Hi, I have a doubt with respect to half-band FIR filters... WHile specifying weights for the remez function the response looks weird when we specify w=[0.5 0.5]. Is there any way to do it?? I thought of another way.. I found out the coefficients by using the remezord function so that the original response meets my specs. After that i stored the coeff in an array [x1] and converted all the odd coeffs. to zero and inserted 0.5 at the median of the array. Is the response plotted with x1 the half-band response??? If anyone knows more information about this or would like to correct me..please let me know. Thanks.. Vivek Venugopal |