Skip to content

Japanese sound contrasts

Japanese sounds: long vowels, small っ and the r row

Japanese pronunciation can look deceptively familiar in Roman letters. Yet a longer vowel or a held consonant can identify a different word, and written r does not tell an English speaker exactly what to do. Start with sound and meaning together. These exercises use Japanese writing, small timing diagrams and listening tasks to make the contrasts concrete without asking you to imitate an entire accent at once.

Reviewed · 8 min read · Flalingo editorial team

1. Hear the word before deciding where to stress it

An English speaker may look at a new word and search for its stressed syllable. For the Japanese words in this guide, first check whether a vowel is long and whether a consonant has an extra timing unit. Stretching one part merely to make the word sound important can obscure the word's intended shape.

The useful counting unit is the mora. The Tokyo University of Foreign Studies explanation counts a short kana unit, the second part of a long vowel, small っ and ん as moraic units. Treat that as a practical timing map. Ordinary conversation is not a metronome, and every sound need not occupy an identical measured number of milliseconds.

Put two counters on the table for おば, meaning aunt, and three for おばあ, the beginning of おばあさん, grandmother. Your next step is to hear the complete words おばさん and おばあさん in reliable audio. Counting helps you notice what is present; it does not replace a spoken model.

2. Keep a long vowel continuous

A long vowel is sustained rather than restarted as two separate vowel attacks. TUFS's long-vowel lesson describes this length contrast and its spelling. In katakana, ー commonly marks length. Hiragana uses vowel characters, with spellings such as おう often representing a long o. Check the particular word instead of applying one reading to every adjacent pair of vowels.

Try おばさん and おばあさん at a comfortable volume. For the longer word, let the a continue through the extra timing slot. Avoid making a sudden gap between two a sounds. Equally, avoid compensating by shouting or adding an English-style glide. The target distinction here is the sustained vowel within the complete word.

Ask a partner to show an aunt card or a grandmother card after hearing your recording. Do not show the written Japanese until they choose. If they cannot tell which you intended, compare both recordings with the same reliable model and check the vowel interval. Your partner's choice gives the sound a communicative consequence.

ContrastMeaningWhat to listen for
おばさん / おばあさんAunt / grandmotherThe extra length on a in the second word.
おじさん / おじいさんUncle / grandfatherThe extra length on i in the second word.
さか / さっかSlope / writerThe held consonant interval before the final ka.
きて / きってCome, in a request form / postage stampThe longer t closure in the second word.

3. Let small っ borrow the following consonant

The small character っ is different from the full-sized つ. In words such as きって, it signals an extended consonant interval, not an extra spoken tsu. TUFS's mora-obstruent lesson distinguishes holding a stop closure from extending a following friction sound. There is not one separate noise that you insert wherever small っ appears.

For きって, prepare the t closure and hold it briefly before releasing into て. Say the whole word again without placing a new vowel between the parts. For さっか, the hold is associated with k. You should not say sa-tsu-ka, and you should not add an uh sound to make the consonants easier.

An English compound such as bookcase may offer a useful feeling of adjacent k sounds for some speakers, but the English word is only a starting cue. Return to Japanese audio immediately. English accents differ, and a familiar English sequence cannot establish the correct duration or rhythm of every Japanese example.

4. Separate hearing a contrast from producing it

Make four picture cards: a slope, a writer, an aunt and a grandmother. First listen to model words and point to a picture. Leave your microphone off. If you cannot reliably select the intended picture, additional listening may be more useful than repeating an uncertain version louder.

Next record each word in a mixed order, leaving enough space to make a choice. Ask someone to select the corresponding cards. In solo practice, return to the recording later without looking at your list, then compare your choices with the initial order. Familiarity with your own recording can help you guess, so treat the solo result as a rough check.

Keep two notes: heard correctly and said clearly. You might hear さか and さっか easily but still shorten the held consonant when speaking. Or you might copy a model accurately while finding unfamiliar recordings difficult. Those are different practice needs; one overall impression of being bad at pronunciation would hide them.

5. Approach らりるれろ as Japanese sounds

Romanised r is a label, not an instruction to use your usual English r. TUFS describes the ra row with a brief tongue-tip contact near the ridge behind the upper teeth. Explore a light contact followed by the vowel. Do not force a prolonged roll, and stop if an articulation exercise feels uncomfortable.

Say ら, り, る, れ and ろ after a model, then place the sound in a short familiar word such as そら, sky. A sound you can manage alone may change when you concentrate on remembering a word. Keep the word's meaning visible so you practise the whole item rather than an endless tongue movement.

Some English accents use a brief tap in words such as city, which may provide a personal point of comparison. Other speakers do not. Avoid declaring Japanese r identical to every English d, l or r. Japanese realisations also vary, so a teacher's specific observation is more helpful than trying to place the sound at an imaginary exact midpoint between two English letters.

6. Read short phrases without losing the contrast

Use the practice card after the isolated words feel recognisable. The first pair can identify relatives or women of different ages, depending on context; the final line puts a held consonant inside a useful request for stamps. The glosses explain the meaning. They are not a Roman-letter pronunciation key, so listen to a trustworthy model for any unfamiliar word.

Record one line, listen and identify the single contrast you intended to preserve. On a second attempt, keep your attention on that contrast while saying the phrase smoothly. Do not exaggerate every mora equally or separate each character with a pause. Your aim is a recognisable phrase with its relevant timing intact.

Say it aloud

Keep the word's timing inside a phrase

7. Ask for feedback that identifies the problem

Instead of asking whether your Japanese sounds native, ask which word the listener heard. Then ask where the uncertainty occurred. Was the vowel too short, was there an added vowel before k, or was a different consonant heard in the r-row word? Write down that answer in ordinary language.

Pitch patterns are another part of Japanese pronunciation, and they vary by region. This guide's length drills do not teach every word's pitch pattern. A recording can preserve vowel length yet still need other work. Keep that distinction explicit so success on one contrast does not become an unsupported claim about your overall speaking ability.

The Japanese reading test reports word accuracy, fluency and completeness. It does not give a phoneme-by-phoneme diagnosis of your r, verify a regional pitch pattern or award a proficiency certificate. Use the result to select something to inspect, then return to the word and its model.

Free pronunciation practiceUse the Japanese pronunciation test for a short reading sample, then review one length contrast from this guide with reliable audio.Test your pronunciation →

8. Carry one contrast into something you want to say

Choose a word you will actually use, perhaps a family term, a place name or an item you need. Confirm its reading, listen and mark the timing feature that matters. Put it into one meaningful sentence. This makes the sound practice easier to revisit than a long list of disconnected syllables.

In a later session, check the same word in a different sentence or from a different speaker. Keep any accepted regional difference separate from a lost length contrast. The schedule is yours to set; the useful habit is returning to a clear question about what you heard and what your listener understood.

Frequently asked questions

Are Japanese long vowels just stressed vowels?

No. Vowel length can distinguish words independently of emphasis. Keep the vowel continuous for its longer interval, and learn the word's pitch pattern separately with audio. Increasing volume alone does not reproduce a length contrast.

Do I pronounce small っ as tsu?

Not in the ordinary word examples here. It indicates a held or extended following consonant. Full-sized つ has its own pronunciation. Check the character size and the word's recording rather than reading both characters the same way.

Must every mora take exactly the same time?

No mechanical stopwatch rule is needed. Mora counting is a useful learning model for preserving important units. Natural duration changes with the sounds and speaking context; use normal speech recordings alongside any slow counting exercise.

Can software tell me exactly how to move my tongue for Japanese r?

The test linked here does not diagnose tongue movement or individual Japanese phonemes. A teacher can listen to a specific example and give an articulation suggestion. Audio comparison and a comfortable brief contact are better starting points than forcing an exaggerated movement.

Sources

  1. TUFS: morasChecked:
  2. TUFS: long vowelsChecked:
  3. TUFS: mora obstruentChecked:
  4. TUFS: ra rowChecked:

Keep practising

Flai Speaking app

Practise Japanese with an AI coach

Flai Speaking helps you speak Japanese every day with short conversations and instant pronunciation feedback.