Forums

Similarity measure for signals

Started by fidaki35 August 8, 2011
 Hello,

 I'm currently working with a new parametric spectral estimation technique
(it achieves better resolution frequency than the FFT for instance) and I
would like to use it so as to find similarities between two or more signals
based on their frequency components.
 I could use the cross correlation or the correlation coefficient but by
this way I do not take advantage of the tool that I have in my disposal. Is
there any way to make a comparison of signals based on the their frequency
values?? 

Thanks,
Kostas


On Mon, 08 Aug 2011 10:45:41 -0500, "fidaki35"
<fidaki35@n_o_s_p_a_m.hotmail.com> wrote:

> Hello, > > I'm currently working with a new parametric spectral estimation technique >(it achieves better resolution frequency than the FFT for instance) and I >would like to use it so as to find similarities between two or more signals >based on their frequency components. > I could use the cross correlation or the correlation coefficient but by >this way I do not take advantage of the tool that I have in my disposal. Is >there any way to make a comparison of signals based on the their frequency >values?? > >Thanks, >Kostas
Is there a reason not to use correlation in the frequency domain? Eric Jacobsen http://www.ericjacobsen.org http://www.dsprelated.com/blogs-1//Eric_Jacobsen.php

fidaki35 wrote:

> Hello, > > I'm currently working with a new parametric spectral estimation technique > (it achieves better resolution frequency than the FFT for instance) and I > would like to use it so as to find similarities between two or more signals > based on their frequency components. > I could use the cross correlation or the correlation coefficient but by > this way I do not take advantage of the tool that I have in my disposal. Is > there any way to make a comparison of signals based on the their frequency > values?? > > Thanks, > Kostas > >