Tones: the pitch is the word
In Cantonese and Mandarin, pitch is not expression — it is spelling. Change the pitch of si and you move from poem to history to time to market. Below: every tone drawn, played on its own, and played back to back against any other, both by a real voice and by the synthesiser (which gives you the pitch with nothing else in the way).
Cantonese — 6 tones — or 9, depending on what you are counting
Both numbers are right, and they answer different questions. Cantonese has SIX distinct pitch shapes: high level, high rising, mid level, low falling, low rising, low level (Jyutping 1–6). The traditional count of NINE comes from classical Chinese phonology, which also splits by syllable type: the three extra 'entering tones' (入聲) are syllables that end in a hard -p, -t or -k and stop dead. Their pitch is not new — it is exactly tone 1, tone 3 and tone 6 — they are just short. So: 9 traditional categories, 6 pitches to learn. Old textbooks and Hong Kong schooling say 9; linguists and Jyutping say 6. Nobody is lying to you.
The 6 pitches
The line is the pitch across the syllable, on the five-level scale phonologists use (5 = top of your range, 1 = bottom). Every ▶ here is the synthesiser: pure pitch, no voice, identical every time — which is what you want when you are comparing two shapes.
Some speakers and older recordings use a high FALLING variant (˥˧) for the same tone — free variation in modern Hong Kong, and the one nuance that genuinely has two forms.
Rises from mid to high — the tone of a surprised question in English.
Flat, in the middle of your range. The reference point for the others.
Starts low and sags lower; often creaky at the bottom.
The one most often confused with tone 2 — both rise, but 5 starts at the bottom and 2 starts in the middle.
Flat and low. Confused with 4 (which falls) and 3 (which is higher).
One syllable, every tone
The classic demonstration: si said 6 ways. Same consonant, same vowel — only the pitch moves, and every one is a different word. You have no Cantonese voice installed, so these play on the synthesiser — pitch is exact, the timbre is a robot.
| Tone | Shape | Word | Means | Hear it |
|---|---|---|---|---|
| 1 high level | 詩 si1 | poem | ||
| 2 high rising | 史 si2 | history | ||
| 3 mid level | 試 si3 | to try | ||
| 4 low falling | 時 si4 | time | ||
| 5 low rising | 市 si5 | market | ||
| 6 low level | 是 si6 | to be |
Compare two, back to back
Tones are easiest to hear against each other, not alone. Pick a pair and it plays them A–B–A–B.
The three 'extra' tones (7, 8, 9)
These are the entering tones — syllables ending in -p, -t or -k, which cut the vowel short. Listen to the pitch: 7 is tone 1's pitch, 8 is tone 3's, 9 is tone 6's. Nothing new to learn except the stop at the end, which is unreleased: your mouth closes and no puff follows.
| Word | Jyutping | Means | Traditional slot | Hear it |
|---|---|---|---|---|
| 色 | sik1 sɪk˥ | colour | traditional 7 · 陰入 — pitch of tone 1 | |
| 錫 | sek3 sɛːk˧ | tin; to kiss | traditional 8 · 中入 — pitch of tone 3 | |
| 食 | sik6 sɪk˨ | to eat | traditional 9 · 陽入 — pitch of tone 6 |
Changed tone (變調 bin3 diu6)
Cantonese often lifts a syllable to tone 2 (high rising) to mean 'the familiar one': 女 neoi5 'woman' → 女 neoi2 'daughter/girl'; 魚 jyu4 'fish' → 魚 jyu2 in a dish name; 糖 tong4 'sugar' → tong2 'sweets'. Shop names, nicknames and food words are full of it. Dictionaries usually list the citation tone, so what you hear on the street will sometimes be a step higher.
Tones 2 and 5 are the real merger risk
Many younger Hong Kong speakers merge low rising (5) toward high rising (2), and low falling (4) toward low level (6). If you cannot hear 5 vs 2 yet, you are in good company — train that pair first.
More: Cantonese spelling → sound · Cantonese grammar & words · practise telling them apart
Mandarin — 4 tones + a neutral one
Mandarin's four tones are famously demonstrated on the syllable ma. The fifth 'neutral' tone is not a pitch at all — it is a syllable with no tone of its own, short and light, taking its height from whatever came before. Pinyin marks tones with accents (mā má mǎ mà) or numbers (ma1 ma2 ma3 ma4); this site accepts both.
The 5 pitches
The line is the pitch across the syllable, on the five-level scale phonologists use (5 = top of your range, 1 = bottom). Every ▶ here is the synthesiser: pure pitch, no voice, identical every time — which is what you want when you are comparing two shapes.
Held high and flat — higher than most beginners pitch it.
Like the English 'huh?' rise.
In real speech, only the LOW part survives unless the syllable is final or stressed — the textbook full dip-and-rise is rarer than you think.
Sharp fall from the top — like a curt English 'No.'
Short, unstressed, pitch borrowed from the syllable before. 媽媽 māma, 東西 dōngxi.
One syllable, every tone
The classic demonstration: ma said 5 ways. Same consonant, same vowel — only the pitch moves, and every one is a different word. You have no Mandarin voice installed, so these play on the synthesiser — pitch is exact, the timbre is a robot.
| Tone | Shape | Word | Means | Hear it |
|---|---|---|---|---|
| 1 high level | 妈 mā | mother | ||
| 2 rising | 麻 má | hemp; numb | ||
| 3 dipping | 马 mǎ | horse | ||
| 4 falling | 骂 mà | to scold | ||
| neutral neutral | 吗 ma | (question particle) |
Compare two, back to back
Tones are easiest to hear against each other, not alone. Pick a pair and it plays them A–B–A–B.
Third tone sandhi
Two third tones in a row: the first becomes a second. 你好 is written nǐ hǎo but said ní hǎo. With three in a row it depends on the phrasing — 我很好 is wǒ hěn hǎo → wó hén hǎo.
不 and 一 change tone
不 bù drops to bú before a fourth tone (不是 bú shì). 一 yī becomes yì before tones 1–3 (一起 yìqǐ) and yí before a fourth tone (一样 yíyàng).
More: Mandarin spelling → sound · Mandarin grammar & words · practise telling them apart
Why the counts differ everywhere you look
Cantonese is the sharpest case. Ask a Hong Kong secondary school and the answer is nine; ask a linguist and it is six. Both are describing the same language:
- Six counts distinct pitch shapes — the thing your mouth has to learn. Jyutping numbers them 1–6.
- Nine is the classical count, which also splits by syllable type. Syllables ending in -p, -t or -k (the 入聲 entering tones) get their own categories — three of them — even though their pitches are simply tones 1, 3 and 6 on a shortened syllable.
- You will also see seven, from older descriptions that kept a separate high-falling tone 1 (˥˧) alongside high-level (˥). Most Hong Kong speakers merged those; hearing either is normal.
Mandarin has the same kind of history without the arithmetic: it lost the entering tones entirely, which is why it has four where Cantonese has six, and why Cantonese preserves rhymes in Tang poetry that no longer rhyme in Mandarin.