Open EuphoriaCelestial opened 3 years ago
The lab files there are "grapheme" level (words are separated by spaces), the orthography used is based off the phones (though the phone level has tones associated to vowels rather than at the ends of syllables). Grapheme level transcriptions of .wav files are fine, you just need a pronunciation dictionary that maps them into phones.
The lab files there are "grapheme" level (words are separated by spaces), the orthography used is based off the phones (though the phone level has tones associated to vowels rather than at the ends of syllables). Grapheme level transcriptions of .wav files are fine, you just need a pronunciation dictionary that maps them into phones.
oh, those lab files are in grapheme level? I dont know Mandarin so I thought the lab files in Example 2 was in phoneme level I am using MFA to generate alignments and TextGrid files for FastSpeech2 and everything need to be in phoneme level. Is MFA able to do this?
Hi, can someone please guide me how to generate .lab files in this Mandarin example corpus in example 2 of the docs? https://drive.google.com/file/d/1zPfwvTE_x7o9iX8J8bzeb0KNHEi3jrgN/edit
I see it basically phoneme level transcriptions of .wav files. I am having grapheme level transcriptions of my .wav files p/s: I think we should add this in the docs too, for beginner like me