/Digital-Filter-Design

My solution for the Filter Design Assignment of the course EE 338: Digital Signal Processing

Primary LanguageMATLABMIT LicenseMIT

Digital Filter Design

My solution for the Filter Design Assignment of the course EE 338: Digital Signal Processing As a part of this assignment, I had to design 6 different filters based a set of specifications given. The detailed information is available in the report{:target="_blank"}.
These are the following filters that have been designed:


Infinite Impulse Response Filters

  1. Bandpass filter using Butterworth design
  2. Bandstop filter using Chebyshev design

Finite Impulse Response Filters

  1. Bandpass filter using Kaiser window
  2. Bandstop filter using Kaiser window

Elliptic Filters

  1. Bandpass filter
  2. Bandstop filter

Some observations

After realising the bandpass and bandstop filters for a given set of specifications in various forms, I was able to make the following conclusions:

  • Realisation using an elliptic approximation gives the least order, i.e, it requires mininum resources. The order of resources required is as follows (for a given set of specifications):
    Elliptic (3) < Chebyshev (4) < Butterworth (8) < Kaiser FIR (50-60)
  • The elliptic filter has the sharpest transition from passband to stopband or vice versa for the same specifications. In particular, the decreasing order of transition sharpness is:
    Elliptic > Chebyshev > Butterworth > Kaiser FIR
  • The FIR realisation gives a linear phase response in the passband region. The nonlinearity of the phase response decreases in the following order:
    Elliptic > Chebyshev > Butterworth > Kaiser FIR