what is decimation filter?
IN discrete signals we have discret samples at discrete intervals. when everr we sample a analogue/continuous signal to make a discrete signal we follow nyquist criterion. some time we oversample to make sure therer is no aliasing. now when we have digital samples we may have huge number of samples and want to reduce them. in decimation by 2 we throw away every 2nd sample. in freq domain decimation leads to spread in spectrum by factor which is actually same as the amount of decimatin. e.g if a digital signal has a highest freq of 1 Mhz when we decimate by 2 it will become 2 Mhz. so there may be a case where there was no aliasing of orig signals but when we decimate they aliase. to cater for this we can first passs the signal thru a pi/M filter and then decimate by M to ensure there is no aliasing although in the process we may loose some high freq components of the original signal. e.g if it is an audio loosing some high freq comp may not even be noticeable to human ear but if we let aliasing to happen the signal may even not be recoverable.
plz press helped me if this was of some help