4 ms·
It doesn't like spelling mistakes. Ask it to say " The Aristoocrats!" with Gilbert Goddfried.
by maxerickson 6y ago
It doesn't like spelling mistakes.
Ask it to say " The Aristoocrats!" with Gilbert Goddfried.
- echelon 6y agoIt's currently sourcing phonemes from a lookup table called CMUdict, which is constructed by Carnegie Mellon [1]. That database has 140,000 entries, but even so, you'd be surprised how many common words are omitted. And of course it is missing terms for things like "pokemon" and "fortnite", which I had to add myself. I don't have generic grapheme -> phoneme/polyphone prediction, but that's something I look to add soon. In my literature review I didn't see anything in this space, so I was thinking I might have to come up with something novel. [1] http://www.speech.cs.cmu.edu/cgi-bin/cmudict http://www.speech.cs.cmu.edu/cgi-bin/cmudict
- nmstoker 6y agoEspeak-ng has pretty decent English word to phoneme translation. You run it in the mode where it just outputs the IPA. The vocabulary can be extended too (as the coverage is good but far from perfect)
- totetsu 6y agoEspeak-ng also has extensive language rules to turn things like letter and abbreviations into pronounceable text https://github.com/espeak-ng/espeak-ng/blob/master/dictsource/en_list https://github.com/espeak-ng/espeak-ng/blob/master/dictsourc... you can get it's ipa output with "speak-ng -q --ipa=3"