160 likes | 379 Views
Multirate Digital Signal Processing Problem : process data sampled at different rates. Typical example : audio data sampled at 8kHz, to be played by a sound card operating at 20kHz; Important Applications : Data Compression., Digital Communication, Spectral Estimation, Filtering.
E N D
Multirate Digital Signal Processing • Problem: process data sampled at different rates. • Typical example: • audio data sampled at 8kHz, to be played by a sound card operating at 20kHz; • Important Applications: Data Compression., Digital Communication, Spectral Estimation, Filtering. Main problem: Preserve the Frequency Spectrum resample
Two Main Operations: DownSampling by D: D UpSampling by I: I zeros
Consider an intermediate operation called Sampling. Define the sampling sequence: Then, given a sequence we define the sampled sequence .
We can relate the frequency spectrum of with the frequency spectrum of . Take the case of . Then we can write Modulation Modeling of Sampling
+ = even samples add odd samples cancel
In general for sampling by D : We can also write in exponential form as
Now we can determine the effect of DownSampling by an integer D. First Notice the following: D D Therefore:
no aliasing! aliased back
LPF D We need to Low Pass Filter before the decimation.
UpSampling by a Factor I: I Frequency Spectrum: we need to filter these out
LPF I
Multirate by a Factor I/D: LPF LPF I D combine the two LPF LPF I D