import librosa
import librosa.display
import matplotlib.pyplot as plt

# 讀取音頻文件
audio_file = 'example_audio.wav'
y, sr = librosa.load(audio_file)

# 繪制波形圖
plt.figure(figsize=(14, 5))
librosa.display.waveplot(y, sr=sr)
plt.title('Waveform')
plt.xlabel('Time (s)')
plt.ylabel('Amplitude')
plt.show()

# 繪制頻譜圖
D = librosa.amplitude_to_db(librosa.stft(y), ref=np.max)
plt.figure(figsize=(14, 5))
librosa.display.specshow(D, sr=sr, x_axis='time', y_axis='log')
plt.colorbar(format='%+2.0f dB')
plt.title('Log-frequency power spectrogram')
plt.show()

import numpy as np
import scipy.io.wavfile as wavfile

# 生成正弦波信號(hào)
duration = 5  # 時(shí)長(zhǎng)為5秒
sampling_freq = 44100  # 采樣頻率
tone_freq = 440  # 音調(diào)頻率為440Hz（A4音）

t = np.linspace(0, duration, int(sampling_freq * duration), endpoint=False)
audio_data = 0.5 * np.sin(2 * np.pi * tone_freq * t)

# 將生成的音頻數(shù)據(jù)保存為WAV文件
wavfile.write('generated_audio.wav', sampling_freq, audio_data.astype(np.float32))

import pyaudio
import numpy as np
import matplotlib.pyplot as plt

# 參數(shù)設(shè)置
CHUNK = 1024  # 每次讀取的音頻幀數(shù)
FORMAT = pyaudio.paInt16  # 音頻格式為16位整數(shù)
CHANNELS = 1  # 單聲道
RATE = 44100  # 采樣頻率

# 創(chuàng)建PyAudio對(duì)象
p = pyaudio.PyAudio()

# 打開音頻流
stream = p.open(format=FORMAT,
                channels=CHANNELS,
                rate=RATE,
                input=True,
                frames_per_buffer=CHUNK)

# 實(shí)時(shí)繪制頻譜圖
plt.ion()  # 打開交互模式
fig, ax = plt.subplots()
x = np.arange(0, 2 * CHUNK, 2)
line, = ax.plot(x, np.random.rand(CHUNK))

while True:
    # 讀取音頻數(shù)據(jù)
    data = stream.read(CHUNK)
    data_np = np.frombuffer(data, dtype=np.int16)
    
    # 計(jì)算頻譜
    spectrum = np.abs(np.fft.fft(data_np))
    spectrum = spectrum[:CHUNK]
    
    # 更新頻譜圖
    line.set_ydata(spectrum)
    ax.set_xlim(0, CHUNK)
    ax.set_ylim(0, 10000)  # 根據(jù)實(shí)際音頻特性調(diào)整縱坐標(biāo)范圍
    fig.canvas.draw()
    fig.canvas.flush_events()

# 關(guān)閉音頻流和PyAudio對(duì)象
stream.stop_stream()
stream.close()
p.terminate()

import numpy as np
from scipy.io import wavfile
from scipy import signal
import matplotlib.pyplot as plt

# 讀取音頻文件
sampling_freq, audio_data = wavfile.read('example_audio.wav')

# 提取單聲道音頻數(shù)據(jù)
audio_data_mono = audio_data[:, 0]

# 設(shè)計(jì)低通濾波器
cutoff_freq = 4000  # 截止頻率為4000Hz
nyquist_freq = 0.5 * sampling_freq
order = 4  # 濾波器階數(shù)
b, a = signal.butter(order, cutoff_freq/nyquist_freq, btype='low')

# 應(yīng)用濾波器
filtered_audio = signal.filtfilt(b, a, audio_data_mono)

# 繪制原始音頻波形和處理后的波形對(duì)比
plt.figure(figsize=(12, 6))

plt.subplot(2, 1, 1)
plt.plot(np.arange(len(audio_data_mono)) / sampling_freq, audio_data_mono, label='Original Audio')
plt.title('Original Audio Waveform')
plt.xlabel('Time (s)')
plt.ylabel('Amplitude')
plt.legend()

plt.subplot(2, 1, 2)
plt.plot(np.arange(len(filtered_audio)) / sampling_freq, filtered_audio, label='Filtered Audio', color='orange')
plt.title('Filtered Audio Waveform (Low-pass Filter)')
plt.xlabel('Time (s)')
plt.ylabel('Amplitude')
plt.legend()

plt.tight_layout()
plt.show()

from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
import librosa
import os

# 準(zhǔn)備數(shù)據(jù)集（假設(shè)每個(gè)文件夾中包含不同類別的音頻文件）
data_dir = 'audio_dataset'
classes = os.listdir(data_dir)
X = []
y = []

for cls in classes:
    cls_dir = os.path.join(data_dir, cls)
    for file in os.listdir(cls_dir):
        file_path = os.path.join(cls_dir, file)
        y.append(cls)
        audio_data, _ = librosa.load(file_path, sr=None)
        mfccs = librosa.feature.mfcc(audio_data, sr=44100, n_mfcc=13)
        X.append(np.mean(mfccs.T, axis=0))

# 將數(shù)據(jù)集分為訓(xùn)練集和測(cè)試集
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# 訓(xùn)練隨機(jī)森林分類器
clf = RandomForestClassifier(n_estimators=100, random_state=42)
clf.fit(X_train, y_train)

# 預(yù)測(cè)并評(píng)估模型
y_pred = clf.predict(X_test)
accuracy = accuracy_score(y_test, y_pred)
print(f'Accuracy: {accuracy:.2f}')