metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: labels
list: int8
- name: label_names
list: string
- name: filename
dtype: string
- name: recording_site
dtype:
class_label:
names:
'0': market
'1': park
'2': square
'3': street
'4': waterfront
splits:
- name: train
num_bytes: 30876280531.6
num_examples: 17500
- name: validation
num_bytes: 6616345505.7
num_examples: 3750
- name: test
num_bytes: 6616345985.7
num_examples: 3750
download_size: 44108296475
dataset_size: 44108972022.99999
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
license: cc-by-4.0
License: Creative Commons Attribution 4.0 International
Source
Rhoddy Viveros-Muñoz, Pablo Huijse, Victor Poblete, Victor Vargas, Diego Espejo, Matthieu Vernier, Diego Vergara, Jorge Arenas, & Enrique Suárez. (2022).
SPASS dataset: A synthetic polyphonic dataset with spatiotemporal labels of sound sources (1.0). Zenodo.
https://doi.org/10.5281/zenodo.7484370