Two implements for Constant Q Transform
SciPy(for sparse matrix representation) and NumPy are required.
fmin: minimum frequecy;fmax: maximum frequecy;bins: number of bins;fs: the rate of sample;wnd: the window function;
The window function wnd receives a parameter N representing the bandwith of window and returns a numpy.array representing the corresponding weight for each position.
Question: why does it come out ValueError: matrices are not aligned?
Answer: the length of sample is small. Padding some 0's to the end of the sample can fix this.