Multi-City dataset

该数据集名为“多城市”，包含了来自10个不同声学场景类别的40小时音频，这些音频是在10个欧洲城市的不同地点录制的。该数据集旨在评估声学场景分类（ASC）在已知和未知城市环境下的性能。此外，数据集还包括从设备A录制的音频，并已转换为单声道音频。训练集包含7242个音频记录，测试集则包含7158个音频记录。整个数据集的规模为40小时，所涉及的任务是声学场景分类。

This dataset is named "Multi-City". It contains 40 hours of audio spanning 10 distinct acoustic scene categories, recorded at various locations across 10 European cities. It is designed to evaluate the performance of Acoustic Scene Classification (ASC) in both known and unknown urban environments. Additionally, the audio recordings from Device A included in this dataset have been converted to mono audio. The training set consists of 7,242 audio recordings, while the test set contains 7,158 audio recordings. The total duration of the entire dataset is 40 hours, and the targeted task is acoustic scene classification.