# Understanding Jyutping and Cantonese Romanization

If you ask a local in Hong Kong how to spell their address in English, they will instinctively give you a specific string of letters: *Kwun Tong*, *Tsim Sha Tsui*, or *Sham Shui Po*. But if you ask that same person to type those words using "Jyutping," you might be met with a blank stare.

Unlike Mandarin, which has universally adopted Pinyin as its standard romanization system, Cantonese has a long, chaotic history of spelling out sounds. For a Cantopop lyricist, music producer, or language learner, understanding how Cantonese is romanized—and how locals actually type—is essential to navigating dictionaries and writing lyrics that flow.

Here is a breakdown of the different Cantonese romanization systems, how Input Method Editors (IMEs) work, and how **canTone** brings order to the chaos.

### The Big Three Systems

Over the last century, academics and missionaries have tried to map the sounds of Cantonese to the English alphabet. While there are dozens of obscure systems, three matter today:

#### 1. Yale (The Legacy Standard)
Developed in the 1940s, Yale was the standard for teaching Cantonese to English speakers for decades. It uses a mix of letters and diacritic marks (like `ā`, `á`, `à`) or numbers to denote tones. While highly accurate, typing diacritics on a standard keyboard is a nightmare, leading to its slow decline in the digital age.

#### 2. Jyutping (The Modern Standard)
Created by the Linguistic Society of Hong Kong (LSHK) in 1993, **Jyutping (粵拼)** is the undisputed standard for modern software, NLP (Natural Language Processing), and digital dictionaries. 
* It uses only standard English letters and numbers (1-6) for tones. 
* It is mathematically consistent.
* For example, "Hong Kong" is `hoeng1 gong2`.

#### 3. Hong Kong Romanisation (The Reality)
This isn't an academic system at all—it’s the spelling conventions used in Hong Kong for street signs, MTR stations, and personal names.
* Instead of the Jyutping `z` or `c`, it uses `ts` or `ch` (e.g., *Tsim Sha Tsui* instead of *Zim1 Saa1 Zeoi1*).
* Instead of `gw`, it often uses `kw` (*Kwun Tong* instead of *Gwun1 Tong4*).
* It does not mark tones at all.

### The Cantonese IME Problem

Because of this historical divide, typing Cantonese on a computer or phone is famously fragmented. 

Historically, locals bypassed phonetics entirely. Most people in Hong Kong learned shape-based typing systems like **Cangjie (倉頡)** or **Quick (速成)**. You don't need to know how a word sounds to type it; you just need to know how to draw it. 

However, as younger generations shift toward phonetic IMEs (typing the sound on a QWERTY keyboard), a massive friction point has emerged. Most modern IMEs demand strict Jyutping. But because native speakers grew up looking at MTR signs, forcing them to type `c-a-a` to get the word for tea (茶) feels entirely unnatural when their brain expects to type `c-h-a`.

### How canTone Handles Jyutping

When we built **canTone**, we knew that in order to do complex mathematical operations—like matching tones to melodies (`0243`) or chaining exact rhymes (`>港`)—the underlying database had to be perfectly structured. 

#### The Backbone: words.hk (粵典)
Our engine relies on the incredible open-source work of various dictionaries like **[words.hk (粵典)](https://words.hk/)**. These provde the strict, pristine Jyutping data that powers our search engine. Without their tireless work documenting modern slang, idioms, and classical vocabulary, canTone's offline melody-matching would simply not be possible.

Every result you see in the app is mapped to strict LSHK Jyutping, ensuring that when you search for a rhyme or a melodic tone contour, the results reflect real life pronunciations. 

#### The Solution: The canTone Custom Vocab IME
But what happens when you need to add a brand new piece of local slang or a project-specific word to your Custom Vocabulary? 

We didn't want to force songwriters to become linguistics professors just to add a word to their dictionary. So, we built our Custom Vocab IME to act as a bridge between the rigid math of Jyutping and the intuitive "Hong Kong Style" spelling.

When you type a phonetic approximation into the canTone custom importer, our fuzzy parser does the heavy lifting:
* If you type `ch` (like in *Cheung Chau*), the IME knows you likely mean the Jyutping `c` or `z`.
* If you type `sh` (like in *Sham Shui Po*), the engine translates it to `s`.
* If you type `kw` (like in *Kwun Tong*), it intelligently maps it to `gw`.

> **For Songwriters:**
> You can type your custom vocabulary using the intuitive, English-approximate spelling you grew up with. The app will automatically generate the strict Jyutping and assign the correct tone numbers. This means your custom slang instantly becomes available to the advanced Rhyme (`>`) and Melody (`0243`) operators, perfectly formatted and ready to be written into your next track.

### The Best of Both Worlds
Accurate romanization is the secret weapon of Cantonese songwriting. It is what allows the canTone engine to predict exactly how a word will sound when sung across a melody. By using *words.hk* as our source of truth for the core dictionary, and providing a flexible, HK-style IME for your personal workspace, you get the absolute precision of Jyutping without the academic headache.