Interface IAudioFeatureExtractor<T>

Namespace

AiDotNet.Interfaces

Assembly

AiDotNet.dll

Defines the contract for audio feature extraction algorithms.

public interface IAudioFeatureExtractor<T>

Type Parameters

T

The numeric type used for calculations.

Remarks

Audio feature extractors transform raw audio waveforms into meaningful representations that can be used for tasks like speech recognition, music analysis, and audio classification.

For Beginners: Audio files are just sequences of numbers representing sound waves. Feature extractors convert these raw numbers into more useful formats:

• MFCC: Captures how humans perceive different frequencies
• Chroma: Represents musical pitch classes (C, C#, D, etc.)
• Spectral: Measures brightness, contrast, and other spectral properties

Properties

FeatureDimension

Gets the number of features produced per frame.

int FeatureDimension { get; }

Property Value

int

Name

Gets the name of this feature extractor.

string Name { get; }

Property Value

string

SampleRate

Gets the sample rate expected by this extractor.

int SampleRate { get; }

Property Value

int

Methods

Extract(Tensor<T>)

Extracts features from an audio waveform.

Tensor<T> Extract(Tensor<T> audio)

Parameters

audio Tensor<T>

The audio waveform as a 1D tensor [samples].

Returns

Tensor<T>

Features as a 2D tensor [frames, features].

Extract(Vector<T>)

Extracts features from an audio waveform.

Matrix<T> Extract(Vector<T> audio)

Parameters

audio Vector<T>

The audio waveform as a Vector.

Returns

Matrix<T>

Features as a Matrix [frames x features].

ExtractAsync(Tensor<T>, CancellationToken)

Extracts features from an audio waveform asynchronously.

Task<Tensor<T>> ExtractAsync(Tensor<T> audio, CancellationToken cancellationToken = default)

Parameters

audio Tensor<T>

The audio waveform.

cancellationToken CancellationToken

Cancellation token.

Returns

Task<Tensor<T>>

Features as a 2D tensor [frames, features].