TFGridNetOptions
Options & Configuration · AiDotNet.Audio.Enhancement
Configuration options for the TF-GridNet (Time-Frequency GridNet) speech enhancement model.
Inheritance: ModelOptions → TFGridNetOptions
For Beginners
TF-GridNet processes audio like reading a grid:
- Audio is represented as a time-frequency grid (spectrogram)
- The network looks across time (left-right) to track how sounds change
- Then it looks across frequency (up-down) to understand the harmonic structure
- By alternating these views, it builds a complete understanding of the audio
Imagine cleaning a dirty photo by first cleaning each row, then each column, and repeating - that is how TF-GridNet cleans audio.
How It Works
TF-GridNet (Wang et al., ICASSP 2023) applies alternating attention along the time and frequency axes in a grid pattern, achieving state-of-the-art performance. On the WSJ0-2mix benchmark it reaches 23.4 dB SI-SNRi, and on DNS Challenge 2020 it achieves PESQ 3.41.
References:
- Paper: “TF-GridNet: Making Time-Frequency Domain Models Great Again” (Wang et al., ICASSP 2023)
- Repository:
https://github.com/espnet/espnet
Constructors
| Constructor | Summary |
|---|---|
TFGridNetOptions() |
Properties
| Property | Summary |
|---|---|
OnnxModelOptions OnnxOptions { get; set; } |
Gets or sets ONNX runtime options. |
double DropoutRate { get; set; } |
Gets or sets the dropout rate. |
double EnhancementStrength { get; set; } |
Gets or sets the enhancement strength (0.0 = no enhancement, 1.0 = maximum). |
double LearningRate { get; set; } |
Gets or sets the initial learning rate. |
double WeightDecay { get; set; } |
Gets or sets the weight decay for regularization. |
int EmbeddingDim { get; set; } |
Gets or sets the embedding dimension for each T-F bin. |
int FftSize { get; set; } |
Gets or sets the FFT window size. |
int HiddenDim { get; set; } |
Gets or sets the hidden dimension for the grid network. |
int HopLength { get; set; } |
Gets or sets the hop length between frames. |
int InterFrameHiddenSize { get; set; } |
Gets or sets the LSTM hidden size for inter-frame (time) processing. |
int IntraFrameHiddenSize { get; set; } |
Gets or sets the LSTM hidden size for intra-frame (frequency) processing. |
int NumAttentionHeads { get; set; } |
Gets or sets the number of attention heads. |
int NumBlocks { get; set; } |
Gets or sets the number of grid blocks (each contains time + frequency attention). |
int NumFreqBins { get; set; } |
Gets or sets the number of frequency bins. |
int NumSources { get; set; } |
Gets or sets the number of sources (1 for enhancement, 2+ for separation). |
int SampleRate { get; set; } |
Gets or sets the audio sample rate in Hz. |
string ModelPath { get; set; } |
Gets or sets the path to a pre-trained ONNX model file. |