Audio Classifier Visualizer

A Python library for visualizing audio classifier outputs, including waveforms, spectrograms, and class probabilities over time.

This library was extracted from the elephant-rumble-inference project, which uses deep learning to detect and classify elephant rumble vocalizations in audio recordings.

Usage

import audio_classifier_visualizer as acv

afv = acv.AudioFileVisualizer("audio_file.wav",
                              start_time=0, end_time=60*5, 
                              feature_rate=feature_rate, class_probabilities=probs, class_labels = class_labels)

plt = afv.visualize_audio_file_fragment(f"Title",  width=20, height=8)

plt.show()

Demo notebook here

Features

• Visualize audio waveforms
• Generate spectrograms with optional color overlays for classification results
• Plot class probabilities over time
• Highly configurable plotting options
• Support for both short and long audio files

Installation

pip install --quiet git+https://github.com/ramayer/audio-classifier-visualizer

License

audio-classifier-visualizer is distributed under the terms of the MIT license.