3084 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 42, NO.
11, NOVEMBER 1994
The Fractional Fourier Transform
and Time-Frequency Representations
Luis B. Almeida, Member, IEEE
Abstract-The fractional Fourier transform (FRFT), which is a
generalization of the classical Fourier transform, was introduced
a number of years ago in the mathematics literature but appears
to have remained largely unknown to the signal processing
community, to which it may, however, be potentially useful. The
FRFT depends on a parameter cy and can be interpreted as a
rotation by an angle a in the time-frequency plane. An FRFT
with a = n-/2 corresponds to the classical Fourier transform, and
an FRFT with Q = 0 corresponds to the identity operator. On the
other hand, the angles of successively performed FRFT’s simply
add up, as do the angles of successive rotations. The FRFT of a
signal can also be interpreted as a decomposition of the signal in
terms of chirps.
In this paper, we briefly introduce the FRFT and a number
of its properties and then present some new results: the in-
terpretation as a rotation in the time-frequency plane, and the
FRFT’s relationships with time-frequency representations such
as the Wigner distribution, the ambiguity function, the short-
time Fourier transform and the spectrogram. These relationships
have a very simple and natural form and support the FRFT’s
interpretation as a rotation operator. Examples of FRFT’s of
some simple signals are given. An example of the application of
the FRFT is also given.
I. INTRODUCTION
F OURIER analysis is one of the most frequently used
tools in signal processing and is frequently used in many
other scientific disciplines. Besides the Fourier transform itself,
time-frequency representations (TFR’s) of signals, such as the
Wigner distribution (WD) [ 11, [3], the ambiguity function [ 11,
[3] the short-time Fourier transform (STFT) [1]-[3] and the
spectrogram [ 11-[3] are often used, e.g., in speech processing,
radar, or quantum physics.
In the mathematics literature, a generalization of the Fourier
transform known as the fractional Fourier transform (FRFT),
was proposed some years ago [4], [SI. Although potentially
useful for signal processing applications, this transform ap-
pears to have remained largely unknown to the signal process-
ing community. Recently, the FRFT has been independently
reinvented by a number of researchers, including this author
[6]-[9]. New results, concerning the FRFT’s interpretation as
a rotation in the time-frequency plane, and its relationships
with time-frequency transforms, have also been independently
found, more or less simultaneously, by this author and by
some of the above referenced ones.
Manuscript received December 6 , 1992; revised January i8, 1994. The
associate editor coordinating the review of this paper and approving it for
publication was Prof. K. M. Buckley.
The author is with INESCAST, Instituto Superior Tecnico, Lisbon, Portugal.
IEEE Log Number 9403739.
1053-587X/94$04.00 0 1994 IEEE
The purpose of this paper is twofold: First, to briefly
introduce the fractional Fourier transform and its main
properties and, second, to present the new results, including
the FRFT’s interpretation as a rotation in the time-frequency
plane and the simple relationships that the FRFT has
with several time-frequency re resentations that support
the interpretation as a rotation. We will also see that
computing the FRFT of a signal corresponds to expressing
it in terms of an orthonormal basis formed by chirps, i.e.,
complex exponentials with linearly varying instantaneous
frequencies.
This paper is organized as follows. Section I1 presents
the FRFT, emphasizing its interpretation as a rotation oper-
ator, lists some of its properties, and gives some examples.
Section I11 discusses the relationships with the Wigner dis-
tribution, the ambiguity function, and other quadratic TFR’s.
Section IV discusses the relationships with the short-time
Fourier transform and the spectrogram. Section V briefly
presents an example of an application. Section VI concludes
the paper.
Notes on the Formalism: We will represent by “j” the
imaginary unit and by a superscript asterisk * the complex
conjugation operation. We will often use square roots of
complex numbers, and we will reserve the square root symbol
‘‘J” to denote the square root whose argument lies in the
interval ] - 5 ~ / 2 , ~ / 2 ] .
11. THE TRANSFORM AND ITS PROPERTIES
A. Definition
In time-frequency representations, one normally uses a
plane with two orthogonal axes corresponding to time and
frequency, respectively (Fig. 1). If we consider a signal x ( t )
represented along the time axis and its Fourier transform X ( w )
represented along the frequency axis, we can view the Fourier
transform operator,’ which we shall designate by F, as a
change in the representation of the signal corresponding to
a counterclockwise axis rotation of ~ / rad.
2 This is consistent
with the result of the repeated application of the F operator,
‘We consider the Fourier transform defined as S(UJ) & = JTm~ ( t )
x p - J d f d t . In the mathematics literature, the Fourier transform is normally
defined with a plus sign in the exponent. We prefer to use the engineering
convention here, although this results in a number of sign differences between
the expressions of this paper and those of the references (which amounts to
a change in the sign of the angle ct in the expressions given ahead for the
FRFT).
ALMEIDA: FRACTIONAL FOURIER TRANSFORM 3085

Note that this kernel is continuous in a, in the generalized

function sense, even at multiples of 7 r , i.e.

lirn K , = K,, for integer n (7)

o+nri

The kernel has the following properties, which will be useful

later in this paper:

Fig. I . Time-frequency plane and a set of coordinates (U.v) rotated by an

angle c1 relative t o the original coordinates (t.-.).

given that FFz(t) = x ( - t ) and that two successive rotations

of 7r/2 of the t axis result in an axis directed along -t. The first three properties are trivial. The derivation of (1 1) is
In this context, one may ask which linear operator would rather long and will not be given here. Property (12) can be
correspond to a rotation by an angle N that is not a multiple derived using (S), (9), (1 I), and the definition of K , for Q = 0.
of 7 r / 2 , or equivalently, what would be a representation of This latter property means that the kernel functions K,(t, U ) ,
the signal along an axis U (Fig. 1) making an angle N with taken as functions o f t with parameter U , form an orthonormal
the time axis. Let us assume, for the moment, that such an set. Given (S), the same can be said if we take U as a variable
operator exists, and let us represent it by RO. This operator and t as a parameter. We also note that for N = 7r/2, the kernel
should have the following properties: coincides with the kernel of the Fourier transform.
The fractional Fourier transform of a function 2, with an
1) Zero rotation: Ro = I (1) angle cy, is defined [4], [ 5 ] as the function R"z = X,, given
2) Consistency with Fourier transform: R"/2 = F (2)
3) Additivity of rotations: RaRO = R04-d (3) by (13) and (14), which appear at the bottom of the page.
The last equation shows that for angles that are not multiples
4) 27r rotation: R2" = 1 (4)
of 7 r , the computation of the fractional Fourier transform
Note that property 4 is a consequence of properties 2 and 3
corresponds to the following steps:
and of the fact that four successive applications of the Fourier
transform correspond to the identity operator: 1) a product by a chirp
2 ) a Fourier transform (with its argument scaled by csccu)
R2" = R4("I2)= F4 = 1 (5) 3) another product by a chirp
The FRFT, which has these properties, is defined by means of 4) a product by a complex amplitude factor.
the transformation kernel2 Since chirps have constant magnitude, this immediately allows
us to make a rather general statement about the existence
K,(t. U)
JF of the transform. In fact, if x ( t ) is in L1, L2, or is a

=i
el- cot a - p t c s r a
generalized function, its product by the chirp is also in L1, L2,
if cy is not a multiple of 7r or is a generalized function, respectively. Therefore, in these
(6) situations, the FRFT of z ( t ) exists in the same conditions in
6(t - U ) if N is a multiple of 271.
6(t + U ) if (Y+T is a multiple of%. which its Fourier transform exists.
It is easy to see that the FRFT, as defined in (13), satisfies
The square root factor that precedes the exponential, in this definition, the properties that were discussed in the introduction. Property
can be written in a number of equivalent forms, some of which appear in the
1 holds directly by definition. Property 2 is a consequence of
references. Note, however, that 141 contained a flaw in the definition of this
factor, a correction having been made in [SI. Another useful form for this the fact that for cy = 7r/2, K , coincides with the kernel of
factor is J 2 ~ ~ ~ " / ( 7 r s i i i c r ) . the Fourier transform. Property 3 is easily derived, using (1 1).
3086 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 42, NO. 11, NOVEMBER 1994

In fact, if we define

Y(U) = X,(u)
Signal Fractional Founer transform with angle a

then ' '(I-')

X,(u-rcosa)e
2
I-ma
2
ma-lm m a

II I
4 /(X(I8,dlr
seca e
-I-iana

if a -TI 1 2
2

IS
l X a ( z ) e ' ~ ' " a dz if a - x I 2 is not a multiple of x

a multiple of x , the classical Fourier transform property

Finally, property 4 is a consequence of properties 2 and 3 as

had been noted in the introduction. From these properties, we
also easily conclude that the inverse of an FRFT with an angle
a is the FRFT with angle -a:
where p = arcm(c2 tana)
3=

z ( t )=
.II,X,(U)K_,(U,t)dU. (20)

From (20), we can see that the FRFT consists of expressing
~ ( ton) a basis formed by the set of functions K-,(u,t)
(with u acting as a parameter for spanning the set of basis
functions). This basis is orthonormal, according to (12). The
basis functions are chirps, i.e., complex exponentials with
linear frequency modulation. For different values of U , they
only differ by a time shift and by a phase factor that depends
on U :
K,(t,u) = e-+ano K,(t - useca.0).
An interesting issue is whether this is the only transform that
satisfies properties 1-4. It is easy to see that if we would
multiply the transform by e4jna with integer n, the four
properties would still hold. However, some of the results
presented below would become more complex, and therefore,
the FRFT appears to be the most natural one among these.
We do not know if there is still any other transform, besides
these, obeying the four properties. However, most probably, if
such a transform exists, its relationships with time-frequency
distributions will not be as simple as those derived below for
z ( t ) is even, X,(u) is also even; if z ( t ) is odd, X,(u) is
also odd.
The scaling property, although somewhat complex, is im-
portant since it shows the effect of a change of units (or a
scaling) of the independent variable t. In the classical Fourier
transform, the effect of such a change is only a corresponding
change of units (or scaling) of the frequency variable and a
scaling of the amplitude. In the FRFT, the effects are a scaling
of the u variable by sin p / ( c sin a ) , a (complex) amplitude
(21) scaling, a product by a chirp, and most important, a change
in the angle at which we compute the transform from (Y to
,O = arctan(c2 t a n a ) . This angle change can be understood
if we think of contracting the time axis, in the time-frequency
plane, by a factor c. As we know from the classical Fourier
transform, we will then also have to expand the frequency axis
by c. With these two operations, the axis along which we were
computing the transform, which was originally at an angle Q
with the time axis, will move to a new position at the angle
,O given above.
The Parseval relation
W 00

the FRFT.
In summary, the FRFT is a linear transform, continuous
in the angle a, which satisfies the basic conditions for being
1, z(t)y*(t)dt =
.I,
is easily derived by expressing x ( t ) in the left-hand side of
X,(u)Y,*(u) du (22)

interpretable as a rotation in the time-frequency plane. this equation as the inverse transform of X , (U ) and then using
(9). A consequence of this equality is the energy-preserving
property of the FRFT
B. Further Properties
oc

L L
DCI

In Table I, we list a number of additional useful properties Ix(t)I2dt = I X , ( U ) ~du.

of the FRFT, which are extensions of the corresponding
properties of the Fourier transform. Their proofs, except for
that of the scaling property (row 8), can be found in [4] and
[ 5 ] . The proof of the scaling property is given in Appendix
A.3 Property 6 of the table is naturally subject to the existence
of z ( t ) / t itself. Property 7 leads to the parity property: If
To our knowledge, this property was first presented in [IO] and [ I l l . Its
derivation was not given there, however, for lack of space, and this is why
we give it in this paper.
~ (23)
The Parseval relation and the energy-preserving property can
also be viewed as consequences of the fact that the FRFT is
based on an set of orthonormal basis functions.
Due to the energy-preserving property of the Fourier trans-
form, the squared magnitude of the Fourier transform of a
signal ~ X ( L Jis ) often
~ ~ called the energy spectrum of the
signal and is interpreted as the distribution of the signal's
energy among the different frequencies ,jut. Although it is less
ALMEIDA: FRACTIONAL FOURIER TRANSFORM 3087

intuitive, ( 3 8 ) also allows us to call IX,(U)~*the fractional TABLE I1

OF SOMECOMMON
TRANSFORMS SIGNALS
energy spectrum of the signal z ( t ) , with angle a , and to
interpret it as the distribution of the signal's energy among Fractional Fourier transform with angle a
i"*
Signal
the different chirps K-,(t, U ) . 6(r-r) ,i~mla-,ur,ca
1 if a is not a multiple of x
If z ( t ) is real, X,(u) enjoys one further property:
? I
x-,( U ) = x; (U) (24)
if a - x I 2 is not a multiple of x

which is a consequence of (9). Note, however, that the

symmetry properties of the Fourier transform for even or odd
real sequences do not extend to the FRIT with arbitrary angle
a since the FRFT of a real function is not in general Her-
mitean. Another important property of the Fourier transform
that does not extend in a simple way to the FRFT is the
convolution theorem.

C. Examples
Table I1 gives the FRFT's of a number of common signals.
An expression for the FRFT of a rectangle can also be
obtained, but that expression is rather complex and does not and therefore
-
appear to be particularly instructive. Instead of presenting
it here, we prefer to give a feeling of the behavior of the
transform by showing the variation of the transform of a
X ( t . w ) = 2P2JWt
.
e-2'JtL
1"'1:
--x
..(T)X;t(-Z

'5lllOCOS~+2JZtSlll~
+ 2tcosn)
rectangle [ z ( t )= 1 for It1 < 2; otherwise, x ( t ) = 01 with the
angle a (Figs. 2 and 3). In Fig. 2, we can see some evidence x K, (T. z ) P J Y ' dr d z (28)
oc
of nonuniform convergence towards x ( t ) as cy i 0, with a
phenomenon similar to the so-called "Gibbs effect" of the
classical Fourier transform.
-
-
1,x ; ( - z +
2e2J.c't 2tcosa)
e - ~ ~s it nzn c o s n + 2 j z t s i r i n

111. RELATIONSHIPS
WITH THE WIGNER DISTRIBUTION,
THE
x s_, oc
z(T)e-2JwTKa(T. z ) d7 d z . (29)
AMBIGUITY FUNCTION,AND OTHER QUADRATICTFRS The inner integral is easily computed using property 2 of
The Wigner distribution (WD) [I], [3] is a very important Table I, and we obtain
time-frequency analysis tool, with applications in such diverse - E

areas as radar and quantum mechanics. Apart from its direct

applications, its importance also stems from the fact that many
~ ( t , w=) 2e2jdt
.I,Xn(z + 2wsincy)X;(-z + 2tcoscu)
p - 2 j ( t 2 + w ' ) s i n a c o ~ a + 2 j z t s i n a - ~ jc zo s~a
time-frequency analysis tools can be expressed in terms of the dz.
WD. We will now see that the WD has a simple relationship
with the FRFT (although the derivation that leads to that result
is somewhat long). We now make the change of variable E =z + 2wsina, and
The WD of a signal :c is defined as come to
x
-
X ( t . w ) = 2e2JWt
.Ix Xn(~)X;(-e + 2tcosn + 2 w s i r i a )
e2j(w' - t 2 ) s i n e cos n+2je(t s i n a - w c o s n - 4 j w t sin2 a
dE.

which can also be expressed, by a suitable change of integra- (31)

tion variable, as
If we make, on the right-hand side. a change of variables
- x
from ( t . w ) to (u.,))),
corresponding to an axis rotation by N ,
X ( t ,w)= 2 e 2 ~ w t
.i, t ~ ) e - (~1 ~~. - (26)
. r ( ~ ) . c * ( 2- ~
as indicated in Fig. 1

Taking into account property 1 of Table I, we can express '11) = t cos cy + w sin (Y (32)
z*(2t - T ) as II = -tsin N +wcosa (33)
cc
z*(2t - T ) =
1,x;(-z+e-
2t c o s n )
2jtLsin n cos n + Z j z t sin n
K,(T. Z) (1%
we arrive at the following result after simplifying the
exponents:

X(t.w)= 2e2Juz' u ~ ) e - ~ j " d E.

Xa(~)X:(2- (34)
(27)
3088
alpha = 0.01
(a)
alpha = 0.2
alpha = Pii4
(e)
Fig. 2. FRFT of a rectangle, computed at various angles. Solid line: real part. Dashed line: imaginary part.
The right-hand side of this equation is the WD of X ,
computed with arguments ( U , U ) . The left-hand side is the WD
of II’ computed with arguments (t.U ) . The equation shows
that the WD of X , coincides with the WD of s if we take
into account the rotation that corresponds to the fact that we
are using different axes on the left- and right-hand sides of
the equation. This is equivalent to saying that the WD of
IEEE TRANSACTIONS O N SIGNAL PROCESSING, VOL. 42, NO. 11, NOVEMBER 1994
alpha = 0.05
(b)
alpha = 0.4
alpha = Pi/2
X , is the WD of z, rotated by an angle -cy, or that it is
simply the WD of z expressed in the new set of coordinates
(U.U). The fact that the FRFT induces a simple rotation of
the WD is remarkable and enhances our view of the FRFT as
a rotation of the axis along which the signal is represented.
On the other hand, this fact also shows that a rotated WD is
still a legal WD. It is well known that the WD has similar
ALMEIDA: FRACTIONAL FOURIER TRANSFORM 3089

Fig. 4. Representation of a swept-frequency system. The input signal is

first down modulated by an exponential with linearly varying instantaneous
frequency, then passed through a time-invariant filter and finally up-modulated
again. The constant c controls the sweep speed of the system.

(MSTFT), which consists simply of the normal STFT with a

suitable phase modification:

5
L(T)lli*(t - T)e-JdTd7. (37)
Fig. 3 . Variation of magnitude of the FRFI’ of a rectangle with the angle 0.
From (37), the MSTFT can also be computed as
- 1 “Jt
X(t.w) = - e - j T X ( v ) W * ( w- v)eJutdv (38)
properties relative to time and frequency. This result shows G

be obtained from their definitions in terms of the WD and - “

from the fact that the FFUT induces a rotation in the WD. For
example, it can be easily shown that the FRFT also induces a
X(t.w) = -e3r
1
Jz;;
ut

1, X,(z)W:(u - z)e-JZt’dz. (39)

simple rotation in the ambiguity function. The right-hand side of this equation is the MSTFT of X,
computed with window W , and with arguments (..TI). The
left-hand side is the MSTFT of 5 computed with window w
Iv. RELATIONSHIPS WITH THE SHORT-TIME and with arguments ( t . ~ )As . in the case of the WD, this
FOURIERTRANSFORM AND THE SPECTROGRAM equation shows that the MSTFT of X , is the same as the
The short-time Fourier transform (STFT) is another impor- MSTFT of z, again taking into account the rotation, i.e., that
tant time-frequency analysis tool. It is frequently used, e.g., it is simply a rotated version of the MSTFT of z or that it is
for speech processing. A related tool is the spectrogram, which the MSTFT of z expressed in the rotated axes ( U , w ) . Again, ~
corresponds to the squared magnitude of the STFT. The STFT this enhances our view of the FRFT as a rotation operator.
of a signal z ( t ) is defined as [l], [3] The spectrogram [l], [3] is simply the squared magnitude
of the STFT and, therefore, of the MSTFT as well. The
X ( t ,w ) = -
S”
-32
~ ( ~ ) w- * (7 t) e - J d Td.r ( 3 5 ) results we obtained on the MSTFT immediately lead us to
conclude that the effect of the fractional Fourier transform on
the spectrogram is identical to the one it has on the MSTFT:
where w ( t ) is a suitably chosen analysis win do^.^ The STFT
can also be computed in a simple way from the Fourier the spectrogram of X, computed with window W, is a rotated
version of the spectrogram of x computed with window IU.
transform of z ( t ) :
, rcc
X ( t , w ) = --&-jut X ( v ) W * ( w- v)eJutdv (36) V. APPLICATION
fi
In this section, we will briefly present, as an example, the
where X and W are the Fourier transforms of z and w , application of the FRFT to the study of swept-frequency filters.
respectively. This equation is similar to (35), except for the These filters are used, for example, in frequency analyzers for
presence of the exponential factor e - J w t . This is an asymmetry high-frequency signals. Swept-frequency filters are linear time-
between time and frequency that we wish to avoid since we varying systems that can be represented in the form shown in
want to deal with rotations in the time-frequency plane. We Fig. 4. They can also be represented by their time-varying
will therefore define a mod$ed short-time Fourier transform
‘When we compute the MSTFT of S,,we have to use the transformed
40ften, in the literature, the windows that are considered are real, and the window TI, as (39) shows. Note that even for a real-valued w, 147n will
complex conjugation of the window is not mentioned in the definition. Here, normally be complex. If we use the Gaussian window ~ ( t=) the
however, we prefer to adopt a more general viewpoint, allowing the window window’s FRFT is equal to to for any cr. Therefore, in this case, the
to be complex. FRFT induces a simple rotation of the MSTFT without any window change.
3090 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 42, NO. 1 I , NOVEMBER 1994

impulse response, h ( t , ~ which

), is the response at time t to VI. CONCLUSION
an input S ( t - T ) . It is easy to see that We have presented an extension of the Fourier transform,
h(t.i-) = e J I ( t L - T L S) ( t - i - ) (40) which is designated fractional Fourier transform. This linear
transform depends on a parameter a and can be interpreted
where g ( t ) is the impulse response of the shift-invariant filter as a rotation by an angle a in the time-frequency plane.
in Fig. 4. Particularly, the following hold:
The output of the swept-frequency filter is given by
1) When a = 7r/2, the FRFT coincides with the conven-
g(t) = 1 -,
x
x ( r ) h ( tr,) d T . (41)
tional Fourier transform.
2) When a = 0, the FRFT is the identity operator.
Let us compute the FRFT of y(t), choosing the angle a = 3) Two successive FRFTs with angles (IY. and p, respec-
-arccot c: tively, are equivalent to a single FRFT with an angle
cy + g.
On the other hand, the FRIT was shown to induce rotations
in various time-frequency transforms, including the Wigner
g(t - +j- cot a - j u t r s c a
distribution and the short-time Fourier transform, further en-
di-dt (42)
hancing its interpretation as a rotation operator.
The FRFT was also shown to correspond to a representation
of the signal on an orthonormal basis formed by chirps, which
are essentially shifted versions of one another. Finally, an
application example was given, showing how the use of the

=.J 1-jcotcu
2T 1, O0 .(+j-
or finally
where G is the (classical) Fourier transform of the impulse
response g. Therefore, G(u csccr) can be called the transfer
function of the swept-frequency filter in the FRFT domain.
The use of the FRFT and of this transfer function allows a
treatment of swept-frequency filters that is very similar to the
classical treatment of time-invariant filters with the Fourier
transform.
It is also worth mentioning that the FRFT is useful for of Table I. Let us consider two signals x and y, which
solving certain classes of differential equations that appear, are related by
for example, in quantum mechanics and optics, and that may
also be useful in signal processing applications (see [4], [SI,
[lo], [111).
cota
FRFT allows a treatment of swept-frequency filters that is very
similar to the classical treatment of shift-invariant filters with
the Fourier transform.
The work presented here opens several areas for further
research. Among these is the study of discrete versions of this
transform. On the other hand, the relationships between the
FRFT and time-frequency transforms suggest that it can be a
useful tool to further study the properties of these transforms
and to develop time-frequency transforms better suited to
specific applications.
APPENDIXA
We shall prove the scaling property given in row 8
Y(t) =

(53)

(54)
ALMEIDA: FRACTIONAL FOURIER TRANSFORM 309 1

and let us compute the FRFT of y: ACKNOWLEDGMENT

The author wishes to acknowledge useful suggestions from
A. F. Santos, J. F. Moura, J. S . Marques, M. D. Ortigueira,
(47) and the anonymous reviewers.
where we have used the integration variable t’ for convenience.
Making the change of variable t = ct‘, we obtain REFERENCES
[ I ] F. Hlawatsch and G. F. Bourdeaux-Bartels, “Linear and quadratic time-
frequency signal representations,” IEEE Signal Processing Mag., vol. 9,
no. 2, pp. 2 1 4 7 , Apr. 1992.
121 M. R. Portnoff, “Time-frequency representation of digital signals and
systems based on short-time fourier analysis,” IEEE Trans. Acoust.,
Speech, Signal Processing, vol. ASSP-28, no. 1, pp. 55-69, Feb.
1980.
[3] L. Cohen, “Time-frequency distributions-A review,” Proc. IEEE, vol.
If we define p = arc cot[(cot a ) / c 2 ] = arctan(c2 tarla), 77, no. 7, pp. 941-981, July 1989.
[4] V. Namias, “The fractional order Fourier transform and its application
we have (50)-(54), which appear at the bottom of the to quantum mechanics,” J. Insr. Math. Appl., vol. 25, pp. 241-265,
previous page. 1980.
[ 5 ] A. C. McBride and F. H. Kerr, “On Namias’ fractional Fourier trans-
forms,” /MA J. Appl. Math., vol. 39, pp. 159-175, 1987.
161 L. B. Almeida, “An introduction to the angular Fourier transform,” in
APPENDIXB Proc. 1993 IEEE Int. Con$ Acoust., Speech, Signal Processing (Min-
To obtain the relationship between the FRFT and the neapolis, MN), April 1993.
171 H. M. Ozatkas, B. Barshan, D. Mendlovic, and L. Onural, “Convolution,
MSTFT, we will start from (37), replacing Z ( T ) by the FRFT filtering, and multiplexing in fractional Fourier domains and their
of X , with angle --cy: relationship to chirp and wavelet transforms,” J. O p t Soc. Amer. A ,
in press.
[8] A. W. Lohmann, “Image rotation, wigner rotation and the fractional
Fourier transform,” J. Opr. Soc. Amer. A, vol. 10, pp. 2181-2186,
1993.
[9] 0. Seger, “Model building and restoration with applications in con-
focal microscopy,” Ph.D. Dissertation No. 301, Linkoping University,
Sweden, 1993.
[IO] T. Alieva, V. L6pez, and L. B. Almeida, “The angular Fourier transform
and wave propagation,” submitted to Phys. Rev. Lett.
x w * ( t - r)eCJd‘dr. (56) [ I I ] T. Alieva, V. L6pez, F. Aguill&L6pez, and L. B. Almeida, “The angular
Fourier transform in optical propagation problems,” J. Modern O p t , in
The inner integral is the complex conjugate of the FRFT with press.
angle a of w ( t - -r)eJwt,taken as a function of r and with
argument z . Using properties 1 and 2 of Table I, we can
conclude that this integral is equal to
W:( -z + t cos a + w sin a )
ej -sin n cos a + j z ( t sin a--w c o s a ) - j w t Sin’ a
(57)
and therefore
-
L,
03
1
X ( t , w ) = -e3
Li
X , ( z )W: ( - z +t cos +w sin a ) Luis B. Almeida (M’87) was born in Lisbon, Por-
&G tugal, in 1950. He received the degree in electri-
e j &$ s i n e c o s a + j z ( t s m a - - w c o s a ) - j w t sin’ adz. cal engineering from the Instituto Superior Tecnico
(IST), Lisbon, in 1972, and received the Ph.D.
degree in signal processing from the Universidade
(58)
Tecnica de Lisboa in 1983. He has been with the
IST since his graduation and is presently Associate
We now make, on the right-hand side, the change of variables
Professor of Signal Theory and of Neural Networks.
(32) and (33) from ( t ,w ) to ( U , U), and after simplification of He is head of the Neural Networks and Signal
the exponents, we obtain Processing Group at INESC, which is a research
institute associated with Portuguese universities and
communications companies.
Dr. Almeida is vice-president of the European Neural Network Society.
In 1985, he received a Senior Award from the IEEE for a paper on the
nonstationary modeling of voiced speech.