forget’ preset. Thus, singers had to limit their sibilance in the recording stage, and there were only basic tools for dealing with a sibilant vocal performance like EQ.
In the light of this problem, the music industry sifted through various de-essing solutions ranging from treble limiters to standalone dynamic controllers to find something reliable and convenient. Finally, the 80s saw the rise of the legendary dbx 902 with VCA gain reduction.
It is possible to EQ away the sibilance from vocals, but it just doesn’t sound as nice as a de-esser. For instance, you can pinpoint the problem frequency that generally lies in the 3-10 kHz range and use a narrow bell filter to attenuate it.
However, this will get rid of that frequency in the entire vocal track, which can make it sound weak or unnatural. To remedy this, you need some form of gain automation. Sadly, that can be a time-consuming and mind-numbing process… unless you enjoy fine-tuning curves for the full track.
You can sidestep the process of gain automation by using de-esser plugins. The primary role of de-essing plugins is to cut down overbearing sibilance in a vocal track. Various VST plugins exist for this purpose, with varying numbers of parameters and complexity.
They essentially work by detecting when a sibilant frequency crosses a set threshold and applying gain reduction to the signal. You can get the job done with any stock plugin in all major Daws or access the advanced capabilities of third party options such as FabFilter Pro-DS and Sonnox Oxford SuprEsser V3.
De-esser plugins will have either wide-band or split-band processing.
A wide-band de-esser essentially acts as a sidechain compressor. Its detection circuit gets triggered by an external signal, but the compression itself is applied to the source signal, which, in this case, is the sound that is being attenuated.
To achieve this specific goal, the sidechain consists of the original signal but with a gain boost around the frequencies where the harsh sibilance is present.
It causes the compressor to be triggered only by the sibilant frequencies. The compressor performs gain reduction on the entire signal whenever the level of these frequencies crosses the set threshold.
In wide-band de-essing, the entire signal is compressed in response to the sibilant frequencies. Attack and release times are pivotal in preventing unnatural pumping effects that make the vocal inaudible.
However, wide-band processing can retain the overall timbre of the voice better since all frequencies are affected equally. Hence, wide-band de-essing is a little harder to dial in for very specific frequencies but can yield more musical results. Most plugins include an option for wide-band processing along with split-band techniques.
In split-band de-essing, instead of applying gain reduction to the entire signal, the signal is split into two parts. Once you decide the target frequency for reducing sibilance, the signal is split into frequencies lower than the target frequency, and frequencies above it.
Now, when the sibilant frequencies in the sidechain cross the threshold, only the higher band in this split is compressed. This only targets the offending high frequencies without affecting the lower frequencies where the ‘body’ of the vocal may be.
Modern versions of split-band de-essers often incorporate three bands: lower than, around, and above the target frequency. This allows for pinpointing the exact range of the harshness and reducing it without affecting the rest of the recording.
In some cases, the width of the sibilant range may be adjustable. Overall, split-band plugins are more capable of surgical de-essing for greater precision without volume discrepancies.
Step by Step Guide to Using a De-Esser
Step One: Isolate the most sibilant part
Play the entire track and note down the time stamp of the significant sibilant events. From the list, pick the part that has the most sibilance and set it to loop with your de-esser activated on the track. Find a phrase with a lot of ‘s’ or ‘sh’ syllables.
If you’re treating cymbals, find a part where there might be bright, loud hi-hats or crash hits. Looping such a section will make it easy to hear what the de-esser is doing.
Step Two: Find the offending frequency
Using your de-esser’s audition function (or equivalent), pinpoint the frequency or range of frequencies that sound the harshest. Using FabFilter’s Pro-DS as an example here, we can set the range and move the ‘Audition’ slider across the frequency range to isolate the harshness.
This plugin also provides a visual aid in the form of yellow illumination of varying intensities to indicate the presence of sibilance. You should rely on your ears to figure out where to target the de-esser, and once you’ve identified the problem area, turn the audition or solo function off.
Step Three: Adjust the sensitivity (threshold)
Set the sensitivity of the de-esser plugin so that it only attenuates the signal when sibilance occurs. Start with the threshold maxed out so that no de-essing takes place. Then, start pulling down the ‘threshold’ until you hear the de-esser kicking in.
In this example in Pro-DS, we see that as we pull the threshold down, the de-esser’s activity is marked in green, along with gain reduction in red visible on the meter to the right.
Step Four: Adjust the strength
Depending on your plugin, the strength or intensity of the de-essing applied might be controlled by the threshold, or by a dedicated ‘strength’ control or similar function. Start by lowering the threshold (or increasing strength) to the point where the vocalist’s sibilant consonants get aggressively attenuated and muffled.
From this point, start backing off the strength until you reach a point where the de-essing eliminates the harshness without muffling the sound. You can use a similar approach while treating harsh cymbals or even full mixes.
Some plugins, like FabFilter Pro-DS shown here, have a ‘range control’ to limit the maximum amount of attenuation that can occur. This ensures that the de-essing doesn’t get too aggressive even when levels far exceed the threshold.
Once you’re satisfied with the frequency selection and amount of de-essing, you can fine-tune the settings in your chosen plugin. As discussed before, you can choose between split-band or wide-band modes depending on your material and preference.
FabFilter Pro-DS provides some extra options. It features dedicated modes optimized for single vocal performances or an ‘All-round’ mode helpful for treating buses or full mixes. Further, a mid-side processing option provides further flexibility and can help tame cymbals in a full mix by de-essing only the side signal (where the cymbals are present) while leaving the mid signal untouched.
Pro-DS also provides a ‘lookahead’ function, which previews the incoming material to prepare to catch very fast transients before they arrive. This, of course, comes at the cost of some latency.
De-Essing Audio Examples
Here I have included some sample tracks from a studio recording, before and after de-essing, to help you understand the process. They demonstrate how to subtly iron out the problematic frequencies in different ways.
Where Should You Place the De-esser in the Vocal Chain?
Really there are no strict rules to de-esser placement, but generally, the de-esser is placed after the EQ and compressor. You don’t want to boost anything (via EQ) once the track has been de-essed as it can cause unwanted sibilance to reappear. As for compression, it will raise the noise floor and make it easy to spot the sibilance. Moreover, if a compressor is placed after a de-essser, it can bring the sibilants back up unless you do hard de-essing.
Secondly, you want sounds to be at their cleanest before they enter the effects plugins of the mix chain. Effects exaggerate the troublesome or problem frequencies. You don’t want to deal with any such exaggeration while trying to de-ess a track. So, as a rule of thumb, I recommend placing the de-esser after EQ and compressor and before other effects.
Nevertheless, mixing is as much an art as it is a science. Some people prefer placing the de-esser before EQ & compression. Others place it at the end of the chain. As a beginner, you can use trial and error to figure out what works for you. If you can justify some other creative or quirky placement trick, I’d love to hear about it in the comments.
De Essers: General Usage Tips
Hear what you don’t want others to
All modern de-essers feature a “listen” option that only plays back the parts that have been removed. This indispensable feature will come in handy when you start using the plugin and will help you with a) being certain that you are only de-essing what needs to be tamed, and b) building a good understanding of your plugin, its range, and capabilities.
Don’t overdo it
If you are new to de-essing, you’ll be tempted to apply more attenuation than required. De-essing is one of those things that benefits from a ‘less is more’ approach. If you slam the signal with an aggressive preset, it will lead to what many people call the ‘Sylvester Effect’. That would be Sylvester the cat, not the Italian Stallion.
This is very distracting and will instantly ruin your song. Remember – our ears are very sensitive to vocals, and we’re very good at detecting unnatural processing. Make sure you de-ess in moderation.
Besides vocals, de-essers can be handy for taming cymbal splash (especially with reverb), finger buzz or “clack” from a guitar recording, as well as bow noise from violas or cellos. Some mixing engineers use specialized and meticulous de-essing on full mixes to filter out overall harshness.
Automation is (still) your friend
There will be times when you have to manually increase the de-ess amount on a particularly strong sibilant event or change the band for a particular phrase to avoid any harshness. It might seem counterintuitive, but don’t be shy about automating here – it will come in handy in more advanced situations.
Over and Out
I’ll leave you with one final tip – de-essers are not sledgehammers. Hold back that heavy-handed itch and remember that it’s too much if you can hear it. Great vocal tracks are a result of a subtle, meticulous, and cumulative layering of tools (and great singing).
You always want a vocal recording to sound natural as possible. When you nail it, your audience won’t even realize how good a job you are doing…and now you know what a bass player feels like!
I don’t expect you to be a master at de-essing after reading this article alone – that takes practice. But I at least hope you’ve got the gist of it and know enough to get sss-started.