Decimation/Interpolation - Anti-Aliasing and Anti-imaging Filter Design in MATLAB

I've got a discrete time signal that I'm looking to decimate and interpolate.

To decimate, I understand that the signal must first be passed through a low pass filter (anti-aliasing filter) before being decimated (taking 1 sample every M as output, where M is the decimation factor). See diagram below

To interpolate, the signal must be filled with zeros according to the factor of interpolation and the signal must then be passed through a low pass (anti-imaging) filter. See diagram below.

How can I go about designing an anti-aliasing and anti-imaging filter using MATLAB? I want to create my filter as an FIR filter.

Any resources would be greatly appreciated!

Answer 1

[Star Strider]

All this is already implemented in the resample function.

Have you experimented with it, or do you simply want to reproduce its results?

[William Greenway]

I actually want to design a code in MATLAB so that I can write a c program that achieves the same, therefore I want to obtain the coefficients of the FIR transfer function. Do you know how this might be possible?

[Star Strider]

There are several ways to design FIR filters in MATLAB, the fir1 function probably being the most commonly used (at least in my experience), although fir2 may be more appropriate for your purposes.

Generically, FIR filter design methods can be found in any signal processing textbook.

[William Greenway]

Makes sense - are there any signal processing textbooks that you can reccommend specific to anti-aliasing and anti-imaging filters?

[Star Strider]

The old stand-bys in signal processing are Oppenheim and Proakis (textbooks I used in grad school in different courses) although there are others, and much can be found online. I’ve not used these specific techniques to design filters since grad school (a while ago), letting MATLAB do everything for me, so you will need to see which is best for your needs.

I would experiment with the coding them in MATLAB and then compare them with the filter designs MATLAB produces as the ‘correct’ result.