6
$\begingroup$

Perhaps a meta-question, but I'll post it here because some of you are bound to possess gold nuggets about continuing education. It might be closed for being subjective but I don't care because this community is small and this could be valuable info for visitors.

Aside from reading DSP.StackExchange questions and answers, I've been doing the following:

  1. MIT OCW Signals and Systems, but Oppenheim's lectures from the 80's and not the 2004 course, actually doing the problem sets, reviewing the solutions, and really understanding where a process comes from

  2. Reviewing circuit analysis diagrams

  3. Taking the concepts from Signals and Systems and plotting / manipulating them in MATLAB

Of course, there's no substitute for working on a project and having your *** on the line for not making it work, but things can only get better if your fundamentals are really solid, right?

Improve this question
$\endgroup$
3
  • 4
    $\begingroup$ some things about DSP (like what IIR vs. FIR is, DFT, quantization error, etc) are timeless concepts that are no different now than in the 80's. maybe check out the USENET newsgroup comp.dsp. (there is a website, dsprelated.com that will get you access to comp.dsp.) $\endgroup$
    – robert bristow-johnson
    Sep 25 '15 at 23:09
  • $\begingroup$ This is a pretty broad question, but it's good for you to try and learn! So, this depends a lot on what you know about math so far. From the CS students I know, about a half have actually taken the same math as math students for the first two semesters (that being Analysis and Linear Algebra), which actually is an excellent base for learning DSP, as these are very familiar with concepts like vector spaces, convergence, differentiability etc. Others have taken Higher Math for CS, and that's less optimal for DSP, but the gaps they have aren't that big. $\endgroup$
    – Marcus Müller
    Sep 30 '17 at 10:21
  • $\begingroup$ So, what you should probably aim for is getting the basics in signal and systems theory. Look around for well-reputed universities with an electrical engineering degree that have a public description (or maybe even publicly available lectures!) about Signals and Systems. Go, figure out the dependencies of exactly that course, and learn them, then learn the contents of that course, is what I'd recommend. $\endgroup$
    – Marcus Müller
    Sep 30 '17 at 10:23
17
$\begingroup$

Tips for DSP self-study huh. Well, ...studying 'signals and systems' is a great idea and having Matlab software means you have the tools to learn an awful lot. I think Dr. Steven Smith's book "The Scientist and Engineer's Guide to Digital Signal Processing", which you can read online for free, is a terrific source of fundamental DSP information. Dr. Smith is a good writer. His book is at: http://www.dspguide.com/pdfbook.htm. Smith is a skilled working DSP engineer. That means he knows which end of the soldering iron is hot.

With regard to a free DSP textbook homework problems, I recently learned that Sophocles J. Orfanidis, the well-known professor with the ECE Department of Rutgers University, has made his signal processing textbook available for downloading on the Internet. The textbook is: "Introduction to Signal Processing" available at: http://eceweb1.rutgers.edu/~orfanidi/intro2sp/. Happily, also available at the above web site are: (1) Errata for the Orfanidis textbook, (2) Orfanidis' Homework Solutions Manual, and (3) the errata for Orfanidis Solutions Manual.

At the risk of appearing tacky, you might consider my "Understanding Digital Signal Processing" textbook. It's been Amazon.com's top-selling DSP book for some years now. My book's been popular not because I'm any more DSP-skilled than other book authors---I can assure you that is not the case---but rather because I have some skill at explaining signal processing concepts from a working engineer's viewpoint rather than a university professor's strictly-mathematical viewpoint. My book costs money. However, I can assure you, I use the royalties from my book to buy milk and bread for my children.

So Anthony, learning the fundamentals of DSP through reading, performing lots and lots of Matlab modeling, and maintaining your enthusiasm, to quote Susan B. Anthony, "Failure is impossible." Good Luck.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Hi Richard, I own your book as well, after working through Signals and Systems, working through your problems is next on my list. I didn't know about the Orfanidis resource, that is great, thank you. $\endgroup$
    – panthyon
    Sep 26 '15 at 14:48
  • 2
    $\begingroup$ @Anthony Parks. If you're interested, you can send me an e-mail at: R_dot_Lyons_at_ieee_dot_com and I'll e-mail the errata for the 3rd edition of my "Understanding DSP" book to you. $\endgroup$
    – Richard Lyons
    Sep 27 '15 at 5:57
3
$\begingroup$

I can recommend online course - Coursera DSP. There are very good introduction in mathematical basis of DSP and review of main DSP themes.

Online courses are symbiose of self-study (study time freedom) and regular education (you will have feedback and you can discuss your problems in forum with another students).

Improve this answer
$\endgroup$
2
$\begingroup$

DSP would be a quite broad subject and where to start would depend on your ability to pick up information, learn, understand and implement it. A basic course for supplying us with the basics to understand DSP would be a 'Signals and Systems' Course. If you have the time and patience to start with the basics you could take this course - https://www.youtube.com/watch?v=c_9JxwuEdqE. It's an IIT BombayX course archived on edx. Else if you would rather pick up on the go, then you could search up online and learn the various important things in DSP. First, to start with, one needs to gather a good understanding about signals, sampling, quantisation, different types of signals- audio, images, etc., how they are represented in time domain and frequency domain (what it actually means), Linearity, Time-invariance, LTI systems, impulse response, convolution, frequency response, fourier series, fourier transform, z-transform. Else if your target is specific like speech processing or image processing, you could rather dive right into the various tutorials available online already and when you dont understand something you can try searching it online, lots of info on these topics are available. True that sometimes its hard to understand and make sense of things when studying on one's own.. but if someone is not available who knows DSP, to teach you.. you could always try these options.. :)

Improve this answer
$\endgroup$
1
  • $\begingroup$ This is a really good exposition to 'signals and systems', which is a first step towards learning DSP. Prof. Gadre's lectures are very systematic, and the challenging problems in each lecture is a step towards higher concepts in DSP $\endgroup$
    – AnVij
    May 22 at 10:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.