Python使用azure语音合成接口

## 安装
```
pip install azure-cognitiveservices-speech
```

## 使用
### 直接播放音频
```python
from azure.cognitiveservices.speech import SpeechConfig, SpeechSynthesizer, AudioDataStream
from config import settings

# 配置
speech_key = settings.speech_key
service_region = settings.service_region
reader = settings.reader

# 配置
speech_config = SpeechConfig(subscription=speech_key, region=service_region)
speech_config.speech_synthesis_voice_name = reader
# 设置为True表示直接播放
audio_config = speechsdk.audio.AudioOutputConfig(use_default_speaker=True)
speech_synthesizer = SpeechSynthesizer(speech_config=speech_config, audio_config=audio_config)

# 播放
def getVoice(text):
    speech_synthesizer.speak_text_async(text).get()
```

### 保存音频到指定文件
第一种方法是直接设置保存路径。
```python
from azure.cognitiveservices.speech import SpeechConfig, SpeechSynthesizer, AudioDataStream
from config import settings

# 配置
speech_key = settings.speech_key
service_region = settings.service_region
reader = settings.reader

# 配置
speech_config = SpeechConfig(subscription=speech_key, region=service_region)
speech_config.speech_synthesis_voice_name = reader
# 设置为False表示不播放，如果设为True则即保存又播放
audio_config = speechsdk.audio.AudioOutputConfig(use_default_speaker=False, filename="./hello.wav")
speech_synthesizer = SpeechSynthesizer(speech_config=speech_config, audio_config=audio_config)

# 保存
def getVoice(text):
    speech_synthesizer.speak_text_async(text).get()
```

第二种方法是先保存到数据流中，再保存到指定文件中，要将`audio_config`的属性值设为`None`，这样扬声器就不会输出音频。如果不设置为`None`，则会默认输出音频。
```python
from azure.cognitiveservices.speech import SpeechConfig, SpeechSynthesizer, AudioDataStream
from config import settings

# 配置
speech_key = settings.speech_key
service_region = settings.service_region
reader = settings.reader

# 配置
speech_config = SpeechConfig(subscription=speech_key, region=service_region)
speech_config.speech_synthesis_voice_name = reader
speech_synthesizer = SpeechSynthesizer(speech_config=speech_config, audio_config=None)

# 保存
def getVoice(text, filename):
    res = speech_synthesizer.speak_text_async(text).get()
    stream = AudioDataStream(res)
    stream.save_to_wav_file(filename)
```

Python使用azure语音合成接口

酷酷番茄