Scribd Logo
Upload

Welcome to Scribd!

  • 30 views

    Math 242 Lecture 4: Pick Up From Last Time..

    Uploaded by

    Andy Nguyen
    This document summarizes key concepts from a math lecture on the Heisenberg uncertainty principle and the sampling theorem. The Heisenberg uncertainty principle establishes a lower bound on the product of the standard deviation of position and momentum for any function. The proof uses properties of Fourier transforms. The sampling theorem states that a bandlimited function (a function whose Fourier transform is compactly supported) is completely determined by its sample values at regularly spaced intervals. The reconstruction formula uses the sinc kernel. The document traces the history of the sampling theorem to several contributors in the early 20th century. It also provides a proof using the Fourier series representation of a bandlimited function's Fourier transform.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Math 242 Lecture 4: Pick Up From Last Time..

    Uploaded by

    Andy Nguyen
    30 views3 pages
    This document summarizes key concepts from a math lecture on the Heisenberg uncertainty principle and the sampling theorem. The Heisenberg uncertainty principle establishes a lower bound on the product of the standard deviation of position and momentum for any function. The proof uses properties of Fourier transforms. The sampling theorem states that a bandlimited function (a function whose Fourier transform is compactly supported) is completely determined by its sample values at regularly spaced intervals. The reconstruction formula uses the sinc kernel. The document traces the history of the sampling theorem to several contributors in the early 20th century. It also provides a proof using the Fourier series representation of a bandlimited function's Fourier transform.

    Original Description:

    ucsd math 242

    Original Title

    lec 4

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document summarizes key concepts from a math lecture on the Heisenberg uncertainty principle and the sampling theorem. The Heisenberg uncertainty principle establishes a lower bound on the product of the standard deviation of position and momentum for any function. The proof uses properties of Fourier transforms. The sampling theorem states that a bandlimited function (a function whose Fourier transform is compactly supported) is completely determined by its sample values at regularly spaced intervals. The reconstruction formula uses the sinc kernel. The document traces the history of the sampling theorem to several contributors in the early 20th century. It also provides a proof using the Fourier series representation of a bandlimited function's Fourier transform.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    30 views3 pages

    Math 242 Lecture 4: Pick Up From Last Time..

    Uploaded by

    Andy Nguyen
    This document summarizes key concepts from a math lecture on the Heisenberg uncertainty principle and the sampling theorem. The Heisenberg uncertainty principle establishes a lower bound on the product of the standard deviation of position and momentum for any function. The proof uses properties of Fourier transforms. The sampling theorem states that a bandlimited function (a function whose Fourier transform is compactly supported) is completely determined by its sample values at regularly spaced intervals. The reconstruction formula uses the sinc kernel. The document traces the history of the sampling theorem to several contributors in the early 20th century. It also provides a proof using the Fourier series representation of a bandlimited function's Fourier transform.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 3

    Andrew Nguyen

    April 14, 2016

    Math 242 Lecture 4


    pick up from last time...
    Definition. a f := (x a)2 | f (x)|2 d x
    R

    .R

    | f (x)|2 d x .

    Theorem (Heisenberg Uncertainty).


    1
    , x 0 , 0 R.
    162

    f L 2 , (x0 f )(0 f)

    We will worry about (x 0 , 0 ) = (0, 0), and then shifts will take care of the rest.
    Proof.
    (I). Suppose that x f L 2 , so that x 2 f 2 = and 0 f = .
    R

    Then if f 0 L 2 , we have f L 2 and 0 f = .


    (II). We can now assume freely that f C (R) P S(R), with x f L 2 , f 0 L 2 . PS stands for
    piecewise smooth.
    b R

    Consider ab x f (x) f 0 (x)d x = x| f (x)|2


    a
    0 as a, b .
    R

    So

    | f (x)|2 + x f 0 (x) f (x) d x , where x| f (x)|2 goes to


    a

    | f (x)|2 f x = 2 Re

    x f (x) f 0 (x)d x.

    And so

    || f

    ||42

    Z
    = 4 Re

    x f (x) f (x)d x

    =4

    So
    || f ||22 || f||22 162

    | f 0 (x)|2 d x

    Z
    2
    2
    2
    2
    2
    x | f (x)| d x 4
    | f ()| d .
    x | f (x)| d x

    x 2 | f (x)|2 d x

    2 | f()|2 d ,

    where x 2 | f (x)|2 d x is just (0 f )|| f ||2 and 2 | f()|2 d is just (0 f )|| f||2 .
    R

    Remark. This is a sharp inequality if youre not gaussian. Gaussians actually achieve equality.

    The Sampling Theorem


    Definition (Band-Limited). We say that f is bandlimited with bandlimit if f is the inverse
    fourier transform of a compactly supportedR function in L 2 (R). In this case we write f B

    F with supp F (, ) such that f (x) = F ()e 2i t d .


    It is customary to say f B if f is compactly supported. This would mean f is equal a.e. to
    a function defined with the other definition.
    Theorem (Sampling theorem). . Suppose f B L 2 (R). (already weve switched to being loose with
    definitions.) Then f is completely determined by its samples at x = k , k Z.
    In fact,
    X

    kZ

    k sin((x k/))
    = f (x).

    (x k/)

    The thing in the parentheses is often called the sinc kernel,


    sin((x k/))
    = sinc(x k)
    (x k/)

    Often attributed by various people, who argued a lot (and got up at conferences and yelled, etc)
    to whittaker (1915), kotelnikov (1933), Ogura (1920), Raabe (1939), Shannon by engineers (1948),
    Someya (1949). Now its just called the Sampling theorem. Also, Nyquists name is involved here
    for some reason? Shannon was the only person on this list really not to do it from a math point of
    view (?)
    Proof. For this proof, supp | f| (/2, /2). f is compactly supported, so we can write its Fourier
    series:
    f() =

    1
    c k p e 2i k/ , || < /2.

    kZ
    X

    Z
    1
    1 /2
    c k = f, p e 2i k/ =
    f ()e 2i k/ d
    /2

    Z
    1
    =p
    f()e 2i k/ d ( B )

    By Fourier inversion,
    =

    1
    f (k).

    Z
    Z /2
    f (x) =
    f()e 2i x d =
    f()e 2i x d

    /2
    Z /2 X
    1 2i k/ 2i x
    =
    ck p e
    e
    d .
    /2 kZ

    Exchange sums and integrals, and use the Fourier inversion to reveal the c k as the sampling points
    of f . The remaining integral over the exponentials can be evaluated easily, and rewritten using
    Eulers identity.
    =

    f (k/)

    kZ

    f (k/)

    kZ

    sin((x + k/))
    .
    (x + k/)

    sin((x k/))
    .
    (x k/)

    The sinc kernel fell out of the proof as an explicit way to reconstruct the original function from
    a finite number of sampling points. However, it is not a very good kernel, since it decays too
    slowly. (/2, /2) is known as the critical bandwidth, but if we pick larger integration limits, say
    (5/2, 5/2), and multiplied our Fourier transform by anything nice and smooth in the Fourier
    world, say g , as long as g () = 1 for || < /2, then g decays fast in the spatial world. Unfortunately
    our new Fourier coefficients will require us to sample five times as fast.
    (In practice, what happens if you dont sample EXACTLY at the right time??)
    Fact: (sin((xk/))/((xk/)) forms an orthonormal basis for the bandlimited functions,
    i.e. B /2 .
    Remark. To obtain x from { f (k/) }kZ , apply the sinc formula.
    INSERT PROOF BY PICTURE HERE

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy