TTS 文字转语音 语音转文字

pyttsx3

pip install pyttsx3


import pyttsx3


def say_text(engine, words, voice):
    # 设置音色
    engine.setProperty('voice', voice)
    engine.say(words)
    engine.runAndWait()
    engine.stop()


def main():
    # 创建对象
    engine = pyttsx3.init()
    # 设置语音速率
    engine.setProperty('rate', 100)
    # 设置语音音量，音量最小为 0，最大为 1
    engine.setProperty('volume', 1.0)
    # words = input('请输入要转换的文本：')
    words = '这是一个小练习'
    say_text(engine, words, 'zh')


if __name__ == '__main__':
    main()

tts

pip install tts

from gtts import gTTS
import os

language = 'zh' # 或 zh-tw
#  slow=False时，语速为Normal
output = gTTS(text="这是中文", lang=language, slow=False)
output.save(filename)

# Play the converted file
os.system(f"start {filename}")

win32com

python -m pip install pyHook
python -m pip install pywin32

import win32com.client
speaker = win32com.client.Dispatch("SAPI.SpVoice")
speaker.Speak("你好，这是中文!")

ttskit

python -m pip install -U ttskit pyworld
python -m pip install ffmpeg

zhtts

python3 -m pip install -U ffmpeg pyworld ttskit  # 会安装tensorflow-cpu版本，但使用时会报CUDA的错
python3 -m pip install tensorflow-gpu

import zhtts
import sounddevice as sd

tts = zhtts.TTS()  # use fastspeech2 by default
text1 = "2020年，这是一个开源的端到端中文语音合成系统"

# 下面是自带的函数，借助Pycharm查看出来的。audio为numpy数组可直接传入播放器。
mel = tts.text2mel(text1)
print(mel.shape, type(mel))
audio = tts.mel2audio(mel)
print(audio, type(audio))
# 下面这里可以先看【6.sounddevice播放音频】
sd.play(audio, samplerate=24000) # samplerate=24000为通过其他包转换为.wav文件，再读取该文件获取的
sd.wait()

souddevice播放音频

python -m pip install soundfile sounddevice

import soundfile
import sounddevice

devs=sd.query_devices()  #返回系统所有的声音设备
print(devs) # 带><的是默认播放设备
# 1 代表设备号；麦克风阵列 (Synaptics Audio)代表设备名称；MME (2 in, 0 out)代表驱动为MME和2个输入通道
"""
   0 Microsoft Sound Mapper - Input, MME (2 in, 0 out)
>  1 麦克风阵列 (Synaptics Audio), MME (2 in, 0 out)
   2 Microsoft Sound Mapper - Output, MME (0 in, 2 out)
<  3 扬声器 (Synaptics Audio), MME (0 in, 2 out)
   4 主声音捕获驱动程序, Windows DirectSound (2 in, 0 out)
   5 麦克风阵列 (Synaptics Audio), Windows DirectSound (2 in, 0 out)
   6 主声音驱动程序, Windows DirectSound (0 in, 2 out)
   7 扬声器 (Synaptics Audio), Windows DirectSound (0 in, 2 out)
   8 扬声器 (Synaptics Audio), Windows WASAPI (0 in, 2 out)
   9 麦克风阵列 (Synaptics Audio), Windows WASAPI (2 in, 0 out)
  10 麦克风阵列 1 (Synaptics Audio capture), Windows WDM-KS (2 in, 0 out)
  11 麦克风阵列 2 (Synaptics Audio capture), Windows WDM-KS (4 in, 0 out)
  12 麦克风阵列 3 (Synaptics Audio capture), Windows WDM-KS (4 in, 0 out)
  13 Output 1 (Synaptics Audio output), Windows WDM-KS (0 in, 2 out)
  14 Output 2 (Synaptics Audio output), Windows WDM-KS (0 in, 8 out)
  15 Input (Synaptics Audio output), Windows WDM-KS (2 in, 0 out)
  16 耳机 (@System32\drivers\bthhfenum.sys,#2;%1 Hands-Free AG Audio%0
;(iGene-U2)), Windows WDM-KS (0 in, 1 out)
  17 耳机 (@System32\drivers\bthhfenum.sys,#2;%1 Hands-Free AG Audio%0
;(iGene-U2)), Windows WDM-KS (1 in, 0 out)
  18 耳机 (), Windows WDM-KS (0 in, 2 out)
"""
data, sampler = soundfile.read(path)
try:
    sd.default.device[1] = sd.default.device[1]	# 可以通过设备号指定播放设备
    sd.play(data, samplerate=sampler) # data为numpy数组，samplerate可能为采样率吧(不太清楚)
except Exception as e:
	print("播放失败")
    sd.wait()

Python语音转文字

import pyttsx3
import io
import sys
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')
 
engine = pyttsx3.init()
engine.setProperty('voice', 'zh')
engine.say('请输入您要播放的文件路径')
engine.runAndWait()
 
path_dir = input('请输入您要播放的文件路径后回车：')
 
 
with open(path_dir, 'r') as f:
    engine = pyttsx3.init()
    engine.setProperty('voice', 'zh')
    lines = f.readlines()
    print('要说的话：',lines)
    for line in lines:
        engine.say(line)
    engine.runAndWait()
 
 
"""Saving Voice to a file"""
engine = pyttsx3.init()
engine.setProperty('voice', 'zh')
# On linux make sure that 'espeak' and 'ffmpeg' are installed
engine.save_to_file('迪迦奥特曼（dijia.top）', 'test.mp3')
engine.runAndWait()
 
engine.stop()

speech 实现语音识别

import speech

while True:

    say = speech.input() # 接收语音

    speech.say("you said:"+say) #说话

    if say == "你好":

        speech.say("How are you?")

    elif say == "天气":

        speech.say("今天天气晴!")

SpeechRecognation

pip install SpeechRecognition