If order of Comb filter is equal to 2, length of impulse response of six stages connected in cascade is 85 then how can i define the length of overall second order six stage Comb filter using a simple formula in MATLAB ?
Is it will be 2*85-1 =169 or else?
If N is filter order it has N+1 is the length of FIR filter. Thus, if order is 2, length is 3, and the lengtho of impuse response of 6 stages is 3 x 6 = 18.
Thanks for reply
Sir actually I am using Comb filter for decimation factor 15 which have transfer function as[ (1+(z^-1)+(z^-2)+........(z^-15))]/(15^6)
After connecting 6 stages of comb filters in cascade using Convolution in MATLAB length of 6 stages of impulse response is 85.
As for order =2 how can i chose length of resultant filter?
In this case am i correct to chose filter length equal to order*length of 6 stages of impulse response
Hi!
I believe that you missintrerpret the order of a filter. The order of a filter with transfer function (1+(z^-1)+(z^-2)+........(z^-15))]/(15^6)
H(z)=(1+(z^-1)+(z^-2)+........(z^-15))]/(15^6
is 15 and its length is 16. In the above transfer function the constant
1/(15^6) = 8.77 * 10^(-8)
is very questinable, as it is way to small, and after six stages would becomoe
4.57 * 10-(43)
On the other hand, with a constant 1/16 instead of 1/(15^6), your transfer function represent moving average and not a comb filter. The transfer function of a comb filter of 15th order would be:
H(z) = 1 + alpha * z^(-15) (where alpha could also be 1, as I beleive is in your case)
As the order of the filter is 15, the length of impulse response of one stage is 15+1
and the length of impulse responese (and also filter length) of 6 stages is 15*6 +1 = 91 (order * no_stages +1)
Sir,
I understood your concept in which I was little bit confused and got proper length of impulse response.
Can i adjust the filter length according to my design to get better results?As i am applying optimization technique on [1 0 0 0] and getting optimized values. When out of four optimized values i have selected only two for using them as input in the formula of calculating FIR filter output.
Thus by choosing (2*length of impulse response -1 )as filter length i am able to get better results but little bit confused in choosing filter length in this way .
Can I chose length in this way?
You should explain what you are doing. What you are optimizing? What otimization technique you are using? What are better results?
In summarry: What do you want to achieve? What is your problem statement?
- Badges
- Science topic