SNU-B36-50E: an inter-floor noise dataset

SNU-B36-50 is an inter-floor noise dataset collected in building 36 at Seoul National University.

It is designed for evaluation of an inter-floor noise type/position classifier.

The dataset is a part of the conference paper:

https://ieeexplore.ieee.org/abstract/document/8521392

And a convolutional neural networks-based inter-floor noise type/position classifier is evaluated with the expanded version (SNU-B36-50E) in a paper titled "Classification of inter-floor noise type/position via convolutional neural network-based supervised learning” (submitted to Applied Sciences).

The inter-floor noises included in the dataset can be classified into 5 noise types and 19 positions.

The noise types are

- A medicine ball falling on the floor from a height of 1.2 m (MB)

- A Hammer dropped from 1.2m above the floor (HD)

- Hammering (HH)

- Chair dragging (CD)

- Running a vacuum cleaner (VC) 

And the positions ([floor][distance from the origin along the X-coordinate]) are

- 1F0m, 1F6m, 1F12m

- 2F0m, 2F6m, 2F12m

- 3F0m, 3F1m, 3F2m, 3F3m, 3F4m, 3F5m, 3F6m, 3F7m, 3F8m, 3F9m, 3F10m, 3F11m, 3F12m

A smartphone microphone was used as a receiver to record the inter-floor noises.

The noises were sampled at 44,100Hz for ~5 s.