Published May 28, 2015 | Version v1
Conference paper Open

Browsing soundscapes

Description

Browsing soundscapes is generally based on the waveform of the audio signal or textual metadata, which may be not informative. The TM-charts provide an efficient tool to represent and compare soundscapes. However, they remain little used probably due to the human annotation they need. In this paper, we describe a new approach to compute charts of soundscapes, that we call Samocharts. The Samocharts are inspired by TM-charts and can be computed without a human annotation. We present two methods for Samochart computation. The first one is based on a segmentation of the signal from a set of predefined sound events. The second one is based on the confidence score of the detection algorithms. We describe two application cases on corpora of field recording from the CIESS and the UrbanSound projects. Finally, Samocharts provide a compact and efficient representation of soundscapes, which can be used in different kind of applications.

Files

18-Guyot-BrowsingSoundscapes.pdf

Files (565.5 kB)

Name Size Download all
md5:6b3ece885dde4bf0667ce38b18414015
565.5 kB Preview Download