I think that poor precision is a problem. I have checked now and
std.numeric.Fft does indeed use floats for the lookup table.
The precision problem could be solved by either changing that to
real or by changing Fft to a template and using whatever the
type
parameter is. Just changing float to real doesn't require
changing
the API, but would probably result in worse performance.
I forgot to mention one more solution (the one that pfft.stdapi
currently uses). The class could also use lazy initialization.
When the fft is called, check if the lookup table for that type
is already created, create it if it isn't, store it for later and
then compute the fft. This would require casting away const in
the fft() method, though.
