python 通过提取声音的特征做对比

python 通过提取声音的特征做对比

1. 环境准备

安装必要的Python库，如librosa（用于音频和音乐分析）、numpy（数学运算）、scipy（信号处理）、matplotlib（绘图）。

pip install librosa numpy scipy matplotlib

2.声音文件的读取

介绍如何使用librosa库读取音频文件。
示例代码：加载音频文件，获取音频信号和采样率。

import librosa

def read_audio_file(file_path):
    y, sr = librosa.load(file_path, sr=None)
    return y, sr

3. 声音特征提取

简要介绍几种常见的音频特征，如MFCC（梅尔频率倒谱系数）、CQT（恒定Q变换）、短时能量、过零率等。
示例代码：使用librosa提取MFCC特征

 recorded_audio, sr_recorded = read_audio_file(r'C:\Users\Desktop\cdc\pythonP56\audio\02 紧急故障.wav')
 local_audio, sr_local = read_audio_file(r'C:\Users\Desktop\cdc\pythonP56\audio\01 热失控.wav')
# 提取MFCC特征
 feature_recorded = np.mean(librosa.feature.mfcc(y=recorded_audio, sr=sr_recorded, n_mfcc=13), axis=1)
 feature_local = np.mean(librosa.feature.mfcc(y=local_audio, sr=sr_local, n_mfcc=13), axis=1)

4. 特征对比策略

介绍几种常见的特征对比方法，如欧氏距离、余弦相似度、动态时间规整（DTW）等。
示例代码：使用计算余弦相似度对比两段音频的MFCC特征。

 # 计算余弦相似度：
    similarity = np.dot(feature_recorded, feature_local) / (
                np.linalg.norm(feature_recorded) * np.linalg.norm(feature_local))
    print("音频1和音频2之间的相似度：", similarity)
    if similarity > 0.9:
        return True
    else:
        return False

5. 完整code

import pyaudio
import wave
import librosa
import numpy as np



# 录制麦克风音频
def record_audio(seconds):
    chunk = 1024
    format = pyaudio.paInt16
    channels = 1
    rate = 44100
    p = pyaudio.PyAudio()
    stream = p.open(format=format,
                    channels=channels,
                    rate=rate,
                    input=True,
                    frames_per_buffer=chunk)
    frames = []
    for i in range(0, int(rate / chunk * seconds)):
        data = stream.read(chunk)
        frames.append(data)
    stream.stop_stream()
    stream.close()
    p.terminate()
    wf = wave.open('recorded_audio.wav', 'wb')
    wf.setnchannels(channels)
    wf.setsampwidth(p.get_sample_size(format))
    wf.setframerate(rate)
    wf.writeframes(b''.join(frames))
    wf.close()


# 读取本地音频文件
def read_audio_file(file_path):
    y, sr = librosa.load(file_path, sr=None)
    return y, sr


# 对比两个音频文件的特征
def compare_audio_files():
    recorded_audio, sr_recorded = read_audio_file(r'C:\Users\Desktop\cdc\pythonP56\audio\02 紧急故障.wav')
    local_audio, sr_local = read_audio_file(r'C:\Users\Desktop\cdc\pythonP56\audio\01 热失控.wav')

    # 提取MFCC特征
    feature_recorded = np.mean(librosa.feature.mfcc(y=recorded_audio, sr=sr_recorded, n_mfcc=13), axis=1)
    feature_local = np.mean(librosa.feature.mfcc(y=local_audio, sr=sr_local, n_mfcc=13), axis=1)

    # 计算余弦相似度：
    similarity = np.dot(feature_recorded, feature_local) / (
                np.linalg.norm(feature_recorded) * np.linalg.norm(feature_local))
    print("音频1和音频2之间的相似度：", similarity)
    if similarity > 0.9:
        return True
    else:
        return False


# 调用函数录制麦克风音频
record_audio(5)

# 对比录制的麦克风音频和本地音频文件
result = compare_audio_files()
print(result)