Re: Envelope Detection


"HelpmaBoab" <FU2@xxxxxxxxxxxx> wrote in message
news:KNIWf.9563$JZ1.355822@xxxxxxxxxxxxxxxxxx
I have read that to get the envelope of a signal we can take its Hilbert
TF
and use this as the imaginary part of a complex array (the real part being
the original signal). Then the envelope is sqrt(re^2 + im^2). I imagine
this
is just creating an I and Q. How accurate is this method and can we use an
FFT (s) to do the Hilbert transform?


Tam



What do you mean by envelope detection? AM demodulation?

If you have baseband I,Q data you can just run it through a cordic function
(or do sqrt(I^2+Q^2)).

The hilbert transform seperates the positive frequencies from the negative
frequencies in a complex signal. I don't see how it would apply to envelope
detection?

-Clark


.

Relevant Pages

  • Re: Envelope Detection
    ... FFT to do the Hilbert transform? ... Tam ... The hilbert transform seperates the positive frequencies from the negative ... When you inverse transform the positive frequencies the original real signal s, ...
    (comp.dsp)
  • Re: Alternative to Hilbert Transform
    ... when taking an FFT, a better approach to generate a Hilbert Transform ... arbitrary phase plus another which is the original signal plus the ... >effective bandwidth of the frequency (i.e. velocity) measurement derived ...
    (comp.dsp)
  • Re: Ways of implementing a Hilbert Transform
    ... john wrote: ... >> I want to implement a Hilbert transform, but i'd like to avoid having ... >> negative frequencies, which is not such an easy concept to master) ... > What I do is take a real FFT of the taps, ...
    (comp.dsp)
  • Re: Ways of implementing a Hilbert Transform
    ... >I want to implement a Hilbert transform, but i'd like to avoid having ... >to deal with a kernel to convolute with, ... One way to generate an analytic signal is to perform the FFT of your ... The result of the inverse FFT will be your desired ...
    (comp.dsp)
  • Re: Ways of implementing a Hilbert Transform
    ... > I want to implement a Hilbert transform, but i'd like to avoid having ... > to deal with a kernel to convolute with, ... > negative frequencies, which is not such an easy concept to master) ... What I do is take a real FFT of the taps, ...
    (comp.dsp)