What does running the filter twice actually buy?

A steeper slope. One section removes twelve decibels per octave above the corner, which leaves enough upper midrange that the result sounds like a tone control; two sections double it to about twenty-four, which is what makes the top end disappear rather than merely fade.

Why six hundred hertz?

Because that is below the range where consonants carry their identity but above the fundamental of most speech. The result is that a voice keeps its rhythm and its pitch while losing the detail that makes words easy to follow, which is the impression of listening through water rather than of listening through a filter.

Can the tremolo be slowed down further?

The rate is fixed at 1.2 Hz, which is a little over one cycle per second. Faster rates would read as vibrato and turn the effect into a modulation rather than a swell, so the single rate keeps the intent unambiguous.

Does it add any reverb or tail?

No, it is a filter followed by a tremolo and nothing else. That is deliberate: a tail would suggest a room, and the point here is a medium rather than a space. If a sense of distance is wanted as well, adding a reverb afterwards is the usual approach.

Will a stereo file lose its stereo image?

No. Each channel is processed on its own with the same settings, so the two sides stay different and the image is preserved; the whole recording simply sounds as though it has been heard through water, with both sides affected equally.

Underwater Effect

You Might Also Need

Telephone Effect

Simulate telephone audio quality

Radio Effect

Simulate AM/FM radio sound

Reverb Simulator

Add realistic room reverberation

Hallway / Room Simulator

Simulate various room acoustics

Underwater Effect buries a recording by removing everything above six hundred hertz and adding a slow amplitude swell. Two cascaded filters and one slow tremolo make the whole effect.

What is Underwater Effect?

Drop a file, set the swell depth, and apply; the low pass stays at six hundred hertz and the tremolo at 1.2 Hz.

What the chain does:

Two cascaded second-order low-pass stages at six hundred hertz, each with a Q of 0.707, which together roll off at about twenty-four decibels per octave.
A tremolo at 1.2 Hz, one swell every eight tenths of a second, slow enough to read as moving water rather than as vibrato.
A depth control from nothing to eighty percent that sets how far the level dips between peaks, defaulting to thirty five.
The measured result: at four kilohertz the residual is about a ten-thousandth of the original level, while two hundred hertz still comes through at roughly a quarter of its own.

Cascading two filters rather than using one is what makes the effect convincing. A single second-order section rolls off at twelve decibels per octave, which leaves a good deal of upper midrange audible, so the result sounds like a tone control turned down rather than like a listener under the surface. Doubling the section doubles the slope, and the difference is measurable: content two octaves above the corner is essentially gone, which is what strips the sibilance and the attack from a voice and leaves only its body behind. The choice of corner matters as much as the slope. Six hundred hertz sits below the range where consonants live but above the fundamental of most speech, so words keep their rhythm and pitch while their intelligibility collapses, which is exactly the impression of hearing someone through water. The tremolo supplies the other half of the cue, and it is deliberately slow: a rate of one and a fifth cycles per second sits in the range of a swell rather than of a vibrato, so the ear reads the level movement as distance instead of as modulation. The depth control governs how far each swell dips below its peak, and at the default the envelope falls to about two thirds of full level, which is audible as movement without making the recording pump.

Where the model stops:

The corner frequency is fixed, so the darkness cannot be dialled in.
Nothing above the corner is recovered, so the removed detail is gone for good.
No resonance, delay or tail, which is what separates it from a reverb.
Each channel is filtered on its own, so a stereo file keeps its image and simply sounds distant.
The tremolo rate is fixed at 1.2 Hz.
Output is sixteen bit WAV.

Use it for a voice heard from beneath water, for a submerged version of a scene, or for any part that should sound muffled and far away without losing its rhythm. When the band should be narrow but clear rather than dark and moving, the telephone effect keeps speech intelligible within a phone line's range, and when the aim is a broadcast receiver's colour rather than the surface of a pool, the radio effect is the closer match. When the distance should come from a room rather than from water, the reverb simulator and the hallway room simulator supply it instead.

How to use Underwater Effect

Drop an audio file onto the panel and let it decode.
Set the tremolo depth, keeping it near the default for a natural swell.
Apply the effect, which low passes at six hundred hertz and then modulates the level.
Download the WAV and check that speech is still paced the way it was.

When to use Underwater Effect vs related tools

Underwater Effect is the tool when a recording should sound submerged, which means losing its top end while keeping its movement. Telephone Effect narrows the band to a phone line but leaves the sound dry and clear, and Radio Effect adds a broadcast receiver's saturation. When the loss of clarity should come from a space rather than from water, Reverb Simulator and Hallway Room Simulator place the sound in a room instead.

Privacy & Security

This tool runs entirely in your browser — no data ever leaves your device. There is no server round-trip, no upload, no logging, and no account required. Your input is processed locally using client-side JavaScript and is never stored, transmitted, or accessible to anyone else. When you close the tab, everything disappears.