The pronunciation of Modern Standard Chinese is based on the Beijing dialect, which is of the Northern, or Mandarin, type.
Whereas English pronunciation is based on accents, Japanese pronunciation is based on pitch.
In a nutshell, the first system's pronunciations are based on Chinese (音読み on'yomi "Chinese reading"), since the Japanese borrowed their kanji (ideographic, or idea-expressing, characters) from the Chinese language centuries ago.
The pronunciation representation is based on the SAMPA standard and also encodes the position of the primary stress.
The teaching and learning method is based on a one-to-one learning process through an oral tradition of memorizing chapters and practicing chanting with Tai Dam pronunciation.
Based on pronunciation from nonstandardized speech sounds, the Wade-Giles system contains like symbols representing different sounds (e.g., Pinyin j, q, zh, and ch are rendered in Wade-Giles as ch and ch'), and different symbols expressing the same sound (ts and tz for Pinyin z).
To get the optimal set of subword units, the pronunciation prediction is first based on syllables, which are then converted to a more specific subword units (similar to morphemes), according to a certain lexicon segmentation that obtains the highest language model score for each pronunciation in the lexicon.
Errors based on pronunciation or sound similarity is another main issue when data is input.
A person sending a text message begins by inputting the pinyin transliteration of a Chinese character, based on its pronunciation, which typically involves entering a few Latin letters using multi-tap.
In 1956 a new system of romanization called Pinyin, based on the pronunciation of the characters in the Beijing dialect, was adopted as an educational instrument to help in the spread of the modern standard language.
