Recognizing the quality of urban sound recordings using hand-crafted and deep audio features

Abstract

Soundscape can be regarded as the auditory landscape, conceived individually or at collaborative level. This paper presents a method for automatic recognition of the soundscape quality of urban recordings. Towards this end, the ATHens Urban Soundscape has been used, which is a dataset of audio recordings of ambient urban sounds, annotated in terms of the corresponding perceived soundscape quality. In order to automatically recognize the soundscape quality, both hand-crafted and deep features have been adopted. Experimental results have demonstrated that the performance of the final classifier that combines hand-crafted and deep context-aware audio features is boosted by almost 2%.