CRNNEventDetectorOptions
Options & Configuration · AiDotNet.Audio.Classification
Configuration options for the CRNN (Convolutional Recurrent Neural Network) Sound Event Detector.
Inheritance: ModelOptions → CRNNEventDetectorOptions
For Beginners
CRNN is a classic and reliable approach to detecting sounds over time. Think of it in two stages:
- CNN stage: Looks at the spectrogram like a picture, finding patterns like “this
frequency pattern looks like a dog bark” or “this shape looks like a piano note”
- RNN stage: Reads the patterns over time like reading a sentence, understanding
“the dog bark started, continued, and then stopped”
This combination makes CRNN good at both identifying what sounds are present AND when they happen. It’s simpler than Transformer-based models but still very effective.
How It Works
CRNN for SED (Cakir et al., 2017) combines convolutional layers for spectral feature extraction with recurrent layers (GRU/LSTM) for temporal modeling. It is the standard baseline architecture for the DCASE Sound Event Detection challenge and achieves strong results on AudioSet, ESC-50, and URBAN-SED benchmarks. The model processes mel spectrograms through CNN blocks, then models temporal dependencies with bidirectional GRU layers, producing frame-level event probabilities.
Constructors
| Constructor | Summary |
|---|---|
CRNNEventDetectorOptions() |
Properties
| Property | Summary |
|---|---|
OnnxModelOptions OnnxOptions { get; set; } |
Gets or sets ONNX runtime options. |
bool Bidirectional { get; set; } |
Gets or sets whether to use bidirectional RNN. |
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 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 CNNKernelSize { get; set; } |
Gets or sets the CNN kernel 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 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 PoolSize { get; set; } |
Gets or sets the CNN pooling size for frequency dimension. |
int RNNHiddenSize { get; set; } |
Gets or sets the RNN hidden size. |
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. |