将文本保存为语音 Python

Question

I am trying save the results of text to speech to a file on windows. I successfully gotten it to speak (using speak.Speak ). However, no such luck with saving files.

The problem is AudioOutputStream isn't being found, despite itbeing listed in the Microsoft docs .

Version information: Windows 10, Python 3.6

Error

Traceback (most recent call last):
  File "...\dd.py", line 87, in <module>
    speak.AudioOutputStream = filestream
  File "...\win32com\client\dynamic.py", line 565, in __setattr__
    self._oleobj_.Invoke(entry.dispid, 0, invoke_type, 0, value)
pywintypes.com_error: (-2147352573, 'Member not found.', None, None)

Code

from win32com.client import Dispatch
import win32api

speak = Dispatch("SAPI.SpVoice")
filestream = Dispatch("SAPI.SpFileStream")
filestream.open("out.wav", 3, False) 
for k in speak.GetAudioOutputs():
    print(k.GetDescription())
speak.AudioOutputStream = filestream
speak.Speak("test")
filestream.close()

Answer 1

I managed to get it working. As there was little response to this post originally, I believe it would be very useful to leave this fix for future visitors.

Use pip install comtypes to install the library. It is a lot more cooperative and less janky than python's native comtype implementation.

import comtypes.client

speak = comtypes.client.CreateObject("SAPI.SpVoice")
filestream = comtypes.client.CreateObject("SAPI.spFileStream")
filestream.open("out.wav", 3, False)
speak.AudioOutputStream = filestream 
speak.Speak("test")
filestream.close()

Answer 2

I found this question but the solution provided by Neil using comtypes didn't work for me and gave me an error.

However, his original solution did work on my system:

• Windows 10 Pro, Version 19.0.19041 Build 19041
• Python 3.8.1 64-bit

I used this to create .wav files for a speaking clock

import win32com.client

phrases = [
    "Zero",
    "One",
    "Two",
    "Three",
    "Four",
    "Five",
    "Six",
    "Seven",
    "Eight",
    "Nine",
    "Ten",
    "Eleven",
    "Twelve",
    "Thirteen",
    "Fourteen",
    "Fifteen",
    "Sixteen",
    "Seventeen", 
    "Eighteen",
    "Nineteen",
    "Twenty",
    "Thirty",
    "Fourty",
    "Fifty",
    "Hundred",
]

speaker = win32com.client.Dispatch("SAPI.SpVoice")
filestream = win32com.client.Dispatch("SAPI.SpFileStream")

def speak(phrase):
    speaker.speak(phrase)
    
def save_voice(phrase):
    filestream.open(".".join([phrase, "wav"]), 3, False)
    speaker.AudioOutputStream = filestream
    speaker.speak(phrase)
    filestream.close()

for phrase in phrases:
    save_voice(phrase)