DCSP-3: Fourier Transform (continuous time) Jianfeng Feng [email protected] feng/dcsp.html.

DCSP-3: Fourier Transform(continuous time)

Jianfeng Feng

[email protected]

http://www.dcs.warwick.ac.uk/~feng/dcsp.html

Two basic laws

• Nyquist-Shannon sampling theorem

• Hartley-Shannon Law

(channel capacity)

Best piece of applied math.

Communication Techniques

Time, frequency and bandwidth (Fourier Transform)

Most signal carried by communication channels are modulated forms of sine waves.

Communication Techniques

Time, frequency and bandwidth (Fourier Transform)

Most signal carried by communication channels are modulated forms of sine waves.

A sine wave is described mathematically by the expression

s(t)=A cos ( t

The quantities A, , are termed the amplitude, frequency and phase of the sine wave.

Communication TechniquesTime, frequency and bandwidth

We can describe this signal in two ways.

One way is to describe its evolution in the time domain, as in the equation above.

Why it works?

(x, y) = x (1,0) + y(0,1)

f(t) = x sin(omega t) + y sin (2 omega t)

time domain vs. frequency domain

Communication TechniquesTime, frequency and bandwidth

We can describe this signal in two ways.

One way is to describe its evolution in the time domain, as in the equation above.

The other way is to describe its frequency content, in frequency domain.

The cosine wave, s(t), has a single frequency, =2 /T where T is the period i.e. S(t+T)=s(t).

This representation is quite general. In fact we have the following theorem due to Fourier.

Any signal x(t) of period T can be represented as the sum of a set of cosinusoidal and sinusoidal waves of different frequencies and phases.

where A0 is the d.c. term, and T is the period of thewaveform. The description of a signal in terms of its constituent

frequencies is called its frequency spectrum.

X = x (1,0)*(1,0)’+y(0,1)*(1,0)’

where A0 is the d.c. term, and T is the period of thewaveform.

where A0 is the d.c. term, and T is the period of thewaveform. The description of a signal in terms of its constituent

frequencies is called its frequency spectrum.

Example 1X(t)=1, 0<t<, 2<t<3, 0 otherwise

Hence X(t) is a signal with a period of 2

Time domain

Frequency domain

Script1_1.m

NoteFrequency (Hz)

FrequencyDistance fromprevious note

Log frequencylog2 f

Log frequencyDistance fromprevious note

A2 110.00 N/A 6.781 N/A

A2# 116.54 6.54 6.8640.0833 (or 1/12)

B2 123.47 6.93 6.948 0.0833

C2 130.81 7.34 7.031 0.0833

C2# 138.59 7.78 7.115 0.0833

D2 146.83 8.24 7.198 0.0833

D2# 155.56 8.73 7.281 0.0833

E2 164.81 9.25 7.365 0.0833

F2 174.61 9.80 7.448 0.0833

F2# 185.00 10.39 7.531 0.0833

G2 196.00 11.00 7.615 0.0833

G2# 207.65 11.65 7.698 0.0833

A3 220.00 12.35 7.781 0.0833

A new way to represent a signal is developed: wavelet analysis

• Fourier1.m

• Script1_2.m

• Script1_3.m

3. Note that the spectrum is continuous now: having power all over the place, rather than discrete as in periodic case.

4. A straightforward application is in data compression.

Fourier's Song• Integrate your function times a complex exponential

It's really not so hard you can do it with your pencilAnd when you're done with this calculationYou've got a brand new function - the Fourier TransformationWhat a prism does to sunlight, what the ear does to soundFourier does to signals, it's the coolest trick aroundNow filtering is easy, you don't need to convolveAll you do is multiply in order to solve.

• From time into frequency - from frequency to time• Every operation in the time domain

Has a Fourier analog - that's what I claimThink of a delay, a simple shift in timeIt becomes a phase rotation - now that's truly sublime!And to differentiate, here's a simple trickJust multiply by J omega, ain't that slick?Integration is the inverse, what you gonna do?Divide instead of multiply - you can do it too.

• From time into frequency - from frequency to time• Let's do some examples... consider a sine

It's mapped to a delta, in frequency - not timeNow take that same delta as a function of timeMapped into frequency - of course - it's a sine!

• Sine x on x is handy, let's call it a sinc.Its Fourier Transform is simpler than you think.You get a pulse that's shaped just like a top hat...Squeeze the pulse thin, and the sinc grows fat.Or make the pulse wide, and the sinc grows dense,The uncertainty principle is just common sense.