De-esser

In-Depth Explanation

De-esser is a specialized dynamic processor that reduces sibilant frequencies in vocal recordings. Sibilance is the harsh "S," "Sh," "T," and "Ch" sounds that occur between 5 and 9 kHz. A de-esser detects these sounds and ducks only that frequency range when they get too loud. This tames harshness without affecting the rest of the vocal tone.

How De-essing Works

A de-esser is essentially a sidechain compressor with a built-in filter. Standard compression reduces the entire signal when any frequency gets loud. A de-esser only reacts to a narrow band of high frequencies. When the singer hits an "S" sound, the energy spikes in the 5 to 9 kHz range. The de-esser detects that spike and reduces the volume of that frequency range for the duration of the sibilant. When the "S" passes, the de-esser releases and the vocal returns to normal.

There are two main types of de-essers:

1. Wideband de-essers: These reduce the entire signal level when sibilance is detected. They are simple to use but can make the vocal duck noticeably when an "S" triggers them. This sounds unnatural if set too aggressively.

2. Split-band de-essers: These only reduce the specific frequency band where sibilance occurs. The rest of the vocal stays at full level. This is more transparent and is the preferred approach for most modern mixing.

Most de-essers have three controls:

• Frequency: The center frequency the de-esser monitors. Set this to where the singer's sibilance is loudest. Female vocals typically sibilate between 7 and 9 kHz. Male vocals typically sibilate between 5 and 7 kHz.

• Threshold: The level at which the de-esser activates. Set it so the de-esser only triggers on actual sibilant sounds, not on normal vocal content.

• Range or reduction: How much the de-esser ducks the signal when triggered. A reduction of 3 to 6 dB is usually enough. More than 8 dB sounds obvious and unnatural.

Some de-essers add a second frequency band for tackling both high sibilance (around 8 kHz) and lower "Sh" sounds (around 4 kHz) independently. These dual-band de-essers give you more precise control over different types of sibilance.

Real-World Example

A mixing engineer works on a lead vocal recorded with a large-diaphragm condenser microphone. The microphone has a presence boost at 8 kHz, which makes the singer's "S" sounds piercing. Every time the singer sings the word "see," a sharp spike at 7.5 kHz jumps 12 dB above the surrounding vocal level. On earbuds, the "S" sounds painful.

The engineer inserts a split-band de-esser on the vocal track. They solo the vocal and sweep the frequency control while the singer hits sibilant sounds. The sibilance is loudest at 7.5 kHz. They set the frequency to 7.5 kHz.

They set the threshold to -24 dB, which is low enough that the de-esser triggers on every "S" but not on normal vocal content. They set the reduction to 5 dB. The de-esser ducks the 7.5 kHz range by 5 dB whenever an "S" occurs.

The result: the "S" sounds are now at the same level as the surrounding vocal. The harshness is gone. The vocal tone is unchanged because the de-esser only affects a narrow band around 7.5 kHz. The rest of the frequency spectrum passes through untouched.

The engineer used the stock de-esser in Logic Pro, which costs $0. They also tried FabFilter Pro-DS ($179) and found it more transparent due to its intelligent detection algorithm, but the stock plugin solved the problem adequately.

Why It Matters for Independent Artists

Sibilance is one of the most common problems in home-recorded vocals. Condenser microphones (the most popular choice for home studios) have high-frequency boosts that exaggerate sibilance. Without de-essing, your vocals sound harsh and amateur on streaming platforms.

Practical advice for independent artists:

• Always de-ess after compression: Compression raises the level of quiet sounds, including sibilance. A vocal that sounded fine before compression often needs de-essing after. Place the de-esser after the compressor in your vocal chain.
• Find the exact sibilance frequency: Solo the vocal and loop a section with obvious "S" sounds. Sweep the de-esser frequency control until you find the spot where the sibilance is loudest. This is usually between 5 and 9 kHz. Setting the frequency correctly is more important than any other setting.
• Use less than you think: A reduction of 3 to 5 dB is usually enough. If you can hear the de-esser working (the vocal sounds like it has a lisp), you are reducing too much. Back off the range control.
• De-ess before reverb: If you send sibilant vocals to a reverb, the reverb amplifies the "S" sounds and creates a harsh, splashy wash. De-ess the vocal before it hits the reverb send. This keeps the reverb clean.

Dynamic EQ can also function as a de-esser. A dynamic EQ band at 7 kHz that ducks 4 dB when the signal exceeds a threshold does the same job. Some engineers prefer dynamic EQ because it offers more precise frequency control than a dedicated de-esser.

For mastering, a de-esser on the master bus can tame harshness across the entire mix. Set the frequency to 8 to 10 kHz and the reduction to 1 to 2 dB. This is a subtle move that removes accumulated harshness from cymbals, synths, and vocals without affecting the overall tone. Read our vocal chains guide and our mixing vs mastering breakdown for complete vocal processing chains.

Related Terms

• Compression - The foundational effect that de-essing builds upon
• Sidechain Compression - The routing technique that makes frequency-selective compression possible
• EQ (Equalization) - Static frequency control, as opposed to the dynamic control a de-esser provides
• Transient - Sibilant sounds are fast transients that de-essers are designed to catch