Looking for large word list file/db (>1k) with IPA transcriptions

Question

I'm building a mobile app to practice/teach IPA. The idea is you look at a word and then input the correct IPA. I've used some dictionary APIs, but they either:

• Don't include the IPA

• If it includes the IPA, it's not using American accent (Oxford API)

• Is super slow (https://www.dictionaryapi.com/)

• Doesn't use standard IPA. Instead of using 'ʊ' uses 'o͝o' . I've never even seen that before and doesn't work with the rest of my app (http://developer.wordnik.com/docs.html#!/word/getTextPronunciations_get_5)

I would rather just use a csv file of at least 1,000 words and store my dictionary locally. Is there somewhere I can get a large list, or a file containing a bunch of words with IPA using American accent? Ideally, someone out there has a file on github.

Answer 1

The CMU Dictionary is easily convertible if you are willing to do a tiny bit of programming. Each phoneme is surrounded by space and there's a pretty obvious mapping from e.g. "AE" is [æ], "AH" is [ə/ʌ], "AY" is [ai], the numbers indicate primary, secondary versus no stress. The advantage of this as a source is that (for English) there are very many ways to transcribe words ([e: ei ɛɪ ɛi ej ej] in "save"). If you want to write that vowel as [ei] then EY becomes [ei]. This is American English, not sure what dialect, but it records "fortune" with [tʃən] and "tune" with [tuwn], that is, it is not the "tyune" dialect. You could extract the most frequent 1000 or 5000 words (they do include a lot of non-English place and personal names).