coqui-ai
/
TTS
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update docstring for clean_gruut_phonemes

pull/506/head
Michael Hansen 2021-06-10 10:08:01 -04:00 committed by Eren Gölge
parent 07e8ff193a
commit da6f6a4a01
1 changed files with 13 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
TTS/tts/utils/text/__init__.py
View File

@ -34,13 +34,21 @@ GRUUT_TRANS_TABLE = str.maketrans("g", "ɡ")
def clean_gruut_phonemes(ph_list):
"""Decompose, substitute, and clean gruut phonemes for TTS.
Parameters:
ph_list (list[str]): list of phonemes from gruut
gruut phonemes may contain any IPA characters (e.g., "ẽ" for the nasalized
"e"), and may be composed of multiple characters (e.g., "aɪ" in the English
"r[i]ce").
TTS phonemes come from a fixed set of symbols, and do not include every
possible variation of every vowel/consonant. Here, we decompose dipthongs,
etc. into single characters and then filter out Unicode combining characters
such as ties. This ensures that (most) phonemes will exist in the TTS symbol
table.
Args:
ph_list (list[str]): list of phonemes from gruut
Returns:
clean_list (list[str]): decomposed/clean list of phonemes for TTS
Dipthongs, etc. are decomposed into single characters
Unicode combining characters are removed (e.g., ties)
clean_list (list[str]): decomposed/clean list of phonemes for TTS
"""
cleaned_phonemes = []