一直关注微软的神经网络自然语音库的开发应用，可惜不是科班出身，一直在用VB.NET,遗憾的是VB.NET 没有提供相关的库，或许有但没找到相关的库（.dll),。网友张老师一直关注微软的神经网络自然语音库，因为相关的语音库较一直用的IVONA语音库的音质要好得多， 并督促俺看能不能也开发一个微软的神经网络自然语音库应用。寻寻觅觅，总于找到了Python edge-tts, 并把相关的开发文本下载下来，其中有个模板：

#!/usr/bin/env python3

"""
Streaming TTS example with subtitles.

This example is similar to the example basic_audio_streaming.py, but it shows
WordBoundary events to create subtitles using SubMaker.
"""

import asyncio

import edge_tts

TEXT = "Hello World!"
VOICE = "en-GB-SoniaNeural"
OUTPUT_FILE = "test.mp3"
WEBVTT_FILE = "test.vtt"

async def _main() -> None:
    communicate = edge_tts.Communicate(TEXT, VOICE)
    submaker = edge_tts.SubMaker()
    with open(OUTPUT_FILE, "wb") as file:
        async for chunk in communicate.stream():
            if chunk["type"] == "audio":
                file.write(chunk["data"])
            elif chunk["type"] == "WordBoundary":
                submaker.create_sub((chunk["offset"], chunk["duration"]), chunk["text"])

    with open(WEBVTT_FILE, "w", encoding="utf-8") as file:
        file.write(submaker.generate_subs())

if __name__ == "__main__":
    loop = asyncio.get_event_loop()
    try:
        loop.run_until_complete(_main())
    finally:
        loop.close()

看到这模板，总算有了设计的方案，只要找到方案替换

TEXT = "Hello World!"
VOICE = "en-GB-SoniaNeural"

这两个，不就是容易了吗？

于是开始改造这模板如下：

import edge_tts
import asyncio
import os
TEXT = ""
with open('C:\\Edgetts\\t2v.txt', 'rb') as f1:
    data = f1.read()
    TEXT = data.decode('utf-8')
    f1.close()
    print(TEXT)
    
VOICE = "" 
with open('C:\\Edgetts\\voice.txt', 'rb') as f2:
    data = f2.read()
    VOICE = data.decode('utf-8')
    f2.close()
    print(VOICE)
    
OUTPUT_FILE ="C:\\Edgetts\\t2v.mp3"  
WEBVTT_FILE ="C:\\Edgetts\\t2v.vtt" 

async def _main(VOICE) -> None:  
    communicate = edge_tts.Communicate(TEXT, VOICE)  
    submaker = edge_tts.SubMaker()  
    with open(OUTPUT_FILE, "wb") as file:  
        async for chunk in communicate.stream():  
            if chunk["type"] == "audio":  
                file.write(chunk["data"])  
            elif chunk["type"] == "WordBoundary":  
                submaker.create_sub((chunk["offset"], chunk["duration"]), chunk["text"])  
  
    with open(WEBVTT_FILE, "w", encoding="utf-8") as file:  
        file.write(submaker.generate_subs())  
  
  
if __name__ == "__main__":  
    asyncio.run(_main(VOICE))

于是在VB.NET 进行相关的文本文件的读写设置：

TextBox1.Text = V1（这个是选择的语音库的名字）
        Dim noterec As StreamWriter
        Dim noteread As FileStream
        noteread = New FileStream("C:\\Edgetts\\voice.txt", FileMode.Create)
        noterec = New StreamWriter(noteread)
        noterec.Write(V1)
        noterec.Close()
        noteread.Close()

解决了语音库的选择，接下来解决输入文本：

T2v = RichTextBox1.Text
            Dim noterec1 As StreamWriter
            Dim noteread1 As FileStream
            noteread1 = New FileStream("C:\\Edgetts\\t2v.txt", FileMode.Create)
            noterec1 = New StreamWriter(noteread1)
            noterec1.Write(T2v)
            noterec1.Close()
            noteread1.Close()
   最后一步是解决调用Python打包生成的EXE 来读取相关的语音库及文本，转成相应的MP3

            Shell("edgefiletts.exe")

于是，一个简单的VB.NET+Python edge-tts 应用就诞生了。