AlexaKorean

Alexa (Amazon Echo) supports English voice input and output only now. But you can provide a phonemic pronunciation or play mp3 clips by using Speech Synthesis Markup Language (SSML). Although SSML supports a limited set of symbols, this script mimics Korean speech as close as possible.

Usage

output1 = AlexaKorean.speak("BBQ양념치킨")
// <say-as interpret-as="characters">BBQ</say-as><phoneme alphabet="ipa" ph="jɑŋ.njʌm.t͡ʃi.kin.">양념치킨</phoneme>
output2 = AlexaKorean.speak("그래 오늘은 안녕", AlexaKorean.IPA) // default: IPA
// <phoneme alphabet="ipa" ph="g.ɹɛ.">그래</phoneme> <phoneme alphabet="ipa" ph="oʊ.nl.n.">오늘은</phoneme> ...
output3 = AlexaKorean.speak("hello world 예제", AlexaKorean.SAMPA)
// hello world <phoneme alphabet="x-sampa" ph="jE.dZE.">예제</phoneme>

format

• "철수{는} 서울{으로} 가서 음료{을} 마신다." is turned into "철수는 서울로 가서 음료를 마신다.". Affected characters are {와}, {과}, {으로}, {로}, {은}, {는}, {을}, {를}, {이} and {가}.
• "123456 -3.141592" speaks numbers.
• "{010-0000-0000} Intel {386}" speaks each digit separately.

You can hear outputs at Test > Voice Simulator in developer console.

Known Issue

• affected 'ㅗ' (hog) and 'ㅜ' (foot) sounds

TODO

• complete a phonemic mapping table
• support ordinal numbers
• support symbols
• support phonological rules (wikipedia, reference)
• tell short vowels from long vowels
• testcases
• JavaScript version
• package, deployable to AWS Lambda
• sample output