Learn the fundamentals of digital signal processing theory and discover the myriad ways DSP makes everyday life more productive and fun.

Loading...

Do curso por École Polytechnique Fédérale de Lausanne

Processamento Digital de Sinais

329 classificações

Learn the fundamentals of digital signal processing theory and discover the myriad ways DSP makes everyday life more productive and fun.

- Paolo PrandoniLecturer

School of Computer and Communication Science - Martin VetterliProfessor

School of Computer and Communication Sciences

Hello, welcome to our online class on digital signal processing.

If you are here it's because you are curious about signal crossing, and

you're curious about this class on signal processing.

>> We know you have your choice of classes out there, so

our job now is to try and convince you to stick with this one.

Let's talk a bit about ourselves.

My name is Martin Mettely.

I'm a professor at EPFL.

I have been doing teaching and researching signal processing for

the last couple of decades, and I still love the topic like on the first day.

>> And my name is Paulo Brandoni, and half of my time I teach here at EPFL

where I got my PhD in the last century, the other half of my time I tried

to run this tech startup, but when I grow up I want to be a rock musician.

>> Okay, so you're the practical guy.

>> And you're supposedly the theory guy, except that you always ask the more and

more practical examples in your classes.

>> And you always ask for more and more theory.

>> And I guess that's what psychiatrists call, overcompensation.

>> Enough about ourselves, let's talk about signal processing actually is.

This is when people usually pull out their fancy cellphone, and

promise that by the end of the class, you'll understand every detail about it.

But, actually,

a simple coffee grinder, is enough to explain what signal processing is.

The beans are the signals,

the grinder is a processor, and the ground coffee is the useful for

useful information that we extract from the signals.

>> All right, but what if you want to grind something other than coffee.

Suppose you want to crush some ice for some margaritas?

>> Okay, this is why we teach digital signal processing,

where we have a universal processor.

>> A computer.

>> Exactly!

And a as long the signals are represented a sequence of numbers,

we can apply all sorts of algorithms to all sorts of different data.

This is why your cell phone is such a powerful device.

>> Aha, so we are bringing out the cell phone after all.

>> You call this a cell phone?

Are you asking for a pay raise?

>> It's alright, I'll update my gear next time.

But the idea here is that, if you have a general purpose computer you

can apply processing to virtually all kind of interesting signals.

>> Okay, but we have not yet explained what the signal actually is.

>> That's true.

So in theory,

a signal is the measurement of a physical phenomena that evolves over time.

>> And in practice, this could be sound pressure, temperature,

light intensity, or a voltage.

>> And a digital signal is a sequence of regular measurements of this phenomenon.

And the cool thing is that,

if we take these measurements often enough, the resulting sequence of numbers

will contain all the relevant information about the original phenomenon.

>> This is where we need a mathematical framework to talk about signal processing.

The convenient framework is linear algebra, in particular, Hilbert space.

>> Aha, did you say Dilbert space?

>> Come on.

>> [LAUGH] >> Let's say that it's very convenient

to see signals as vectors embedded it in a vector space.

>> Okay, so you said the V word, and

I thought this video was going to be all lame jokes and fun.

And now, people are going to look inside our blog and see that it is full of math

and formulas, a nd we're going to lose half of our audience, at least.

>> Let me say two things about this.

First, the book is available for free, and one doesn't look for

a gift horse in the mouth.

Second, theorems are extremely important, they're actually life or death.

>> Okay, sell it to me.

>> Take, for example, a CAT scan >> A machine takes a bunch of x-rays, and

the fancy DSP algorithm makes a three dimensional reconstruction of your brain.

>> Like this one, which is me, by the way.

>> Exactly.

Now, if a medical doctor wants to produce a diagnosis based on the reconstruction,

you want to make sure the reconstruction is correct and unique.

>> Sure, I don't want to end up with someone else's brain.

>> And signal processing theorems give you exactly this guarantee.

>> Okay, but then we can say, that all the math and the theory that we will study,

will always be at the service of practical applications.

>> Exactly. And we went through great pain and

efforts to produce a series of videos called, Signal of the Day, which present

interesting and famous signal processing sequences and relevant algorithms.

>> And remember, your PC is a fully functional signal processing lab, so

we will encourage you to write code to test the theory that we will teach you,

and also to develop some cool applications.

>> For example, you learn how to program an MP3 player.

>> Well, I guess our sales pitch is over.

>> We hope you enjoy the class.

O Coursera proporciona acesso universal à melhor educação do mundo fazendo parcerias com as melhores universidades e organizações para oferecer cursos on-line.