How do i use a FFT to acquire a nice spectrum?

[GG_Jialat]

I have tried using different FFT spectrum on my program but i am not too sure which FFT to use as there are many different FFT in labview. I need some guidance. i wanna use the FFT to extract value from my graph so that i can filter away those unwanted value. Heres the vi i did.

[altenbach]

You did not include a datafile.

What "value" do you want to extract and what "values" are unwanted? Are you looking at filtering, tone extraction, etc.? You need to be much more specific!

Just looking at your code; What is the purpose of the lower while loop? Since the input never changes, the loop is not needed at all. It would just recalulate the same thing over and over again, consuming all CPU. Once is enough! The upper loop has some controls, that could be operated interactively. Sill, I doubt you can change them every nanosecond, so that loop needs a small wait too (e.g. 50ms).

[GG_Jialat]

I want to get the frequency spectrum of my wave in a log graph. But i am not sure which application to use in order to change my graph from time domain to frequency domain. I am not sure if u understand what i am saying but i need to get a frequency spectrum out from the waveform i have. Which FFT should i use in order to aquire the spectrum? Thanx! 

[GG_Jialat]

It looks something like this. Is it possible to do it? 

[GG_Jialat]

I tried to implenment the FFT but i still could not get the spectrum i wanted. Need help plz! >.< Here the vi i did.

[altenbach]

You still did not include a measurement file and you still don't have any delays in the loops.

[GG_Jialat]

What do u mean by measurement file and a delay?? The measurement file is the waveform itself? And why do we need a delay? Can u do a demo for me? thanx!

[altenbach]

Your first operation is reading the file "C:\Documents and Settings\sasca205\Desktop\gerald\database\100m.tdms", which we definitely don't have. You just have to believe me here. 😄

Any toplevel loop should contain a small wait, else each loop will spin millions of times per second, consuming all CPU it can possibly get. There is no need to spin the loops faster than you can enter new control values.

I alos don't understand why you need two loops. One seems to be sufficient.

[GG_Jialat]

Ops sorry! I cant send the file i send you the link. http://www.physionet.org/physiobank/database/mitdb/ Its under the signals

[GG_Jialat]

I used the biomedical kit to transform the dat file to tdms file 🙂