I believe that in FIR there are too many poles placed at z=0. and are of higher order. ie large number of ploes are at same location. thats the reason that they are difficult to implement.
May be I should read some on FIR filters, but to me; if the poles are at zero then that means there are no poles to implement lets say you have a filter like : 1+2z-²+5z-³. That is a filter where the poles are all at zero, init? So its easy to implement; Take a couple of delay elements, dont forget to take the weights into account and thats it.
I may not be right because I havent read anything about filters, so could you tell me a reference where it says the FIR with all poles at zero is not easy to implement?
The only difficult thing with FIR filters is the availability of required resources to my opinion. Apart from that, they are easy to implement. The minimum order depends on intended filter characteristics and relative sampling rate fc/fs as well. Thus a simple FIR filter may require a high order for fs >> fc.
The most important thing is you gain phase linearity in FIR filters, at the cost of larger degree. Besides, they're so easily implemented with Multiply and Accumulate blocks and a Tapped Delay Line. They are feedforward and that makes their implementation so easy, and also stable.
bhq,
See the article "A New Estimation Formula for Minimum Filter length of Optimum FIR Digital Filters".
ieeexplore.ieee.org/iel4/5224/14216/00652198.pdf
You must be an IEEE member to download this article.
Regards,
Kral