DESED 2020Task4 dataset

Citation Author(s):
Nicolas
Turpault
Romain
Serizel
Ankit
Shah
Justin
Salamon
Submitted by:
Jian Zhuang
Last updated:
Fri, 09/01/2023 - 04:51
DOI:
10.21227/ptnf-cz34
Data Format:
*.wav
Links:
DCASE_2020
License:
Creative Commons Attribution
76 Views
Categories:
Signal Processing
Analog signal processing
Keywords:
DCASE, IEEE, Data Challenge, IEEE Signal Processing Society
0
0 ratings - Please login to submit your rating.
ACCESS DATASET CITE SHARE/EMBED

Abstract 

DESED dataset is the dataset that was used in DCASE 2019 task 4. The dataset for this task is composed of 10 sec audio clips recorded in domestic environment or synthesized using Scaper to simulate a domestic environment. The task focuses on 10 class of sound events that represent a subset of Audioset (not all the classes are present in Audioset, some classes of sound events are including several classes from Audioset):

  • Speech Speech
  • Dog Dog
  • Cat Cat
  • Alarm/bell/ringing Alarm_bell_ringing
  • Dishes Dishes
  • Frying Frying
  • Blender Blender
  • Running water Running_water
  • Vacuum cleaner Vacuum_cleaner
  • Electric shaver/toothbrush Electric_shaver_toothbrush

More information about this dataset and how to generate synthetic soundscapes can be found on DESED website.

Instructions: 

There are 3 different datasets:

  • Recorded soundscapes (a.k.a. real).
  • Soundbank to generate synthetic soundscapes.
  • Public evaluation (recorded soundscapes) (a.k.a., Youtube in DCASE19, Vimeo is not available): DESED public eval

DESED dataset is for now composed of 10 event classes in domestic environment.

You can

  • Use only the real dataset.
  • Use the soundbank to create your own synthetic soundscapes. (generate new mixtures using Scaper [1])
  • Reproduce the soundscapes made for DCASE task 4.

Dataset Files

    Files have not been uploaded for this dataset

    Documentation

    AttachmentSize
    File README.md11.56 KB

    QUESTIONS?