AudioNoise<T>

Models & Types · AiDotNet.Augmentation.Audio

Adds background noise to audio data.

Inheritance: AugmentationBase<T, Tensor<T>> → AudioAugmenterBase<T> → AudioNoise<T>

For Beginners

This augmentation adds random noise to audio, simulating real-world recording conditions like background hum, ambient sounds, or electronic interference. This helps models become robust to noisy inputs.

How It Works

SNR (Signal-to-Noise Ratio):

  • Higher SNR = less noise (cleaner audio)
  • Lower SNR = more noise (noisier audio)
  • 20 dB = barely audible noise
  • 10 dB = noticeable noise
  • 0 dB = signal and noise are equal

Constructors

Constructor Summary
AudioNoise(double minSnrDb, double maxSnrDb, double probability, int sampleRate) Creates a new audio noise augmentation.

Properties

Property Summary
NoiseType NoiseType { get; set; } Gets or sets the type of noise to add.
double MaxSnrDb { get; } Gets the maximum signal-to-noise ratio in dB.
double MinSnrDb { get; } Gets the minimum signal-to-noise ratio in dB.

Methods

Method Summary
IDictionary<string, object> GetParameters()