FPGA Central

Hai! every one.
This question is basic for DSP experts.I try to design digital butte
worth filter(lowpass) in C.First I converted given digital specs to analo
specs.after that i caluculated prewarped analog specs.I found order, cutof
frequency of analog butterworth filter.After that I used butterworth filte
tables to find normalized denominator coefficients of analog filter.Now
tried to convert these analog filter coefficients to digital filte
coefficients, using bilinear transfermation.I know that we need to replac
's' with 2/T*(z-1)/(z+1) to get equivalent digital filter.
Here I have two problems.
i>It's difficult to write C code, to get polynomial coefficients in
domain.So instead of like this,is there any other way of getting digita
filter coefficients?(using poles....etc)
ii>In analog filters, I used butterworth filter tables, instead of tha
can we caluculate directly protype denominator coefficients(cutof
frequency=1)?
I don't have depth knowledge of DSP concepts.If any body knows answers t
my questions, please let me know.
Thank you.
vyshali

vyshali wrote:
> Hai! every one.
> This question is basic for DSP experts.I try to design digital butter
> worth filter(lowpass) in C.First I converted given digital specs to analog
> specs.after that i caluculated prewarped analog specs.I found order, cutoff
> frequency of analog butterworth filter.After that I used butterworth filter
> tables to find normalized denominator coefficients of analog filter.Now i
> tried to convert these analog filter coefficients to digital filter
> coefficients, using bilinear transfermation.I know that we need to replace
> 's' with 2/T*(z-1)/(z+1) to get equivalent digital filter.
> Here I have two problems.
> i>It's difficult to write C code, to get polynomial coefficients in z
> domain.So instead of like this,is there any other way of getting digital
> filter coefficients?(using poles....etc)

Divide the filter into a casrcade of biquads, and transforme each
biquad analytically.

> ii>In analog filters, I used butterworth filter tables, instead of that
> can we caluculate directly protype denominator coefficients(cutoff
> frequency=1)?

Yes. There is the book by Antoniou that takes you through all
the steps.

> I don't have depth knowledge of DSP concepts.If any body knows answers to
> my questions, please let me know.

You have done a lot already. Find the book by Antoniou and you'll
reach your goal.

Rune