FIR Cookbook?
in [DSP]
|
From: Rune Allnor on 12 Dec 2006 10:30 jeff227 skrev: > >Ron N. > >You should explore other window > >functions (von Hann, Blackman, etc.). > > > Ron, yes I understand there are window influences and other areas to > tweak. Thank you for the heads up. > > Now, what I don't understand is your use of BW and Fc in the code above - > it seems to have a different meaning than what I have seen elsewhere. Is > "BW" the transition band or something else? In other Window FIRs I have > have worked with the "cutoff" (-6dB) point has been at Fc + 1/2Tr. This > doesn't seem to be the case with the above code. The cutoff point I > observed in a gain plot is at Fc + Tr. So I'm confused as to what "Fc" > and "BW" mean in this code. .... and I am a bit confused about what Tr means. Defining the acronyms might help others to understand. > I was under the impression (probably wrong) that the rolloff response of a > "window based" FIR was determined soley by the window selection and the > number of taps. The term "rolloff" does not have a well with window FIRs. The main design factors are the transition band between passband and stop- band corner frequencies, and the corner frequency of the passband. The width of the transition band determines the number of coefficients in the filter, the corner frequency determines certain computational factors in designing the filter. > In your code there are 3 variables - BW, Fc and Window. > Does this mean passband ripple is adjustable or ???? It means that except for the Kaiser window, the user has no contol of stop-band attenuation or passband ripple. Choose a window function at your own peril; it is you as designer, who is responsible for checking that the filter attenuates the stopband sufficiently well. Rune
From: Rune Allnor on 12 Dec 2006 10:32 Typo to be corrected: Rune Allnor skrev: > The term "rolloff" does not have a well with window FIRs. Should be "Does not have a well-defined meaning when used with wiundow FIRs". Rune
From: jeff227 on 12 Dec 2006 11:04 ... and I am a bit confused about what Tr means. Defining the acronyms >might help others to understand. Tr = "transition band". In a "Window" type FIR the -6dB point is usually the midpoint of this region. Fc is where the rolloff begins. So the "cutoff" point for a LPF is Fc + 1/2Tr. The "BW" variable in Ron's code does not behave in this way so I am wondering what it means. >> In your code there are 3 variables - BW, Fc and Window. >> Does this mean passband ripple is adjustable or ???? > >It means that except for the Kaiser window, the user has no contol >of stop-band attenuation or passband ripple. Choose a window >function at your own peril; it is you as designer, who is responsible >for checking that the filter attenuates the stopband sufficiently well. Agree. I understand the window choice determines the transition band shape (trading off steepness for stopband attenuation). So how is it possible to define BW AND the Window? That was my question - what does the variable "BW" mean in Ron's code above? It does not appear to be the transition bandwidth.
From: Jerry Avins on 12 Dec 2006 12:04 jeff227 wrote: ... > So how is it possible to define BW AND the Window? By choosing the number of taps to meet the spec. ... Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
From: jeff227 on 12 Dec 2006 13:15
>> So how is it possible to define BW AND the Window? > >By choosing the number of taps to meet the spec. Yes, I know but please read the previous posts - nTaps, Fc, BW, & Window are ALL input variables to this "Windowed Sinc" coefficient generator. My question was what BW does in this routine since "Fc", "nTaps" and "Window" normally *define* BW. |