Is there a "DIGITAL" equivalent of "Q"?
in [DSP]
|
Prev: Rotate 2D Gaussian
Next: signal fades in noise
From: Steve Pope on 28 Apr 2010 14:55 Tim Wescott <tim(a)seemywebsite.now> wrote: >Steve Pope wrote: >> And, even in a digital domain, it's still the same -- resonant >> frequency divided by the 3 dB bandwidth is still Q. Even if >> you're close to Fs/2, and the filter shape looks funny. >But in the Laplace domain Q is pretty easy to pick off of a 2nd-order >characteristic polynomial; this is not so easy and direct in the >z-transform world (unless you know something I don't; in which case I'd >be happy if you corrected me). As far as I know, it's not so easy. If there's a good way to describe how "resonant" a digital filter is, that would be interesting to know. Steve
From: Clay on 28 Apr 2010 15:59 On Apr 26, 3:55 pm, Richard Owlett <rowl...(a)pcnetinc.com> wrote: > Please note quotation marks in subject ;) > Also, I'm not the oldest on group --- BUT > my father operated a *LEGAL* land based spark gap xmtr > > All that to say that I think in "linear passive discrete" domain > rather than in "digital" domain. > > I have a "filter" problem. > I have a reasonable idea on how to implement it. > *UNFORTUNATELY* requires HENRY's and FARADs ;/ > I can write and solve the associated mesh equations > My solution will obviously be a subset of SPICE > > BUT will I be able to describe either > PROBLEM or SOLUTION > to those educated in digital domain? > Hello Richard, Certainly if you can describe your problem in terms of poles, zeroes, and filter order (i.e., "s" equations), a digital designer can then create an approximation to it. A common example concerns the A,B,C, or D weighting filters in acoustics. They are described using analog terms. And a few here have talked about how to go from there to digital approximations. So when you arrive at your filter, then the same approaches can be used to find digital approximations to your filter. At the following link, you will see some common "s" equations used for audio filters. http://en.wikipedia.org/wiki/A-weighting While it is true that these transfer functions are they themeselves approximations to human phenomina, a digital guy will find an approximation to them. See Al Clark's paper here for a method to make a digital filter have a transfer function that matches (within reason) a given mangitude function: http://www.compdsp.com/presentations/Clark/Magnitude%20squared%20method%20to%20solve%20a%20collection%20of%20arbitrary%20functions.pdf IHTH, Clay
From: Richard Owlett on 28 Apr 2010 16:23 Mark wrote: > On Apr 28, 5:37 am, Jerry Avins <j...(a)ieee.org> wrote: >> On 4/28/2010 3:44 AM, Magnum wrote: >> >>> "cassiope"<f...(a)u.washington.edu> wrote in message >>> news:a1905334-e7ac-4421-996b-c4ebdf636e7a(a)u31g2000yqb.googlegroups.com... >>>> Perhaps it would be useful if you would tell us how you define Q in >>>> the analog domain (there's more than one way). >>> ISTR that each way is a representation of the same formula. >> Given high-enough Q. 3 is certainly high enough for folk music. >> >> Jerry >> -- >> "I view the progress of science as ... the slow erosion of the tendency >> to dichotomize." --Barbara Smuts, U. Mich. >> ����������������������������������������������������������������������� > > in the original post the op set the context of the question as a > "filter problem" I'm trying to model a system. My initial tack was a set of filters. Not sure that will be a success. > > Often folks will say they need a high Q filter when they really mean > they need a high selectivity filter i.e. a filter with small > transition regions. I was explicitly thinking of an RLC network as a band pass filter. > > This is not really Q but more closely related to filter ORDER. But Q can define the shape of the transfer function. I completed 3 yrs of a BSEE 40 some years ago and used almost none of the network theory since. > > My guess is the op is asking about selectivity and order which is why > I tried to steer him to "taps". > > Mark >
From: glen herrmannsfeldt on 28 Apr 2010 16:34 Clay <clay(a)claysturner.com> wrote: (snip) > Certainly if you can describe your problem in terms of poles, zeroes, > and filter order (i.e., "s" equations), a digital designer can then > create an approximation to it. > A common example concerns the A,B,C, or D weighting filters in > acoustics. They are described using analog terms. And a few here have > talked about how to go from there to digital approximations. So when > you arrive at your filter, then the same approaches can be used to > find digital approximations to your filter. > At the following link, you will see some common "s" equations used for > audio filters. > http://en.wikipedia.org/wiki/A-weighting I wonder if it would be reasonable to have standard digital filters for these. That would obviously require that they be defined for specific sample rates, though. Well, maybe standard filters at 44.1kHz and 48kHz would be enough. -- glen
From: Mark on 28 Apr 2010 17:04
> > > > Often folks will say they need a high Q filter when they really mean > > they need a high selectivity filter i.e. a filter with small > > transition regions. > > I was explicitly thinking of an RLC network as a band pass filter. > > So you are talking about a one pole (one complex pole pair) filter? Mark |