International Audio Laboratories Erlangen

AID: Open-Source Anechoic Interferer Dataset

Welcome to the website accompanying the paper "AID: Open-Source Anechoic Interferer Dataset" by Philipp Götz, Cagdas Tuna, Andreas Walther, and Emanuël A. P. Habets

Abstract

A dataset of anechoic recordings of various sound sources encountered in domestic environments is presented. The dataset is intended to be a resource of non-stationary, environmental noise signals that, when convolved with acoustic impulse responses, may be used to simulate complex acoustic scenes. Additionally, a Python library is provided, which implements a signal generator for random mixtures of the recordings in the dataset, that may be used as non-stationary interference signals.

Signal Generator: Examples

The following example mixtures were generated with the number of concurrently playing samples set to 2.