FSDnoisy18k

sounds

General information

Label Value Description
Name FSDnoisy18k Full dataset name
ID sounds/fsdnoisy18k Datalist id for external indexing
Abbreviation FSDnoisy18k Official dataset abbreviation, e.g. one used in the original paper
Provider UPF
Year 2019 Dataset release year
Modalities Audio Data modalities included in the dataset
Collection name FSD Common name for all related datasets, used to group datasets coming from same source
Research domain Tagging Weak annotation Noisy labels Related domains, e.g., Scenes, Mobile devices, Audio-visual, Open set, Ambient noise, Unlabelled, Multiple sensors, SED, SELD, Tagging, FL, Strong annotation, Weak annotation, Unlabelled, Multi-annotator
License Creative Commons
Download Download (9.5GB)
Companion site Site Link to the companion site for the dataset
Citation [Fonseca2019] Learning Sound Event Classifiers from Web Audio with Noisy Labels

Audio

Label Value Description

Data

Data type Audio Possible values: Audio | Features
File format
File format type Constant Possible values: Constant | Variable
File format wav Possible value: wav | aiff | flac | mp3 | aac | ogg
Channels
Setup Mono Possible values: Mono | Stereo | Binaural | Ambisonic | Array | Multi-Channel | Variable
Number of channels 1
Material
Source Freesound Possible values: Original | Youtube | Freesound | Online | Crowdsourced | [Dataset name]

Content

Content type Freefield Possible values: Freefield | Synthetic | Isolated

Recording

Files

Count 18533 files Total number of files
Total duration (minutes) 2550 min Total duration of the dataset in minutes
File length Variable Characterization of the file lengths, possible values: Constant | Quasi-constant | Variable

Meta

Label Value Description
Types Tag List of meta data types provided for the data, possible values: Event, Tag, Scene, Caption, Geolocation, Spatial location, Annotator, Timestamp, Presence, Proximity, etc.

Scene

Event

Classes 20 Number of event classes
Annotation
Type Weak Possible values: Strong | Weak | Location | None
Source Crowdsourced Possible values: Experts | Crowdsourced | Synthetic | Metadata | Automatic
Annotations per item 1 How many annotations there are available per item (possible multi-annotator setup)
Labelled amount (%) 100 % Percentage of all data, amount of data which is labelled
Validated amount (%) 14.671127179 % Percentage of all data, amount of data which is validated by human
Strong annotations amount (%) 0 % Percentage of all data, amount of data which has strong annotations
Overlapping event instances Yes
Labeling
Hierarchical Yes
Ontology name Yes
Instance
Count 18533 Count of all event instances in the dataset
Average instances per class 926.65 Average per class instance count

Cross-validation setup

Label Value Description
Provided Yes
Folds 1
Sets Train Test Set types provided in the split, possible values: Train | Test | Val | Dev | Eval

Baseline

Label Value Description
Download Download Link to baseline system source code
Citation [Fonseca2019] Paper to cite for the baseline