Birdvox-Dcase-20K: A Dataset For Bird Audio Detection In 10-Second Clips

BirdVox-DCASE-20k: a dataset for bird audio detection in 10-second clips<br>
=====================================================<br>
Version 1.0, March 2018.

<br>
Created By<br>
-------------

Vincent Lostanlen (1, 2, 3), Justin Salamon (2, 3), Andrew Farnsworth (1), Steve Kelling (1), and Juan Pablo Bello (2, 3).

(1): Cornell Lab of Ornithology (CLO)<br>
(2): Center for Urban Science and Progress, New York University<br>
(3): Music and Audio Research Lab, New York University

https://wp.nyu.edu/birdvox

Description<br>
--------------

The BirdVox-DCASE-20k dataset contains 20,000 ten-second audio recordings. These recordings come from ROBIN autonomous recording units, placed near Ithaca, NY, USA during the fall 2015. They were captured on the night of September 23rd, 2015, by six different sensors, originally numbered 1, 2, 3, 5, 7, and 10.

Out of these 20,000 recording, 10,017 (50.09%) contain at least one bird vocalization (either song, call, or chatter).

The dataset is a derivative work of the BirdVox-full-night dataset [1], containing almost as much data but formatted into ten-second excerpts rather than ten-hour full night recordings.

In addition, the BirdVox-DCASE-20k dataset is provided as a development set in the context of the "Bird Audio Detection" challenge, organized by DCASE (Detection and Classification of Acoustic Scenes and Events) and the IEEE Signal Processing Society.

The dataset can be used, among other things, for the development and evaluation of bioacoustic classification models.

<br>
We refer the reader to [1] for details on the distribution of the data and [2] for details on the hardware of ROBIN recording units.

[1] V. Lostanlen, J. Salamon, A. Farnsworth, S. Kelling, J.P. Bello. "BirdVox-full-night: a dataset and benchmark for avian flight call detection", Proc. IEEE ICASSP, 2018.

[2] J. Salamon, J. P. Bello, A. Farnsworth, M. Robbins, S. Keen, H. Klinck, and S. Kelling. Towards the Automatic Classification of Avian Flight Calls for Bioacoustic Monitoring. PLoS One, 2016.

Data Files<br>
------------

The wav folder contains the recordings as WAV files, sampled at 44,1 kHz, with a single channel (mono). The original sample rate was 24 kHz.

The name of each wav file is a random 128-bit UUID (Universal Unique IDentifier) string, which is randomized with respect to the origin of the recording in BirdVox-full-night, both in terms of time (UTC hour at the start of the excerpt) and space (location of the sensor).

The origin of each 10-second excerpt is known by the challenge organizers, but not disclosed to the participants.

Metadata Files<br>
--------------

A table containing a binary label "hasbird" associated to every recording in BirdVox-DCASE-20k is available on the website of the DCASE "Bird Audio Detection" challenge: http://machine-listening.eecs.qmul.ac.uk/bird-audio-detection-challenge/

These labels were automatically derived from the annotations of avian flight call events in the BirdVox-full-night dataset.

If your evaluation procedure requires the precise timestamps of each avian flight call (at a fine time scale of 50 ms), and is agnostic to non-flight call avian vocalizations (e.g. geese, crows, owls, etc.), we kindly suggest you to use the BirdVox-full-night dataset rather than BirdVox-DCASE-20k: wp.nyu.edu/birdvox/birdvox-full-night

On the other hand, if your evaluation procedure encompasses all avian vocalizations, and is performed at a coarse time scale of 10 seconds, then BirdVox-DCASE-20k is the appropriate dataset.

The annotation campaign of avian flight calls in BirdVox-full-night was performed by Andrew Farnsworth and lasted 102 hours.

The additional annotation campaign of non-flight call avian vocalizations was performed by Vincent Lostanlen and lasted 10 hours.

The accuracy of the labeling is estimated to be somewhere between 99.5% (100 mislabelings) and 99.95% (10 mislabelings).

<br>
Please Acknowledge BirdVox-DCASE-20k in Academic Research<br>
--------------------------------------------------------------------------------

When BirdVox-70k is used for academic research, we would highly appreciate it if scientific publications of works partly based on this dataset cite the following publication:

V. Lostanlen, J. Salamon, A. Farnsworth, S. Kelling, J. Bello. "BirdVox-full-night: a dataset and benchmark for avian flight call detection", Proc. IEEE ICASSP, 2018.

@inproceedings{lostanlen2018icassp,<br>
title = {BirdVox-full-night: a dataset and benchmark for avian flight call detection},<br>
author = {Lostanlen, Vincent and Salamon, Justin and Farnsworth, Andrew and Kelling, Steve and Bello, Juan Pablo},<br>
booktitle = {Proc. IEEE ICASSP},<br>
year = {2018},<br>
published = {IEEE},<br>
venue = {Calgary, Canada},<br>
month = {April},<br>
}

The creation of this dataset was supported by NSF grants 1125098 (BIRDCAST) and 1633259 (BIRDVOX), a Google Faculty Award, the Leon Levy Foundation, and two anonymous donors.

Conditions of Use<br>
---------------------

Dataset created by Vincent Lostanlen, Justin Salamon, Andrew Farnsworth, Steve Kelling, and Juan Pablo Bello.

The BirdVox-DCASE-20k dataset is offered free of charge under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) license:<br>
https://creativecommons.org/licenses/by/4.0/

The dataset and its contents are made available on an "as is" basis and without warranties of any kind, including without limitation satisfactory quality and conformity, merchantability, fitness for a particular purpose, accuracy or completeness, or absence of errors. Subject to any liability that may not be excluded or limited by law, Cornell Lab of Ornithology is not liable for, and expressly excludes all liability for, loss or damage however and whenever caused to anyone by any use of the BirdVox-DCASE-20k dataset or any part of it.

Feedback<br>
-----------

Please help us improve BirdVox-DCASE-20k by sending your feedback to:<br>
* Vincent Lostanlen: [e-mail] for feedback regarding data pre-processing,<br>
* Andrew Farnsworth: [e-mail] for feedback regarding data collection and ornithology, or<br>
* Dan Stowell: [e-mail] for feedback regarding the DCASE "Bird Audio Detection" challenge.

In case of a problem, please include as many details as possible.

<br>
Acknowledgements<br>
------------------------

We thank Jessie Barry, Ian Davies, Tom Fredericks, Jeff Gerbracht, Sara Keen, Holger Klinck, Anne Klingensmith, Ray Mack, Peter Marchetto, Ed Moore, Matt Robbins, Ken Rosenberg, and Chris Tessaglia-Hymes for designing autonomous recording units and collecting data.<br>
We acknowledge that the land on which the data was collected is the unceded territory of the Cayuga nation, which is part of the Haudenosaunee (Iroquois) confederacy.

BirdVox-DCASE-20k：面向10秒音频片段的鸟类音频检测数据集
=====================================================
版本1.0，2018年3月。

创作者
-------------
文森特·洛斯塔伦（Vincent Lostanlen）(1,2,3)、贾斯汀·萨拉蒙（Justin Salamon）(2,3)、安德鲁·法恩斯沃思（Andrew Farnsworth）(1)、史蒂夫·凯林（Steve Kelling）(1)、胡安·巴勃罗·贝略（Juan Pablo Bello）(2,3)。

(1)：康奈尔鸟类学实验室（Cornell Lab of Ornithology, CLO）
(2)：纽约大学城市科学与进步中心
(3)：纽约大学音乐与音频研究实验室

官网：https://wp.nyu.edu/birdvox

数据集概述
--------------
BirdVox-DCASE-20k数据集包含20000条10秒长度的音频录音。这些录音源自**ROBIN自主录音设备（ROBIN autonomous recording units）**，于2015年秋季部署在美国纽约州伊萨卡附近。本次采集工作于2015年9月23日晚间完成，共使用6台传感器，原始编号分别为1、2、3、5、7、10。

在全部20000条录音中，有10017条（占比50.09%）包含至少一种鸟类鸣唱（涵盖鸣曲、鸣叫或啭鸣）。

本数据集是BirdVox-full-night数据集[1]的衍生版本：二者数据体量相近，但BirdVox-full-night以10小时完整夜间录音为存储单元，而本数据集将其切分为10秒的音频片段。

此外，本数据集作为“鸟类音频检测”挑战赛的开发集发布，该赛事由**DCASE（声学场景与事件检测与分类，Detection and Classification of Acoustic Scenes and Events）**与IEEE信号处理学会联合主办。

本数据集可用于鸟类声学分类模型的开发与评估等诸多生物声学研究场景。

如需了解数据分发细节，请参考文献[1]；如需了解ROBIN录音设备的硬件参数，请参考文献[2]。
[1] V. Lostanlen, J. Salamon, A. Farnsworth, S. Kelling, J.P. Bello. "BirdVox-full-night: a dataset and benchmark for avian flight call detection", Proc. IEEE ICASSP, 2018.
[2] J. Salamon, J. P. Bello, A. Farnsworth, M. Robbins, S. Keen, H. Klinck, and S. Kelling. Towards the Automatic Classification of Avian Flight Calls for Bioacoustic Monitoring. PLoS One, 2016.

数据文件
------------
wav文件夹内存储为WAV（波形音频格式，Waveform Audio File Format）格式的录音文件，采样率为44.1kHz，单声道（mono）。原始采样率为24kHz。

每条WAV文件的文件名均为随机生成的128位**UUID（通用唯一识别码，Universal Unique IDentifier）**字符串，该标识符会根据录音在BirdVox-full-night中的来源（包括片段起始时刻的UTC（协调世界时，Coordinated Universal Time）时间、传感器部署位置）进行随机化处理。

每条10秒音频片段的来源信息仅为挑战赛主办方掌握，不会向参赛选手公开。

元数据文件
--------------
包含每条录音对应的二进制标签“hasbird”的元数据表，可在DCASE“鸟类音频检测”挑战赛官网获取：http://machine-listening.eecs.qmul.ac.uk/bird-audio-detection-challenge/

该标签源自BirdVox-full-night数据集中的鸟类飞行鸣叫声事件标注。

如果您的评估流程需要精确到50毫秒级的鸟类飞行鸣叫声时间戳，且仅关注飞行鸣叫声而非其他鸟类鸣唱（例如鹅、乌鸦、猫头鹰等的鸣叫声），我们建议您使用BirdVox-full-night数据集而非本数据集：wp.nyu.edu/birdvox/birdvox-full-night

反之，如果您的评估流程覆盖所有鸟类鸣唱，且仅需要10秒级的粗粒度时间划分，则BirdVox-DCASE-20k是合适的选择。

BirdVox-full-night数据集中的鸟类飞行鸣叫声标注工作由安德鲁·法恩斯沃思（Andrew Farnsworth）完成，耗时102小时。
针对非飞行鸣鸟类叫声的额外标注工作由文森特·洛斯塔伦（Vincent Lostanlen）完成，耗时10小时。
标注的准确率估计在99.5%（100处标注错误）至99.95%（10处标注错误）之间。

学术研究中致谢本数据集
--------------------------------------------------------------------------------
若您在学术研究中使用BirdVox-70k，我们恳请您在部分基于本数据集的研究成果的科学出版物中引用以下文献：
V. Lostanlen, J. Salamon, A. Farnsworth, S. Kelling, J. Bello. "BirdVox-full-night: a dataset and benchmark for avian flight call detection", Proc. IEEE ICASSP, 2018.

@inproceedings{lostanlen2018icassp,
title = {BirdVox-full-night: a dataset and benchmark for avian flight call detection},
author = {Lostanlen, Vincent and Salamon, Justin and Farnsworth, Andrew and Kelling, Steve and Bello, Juan Pablo},
booktitle = {Proc. IEEE ICASSP},
year = {2018},
published = {IEEE},
venue = {Calgary, Canada},
month = {April},
}

本数据集的创建得到了美国国家科学基金会（National Science Foundation, NSF）项目编号1125098（BIRDCAST）与1633259（BIRDVOX）、Google Faculty Award、Leon Levy基金会以及两位匿名捐赠者的支持。

使用条款
---------------------
本数据集由文森特·洛斯塔伦（Vincent Lostanlen）、贾斯汀·萨拉蒙（Justin Salamon）、安德鲁·法恩斯沃思（Andrew Farnsworth）、史蒂夫·凯林（Steve Kelling）与胡安·巴勃罗·贝略（Juan Pablo Bello）共同创建。
BirdVox-DCASE-20k数据集遵循**知识共享署名4.0国际许可（Creative Commons Attribution 4.0 International, CC BY 4.0）**条款免费发布：https://creativecommons.org/licenses/by/4.0/

本数据集及相关内容按“现状”提供，不附带任何形式的明示或默示担保，包括但不限于对商品性、特定用途适用性、准确性或完整性以及无错误的担保。在法律允许的最大范围内，康奈尔鸟类学实验室不对因使用本数据集或其任何部分而导致的任何损失或损害承担责任，并明确排除所有相关责任。

反馈与建议
-----------
请通过以下联系方式向我们反馈，以帮助我们改进BirdVox-DCASE-20k：
* 文森特·洛斯塔伦（[e-mail]）：针对数据预处理相关的反馈
* 安德鲁·法恩斯沃思（[e-mail]）：针对数据采集与鸟类学相关的反馈
* 丹·斯托威尔（[e-mail]）：针对DCASE“鸟类音频检测”挑战赛相关的反馈

如遇问题，请尽可能提供详细信息。

致谢
------------------------
我们感谢杰西·巴里（Jessie Barry）、伊恩·戴维斯（Ian Davies）、汤姆·弗雷德里克斯（Tom Fredericks）、杰夫·格布拉赫特（Jeff Gerbracht）、萨拉·基恩（Sara Keen）、霍尔格·克林克（Holger Klinck）、安妮·克林根史密斯（Anne Klingensmith）、雷·马克（Ray Mack）、彼得·马切托（Peter Marchetto）、埃德·摩尔（Ed Moore）、马特·罗宾斯（Matt Robbins）、肯·罗森伯格（Ken Rosenberg）以及克里斯·特萨利亚-海姆斯（Chris Tessaglia-Hymes），感谢他们参与自主录音设备的设计与数据采集工作。
我们特此声明，数据采集所在的土地为卡尤加族（Cayuga nation）的未割让领土，该区域隶属于豪德诺索尼（Haudenosaunee，又称易洛魁）联盟。