python3实现语音转文字(语音识别)和文字转语音(语音合成)

Python3实现语音识别和语音合成

本文将分享如何使用Python3实现语音识别和语音合成的过程，主要使用的是Google Speech API和Google Text-to-Speech API。

安装依赖

在开始之前需要安装以下库：

pip install google-cloud-speech google-cloud-texttospeech pyaudio

同时需要安装Google的API，我们需要创建一个Google Cloud Platform帐户并为它启用Google Cloud Speech-to-Text API和Google Cloud Text-to-Speech API。获取授权文件后将其放入项目目录中。

实现语音识别

import io
import os

# 导入语音识别库
from google.cloud import speech
from google.cloud.speech import enums
from google.cloud.speech import types

# 启用授权
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "your_auth_file.json"

# 初始化语音识别客户端
client = speech.SpeechClient()

# 从音频文件获取语音内容并进行识别
def transcribe_file(speech_file):
    with io.open(speech_file, 'rb') as f:
        content = f.read()
    audio = types.RecognitionAudio(content=content)
    config = types.RecognitionConfig(
        encoding=enums.RecognitionConfig.AudioEncoding.LINEAR16,
        sample_rate_hertz=16000, # 采样率需要与音频文件的采样率匹配
        language_code='zh-CN') # 语言设置为中文
    response = client.recognize(config, audio)
    for result in response.results:
        return result.alternatives[0].transcript # 返回最佳识别结果

# 按下回车后录制音频并识别
def recognize_speech():
    input("Press Enter to start recording...")
    CHUNK = 1024
    FORMAT = pyaudio.paInt16
    CHANNELS = 1
    RATE = 16000
    RECORD_SECONDS = 5
    WAVE_OUTPUT_FILENAME = "test.wav"

    p = pyaudio.PyAudio()

    stream = p.open(format=FORMAT,
                    channels=CHANNELS,
                    rate=RATE,
                    input=True,
                    frames_per_buffer=CHUNK)

    frames = []

    print("Recording...")

    for i in range(0, int(RATE / CHUNK * RECORD_SECONDS)):
        data = stream.read(CHUNK)
        frames.append(data)

    print("Finished recording.")

    stream.stop_stream()
    stream.close()
    p.terminate()

    wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
    wf.setnchannels(CHANNELS)
    wf.setsampwidth(p.get_sample_size(FORMAT))
    wf.setframerate(RATE)
    wf.writeframes(b''.join(frames))
    wf.close()

    print("Transcribing...")
    text = transcribe_file(WAVE_OUTPUT_FILENAME)
    print("Transcription:", text)

# 调用录音函数
recognize_speech()

实现了一个简单的语音识别程序，同时可以录制音频输入（通过按下回车键开始录制，录制5秒钟后自动停止），并输出识别的文字结果。

实现语音合成

from google.cloud import texttospeech

# 启用授权
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "your_auth_file.json"

# 初始化语音合成客户端
client = texttospeech.TextToSpeechClient()

# 保存TTS合成后的语音
def save_audio(synthesis_input, voice, audio_config, output_file):
    response = client.synthesize_speech(synthesis_input, voice, audio_config)

    with open(output_file, 'wb') as out:
        out.write(response.audio_content)
        print('Audio content saved to file {output_file}')

# 合成指定文字并保存
def synthesize_text(text, output_file):
    input_text = texttospeech.SynthesisInput(text=text)
    voice = texttospeech.VoiceSelectionParams(
        language_code='zh-CN', # 语言设置为中文
        name='zh-CN-Wavenet-D') # 选择语音类型为Wavenet
    audio_config = texttospeech.AudioConfig(
        audio_encoding=texttospeech.AudioEncoding.MP3)
    save_audio(input_text, voice, audio_config, output_file)

# 调用语音合成函数
synthesize_text("你好，很高兴认识你", "output.mp3")

这是一个简单的语音合成程序，将输入的文字转为语音并输出为MP3文件。

以上两个示例程序都是使用Google的API，但是其他厂商如阿里云、腾讯云、百度云等也提供了类似的API，开发者可以根据自己的需要进行选择。