github/NarratoAI
1
0
Fork 0
mirror of https://github.com/linyqh/NarratoAI.git synced 2025-12-11 10:32:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(tts): 添加多引擎TTS支持并重构语音设置界面

Browse Source 
- 新增Azure Speech Services和Edge TTS引擎支持
- 重构语音设置界面，支持不同引擎的独立配置
- 添加引擎选择器和详细说明
- 更新requirements.txt添加azure-cognitiveservices-speech依赖
- 改进音色名称验证逻辑
This commit is contained in:
linyq 2025-08-03 18:45:33 +08:00
parent e59fd6030c
commit 062d317261
4 changed files with 507 additions and 78 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
app/services/voice.py
View File

@ -1058,6 +1058,27 @@ def is_azure_v2_voice(voice_name: str):
return ""
def should_use_azure_speech_services(voice_name: str) -> bool:
"""判断音色是否应该使用Azure Speech Services"""
if not voice_name or is_soulvoice_voice(voice_name):
return False
voice_name = voice_name.strip()
# 如果是带-V2后缀的，肯定是Azure Speech Services
if voice_name.endswith("-V2"):
return True
# 检查是否为Azure官方音色格式 (如: zh-CN-YunzeNeural)
# Azure音色通常格式为: [语言]-[地区]-[名称]Neural
import re
pattern = r'^[a-z]{2}-[A-Z]{2}-\w+Neural$'
if re.match(pattern, voice_name):
return True
return False
def tts(
text: str, voice_name: str, voice_rate: float, voice_pitch: float, voice_file: str
) -> Union[SubMaker, None]:
@ -1065,11 +1086,11 @@ def tts(
if is_soulvoice_voice(voice_name):
return soulvoice_tts(text, voice_name, voice_file, speed=voice_rate)
# 检查是否为 Azure V2 引擎
if is_azure_v2_voice(voice_name):
# 检查是否应该使用 Azure Speech Services
if should_use_azure_speech_services(voice_name):
return azure_tts_v2(text, voice_name, voice_file)
# 默认使用 Azure V1 引擎
# 默认使用 Edge TTS (Azure V1)
return azure_tts_v1(text, voice_name, voice_rate, voice_pitch, voice_file)
@ -1140,12 +1161,22 @@ def azure_tts_v1(
def azure_tts_v2(text: str, voice_name: str, voice_file: str) -> Union[SubMaker, None]:
voice_name = is_azure_v2_voice(voice_name)
if not voice_name:
logger.error(f"invalid voice name: {voice_name}")
raise ValueError(f"invalid voice name: {voice_name}")
# 直接使用官方音色名称，不需要V2后缀验证
# Azure Speech Services 的音色名称如: zh-CN-YunzeNeural, en-US-AvaMultilingualNeural
processed_voice_name = voice_name.strip()
if not processed_voice_name:
logger.error(f"invalid voice name: {voice_name} (empty)")
raise ValueError(f"invalid voice name: {voice_name} (empty)")
text = text.strip()
# 检查Azure Speech SDK是否可用
try:
import azure.cognitiveservices.speech as speechsdk
except ImportError as e:
logger.error("Azure Speech SDK 未安装。请运行: pip install azure-cognitiveservices-speech")
logger.error("或者使用 Edge TTS 引擎作为替代方案")
return None
def _format_duration_to_offset(duration) -> int:
if isinstance(duration, str):
time_obj = datetime.strptime(duration, "%H:%M:%S.%f")
@ -1164,9 +1195,7 @@ def azure_tts_v2(text: str, voice_name: str, voice_file: str) -> Union[SubMaker,
for i in range(3):
try:
logger.info(f"start, voice name: {voice_name}, try: {i + 1}")
import azure.cognitiveservices.speech as speechsdk
logger.info(f"start, voice name: {processed_voice_name}, try: {i + 1}")
sub_maker = SubMaker()
@ -1185,7 +1214,7 @@ def azure_tts_v2(text: str, voice_name: str, voice_file: str) -> Union[SubMaker,
speech_config = speechsdk.SpeechConfig(
subscription=speech_key, region=service_region
)
speech_config.speech_synthesis_voice_name = voice_name
speech_config.speech_synthesis_voice_name = processed_voice_name
# speech_config.set_property(property_id=speechsdk.PropertyId.SpeechServiceResponse_RequestSentenceBoundary,
# value='true')
speech_config.set_property(

16
config.example.toml
View File

@ -92,6 +92,22 @@
# 默认模型（可选）
model = "FunAudioLLM/CosyVoice2-0.5B"
[ui]
# TTS引擎选择 (edge_tts, azure_speech, soulvoice)
tts_engine = "edge_tts"
# Edge TTS 配置
edge_voice_name = "zh-CN-XiaoyiNeural-Female"
edge_volume = 80
edge_rate = 1.0
edge_pitch = 0
# Azure Speech Services 配置
azure_voice_name = "zh-CN-XiaoyiNeural-Female"
azure_volume = 80
azure_rate = 1.0
azure_pitch = 0
[proxy]
# clash 默认地址：http://127.0.0.1:7890
http = ""

2
requirements.txt
View File

@ -29,7 +29,7 @@ google-generativeai>=0.8.5
# python-multipart~=0.0.9
# redis==5.0.3
# opencv-python~=4.10.0.84
# azure-cognitiveservices-speech~=1.37.0
azure-cognitiveservices-speech~=1.37.0
# git-changelog~=2.5.2
# watchdog==5.0.2
# pydub==0.25.1

516
webui/components/audio_settings.py
View File

@ -19,6 +19,53 @@ def get_soulvoice_voices():
return ["soulvoice:custom"]
def get_tts_engine_options():
"""获取TTS引擎选项"""
return {
"edge_tts": "Edge TTS",
"azure_speech": "Azure Speech Services",
"soulvoice": "SoulVoice"
}
def get_tts_engine_descriptions():
"""获取TTS引擎详细描述"""
return {
"edge_tts": {
"title": "Edge TTS",
"features": "完全免费，但服务稳定性一般，不支持语音克隆功能",
"use_case": "测试和轻量级使用",
"registration": None
},
"azure_speech": {
"title": "Azure Speech Services",
"features": "提供一定免费额度，超出后按量付费，需要绑定海外信用卡",
"use_case": "企业级应用，需要稳定服务",
"registration": "https://portal.azure.com/#view/Microsoft_Azure_ProjectOxford/CognitiveServicesHub/~/SpeechServices"
},
"soulvoice": {
"title": "SoulVoice",
"features": "提供免费额度，支持语音克隆，支持微信购买额度，无需信用卡，性价比极高",
"use_case": "个人用户和中小企业，需要语音克隆功能",
"registration": "https://soulvoice.scsmtech.cn/"
}
}
def is_valid_azure_voice_name(voice_name: str) -> bool:
"""检查是否为有效的Azure音色名称格式"""
if not voice_name or not isinstance(voice_name, str):
return False
voice_name = voice_name.strip()
# Azure音色名称通常格式为: [语言]-[地区]-[名称]Neural
# 例如: zh-CN-YunzeNeural, en-US-AvaMultilingualNeural
import re
pattern = r'^[a-z]{2}-[A-Z]{2}-\w+Neural$'
return bool(re.match(pattern, voice_name))
def render_audio_panel(tr):
"""渲染音频设置面板"""
with st.container(border=True):
@ -33,46 +80,91 @@ def render_audio_panel(tr):
def render_tts_settings(tr):
"""渲染TTS(文本转语音)设置"""
# 1. TTS引擎选择器
# st.subheader("🎤 TTS引擎选择")
engine_options = get_tts_engine_options()
engine_descriptions = get_tts_engine_descriptions()
# 获取保存的TTS引擎设置
saved_tts_engine = config.ui.get("tts_engine", "edge_tts")
# 确保保存的引擎在可用选项中
if saved_tts_engine not in engine_options:
saved_tts_engine = "edge_tts"
# TTS引擎选择下拉框
selected_engine = st.selectbox(
"选择TTS引擎",
options=list(engine_options.keys()),
format_func=lambda x: engine_options[x],
index=list(engine_options.keys()).index(saved_tts_engine),
help="选择您要使用的文本转语音引擎"
)
# 保存TTS引擎选择
config.ui["tts_engine"] = selected_engine
# 2. 显示引擎详细说明
if selected_engine in engine_descriptions:
desc = engine_descriptions[selected_engine]
with st.expander(f"📋 {desc['title']} 详细说明", expanded=True):
st.markdown(f"**特点：** {desc['features']}")
st.markdown(f"**适用场景：** {desc['use_case']}")
if desc['registration']:
st.markdown(f"**注册地址：** [{desc['registration']}]({desc['registration']})")
# 3. 根据选择的引擎渲染对应的配置界面
# st.subheader("⚙️ 引擎配置")
if selected_engine == "edge_tts":
render_edge_tts_settings(tr)
elif selected_engine == "azure_speech":
render_azure_speech_settings(tr)
elif selected_engine == "soulvoice":
render_soulvoice_engine_settings(tr)
# 4. 试听功能
render_voice_preview_new(tr, selected_engine)
def render_edge_tts_settings(tr):
"""渲染 Edge TTS 引擎设置"""
# 获取支持的语音列表
support_locales = ["zh-CN", "en-US"]
azure_voices = voice.get_all_azure_voices(filter_locals=support_locales)
all_voices = voice.get_all_azure_voices(filter_locals=support_locales)
# 添加 SoulVoice 语音选项
soulvoice_voices = get_soulvoice_voices()
# 合并所有语音选项
all_voices = azure_voices + soulvoice_voices
# 只保留标准版本的语音（Edge TTS专用，不包含V2）
edge_voices = [v for v in all_voices if "-V2" not in v]
# 创建友好的显示名称
friendly_names = {}
# Azure 语音的友好名称
for v in azure_voices:
for v in edge_voices:
friendly_names[v] = v.replace("Female", tr("Female")).replace("Male", tr("Male")).replace("Neural", "")
# SoulVoice 语音的友好名称
for v in soulvoice_voices:
friendly_names[v] = "SoulVoice (自定义音色)"
# 获取保存的语音设置
saved_voice_name = config.ui.get("voice_name", "")
saved_voice_name_index = 0
saved_voice_name = config.ui.get("edge_voice_name", "zh-CN-XiaoxiaoNeural-Female")
if saved_voice_name in friendly_names:
saved_voice_name_index = list(friendly_names.keys()).index(saved_voice_name)
else:
# 如果没有保存的设置，选择与UI语言匹配的第一个语音
for i, v in enumerate(all_voices):
if (v.lower().startswith(st.session_state["ui_language"].lower())
and "V2" not in v and not v.startswith("soulvoice:")):
saved_voice_name_index = i
# 确保保存的音色在可用列表中
if saved_voice_name not in friendly_names:
# 选择与UI语言匹配的第一个语音
for v in edge_voices:
if v.lower().startswith(st.session_state.get("ui_language", "zh-CN").lower()):
saved_voice_name = v
break
else:
# 如果没找到匹配的，使用第一个
saved_voice_name = edge_voices[0] if edge_voices else ""
# 语音选择下拉框
# 音色选择下拉框（Edge TTS音色相对较少，保留下拉框）
selected_friendly_name = st.selectbox(
tr("Speech Synthesis"),
"音色选择",
options=list(friendly_names.values()),
index=saved_voice_name_index,
index=list(friendly_names.keys()).index(saved_voice_name) if saved_voice_name in friendly_names else 0,
help="选择Edge TTS音色"
)
# 获取实际的语音名称
@ -80,34 +172,342 @@ def render_tts_settings(tr):
list(friendly_names.values()).index(selected_friendly_name)
]
# 如果选择的是 SoulVoice 自定义选项，使用配置的音色 URI
if voice_name == "soulvoice:custom":
custom_voice_uri = config.soulvoice.get("voice_uri", "")
if custom_voice_uri:
# 确保音色 URI 有正确的前缀
if not custom_voice_uri.startswith("soulvoice:") and not custom_voice_uri.startswith("speech:"):
voice_name = f"soulvoice:{custom_voice_uri}"
# 显示音色信息
with st.expander("💡 Edge TTS 音色说明", expanded=False):
st.write("**中文音色：**")
zh_voices = [v for v in edge_voices if v.startswith("zh-CN")]
for v in zh_voices:
gender = "女声" if "Female" in v else "男声"
name = v.replace("-Female", "").replace("-Male", "").replace("zh-CN-", "").replace("Neural", "")
st.write(f"• {name} ({gender})")
st.write("")
st.write("**英文音色：**")
en_voices = [v for v in edge_voices if v.startswith("en-US")][:5] # 只显示前5个
for v in en_voices:
gender = "女声" if "Female" in v else "男声"
name = v.replace("-Female", "").replace("-Male", "").replace("en-US-", "").replace("Neural", "")
st.write(f"• {name} ({gender})")
if len([v for v in edge_voices if v.startswith("en-US")]) > 5:
st.write("• ... 更多英文音色")
config.ui["edge_voice_name"] = voice_name
config.ui["voice_name"] = voice_name # 兼容性
# 音量调节
voice_volume = st.slider(
"音量调节",
min_value=0,
max_value=100,
value=int(config.ui.get("edge_volume", 80)),
step=1,
help="调节语音音量 (0-100)"
)
config.ui["edge_volume"] = voice_volume
st.session_state['voice_volume'] = voice_volume / 100.0
# 语速调节
voice_rate = st.slider(
"语速调节",
min_value=0.5,
max_value=2.0,
value=config.ui.get("edge_rate", 1.0),
step=0.1,
help="调节语音速度 (0.5-2.0倍速)"
)
config.ui["edge_rate"] = voice_rate
st.session_state['voice_rate'] = voice_rate
# 语调调节
voice_pitch = st.slider(
"语调调节",
min_value=-50,
max_value=50,
value=int(config.ui.get("edge_pitch", 0)),
step=5,
help="调节语音音调 (-50%到+50%)"
)
config.ui["edge_pitch"] = voice_pitch
# 转换为比例值
st.session_state['voice_pitch'] = 1.0 + (voice_pitch / 100.0)
def render_azure_speech_settings(tr):
"""渲染 Azure Speech Services 引擎设置"""
# 服务区域配置
azure_speech_region = st.text_input(
"服务区域",
value=config.azure.get("speech_region", ""),
placeholder="例如：eastus",
help="Azure Speech Services 服务区域，如：eastus, westus2, eastasia 等"
)
# API Key配置
azure_speech_key = st.text_input(
"API Key",
value=config.azure.get("speech_key", ""),
type="password",
help="Azure Speech Services API 密钥"
)
# 保存Azure配置
config.azure["speech_region"] = azure_speech_region
config.azure["speech_key"] = azure_speech_key
# 音色名称输入框
saved_voice_name = config.ui.get("azure_voice_name", "zh-CN-XiaoxiaoMultilingualNeural")
# 音色名称输入
voice_name = st.text_input(
"音色名称",
value=saved_voice_name,
help="输入Azure Speech Services音色名称，直接使用官方音色名称即可。例如：zh-CN-YunzeNeural",
placeholder="zh-CN-YunzeNeural"
)
# 显示常用音色示例
with st.expander("💡 常用音色参考", expanded=False):
st.write("**中文音色：**")
st.write("• zh-CN-XiaoxiaoMultilingualNeural (女声，多语言)")
st.write("• zh-CN-YunzeNeural (男声)")
st.write("• zh-CN-YunxiNeural (男声)")
st.write("• zh-CN-XiaochenNeural (女声)")
st.write("")
st.write("**英文音色：**")
st.write("• en-US-AndrewMultilingualNeural (男声，多语言)")
st.write("• en-US-AvaMultilingualNeural (女声，多语言)")
st.write("• en-US-BrianMultilingualNeural (男声，多语言)")
st.write("• en-US-EmmaMultilingualNeural (女声，多语言)")
st.write("")
st.info("💡 更多音色请参考 [Azure Speech Services 官方文档](https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/language-support)")
# 快速选择按钮
st.write("**快速选择：**")
cols = st.columns(3)
with cols[0]:
if st.button("中文女声", help="zh-CN-XiaoxiaoMultilingualNeural"):
voice_name = "zh-CN-XiaoxiaoMultilingualNeural"
st.rerun()
with cols[1]:
if st.button("中文男声", help="zh-CN-YunzeNeural"):
voice_name = "zh-CN-YunzeNeural"
st.rerun()
with cols[2]:
if st.button("英文女声", help="en-US-AvaMultilingualNeural"):
voice_name = "en-US-AvaMultilingualNeural"
st.rerun()
# 验证音色名称并显示状态
if voice_name.strip():
# 检查是否为有效的Azure音色格式
if is_valid_azure_voice_name(voice_name):
st.success(f"✅ 音色名称有效: {voice_name}")
else:
st.warning(f"⚠️ 音色名称格式可能不正确: {voice_name}")
st.info("💡 Azure音色名称通常格式为: [语言]-[地区]-[名称]Neural")
# 保存配置
config.ui["azure_voice_name"] = voice_name
config.ui["voice_name"] = voice_name # 兼容性
# 音量调节
voice_volume = st.slider(
"音量调节",
min_value=0,
max_value=100,
value=int(config.ui.get("azure_volume", 80)),
step=1,
help="调节语音音量 (0-100)"
)
config.ui["azure_volume"] = voice_volume
st.session_state['voice_volume'] = voice_volume / 100.0
# 语速调节
voice_rate = st.slider(
"语速调节",
min_value=0.5,
max_value=2.0,
value=config.ui.get("azure_rate", 1.0),
step=0.1,
help="调节语音速度 (0.5-2.0倍速)"
)
config.ui["azure_rate"] = voice_rate
st.session_state['voice_rate'] = voice_rate
# 语调调节
voice_pitch = st.slider(
"语调调节",
min_value=-50,
max_value=50,
value=int(config.ui.get("azure_pitch", 0)),
step=5,
help="调节语音音调 (-50%到+50%)"
)
config.ui["azure_pitch"] = voice_pitch
# 转换为比例值
st.session_state['voice_pitch'] = 1.0 + (voice_pitch / 100.0)
# 显示配置状态
if azure_speech_region and azure_speech_key:
st.success("✅ Azure Speech Services 配置已设置")
elif not azure_speech_region:
st.warning("⚠️ 请配置服务区域")
elif not azure_speech_key:
st.warning("⚠️ 请配置 API Key")
def render_soulvoice_engine_settings(tr):
"""渲染 SoulVoice 引擎设置"""
# API Key 输入
api_key = st.text_input(
"API Key",
value=config.soulvoice.get("api_key", ""),
type="password",
help="请输入您的 SoulVoice API 密钥"
)
# 音色 URI 输入
voice_uri = st.text_input(
"音色URI",
value=config.soulvoice.get("voice_uri", "speech:mcg3fdnx:clzkyf4vy00e5qr6hywum4u84:bzznlkuhcjzpbosexitr"),
help="请输入 SoulVoice 音色标识符",
placeholder="speech:mcg3fdnx:clzkyf4vy00e5qr6hywum4u84:bzznlkuhcjzpbosexitr"
)
# 模型名称选择
model_options = [
"FunAudioLLM/CosyVoice2-0.5B",
"FunAudioLLM/CosyVoice-300M",
"FunAudioLLM/CosyVoice-300M-SFT",
"FunAudioLLM/CosyVoice-300M-Instruct"
]
saved_model = config.soulvoice.get("model", "FunAudioLLM/CosyVoice2-0.5B")
if saved_model not in model_options:
model_options.append(saved_model)
model = st.selectbox(
"模型名称",
options=model_options,
index=model_options.index(saved_model),
help="选择使用的 TTS 模型"
)
# 高级设置
with st.expander("高级设置", expanded=False):
api_url = st.text_input(
"API 地址",
value=config.soulvoice.get("api_url", "https://tts.scsmtech.cn/tts"),
help="SoulVoice API 接口地址"
)
# 保存配置
config.soulvoice["api_key"] = api_key
config.soulvoice["voice_uri"] = voice_uri
config.soulvoice["model"] = model
config.soulvoice["api_url"] = api_url
# 设置兼容性配置
if voice_uri:
# 确保音色 URI 有正确的前缀
if not voice_uri.startswith("soulvoice:") and not voice_uri.startswith("speech:"):
voice_name = f"soulvoice:{voice_uri}"
else:
voice_name = voice_uri if voice_uri.startswith("soulvoice:") else f"soulvoice:{voice_uri}"
config.ui["voice_name"] = voice_name
# 显示配置状态
if api_key and voice_uri:
st.success("✅ SoulVoice 配置已设置")
elif not api_key:
st.warning("⚠️ 请配置 SoulVoice API Key")
elif not voice_uri:
st.warning("⚠️ 请配置音色 URI")
def render_voice_preview_new(tr, selected_engine):
"""渲染新的语音试听功能"""
if st.button("🎵 试听语音合成", use_container_width=True):
play_content = "感谢关注 NarratoAI，有任何问题或建议，可以关注微信公众号，求助或讨论"
# 根据选择的引擎获取对应的语音配置
voice_name = ""
voice_rate = 1.0
voice_pitch = 1.0
if selected_engine == "edge_tts":
voice_name = config.ui.get("edge_voice_name", "zh-CN-XiaoyiNeural-Female")
voice_rate = config.ui.get("edge_rate", 1.0)
voice_pitch = 1.0 + (config.ui.get("edge_pitch", 0) / 100.0)
elif selected_engine == "azure_speech":
voice_name = config.ui.get("azure_voice_name", "zh-CN-XiaoxiaoMultilingualNeural")
voice_rate = config.ui.get("azure_rate", 1.0)
voice_pitch = 1.0 + (config.ui.get("azure_pitch", 0) / 100.0)
elif selected_engine == "soulvoice":
voice_uri = config.soulvoice.get("voice_uri", "")
if voice_uri:
if not voice_uri.startswith("soulvoice:") and not voice_uri.startswith("speech:"):
voice_name = f"soulvoice:{voice_uri}"
else:
voice_name = voice_uri if voice_uri.startswith("soulvoice:") else f"soulvoice:{voice_uri}"
voice_rate = 1.0 # SoulVoice 使用默认语速
voice_pitch = 1.0 # SoulVoice 不支持音调调节
if not voice_name:
st.error("请先配置语音设置")
return
with st.spinner("正在合成语音..."):
temp_dir = utils.storage_dir("temp", create=True)
audio_file = os.path.join(temp_dir, f"tmp-voice-{str(uuid4())}.mp3")
sub_maker = voice.tts(
text=play_content,
voice_name=voice_name,
voice_rate=voice_rate,
voice_pitch=voice_pitch,
voice_file=audio_file,
)
if sub_maker and os.path.exists(audio_file):
st.success("✅ 语音合成成功！")
# 播放音频
with open(audio_file, 'rb') as audio_file_obj:
audio_bytes = audio_file_obj.read()
st.audio(audio_bytes, format='audio/mp3')
# 清理临时文件
try:
os.remove(audio_file)
except:
pass
else:
voice_name = custom_voice_uri if custom_voice_uri.startswith("soulvoice:") else f"soulvoice:{custom_voice_uri}"
st.error("❌ 语音合成失败，请检查配置")
# 保存设置
config.ui["voice_name"] = voice_name
# 根据语音类型渲染不同的设置
if voice.is_soulvoice_voice(voice_name):
render_soulvoice_settings(tr)
elif voice.is_azure_v2_voice(voice_name):
render_azure_v2_settings(tr)
def render_azure_v2_settings(tr):
"""渲染Azure V2语音设置（保留兼容性）"""
saved_azure_speech_region = config.azure.get("speech_region", "")
saved_azure_speech_key = config.azure.get("speech_key", "")
# 语音参数设置
render_voice_parameters(tr, voice_name)
azure_speech_region = st.text_input(
tr("Speech Region"),
value=saved_azure_speech_region
)
azure_speech_key = st.text_input(
tr("Speech Key"),
value=saved_azure_speech_key,
type="password"
)
# 试听按钮
render_voice_preview(tr, voice_name)
config.azure["speech_region"] = azure_speech_region
config.azure["speech_key"] = azure_speech_key
def render_soulvoice_settings(tr):
"""渲染 SoulVoice 语音设置"""
"""渲染 SoulVoice 语音设置（保留兼容性）"""
saved_api_key = config.soulvoice.get("api_key", "")
saved_api_url = config.soulvoice.get("api_url", "https://tts.scsmtech.cn/tts")
saved_model = config.soulvoice.get("model", "FunAudioLLM/CosyVoice2-0.5B")
@ -158,27 +558,8 @@ def render_soulvoice_settings(tr):
st.warning("⚠️ 请配置音色 URI")
def render_azure_v2_settings(tr):
"""渲染Azure V2语音设置"""
saved_azure_speech_region = config.azure.get("speech_region", "")
saved_azure_speech_key = config.azure.get("speech_key", "")
azure_speech_region = st.text_input(
tr("Speech Region"),
value=saved_azure_speech_region
)
azure_speech_key = st.text_input(
tr("Speech Key"),
value=saved_azure_speech_key,
type="password"
)
config.azure["speech_region"] = azure_speech_region
config.azure["speech_key"] = azure_speech_key
def render_voice_parameters(tr, voice_name):
"""渲染语音参数设置"""
"""渲染语音参数设置（保留兼容性）"""
# 音量 - 使用统一的默认值
voice_volume = st.slider(
tr("Speech Volume"),
@ -260,9 +641,12 @@ def render_voice_preview(tr, voice_name):
)
if sub_maker and os.path.exists(audio_file):
st.success(tr("Voice synthesis successful"))
st.audio(audio_file, format="audio/mp3")
if os.path.exists(audio_file):
os.remove(audio_file)
else:
st.error(tr("Voice synthesis failed"))
def render_bgm_settings(tr):