Language-based audio retrieval DCASE 2022 evaluation dataset

This is the <strong>evaluation dataset</strong> for Task 6 (Subtask B), <strong>Language-based Audio Retrieval</strong>, in <strong>DCASE 2022</strong> Challenge. This evaluation dataset is meant to be used for the purposes of the Subtask B in the Task 6 at the scientific challenge 2022. This dataset is not meant to be used for developing language-based audio retrieval methods. For developing language-based audio retrieval methods, you should use the development dataset, i.e., the Clotho v2.1 dataset, which can be found also in Zenodo, at: https://zenodo.org/record/4783391. <strong>== License ==</strong> The audio files in the archives: retrieval_audio.7z and the associated meta-data in the CSV file: retrieval_audio_metadata.csv are under the corresponding licenses of Freesound [1] platform, mentioned explicitly in the CSV file for each of the audio files. That is, each audio file in the 7z archives is listed in the CSV file with the meta-data. The meta-data for each file are: File name Keywords URL for the orignal audio file Start and end samples for the excerpt that is used in the dataset Uploader/user in the Freesound platform (manufacturer) Link to the license of the file The caption queries in the file: retrieval_captions.csv are under the Tampere University license, described in the LICENSE file. <strong>==References==</strong> [1] Frederic Font, Gerard Roma, and Xavier Serra. 2013. Freesound technical demo. In Proceedings of the 21st ACM international conference on Multimedia (MM '13). ACM, New York, NY, USA, 411-412. DOI: https://doi.org/10.1145/2502081.2502245

本数据集为**DCASE 2022挑战赛**任务6（子任务B：基于语言的音频检索（Language-based Audio Retrieval））的**评测数据集**。本数据集仅用于2022年科学挑战赛任务6的子任务B相关评测，不得用于开发基于语言的音频检索方法。若需开发此类方法，请使用开发数据集Clotho v2.1，该数据集同样可在Zenodo平台获取，链接为：https://zenodo.org/record/4783391。

**== 许可证 ==**
压缩包`retrieval_audio.7z`内的音频文件，以及CSV文件`retrieval_audio_metadata.csv`中的相关元数据，均遵循Freesound平台（Freesound）的对应许可协议，每项音频文件的具体许可信息已在CSV文件中明确标注。具体而言，7z压缩包内的所有音频文件均已在CSV文件中列出并附带元数据，单条音频文件的元数据包含：文件名、关键词、原始音频文件的URL、本数据集中所采用的音频片段的起始与结束采样点、Freesound平台的上传者/使用者（版权方）、该音频文件的许可协议链接。
而`retrieval_captions.csv`文件中的描述查询文本，则遵循坦佩雷大学（Tampere University）许可协议，具体条款可在`LICENSE`文件中查阅。

**== 参考文献 ==**
[1] Frederic Font、Gerard Roma与Xavier Serra. 2013. Freesound技术演示. 收录于第21届ACM国际多媒体会议（MM '13）论文集. ACM，美国纽约州纽约市，第411-412页. DOI: https://doi.org/10.1145/2502081.2502245