Python Speech-to-Text

Making speech-to-text in python is easier than other languages, just call pyttsx3 (module). But the voice recognition function of WIN11 system must be installed first

Specific viewing method: open the control panel, search for "speech recognition", if the following icon appears

It means that the computer has a speech recognition module installed, otherwise it will be useless to toss. In particular, it is necessary to test whether the "text-to-speech conversion" function is useful, and it is useless if it does not work.

After the voice recognition module is installed on the computer

Then install the pypiwin32 module

Finally install the pyttsx3 module

import pyttsx3

engine = pyttsx3.init() # create object

engine.say('Hello, world!') # read the text

engine.say('Hello, world!') # read the text

engine.runAndWait() # Execute and wait for the voice to end

import pyttsx3

# create object

engine = pyttsx3.init()

# Get the current voice rate

rate = engine.getProperty('rate')

print(f'voice rate: {rate}')

# Set new speech rate

engine.setProperty('rate', 150)

# Get the current voice volume

volume = engine.getProperty('volume')

print(f'Voice volume: {volume}')

# Set the new voice volume, the minimum volume is 0 and the maximum volume is 1

engine.setProperty('volume', 1.0)

# Get the detailed information of the current voice and sound# I also found the example code here, which feels very contradictory, and the final voice is still a female voice

voices = engine.getProperty('voices')

print(f'Voice voice details: {voices}')

# Set the current voice as female, the current voice cannot read Chinese

engine.setProperty('voice', voices[1].id)

# Set the current voice voice as male, and the current voice can read Chinese

engine.setProperty('voice', voices[0].id)

# Get the current voice sound

voice = engine.getProperty('voice')

print(f'voice sound: {voice}')

# voice text

words = input('Please enter the words to say:')

# Speak the audio text

a = engine.say(words)

engine.runAndWait()

engine.stop()

# save the audio

engine.save_to_file("text.mp3")

engine.runAndWait()

Guess you like

Origin blog.csdn.net/stupid250/article/details/128690235