Last active
April 13, 2017 16:35
-
-
Save GregaMohorko/356a60d269e8f403af151b1504034b26 to your computer and use it in GitHub Desktop.
Audio signal analysis to get information about frequencies.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
using System; | |
using NAudio.Dsp; | |
using NAudio.Wave; | |
using NAudio.Wave.SampleProviders; | |
namespace GM.SP.Audio | |
{ | |
public static class SignalAnalysis | |
{ | |
public class Frequency | |
{ | |
public float Hz { get; set; } | |
public float Amplitude { get; set; } | |
public float Phase { get; set; } | |
} | |
public static Frequency[] AnalyzeAudioFile(string audioFile) | |
{ | |
WaveStream audioWaveProvider = null; | |
try { | |
// open audio file | |
if(audioFile.ToLower().EndsWith(".mp3")) | |
audioWaveProvider = new Mp3FileReader(audioFile); | |
else if(audioFile.ToLower().EndsWith(".wav")) | |
audioWaveProvider = new WaveFileReader(audioFile); | |
else | |
throw new Exception("Unsupported format."); | |
// create a sample provider | |
ISampleProvider audioSampleProvider; | |
switch(audioWaveProvider.WaveFormat.BitsPerSample) { | |
case 8: | |
audioSampleProvider = new Pcm8BitToSampleProvider(audioWaveProvider); | |
break; | |
case 16: | |
audioSampleProvider = new Pcm16BitToSampleProvider(audioWaveProvider); | |
break; | |
case 24: | |
audioSampleProvider = new Pcm24BitToSampleProvider(audioWaveProvider); | |
break; | |
case 32: | |
audioSampleProvider = new Pcm32BitToSampleProvider(audioWaveProvider); | |
break; | |
default: | |
throw new Exception("Unsupported bits per sample in audio file."); | |
} | |
// read all samples (includes left and right channel in case of stereo) | |
float[] allSamples = new float[audioWaveProvider.Length / (audioWaveProvider.WaveFormat.BitsPerSample / 8)]; | |
audioSampleProvider.Read(allSamples, 0, allSamples.Length); | |
// unzip to one channel samples that are to be analyzed | |
float[] samples; | |
if(audioWaveProvider.WaveFormat.Channels == 2) | |
samples = AudioUtility.Unzip(allSamples).Item1; | |
else | |
samples = allSamples; | |
return AnalyzeAudio(samples,audioWaveProvider.WaveFormat.SampleRate); | |
} finally { | |
audioWaveProvider?.Dispose(); | |
} | |
} | |
/// <summary> | |
/// If the count of samples is not a power of 2, it will be padded by zeros until the next larger power of 2. | |
/// </summary> | |
public static Frequency[] AnalyzeAudio(float[] samples,int sampleRate) | |
{ | |
// size must be a power of 2 | |
int size = 1; | |
int m = 0; | |
while(size < samples.Length) { | |
size <<= 1; | |
++m; | |
} | |
// create complex numbers from floats (X is real, Y is imag) | |
Complex[] fftResults = new Complex[size]; | |
for(int i = samples.Length - 1; i >= 0; --i) | |
fftResults[i] = new Complex() { X = samples[i], Y = 0 }; | |
// fourier transform | |
FastFourierTransform.FFT(true, m, fftResults); | |
float frequencyStep = sampleRate / (float)size; | |
Frequency[] frequencies = new Frequency[size / 2]; | |
for(int i = 0; i < frequencies.Length; ++i) { | |
Complex v = fftResults[i]; | |
Frequency f = new Frequency(); | |
f.Hz = (i+1) * frequencyStep; | |
f.Amplitude = (float)Math.Sqrt(Math.Pow(v.X, 2) + Math.Pow(v.Y, 2)); | |
f.Phase = (float)Math.Atan(v.Y / v.X); | |
frequencies[i] = f; | |
} | |
return frequencies; | |
} | |
} | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment