Automatically Extract Audio Features

Feature extraction is the process of converting data to a numeric array that can be easily used to compute similarity measures, classification and training. This example shows automatic feature extraction for audio recordings.

Collect all of the audio signals in ExampleData.

Use FeatureSpacePlot to embed the signals in a semantically meaningful 2D feature space.

Use FeatureNearest to define a function that returns the nearest signal by comparing the automatically extracted features.

Consider a recording of a dog bark.

Retrieve the signal that is closest to it in the feature space.

Create an automatic feature extractor using FeatureExtraction from the example data collection.

Visualize the computed features.