The research behind Harmonic Recovery

The idea in one paragraph

If one ear hears pitch wrongly and the other hears it correctly, perhaps the correct ear can act as a reference that the brain uses to recalibrate the other. Harmonic Recovery plays your music clearly in the affected ear while the same music fades in and out in your other ear as a pitch hint. The hint gives the correct pitch to compare against, then steps back so the affected ear has to hold it on its own. Over time, the training levels make the hint rarer.

Why it might work

Adult pitch perception can recalibrate

The strongest evidence comes from cochlear implants. The pitch an implant electrode produces often doesn't match what the listener expects at first. Studies that followed implant users found the heard pitch shifted over months and years, in some cases by two octaves or more. In users with some natural hearing, it shifted toward the pitch their natural hearing reported: the brain seems to recalibrate one input against a trusted reference.

Pitch hearing improves with practice

People's ability to tell close pitches apart improves measurably with practice, a well-studied kind of "perceptual learning". Active practice with feedback, such as saying whether a tone was higher or lower and being told the answer, generally trains better than passive listening. That's why the app includes short drills as well as music.

Vision offers a parallel, and a warning

Some treatments for amblyopia ("lazy eye") turn down the strong eye's input so the brain has to rely on the weak eye. It's the same shape of idea. But it's also a warning: in a large randomized trial, a binocular iPad game built on this idea did worse than ordinary patching. Plausible mechanisms don't guarantee results.

Why it might not

• It hasn't been tested. No published study has tested a pitch hint, or any home practice method, for diplacusis.
• The brain might simply defer to the good ear while the hint plays, without retraining the affected ear at all.
• The brain might learn to tolerate the mismatch rather than fix it. People with hearing loss can merge tones from the two ears into one sound even when they're far apart in pitch. Practice could strengthen that merging instead of correcting the pitch.
• Diplacusis often changes on its own, as the underlying hearing loss recovers or fluctuates. Without a comparison group, it's impossible to say whether any change came from practice.

So why measure?

Because the method is unproven, the app doesn't assume it works for you. It measures:

• A listening setup for each pair of headphones finds each ear's quietest audible level. It's a short version of the clinical method, with silent catch trials, so tones play at a level suited to each ear and measurements stay comparable.
• The hearing check matches loudness first, since pitch shifts with loudness in affected ears. It then pitch-matches at six frequencies, twice each, playing the ears one after the other so the brain can't merge the tones.
• Changes are only called real when they're bigger than your own variation between repeated answers. Otherwise the app says "no clear change".
• Drills adapt with a standard staircase method, settling where you're right about 71% of the time.

The result is an honest answer to the only question that matters to you: is your mismatch changing? Your measurements stay on your phone.

What would change our minds

If users' measurements show no change beyond normal variation after weeks of practice, that's evidence against the method, and we'd say so. We'd welcome working with researchers on a proper study. If you're an audiologist or hearing researcher, get in touch.