What does Audio De-Esser do?
Audio De-Esser turns down a bright band when that band gets loud. The file is split into a dry path and a side path. The side path is a high-pass at the frequency you choose, followed by a low-pass at 12 kHz, so the detector is listening to sibilance rather than the whole voice. When that band crosses the threshold, a compressor reduces the dry voice. Amount is the ratio. Attack and release are how fast the reduction starts and lets go. The frequency control is hertz, from 2 kHz to 12 kHz. A light setting catches only the harshest peaks. A strong setting can dull the voice because it also reduces brightness you may have wanted. This is not a full vocal cleanup, and it will not remove noise, room tone or distortion. Compare the before and after players. If the source already peaks near full scale, lower the amount before you export.
How to use Audio De-Esser
- 1Upload a voice recording and look at the waveform.
- 2Start with Light or Voice, or set frequency, threshold and amount yourself.
- 3Export and compare the after player with the original.
- 4Cancel stops the local job. Strong settings can dull the voice.
Supported inputs and outputs
Inputs
Local audio the browser can decode, such as MP3, WAV, M4A or OGG
Outputs
MP3, WAV, M4A
Practical uses
- Softening harsh S and SH sounds in a narration
- Trying a lighter pass before a stronger one
- Checking a voice recording against a waveform before export
Browser and media limitations
- A strong amount can reduce brightness and make the voice sound dull.
- This does not remove noise, echo, clipping or a whole mix problem.
- The band is fixed. It does not find S sounds by recognizing speech.
- The media engine downloads on first export.
How local processing works
Your media is processed locally in your browser and is not uploaded to Tubelexity for processing. The media engine downloads on first export. Page navigation and analytics still use network requests.
Sibilance, sidechain reduction and dulling
Sibilance is the bright energy in S and SH sounds, often from about 5 kHz upward. The side path isolates that band in hertz and uses it as the signal that tells the compressor when to turn the voice down. The rest of the voice is not filtered away.
If the amount is high, ordinary brightness is reduced along with the harsh peaks. That is overprocessing. Use the after player and back the amount off if the voice sounds muffled.
Frequently asked questions
Which frequencies are treated as sibilance?
The side path starts at the frequency you set, between 2 kHz and 12 kHz, and stops at 12 kHz. Speech S sounds often sit in that bright range.
Will a strong preset clean the voice completely?
No. Strong reduction can dull the voice. It does not remove noise or fix a clipped recording.
What do threshold and amount do?
Threshold is how loud the bright band must be before reduction starts. Amount is how hard that band turns the voice down.
Is the recording uploaded?
No. The file stays in this browser tab. The media engine downloads on first export.