Instantaneous Frequency with I-Q demodulation

On Mar 5, 7:06&#4294967295;am, "freedreamer" <freedrea...@email.it> wrote:
> I have a few questions, maybe you can help me. I have 2 signal I,Q @ 40khz
> and the task is to obtain the instantaneous frequency. What i did in matlab
> is this: first of all i found the phase with atan2, i unwrapped the phase
> and I derivated in order to have the frequency (with command diff).
>
> A first issue is here: I had to average the phase to see somenthings in
> frequency 'cause otherwise seems very noisy the istantaneous freq ( a lot a
> spurious peaks). Is there a method to optimize this average or to avoid?
> (now i use a very long moving average filter)

You didn't include the bandwidth of the data in your samples.
If it's not low-pass filtered sufficiently, the phase change of
any peaks in the noise spectrum could overwhelm the phase change
due to your signal.  A properly designed FIR filter would
probably work better than the rectangular window filter
implicit in your moving average.  An interpolating filter
might be even better if you really wanted "instantaneous"
frequency on a continuum.


--
IMHO. YMMV.
rhn A.T nicholson d.0.t C-o-M
 http://www.nicholson.com/rhn/dsp.html