Using ffaudIO (most likely), performantly generate spectrograms from audio file(s), with an option for doing so in a streaming/chunked way. Speed will be key to this; possibly compute on the GPU, and parallelize IO and the FFT.
These are in-memory spectrograms (ultimately NumPy arrays), but it would be good to have a built-in visualization function too. (But that's a rabbit hole for sure.)
Using ffaudIO (most likely), performantly generate spectrograms from audio file(s), with an option for doing so in a streaming/chunked way. Speed will be key to this; possibly compute on the GPU, and parallelize IO and the FFT.
These are in-memory spectrograms (ultimately NumPy arrays), but it would be good to have a built-in visualization function too. (But that's a rabbit hole for sure.)