Paul wavelet problems

[paulthebaker]

There may be a bug in the Paul wavelet implementation. The ICWT reconstructions have problems for particular choices of m (e.g. m = 6, 7, 10).

This could also be a problem with the ICWT implementation.

[paulthebaker]

This will be put on hold until issue #22 is sorted.

In the meantime, one should only use the Morlet-Gabor wavelets and take extreme care in defining one's own basis.