Signal Conditioning

cancel
Showing results for 
Search instead for 
Did you mean: 

inverse filter

Dear All,

               

             I have attached VI with email , I have to design inverse filter of the band pass filter to recover the original signal back , Is there way to implement in the labVIEW.

 

In my VI

 

1) Input square wave frequency 15KHz

2) Filter band pass filter 20 to 20KHz 

 

 

Best Regards

Nabz

          

0 Kudos
Message 1 of 4
(5,997 Views)

Nabz,

 

Your request does not make much sense.  After the signal is filtered, components outside the filter bandpass are no longer present. Because all information about those components is lost, no possible filter can regenerate them.

 

What is it that you are really trying to do?

 

Lynn

0 Kudos
Message 2 of 4
(5,994 Views)
I know it is possible mathamaticly, if you know in filtering you have multiplied signal with some thing to get that response in inverse we have to divide the response to get orginal back. but how is labVIEW?
0 Kudos
Message 3 of 4
(5,986 Views)

You are right that it is possible mathematically, at least for linear, time-invariant filters.  I got to thinking about how it might be possible to demonstrate this in LV.

 

The Butterworth Filter Coefficients.vi (and related VIs for other types of filters) generates sets of coefficients which represent the transfer fucntion. By inverting the transfer function it should be possible to create the inverse filter. Read the help files for details.  And it works.  At least for simple filters. At some point the attenuation will get too large and numerical errors will prevent good reconstruction.  I did not try to extend this to high order filters or filters with complicated passband specifications.  While it will work with those in principle, the errors will accumulate faster.

 

With real filters noise plays much the same role as the numeric errors in the digitized filters. The Nyquist limit will also play a role for signals which have large harmonic content.

 

Lynn

0 Kudos
Message 4 of 4
(5,978 Views)