license: cc-by-sa-4.0
dataset_info:
features:
- name: video_id
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: labels
sequence: string
- name: human_labels
sequence: string
splits:
- name: test
num_bytes: 1960420240.25
num_examples: 6182
- name: train
num_bytes: 4999231653.125
num_examples: 15759
download_size: 6813053099
dataset_size: 6959651893.375
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
- split: train
path: data/train-*
Re-Upload
This repository is a re-upload of akgphysics/AudioSet in Parquet format, with all audio resampled to 16 KHz using torchaudio.transforms.Resample.
Author's Description
Audio Set: An ontology and human-labeled dataset for audio events
Audio event recognition, the human-like ability to identify and relate sounds from audio, is a nascent problem in machine perception. Comparable problems such as object detection in images have reaped enormous benefits from comprehensive datasets - principally ImageNet. This paper describes the creation of Audio Set, a large-scale dataset of manually-annotated audio events that endeavors to bridge the gap in data availability between image and audio research. Using a carefully structured hierarchical ontology of 632 audio classes guided by the literature and manual curation, we collect data from human labelers to probe the presence of specific audio classes in 10 second segments of YouTube videos. Segments are proposed for labeling using searches based on metadata, context (e.g., links), and content analysis. The result is a dataset of unprecedented breadth and size that will, we hope, substantially stimulate the development of high-performance audio event recognizers.
Jort F. Gemmeke; Daniel P. W. Ellis; Dylan Freedman; Aren Jansen; Wade Lawrence; R. Channing Moore; Manoj Plakal; Marvin Ritter et al., 10.1109/ICASSP.2017.7952261
License
- The AudioSet labels are published under CC-BY-4.0 (attribution.)
- The audio under the AudioSet ontology is published under CC-BY-SA-4.0 (attribution, share-alike.)
- Individual contribution information may be viewed at research.google.com.
To summarize the license:
- if you use the audio from this dataset, you must share your contributions under CC-BY-SA and cite the original authors of that audio.
- if you use the labels from this dataset, you must include a citation for Google research, exampled below.
Note: The above is not legal advice, nor does it replace the need for you to read the entirety of the licenses above before making a decision on how to use this dataset. If you're unsure whether or not your usage is allowable as per the license terms, please consult a legal professional.
Citation
@inproceedings{jort_audioset_2017,
title = {Audio Set: An ontology and human-labeled dataset for audio events},
author = {Jort F. Gemmeke and Daniel P. W. Ellis and Dylan Freedman and Aren Jansen and Wade Lawrence and R. Channing Moore and Manoj Plakal and Marvin Ritter},
year = {2017},
booktitle = {Proc. IEEE ICASSP 2017},
address = {New Orleans, LA}
}