Audio Filter
Pick an audio file, choose low-pass, high-pass, band-pass, or notch, set the frequency, and filter it in your browser — nothing is uploaded.
About this tool
Audio Filter applies one of the four classic frequency filters to a recording:
- Low-pass — keeps everything below the cutoff frequency and attenuates the highs above it. Use it to soften harsh, sibilant, or hissy top end, or to make a clip sound muffled/underwater on purpose.
- High-pass (low-cut) — keeps everything above the cutoff and attenuates the lows below it. Use it to strip rumble, mains hum, HVAC drone, plosives, and mic handling noise.
- Band-pass — keeps only a band of frequencies centred on your frequency, as wide as the band width, and cuts everything outside it. A ~300–3400 Hz band is the classic telephone / walkie-talkie voice sound.
- Notch (band-reject) — the opposite of band-pass: it removes a narrow band centred on your frequency while everything else passes. Perfect for surgically killing a single tone like 50/60 Hz mains hum or a whistle.
Pick an audio file, choose the filter type, set the frequency (and, for band-pass or notch, the band width), choose an output format, and download the result. Everything runs locally in your browser with a WebAssembly build of ffmpeg: your audio is never uploaded to a server.
Controls
- Filter type — low-pass, high-pass, band-pass, or notch (see above).
- Frequency (Hz) — for low-pass / high-pass this is the corner (cutoff) frequency; for band-pass / notch it is the centre of the band. Range 20–20000 Hz, default 1000.
- Band width (Hz) — how wide the band is, for band-pass and notch only (ignored by low-pass and high-pass). A narrow width like 20 makes a surgical notch; a wide one like 3000 passes a broad band. Range 1–10000 Hz, default 200.
- Output format — mp3 (192 kbps, default), wav, ogg, flac, or m4a.
Worked example
Upload a spoken-word recording interview.wav with a steady 60 Hz mains hum.
Choose Notch, set frequency 60, band width 20, and pick wav. The tool
runs bandreject=f=60:width_type=h:w=20 and returns interview-filtered.wav — the
same voice with the hum gone. Want the classic telephone sound instead? Switch to
Band-pass, frequency 1000, band width 2800, and you get
interview-filtered.mp3 with only the mid-band voice passing through.
FAQ
What's the difference between low-pass, high-pass, band-pass, and notch?
They differ in which frequencies they let through. A low-pass passes lows and cuts highs above the cutoff. A high-pass passes highs and cuts lows below the cutoff. A band-pass passes only a band of frequencies centred on your frequency and cuts everything else. A notch (band-reject) does the reverse of a band-pass: it removes just that band and passes everything else. Low-pass and high-pass are set by a single cutoff; band-pass and notch also need a band width.
What frequency should I choose?
It depends on the filter and what you're targeting. To cut rumble with a high-pass, start around 80–120 Hz. To tame harsh highs with a low-pass, try 3000–8000 Hz. For a telephone-style band-pass, centre near 1000 Hz with a wide band (~2800 Hz) so roughly 300–3400 Hz passes. To kill mains hum with a notch, set the frequency to your local mains frequency — 60 Hz (Americas) or 50 Hz (most of Europe/Asia) — with a small band width. Frequencies must fall in the 20–20000 Hz audible range.
What does the band width do, and why is it greyed-out for low/high-pass?
Band width sets how wide the affected band is, in Hz, and it only applies to band-pass and notch — those two filters act on a band centred on your frequency. Low-pass and high-pass are defined by a single cutoff with no band, so the width value is ignored for them. A narrow width (e.g. 20 Hz) makes a surgical notch that removes just one tone; a wide width (e.g. 3000 Hz) passes or rejects a broad range. Range 1–10000 Hz, default 200.
Can this remove hiss or background chatter?
Only if the unwanted sound sits in a specific frequency range. A steady tone like mains hum or a whistle can be removed with a notch; broadband hiss and mid/ high noise overlap the frequencies you want to keep, so a simple filter will dull the whole recording rather than cleanly remove them — a spectral denoiser is the right tool there. One-off sounds (a cough, a click, a door slam) can't be targeted by frequency at all.
Is my audio uploaded anywhere? What are the limits?
No. Processing happens entirely in your browser via a WebAssembly ffmpeg build — the file never leaves your device. Practical limits: input up to about 10 MB, output up to about 10 MB. Input can be any common audio format (mp3, wav, m4a, ogg, flac); output is mp3, wav, ogg, flac, or m4a. Filtering re-encodes the audio, so a lossless stream copy isn't possible. Very long or high-bitrate files may hit the size cap — trim or compress first.
Developer & Automation Access
Run it from the terminal
Same engine as this page, headless — via the gizza CLI:
gizza tool audio-filter 'url=https://example.com/input' 'type=lowpass' 'frequency=1000' 'width=200' 'format=mp3'New to the CLI? Get gizza →
Open it by URL
Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:
https://gizza.ai/tools/audio-filter/?url=https://example.com/input&type=lowpass&frequency=1000&width=200&format=mp3Machine-readable descriptor: tool.json — title + parameters JSON Schema for agents.