我需要提取wav文件的幅度,并希望将其作为一个简单的命令行应用程序。有什么简单的方法可以做到这一点?跨平台会很棒。至少需要在Windows上工作。
你可以download an audio file for testing from soundcloud(或从https://dl.dropboxusercontent.com/u/313647/hosted/one_pos_to_neg_crossing.wav直到soundcloud处理它。)
一些潜在用途的图书馆:
答案 0 :(得分:2)
这是我使用Accord.NET提出的解决方案:
public IEnumerable<float> AmplitudesFromFile(string fileName)
{
Accord.Audio.Formats.WaveDecoder sourceDecoder = new Accord.Audio.Formats.WaveDecoder(fileName);
Signal sourceSignal = sourceDecoder.Decode();
for (int i = 0; i < sourceSignal.Samples; i++) yield return sourceSignal.GetSample(1, i);
}
答案 1 :(得分:0)
这是我想出如何用NAudio做的第一种方式。
>SoundToAmplitudes.exe w:\materials\audio\one_pos_to_neg_crossing.wav
0.04284668
-0.005615234
-0.1177368
using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using NAudio.Wave;
namespace SoundToAmplitudes {
class Program {
private static int Main(string[] args) {
#if DEBUG
Console.WriteLine(String.Join(", ", args));
args = new[] { @"W:\materials\audio\one_pos_to_neg_crossing.wav" };
#endif
return Cli(args);
}
static int Cli(string[] args) {
string fileName = args[0];
var soundFile = new FileInfo(fileName);
foreach (float s in AmplitudesFromFile(soundFile)) {
Console.WriteLine(s);
}
//Console.WriteLine();
#if DEBUG
Console.Read();
#endif
return 0;
}
public static IEnumerable<float> AmplitudesFromFile(FileInfo soundFile) {
var reader = new AudioFileReader(soundFile.FullName);
int count = 4096; // arbitrary
float[] buffer = new float[count];
int offset = 0;
int numRead = 0;
while ((numRead = reader.Read(buffer, offset, count)) > 0) {
foreach (float amp in buffer.Take(numRead)) {
yield return amp;
}
}
}
}
}
答案 2 :(得分:0)
这是在python中使用scipy)的一种方法,虽然我假设wav文件是16位。
W:\>python amplitudes.py w:\materials\audio\one_pos_to_neg_crossing.wav
0.0428479873043
-0.00561540574358
-0.117740409558
"""Usage:
amplitudes WAV_FILE
Returns the (linear) amplitude of the signal inside the wav file, sample by sample.
"""
from __future__ import division
import docopt
import scipy.io.wavfile
MAX_WAV16_AMP = 32767 # = 2**15-1 # because wav16 is signed (wav8 isn't)
def main():
args = docopt.docopt(__doc__)
one_pos_to_neg_crossing_path = r"W:\materials\audio\one_pos_to_neg_crossing.wav"
if False: args['WAV_FILE'] = one_pos_to_neg_crossing_path
rate, amp_arr = scipy.io.wavfile.read(args['WAV_FILE'])
for amp in (amp_arr / MAX_WAV16_AMP):
print amp
if __name__ == '__main__':
main()