树莓派——语音唤醒

#!/usr/bin/env python3.10

import pvporcupine

import pyaudio

import struct

import wave

porcupine_key = "xxxx"

porcupine_model = '/file/model/hello-chat_en_raspberry-pi_v3_0_0.ppn'

def keyword_wake_up():

    porcupine = pvporcupine.create(access_key=porcupine_key, keyword_paths=[porcupine_model])

    kws_audio = pyaudio.PyAudio()

    audio_stream = kws_audio.open(

        rate=porcupine.sample_rate,

        channels=1,

        format=pyaudio.paInt16,

        input=True,

        frames_per_buffer=porcupine.frame_length,

        input_device_index=None,

    )

    print("等待唤醒中,唤醒词:hello chat...")

    recording = []

    while True:

        pcm = audio_stream.read(porcupine.frame_length)

        _pcm = struct.unpack_from("h" * porcupine.frame_length, pcm)

        keyword_index = porcupine.process(_pcm)

        recording.extend(_pcm)

        if keyword_index >= 0:

            print("唤醒了捏！")

            save_audio(recording)

            return

    audio_stream.stop_stream()

    audio_stream.close()

    porcupine.delete()

    kws_audio.terminate()

def save_audio(data, filename="wake_word.wav"):

    wf = wave.open(filename, 'wb')

    wf.setnchannels(1)

    wf.setsampwidth(2)

    wf.setframerate(16000)  # Adjust the sample rate as needed

    wf.writeframes(b''.join(struct.pack('h', sample) for sample in data))

    wf.close()

keyword_wake_up()

# -*- coding: utf-8 -*-

from aip import AipSpeech

import pvporcupine

import pyaudio

import struct

import wave

import json

import base64

import io

import pvcobra

import time

import os

porcupine_key = "cinmq/v7vHEzd3vrbTD9I24KiGgxbmUBXjxCcgG8kGnx8l48h57L6g=="

porcupine_model = '../file/model/hello-chat_en_raspberry-pi_v3_0_0.ppn'

def get_file_content(filePath):  # filePath  待读取文件名



    with open(filePath, 'rb') as fp:



        return fp.read()

def sound_record():

    porcupine = pvporcupine.create(access_key=porcupine_key, keyword_paths=[porcupine_model])

    cobra = pvcobra.create(access_key=porcupine_key)

    kws_audio = pyaudio.PyAudio()

    audio_stream = kws_audio.open(

        rate=porcupine.sample_rate,

        channels=1,

        format=pyaudio.paInt16,

        input=True,

        frames_per_buffer=porcupine.frame_length,

        input_device_index=None,

    )

    is_voiced = 1

    print("开始录音...")

    recording = []

    silence_count = 0



    while is_voiced:

        pcm = audio_stream.read(porcupine.frame_length)

        _pcm = struct.unpack_from("h" * porcupine.frame_length, pcm)

        is_voiced = cobra.process(_pcm)

        print(is_voiced,silence_count)

        silence_count = 0 if is_voiced > 0.5 else silence_count  + 1

        if silence_count <= 100:

            if silence_count < 70:

                recording.extend(_pcm)

        else:

            break

    filename = save_audio(recording)

    audio_stream.stop_stream()

    audio_stream.close()

    porcupine.delete()

    kws_audio.terminate()

    # 保存录音结果为WAV文件

    print(f"录音已保存:recorded_audio.wav")

    return filename

 

def save_audio(data, filename="../file/tmp/recorded_audio.wav"):

    wf = wave.open(filename, 'wb')

    wf.setnchannels(1)

    wf.setsampwidth(2)

    wf.setframerate(16000)  # Adjust the sample rate as needed

    wf.writeframes(b''.join(struct.pack('h', sample) for sample in data))

    wf.close()

    return filename

    

'''凭据从环境变量读取，不要写入源码或公开笔记。'''

APP_ID = os.environ['BAIDU_SPEECH_APP_ID']

API_KEY = os.environ['BAIDU_SPEECH_API_KEY']

SECRET_KEY = os.environ['BAIDU_SPEECH_SECRET_KEY']

textPath = '../file/tmp/text.txt'

# 新建一个AipSpeech

client = AipSpeech(APP_ID, API_KEY, SECRET_KEY)

def baidu_stt(filename):

    result = client.asr(get_file_content(filename), 'wav', 16000, {'dev_pid': 1536})

    if result['err_msg'] == 'success.':

        print("stt successful")

        word = result['result'][0].encode('utf-8')  # utf-8编码

        wordStr = word.decode('utf-8')  # Decode the bytes to a string

        if wordStr != '':

            if wordStr[-3:] == '，':

                print(wordStr[:-3])

                with open(textPath, 'w', encoding='utf-8') as f:

                    f.write(wordStr[:-3])

            else:

                print(wordStr)

                with open(textPath, 'w', encoding='utf-8') as f:

                    f.write(wordStr)

        else:

            print("音频文件不存在或格式错误")

    else:

        print("错误")

    return wordStr

def listen(model: str = "baidu"):

    filepath = sound_record()

    if model == "baidu":

        user_words = baidu_stt(filepath)

    else:

        user_words = "Unsupported speech recognition model"

    if user_words == "":

        print("你什么都没说")

    else:

        print("你说了: ", user_words)

    return user_words

listen()