What Is Transliteration?

A concrete example

The Chinese characters 你好 mean "hello". Transliterated into the Latin alphabet, they become:

nǐ hǎo is not an English word, and it does not mean anything in English. It is a way of writing the Chinese pronunciation using letters an English reader can pronounce. The translation would be hello; the transliteration is nǐ hǎo.

Transliteration vs translation

This is the distinction people most often ask about. Translation carries meaning across languages; transliteration carries sound across writing systems.

Comparison of transliteration and translation

Transliteration	Translation
Changes the writing system	Changes the language
Leaves the words as they are	Replaces the words with different ones
Aims to reflect pronunciation	Aims to convey meaning
你好 → nǐ hǎo	你好 → hello

Transliteration vs transcription

These two are easier to confuse, because both deal with sound. The difference is what they start from.

Transliteration starts from written text and maps characters from one script to another, ideally so the original spelling can be reconstructed. Transcription starts from speech and records how something is actually pronounced, often using a phonetic alphabet such as the IPA, without concern for the original spelling.

In practice the line blurs: most everyday "transliteration" systems, including the ones this tool uses, are partly phonetic and do not allow perfect reconstruction of the original spelling.

What about romanization?

Romanization is transliteration where the target script is specifically the Latin (Roman) alphabet. Every conversion on this site is a romanization: Chinese into Pinyin, Japanese kana into romaji, Cyrillic into Latin letters, and so on. Transliteration is the broader term and can go between any two writing systems, such as Cyrillic into Greek.

Why transliteration is rarely exact

Writing systems do not encode the same information, so something is usually lost or guessed. Three examples of what makes it hard:

• Missing vowels. Arabic and Hebrew normally omit short vowels in writing. Without them, the consonants can be transliterated with confidence but the vowels cannot be recovered from the text alone.
• Characters with more than one reading. A Chinese character can be pronounced differently depending on the word it appears in: 行 is háng in 银行 (bank) but xíng elsewhere. Getting this right needs a dictionary, not a character table.
• Sounds with no equivalent. Many languages have sounds the Latin alphabet has no letter for, which is why romanization systems rely on diacritics like ṇ, ś or ā, and why competing systems can spell the same word differently.

Common romanization systems

Most languages have one or more established standards. These are the ones this tool follows:

Romanization systems by language

Language	System	Example
Chinese	Hanyu Pinyin	你好 → nǐ hǎo
Japanese	Hepburn	コンピューター → konpyūtā
Korean	Revised Romanization	안녕하세요 → annyeonghaseyo
Russian	BGN/PCGN	Привет → Privet
Indic scripts	ISO 15919	வணக்கம் → vaṇakkam