Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://hea-www.harvard.edu/AstroStat/astro193/scientificamerican0689-86.pdf Äàòà èçìåíåíèÿ: Mon Mar 30 19:38:26 2015 Äàòà èíäåêñèðîâàíèÿ: Sun Apr 10 13:23:59 2016 Êîäèðîâêà:

The Fourier Transform
DNA's double helix, the sunspot cycle and the sawtooth signals of electronics can be reduced mathematically to a series . of undulating curves. This idea underlies a powerful analytical tool
by Ronald N. Bracewell

intellect can execute only after years of mathematical education. The ear formulates a transform by converting sound-the waves of pressure travel ing through time and the atmos phere-into a spectrum, a description of the sound as a series of volumes at distinct pitches. The brain turns this information into perceived sound. Similar operations can be done by mathematical methods on sound waves or virtually any other fluctu ating phenomenon, from light waves to ocean tides to solar cycles. These mathematical tools can decompose functions representing such fluctu ations into a set of sinusoidal com ponents-undulating curves that vary from a maximum to a minimum and back, much like the heights of ocean waves. The Fourier transform is a function that describes the amplitude and phase of each sinusoid, which corresponds to a specific frequency. (Amplitude describes the height of the sinusoid; phase specifies the start ing point in the sinusoid's cycle.) The Fourier transform has become a powerful tool in diverse fields of sci ence. In some cases, the Fourier trans form can provide a means of solving unwieldy equations that describe dy namic responses to electricity, heat or light. In other cases, it can identify the regular contributions to a fluctuating

T

signal, thereby helping to make sense ten. The ear automatically per o calculate a transform, just the forms the calculation, which lis of observations in astronomy, medi cine and chemistry.

Returning to France relatively un scathed, Fourier focused on math ematical matters as professor of anal ysis at the Polytechnic School, but in 1802 he again entered Napoleon's

is named, Baron Jean-Baptiste-Joseph Fourier. Fourier was not merely inter ested in heat; he was obsessed by it. He kept his home in Grenoble so uncomfortably hot that visitors of ten complained. At the same time he would cloak himself in heavy coats. Perhaps it was the lure of a warm climate that in 1798 drew Fourier to join the retinue of 165 savants that accompanied Napoleon's expedition to Egypt. While Napoleon was fighting Syr ians in Palestine, repelling the Turks from Egypt and hunting the Mameluke chief, Murad Bey, the French scientists undertook ambitious studies in ge ography, archaeology, medicine, agri culture and natural history. Fourier was appointed secretary of a scientific body known as the Institute of Egypt. He discharged administrative duties with such competence that he re ceived many diplomatic assignments. Yet he was still able to conduct inten sive research on Egyptian antiquities and contemplate a theory about the roots of algebraic equations. Shortly before the French were driv en from Egypt in 1801, Fourier and his colleagues set sail for France. The commander of the British fleet, Admiral Sir Sidney Smith, promptly seized their ship along with its cargo of Egyptian documents and relics. In the honorable spirit of the time, Smith put the scientists ashore unharmed in Alexandria. The English commander eventually traveled to Paris to return the confiscated material-except for the Rosetta stone (the key to Egyptian hieroglyphics), which stands today in the British Museum as a monument to Napoleon's military defeat and his contribution to Egyptology.

T

technique from the mathema he world whom the about the tician forfirst learned transform

service. Fourier became the prefect of the Isere department. While attempt ing to repair the disruptions remain ing from the Revolution of 1789, he built the French section of the road to Turin and drained 80,000 square ki lometers of malarial swamp. During this time he derived an equation that described the conduction of heat in solid bodies. By 1807 Fourier had in vented a method for solving the equa tion: the Fourier transform.

instances of heat conduction. A ourier applied his mathemati particularly instructive example that avoids computational complications is the flow of heat around an anchor ring-an iron ring that attaches a ship's anchor to its chain-that has been thrust halfway into a fire. When part of the circumference becomes red hot, the ring is withdrawn. Before much heat is lost to the air, the ring is buried in fine, insulating sand, and the temperature is measured around the outer curve

F

cal technique to explain many

page 88).

[see illustration on

Initially the temperature distribu tion is irregular : part of the ring is uniformly cool, and part is uniform ly hot; in between the temperature abruptly shifts. As heat is conducted from the hot region to the cool re gion, however, the distribution begins to smooth out. Soon the temperature distribution of heat around the ring reaches a sinusoidal form: a plot of the temperature rises and falls evenly, like an S curve, in exactly the way sine and cosine functions vary. The sinus oid gradually flattens until the whole ring arrives at a constant temperature. Fourier proposed that the initial, irregular distribution could be bro ken down into many simple sinusoids

RONALD N. BRACEWELL has been on the faculty of the electrical engineering department at Stanford University since
1955. He was educated at the Univer

sity of Sydney and the Cavendish lab oratory in Cambridge, England, where he got his Ph.D. His research interests have ranged over microwave radar, ion ospheric physics and radio astronomy. At Stanford he is a member of the Space, Telecommunications and Radioscience Laboratory and is professor of comput er science, by courtesy.

86

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC


that had their own maximum temper ature and phase, that is, relative po sition around the ring. Furthermore, each sinusoidal component varied from a maximum to a minimum and back an integral number of times in a single rotation around the ring. The

one-cycle variation became known as the fundamental harmonic, whereas variations with two, three or more cycles in a single rotation became the second, third and higher harmon ics. The mathematical function that describes the maximum temperature

and position, or phase, for each of the harmonics is the Fourier transform of the temperature distribution. Fourier had traded a single distribution that was difficult to describe mathemati cally for a more manageable series of full-period sine and cosine functions

SUNBEAM resolved into a spectrum provides a physical anal ogy for mathematical transforms (top). The sunlight entering the prism varies in strength from moment to moment

an amplitude at each frequency. Thus, a function of strength versus time has been transformed into a function of ampli tude versus frequency. The Fourier transform can represent a time-varying signal as a function of frequency and ampli tude, but the transform also provides information about phase.

(bottom).

The light leaving the prism has been separated in space ito pure colors, or frequencies. The intensity of each color implies

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC

87


a

that when added together would make up the original distribution. In applying this analysis to the con duction of heat around the ring, Fouri er reasoned that the greater the num ber of periods of a sinusoidal compo nent, the more rapidly it will decay. One can follow his reasoning by ex amining the relation between the fun

90

270

b
o

180

damental and the second harmonic

(L' M!.

;

W

,,,,.,,

','.'

" ... ...) li.u ,.

90

L . .,. ..... ""." ;,.." . . .. .... . ..180 .;o-L.
,.

..

·· ".." .....'".". >4,,"'"

POSITION

"".1 ., 270

w,· ... . .. · ,·

",

... . ..

.

· ·............ .·.·

,

360

,J

of the temperature distribution. The temperature of the second harmonic varies from hot to cool twice around the circumference of the ring, where as the fundamental varies only once. Therefore, the distance that the heat must travel from hot peak to cool trough is only half as far for the sec ond harmonic as it is for the fun damental. Furthermore, the tempera ture gradient in the second harmonic is twice as steep as it is in the funda mental variation. Because twice the heat flow occupies half the distance,

c

UJ 0::: :::> « 0::: UJ Q.

t

:::E
UJ

0

180 POSITION d

360

the second harmonic will die out four times faster than will the fundamental. Higher harmonics will decay even more rapidly. Hence, it is only a single sinusoidal distribution of the funda mental variation that persists as the temperature of the ring approaches equilibrium. Fourier believed that the evolution over time of any initial heat distribution could be computed by his technique.

adhered. In the early 19th century, many extraordinary Parisian mathema ticians, including Lagrange, Laplace, Legendre, Biot and Poisson, could not accept Fourier's claim that any ini tial temperature distribution could be decomposed into a simple arithmetic sum that consisted of a fundamental variation and its higher-frequency har monics. Leonhard Euler also found fault with Fourier's ideas, although he had already proposed that some

F

mathematical theories to which his contemporaries adamantly ourier's analysis challenged the

180 POSITION

TEMPERATURE OF AN IRON RING was one of the first phenomena analyzed by Fourier's technique. One distribution of heat around a ring is shown (a); bright· er color represents hotter areas. To be gin the analysis, the ring is "uncoiled"

e

t
UJ

(b),

and the temperature is measured at

every point, yielding a temperature distribution around the circumference (e). Then the temperature distribution is decomposed into many sinusoidal curves having one, two, three or more cycles ed together






UJ

(d).

When 16 of the curves are simply add_______---, :-::-::-_____________

o ______
88

(solid line in e),

they yield a

180 POSITION

360

good approximation of the original temperature distribution

(broken line in e).

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC


0-

180
POSITION

180
POSITION

360

180
POSITION

360

CONDUCTION OF HEAT through an iron ring causes the tem perature distribution to change over time

(left).

ic

Just as the

ic (right), are shown. Fourier determined that the second har monic will decay four times faster than the first harmonic and higher harmonics will decay even faster. Because the first harmonic persists the longest, the overall temperature distri bution approaches the sinusoidal shape of the first harmonic.

(middle),

and the two-cycle distribution, or second harmon

temperature distribution at any instant can be described as a series of sinusoidal curves, the evolution of a temperature distribution over time can be described in terms of changes in the sinusoids. The one-cycle distribution, or first harmon·

functions could be represented as a sum of sine functions. And so when Fourier made this claim at a meeting of the French Academy of Sciences, Lagrange stood up and held it to be impossible. Even under these circumstances the Academy could not ignore the signif icance of Fourier's results, and it awarded him a prize for his mathe matical theory of the laws of heat propagation and his comparison of the results of his theory with pre cise experiments. The award was an nounced, however, with the following caveat: "The novelty of the subject, together with its importance, has de cided us to award the prize, while nevertheless observing that the man ner in which the author arrives at his equations is not without difficulties, and that his analysiS for integrating them still leaves something to be de sired both as to generality and even as to rigor." The great uneasiness with which Fourier's colleagues regarded his work caused its publication to be delayed until 18 15 . In fact, it was not com pletely described until the 1822 publi cation of his book,

a discontinuous function described as resulting from a combination of or dinary, continuous functions, such as linear, quadratic, exponential and si nusoidal functions. If Fourier was cor rect, however, a sum of an infinite number of sinusoids would converge to represent accurately a function with jumps, even with many jumps. At the time this seemed patently absurd. In spite of these objections many workers, including the mathemati cian Sophie Germain and the engineer Claude Navier, began extending Fou rier's work beyond the field of heat analysis. Yet mathematicians contin ued to be plagued by the question of whether a sum of sinusoidal functions would converge to represent a discon tinuous function accurately.

the ebb and flow of the tides. Lord Kelvin had invented an analogue com puter for providing information about the tides to the crews of merchant and naval vessels. First sets of am plitudes and phases were calculat ed manually from a record of tidal heights and corresponding times that had been painstakingly measured dur ing the course of a year in a particu lar harbor. Each amplitude and phase repre sented a sinusoidal component of the tidal-height function and revealed one of the periodiC contributions to the tide. Then the results were fed into Lord Kelvin's computer, which synthe sized a curve predicting the heights of the tide for the next year. Tidal curves were soon produced for ports all over the world. It seemed obvious that a tide

sider the classic example: will you ever arrive at a wall if with each step you travel half of the remaining distance? The first step will bring your toe to the halfway mark, the second, three quar ters of the way, and at the end of the fifth step you are almost 97 percent of the way there. Clearly this is almost as good as reaching the wall, but no matter how many steps you take, you will never quite reach it. You could prove mathematically, however, that you would ultimately get closer to the wall than any distance nominated in advance. (The demonstration is equiv alent to showing that the sum of a half, a fourth, an eighth, a 16th and so on approaches one.) The question of the convergence of Fourier series emerged again late in the 19th century in efforts to predict

T

es whenever an infinite series of he questionto be added up. Con numbers is of convergence aris

predicting machine with more parts could process more amplitudes and phases and thus would make better predictions. This turned out not to be completely true if the mathematical function to be processed contained a steep jump, that is, it described an essentially discontinuous function. Suppose such a function was re duced into a small set of amplitudes and phases-that is, just a few Fourier coefficients. The original function can then be reconstructed from the sinus oidal components corresponding to the coefficients, and the error between the original function and the recon structed function can be measured at each point. The error-finding proce dure is repeated, each time computing more coefficients and incorporating them into the reconstruction. In every case, the value of the maximum error does not diminish. On the other hand,

ory of Heat.

The Analytical The
approach function

Objections apparently

to Fourier's discontinuous

focused on the proposition that an could be represented by a sum of sinusoidal functions, which are con tinuous. Discontinuous functions de scribe broken curves or lines. For in stance, a function called the Heaviside step function is zero on the left and jumps to one on the right. (Such a function can describe the flow of cur rent when a switch is turned on.) Fou rier's contemporaries had never seen

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC

89


SCHWINGER

S EIN S T EIN'

WINFR EE

BI OLOGICAL
STRUCTURES AN

GO RDO N

WAT ERMA N

ANIMAL NAV I

HOBSON

S LEE P

FRIEDMAN

SUN AND

EA

S I EV ER

SA N D

HUB E L

EYE,

BRAI N, A

GO U L D GO U LD.

THE H O N EY

MENARD

ISLANDS
ATK INS

MOLECU LES
MES EA MORRIS ON

PO WER S OF

T


Join the most exciting book club in science where the only commitment you make is to reward your own cur iosity
This is the Scientific American Library, where the mysteries of the human and natural worlds unfold before your eyes in lavishly illustrated volumes you'll want to read and refer to for years to come.
Here's why 75,000 readers like you have already become members of the Scientific American Library
· 15-day Free Trial You will receive a new volume to examine approximately every two months. (Your first volume is shipped immediately via UP S at no extra charge.) You have 15 days to look over each volume. If you decide to keep it, send in your payment. Other wise, simply return the book in the same shipping carton and we'll immediately credit your account. · No Prepayment and No Minimum Purchase Send no money in advance. You pay only for the volumes you keep. You may choose as many or as few volumes as you like: there is no yearly minimum. · Low Members' Price You may purchase any of these handsome volumes for only $24.9525% below the bookstore price. · No Strings, No Risk,

A one-of-a-kind book club for those who take science seriously.

Each volume in the Scientific American Library was created exclusively for this series, written by preeminent scientists-including many Nobel laureates-who describe their work in an exciting, accessible way.
Choose your premier volume
· Extinction, Steven M. Stanley. In this geologic detective story, a renowned paleobiologist ex plores the causes of the mass extinctions that have periodi cally decimated thousands of thriving species. "Smack up to date scientifically ...the best over view of the changing relationship between the physical environment and life .... Enthralling ...." -John Gribbon, New Scientist

No Commitment You may cancel your obligation-free membership at any time for any reason, without notice, without penalty.
CREDITS (top to bottom)-From Sun and Earth, the "ballerina skirt" model of the solar-imerpianetary current sheet. From The Science of Structures and Materials, bubble formation due to the surface tension 0/ a liquid; courtesy Stephen Dalton, NHPA. From The Honey Bee, drone. queen, and worker; drawing by John Hatzakis. From Einstein's Legacy, Chaplin and Einstein in Hollywood, 1931; COllrtesy The Granger Collection. From Drugs and The Brain. major anatomical divisions of the brain; painting by Bill Andrews.

· Molecules, P. w. Atkins. A distin. guished chemist takes us below the surface of visible reality to show how molecules make soap "soapy;' deter mine the taste of barbecue, and give rise to the brilliant colors of fall foliage. "Undoubtedly the most beautiful chemistry book ever written .... It is more than a book of facts; Molecules is a work of art." -John Emsley, New Scientist

Join us. You have nothing to lose. And a world of scientific discovery awaits.
To become a member of the Scientific American Library, simply detach and mail the facing reply card. Or write to: Scientific American Library, P.O. Box 646, Holmes, PA 19043
© 1989 SCIENTIFIC AMERICAN, INC


the error becomes confined to a re gion that gradually shrinks around the discontinuity, so that ultimately at any given point the error approaches zero. Josiah Willard Gibbs of Yale Universi ty confirmed this result theoretically in 1899. Fourier analysis is still not appli cable to unusual functions, such as those possessing an infinite number of infinite jumps in a finite interval. By and large, however, a Fourier se ries will converge if its original func tion represents the measurement of a physical quantity. Vast areas of new mathematics have been developed from investigations of whether the Fourier series of a partic ular function converges. One exam ple is the theory of generalized func tions, which is associated with George F.

basis for the Heaviside step function and the Dirac delta function; the latter describes a unit of area concentrat ed at a point. The theory enabled the Fourier transform to be applied to solve equations that involved such in tuitively accepted concepts as point mass, point charge, magnetic dipoles and the concentration of a load on a beam. ter almost two centuries of de velopment, the theory behind the Fourier transform is firm ly established and well understood. As we have seen, Fourier analysis breaks down a function in space or time into sinusoidal components that have varying frequencies, amplitudes and phases. The Fourier transform is a function that represents the ampli tude and phase at each frequency. The transform can be derived by two dif ferent mathematical methods, one ap-

plied when the original function is continuous and the other when it con sists of many discrete measurements. If the function is made from a list of values at discrete intervals, it can be broken down into a series of sinusoi dal functions at discrete frequencies, which range from a lowest frequen cy, the fundamental, through a series of frequencies that are two, three or more times the fundamental. Such a sum of sinusoids is called the Fou rier series. If the original function provides a value for every real number, then it is decomposed into sinusoidal func tions at all frequencies, which are combined by means of an operation called the Fourier integral. The trans form is neither the series nor the inte gral. In the case of the discrete func tion, it is the frequency-dependent list of amplitudes and phases appearing in the Fourier series; in the case of the



j.

Temple of England, Jan G. Mikusin

ski of Poland and Laurent Schwartz of France. It established in 1945 a firm

FERREL TIDE PREDICTOR, an analogue computer built in the late 19th century, performed Fourier synthesis to forecast the ebb and flow of the tides. Data that were collected on tidal heights at a particular harbor could be reduced by hand calcu lations into a set of numbers, each one representing a period-

ic contribution to the tide, such as the gravitational pull of the moon. The numbers for a specific port could then be fed into the Ferrel Tide Predictor by twisting knobs on the back of the machine (left). When a time was set on the front of the machine

(right), the predicted height of the tide could be read off a dial.

92

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC


unbroken function, it is the function of frequency that results when the Fourier integral is evaluated. Regardless of the manner in which the transform is derived, it is neces sary to specify two numbers at each frequency. These might be the ampli tude and phase; however, other num ber pairs could encode the same infor mation. These values can be expressed as a single complex number. (A com plex number is the sum of one real number and another real number mul tiplied by the square root of negative one. This representation is very popu ) lar because it invites the use of com plex algebra. The theory of complex al gebra and the Fourier transform have become indispensable in the numeri cal calculations needed to design elec trical circuits, analyze mechanical vi brations and study wave propagation. Representing an original function by its complex Fourier transform leads to computational advantages. A typi cal problem is to ascertain the cur rent that flows when a known volt age is applied to a circuit. The direct method involves solving a differential equation that relates the voltage and current functions. The Fourier trans forms of the voltage and current func tion, in contrast, can be related by an equation whose solution is trivial.

Just how much arithmetic had to be performed depended on the num ber of data points needed to describe the wave. The number of additions was comparable to the number of points, and the number of multiplica tions equaled the number of points squared. For example, analyzing a wave specified by 1,000 points taken at regular intervals required on the order of 1,000 additions and one mil lion multiplications. Such calculations became more fea sible as computers and programs were developed to implement new methods of Fourier analysis. One was developed in 1965 by James W. Cooley of illM's Thomas]. Watson Research Center and John W. Tukey of the Bell Telephone Laboratories in Murray Hill, N. Their work led to the development ]. of a program known as the fast Fourier transform. The fast Fourier transform saves time by decreasing the number of multiplications needed to analyze a curve. At the time, the amount of mul tiplication was emphasized simply be cause multiplication was slow with respect to other computer operations, such as addition and fetching and storing data. The fast Fourier transform divides a curve into a large number of equally spaced samples. The number of multi plications needed to analyze a curve

back long before Cooley and Tukey's work to the astronomer Carl Fried rich Gauss. Gauss wanted to calculate asteroidal and cometary orbits from only a few observations. After discov ering a solution, he found a way to reduce the complexity of the calcula tions based on principles similar to those of the fast Fourier transform. In an 1805 paper describing the work, Gauss wrote: "Experience will teach the user that this method will greatly lessen the tedium of mechanical cal culation." Thus, the challenge of celes tial motions not only gave us calculus and the three laws of motion but also stimulated the discovery of a modern computing tool.

come comfortable with the represen tation of sinusoids. The convenience of representing the Fourier transform as a complex function lets us forget that the underlying sinusoidal com ponents are real and not necessarily complex. This habit of thought has obscured the significance of and re tarded the adoption of a transform similar to Fourier's that was conceived by Ralph V. L. Hartley in 1942. Working in the research laboratory of the Western Electric Company, Hart ley directed the early development of radio receivers for a transatlantic ra diotelephone and invented the Hartley oscillating circuit. During World War I Hartley investigated how a listener, through mechanisms in the ear and brain, perceives the direction from which a sound emanates. Working at Bell Laboratories after the war, Hartley was the first to formulate an impor tant principle of information technol ogy that states that the total amount of information a system can transmit is proportional to the product of the frequency range the system transmits and the time during which the system is available for transmission. In 1929 group because of illness. As his health improved he devoted himself to the theoretical studies that led to the Hartley transform. . The Hartley transform is an alterna tive means of analyzing a given func tion in terms of sinusoids. It differs from the Fourier transform in a rather simple manner. Whereas the Fourier transform involves real and imaginary numbers and a complex sum of sinus oidal functions, the Hartley transform involves only real numbers and a real sum of sinusoidal functions. In 1984 I developed an algorithm

P

trinated with complex algebra hysicists and engineers, indoc early in their education, have be

ing techniques of moving free oday the study forFourier trans ly between functions and their trans forms. Analytical methods can be ap plied to evaluate the Fourier integral and produce the transform. Although these methods may be difficult for ordinary practitioners, many Fourier integrals have been found and are list ed in tables of reference. These meth ods can be supplemented by learning a handful of theorems pertaining to transforms. With the aid of these the orems more or less complicated wave forms can be handled by reduction to simpler components. Fortunately numerical methods are available for computing Fourier trans forms of functions whose forms are based on experimental data or whose Fourier integrals are not easily evalu ated and are not found in tables. Be fore electronic computers, numerical calculation of a transform was rather tedious, because such a large amount of arithmetic had to be performed with paper and pencil. The time re quired could be reduced somewhat by forms and schedules that guided in vestigators through the calculations, but the labor involved could still be daunting.

T

forms consists largely of acquir

decreases by one half when the num ber of samples is halved. For exam ple, a 16-sample curve would ordinar ily take 16 squared, or 25 6, multipli cations. But suppose the curve was halved into two pieces of eight points each. The number of multiplications needed to analyze each segment is eight squared, or 64. For the two seg ments the total is 128, or half the number required before. If halving the given sequence yields a twofold gain, why not continue with the strategy? Continued subdivision leaves eight irreducible pieces of two points each. The Fourier transforms of ed without any multiplic;ations, but multiplication is required in the proc ess of combining the two-point trans forms to construct the whole trans form. First, eight two-point transforms are combined into four four-point transforms, then into two eight-point transforms and finally into the desired 16-point transform. These three stag es that combine the pieces each call for 16 multiplications, and so the total number of multiplications will be 48, which is 3/ 16 of the original 25 6. This strategy for reducing the num ber of computations can be traced

these two-point pieces can be comput- . Hartley gave up the directIon of his

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC

93


THE FOURIER AND HARTlEY TRANSFORMS
The Fourier and Hartley transforms convert functions of time into functions of frequency that encode phase and amplitude information. The graphs below represent the unbroken function g(t) and the discrete function g(T), where t is time and T is a number designated at each data point.

for a fast Hartley transform. The dif ference in computation time b.etween the fast Hartley transform and the fast Fourier transform depends on the computer and the programming lan guage and style. If these factors are kept constant and no oversights are made in the programming, programs for the fast Hartley transform run

z o

(3 z
w..

z o

more quickly than those for the fast Fourier transform. Although both pro grams need the same amount of time to retrieve the data, provide for trigo nometric functions and perform other preliminaries, the time spent on the stages of the Hartley transform is half that required by the Fourier. It was not clear at first, however, that the Hartley transform provided the same information as the Fourier trans form. Therefore, when the first pro grams were developed for comput ing the Hartley transform, an extra step was provided to convert it to the more familiar Fourier. Workers soon realized, however, that intensi ties and phases can be deduced direct ly from the Hartley transform without the need for the additional step. Fur ther reflection revealed that either kind of transform furnishes at each frequency a pair of numbers that rep resents a physical oscillation in ampli tude and phase. Yet another reservation about the Hartley transform was that the Fou rier transform described physical phe nomena more naturally. Many phe nomena, such as the response of a simple system to vibration, are com monly described by a complex sum of sinusoidal functions, which is the hallmark of the Fourier transform. It might seem, therefore, that Fourier transforms are more suitable for de scribing the behavior of nature. Such a conclusion is in fact more
a

::;)

(3 z

::;)

w..

z

UJ

UJ

o cr: co Z ::;)

cr: U Vl

t;:;
o

Both functions start at zero, jump to a positive value and then decay exponentially. The definition of the Fourier transform for the unbroken function is an infinite integral, F(f), whereas the definition for the discrete function is a finite sum, F(v).

F(f) =

00

·00

f g(t)

(cos

2TTft -I sin 2TTft) dt

F( v)

=

J..
n

n-l

L g(t) (cos 2TTVT -I sin 2TTVT) T=O

Here f is frequency, v is related to frequency, n is the total number of samples and 1 is the imaginary number equal to the square root of -1. The integral representation is more suited to theoretical manipulations, whereas the finite-sum representation is more suited to computer applications. The Hartley transform and discrete Hartley transform have similar definitions.

H(f) =

-00

j

g(t) (cos 2TTft

+ sin

2TTft) dt

H( v)

=

J..
n

n-l

L g(t) (cos 2TTVT T=O

+ sin

2TTVT)

Even though the only notational difference between the Fourier and Hartley definitions is a factor -I in front of the sine function, the fact that the Fourier transform has real and imaginary parts makes the representations of the Fourier and Hartley transforms quite different. The discrete Fourier and discrete Hartley transforms have essentially the same shape as their unbroken counterparts.

reflection of our mathematical up

bringing than it is of nature. After all, when physical objects are measured,
FREQUENCY FREQUENCY

they provide data in real numbers, not complex ones. The advent of the fast Hartley trans form has made obsolete certain adap tations of the fast Fourier transform, such as those used for eliminating noise from digitally recorded music. These adaptations require two pro grams: one of them transforms real functions into the complex Fourier domain, whereas the other converts complex functions from the Fourier

Although the graphs look different, the phase and amplitude information that can be deduced from the Fourier and Hartley transforms is the same, as shown below.

FREQUENCY Fourier amplitude is the square root of the sum of the squares of the real and imaginary parts. Hartley amplitude is the square root of the sum of the squares of H(-v) and H(v). Fourier phase is the arc tangent of the imaginary part divided by the real part, and Hartley phase is 45 degrees added to the arc tangent of H(-v) divided by H(v).

domain into real functions. High-fre quency noise in digitally recorded mu sic can be eliminated by filtering out portions of the transform produced by the first program. The second pro gram then converts the changed trans-

94

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC


FOURIER ANALYSIS can transform X-ray diffraction patterns into molecular models_ X rays scatter off the electrons in a virus, for example, to produce patterns on film (left)- These patterns represent part of the Fourier transform of the virus's

molecular structure. If the process of transformation is re versed, the distribution of electrons, and therefore atoms, can be deduced (middle)_ From these distributions, models of the virus are made (right). Here colors indicate different proteins.

form back into an improved musical signal. Although these ingenious pro grams run individually at speeds rival ing the fast Hartley transform, a sin gle Hartley program suffices for both transforming a real function into a Hartley transform and converting the transform, after the desired filtering, back to a real function. Therefore, ex tra computer memory for storing two programs is not required.

jects taken in space by means of Fou rier analysis. Planetary probes and earth-orbiting satellites transmit imag es to the earth as a series of radio im pulses. A computer transforms these impulses by Fourier techniques. The computer then adjusts various com ponents of each transform to enhance certain features and remove others much as noise can be removed from the Fourier transforms of recorded music. Finally, the altered data are converted back to reconstruct the im

diffused into general antenna technol ogy. NASA commended the maps of the sun for contributing to the safety of the lunar astronauts. I have also applied the Hartley trans form to other studies. Recently my colleague John D. Villasenor and I de scribed an optical method for finding the Hartley transform, a development that enables Fourier phase and ampli tude to be encoded in a single real image. We have also developed a de vice that constructs the Hartley trans form using microwaves. I am now writ ing papers on solar physics in which transform techniques underlie new ways of analyzing data from sunspot counts and from the thickness of sedi mentary layers on the earth. The wide use of Fourier's meth od and related analytical techniques makes what Lord Kelvin said in 1867 just as true today: "Fourier's theorem is not only one of the most beautiful results of modern analysis, but it may be said to furnish an indispensable instrument in the treatment of nearly every recondite question in modern physics."

with

I

rier and Hartley transforms have n the most general terms, Fou been applied in fields that contend fluctuating phenomena. Their

age. This process can sharpen focus, filter out background fog and change contrast. The Fourier transform is also valu able in plasma phYSiCS, semiconduc tor phYSiCS, microwave acoustics, seis mography, oceanography, radar map ping and medical imaging. Among the many applications in chemistry is the use of the Fourier-transform spec trometer for chemical analysis. Fourier analysis has proved valuable in my own work in two-dimension al imaging. In 195 6 I stumbled on a "projection slice" theorem that yield ed a way to reconstruct images from strip integrals, a problem now wide ly known as tomographic reconstruc tion. Later, I hit on the "modified back projection" algorithm, now universally used in computer-assisted X-ray to mography, or CAT scanning. I was also interested in reconstruct ing images based on data from radio astronomy. I wanted to pinpoint sour ces of radio waves on the sun's sur face, so I applied transform methods to the design of a scanning radio tele scope that made daily microwave tem perature maps of the sun for 11 years. The methods led to the first antenna with a beam sharper than the resolu tion of the human eye and have since

field of application is thus very broad indeed. Many applications exist in biolo gy. In fact, the double-helix form of DNA was discovered in 1962 through X-ray diffraction techniques and Fou rier analysis. A beam of X rays was fo cused on a crystal of DNA strands, and the X rays were diffracted by the mol ecules of the DNA and recorded on film. This diffraction pattern provided the amplitude information of the crys tal structure's Fourier transform. The phase information, which the photo graphs alone did not provide, was de duced by comparing the DNA diffrac tion pattern with patterns produced by similar chemicals. From the X-ray intensity and phase information in the Fourier transform, biologists worked back to a crystal structure-the origi nal function. In recent years, X-ray dif fraction studies combined with such "reverse" Fourier analysis have re vealed the structure of many other biological molecules and more com plex structures, such as viruses. The National Aeronautics and Space Administration improves the clarity and detail of pictures of celestial ob-

RJRTHER RFADING
JOSEPH FOURIER: THE MAN AND THE PHYSICIST.

John

Herviel.

Clarendon

Press, 1975.
THE FOURIER TRANSFORM AND ITS AP PLICATIONS, SECOND EDITION, REVISED.

Ronald N. Bracewell. McGraw-Hili Book Company, 1986.
THE HARTLEY TRANSFORM. Ronald N.

Bracewell.
1986.

Oxford

University

Press,

OPTICAL PHASE OBTAINED BY ANALOGUE HARTLEY TRANSFORMATION. John Vil

lasenor and R. N. Bracewell in cember

Nature,

Vol. 330, No. 6150, pages 735-737; De

24, 1987.

SCIENTIFIC AMERICAN June 1989
© 1989 SCIENTIFIC AMERICAN, INC

95