Hi everyone! I write here, after having googled for some days, because I'm looking for some advices. I'm working, for my thesi, on a lowpass FIR filter for its implementation in a Altera FPGA (my tutor is on holiday and I haven't idea of exact model of this, but I think it is a very good board). Now, I choose to design an Interpolated filter, with a cascade of 2 FIR with symmetric transposed structure. I've discard direct form structure because the filter is long (about 100taps) and I need a clock frequency of about 50MHz. I know there are others structures, like second order cells, or bufferized direct form, but I think transposed choice is better for area occupation and pipeline improvements. Is it correct?
Then, the filter is a base band filter, with 2 input signals, real part and imaginary part. I think to develop the filter for real part and replicate it for the imaginary part. Are there better ways to do this?
And, last but not least, I would use RADn or other algorithm for avoiding multiplicators, because the filter is with constant coefficients. Did some of you use any of these algorithms?
Thank you very much to all the people to answer this message. Best regards Dario
PS: I'm sorry for my written English...