Rate of those recordings (feature = "smp", 0 to 5, histogram)
Recording per year (histogram)
Duration (histogram)
Frequency (44100 vs. 48000 Hz)
Song / call
Visualisation:
Mapping in Brazil (done by @Tfcosendey already) - see if we can improve some elements. Ex: using different colors to map different generics
Explain to general audience how to visualise an audio file (using librosa and Pytorch)
Oscillogram (time x amplitude (decibels)): "time-domain representation of an audio signal, which means that it shows how the amplitude of the signal changes over time...."
Spectrogram (time x frequency): "a spectrogram is a frequency-domain representation of an audio signal"
Mel Spectrogram: "similar to a regular spectrogram, but it uses a frequency scale. They capture the essential features of the audio and are often the most suitable way to input audio data into deep learning models"
--> Show 1x Mel spec per bird category.
Basic figures about the dataset:
Charts / Graphs / Histogram:
Visualisation:
Explain to general audience how to visualise an audio file (using librosa and Pytorch)