FDYSEDOptions

Options & Configuration · AiDotNet.Audio.Classification

Configuration options for the FDY-SED (Frequency Dynamic Sound Event Detection) model.

Inheritance: ModelOptions → FDYSEDOptions

For Beginners

FDY-SED is like having a team of specialized listeners, where each listener is an expert at hearing different pitch ranges:

  • One listener is good at hearing low rumbles (bass) like washing machines or traffic
  • Another is good at mid-range sounds like speech or dog barks
  • Another specializes in high-pitched sounds like bird chirps or alarms

Instead of using the same “ear” for all frequencies (like a standard CNN), FDY-SED adapts its filters based on which frequency range it’s analyzing. This makes it more accurate at detecting sounds that have very different spectral characteristics.

How It Works

FDY-SED (Nam et al., ICASSP 2022) introduces frequency-dynamic convolutions for sound event detection, where convolution kernels are dynamically generated based on the frequency band being processed. This allows the model to apply different processing strategies to different frequency ranges (e.g., bass frequencies need different patterns than treble). FDY-SED achieves DCASE challenge-winning results on the DESED dataset for domestic sound event detection.

Constructors

Constructor Summary
FDYSEDOptions()

Properties

Property Summary
OnnxModelOptions OnnxOptions { get; set; } Gets or sets ONNX runtime options.
double DetectionWindowSize { get; set; } Gets or sets the window size in seconds for detection.
double DropoutRate { get; set; } Gets or sets the dropout rate.
double LabelSmoothing { get; set; } Gets or sets the label smoothing factor.
double LearningRate { get; set; } Gets or sets the learning rate.
double MeanTeacherWeight { get; set; } Gets or sets the mean teacher consistency weight for semi-supervised training.
double Threshold { get; set; } Gets or sets the confidence threshold for event detection.
double WindowOverlap { get; set; } Gets or sets the window overlap ratio (0-1).
int EmbeddingDim { get; set; } Gets or sets the embedding dimension after CNN blocks.
int FDYKernelHiddenSize { get; set; } Gets or sets the frequency dynamic kernel generation hidden size.
int FMax { get; set; } Gets or sets the maximum frequency for mel filterbank.
int FMin { get; set; } Gets or sets the minimum frequency for mel filterbank.
int FftSize { get; set; } Gets or sets the FFT window size.
int HopLength { get; set; } Gets or sets the hop length between FFT frames.
int NumFrequencyGroups { get; set; } Gets or sets the number of frequency dynamic convolution groups.
int NumMels { get; set; } Gets or sets the number of mel filterbank channels.
int NumRNNLayers { get; set; } Gets or sets the number of RNN layers.
int RNNHiddenSize { get; set; } Gets or sets the RNN hidden size for temporal modeling.
int SampleRate { get; set; } Gets or sets the audio sample rate in Hz.
int[] CNNChannels { get; set; } Gets or sets the number of CNN channels per block.
string ModelPath { get; set; } Gets or sets the path to a pre-trained ONNX model file.
string Variant { get; set; } Gets or sets the model variant (“small”, “base”, “large”).
string[] CustomLabels { get; set; } Gets or sets custom event labels.