Online de-esser: tame the “s” sounds in your voice
Reduce the sibilance (“s”, “z”, “sh”) in voices and podcasts only while it sounds, without dulling the voice. In Auto, the threshold is set for each recording.
Drag your audio files here or click to choose them
- .mp3
- .wav
- .flac
- .ogg
- .m4a
- .aac
Use your Google Drive without giving up your privacy
Pick files from your Drive, process them here and save the result to any folder you like. Everything happens in your browser: ShiftAny has no servers of its own.
- Your Google Drive
- Your browserAll processing happens here
- Your Drive or your device
- Your files travel straight between Google and your browser. They never touch a server of ours, because we don’t have any.
- Processing runs on your device, exactly as it does with a file from your computer.
- The app can only access the files you pick and the ones it saves to your Drive. Never the rest.
- Access expires after one hour and is never stored. This browser only remembers your name and email so you can reconnect in one click; you can disconnect at any time.
- Processing
- In your browser
- Upload
- No upload to ShiftAny
- Account
- No sign-up
- Price
- Free
- Input
- MP3, WAV, FLAC, OGG, M4A, AAC
- Output
- MP3, WAV
- Platform
- Web browser
With the quality of SH Core, our own engine: what you hear is exactly what you download.
How to use “De-esser”
- Add one or more voice recordings.
- Keep Auto ticked and choose the strength: Gentle, Medium or Strong. You will see how much it will reduce in your audio.
- Click “Listen 10 s”; under “Advanced options” you can hear only what is being removed.
- Choose MP3 or WAV, click “Remove sibilance” and download the result.
How the de-esser works
It splits the voice into two bands at the crossover frequency (6 kHz by default) with a Linkwitz-Riley filter, whose bands sum flat: while there are no “s” sounds, the voice sounds exactly the same. When the upper band goes over the threshold, only that band is turned down, for as long as the “s” lasts, with a half-millisecond attack and without touching the body of the voice. In Auto, each file’s threshold is set relative to the level of its voice and to the normal level of sibilance, spoken (16 studio voices) or sung (20 singers), so a voice that already sounds good barely changes. Measured on speech: with deliberately exaggerated “s” sounds, the “Medium” strength removes half of the excess; on a well-recorded voice it acts less than 3% of the time. It is the de-esser of SH Core, ShiftAny’s own engine.
Are my files uploaded?
No. ShiftAny Audio processes your audio directly on your device: this tool runs in your browser and never sends your files to any server.
Limits and metadata
ShiftAny sets no duration or size limits: the real ceiling is your device’s memory, because the audio is decompressed to be processed and an hour of stereo takes more than 1 GB. With very long files, we warn you before loading them. The new file contains only the audio: tags (title, artist, cover art) are not copied.
Frequently asked questions
What is a de-esser?
A compressor that only listens to, and only lowers, the high frequencies of “s”, “z” and “sh” sounds, and only while they sound. An equalizer would lower them all the time and dull the voice.
Which strength should I choose?
“Medium” for most voices. “Gentle” if only the odd “s” bothers you; “Strong” for very sibilant voices or a very bright microphone. Listen first: if the voice starts to lisp, lower the strength.
Does it work with singing or with music?
With singing, yes: choose “Voice type: Sung voice”, because when singing the “s” sounds sit much lower relative to the voice than when speaking. It is validated with classically trained singers; on heavily compressed pop it may reduce slightly too much. On a full mix it would also turn down the cymbals, so it is better applied to the vocal track.
Why not use the “Audio Equalizer” for harsh “s” sounds?
Because the “Audio Equalizer” would lower the highs for the whole recording and the voice would sound dull. The de-esser only acts at the moment of each “s”.