Dimitris Papavasiliou writes: > 3) Well I suppose there's a good reason for this but the halfcomplex > arrays produced by radix2 and mixed-radix routines are different.
The Radix-2 ffts are designed to operate inplace without using any extra memory, so the layout is a consequence of that constraint and the algorithm. If memory is not an issue it is fine to use the Mixed Radix routines for size N=2^k to get the same layout. There is no requirement to use radix 2 routines in that case. -- Brian Gough _______________________________________________ Help-gsl mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-gsl
