public MainWindow()
{
InitializeComponent();
try
{
this.logger = Gnosis.Utilities.Log4NetLogger.GetDefaultLogger(typeof(MainWindow));
}
catch (Exception loggerEx)
{
throw new ApplicationException("Could not initialize logger", loggerEx);
}
try
{
logger.Info("Initializing Alexandria");
mediaFactory = new MediaFactory(logger);
securityContext = new SecurityContext(mediaFactory);
tagTypeFactory = new TagTypeFactory();
mediaRepository = new SQLiteMediaRepository(logger, mediaFactory);
mediaRepository.Initialize();
linkRepository = new SQLiteLinkRepository(logger);
linkRepository.Initialize();
tagRepository = new SQLiteTagRepository(logger, tagTypeFactory);
tagRepository.Initialize();
metadataRepository = new SQLiteMetadataRepository(logger, securityContext, mediaFactory);
metadataRepository.Initialize();
marqueeRepository = new SQLiteMarqueeRepository(logger);
audioStreamFactory = new AudioStreamFactory();
videoPlayer = new Gnosis.Video.Vlc.VideoPlayerControl();
videoPlayer.Initialize(logger, () => GetVideoHost());
catalogController = new CatalogController(logger, securityContext, mediaFactory, mediaRepository, linkRepository, tagRepository, metadataRepository, audioStreamFactory);
spiderFactory = new SpiderFactory(logger, securityContext, mediaFactory, linkRepository, tagRepository, mediaRepository, metadataRepository, audioStreamFactory);
metadataController = new MediaItemController(logger, securityContext, mediaFactory, linkRepository, tagRepository, metadataRepository);
taskController = new TaskController(logger, mediaFactory, videoPlayer, spiderFactory, metadataController, marqueeRepository, metadataRepository);
tagController = new TagController(logger, tagRepository);
commandController = new CommandController(logger);
taskResultView.Initialize(logger, securityContext, mediaFactory, metadataController, taskController, tagController, videoPlayer);
//taskManagerView.Initialize(logger, taskController, taskResultView);
searchView.Initialize(logger, taskController, taskResultView);
commandView.Initialize(logger, commandController, taskController, taskResultView);
ScreenSaver.Disable();
}
catch (Exception ex)
{
logger.Error("MainWindow.ctor", ex);
}
}
private void AddConcatenatedFiles(List <FileInfo> fileInfos)
{
if (fileInfos.Select(fi => AudioStreamFactory.IsSupportedFile(fi.FullName)).Count() == fileInfos.Count)
{
AudioTrack audioTrack = new AudioTrack(fileInfos.ToArray());
trackList.Add(audioTrack);
}
}
private void AddFile(string fileName)
{
if (AudioStreamFactory.IsSupportedFile(fileName))
{
AudioTrack audioTrack = new AudioTrack(new FileInfo(fileName));
trackListBox.Items.Add(audioTrack);
}
}
private void AddFile(FileInfo fileInfo)
{
try {
AudioStreamFactory.IsSupportedFileOrThrow(fileInfo.FullName);
AudioTrack audioTrack = new AudioTrack(fileInfo);
trackList.Add(audioTrack);
} catch (Exception e) {
MessageBox.Show(this, e.Message, "Cannot open file: " + fileInfo.Name, MessageBoxButton.OK, MessageBoxImage.Error);
}
}
private void SetAudioTrack(AudioTrack audioTrack)
{
UnsetAudioTrack();
this.audioTrack = audioTrack;
audioStream = AudioStreamFactory.FromAudioTrackForGUI(audioTrack);
audioStream.WaveformChanged += OnAudioStreamWaveformChanged;
audioTrack.LengthChanged += OnAudioTrackLengthChanged;
audioTrack.VolumeChanged += OnAudioTrackVolumeChanged;
audioTrack.BalanceChanged += OnAudioTrackBalanceChanged;
}
public MainWindow()
{
InitializeComponent();
// Use PFFFT as FFT implementation
FFTFactory.Factory = new Aurio.PFFFT.FFTFactory();
// Use Soxr as resampler implementation
ResamplerFactory.Factory = new Aurio.Soxr.ResamplerFactory();
// Use FFmpeg for file reading/decoding
AudioStreamFactory.AddFactory(new FFmpegAudioStreamFactory());
}
public void Generate()
{
IAudioStream audioStream = inputTrack.File ?
AudioStreamFactory.FromFileInfoIeee32(inputTrack.FileInfo) :
inputTrack.Stream;
audioStream = new MonoStream(audioStream);
audioStream = new ResamplingStream(audioStream, ResamplingQuality.Medium, profile.SampleRate);
STFT stft = new STFT(audioStream, profile.FrameSize, profile.FrameStep, WindowType.Hann, STFT.OutputFormat.Decibel, this.bufferSize);
index = 0;
indices = stft.WindowCount;
frameBuffer = new float[profile.FrameSize / 2];
List <SubFingerprint> subFingerprints = new List <SubFingerprint>();
while (stft.HasNext())
{
// Get FFT spectrum
stft.ReadFrame(frameBuffer);
// Sum FFT bins into target frequency bands
profile.MapFrequencies(frameBuffer, bands);
CalculateSubFingerprint(bandsPrev, bands, subFingerprints);
CommonUtil.Swap <float[]>(ref bands, ref bandsPrev);
index++;
// Output subfingerprints every once in a while
if (index % this.eventInterval == 0 && SubFingerprintsGenerated != null)
{
SubFingerprintsGenerated(this, new SubFingerprintsGeneratedEventArgs(inputTrack, subFingerprints, index, indices));
subFingerprints.Clear();
}
}
// Output remaining subfingerprints
if (SubFingerprintsGenerated != null)
{
SubFingerprintsGenerated(this, new SubFingerprintsGeneratedEventArgs(inputTrack, subFingerprints, index, indices));
}
if (Completed != null)
{
Completed(this, EventArgs.Empty);
}
audioStream.Close();
}
public IAudioStream CreateAudioStream()
{
IAudioStream stream = null;
if (MultiFile)
{
stream = new ConcatenationStream(FileInfos.Select(fi => AudioStreamFactory.FromFileInfoIeee32(fi)).ToArray());
}
else
{
stream = AudioStreamFactory.FromFileInfoIeee32(FileInfo);
}
return(new TimeWarpStream(stream, timeWarps));
}
public IAudioStream OpenFile(FileInfo fileInfo)
{
if (FFmpegSourceStream.WaveProxySuggested(fileInfo))
{
Console.WriteLine("File format with known seek problems, creating proxy file...");
return(AudioStreamFactory.FromFileInfo(FFmpegSourceStream.CreateWaveProxy(fileInfo)));
}
else
{
try
{
FFmpegSourceStream stream = new FFmpegSourceStream(fileInfo);
// Make a seek to test if it works or if it throws an exception
stream.Position = 0;
return(stream);
}
catch (FFmpegSourceStream.FileNotSeekableException)
{
/*
* This exception gets thrown if a file is not seekable and therefore cannot
* provide all the functionality that is needed for an IAudioStream, although
* the problem could be solved by creating a seek index. See FFmpegSourceStream
* for further information.
*
* For now, we create a WAV proxy file, because it is easier (consumes
* additional space though).
*/
Console.WriteLine("File not seekable, creating proxy file...");
return(AudioStreamFactory.FromFileInfo(FFmpegSourceStream.CreateWaveProxy(fileInfo)));
}
catch (FFmpegSourceStream.FileSeekException)
{
/*
* This exception gets thrown if a file should be seekable but seeking still does
* not work correctly. We also create a proxy in this case.
*/
Console.WriteLine("File test seek failed, creating proxy file...");
return(AudioStreamFactory.FromFileInfo(FFmpegSourceStream.CreateWaveProxy(fileInfo)));
}
catch (DllNotFoundException e)
{
throw new DllNotFoundException("Cannot open file through FFmpeg: DLL missing", e);
}
}
}
private static void Main(string[] args)
{
// Use PFFFT as FFT implementation
FFTFactory.Factory = new Aurio.PFFFT.FFTFactory();
// Use Soxr as resampler implementation
ResamplerFactory.Factory = new Aurio.Soxr.ResamplerFactory();
// Use FFmpeg for file reading/decoding
AudioStreamFactory.AddFactory(new FFmpegAudioStreamFactory());
try
{
// read config file
Dictionary <string, double> mapping = ReadConfig(args[0]);
// read input dir
DirectoryInfo indir = new DirectoryInfo(args[1]);
// read output dir
DirectoryInfo outdir = new DirectoryInfo(args[2]);
if (!outdir.Exists)
{
outdir.Create();
}
Process(mapping, indir, outdir);
}
catch (Exception e)
{
Console.WriteLine("Batch Resampling Tool to change the playback speed of audio files");
Console.WriteLine();
Console.WriteLine("Error: " + e.Message);
Console.WriteLine();
var info = "Usage: BatchResampler mappingFile inputDirectory outputDirectory" + Environment.NewLine +
Environment.NewLine +
"The mappingFile is a text file with at least one line of the pattern 'filenamePattern;resamplingFactor' (without quotes), " +
"where the filenamePattern can be a filename or a filename wildcard expression to be matched in the input directory, and the " +
"resamplingFactor is a floating point value specifying the input:output resampling ratio with which the files will be written" +
"to the output directory. The simplest mappingFile would be '*;1', which means that all files in the input directory will be" +
"resampled with a factor of 1.0 (i.e. no resampling at all) and written to the output directory. The nominal sampling rate of a file" +
"stays untouched, leading to altered playback speeds." + Environment.NewLine +
"A mappingFile with the content '*.wav;0.5' speeds up all wave audio files to 200% playback speed because they are getting resampled " +
"to half of their effective sampling rate while the nominal playback sample rate stays the same, cutting their playback duration in half.";
Console.WriteLine(info);
}
}
private void bw_DoWork(object sender, DoWorkEventArgs e)
{
BackgroundWorker worker = sender as BackgroundWorker;
Parameters parameters = (Parameters)e.Argument;
var streams = parameters.SourceFiles.Select(fi => AudioStreamFactory.FromFileInfo(fi));
// load source stream
var sourceStream = new ConcatenationStream(streams.ToArray());
var firstFile = parameters.SourceFiles.First();
string targetFileNamePrefix = firstFile.FullName.Remove(firstFile.FullName.Length - firstFile.Extension.Length);
string targetFileNameSuffix = firstFile.Extension;
int partCount = 0;
Random random = new Random();
CropStream cropStream = new CropStream(sourceStream, 0, 0);
while (sourceStream.Position < sourceStream.Length)
{
partCount++;
int length = random.Next(parameters.MinLength, parameters.MaxLength); // length in seconds of the current part to write
long byteLength = TimeUtil.TimeSpanToBytes(new TimeSpan(TimeUtil.SECS_TO_TICKS * length), cropStream.Properties);
Debug.WriteLine("writing part " + partCount + " (" + length + " secs = " + byteLength + " bytes)");
Debug.WriteLine("before: " + cropStream.Begin + " / " + cropStream.End + " / " + cropStream.Position + " / " + sourceStream.Position);
cropStream.Begin = cropStream.End;
cropStream.End += sourceStream.Length - cropStream.Begin < byteLength ? sourceStream.Length - cropStream.Begin : byteLength;
cropStream.Position = 0;
Debug.WriteLine("after : " + cropStream.Begin + " / " + cropStream.End + " / " + cropStream.Position + " / " + sourceStream.Position);
AudioStreamFactory.WriteToFile(cropStream, String.Format("{0}.part{1:000}{2}", targetFileNamePrefix, partCount, targetFileNameSuffix));
if (worker.CancellationPending)
{
e.Cancel = true;
Debug.WriteLine("canceled");
return;
}
worker.ReportProgress((int)((double)sourceStream.Position / sourceStream.Length * 100));
}
Debug.WriteLine("finished");
}
static void Process(Dictionary <string, double> mapping, DirectoryInfo indir, DirectoryInfo outdir)
{
Dictionary <FileInfo, double> fileMapping = new Dictionary <FileInfo, double>();
foreach (string fileNamePattern in mapping.Keys)
{
double factor = mapping[fileNamePattern];
foreach (FileInfo fileInfo in indir.EnumerateFiles(fileNamePattern))
{
fileMapping.Add(fileInfo, factor);
}
}
Parallel.ForEach <FileInfo>(fileMapping.Keys, (fileInfo) => {
double factor = fileMapping[fileInfo];
FileInfo outputFileInfo = new FileInfo(Path.Combine(outdir.FullName, fileInfo.Name));
if (outputFileInfo.Exists)
{
Console.WriteLine(fileInfo.Name + " SKIP (file already existing)");
return;
}
Console.WriteLine(fileInfo.Name);
try
{
IAudioStream inputStream = AudioStreamFactory.FromFileInfoIeee32(fileInfo);
IAudioStream resamplingStream = new ResamplingStream(inputStream, ResamplingQuality.VeryHigh, factor);
MixerStream sampleRateResetStream = new MixerStream(resamplingStream.Properties.Channels, inputStream.Properties.SampleRate);
sampleRateResetStream.Add(resamplingStream);
IAudioStream outputStream = sampleRateResetStream;
AudioStreamFactory.WriteToFile(outputStream, outputFileInfo.FullName);
}
catch (Exception e)
{
Console.WriteLine("Error processing " + fileInfo.Name + ": " + e.Message);
}
});
}
public AudioTrack(FileInfo[] fileInfos, bool initialize)
: base(fileInfos)
{
this.TimeWarps = new TimeWarpCollection();
if (initialize)
{
using (IAudioStream stream = AudioStreamFactory.FromFileInfo(FileInfo)) {
sourceProperties = stream.Properties;
if (MultiFile)
{
// For multi-file tracks, we need to get a concatenated stream of all files for the length
InitializeLength();
}
else
{
// Single-file tracks can just reuse this stream to get the length
InitializeLength(stream);
}
}
}
}
private void SetAudioTrack(AudioTrack audioTrack)
{
UnsetAudioTrack();
// init renderers
waveformBitmapRenderers = new WaveformBitmapRenderer[audioTrack.SourceProperties.Channels];
for (int i = 0; i < waveformBitmapRenderers.Length; i++)
{
waveformBitmapRenderers[i] = new WaveformBitmapRenderer();
}
waveformGeometryRenderers = new WaveformGeometryRenderer[audioTrack.SourceProperties.Channels];
for (int i = 0; i < waveformGeometryRenderers.Length; i++)
{
waveformGeometryRenderers[i] = new WaveformGeometryRenderer();
}
this.audioTrack = audioTrack;
audioStream = AudioStreamFactory.FromAudioTrackForGUI(audioTrack);
audioStream.WaveformChanged += OnAudioStreamWaveformChanged;
audioTrack.LengthChanged += OnAudioTrackLengthChanged;
audioTrack.VolumeChanged += OnAudioTrackVolumeChanged;
audioTrack.BalanceChanged += OnAudioTrackBalanceChanged;
}
public void Generate(AudioTrack track)
{
IAudioStream audioStream = new ResamplingStream(
new MonoStream(AudioStreamFactory.FromFileInfoIeee32(track.FileInfo)),
ResamplingQuality.Medium, profile.SamplingRate);
STFT stft = new STFT(audioStream, profile.WindowSize, profile.HopSize, WindowType.Hann, STFT.OutputFormat.Decibel);
int index = 0;
int indices = stft.WindowCount;
int processedFrames = 0;
float[] spectrum = new float[profile.WindowSize / 2];
float[] smoothedSpectrum = new float[spectrum.Length - profile.SpectrumSmoothingLength + 1]; // the smooved frequency spectrum of the current frame
var spectrumSmoother = new SimpleMovingAverage(profile.SpectrumSmoothingLength);
float[] spectrumTemporalAverage = new float[spectrum.Length]; // a running average of each spectrum bin over time
float[] spectrumResidual = new float[spectrum.Length]; // the difference between the current spectrum and the moving average spectrum
var peakHistory = new PeakHistory(1 + profile.TargetZoneDistance + profile.TargetZoneLength, spectrum.Length / 2);
var peakPairs = new List <PeakPair>(profile.PeaksPerFrame * profile.PeakFanout); // keep a single instance of the list to avoid instantiation overhead
var subFingerprints = new List <SubFingerprint>();
while (stft.HasNext())
{
// Get the FFT spectrum
stft.ReadFrame(spectrum);
// Skip frames whose average spectrum volume is below the threshold
// This skips silent frames (zero samples) that only contain very low noise from the FFT
// and that would screw up the temporal spectrum average below for the following frames.
if (spectrum.Average() < spectrumMinThreshold)
{
index++;
continue;
}
// Smooth the frequency spectrum to remove small peaks
if (profile.SpectrumSmoothingLength > 0)
{
spectrumSmoother.Clear();
for (int i = 0; i < spectrum.Length; i++)
{
var avg = spectrumSmoother.Add(spectrum[i]);
if (i >= profile.SpectrumSmoothingLength)
{
smoothedSpectrum[i - profile.SpectrumSmoothingLength] = avg;
}
}
}
// Update the temporal moving bin average
if (processedFrames == 0)
{
// Init averages on first frame
for (int i = 0; i < spectrum.Length; i++)
{
spectrumTemporalAverage[i] = spectrum[i];
}
}
else
{
// Update averages on all subsequent frames
for (int i = 0; i < spectrum.Length; i++)
{
spectrumTemporalAverage[i] = ExponentialMovingAverage.UpdateMovingAverage(
spectrumTemporalAverage[i], profile.SpectrumTemporalSmoothingCoefficient, spectrum[i]);
}
}
// Calculate the residual
// The residual is the difference of the current spectrum to the temporal average spectrum. The higher
// a bin residual is, the steeper the increase in energy in that peak.
for (int i = 0; i < spectrum.Length; i++)
{
spectrumResidual[i] = spectrum[i] - spectrumTemporalAverage[i] - 90f;
}
// Find local peaks in the residual
// The advantage of finding peaks in the residual instead of the spectrum is that spectrum energy is usually
// concentrated in the low frequencies, resulting in a clustering of the highest peaks in the lows. Getting
// peaks from the residual distributes the peaks more evenly across the spectrum.
var peaks = peakHistory.List; // take oldest list,
peaks.Clear(); // clear it, and
FindLocalMaxima(spectrumResidual, peaks); // refill with new peaks
// Pick the largest n peaks
int numMaxima = Math.Min(peaks.Count, profile.PeaksPerFrame);
if (numMaxima > 0)
{
peaks.Sort((p1, p2) => p1.Value == p2.Value ? 0 : p1.Value < p2.Value ? 1 : -1); // order peaks by height
if (peaks.Count > numMaxima)
{
peaks.RemoveRange(numMaxima, peaks.Count - numMaxima); // select the n tallest peaks by deleting the rest
}
peaks.Sort((p1, p2) => p1.Index == p2.Index ? 0 : p1.Index < p2.Index ? -1 : 1); // sort peaks by index (not really necessary)
}
peakHistory.Add(index, peaks);
if (FrameProcessed != null)
{
// Mark peaks as 0dB for spectrogram display purposes
foreach (var peak in peaks)
{
spectrum[peak.Index] = 0;
spectrumResidual[peak.Index] = 0;
}
FrameProcessed(this, new FrameProcessedEventArgs {
AudioTrack = track, Index = index, Indices = indices,
Spectrum = spectrum, SpectrumResidual = spectrumResidual
});
}
processedFrames++;
index++;
if (processedFrames >= peakHistory.Length)
{
peakPairs.Clear();
FindPairsWithMaxEnergy(peakHistory, peakPairs);
ConvertPairsToSubFingerprints(peakPairs, subFingerprints);
}
if (subFingerprints.Count > 512)
{
FireFingerprintHashesGenerated(track, indices, subFingerprints);
subFingerprints.Clear();
}
}
// Flush the remaining peaks of the last frames from the history to get all remaining pairs
for (int i = 0; i < profile.TargetZoneLength; i++)
{
var peaks = peakHistory.List;
peaks.Clear();
peakHistory.Add(-1, peaks);
peakPairs.Clear();
FindPairsWithMaxEnergy(peakHistory, peakPairs);
ConvertPairsToSubFingerprints(peakPairs, subFingerprints);
}
FireFingerprintHashesGenerated(track, indices, subFingerprints);
audioStream.Close();
}
/// <summary>
/// This method generates hash codes from an audio stream in a streaming fashion,
/// which means that it only maintains a small constant-size state and can process
/// streams of arbitrary length.
///
/// Here is a scheme of the data processing flow. After the subband splitting
/// stage, every subband is processed independently.
///
/// +-----------------+ +--------------+ +-----------+
/// audio stream +---> mono conversion +---> downsampling +---> whitening |
/// +-----------------+ +--------------+ +---------+-+
/// |
/// +-------------------+ +------------------+ |
/// | subband splitting <---+ subband analysis <---+
/// +--+---+---+---+----+ +------------------+
/// | | | |
/// | v v v
/// | ... ... ...
/// |
/// | +------------------+ +-----------------+
/// +---> RMS downsampling +---> onset detection |
/// +------------------+ +----------+------+
/// |
/// +-----------------+ |
/// hash codes <-------------------+ hash generation <----------+
/// +-----------------+
///
/// The hash codes from the hash generators of each band are then sent though a
/// sorter which brings them into sequential temporal order before they are stored
/// in the final list.
/// </summary>
/// <param name="track"></param>
public void Generate(AudioTrack track)
{
IAudioStream audioStream = new ResamplingStream(
new MonoStream(AudioStreamFactory.FromFileInfoIeee32(track.FileInfo)),
ResamplingQuality.Medium, profile.SamplingRate);
var whiteningStream = new WhiteningStream(audioStream,
profile.WhiteningNumPoles, profile.WhiteningDecaySecs, profile.WhiteningBlockLength);
var subbandAnalyzer = new SubbandAnalyzer(whiteningStream);
float[] analyzedFrame = new float[profile.SubBands];
var bandAnalyzers = new BandAnalyzer[profile.SubBands];
for (int i = 0; i < profile.SubBands; i++)
{
bandAnalyzers[i] = new BandAnalyzer(profile, i);
}
List <SubFingerprint> hashes = new List <SubFingerprint>();
HashTimeSorter hashSorter = new HashTimeSorter(profile.SubBands);
var sw = new Stopwatch();
sw.Start();
int totalFrames = subbandAnalyzer.WindowCount;
int currentFrame = 0;
while (subbandAnalyzer.HasNext())
{
subbandAnalyzer.ReadFrame(analyzedFrame);
for (int i = 0; i < profile.SubBands; i++)
{
bandAnalyzers[i].ProcessSample(analyzedFrame[i], hashSorter.Queues[i]);
}
if (currentFrame % 4096 == 0)
{
hashSorter.Fill(hashes, false);
if (SubFingerprintsGenerated != null)
{
SubFingerprintsGenerated(this, new SubFingerprintsGeneratedEventArgs(track, hashes, currentFrame, totalFrames));
hashes.Clear();
}
}
currentFrame++;
}
for (int i = 0; i < bandAnalyzers.Length; i++)
{
bandAnalyzers[i].Flush(hashSorter.Queues[i]);
}
hashSorter.Fill(hashes, true);
if (SubFingerprintsGenerated != null)
{
SubFingerprintsGenerated(this, new SubFingerprintsGeneratedEventArgs(track, hashes, currentFrame, totalFrames));
hashes.Clear();
}
sw.Stop();
audioStream.Close();
Console.WriteLine("time: " + sw.Elapsed);
}
public float Run()
{
IAudioStream audioStream = new ResamplingStream(
new MonoStream(AudioStreamFactory.FromFileInfoIeee32(audioTrack.FileInfo)),
ResamplingQuality.Medium, 11000);
ContinuousFrequencyActivationQuantifier cfaq = new ContinuousFrequencyActivationQuantifier(audioStream);
float[] cfaValue = new float[1];
float[] cfaValues = new float[cfaq.WindowCount];
Label[] cfaLabels = new Label[cfaq.WindowCount];
int count = 0;
int musicCount = 0;
while (cfaq.HasNext())
{
cfaq.ReadFrame(cfaValue);
cfaValues[count] = cfaValue[0];
if (cfaValue[0] > threshold)
{
musicCount++;
cfaLabels[count] = Label.MUSIC;
}
Console.WriteLine("cfa {0,3}% {3} {1,5:0.00} {2}", (int)(Math.Round((float)count++ / cfaq.WindowCount * 100)), cfaValue[0], cfaValue[0] > threshold ? "MUSIC" : "", TimeUtil.BytesToTimeSpan(audioStream.Position, audioStream.Properties));
}
audioStream.Close();
if (smoothing)
{
// 3.3 Smoothing
/* majority filtering with sliding window ~5 secs
* 1 frame = ~2,4 secs, at least 3 frames are needed for majority filtering -> 3 * ~2,4 secs = ~7,2 secs */
// filter out single NO_MUSIC frames
for (int i = 2; i < cfaLabels.Length; i++)
{
if (cfaLabels[i - 2] == Label.MUSIC && cfaLabels[i - 1] == Label.NO_MUSIC && cfaLabels[i] == Label.MUSIC)
{
cfaLabels[i - 1] = Label.MUSIC;
}
}
// filter out single MUSIC frames
for (int i = 2; i < cfaLabels.Length; i++)
{
if (cfaLabels[i - 2] == Label.NO_MUSIC && cfaLabels[i - 1] == Label.MUSIC && cfaLabels[i] == Label.NO_MUSIC)
{
cfaLabels[i - 1] = Label.NO_MUSIC;
}
}
// swap ~5 secs NO_MUSIC segments to MUSIC
for (int i = 3; i < cfaLabels.Length; i++)
{
if (cfaLabels[i - 3] == Label.MUSIC && cfaLabels[i - 2] == Label.NO_MUSIC && cfaLabels[i - 1] == Label.NO_MUSIC && cfaLabels[i] == Label.MUSIC)
{
cfaLabels[i - 1] = Label.MUSIC;
cfaLabels[i - 2] = Label.MUSIC;
}
}
// swap ~5 secs NMUSIC segments to NO_MUSIC
for (int i = 3; i < cfaLabels.Length; i++)
{
if (cfaLabels[i - 3] == Label.NO_MUSIC && cfaLabels[i - 2] == Label.MUSIC && cfaLabels[i - 1] == Label.MUSIC && cfaLabels[i] == Label.NO_MUSIC)
{
cfaLabels[i - 1] = Label.NO_MUSIC;
cfaLabels[i - 2] = Label.NO_MUSIC;
}
}
}
float musicRatio = (float)musicCount / count;
float musicRatioSmoothed = -1f;
Console.WriteLine("'" + audioTrack.FileInfo.FullName + "' contains " + ((int)(Math.Round(musicRatio * 100))) + "% music");
if (smoothing)
{
musicCount = cfaLabels.Count <Label>(l => l == Label.MUSIC);
musicRatioSmoothed = (float)musicCount / count;
Console.WriteLine("smoothed: " + ((int)(Math.Round(musicRatioSmoothed * 100))) + "% music");
}
if (writeLog)
{
FileInfo logFile = new FileInfo(audioTrack.FileInfo.FullName + ".music");
StreamWriter writer = logFile.CreateText();
writer.WriteLine(musicRatio + "; " + musicRatioSmoothed);
writer.WriteLine(threshold);
for (int i = 0; i < cfaValues.Length; i++)
{
writer.WriteLine("{0:0.00000}; {1}; \t{2}", cfaValues[i], cfaValues[i] > threshold ? Label.MUSIC : Label.NO_MUSIC, cfaLabels[i]);
}
writer.Flush();
writer.Close();
}
return(0);
}
public void Generate(AudioTrack track)
{
IAudioStream audioStream = new ResamplingStream(
new MonoStream(AudioStreamFactory.FromFileInfoIeee32(track.FileInfo)),
ResamplingQuality.Medium, profile.SamplingRate);
var chroma = new Chroma(audioStream, profile.WindowSize, profile.HopSize, profile.WindowType,
profile.ChromaMinFrequency, profile.ChromaMaxFrequency, false, profile.ChromaMappingMode);
float[] chromaFrame;
var chromaBuffer = new RingBuffer <float[]>(profile.ChromaFilterCoefficients.Length);
var chromaFilterCoefficients = profile.ChromaFilterCoefficients;
var filteredChromaFrame = new double[Chroma.Bins];
var classifiers = profile.Classifiers;
var maxFilterWidth = classifiers.Max(c => c.Filter.Width);
var integralImage = new IntegralImage(maxFilterWidth, Chroma.Bins);
int index = 0;
int indices = chroma.WindowCount;
var subFingerprints = new List <SubFingerprint>();
while (chroma.HasNext())
{
// Get chroma frame buffer
// When the chroma buffer is full, we can take and reuse the oldest array
chromaFrame = chromaBuffer.Count == chromaBuffer.Length ? chromaBuffer[0] : new float[Chroma.Bins];
// Read chroma frame into buffer
chroma.ReadFrame(chromaFrame);
// ChromaFilter
chromaBuffer.Add(chromaFrame);
if (chromaBuffer.Count < chromaBuffer.Length)
{
// Wait for the buffer to fill completely for the filtering to start
continue;
}
Array.Clear(filteredChromaFrame, 0, filteredChromaFrame.Length);
for (int i = 0; i < chromaFilterCoefficients.Length; i++)
{
var frame = chromaBuffer[i];
for (int j = 0; j < frame.Length; j++)
{
filteredChromaFrame[j] += frame[j] * chromaFilterCoefficients[i];
}
}
// ChromaNormalizer
double euclideanNorm = 0;
for (int i = 0; i < filteredChromaFrame.Length; i++)
{
var value = filteredChromaFrame[i];
euclideanNorm += value * value;
}
euclideanNorm = Math.Sqrt(euclideanNorm);
if (euclideanNorm < profile.ChromaNormalizationThreshold)
{
Array.Clear(filteredChromaFrame, 0, filteredChromaFrame.Length);
}
else
{
for (int i = 0; i < filteredChromaFrame.Length; i++)
{
filteredChromaFrame[i] /= euclideanNorm;
}
}
// ImageBuilder
// ... just add one feature vector after another as rows to the image
integralImage.AddColumn(filteredChromaFrame);
// FingerprintCalculator
if (integralImage.Columns < maxFilterWidth)
{
// Wait for the image to fill completely before hashes can be generated
continue;
}
// Calculate subfingerprint hash
uint hash = 0;
for (int i = 0; i < classifiers.Length; i++)
{
hash = (hash << 2) | grayCodeMapping[classifiers[i].Classify(integralImage, 0)];
}
// We have a SubFingerprint@frameTime
subFingerprints.Add(new SubFingerprint(index, new SubFingerprintHash(hash), false));
index++;
if (index % 512 == 0 && SubFingerprintsGenerated != null)
{
SubFingerprintsGenerated(this, new SubFingerprintsGeneratedEventArgs(track, subFingerprints, index, indices));
subFingerprints.Clear();
}
}
if (SubFingerprintsGenerated != null)
{
SubFingerprintsGenerated(this, new SubFingerprintsGeneratedEventArgs(track, subFingerprints, index, indices));
}
if (Completed != null)
{
Completed(this, EventArgs.Empty);
}
audioStream.Close();
}
static void Main(string[] args)
{
if (args.Length == 0)
{
Console.WriteLine("no file(s) specified");
Console.WriteLine();
Console.WriteLine("Usage: MusicDetector file_1 file_2 ... file_n");
Console.WriteLine("");
Console.WriteLine("This tool expects at least one file, either specified as filename or " +
"filename wildcard pattern. It scans all files for music content and writes " +
"detection results in separate text files named {file_x}.music. Errors are logged " +
"to a seperate error.log text file.");
return;
}
// Use PFFFT as FFT implementation
FFTFactory.Factory = new Aurio.PFFFT.FFTFactory();
// Use Soxr as resampler implementation
ResamplerFactory.Factory = new Aurio.Soxr.ResamplerFactory();
// Use FFmpeg for file reading/decoding
AudioStreamFactory.AddFactory(new FFmpegAudioStreamFactory());
Queue <FileInfo> scanQueue = new Queue <FileInfo>();
foreach (string file in args)
{
if (file.Contains("*"))
{
int slashIndex = file.LastIndexOf(@"\");
var di = new DirectoryInfo(slashIndex > -1 ? file.Substring(0, slashIndex) : ".");
foreach (var fi in di.GetFiles(file.Substring(slashIndex > -1 ? slashIndex + 1 : 0), SearchOption.TopDirectoryOnly))
{
scanQueue.Enqueue(fi);
}
}
else
{
scanQueue.Enqueue(new FileInfo(file));
}
}
FileInfo errorLogFileInfo = new FileInfo("error.log");
StreamWriter errorLogWriter = errorLogFileInfo.AppendText();
foreach (var fi in scanQueue)
{
try {
if (fi.Exists && fi.Length > 0)
{
new CFA(new AudioTrack(fi), CFA.DEFAULT_THRESHOLD, true, true).Run();
}
else
{
throw new FileNotFoundException(String.Format("file '{0}' not existing or empty, skipping", fi.FullName));
}
}
catch (FileNotFoundException e) {
Console.WriteLine(DateTime.Now + " " + e.Message);
errorLogWriter.WriteLine(DateTime.Now + " " + e.Message);
}
catch (Exception e) {
Console.WriteLine(DateTime.Now + " " + e.ToString());
Console.WriteLine();
errorLogWriter.WriteLine(DateTime.Now + " " + fi.FullName);
errorLogWriter.WriteLine(e.ToString());
}
}
errorLogWriter.Flush();
errorLogWriter.Close();
}
public MainWindow()
{
// Use PFFFT as FFT implementation
FFTFactory.Factory = new Aurio.PFFFT.FFTFactory();
// Use Soxr as resampler implementation
ResamplerFactory.Factory = new Aurio.Soxr.ResamplerFactory();
// Use FFmpeg for file reading/decoding
AudioStreamFactory.AddFactory(new FFmpegAudioStreamFactory());
recentProjects = new RecentProjects();
/*
* The menu items of the recently opened project are handled here in code behind
* because in XAML (ItemsSource/CompositeCollection/CollectionViewSource/CollectionContainer)
* there's a few problems I spent too much time with and could not solve. This
* programmatic solution works perfectly.
*
* - When overriding the ItemContainerStyle, which is necessary to automatically
* set the Header and Command, the style breaks... this is probably because of
* the weird style combination I'm using in this app and wouldn't happen with
* the default style (setting ItemContainerStyle replaces the style in the locally
* included style file and falls back to proerties of the default style for the app).
* - When setting the header text through the style, the header text of all statically
* defined MenuItems gets overwritten because they don't have the header directly
* set but get the header text from the associated command. So the header text
* overrules the command text.
*/
recentProjects.MenuEntries.CollectionChanged += delegate(object sender, System.Collections.Specialized.NotifyCollectionChangedEventArgs e) {
// Always rebuilt the whole menu... happens very seldomly and shouldn't be a noticable performance impact
int separatorIndex = FileMenu.Items.IndexOf(RecentSeparator);
// Clear old entries
if (FileMenu.Items.Count > (separatorIndex + 1))
{
for (int x = FileMenu.Items.Count - 1; x > separatorIndex; x--)
{
FileMenu.Items.RemoveAt(x);
}
}
Debug.Assert(FileMenu.Items.Count == separatorIndex + 1, "wrong menu delete count");
// Add new entries
int count = 0;
foreach (RecentProjects.RecentEntry entry in recentProjects.MenuEntries)
{
// Determine if this item represents a real project to be numbered
bool projectEntry = entry.Parameter != null && entry.Parameter != RecentProjects.ClearCommand;
FileMenu.Items.Add(new MenuItem {
Header = (projectEntry ? (++count) + " " : "") + entry.Title.Replace("_", "__"),
IsEnabled = entry.Enabled,
Command = Commands.FileOpenRecentProject,
CommandParameter = entry.Parameter
});
}
Debug.Assert(FileMenu.Items.Count == separatorIndex + 1 + recentProjects.MenuEntries.Count, "wrong menu item count");
};
project = new Project();
trackList = new TrackList <AudioTrack>();
InitializeComponent();
recentProjects.Load();
}
/// <summary>
/// Creates a Wave format proxy file in the same directory and with the same name as the specified file,
/// if no storage directory is specified (i.e. if it is null). If a storage directory is specified, the proxy
/// file will be stored in the specified directory with a hashed file name to avoid name collisions and
/// file overwrites. The story directory option is convenient for the usage of temporary or working directories.
/// </summary>
/// <param name="fileInfo">the file for which a proxy file should be created</param>
/// <param name="storageDirectory">optional directory where the proxy file will be stored, can be null</param>
/// <returns>the FileInfo of the proxy file</returns>
public static FileInfo CreateWaveProxy(FileInfo fileInfo, DirectoryInfo storageDirectory)
{
FileInfo outputFileInfo;
if (storageDirectory == null)
{
// Without a storage directory, store the proxy file beside the original file
outputFileInfo = new FileInfo(fileInfo.FullName + ".ffproxy.wav");
}
else
{
// With a storage directory specified, store the proxy file with a hashed name
// (to avoid name collision / overwrites) in the target directory (e.g. a temp or working directory)
using (var sha256 = SHA256.Create()) {
byte[] hash = sha256.ComputeHash(Encoding.Unicode.GetBytes(fileInfo.FullName));
string hashString = BitConverter.ToString(hash).Replace("-", "").ToLowerInvariant();
outputFileInfo = new FileInfo(Path.Combine(storageDirectory.FullName, hashString + ".ffproxy.wav"));
}
}
if (outputFileInfo.Exists)
{
Console.WriteLine("Proxy already existing, using " + outputFileInfo.Name);
return(outputFileInfo);
}
var reader = new FFmpegReader(fileInfo, FFmpeg.Type.Audio);
var writer = new MemoryWriterStream(new AudioProperties(
reader.AudioOutputConfig.format.channels,
reader.AudioOutputConfig.format.sample_rate,
reader.AudioOutputConfig.format.sample_size * 8,
reader.AudioOutputConfig.format.sample_size == 4 ? AudioFormat.IEEE : AudioFormat.LPCM));
int output_buffer_size = reader.AudioOutputConfig.frame_size * writer.SampleBlockSize;
byte[] output_buffer = new byte[output_buffer_size];
int samplesRead;
long timestamp;
FFmpeg.Type type;
// sequentially read samples from decoder and write it to wav file
while ((samplesRead = reader.ReadFrame(out timestamp, output_buffer, output_buffer_size, out type)) > 0)
{
int bytesRead = samplesRead * writer.SampleBlockSize;
writer.Write(output_buffer, 0, bytesRead);
}
reader.Dispose();
writer.Position = 0;
AudioStreamFactory.WriteToFile(writer, outputFileInfo.FullName);
writer.Close();
return(outputFileInfo);
}
