<< Chapter < Page | Chapter >> Page > |
Previously, polyphase interpolation and decimation were derived from the Noble identities and motivated for reasons of computationalefficiency. Here we present a different interpretation of the (ideal) polyphase filter.
Assume that is an ideal lowpass filter with gain , cutoff , and constant group delay of :
Recall that the polyphase filters are defined as
In other words, is an advanced (by samples) and downsampled (by factor ) version of (see ).
The DTFT of the polyphase filter impulse response is then
Thus, the ideal polyphase filter has a constant magnitude response of one and a constant group delay of samples. The implication is that if the input to the polyphase filter is the unaliased -sampled representation , then the output of the filter would be the unaliased -sampled representation (see ).
shows the role of polyphase interpolation filters assume zero-delay( ) processing. Essentially, the filter interpolates the waveform -way between consecutive input samples. The polyphase outputs are then interleaved to create the output stream. Assuming that is bandlimited to , perfect polyphase filtering yields a perfectly interpolated output. In practice, we use the casual FIRapproximations of the polyphase filters (which which correspond to some casual FIR approximation of the master filter ).
Notification Switch
Would you like to follow the 'Digital signal processing (ohio state ee700)' conversation and receive update notifications?