GTZAN音乐流派数据集是一个包含1000个各为30秒的音频轨道的集合，共包含10种音乐流派，每种流派有100首曲目。这些曲目都是单声道、16位、22050Hz采样率的WAV格式音频文件。该数据集由George Tzanetakis和Perry Cook于2002年创建，用于自动音乐流派分类研究，并在音乐信息检索领域被广泛用作标准基准数据集。

Hugging Face2025-09-13 更新410

Audio Classifier Dataset

音频分类

声音场景识别

Diverse Soundscape: Train Your Model with Real-World Audio Clips

kaggle2024-03-22 更新110

bird1401_pair6_son2_all_songs

动物声音识别

生物声学

sound file

Dryad2013-01-01 更新40

Maqam478: Qur'anic Recitations in 8 different Maqams

音频分类

阿拉伯音乐

A dataset conatining 478 Qur'anic audio recitations in 8 different maqams . The dataset is intended to be used for the purpose of maqam classification.

DataCite Commons2021-02-07 更新110

VTL Calculator from Dogs perceive and spontaneously normalize formant-related speaker and vowel differences in human speech sounds

动物声音识别

声道长度估计

This vocal tract length calculator is used for estimating the vocal tract length of speakers from h-vowel-d words.

NIAID Data Ecosystem80