Skip to content

A sound tool for tinnitus

I have tinnitus, so I built this for my own ears and made it public for anyone who wants it. Layer noise and tones, find your pitch, and try the guided modes the research describes.

Stopped

50%
Sleep timer (minutes)

Three looping layers (a band of noise, a main tone and a second tone), each with its own frequency, volume and left/right balance.

Noise

2 kHz
1.0 octaves
30%
Center

Main tone

4 kHz
30%
Center
Waveform

Second tone

6 kHz
30%
Center
Waveform

What the research says

The idea behind most of these modes goes like this. When hearing drops at a certain pitch, the brain cells tuned to that pitch lose their input and start firing together, and the ring you hear is that synchronized activity. Sound on either side of the pitch is meant to break the synchrony up, which is where the flanking tones and the notched noise come from.

Acoustic Coordinated Reset plays four tones around your matched pitch in a shuffled order. The developer's own placebo-controlled trial of 63 people reported less loudness and less annoyance, and an independent UK double-blind trial of 100 adults found no difference from placebo. Those two results disagree, and nobody has settled it yet.

Notched music, where the music has about an octave removed around your pitch, hasn't held up either. Two 2024 meta-analyses found no significant benefit over ordinary music.

Residual inhibition is the most dependable of the bunch. About a minute of noise near your pitch can quiet the ringing for anywhere from a few seconds to a few minutes, and it shows up in most people who try it. It's brief, so it works better as a way to learn how your own ringing behaves than as a treatment.

Masking is what clinical guidelines actually support. Steady background sound makes the ringing less noticeable and less distressing, and that's the job the noise layer does here whether or not any of the theory above holds. For the distress itself, counseling and cognitive behavioral therapy have the strongest support of anything.

Among the sound-based approaches, the best controlled evidence belongs to Lenire, an FDA-cleared device that pairs sound with mild tongue stimulation, mainly in moderate to severe cases. This page has no tongue stimulation and isn't a substitute for it.

So why build it? The experiment is cheap, it runs on my own ears, the noise layer is plain masking even if everything else turns out to be a wash, and I wanted to hear for myself what the research describes. Try it at a low volume and see what happens with yours!