| Re: FFT-bug???? (PV-Wave) [message #10344] |
Wed, 19 November 1997 00:00 |
steinhh
Messages: 260
Registered: June 1994
|
Senior Member |
|
|
In article <879843126.431506@curry.top.tld>
Franz.Dollinger@mchp.siemens.de (Franz Dollinger) writes:
> The absolute value is identical for Fortran and PV-Wave, but the
> phase of the PV-Wave output is wrong.
This may simply be a different convention as to what is regarded
as the "central pixel" in the (periodic!) pattern that you are
FFT'ing.
E.g., given an array a = fltarr(128), IDL will consider the a(0)
as the "central" pixel (the pixel at the "origin") - thus a
discrete "delta function" would be represented by
a = fltarr(128)
a(0) = 1.0
Analogously, the zero frequency component of an FFT (in IDL) is
stored in a(0).
I suspect that some libraries implementing FFT has diverging
definitions - e.g., regarding a(63) as the central element (but
still storing the zero *frequency* component in a(0) - ??)
Depending on the exact definition, such a difference may cause
only the phase of a transform to be wrong (since the original
data seems to have been shifted).
Regards,
Stein Vidar
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
