Everyday sound datasets DCASE Datalist

Name Collection
name
Related
datasets
Provider Abbreviation D S License Size Year Cite Paper title General
domains
Data
modalities
Total
duration
(min)
Files Length
consistency
File
length (sec)
Content
type
Scene
content
Unique
event
instances
in synthetic
mixtures
Recording
setup type
Recording
setups
Recording
spot type
Data
type
Material
source
Variability
source
Audio
type
Format Lossy
compression
Bit
rate
Sampling
rate
Channel
setup
Channels Meta
types
Scene
classes
Scene
class
balance
Scene
class
list
Event
classes
Event
list
Event
instance
count
Event
instance
per class
Event
class
balance
Event
annotation
type
Event
annotation
source
Event
ann.
overlapping
instances
Event
ann.
labelled
Event
ann.
validated
Event
ann.
strong
Event
labeling /
hierarchical
Event
labeling /
ontology
Data
split
Split
sets
Split
folds
Baseline Baseline
cite
Evaluation campaigns Comments


Everyday sounds

This data list pulls together various type of datasets containing everyday sounds. These datasets are suitable for research focusing on sound event detection, sound event detection and localization, or audio tagging. A sound event corresponds to an audio segment that is attributed to a specific sound source and that is perceived as an entity. Sound event has start and end timestamps along with a textual label that is related to the sound source. Some datasets in this list contains either the strong annotations, annotations with start and end timestamps, or weak annotations, annotations with sound presence at clip/time-segment level.