Luckily, GarageBand has a built-in de-esser (that’s literally what they’re called), to take care of this. Hit B to open up the Smart Controls, then drop in an instance of GarageBand’s DeEsser in the plugin chain. TL;DW: DeEss is the best de-essing solution there is. A go-to utility plugin. Now it can be told: I’m just back from PatreCon! I’ll have more of that (probably just on my website and youtube, as it won’t be a product release) but let’s just say: what I’m doing with Airwindows has powerful motives. Spitfish is a free de esser VST, Audio Unit plugin developed by digitalfishphones. Compatible OS (s): Windows 64b, macOS. Spitfish is a dead-easy to use de-esser, aimed mainly at mono or stereo vocal tracks. Just like classic analog de-essers, this unit dynamically filters out harsh, annoying s-like sounds that would otherwise ‘spit’ in.

Article Content

The human voice is a wonderfully complex instrument. A typical sentence — whether its spoken or sung, is actually composed of several sonic components, some generated in the lungs and chest cavity, others a result of the resonance of the larynx (which we also refer to as our “voice box”), and even more that occur higher up in the anatomy (think tongue, lips, etc.)

When recording the human voice, it’s essential to be mindful of how each of these components is working together. Some microphones do a better job of capturing and reproducing the rich, full, low frequencies of a voice, others are airy, silky, and articulate.

Context is everything, of course, and whereas one of my favorite microphones, the AKG C414 XLII works like a dream on certain singers, I’ve found it to be overly “bright” on other performers. When I say bright, I’m referring to how the microphone reproduces high frequencies, specifically, “esses”.

Esses can be generated by a human voice when the tongue is placed on the upper palate, directly behind the top teeth. When air is generated and passes through, the resulting sound consists of mostly high frequencies. Different voices have different tendencies, and combining a vocalist who generates an abundance of high frequencies with a microphone that was designed with an emphasis on capturing these same frequencies can result in a displeasing sound.

Ideally, the recording engineer will get it right at the source, but in some cases, the mix engineer simply has to make do with what they’ve been provided — moreover, with modern music production and mixing aesthetics tending to call for liberal usage of compression, saturation, and additive EQ (especially in the high frequencies), sibilance can quickly become a problem.

There isn’t a singular solution to this issue, but a “de-esser” is one of the most common tools used to tame harshness in vocals. Traditionally, a de-esser works as a compressor that has been fed an EQ boost via a sidechain, making it more responsive to the frequency range at which the EQ boost has been applied. Once the de-esser detects that the amplitude has surpassed a set threshold, attenuation occurs.

The SPL De-Esser plugins also contain features not found on the original hardware, such as an M/S Stereo mode that allows you to de-ess a full mix and without losing the high frequencies of cymbals, guitars or keyboards. This makes the SPL De-Esser a valuable tool during mastering.

Beneath is a breakdown of some of the terminology you’ll typically find on a de-esser and a roundup of my favorite plugins used for de-essing. I want to be careful to clarify that not all of these are designed specifically for the task of de-essing, but plugins have come such a long way in a short time, and several of these handle sibilance brilliantly.

De-Esser Controls and Terminology

Threshold

Much like the threshold control on a compressor, this determines the point at which the processing takes place. Once the amplitude of the signal “crosses” the threshold, de-essing will occur.

Range

This determines how much de-essing will take place once the amplitude crosses the threshold. Lower values may sound more subtle and natural but might not do the job entirely. Higher values can create undesired artifacts and sometimes leave the vocalist sounding “lispy”

Lookahead

This feature allows the de-esser to predict incoming sibilance by “looking ahead” a certain amount of milliseconds. For rapid-fire vocal performances, this might be a useful feature to handle sibilance. I’ve found lookahead to be helpful when mixing several Hip-Hop projects.

Wide vs. Split-band

Wide-band is great for general de-essing tasks and is a great starting point. It will attenuate all areas of the frequency spectrum evenly. Split-band will attenuate a smaller range of the frequency spectrum, generally the high frequencies where sibilance occurs. I suggest experimenting with both when mixing, and discovering which works best for the specific performance.

Some de-essers feature a function that allows you to hear only what the processor is doing, which can be helpful for determining if the de-esser is working at the correct area of the frequency spectrum. This function is sometimes identified as “audition”, “delta”, or “monitor”.

Additionally, certain de-essers allow you to choose at which frequency range the sidechain EQ boost is applied, so keep an eye out for this control, which is sometimes named differently depending on which plugin you’re using.

My Favorite De-Esser Plugins …

1. oeksound soothe2

The original soothe plugin from oeksound was a subtle game-changer for me. Normally, the plugins that really excite me are the ones that add some unique characteristic — saturators that growl, equalizers with character, or compressors that really smack. But the original soothe (and its recently-released successor) are subtractive by nature, and were designed to remove unwanted frequencies, and not really add anything whatsoever.

Described s a dynamic resonance suppressor, soothe2 will only attenuate unwanted frequencies when it detects an abundance of them. The user places individual bands at the offending frequencies, and then dials in the depth, sharpness, and selectivity controls to transparently attenuate, resulting in a sound that keeps what you want, and removes what you don’t. Soothe was a mainstay for me on mix-buss, sub-groups, and when mixing vocals, and soothe2 is even more powerful. It was re-coded from the ground up and sounds cleaner and more transparent than its predecessor.

soothe 2 is a fantastic update to the original, with new features and capabilities that include:

De Esser For Garageband 2

The ability to affect the entire human hearing spectrum, rather than just the high-mids and highs
You can now work in mid/side mode, which is great for mix-buss and mastering duties
Attack and release parameters, which makes working on material with prominent transients much easier
A new ‘soft mode’ which allows for more transparent processing

Mix Tip

The new mid-side mode is so useful. I typically use mid-side equalization when processing the mix buss to enhance clarity and space. I now use soothe in this way as well, hunting down harshness that only exists straight up the center of the mix, and attenuating that that frequency range as to not alter the content on the sides. I then find unwanted resonances in the sides, and attenuate there, preserving what is up the middle. It’s a subtle process, but one that truly improves the depth and width of mixes and/or masters.

How to Tame Harshness in a Mix with oeksound soothe

2. FabFilter Pro-DS

The transparent sound quality, helpful visual feedback, and vers