Re: "Discrete Hartley transform" vs " Discrete Fouuier transform"

On Oct 17, 8:21=A0pm, Eric Jacobsen <eric.jacob...@[EMAIL PROTECTED]
> wrote:
> I understand this. =A0You seem to be arguing that because a bunch of
> papers came to some conclusion that that conclusion must be true for
> everyone, everywhere.

No, I'm arguing that your claims of "substantial" advantages have no
sup****t, and your examples of supposed advantages (lower memory for
trig tables, more sequential access, reduced memory reordering) were
demonstrably false.  It is impossible to prove a negative, but since
you are the one claiming substantial advantages, the burden of proof
is on you, especially since people have tried and failed to
convincingly demonstrate such advantages for 20 years.

> Given that the two transforms are different, it would seem pretty
> logical that the differences would lead to advantages one way or other
> depending on different cir***stances.

If you actually look in detail at the algorithms, they are in fact
extremely similar. This is why people who have studied the problem in
detail have concluded that there don't seem to be significant
advantages one way or the other.  You have yet to correctly identify
one significant difference.

I would be very happy to learn of a way to use the DHT to gain a
substantial advantage over real-data FFTs.  FFTs of real data are
annoying asymmetric transforms and the diversity of data formats for
their redundant complex outputs are a PITA.  But unsup****ted claims of
substantial computational or memory savings, which have proved false
again and again for two decades, are not helpful.

Regards,
Steven G. Johnson