/// <summary>
/// Prepare an audio file. This will be a single segment of a larger audio file,
/// modified based on the provided settings.
/// </summary>
/// <param name="outputDirectory">
/// The analysis Base Directory.
/// </param>
/// <param name="source">
/// The source audio file.
/// </param>
/// <param name="outputMediaType">
/// The output Media Type.
/// </param>
/// <param name="startOffset">
/// The start Offset from start of entire original file.
/// </param>
/// <param name="endOffset">
/// The end Offset from start of entire original file.
/// </param>
/// <param name="targetSampleRateHz">
/// The target Sample Rate Hz.
/// </param>
/// <returns>
/// The prepared file. The returned FileSegment will have the targetFile and OriginalFileDuration set -
/// these are the path to the segmented file and the duration of the segmented file.
/// The start and end offsets will not be set.
/// </returns>
public async Task <FileSegment> PrepareFile(
DirectoryInfo outputDirectory,
string source,
string outputMediaType,
TimeSpan startOffset,
TimeSpan endOffset,
int targetSampleRateHz)
{
throw new NotImplementedException();
var request = new AudioUtilityRequest
{
OffsetStart = startOffset,
OffsetEnd = endOffset,
TargetSampleRate = targetSampleRateHz,
};
var preparedFile = AudioFilePreparer.PrepareFile(
outputDirectory,
source.ToFileInfo(),
outputMediaType,
request,
TempFileHelper.TempDir());
return(new FileSegment(
preparedFile.TargetInfo.SourceFile,
preparedFile.TargetInfo.SampleRate.Value,
preparedFile.TargetInfo.Duration.Value));
}
/// <summary>
/// Prepare an audio file. This will be a single segment of a larger audio file, modified based on the analysisSettings.
/// </summary>
/// <param name="outputDirectory">
/// The analysis Base Directory.
/// </param>
/// <param name="source">
/// The source audio file.
/// </param>
/// <param name="outputMediaType">
/// The output Media Type.
/// </param>
/// <param name="startOffset">
/// The start Offset from start of entire original file.
/// </param>
/// <param name="endOffset">
/// The end Offset from start of entire original file.
/// </param>
/// <param name="targetSampleRateHz">
/// The target Sample Rate Hz.
/// </param>
/// <returns>
/// The prepared file. The returned FileSegment will have the targetFile and OriginalFileDuration set -
/// these are the path to the segmented file and the duration of the segmented file.
/// The start and end offsets will not be set.
/// </returns>
public async Task <FileSegment> PrepareFile(
DirectoryInfo outputDirectory,
string source,
string outputMediaType,
TimeSpan startOffset,
TimeSpan endOffset,
int targetSampleRateHz)
{
return(await Task.Run(() =>
{
FileInfo sourceFileInfo = source.ToFileInfo();
var request = new AudioUtilityRequest
{
OffsetStart = startOffset,
OffsetEnd = endOffset,
TargetSampleRate = targetSampleRateHz,
};
var preparedFile = AudioFilePreparer.PrepareFile(
outputDirectory,
sourceFileInfo,
outputMediaType,
request,
TempFileHelper.TempDir(),
oldFormat: this.useOldNamingFormat);
return new FileSegment(
preparedFile.TargetInfo.SourceFile,
preparedFile.TargetInfo.SampleRate.Value,
preparedFile.TargetInfo.Duration.Value);
}));
}
public static void SeparateChannels(
Arguments args,
FileInfo ipFile,
out double[] samplesL,
out double[] samplesR,
out double epsilon)
{
//you'd then use wavreader on the resulting preparedFile
//the channel select functionality does not currently exist in AnalyzeLongRecording. I need to add it.
var request = new AudioUtilityRequest
{
OffsetStart = args.StartOffset,
OffsetEnd = args.EndOffset,
TargetSampleRate = args.SamplingRate,
Channels = new[] { 1, 2 },
MixDownToMono = false,
};
var audioFile = AudioFilePreparer.PrepareFile(args.OpDir, ipFile, args.OutputMediaType, request, args.OpDir).TargetInfo.SourceFile;
var wavReader = new WavReader(audioFile);
var recording = new AudioRecording(wavReader);
samplesL = recording.WavReader.GetChannel(0);
samplesR = recording.WavReader.GetChannel(1);
epsilon = Math.Pow(0.5, recording.BitsPerSample - 1);
}
/// <summary>
/// A WRAPPER AROUND THE analyser.Analyze(analysisSettings) METHOD
/// To be called as an executable with command line arguments.
/// </summary>
public static void Execute(Arguments arguments)
{
Contract.Requires(arguments != null);
var(analysisSettings, segmentSettings) = arguments.ToAnalysisSettings();
TimeSpan offsetStart = TimeSpan.FromSeconds(arguments.Start ?? 0);
TimeSpan duration = TimeSpan.FromSeconds(arguments.Duration ?? 0);
int resampleRate = ConfigDictionary.GetInt(AnalysisKeys.ResampleRate, analysisSettings.ConfigDict);
// EXTRACT THE REQUIRED RECORDING SEGMENT
FileInfo tempF = segmentSettings.SegmentAudioFile;
if (tempF.Exists)
{
tempF.Delete();
}
if (duration == TimeSpan.Zero)
{
// Process entire file
AudioFilePreparer.PrepareFile(arguments.Source, tempF, new AudioUtilityRequest {
TargetSampleRate = resampleRate
}, analysisSettings.AnalysisTempDirectoryFallback);
////var fiSegment = AudioFilePreparer.PrepareFile(diOutputDir, fiSourceFile, , Human2.RESAMPLE_RATE);
}
else
{
AudioFilePreparer.PrepareFile(arguments.Source, tempF, new AudioUtilityRequest {
TargetSampleRate = resampleRate, OffsetStart = offsetStart, OffsetEnd = offsetStart.Add(duration)
}, analysisSettings.AnalysisTempDirectoryFallback);
////var fiSegmentOfSourceFile = AudioFilePreparer.PrepareFile(diOutputDir, new FileInfo(recordingPath), MediaTypes.MediaTypeWav, TimeSpan.FromMinutes(2), TimeSpan.FromMinutes(3), RESAMPLE_RATE);
}
//DO THE ANALYSIS
// #############################################################################################################################################
// BROKEN!
throw new NotImplementedException("Broken in code updates");
IAnalyser2 analyser = null; //new Rain_OBSOLETE();
AnalysisResult2 result = analyser.Analyze <FileInfo>(analysisSettings, null /*broken */);
/*DataTable dt = result.Data;
* //#############################################################################################################################################
*
* // ADD IN ADDITIONAL INFO TO RESULTS TABLE
* if (dt != null)
* {
* int iter = 0; // dummy - iteration number would ordinarily be available at this point.
* int startMinute = (int)offsetStart.TotalMinutes;
* foreach (DataRow row in dt.Rows)
* {
* row[InitialiseIndexProperties.KEYRankOrder] = iter;
* row[InitialiseIndexProperties.KEYStartMinute] = startMinute;
* row[InitialiseIndexProperties.KEYSegmentDuration] = result.AudioDuration.TotalSeconds;
* }
*
* CsvTools.DataTable2CSV(dt, segmentSettings.SegmentSummaryIndicesFile.FullName);
* //DataTableTools.WriteTable2Console(dt);
* }*/
}
/// <summary>
/// A WRAPPER AROUND THE analyzer.Analyze(analysisSettings) METHOD
/// To be called as an executable with command line arguments.
/// </summary>
/// <param name="arguments">
/// The command line arguments.
/// </param>
public static void Execute(Arguments arguments)
{
Contract.Requires(arguments != null);
var(analysisSettings, segmentSettings) = arguments.ToAnalysisSettings();
TimeSpan start = TimeSpan.FromSeconds(arguments.Start ?? 0);
TimeSpan duration = TimeSpan.FromSeconds(arguments.Duration ?? 0);
// EXTRACT THE REQUIRED RECORDING SEGMENT
FileInfo tempF = segmentSettings.SegmentAudioFile;
if (duration == TimeSpan.Zero)
{
// Process entire file
AudioFilePreparer.PrepareFile(
arguments.Source,
tempF,
new AudioUtilityRequest {
TargetSampleRate = ResampleRate
},
analysisSettings.AnalysisTempDirectoryFallback);
}
else
{
AudioFilePreparer.PrepareFile(
arguments.Source,
tempF,
new AudioUtilityRequest
{
TargetSampleRate = ResampleRate,
OffsetStart = start,
OffsetEnd = start.Add(duration),
},
analysisSettings.AnalysisTempDirectoryFallback);
}
// DO THE ANALYSIS
/* ############################################################################################################################################# */
IAnalyser2 analyser = new LitoriaFallax_OBSOLETE();
//IAnalyser2 analyser = new Canetoad();
analyser.BeforeAnalyze(analysisSettings);
AnalysisResult2 result = analyser.Analyze(analysisSettings, segmentSettings);
/* ############################################################################################################################################# */
if (result.Events.Length > 0)
{
LoggedConsole.WriteLine("{0} events found", result.Events.Length);
}
else
{
LoggedConsole.WriteLine("No events found");
}
}
/// <summary>
/// This entrypoint should be used for testing short files (less than 2 minutes)
/// </summary>
public static void Execute(Arguments arguments)
{
MainEntry.WarnIfDevleoperEntryUsed("EventRecognizer entry does not do any audio maniuplation.");
Log.Info("Running event recognizer");
var sourceAudio = arguments.Source;
var configFile = arguments.Config.ToFileInfo();
var outputDirectory = arguments.Output;
if (configFile == null)
{
throw new FileNotFoundException("No config file argument provided");
}
else if (!configFile.Exists)
{
Log.Warn($"Config file {configFile.FullName} not found... attempting to resolve config file");
configFile = ConfigFile.Resolve(configFile.Name, Directory.GetCurrentDirectory().ToDirectoryInfo());
}
LoggedConsole.WriteLine("# Recording file: " + sourceAudio.FullName);
LoggedConsole.WriteLine("# Configuration file: " + configFile);
LoggedConsole.WriteLine("# Output folder: " + outputDirectory);
// find an appropriate event IAnalyzer
IAnalyser2 recognizer = AnalyseLongRecording.FindAndCheckAnalyser <IEventRecognizer>(
arguments.AnalysisIdentifier,
configFile.Name);
Log.Info("Attempting to run recognizer: " + recognizer.Identifier);
Log.Info("Reading configuration file");
Config configuration = ConfigFile.Deserialize <RecognizerBase.RecognizerConfig>(configFile);
// get default settings
AnalysisSettings analysisSettings = recognizer.DefaultSettings;
// convert arguments to analysis settings
analysisSettings = arguments.ToAnalysisSettings(
analysisSettings,
outputIntermediate: true,
resultSubDirectory: recognizer.Identifier,
configuration: configuration);
// Enable this if you want the Config file ResampleRate parameter to work.
// Generally however the ResampleRate should remain at 22050Hz for all recognizers.
//analysisSettings.AnalysisTargetSampleRate = (int) configuration[AnalysisKeys.ResampleRate];
// get transform input audio file - if needed
Log.Info("Querying source audio file");
var audioUtilityRequest = new AudioUtilityRequest()
{
TargetSampleRate = analysisSettings.AnalysisTargetSampleRate,
};
var preparedFile = AudioFilePreparer.PrepareFile(
outputDirectory,
sourceAudio,
MediaTypes.MediaTypeWav,
audioUtilityRequest,
outputDirectory);
var source = preparedFile.SourceInfo.ToSegment();
var prepared = preparedFile.TargetInfo.ToSegment(FileSegment.FileDateBehavior.None);
var segmentSettings = new SegmentSettings <FileInfo>(
analysisSettings,
source,
(analysisSettings.AnalysisOutputDirectory, analysisSettings.AnalysisTempDirectory),
prepared);
if (preparedFile.TargetInfo.SampleRate.Value != analysisSettings.AnalysisTargetSampleRate)
{
Log.Warn("Input audio sample rate does not match target sample rate");
}
// Execute a pre analyzer hook
recognizer.BeforeAnalyze(analysisSettings);
// execute actual analysis - output data will be written
Log.Info("Running recognizer: " + recognizer.Identifier);
AnalysisResult2 results = recognizer.Analyze(analysisSettings, segmentSettings);
// run summarize code - output data can be written
Log.Info("Running recognizer summary: " + recognizer.Identifier);
recognizer.SummariseResults(
analysisSettings,
source,
results.Events,
results.SummaryIndices,
results.SpectralIndices,
new[] { results });
//Log.Info("Recognizer run, saving extra results");
// TODO: Michael, output anything else as you wish.
Log.Debug("Clean up temporary files");
if (source.Source.FullName != prepared.Source.FullName)
{
prepared.Source.Delete();
}
int eventCount = results?.Events?.Length ?? 0;
Log.Info($"Number of detected events: {eventCount}");
Log.Success(recognizer.Identifier + " recognizer has completed");
}
/// <summary>
/// A WRAPPER AROUND THE analyser.Analyse(analysisSettings) METHOD
/// To be called as an executable with command line arguments.
/// </summary>
/// <param name="sourcePath"></param>
/// <param name="configPath"></param>
/// <param name="outputPath"></param>
public static int Execute(string[] args)
{
int status = 0;
if (args.Length < 4)
{
Console.WriteLine("Require at least 4 command line arguments.");
status = 1;
return(status);
}
//GET FIRST THREE OBLIGATORY COMMAND LINE ARGUMENTS
string recordingPath = args[0];
string configPath = args[1];
string outputDir = args[2];
FileInfo fiSource = new FileInfo(recordingPath);
if (!fiSource.Exists)
{
Console.WriteLine("Source file does not exist: " + recordingPath);
status = 2;
return(status);
}
FileInfo fiConfig = new FileInfo(configPath);
if (!fiConfig.Exists)
{
Console.WriteLine("Config file does not exist: " + configPath);
status = 2;
return(status);
}
DirectoryInfo diOP = new DirectoryInfo(outputDir);
if (!diOP.Exists)
{
Console.WriteLine("Output directory does not exist: " + outputDir);
status = 2;
return(status);
}
//INIT SETTINGS
AnalysisSettings analysisSettings = new AnalysisSettings();
analysisSettings.ConfigFile = fiConfig;
analysisSettings.AnalysisRunDirectory = diOP;
analysisSettings.AudioFile = null;
analysisSettings.EventsFile = null;
analysisSettings.IndicesFile = null;
analysisSettings.ImageFile = null;
TimeSpan tsStart = new TimeSpan(0, 0, 0);
TimeSpan tsDuration = new TimeSpan(0, 0, 0);
var configuration = new ConfigDictionary(analysisSettings.ConfigFile.FullName);
analysisSettings.ConfigDict = configuration.GetTable();
//PROCESS REMAINDER OF THE OPTIONAL COMMAND LINE ARGUMENTS
for (int i = 3; i < args.Length; i++)
{
string[] parts = args[i].Split(':');
if (parts[0].StartsWith("-tmpwav"))
{
var outputWavPath = Path.Combine(outputDir, parts[1]);
analysisSettings.AudioFile = new FileInfo(outputWavPath);
}
else
if (parts[0].StartsWith("-events"))
{
string eventsPath = Path.Combine(outputDir, parts[1]);
analysisSettings.EventsFile = new FileInfo(eventsPath);
}
else
if (parts[0].StartsWith("-indices"))
{
string indicesPath = Path.Combine(outputDir, parts[1]);
analysisSettings.IndicesFile = new FileInfo(indicesPath);
}
else
if (parts[0].StartsWith("-sgram"))
{
string sonoImagePath = Path.Combine(outputDir, parts[1]);
analysisSettings.ImageFile = new FileInfo(sonoImagePath);
}
else
if (parts[0].StartsWith("-start"))
{
int s = Int32.Parse(parts[1]);
tsStart = new TimeSpan(0, 0, s);
}
else
if (parts[0].StartsWith("-duration"))
{
int s = Int32.Parse(parts[1]);
tsDuration = new TimeSpan(0, 0, s);
if (tsDuration.TotalMinutes > 10)
{
Console.WriteLine("Segment duration cannot exceed 10 minutes.");
status = 3;
return(status);
}
}
}
//EXTRACT THE REQUIRED RECORDING SEGMENT
FileInfo tempF = analysisSettings.AudioFile;
if (tsDuration.TotalSeconds == 0) //Process entire file
{
AudioFilePreparer.PrepareFile(fiSource, tempF, new AudioUtilityRequest {
SampleRate = LSKiwiHelper.RESAMPLE_RATE
});
//var fiSegment = AudioFilePreparer.PrepareFile(diOutputDir, fiSourceFile, , Human2.RESAMPLE_RATE);
}
else
{
AudioFilePreparer.PrepareFile(fiSource, tempF, new AudioUtilityRequest {
SampleRate = LSKiwiHelper.RESAMPLE_RATE, OffsetStart = tsStart, OffsetEnd = tsStart.Add(tsDuration)
});
//var fiSegmentOfSourceFile = AudioFilePreparer.PrepareFile(diOutputDir, new FileInfo(recordingPath), MediaTypes.MediaTypeWav, TimeSpan.FromMinutes(2), TimeSpan.FromMinutes(3), RESAMPLE_RATE);
}
//DO THE ANALYSIS
//#############################################################################################################################################
IAnalyser analyser = new LSKiwi2();
AnalysisResult result = analyser.Analyse(analysisSettings);
DataTable dt = result.Data;
//#############################################################################################################################################
//ADD IN ADDITIONAL INFO TO RESULTS TABLE
if (dt != null)
{
AddContext2Table(dt, tsStart, result.AudioDuration);
CsvTools.DataTable2CSV(dt, analysisSettings.EventsFile.FullName);
//DataTableTools.WriteTable(augmentedTable);
}
else
{
return(-993); //error!!
}
return(status);
}
/// <summary>
/// A WRAPPER AROUND THE analyzer.Analyze(analysisSettings) METHOD
/// To be called as an executable with command line arguments.
/// </summary>
/// <param name="arguments">
/// The command line arguments.
/// </param>
public static void Execute(Arguments arguments)
{
Contract.Requires(arguments != null);
TimeSpan start = TimeSpan.FromSeconds(arguments.Start ?? 0);
TimeSpan duration = TimeSpan.FromSeconds(arguments.Duration ?? 0);
// EXTRACT THE REQUIRED RECORDING SEGMENT
var audioUtilityRequest = new AudioUtilityRequest {
TargetSampleRate = RESAMPLE_RATE
};
if (duration == TimeSpan.Zero)
{
// Process entire file
audioUtilityRequest = new AudioUtilityRequest {
TargetSampleRate = RESAMPLE_RATE
};
}
else
{
audioUtilityRequest = new AudioUtilityRequest
{
TargetSampleRate = RESAMPLE_RATE,
OffsetStart = start,
OffsetEnd = start.Add(duration),
};
}
var preparedFile = AudioFilePreparer.PrepareFile(
arguments.Output,
arguments.Source,
MediaTypes.MediaTypeWav,
audioUtilityRequest,
arguments.Output);
var(analysisSettings, segmentSettings) = arguments.ToAnalysisSettings(
sourceSegment: preparedFile.SourceInfo.ToSegment(),
preparedSegment: preparedFile.TargetInfo.ToSegment());
// DO THE ANALYSIS
/* ############################################################################################################################################# */
IAnalyser2 analyser = new CanetoadOld_OBSOLETE();
analyser.BeforeAnalyze(analysisSettings);
AnalysisResult2 result = analyser.Analyze(analysisSettings, segmentSettings);
/* ############################################################################################################################################# */
if (result.Events.Length > 0)
{
LoggedConsole.WriteLine("{0} events found", result.Events.Length);
if (Log.IsDebugEnabled)
{
var firstEvent = (AcousticEvent)result.Events.First();
Log.Debug($"Event 0 profile: start={firstEvent.TimeStart}, duration={firstEvent.TimeStart - firstEvent.TimeEnd}");
}
}
else
{
LoggedConsole.WriteLine("No events found");
}
}
public static void Execute(Arguments arguments)
{
const string Title = "# DETERMINING SIGNAL TO NOISE RATIO IN RECORDING";
string date = "# DATE AND TIME: " + DateTime.Now;
Log.WriteLine(Title);
Log.WriteLine(date);
Log.Verbosity = 1;
var input = arguments.Source;
var sourceFileName = input.Name;
var outputDir = arguments.Output;
var fileNameWithoutExtension = Path.GetFileNameWithoutExtension(input.FullName);
var outputTxtPath = Path.Combine(outputDir.FullName, fileNameWithoutExtension + ".txt").ToFileInfo();
Log.WriteIfVerbose("# Recording file: " + input.FullName);
Log.WriteIfVerbose("# Config file: " + arguments.Config);
Log.WriteIfVerbose("# Output folder =" + outputDir.FullName);
FileTools.WriteTextFile(outputTxtPath.FullName, date + "\n# Recording file: " + input.FullName);
//READ PARAMETER VALUES FROM INI FILE
// load YAML configuration
Config configuration = ConfigFile.Deserialize(arguments.Config);
//ii: SET SONOGRAM CONFIGURATION
SonogramConfig sonoConfig = new SonogramConfig(); //default values config
sonoConfig.SourceFName = input.FullName;
sonoConfig.WindowSize = configuration.GetIntOrNull(AnalysisKeys.KeyFrameSize) ?? 512;
sonoConfig.WindowOverlap = configuration.GetDoubleOrNull(AnalysisKeys.FrameOverlap) ?? 0.5;
sonoConfig.WindowFunction = configuration[AnalysisKeys.KeyWindowFunction];
sonoConfig.NPointSmoothFFT = configuration.GetIntOrNull(AnalysisKeys.KeyNPointSmoothFft) ?? 256;
sonoConfig.NoiseReductionType = SNR.KeyToNoiseReductionType(configuration[AnalysisKeys.NoiseReductionType]);
int minHz = configuration.GetIntOrNull("MIN_HZ") ?? 0;
int maxHz = configuration.GetIntOrNull("MAX_HZ") ?? 11050;
double segK1 = configuration.GetDoubleOrNull("SEGMENTATION_THRESHOLD_K1") ?? 0;
double segK2 = configuration.GetDoubleOrNull("SEGMENTATION_THRESHOLD_K2") ?? 0;
double latency = configuration.GetDoubleOrNull("K1_K2_LATENCY") ?? 0;
double vocalGap = configuration.GetDoubleOrNull("VOCAL_GAP") ?? 0;
double minVocalLength = configuration.GetDoubleOrNull("MIN_VOCAL_DURATION") ?? 0;
//bool DRAW_SONOGRAMS = (bool?)configuration.DrawSonograms ?? true; //options to draw sonogram
//double intensityThreshold = Acoustics.AED.Default.intensityThreshold;
//if (dict.ContainsKey(key_AED_INTENSITY_THRESHOLD)) intensityThreshold = Double.Parse(dict[key_AED_INTENSITY_THRESHOLD]);
//int smallAreaThreshold = Acoustics.AED.Default.smallAreaThreshold;
//if( dict.ContainsKey(key_AED_SMALL_AREA_THRESHOLD)) smallAreaThreshold = Int32.Parse(dict[key_AED_SMALL_AREA_THRESHOLD]);
// COnvert input recording into wav
var convertParameters = new AudioUtilityRequest {
TargetSampleRate = 17640
};
var fileToAnalyse = new FileInfo(Path.Combine(outputDir.FullName, "temp.wav"));
if (File.Exists(fileToAnalyse.FullName))
{
File.Delete(fileToAnalyse.FullName);
}
var convertedFileInfo = AudioFilePreparer.PrepareFile(
input,
fileToAnalyse,
convertParameters,
outputDir);
// (A) ##########################################################################################################################
AudioRecording recording = new AudioRecording(fileToAnalyse.FullName);
int signalLength = recording.WavReader.Samples.Length;
TimeSpan wavDuration = TimeSpan.FromSeconds(recording.WavReader.Time.TotalSeconds);
double frameDurationInSeconds = sonoConfig.WindowSize / (double)recording.SampleRate;
TimeSpan frameDuration = TimeSpan.FromTicks((long)(frameDurationInSeconds * TimeSpan.TicksPerSecond));
int stepSize = (int)Math.Floor(sonoConfig.WindowSize * (1 - sonoConfig.WindowOverlap));
double stepDurationInSeconds = sonoConfig.WindowSize * (1 - sonoConfig.WindowOverlap)
/ recording.SampleRate;
TimeSpan stepDuration = TimeSpan.FromTicks((long)(stepDurationInSeconds * TimeSpan.TicksPerSecond));
double framesPerSecond = 1 / stepDuration.TotalSeconds;
int frameCount = signalLength / stepSize;
// (B) ################################## EXTRACT ENVELOPE and SPECTROGRAM ##################################
var dspOutput = DSP_Frames.ExtractEnvelopeAndFfts(
recording,
sonoConfig.WindowSize,
sonoConfig.WindowOverlap);
//double[] avAbsolute = dspOutput.Average; //average absolute value over the minute recording
// (C) ################################## GET SIGNAL WAVEFORM ##################################
double[] signalEnvelope = dspOutput.Envelope;
double avSignalEnvelope = signalEnvelope.Average();
// (D) ################################## GET Amplitude Spectrogram ##################################
double[,] amplitudeSpectrogram = dspOutput.AmplitudeSpectrogram; // get amplitude spectrogram.
// (E) ################################## Generate deciBel spectrogram from amplitude spectrogram
double epsilon = Math.Pow(0.5, recording.BitsPerSample - 1);
double[,] deciBelSpectrogram = MFCCStuff.DecibelSpectra(
dspOutput.AmplitudeSpectrogram,
dspOutput.WindowPower,
recording.SampleRate,
epsilon);
LoggedConsole.WriteLine("# Finished calculating decibel spectrogram.");
StringBuilder sb = new StringBuilder();
sb.AppendLine("\nSIGNAL PARAMETERS");
sb.AppendLine("Signal Duration =" + wavDuration);
sb.AppendLine("Sample Rate =" + recording.SampleRate);
sb.AppendLine("Min Signal Value =" + dspOutput.MinSignalValue);
sb.AppendLine("Max Signal Value =" + dspOutput.MaxSignalValue);
sb.AppendLine("Max Absolute Ampl =" + signalEnvelope.Max().ToString("F3") + " (See Note 1)");
sb.AppendLine("Epsilon Ampl (1 bit)=" + epsilon);
sb.AppendLine("\nFRAME PARAMETERS");
sb.AppendLine("Window Size =" + sonoConfig.WindowSize);
sb.AppendLine("Frame Count =" + frameCount);
sb.AppendLine("Envelope length=" + signalEnvelope.Length);
sb.AppendLine("Frame Duration =" + frameDuration.TotalMilliseconds.ToString("F3") + " ms");
sb.AppendLine("Frame overlap =" + sonoConfig.WindowOverlap);
sb.AppendLine("Step Size =" + stepSize);
sb.AppendLine("Step duration =" + stepDuration.TotalMilliseconds.ToString("F3") + " ms");
sb.AppendLine("Frames Per Sec =" + framesPerSecond.ToString("F1"));
sb.AppendLine("\nFREQUENCY PARAMETERS");
sb.AppendLine("Nyquist Freq =" + dspOutput.NyquistFreq + " Hz");
sb.AppendLine("Freq Bin Width =" + dspOutput.FreqBinWidth.ToString("F2") + " Hz");
sb.AppendLine("Nyquist Bin =" + dspOutput.NyquistBin);
sb.AppendLine("\nENERGY PARAMETERS");
double val = dspOutput.FrameEnergy.Min();
sb.AppendLine(
"Minimum dB / frame =" + (10 * Math.Log10(val)).ToString("F2") + " (See Notes 2, 3 & 4)");
val = dspOutput.FrameEnergy.Max();
sb.AppendLine("Maximum dB / frame =" + (10 * Math.Log10(val)).ToString("F2"));
sb.AppendLine("\ndB NOISE SUBTRACTION");
double noiseRange = 2.0;
//sb.AppendLine("Noise (estimate of mode) =" + sonogram.SnrData.NoiseSubtracted.ToString("F3") + " dB (See Note 5)");
//double noiseSpan = sonogram.SnrData.NoiseRange;
//sb.AppendLine("Noise range =" + noiseSpan.ToString("F2") + " to +" + (noiseSpan * -1).ToString("F2") + " dB (See Note 6)");
//sb.AppendLine("SNR (max frame-noise) =" + sonogram.SnrData.Snr.ToString("F2") + " dB (See Note 7)");
//sb.Append("\nSEGMENTATION PARAMETERS");
//sb.Append("Segment Thresholds K1: {0:f2}. K2: {1:f2} (See Note 8)", segK1, segK2);
//sb.Append("# Event Count = " + predictedEvents.Count());
FileTools.Append2TextFile(outputTxtPath.FullName, sb.ToString());
FileTools.Append2TextFile(outputTxtPath.FullName, GetSNRNotes(noiseRange).ToString());
// (F) ################################## DRAW IMAGE 1: original spectorgram
Log.WriteLine("# Start drawing noise reduced sonograms.");
TimeSpan X_AxisInterval = TimeSpan.FromSeconds(1);
//int Y_AxisInterval = (int)Math.Round(1000 / dspOutput.FreqBinWidth);
int nyquist = recording.SampleRate / 2;
int hzInterval = 1000;
var image1 = DrawSonogram(deciBelSpectrogram, wavDuration, X_AxisInterval, stepDuration, nyquist, hzInterval);
// (G) ################################## Calculate modal background noise spectrum in decibels
//double SD_COUNT = -0.5; // number of SDs above the mean for noise removal
//NoiseReductionType nrt = NoiseReductionType.MODAL;
//System.Tuple<double[,], double[]> tuple = SNR.NoiseReduce(deciBelSpectrogram, nrt, SD_COUNT);
//double upperPercentileBound = 0.2; // lowest percentile for noise removal
//NoiseReductionType nrt = NoiseReductionType.LOWEST_PERCENTILE;
//System.Tuple<double[,], double[]> tuple = SNR.NoiseReduce(deciBelSpectrogram, nrt, upperPercentileBound);
// (H) ################################## Calculate BRIGGS noise removal from amplitude spectrum
int percentileBound = 20; // low energy percentile for noise removal
//double binaryThreshold = 0.6; //works for higher SNR recordings
double binaryThreshold = 0.4; //works for lower SNR recordings
//double binaryThreshold = 0.3; //works for lower SNR recordings
double[,] m = NoiseRemoval_Briggs.BriggsNoiseFilterAndGetMask(
amplitudeSpectrogram,
percentileBound,
binaryThreshold);
string title = "TITLE";
var image2 = NoiseRemoval_Briggs.DrawSonogram(
m,
wavDuration,
X_AxisInterval,
stepDuration,
nyquist, hzInterval,
title);
//Image image2 = NoiseRemoval_Briggs.BriggsNoiseFilterAndGetSonograms(amplitudeSpectrogram, upperPercentileBound, binaryThreshold,
// wavDuration, X_AxisInterval, stepDuration, Y_AxisInterval);
// (I) ################################## Calculate MEDIAN noise removal from amplitude spectrum
//double upperPercentileBound = 0.8; // lowest percentile for noise removal
//NoiseReductionType nrt = NoiseReductionType.MEDIAN;
//System.Tuple<double[,], double[]> tuple = SNR.NoiseReduce(deciBelSpectrogram, nrt, upperPercentileBound);
//double[,] noiseReducedSpectrogram1 = tuple.Item1; //
//double[] noiseProfile = tuple.Item2; // smoothed modal profile
//SNR.NoiseProfile dBProfile = SNR.CalculateNoiseProfile(deciBelSpectrogram, SD_COUNT); // calculate noise value for each freq bin.
//double[] noiseProfile = DataTools.filterMovingAverage(dBProfile.noiseThresholds, 7); // smooth modal profile
//double[,] noiseReducedSpectrogram1 = SNR.TruncateBgNoiseFromSpectrogram(deciBelSpectrogram, dBProfile.noiseThresholds);
//Image image2 = DrawSonogram(noiseReducedSpectrogram1, wavDuration, X_AxisInterval, stepDuration, Y_AxisInterval);
var combinedImage = ImageTools.CombineImagesVertically(image1, image2);
string imagePath = Path.Combine(outputDir.FullName, fileNameWithoutExtension + ".png");
combinedImage.Save(imagePath);
// (G) ################################## Calculate modal background noise spectrum in decibels
Log.WriteLine("# Finished recording:- " + input.Name);
}
/// <summary>
/// A WRAPPER AROUND THE Analysis() METHOD
/// To be called as an executable with command line arguments.
/// </summary>
public static void Execute(Arguments arguments)
{
Contract.Requires(arguments != null);
TimeSpan tsStart = TimeSpan.FromSeconds(arguments.Start ?? 0);
TimeSpan tsDuration = TimeSpan.FromSeconds(arguments.Duration ?? 0);
string outputDir = arguments.Output.FullName;
// EXTRACT THE REQUIRED RECORDING SEGMENT
FileInfo sourceF = arguments.Source;
FileInfo tempF = TempFileHelper.NewTempFile(arguments.Output);
if (tempF.Exists)
{
tempF.Delete();
}
// GET INFO ABOUT THE SOURCE and the TARGET files - esp need the sampling rate
AudioUtilityModifiedInfo preparedFile;
if (tsDuration == TimeSpan.Zero) // Process entire file
{
preparedFile = AudioFilePreparer.PrepareFile(sourceF, tempF, new AudioUtilityRequest {
TargetSampleRate = ResampleRate
}, arguments.Output);
}
else
{
preparedFile = AudioFilePreparer.PrepareFile(sourceF, tempF, new AudioUtilityRequest {
TargetSampleRate = ResampleRate, OffsetStart = tsStart, OffsetEnd = tsStart.Add(tsDuration)
}, arguments.Output);
}
var(analysisSettings, segmentSettings) = arguments.ToAnalysisSettings(
sourceSegment: preparedFile.SourceInfo.ToSegment(),
preparedSegment: preparedFile.TargetInfo.ToSegment());
//get the data file to identify frog calls. Check it exists and then store full path in dictionary.
string frogParametersPath = analysisSettings.ConfigDict[key_FROG_DATA];
FileInfo fi_FrogData = new FileInfo(Path.Combine(outputDir, frogParametersPath));
if (!fi_FrogData.Exists)
{
LoggedConsole.WriteLine("INVALID PATH: " + fi_FrogData.FullName);
LoggedConsole.WriteLine("The config file must contain the name of a valid .csv file (containing frog call parameters) located in same directory as the .cfg file.");
LoggedConsole.WriteLine("For example, use Key/Value pair: FROG_DATA_FILE=FrogDataAndCompilationFile.csv");
throw new InvalidOperationException();
}
analysisSettings.ConfigDict[key_FROG_DATA] = fi_FrogData.FullName; // store full path in the dictionary.
// DO THE ANALYSIS
// #############################################################################################################################################
IAnalyser2 analyser = new Frogs_OBSOLETE();
AnalysisResult2 result = analyser.Analyze(analysisSettings, segmentSettings);
throw new NotImplementedException("Intentionally broken");
/*
* DataTable dt = result.Data;
* if (dt == null) { throw new InvalidOperationException("Data table of results is null"); }
* // #############################################################################################################################################
*
* // ADD IN ADDITIONAL INFO TO RESULTS TABLE
* AddContext2Table(dt, tsStart, result.AudioDuration);
* CsvTools.DataTable2CSV(dt, segmentSettings.SegmentEventsFile.FullName);
* // DataTableTools.WriteTable(augmentedTable);
*/
}
