Python音频文件读写教程详解

from scipy.io import wavfile
<p>sample_rate, audio_data = wavfile.read('audio.wav')
print("采样率:", sample_rate)
print("音频数据形状:", audio_data.shape)</p>

import numpy as np
from scipy.io import wavfile
<h1>生成一个简单的正弦波作为示例</h1><p>duration = 2  # 秒
frequency = 440  # Hz
t = np.linspace(0, duration, int(sample_rate <em> duration), False)
tone = np.sin(2 </em> np.pi <em> frequency </em> t) * 32767  # 转为 16 位整数范围
tone = tone.astype(np.int16)</p><p>wavfile.write('output.wav', sample_rate, tone)</p>

import soundfile as sf
<p>audio_data, sample_rate = sf.read('audio.flac')
print("采样率:", sample_rate)
print("通道数:", audio_data.shape[1] if len(audio_data.shape) > 1 else 1)</p>

sf.write('output.ogg', audio_data, sample_rate, format='OGG')

from pydub import AudioSegment
<p>audio = AudioSegment.from_mp3("audio.mp3")</p><h1>其他格式可用 from_wav, from_ogg 等，或统一用 from_file</h1><h1>audio = AudioSegment.from_file("audio.m4a", format="m4a")</h1><h1>获取参数</h1><p>sample_rate = audio.frame_rate
channels = audio.channels
duration_ms = len(audio)</p>

audio.export("output.wav", format="wav")

import librosa
<p>audio_data, sample_rate = librosa.load('audio.mp3', sr=22050)  # 默认重采样到 22050 Hz
print("采样率:", sample_rate)
print("时长（秒）:", len(audio_data) / sample_rate)</p>

import soundfile as sf
<p>sf.write('reconstructed.wav', audio_data, sample_rate)</p>