From 107d45ee5cb8d98fc707e9bb442434385689edab Mon Sep 17 00:00:00 2001 From: wangjie Date: Fri, 29 Mar 2024 10:29:38 +0800 Subject: [PATCH] enhance chinese pinyin dict --- TTS/tts/utils/text/chinese_mandarin/phonemizer.py | 4 ++++ TTS/tts/utils/text/phonemizers/zh_cn_phonemizer.py | 3 ++- requirements.txt | 1 + 3 files changed, 7 insertions(+), 1 deletion(-) diff --git a/TTS/tts/utils/text/chinese_mandarin/phonemizer.py b/TTS/tts/utils/text/chinese_mandarin/phonemizer.py index 727c881e..b0c45606 100644 --- a/TTS/tts/utils/text/chinese_mandarin/phonemizer.py +++ b/TTS/tts/utils/text/chinese_mandarin/phonemizer.py @@ -4,8 +4,12 @@ import jieba import pypinyin from .pinyinToPhonemes import PINYIN_DICT +from pypinyin_dict.pinyin_data import kxhc1983 +def enhance_dict(): + kxhc1983.load() + def _chinese_character_to_pinyin(text: str) -> List[str]: pinyins = pypinyin.pinyin(text, style=pypinyin.Style.TONE3, heteronym=False, neutral_tone_with_five=True) pinyins_flat_list = [item for sublist in pinyins for item in sublist] diff --git a/TTS/tts/utils/text/phonemizers/zh_cn_phonemizer.py b/TTS/tts/utils/text/phonemizers/zh_cn_phonemizer.py index 41480c41..a314e17d 100644 --- a/TTS/tts/utils/text/phonemizers/zh_cn_phonemizer.py +++ b/TTS/tts/utils/text/phonemizers/zh_cn_phonemizer.py @@ -1,6 +1,6 @@ from typing import Dict -from TTS.tts.utils.text.chinese_mandarin.phonemizer import chinese_text_to_phonemes +from TTS.tts.utils.text.chinese_mandarin.phonemizer import chinese_text_to_phonemes, enhance_dict from TTS.tts.utils.text.phonemizers.base import BasePhonemizer _DEF_ZH_PUNCS = "、.,[]()?!〽~『』「」【】" @@ -27,6 +27,7 @@ class ZH_CN_Phonemizer(BasePhonemizer): def __init__(self, punctuations=_DEF_ZH_PUNCS, keep_puncs=False, **kwargs): # pylint: disable=unused-argument super().__init__(self.language, punctuations=punctuations, keep_puncs=keep_puncs) + enhance_dict() @staticmethod def name(): diff --git a/requirements.txt b/requirements.txt index 2944e6fa..e6b512fe 100644 --- a/requirements.txt +++ b/requirements.txt @@ -34,6 +34,7 @@ coqpit>=0.0.16 # chinese g2p deps jieba pypinyin +pypinyin-dict # korean hangul_romanize # gruut+supported langs