Identification of Beehive Piping Audio Signals

Citation Author(s):
Agnieszka
Orlowska
IBISC - Univ. Evry/Paris-Saclay
Dominique
Fourer
IBISC - Univ. Evry/Paris-Saclay
Submitted by:
Dominique Fourer
Last updated:
Mon, 10/04/2021 - 08:16
DOI:
10.21227/53mq-g936
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

We introduce a novel dataset of bee piping audio signals which was built by collecting 44 different recordings which were published by various beekeepers on the YouTube platform.Each recording has a duration varying from 2 to 13 seconds and is annotated according to the beekeeper comment respectively as Tooting or Quacking.We extracted the audio using ``YouTube soundtrack extraction'' from 14 distinct videos from which the signal is stored without a loss of quality into a WAVE file with a sampling frequency of F_s=22.05 kHz and a sample precision of 16 bits.After removing the silent frames, the resulting dataset contains 36 tooting and 8 quacking signals which correspond to a duration of 145 seconds for tooting and 60 seconds for quacking (total 205 seconds).For copyright reasons, we only made publicly available the short-time Fourier transforms matrices and the timbre descriptors computed using a matlab implementation of the timbre toolbox proposed by Peeters et al. in 2011.A more detailed description of the dataset containing the links of the original youtube videos can be found

Instructions: 

The files can be loaded in matlab or octave.

The stft.mat files contain the STFT of each signal stored as a complex-valued matrix Sw.

The ttb.mat files contain the 164 timbre coefficients computed stored as a vector ttb_vec (the feature names are stored in the feature_name variable).

The matlab scripts used to generate the dataset from the original wav files are provided in the piping_mfiles.zip archive.