first attempt of LTAS (long term average spectro) creation , taking profit from the spectrogram computation , for each wav we also save a reduced spectro (for the moment , keeping only one welch per file), ie
Sxx=Sxx.mean(axis=1),
and save them as npz in the folder joinpath(OSMOSE_PATH.LTAS,"intermediate_features") , then you can use the method build_LTAS (future class ? ) to aggregate them over different time resolutions
first attempt of LTAS (long term average spectro) creation , taking profit from the spectrogram computation , for each wav we also save a reduced spectro (for the moment , keeping only one welch per file), ie
Sxx=Sxx.mean(axis=1),
and save them as npz in the folder joinpath(OSMOSE_PATH.LTAS,"intermediate_features") , then you can use the method build_LTAS (future class ? ) to aggregate them over different time resolutions