Japanese has fewer sounds than English, and most of them are easy for English speakers. What takes practice is the rhythm: Japanese is spoken in even beats, and getting the number of beats right matters more than any single sound.
Every vocabulary card shows the word, its reading in kana (hiragana or katakana), and — in the first twenty lessons — the reading in romaji, the Latin alphabet. Kana are the real key to pronunciation: each kana is one beat. Press 🔊 next to any example to hear it.
If you only remember five things, make it these. Each one is covered in more detail further down.
Japanese is timed in beats, not syllables. Each kana takes the same amount of time — including ん, the small っ, and the second half of a long vowel. Count them: とうきょう is four beats, not two.
A long vowel lasts two beats, a short one one beat — and the difference changes the meaning.
a, i, u, e, o — short, clear and steady, much like Spanish. They never slide the way English 'go' slides into 'oo'.
Somewhere between English r, l and d: one quick flick of the tongue. Never the English r.
Japanese words have a melody of high and low beats instead of a strongly stressed syllable. The cards don't mark it, so listen and copy. Sometimes it is all that tells two words apart.
Japanese has five vowels, always in the order a, i, u, e, o. Keep them short and keep your mouth still.
Like the a in 'father', but shorter.
Like 'ee' in 'see', but shorter.
Not the 'oo' of 'food': the lips stay relaxed and unrounded, so it is a softer, flatter sound.
Like the e in 'bed'.
Like the o in 'go', but pure — don't let it slide into 'oo'.
Between voiceless consonants (k, s, t, h, p), and at the end of です and ます, i and u are often whispered or nearly silent. です sounds like 'dess', すき like 'ski'.
A long vowel is simply the same vowel held for a second beat. In hiragana it is written with an extra vowel kana; in katakana, with a dash ー. In romaji the cards write it double: oo, uu, ii, aa — or ou and ei, the way the kana spell it.
A long o — two beats of the same sound. おう is not 'o' plus 'u': the う just lengthens the o.
Usually a long e, ee: せんせい sounds like 'sensee'. Said slowly and carefully, the i can be heard.
In katakana, the dash makes the vowel before it long.
Each pair differs only in vowel length.
Most Japanese consonants are close to English — k, s, t, n, h, m, y, w, g, z, d, b and p. These are the ones to watch.
A single light tap of the tongue tip just behind the top teeth — like the tt in American 'better', or a quick d. Between English r and l, and not quite either.
Not an English f: no teeth on the lip. Blow gently through slightly parted lips, as if cooling soup.
The 'ts' of 'cats' — but at the start of a word too, which English never does.
Like 'shee', 'chee' and 'jee', made with the tongue a little flatter than in English.
A breathier h than English — like the h at the start of 'huge'.
Always the hard g of 'go', never the j of 'gem'. In the middle of a word some speakers soften it toward 'ng'.
A small ゃ, ゅ or ょ joins the kana before it into one beat: きょ is 'kyo', said quickly as one sound. A full-size よ would be a second beat.
Two kana take a full beat without having a vowel of their own. They are the ones English speakers most often rush.
A full beat of nasal sound. It takes its shape from what follows: 'm' before b, p and m; 'ng' before k and g; a nasal hum at the end of a word. The romaji always writes n — sanpo is said 'sampo'.
A beat of silence: hold your mouth ready for the next consonant, pause one beat, then release. The romaji doubles that consonant: kitte, chotto.
Where English stresses a syllable by making it louder and longer, Japanese marks a word by pitch: each beat is high or low, and the pattern belongs to the word. Most words rise after the first beat and stay up; others drop after a particular beat. Loudness and length stay even throughout.
The cards don't mark pitch, and you don't need to memorize patterns as a beginner — a wrong pitch rarely stops you from being understood. But copying the melody of the recordings, rather than adding English-style stress, is what makes Japanese sound natural.
Written with the same kana, told apart by pitch alone: chopsticks start high and drop; bridge starts low and rises.
Japanese spelling is almost perfectly regular, with one famous exception: three particles keep old spellings. The romaji on the cards writes them as they sound.
As the topic particle, は is said 'wa', not 'ha'.
The object particle を is said 'o'.
As the direction particle, へ is said 'e', not 'he'.
Each card shows the word as it is normally written — often with kanji — then its reading in kana, and in lessons 1–20 its romaji. Once you can read kana, trust the kana: they show every beat. The romaji follows the Hepburn system, with these conventions.
One kana (counting ゃ ゅ ょ together with the kana before them) is one beat. The kana reading is the most reliable guide to pronunciation on the card.
Hepburn spellings of し, ち, つ, ふ and じ — written as an English speaker would hear them (see Consonants).
Long vowels are written double, following the kana, rather than with a line over the letter. ou and ei are usually said as long o and long e.
A doubled consonant is the small っ: a beat of pause before it.
Always ん, even before b and p, where it is said m (see ん and the small っ).
The particles は, を and へ, written as they sound.
Names of people and places begin with a capital, and a hyphen joins a suffix such as -jin ('person from') to a name: Amerika-jin.
Tap a finger for each kana as you say a word — one tap for ん, one for っ, two for a long vowel. If your taps and the recording don't match, the rhythm is off.
Listen for which beats are high and which are low, and copy that instead of stressing a syllable. Keep every beat the same loudness.
In any lesson, the pronunciation settings let you slow the audio down without changing its pitch — the easiest way to hear long vowels and the pause of っ.
The role-play in each lesson has you speak your character's lines aloud. Rhythm and melody are learned in sentences.