mirror of
https://github.com/KolibriOS/kolibrios.git
synced 2024-12-23 23:26:49 +03:00
29 lines
1.2 KiB
Plaintext
29 lines
1.2 KiB
Plaintext
|
This is a draft version of my Fast Hartley Transform (FHT) routine for KolibriOS
|
||
|
|
|
||
|
|
Hartley transform is a real-basis version of well-known Fourier transform:
|
||
|
|
|
||
|
|
1) basis function: cas(x) = cos(x) + sin(x);
|
||
|
|
2) forward transform: H(f) = sum(k=0..N-1) [X(k)*cas(kf/(2*pi*N))]
|
||
|
|
3) reverse transform: X(k) = 1/N * sum(f=0..N-1) [H(f)*cas(kf/(2*pi*N))]
|
||
|
|
|
||
|
|
FHT is known to be faster than most conventional fast Fourier transform (FHT) methods.
|
||
|
|
It also uses half-length arrays due to no need of imaginary data storage.
|
||
|
|
|
||
|
|
FHT can be easily converted to FFT (and back) with no loss of information.
|
||
|
|
Most of general tasks FFT used for (correlation, convolution, energy spectra, noise
|
||
|
|
filtration, differential math, phase detection ect.) may be done directly with FHT.
|
||
|
|
|
||
|
|
====================================================================================
|
||
|
|
|
||
|
|
Copyright (C) A. Jerdev 1999, 2003 and 2010.
|
||
|
|
|
||
|
|
The code can be used, changed and redistributed in any KolibriOS application
|
||
|
|
with only two limitations:
|
||
|
|
|
||
|
|
1) the author's name and copyright information cannot be deleted or changed;
|
||
|
|
|
||
|
|
2) the code is not allowed to be ported to or distributed with other operation systems.
|
||
|
|
|
||
|
|
18/09/2010
|
||
|
|
Artem Jerdev <kolibri@jerdev.co.uk>
|