升级macOS Ventura后pyttsx3获取语音列表报KeyError: 'VoiceAge'求助

问题背景 #

原代码在macOS Monterey + Python 3.11环境运行正常，升级至macOS Ventura后，执行engine.getProperty('voices')时出现KeyError: 'VoiceAge'错误，精简代码可复现问题：

import pyttsx3
engine = pyttsx3.init('nsss')
voices = engine.getProperty('voices')

报错信息 #

File "./voice assistant/tts.py", line 18, in <module>
    voices = engine.getProperty('voices')
  File "./voice assistant/venv/lib/python3.10/site-packages/pyttsx3/engine.py", line 146, in getProperty
    return self.proxy.getProperty(name)
  File "./voice assistant/venv/lib/python3.10/site-packages/pyttsx3/driver.py", line 173, in getProperty
    return self._driver.getProperty(name)
  File "./voice assistant/venv/lib/python3.10/site-packages/pyttsx3/drivers/nsss.py", line 69, in getProperty
    return [self._toVoice(NSSpeechSynthesizer.attributesForVoice_(v))
  File "./voice assistant/venv/lib/python3.10/site-packages/pyttsx3/drivers/nsss.py", line 69, in <listcomp>
    return [self._toVoice(NSSpeechSynthesizer.attributesForVoice_(v))
  File "./voice assistant/venv/lib/python3.10/site-packages/pyttsx3/drivers/nsss.py", line 64, in _toVoice
    attr['VoiceAge'])
  File "./voice assistant/venv/lib/python3.10/site-packages/objc/_convenience_mapping.py", line 18, in __getitem__objectForKey_
    return container_unwrap(res, KeyError, key)
  File "./voice assistant/venv/lib/python3.10/site-packages/objc/_convenience.py", line 122, in container_unwrap
    raise exc_type(*exc_args)
KeyError: 'VoiceAge'

已尝试无效操作 #

• 重装pyttsx3
• 使用sudo运行代码
• 改用pyttsx3.init('dummy')初始化（无报错但无法发声）

解决方案 #

方案1：修改pyttsx3的nsss驱动文件 #

错误原因是macOS Ventura中部分系统语音的属性不再包含VoiceAge字段，而pyttsx3的nsss驱动直接通过键取值导致报错。修改驱动文件处理缺失字段：

1. 找到虚拟环境中nsss.py的路径：./voice assistant/venv/lib/python3.10/site-packages/pyttsx3/drivers/nsss.py
2. 打开文件，找到_toVoice函数（约60行处）
3. 将原代码中直接取值的部分改为用get方法设置默认值：
   原代码：

return Voice(
    id=attr['VoiceIdentifier'],
    name=attr['VoiceName'],
    languages=attr['VoiceLanguages'],
    gender=attr['VoiceGender'],
    age=attr['VoiceAge'])

修改为：

return Voice(
    id=attr['VoiceIdentifier'],
    name=attr['VoiceName'],
    languages=attr['VoiceLanguages'],
    gender=attr.get('VoiceGender', 'VoiceGenderNeutral'),
    age=attr.get('VoiceAge', 0))

4. 保存文件后重新运行代码即可正常获取语音列表。

方案2：改用原生NSSpeechSynthesizer接口（推荐） #

绕过pyttsx3的兼容性问题，直接通过pyobjc调用macOS原生语音合成接口：

1. 安装依赖：

pip install pyobjc

2. 替代代码示例（支持选择指定语音）：

from Foundation import NSSpeechSynthesizer

def va_speak(phrase, voice_id=None):
    if voice_id:
        synth = NSSpeechSynthesizer.alloc().initWithVoice_(voice_id)
    else:
        synth = NSSpeechSynthesizer.alloc().init()
    # 设置语速（范围100-300，默认175）
    synth.setRate_(150)
    # 设置音量（范围0-1，默认1）
    synth.setVolume_(1.0)
    synth.startSpeakingString_(phrase)
    while synth.isSpeaking():
        pass

# 获取所有可用语音并筛选en_US男性语音
voices = NSSpeechSynthesizer.availableVoices()
target_voice = None
for voice in voices:
    attrs = NSSpeechSynthesizer.attributesForVoice_(voice)
    if attrs.get('VoiceLanguages') == ['en_US'] and attrs.get('VoiceGender') == 'VoiceGenderMale':
        target_voice = voice
        break

# 使用指定语音发声
va_speak("Hello there", target_voice)

内容的提问来源于stack exchange，提问作者Simple_Mac User