On Tue, Apr 5, 2011 at 3:17 PM, David Ogier
<david.ogier@xxxxxxxxx> wrote:
I think that I wasn't very clear on the previous help request.
I
know the frequency and period that I I intend to use. I do however need to
extract that 4 hours signal from the existing data set so that I can
have two data sets, one original with all it's various frequencies and
the other, only the data that is influenced by the 4 hourly
oscillation. That's why I did the FFT first to see where are areas are
that have the greatest 4 hourly wave activity, the 4 hour filtered data
should correspond roughly to this FFT data.
is this even possible?
many thanks again for your time
daveOn 4 April 2011 18:00, David Ogier
<david.ogier@xxxxxxxxx> wrote:
Hi there,
I hope you have some extra experience with Fourier analysis as I'm running out of ideas.
I have a time series (hourly for a month) and I have done a FFTA on it with the following
let FFT_temp = ffta(temp[l=1:744]) !simple enough
and I now need to isolate the 4 hourly frequency and return only it back to the hourly temperature data (ie I need 744 time steps again 24 hours * 31 days)
I've tried FFT_INVERSE but I don't know how to specify the frequency or period I need,
let sample_re = FFT_RE(temp[d=1])
let sample_im = FFT_IM(temp[d=1])
let invert_ts = FFT_INVERSE(sample_re, sample_im) ! this returns the full original data set and imposing l limits to the sample_re or sample_im seems to return wrong data
I've also been trying LSL_LOWPASS but I'm not sure on the functioning and is returning strange data that looks incorrect.
Is there something simple that I'm doing wrong?
Many thanks for your time.
Cheers
Dave