When-in-Rome
When-in-Rome copied to clipboard
Reimplement pitch profiles (AKA distributions) with np.arrays
Operations on numpy arrays are much faster and also the code is much more compact because it avoids explicit for loops. Swapping is not a trivial task because it would require carefully examining every place where pitch distributions are used and making the conversion. A first step could be to implement functions to input and output lists, but internally use np.arrays.
A first leg of this work is proposed in #73