Korean Phonetics

About
Unlike previous English and Japanese Vocaloids, the Korean engine is only partially based on X-SAMPA notation. This is due to the nature of hangul, which already functions as a phonetic alphabet. The voicebank's coding uses romanized hangul for the consonants and X-SAMPA based coding for the vowels. A possible reason why romanization is used for the coding of the consonants is due to the korean language's many allophones. Non-native speakers will find the Korean Phonetics are more complex then either English or Japanese to work with and it is essential that they study how to the Korean language works when using the engine.

Consonant Assimilation
When entering notes with hangul input, the Vocaloid 3 editor automatically detects and edits the phonetic values for cases in which consonant-to-consonant assimilation occurs. However, a flaw with this system is that the editor changes phonetic values regardless of the distance between notes. For example, "언어" (language) entered into the editor will produce [7][n 7]. However, when the two syllables are seperated with a pause inbetween, the correct pronunciation should be [7 np][7]

Vowel Assimilation
Like Japanese, consonants in Korean become palatalized before /i/ and /j/. However, unlike Japanese, Korean consonants also become labialized before /o/, /u/, and /w/ as well.