Tone-melody matching

Level 6Sounds Like hip3 jam1

協音 (hip3 jam1) is the essential practice in Cantonese songwriting where the 6 tones of the language must align with the rising and falling notes of a melody to ensure the lyrics remain intelligible.

How It Works

In Cantonese music, the relationship between lyrics and melody is much stricter than in Mandarin or English. While Mandarin songs largely ignore or overwrite lexical tones in favor of the melody, Cantonese relies so heavily on pitch height for word identity that singing off-tone creates unintelligible 'twisted sounds' known as 拗音 (aau3 jam1, or sometimes aau2 jam1), which practically means 'off-key' or 'mispronounced due to melody'. To avoid this, lyricists use 協音 (hip3 jam1). Note that 協音 is also written 叶音 (hip3 jam1), the traditional linguistic term — 叶 here is the ancient variant of 協 meaning "to harmonise", not 葉 (jip6, "leaf"), which only later borrowed the same simplified form. Colloquially, this is called 啱音 (ngaam1 jam1, literally "correct/fitting sound").

The 0243 System:
Formalized and popularized in the 1980s by renowned Hong Kong music critic and lyricist Wong Chi-wah (黃志華), the name '0243' is a mnemonic based on the digits 零 (ling4, Tone 4), 二 (ji6, Tone 6), 四 (sei3, Tone 3), and 三 (saam1, Tone 1). Crucially, these digits are not in mathematical order. Beginners must remember that the pitch hierarchy from lowest to highest is strictly 0 < 2 < 4 < 3. These represent four relative pitch levels, which can be roughly mapped to Solfège syllables. This does not mean a song must be sung in a specific key with exact Do-Re-Mi notes, but rather it dictates the intervals and direction between adjacent syllables:

  • Lowest Pitch (0 ≈ Do): Tone 4 (Low Falling). Note that in modern Cantonese, and especially in music, Tone 4 functions practically as an Extra-Low Level tone (11), which is why it maps so easily to a single, flat, low musical note.
  • Mid-Low Pitch (2 ≈ Re): Tone 6 (Low Level) and Tone 9 (Low Checked).
  • High-Mid Pitch (4 ≈ Mi): Tone 3 (Mid Level) and Tone 8 (Mid Checked).
  • Highest Pitch (3 ≈ Fa/So): Tone 1 (High Level) and Tone 7 (High Checked).

While checked tones (Tones 7, 8, and 9) map to these pitch groups, they have a crucial musical restriction: because words ending in -p, -t, or -k are physically abrupt, they cannot be sustained. Therefore, they are normally not placed on long, held musical notes, even if the pitch matches. For example, the word 大都會 (daai6 dou1 wui6) has the tones 6-1-6, which maps to the 0243 pitch levels 2-3-2 (Re-So-Re).

Rising Tones and the Full 0-9 System:
A brilliant feature of Cantonese is that the digits 0 to 9 can be used as mnemonics because their own pronunciations perfectly cover the 9 tones. While 0, 2, 4, and 3 cover the flat tones, 九 (gau2, Tone 2) and 五 (ng5, Tone 5) are the standard mnemonics used to map the rising tones. Tones 2 and 5 are unique because they are contour tones: the pitch slides upward while the syllable is being said. That is why they sit outside the base 0243 ladder — 0243 maps each tone to a single flat musical note, and a tone whose pitch moves cannot be pinned to one note. It also lets the melody 'jump' or slide upward on those syllables. For the checked tones, 七 (cat1, Tone 7) and 八 (baat3, Tone 8) are used. Modern lyricists often use the extended 02493 system, where the digit '9' (九) specifically represents Tone 2 to distinguish its rising contour from the flat Tone 1.

The Golden Rule:
If the melody moves upward, the lexical tone of the second word should be higher in the 0243 hierarchy. If the melody moves downward, the tone should be lower. If the melody stays flat, the tones should belong to the same pitch group. Proficiency in this system allows Cantonese listeners to understand song lyrics much more easily without relying solely on printed text, though homophones and poetic phrasing still make lyric sheets useful.

The Lyricist's Workflow:
In practice, a lyricist starts by listening to a raw melody. They then transcribe the melodic contour into a string of these pitch digits. For example, a melodic phrase might be transcribed as 3-4-2-0-2-4-3. Once the melody is mapped out in numbers, the lyricist brainstorms Cantonese characters whose natural tones match that exact number string, ensuring the resulting lyrics are both poetic and perfectly 協音.

Examples (6)

  1. 除咗〈帝女花〉之外,「零二四三」(ling4 ji6 sei3 saam1)同「三四二零」(saam1 sei3 ji6 ling4)就係協音嘅最好例子。

    Besides 'The Floral Princess', '0-2-4-3' and '3-4-2-0' are the best examples of tone-melody matching.

    Word by Word: Besides Floral Princess outside, 0-2-4-3 and 3-4-2-0 exactly is harmonize-sound possessive best example.

    These two sequences are used by lyricists to test melodies. '0-2-4-3' (零二四三) follows an ascending musical scale (Low → Mid-Low → Mid → High), while '3-4-2-0' (三四二零) follows a descending one, perfectly matching the relative pitch of the Cantonese tones.

  2. 「我愛你」(ngo5 oi3 nei5) 呢三個字唔可以唱同一個音。

    The three words in 'I love you' (ngo5 oi3 nei5) cannot be sung on the same musical note.

    Word by Word: I love you these three CL word not can sing same one CL note.

    In 協音, 'I' (ngo5, Tone 5), 'love' (oi3, Tone 3), and 'you' (nei5, Tone 5) do not all share one pitch: Tone 3 sits higher than Tone 5. If a composer puts them all on the same note, it results in 拗音 (aau3 jam1), making the lyrics sound like nonsense or different words entirely.

  3. 好兒童 (hou2 ji4 tung4)

    Good children

    Word by Word: Good child-ren

    In the movie McDull, this is used as a joke. Because the melody doesn't match the tones, 'Good children' (hou2 ji4 tung4) sounds like 'Sore ears' (hou2 ji5 tung3). This is a classic case of 拗音 (aau3 jam1).

  4. 林子祥嘅《數字人生》係協音嘅好例子。

    George Lam's 'Numbers and Life' is a good example of tone-melody matching.

    Word by Word: Lam Tsz-cheung possessive Numbers Life is harmonize-sound possessive good example.

    George Lam's song 'Numbers and Life' (sou3 zi6 jan4 sang1) is famous for singing strings of numbers where the melody matches the Cantonese tones of each digit, demonstrating the 0243 system in action.

  5. 呢首歌唔啱音。

    This song's lyrics don't match the melody.

    Word by Word: This CL song not correct sound.

    啱音 (ngaam1 jam1) is the colloquial way to say a song follows the principle of 協音 (hip3 jam1).

  6. 填詞人要考慮協音。

    Lyricists need to consider tone-melody matching.

    Word by Word: Fill-word-person need consider harmonize-sound.

    填詞人 (tin4 ci4 jan4) refers to a professional lyricist, for whom 協音 is a fundamental skill to avoid 拗音 (aau3 jam1).

Common Mistakes

  • Assuming singing overrides tones like in Mandarin.

    唱「我哋」(ngo5 dei6) 嘅時候,音符一路向上行。
    唱「我哋」(ngo5 dei6) 嘅時候,音符跟住字調低低哋行。

    Cantonese lyrics are normally written so that each word’s tone follows the melody. If a low-tone word like 我哋 (ngo5 dei6) is sung on rising high notes, listeners may hear a different word.

  • Confusing 協音 (hip3 jam1) with 協韻 (hip3 wan5).

    呢首歌嘅歌詞好協音,每句尾字都押同一個韻。
    呢首歌嘅歌詞好協韻,每句尾字都押同一個韻。

    協音 refers to matching tones to melody. 協韻 (hip3 wan5) or 叶韻 (hip3 wan5) refers to rhyming (matching the final sounds of words).

How It Connects

Tap a pattern to see what it builds on and what it leads to.

Tone-melody matchingLevel 6

Learn This First

Tone-melody matching (hip3 jam1) is a high-level concept in Cantonese music and linguistics. To understand it, a learner must first have a solid grasp of the Cantonese tone system and how tones are categorized (high, mid, low, rising). It also requires an awareness of the specific register used in Cantopop lyrics, where the linguistic tones must correspond to the melodic contour to ensure the lyrics remain intelligible when sung.

Learn Next

Patterns that build on this one.