About Japanese Pitch Accent Trainer

Japanese Pitch Accent Trainer coaches Japanese pitch accent with live feedback: 18 two-mora words across the pattern families, then 12 longer words covering the hard contrasts. Your microphone listens, your pitch is checked mora by mora against the target pattern, and one named hint at a time tells you exactly what to fix.

How it works

• Your browser samples the microphone and measures your pitch with autocorrelation, a standard signal-processing method: it finds how often your voice waveform repeats. No model is downloaded, and no audio is sent anywhere.
• Calibrate first. Pitch accent is high versus low inside YOUR range. Press Set my voice range and slide from your lowest to your highest comfortable note. The app then places the high/low line in the middle of your range: a mora clearly above the line reads as high, clearly below as low. Recalibrate any time in Settings; your range is stored only on this device.
• For each word, the app splits your utterance into morae by equal time division (Japanese morae are near-equal in length), takes the median pitch of each mora, and reads the high/low sequence. The drop position is read straight off that sequence: the first high-to-low step is the accent kernel. A mora landing near the middle of your range makes the reading unclear rather than guessed.
• Press Check my pitch after saying the word plus ga. If the pattern matches, you score. If not, you get one named hint, such as: The drop should come after mora 2, on ga: yours came after mora 1. Hold the high morae one mora longer before you drop.
• Hear the target plays the target pattern as a stepped tone, one even step per mora, placed inside your own calibrated range, so you can hear the pattern before you say it.
• Longer words unlock at 50 points. They are three-mora words plus ga: four morae to hold even.
• Pitch tracking approximates patterns. Background noise, a quiet voice, a creaky voice, a cold, or a devoiced mora can confuse it. If a native speaker says your pattern sounds right, trust them over the app.
• Native-speaker review sought: we are seeking review of this app by native Japanese speakers and Japanese teachers, in parallel with its release. Review is not a gate: the app is live while review is sought, and corrections will be folded in as they arrive.

The four patterns

• Heiban, type 0, flat: low, then high all the way, and a following particle stays high. Example: ame ga, candy. No drop.
• Atamadaka, type 1, head-high: high on the first mora, then an immediate drop, and low after that. Example: ame ga, rain.
• Nakadaka, type 2 and up, middle-high: low, rise, then a drop in the middle of the word. Example: kokoro ga, heart, dropping after mora 2.
• Odaka, type equal to the word's mora count, tail-high: low, high through the word, and the drop lands on a following particle. Example: otoko ga, man, dropping on ga.

The type number is the dictionary convention: the mora after which the pitch drops, counting from the start of the word. Dictionaries and apps like OJAD mark it as a small number next to the word: hashi [1] is chopsticks, hashi [2] is bridge, hashi [0] is edge.

The same sounds with a different pattern are often a different word. That is why pitch accent matters: say the wrong pattern and you may say a different word entirely. Getting it wrong is rarely a barrier to being understood, context usually rescues you, but copying the pattern is what makes speech sound natural.

Why every item ends in ga

Odaka and heiban sound identical on the word alone: both are low, then high to the end. The difference appears only when something follows the word. After a heiban word, a particle stays high. After an odaka word, a particle drops low. So every item in this app is practised as "word + ga": hashi ga, ame ga, otoko ga. The particle is the last mora the app checks, and it is what separates odaka from heiban. This is also how the pattern works in real speech: the pitch of what follows a word depends on the word's pattern.

One more honest limit: this app grades one word plus ga, said in isolation. Real sentences reshape pitch: downstep lowers later phrases, phrasing groups words together, and question intonation rises at the end. And regional dialects use different patterns for the same words: in Kansai, hashi chopsticks and bridge swap patterns. The app teaches standard Tokyo citation forms, which are the patterns dictionaries list.

The confusable pairs

Two confusions trip learners up most. Odaka vs heiban: identical without a particle, separated only by ga. Say the word alone and no one, including this app, can tell them apart. Atamadaka vs everything else: only atamadaka starts high. If your first mora is high, the pattern can only be atamadaka; if it is low, atamadaka is already ruled out. When one mora lands in the middle of your range, the app says so honestly instead of guessing, and names the difference to aim for.

Learn more

These links open other websites in a new tab. Japanese Pitch Accent Trainer isn't affiliated with any of them and sends them nothing.

Privacy

• Your voice never leaves your device. It is analysed in your browser frame by frame and immediately discarded. Nothing is recorded, uploaded or shared.
• No accounts, ads, cookies, analytics or trackers. The app loads nothing from other websites: the code is served from this site, and a security policy blocks connections anywhere else.
• Stored on your device only: which items you've tried, your points, badges, practice days and settings, your calibrated voice range, and whether you've seen the welcome message, in your browser's local storage. Clear it any time in Settings → Clear progress, or with the Offline proof → Delete all my data button, or by clearing this site's data in your browser.
• Offline copy: after your first visit, your browser keeps a copy of the app so it works offline. Clearing site data removes it.

Disclaimer

This is a home practice tool, not a teacher, therapist or safety device. Everything runs on your device: no video, audio or images are sent anywhere. Feedback is automatic and can be wrong. For real instruction, learn from a qualified Japanese teacher or a native-speaker community, linked on the Resources page.

Pitch patterns taught here are standard Tokyo citation forms. Regional dialects differ, and real sentences reshape pitch through downstep, phrasing, and intonation. The app grades one word plus ga in isolation; it does not grade sentences, and it does not check grammar, vocabulary, or meaning.

Accessibility

All controls are keyboard operable, the live check announces results to screen readers, and celebrations respect reduced-motion settings. The pitch display is a visual aid only: every result is also given as text, and the target pattern can be heard with Hear the target.

Credits

Japanese Pitch Accent Trainer is built by Toolantern. Pitch detection (autocorrelation), accent classification, voice-range calibration, and the synthetic reference patterns are original code written for this app in js/pitch.js. All lesson content is original. Pattern assignments follow the standard Tokyo dictionary tradition (NHK accent dictionary, OJAD).