need help on FFT

[fenny]

IS there anybody who can help me to verify a FFT result?
I 'm translating a program from ASYST to labview.

I got result different from the old program. could you please make a small labview program to do FFT to the data i attached and see what is the FFT outcome? I want to verify my result. thanks!

[djadjok]

Hmm... Not an answer, just some thoughts...
Are you saw already the NI FFT challenge discussions? There are a different readings for FFT result normalisation for different math packages. May this can help you...

[fenny]

After reading the NI FFT CHALLENGE discussions, I'm more doubtful about": if the algorithms of the labview and the ASYST are the same". because i got normal result when i perform ":fft(vv1)/fft(pp1)(the difference canceled out in this case), but different result when i perform only one fft(vv1) or fft(PP1).

[fenny]

if the algorithms are different (as dicussed in NI FFT CHALLENGE, THERE IS A COEFFIENT DIFFERENCE), can we make changes to the c-code of FFT of LABVIEW to achieve the exact FFT?

[Scott Seidman]

fenny wrote in news:506500000005000000DC010100-
1042324653000@exchange.ni.com:

> After reading the NI FFT CHALLENGE discussions, I'm more doubtful
> about": if the algorithms of the labview and the ASYST are the same".
> because i got normal result when i perform ":fft(vv1)/fft(pp1)(the
> difference canceled out in this case), but different result when i
> perform only one fft(vv1) or fft(PP1).

Wow, I haven't heard of anyone using ASYST in years.

I doubt that the algorithms are the same. If I recall, ASYST was already
considered a dead app by the time non radix 2 algorithms were becoming
popular. Also, Asyst Software Technologies was purchased by Keithley, an
NI competitor!

--
Scott
Reverse first field of address to repl
y

[djadjok]

I regret really, but I have no any experience working with ASYST and also I must admit that I'm not at all very familar with FFT algorithms, that's why I've only try to give you an advice as a reference to one of the NI coding challenges.
But, anywhere, there is no possibility to change existing NI FFT c-code. At the same time, I assume that you can apply changes to the resulting FFT data after the transform (i.e.renormalize it, may be I'm wrong but it seems to me that it is possible to do this at this point). If you prefer, you can also try to implement your own FFT code (preferably also using c, as if you'll do this in labview, this will work much slower), but it doesn't look like a necessary work.
Sorry for my poor english

[fenny]

yes ASYST is dead, that's why I 'm translating the program from ASYST to LABVIEW. my problem is complicated because i don't even have the complete manuls of ASYST. I have to guess some of it's functions.(AND I have only half of the mathematics of the program) DO you happen to know what" smooth " is doing in ASYST? it was used to some frequency domain signal.