We need to figure out how to compress our audio analysis data into numbers that can produce a clear correspondence between visualization and audio. This is a fairly difficult problem, and very crucial to the app's usefulness, but it should be doable with measures like spectral centroids.
We need to figure out how to compress our audio analysis data into numbers that can produce a clear correspondence between visualization and audio. This is a fairly difficult problem, and very crucial to the app's usefulness, but it should be doable with measures like spectral centroids.