Skip to main content


Table 4 Top four results from Soundwalks data set for text-based retrieval with out of vocabulary query "rail". Parenthetical descriptions are not actual tags, but provided to give an idea of the acoustic content of the sound files.

From: An Ontological Framework for Retrieving Environmental Sounds Using Semantics and Acoustic Content

Posterior probability Node path Relevant
0.19 railtrainsegment94.wav (train bell)segment165.wav (traffic/train horn) No
0.17 railvoicesegment136.wav (pa announcement)segment133.wav (pa announcement) Yes
0.15 railtrainsegment40.wav (train brakes)segment30.wav (train bell/brakes) Yes
0.09 railtrainsegmen40.wav (train brakes)segment147.wav (train horn) Yes