>>>>> "Wolfgang" == Wolfgang Jenkner <wjenkner@inode.at> writes:
Wolfgang> Raymond Toy <toy@rtp.ericsson.se> writes:
>> I've been unable to create
>> an array that is acceptable to the fft routines. I tried
>> array(ar,4), arraymake(ar,[4]), and make_array. The resulting arrays
>> are rejected by fft.
Wolfgang> Apparently, different people at different times had different ideas
Wolfgang> of what a macsyma array should look like. I think this is the present
Wolfgang> situation (I wrote this up for another patch I haven't sent yet, sorry
Wolfgang> for the formal tone of it):
[snip]
Wolfgang> Here is a rough outline of a possible solution:
[snip]
Thanks for the code. I've added this to fft.lisp with my FFT routines
and for the simple test cases I've tried, this produces the desired
results. Only 1D arrays and transforms are supported and only if
the length is a power of 2. Could add a brute force DFT for the other
cases, but I think I'll leave that for now.
I do, however, think that these routines shouldn't modify their
inputs, but I think I'll leave them as is for now.
I'll check in a version shortly.
Ray