Hello, I have a question that why can't some IIR filter design methods like Impulse Invariant & Approximation of derivatives be used (or have poor response) as band-pass, notch and high-pass filters I think the reason for approximation of derivatives is that we can't go beyond fs/4 on the Z-plane, so high-pass is out. But what is the issue with band-stop and band-pass for a small pass-band? Can some one explain? Robert, Glen?
Why some IIR filter methods can't be used for High Pass Filters?
Started by ●October 16, 2009
Reply by ●October 16, 20092009-10-16
commengr wrote:> Hello, > > I have a question that why can't some IIR filter design methods like > Impulse Invariant & Approximation of derivatives be used (or have poor > response) as band-pass, notch and high-pass filtersMethods do work. It is just you who doesn't know how to use them properly. VLV
Reply by ●October 16, 20092009-10-16
On Oct 16, 9:19=A0am, "commengr" <communications_engin...@yahoo.com> wrote:> Hello, > > I have a question that why can't some IIR filter design methods like > Impulse Invariant & Approximation of derivatives be used (or have poor > response) as band-pass, notch and high-pass filters > > I think the reason for approximation of derivatives is that we can't go > beyond fs/4 on the Z-plane, so high-pass is out. But what is the issue wi=th> band-stop and band-pass for a small pass-band? > > Can some one explain? Robert, Glen?Aliasing is your problem if you blindly try to sample the impulse response of a highpas filter. Same issue with derivatives.
Reply by ●October 16, 20092009-10-16
On Oct 16, 2:01=A0pm, Clay <c...@claysturner.com> wrote:> On Oct 16, 9:19=A0am, "commengr" <communications_engin...@yahoo.com> > wrote: > > > Hello, > > > I have a question that why can't some IIR filter design methods like > > Impulse Invariant & Approximation of derivatives be used (or have poor > > response) as band-pass, notch and high-pass filters > > > I think the reason for approximation of derivatives is that we can't go > > beyond fs/4 on the Z-plane, so high-pass is out. But what is the issue =with> > band-stop and band-pass for a small pass-band? > > > Can some one explain? Robert, Glen? > > Aliasing is your problem if you blindly try to sample the impulse > response of a highpas filter. Same issue with derivatives.what's being aliased is not the spectrum of your signal. an IIR filter designed by either method is still a legit LTI discrete-time system. but the frequency response that corresponds to the continuous- time impulse response that you've sampled will be an aliased version of the continuous-time frequency response. r b-j
Reply by ●October 17, 20092009-10-17
> > >commengr wrote: > >> Hello, >> >> I have a question that why can't some IIR filter design methods like >> Impulse Invariant & Approximation of derivatives be used (or have poor >> response) as band-pass, notch and high-pass filters > >Methods do work. It is just you who doesn't know how to use themproperly.> >VLV >Look Vladimir, design a high pass digital IIR filter using the Approximation of derivatives method and we can have experts here as critics
Reply by ●October 17, 20092009-10-17
>On Oct 16, 2:01=A0pm, Clay <c...@claysturner.com> wrote: >> On Oct 16, 9:19=A0am, "commengr" <communications_engin...@yahoo.com> >> wrote: >> >> > Hello, >> >> > I have a question that why can't some IIR filter design methods like >> > Impulse Invariant & Approximation of derivatives be used (or havepoor>> > response) as band-pass, notch and high-pass filters >> >> > I think the reason for approximation of derivatives is that we can'tgo>> > beyond fs/4 on the Z-plane, so high-pass is out. But what is theissue =>with >> > band-stop and band-pass for a small pass-band? >> >> > Can some one explain? Robert, Glen? >> >> Aliasing is your problem if you blindly try to sample the impulse >> response of a highpas filter. Same issue with derivatives. > >what's being aliased is not the spectrum of your signal. an IIR >filter designed by either method is still a legit LTI discrete-time >system. but the frequency response that corresponds to the continuous- >time impulse response that you've sampled will be an aliased version >of the continuous-time frequency response. >Effectively meaning that we lose the attenuation of the filter where we have aliasing in frequency domain?>r b-j >Syed
Reply by ●October 18, 20092009-10-18
> > Methods do work. It is just you who doesn't know how to use >them > properly.Omg, you are so arrogant....
Reply by ●October 18, 20092009-10-18
>> >> Methods do work. It is just you who doesn't know how to use >them >> properly. > > >Omg, you are so arrogant.... > >I think Vladimir is a wanna be Simon Cowell... of compdsp
Reply by ●October 18, 20092009-10-18
commengr wrote:>>> Methods do work. It is just you who doesn't know how to use >them >>> properly. >> >> Omg, you are so arrogant.... >> >> > > I think Vladimir is a wanna be Simon Cowell... of compdspNevertheless, Vladimir is correct. The method works when it's properly applied. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 18, 20092009-10-18
On Sun, 18 Oct 2009 09:35:33 +0200, Jake wrote:>> Methods do work. It is just you who doesn't know how to use >them >> properly. > > > Omg, you are so arrogant....He is. He's also very often right, so if you can excuse his arrogance long enough to glean value from what he says it's worth your while. -- www.wescottdesign.com
