What a footstep sounds like to a floor

Interactive page: geophone footstep recordings you can see, zoom into, view as a Morlet wavelet scalogram or spectrogram, and listen to.

1. Arrange the recordings

data/environment/e1/   one .mat file   (same walker on floor 1)
data/environment/e2/   one .mat file   (same walker on floor 2)
data/environment/e3/   one .mat file   (same walker on floor 3)
data/gender/male/      one .mat file   (same floor as female)
data/gender/female/    one .mat file

If a folder has several files, the first one alphabetically is used. Open prepare_data.py and edit GROUPS to choose a file, rename the floors (e.g. “Wooden floor”), or change how many seconds are kept (MAX_SECONDS).

2. Build the data

pip install numpy scipy
python prepare_data.py

This writes assets/environment.js and assets/gender.js.

3. Put it online (GitHub Pages)

Upload index.html, assets/, .gitignore and prepare_data.py. Do not upload data/: .gitignore already keeps the raw recordings out of git. Then Settings → Pages → Deploy from a branch → main → / (root).

Presenting

Space plays all, 1–5 plays one recording, B toggles blind listening. index.html also works offline straight from a USB stick.