<p>So I’m a physics (applied physics) student interested in taking some systems/signals classes through the EE department at my school. There are 3-4 separate classes i’m interested in taking:</p>
<p>continuous systems/signals
discrete systems/signals
digital signal processing
digital signal processing lab</p>
<p>I’m not an EE major, so I can’t take the class in the usual order (as above).
All I’m asking is if the order matters. I’m stuck between a few course orders</p>
<ol>
<li>continuous systems/signals -> discrete systems/signals -> graduate and learn DSP on my own/job/gradschool</li>
<li>discrete systems/signals -> digital signal processing -> continuous systems/signals & digital signal processing lab </li>
<li>discrete systems/signals -> 2 classes in numerical analysis -> 1 class numerical differ eq (partial and ordinary). </li>
</ol>
<p>or some weird but do-able/useful combination above.</p>
<p>I just don’t want to do route 2, get into DSP class and be confronted with lots of stuff from the continuous systems. also, I’m also curious how useful continuous systems are. I’ve tried looking up examples via google, but all Ive ever seen was “an example of continuous systems is the AC outlet”. </p>
<p>Additionally, im curious how “easy” certain topics (DSP and numerical analysis specifically) are to learn outside of a classroom.
ie. if someone said they learned to speak japanese/program in C on their own, I’d probably believe them. on the other hand, if someone said they learned quantum physics/fluid mechanics on their own, i’d be suspicious…</p>
<p>Lastly, how useful would it be to take a class in statistics/engr statistics if I went into signals? I would still have to take stat mech (at the same time as the DSP class if route 2/before any systems class if route 1).</p>
<p>You should take discrete systems before discrete signal processing.
And maybe continuous signals before discrete.
But I think if the school lets you sign up (ie you have prereqs) then it doesn’t matter.</p>
<p>This is interesting. At most schools discrete and continuous are taught in the same course in Linear Systems because the techniques are really the same at a certain level. I would think that you would want to understand the techniques at that level. </p>
<p>I think what you want to do is an excellent idea. </p>
<p>For you, I would recommend that you do option 2, but get the book Signals and Systems by Oppenheim and Willsky. I think that you can learn continuous on your own as follows: While you are learning concepts in discrete, see if you can figure out the same concept in continuous. Since you are a physics major, you are likely to have seen some Fourier techniques before in solving electrostatics problems and have a pretty good math background. I think it’s a much richer subject if you do continuous and discrete together.</p>
<p>continuous and discrete are different classes because my school is the quarter system. i cant imagine cramming both subjects into 10 weeks. continuous is offered in the winter, discrete in the spring and digital in the fall. im also curious how useful it would be to go through discrete and digital then “come back” to continuous. I’m also curious how useful numerical analysis would be to signal processing, and if it could be “more fundamental”.
I’d rather learn something at a more fundamental level, and build up
(no, i wont take the math class on fourier analysis…)</p>
<p>the reason why part of me thinks i can neglect continuous signal processing is because all the comp engr (not EE!) only have to take discrete signals (and take take dsp lecture/lab as electives). </p>
<p>also, any voices on the statistics thing i mentioned?</p>
<p>In a sense the stuff from “continuous” looks at techniques for linear systems described by linear constant coefficient differential equations. The stuff from “discrete” looks at techniques for linear systems described by linear constant coefficient difference equations. The techniques are very similar. Once you learn one, learning the other is pretty straightforward. </p>
<p>Therefore, I don’t think that you will have a problem doing either discrete>digital>continuous or continuous>discrete>digital. I don’t think a numerical analysis class is really related or more fundamental. In some ways, continuous might be more fundamental than discrete because I suspect that you have more experience with differential equations than difference equations (also called recurrence equations). If you studied differential equations and Laplace transforms in diffEq, than continuous is the natural follow-on to that. You can learn the basics from that and then do discrete. I think it’s important to understand the relationship between continuous and discrete though.</p>
<p>I have a bit of experience with difference equations (although I realized we always called them recurrence). The reason I thought numerical analysis would be “fundamental” is cause my experience with recurrence has always been through series solutions, something thats covered a lot in numerical analysis classes.
I’m also curious in classes that do a lot of programming. I know all the numerical classes at my school use matlab out the wa-zo. although DSP probably does too.</p>
<p>and the main problem i have taking continuous is that I cant take it until midway through my senior year…
at the very least, ill take discrete. im still not 100% on digital or continuous… any other view points?</p>
<p>Continuous systems are really useful. Below I’ll give one example of an important subject where linear systems theory is often applied.</p>
<p>Maxwell’s equations can be viewed as a linear system with charge and current sources as input and the field generated as output. The fields, except possibly at boundaries between different materials, are continuous.</p>
<p>Maxwell’s equations are important for quite a few technologies . . . fiber optic & radio wave communication, optical imaging systems, high-speed integrated circuits . . .</p>
<p>interesting. I was talking with a TA of mine and he mentioned that EVERYTHING in science is technically continuous (except for spin states in quantum systems) due everything being an ensemble of something.
Of course, a problem that comes to mind is what is the system actually doing vs. how we are modeling it; re modeling maxwell as continuous system…</p>
<p>STUPID QUESTION: do continuum mechanics and continuous signals overlap in how they teach subject matter? (i know taking continuum mech before gen relativity is good because it involves putting space and time on the same continuum, similar for signals?)</p>