FFTW

Hmm .... FFT was never my strong point ..... code is .... so I was rather hopping this would be a code question but since it's a maths question that's me right out of my league ... sorry Al!

Bart.

Al --

A DFT is reversible; therefore it must also encode the amplitude of each frequency component (as well as its phase). I'm nearly a complete neophyte on FFTs, so you'll have to take this with a grain of salt -- but I think your results are the nature of the beast.

-- Jeff.

Hi,

ok I was a bit off

The inputs are real alright, however the output array is complex, a+ib. To plot the output array you plot the magnitude of the value sqrt(a*a+b*b).

Now wheres that pulsar? :lol:

I was just going to post that Al!

I was just going to post that Al!

Guess the telapathy worked better

Hi,

Just to let you know that there was a point to this topic!

Say you have some data from RXTE, which consists of the exact timing (to tens of microseconds) of when photons are detected from a pulsar (plus noise, ambient light etc.). Now split all this into 'bins' of 0.01s, so we are counting the number of photons that arrive in a sequence of 0.01 seconds. The plot of that looks like this:



So we get between 0 and 9 counts per sample. Great but it doesn't even look like there is a pattern in there. Now we put that through the fft transform which will show us what frequencies are in there:



Now something looks obvious! The peak is at 38245, which means that it has 38245 cycles during out timespan, which equates to a source with a frequency of 19.814792 Hz

By looking at later observations to see how the pulsar frequency has changed we can work out the age and magnetic field of the object.

Cheers,
~Al