This preview shows page 1. Sign up to view the full content.
Unformatted text preview: j/2 This is the same as negating the frequency. As the frequency of a sine
increases past its Nyquist rate, it wraps around to negative frequency.
EE102A:Signal Processing and Linear Systems I; Win 0910, Pauly 31 Minimizing Aliasing
Once a signal has been sampled, there is no way to eliminate aliasing (unless
we have some other information about the signal).
Aliasing is minimized by ﬁrst lowpass ﬁltering the signal, then sampling:
f (t ) f¯(t ) h(t )
−!0 /2 !0 /2 ×
!T (t ) f (t ) h(t )
−!0 /2 !0 /2 Lowpass or
Antialiasing Filter
EE102A:Signal Processing and Linear Systems I; Win 0910, Pauly f¯(t )
t = nT Sampler 32 Any practical antialiasing ﬁlter will not be identically zero outside of its
passband. Some aliasing will always occur.
By bandlimiting with the antialiasing ﬁlter, we are choosing to distort the
signal in a known way.
This is usually preferable to sampling the nonbandlimited signal, and having
unknown artifacts from aliasing.
Bandlimiting also suppresses noise. EE102A:Signal Processing and Linear Systems I; Win 0910, Pauly 33 Fourier Series and Fourier Transforms
Frequency domain representation of signals simpliﬁes signal processing,
• Convolution becomes multiplication
• Spectral representation makes signal content easier to understand
• Modulation/demodulation
• Sampling
Often, Fourier transforms are the only tool you will need.
Some problems involve signals that don’t have Fourier transforms, like
increasing exponentials (ﬁnance, populations, dynamic systems).
For these we need a more powerful tool.
EE102A:Signal Processing and Linear Systems I; Win 0910, Pauly 34...
View
Full
Document
This note was uploaded on 09/28/2013 for the course EE 108b taught by Professor Mukamel during the Fall '13 term at Singapore Stanford Partnership.
 Fall '13
 Mukamel
 Signal Processing

Click to edit the document details