Speaker Resonance Finder Stop Guessing Which Frequency Works
Resonance Finder
Measure · 100–500 Hz
Plays a sweep and listens back with your microphone to find the frequency your speaker outputs most strongly. Nothing is recorded or uploaded — the audio never leaves this page.
How to read the result
Read the position of the peak, not the decibel values. The microphone and room stay constant across the sweep, which makes the shape of the curve reliable even though the absolute levels are uncalibrated.
For the cleanest curve: hard surface, quiet room, 70–80% volume, speaker facing up toward the microphone.
Run it twice. A peak that lands in the same place both times is your speaker.
The Resonance Finder identifies the frequency your speaker outputs most strongly. A sweep runs from 100 Hz to 500 Hz while your microphone reads the level at each step, and the peak of the resulting curve is the frequency that produces the most acoustic output on your device. That frequency is the one that shifts trapped water fastest.
Every speaker has a strongest frequency
A speaker does not reproduce every frequency equally. Driver size, the stiffness of the suspension and the volume of the cavity behind the diaphragm all combine to make some frequencies come out louder than others for the same input. For the small drivers in phones, the strongest output sits in the low hundreds of hertz — which is why 165 Hz became the standard tone for shifting water.
Individual handsets vary around that. A speaker whose strongest output lands at 210 Hz will move noticeably more air there than at 165 Hz, and more air movement means more pressure against trapped water.
Fifteen seconds of sweeping tells you which frequency to use on your device. Take the number into the tone generator and run that instead.
How to get a reading you can trust
| Condition | Why it matters |
|---|---|
| Quiet room | Background noise raises the floor across the whole sweep and flattens the peak you are looking for. |
| Hard surface, speaker facing up | Soft surfaces absorb output unevenly by frequency, which bends the curve. |
| 70–80% volume | Loud enough to dominate the noise floor, quiet enough that the driver is not distorting. |
| Dry speaker | Measure a clean speaker. Water in the grille damps the response and moves the apparent peak. |
| Run it twice | A repeatable peak is a real peak. A peak that moves 100 Hz between runs is the room, not the speaker. |
The absolute levels are not lab figures, and the tool says so — but the relative shape within a single run is real, because the microphone and the room stay constant across the sweep.
Resonance finder questions
What frequency is best for ejecting water from my speaker?
The one your speaker outputs most strongly, which varies between handsets. 165 Hz is the standard starting point and works well on most phones. Run the sweep above and it names your device's peak and shows how much stronger it is than the standard tone.
How does the resonance finder work?
It plays a tone that steps from 100 Hz to 500 Hz through your speaker while reading the level your microphone picks up at each step. Plotting level against frequency produces a response curve, and the highest point on that curve is the frequency your speaker drives hardest.
Does this upload my microphone audio anywhere?
No. The microphone stream connects directly to an analyser inside your browser and is read as frequency levels only. Nothing is recorded, stored or transmitted, and the stream is released the moment the sweep finishes. The measurement runs entirely in the page. Levels are read as numbers from an analyser and plotted, so there is no audio buffer to store even temporarily.
Why does my result change between runs?
Room reflections, background noise and how you hold the phone all shift the reading. Run it twice on a hard surface in a quiet room. A peak that lands within about 20 Hz both times is your speaker; one that jumps around is the room.
How accurate is the measurement?
The relative shape of the curve is reliable, because your microphone and room stay constant across the sweep. Absolute levels are not calibrated figures, so read the peak position rather than the decibel values. Treat it as a comparison rather than a specification. It reliably tells you that your speaker is stronger at one frequency than another, which is what you need to pick a cleaning tone.
What do I do with the frequency it gives me?
Enter it in the tone generator and run that instead of the standard tone, with the speaker facing the floor. If your peak sits close to 165 Hz, the standard water-eject cycle is already doing the job and there is nothing to change.
Sources
- MDN — AnalyserNode, the interface used to read levels
- MDN — getUserMedia, the microphone interface used for the sweep
Last verified July 2026.