It would be nice to be able to use this fast and low memory usage library for larger number of samples.