This is a Fast Fourier Transform program.
----begin documentation----
This 82u-file was prepared by Mikael Bonnier, http://www.df.lth.se.orbin.se/~mikaelb/.

			   *************************
			   * FFT.82G for the TI-82 *
			   *************************

ACKNOWLEDGMENT

FFT.82G is a group file containing the programs FFT.82P and XX.82P.  These
programs were written by Ralph Payne of the TI Graphics Team and are released
to the public domain.  You may copy and change these programs.

INTRODUCTION

This pair of "82" programs illustrate the use of the Fast Fourier Transform.
The Fast Fourier Transform (FFT) is a clever implementation of the Discrete
Fourier Transform and is well described in numerous publications. This
particular code implements a simple radix two algorithm with coefficient
recursion. It assumes that the real part of the time series is in L1 and
that the imaginary part is in L2. The process is done in place and
the positive frequencies are in the lower half of L1 and L2 and the negative
frequencies are in the upper half. The maximum length is 64.

HOW TO USE FFT

A driver program, FFTDEMO, prompts for the length (N), two frequencies, and a
noise amplitude. The program assumes that the transform length is one second,
so if the frequency is greater than one half the length aliasing will occur.
If the frequency is not an integer then scallop loss will occur, but may
not be obvious because the display is automatically scaled. The frequency
is of unit amplitude so the noise that is added is a multiple of one. This
value can be varied to see how the FFT can "pull" a signal out of the noise
The driver program sets the parametric mode and changes the window settings.
It also stores new equations to the first two parametric equation pairs.
Since the FFT expects input in L1 and L2 and operates in place, XX puts a
time series in L1 and 0's in L2 and copies L1 to L5. Since the imaginary
vector is 0, the transform will be symmetric. (If the input data is complex
the transform will not, in general, be symmetric.)  The magnitudes are stored
in L4. L4 and L5 are then scaled to fit on the Display. The upper half of the
display is the time series and the lower half is the frequency domain.
Program FFTDEMO, which calls FFT, takes about 25 seconds to execute a 32 
length transform and about 50 seconds to execute a 64 length transform.
----begin ascii----
\START82\
\COMMENT=Program file dated 08/13/96, 16:52
\NAME=FFT
\FILE=FFT.82P
:int (ln N/ln 2+.5)\->\M
:1\->\J
:For(I,1,N-1)
:If I>J
:Then
:\L1\(J)\->\T
:\L1\(I)\->\\L1\(J)
:T\->\\L1\(I)
:\L2\(J)\->\T
:\L2\(I)\->\\L2\(J)
:T\->\\L2\(I)
:End
:N/2\->\K
:While K<J
:J-K\->\J
:K/2\->\K
:End
:J+K\->\J
:End
:For(I,1,N,2)
:I+1\->\P
:\L1\(P)\->\U
:\L2\(P)\->\V
:\L1\(I)-U\->\\L1\(P)
:\L2\(I)-V\->\\L2\(P)
:\L1\(I)+U\->\\L1\(I)
:\L2\(I)+V\->\\L2\(I)
:End
:2\->\E
:For(L,2,M)
:E\->\S:2E\->\E
:1\->\U:0\->\V
:2*\pi\/E\->\A
:cos A\->\W
:sin A\->\X
:For(J,1,S)
:For(I,J,N,E)
:I+S\->\P
:U*\L1\(P)-V*\L2\(P)\->\B
:U*\L2\(P)+V*\L1\(P)\->\C
:\L1\(I)-B\->\\L1\(P)
:\L2\(I)-C\->\\L2\(P)
:\L1\(I)+B\->\\L1\(I)
:\L2\(I)+C\->\\L2\(I)
:End
:W*U-X*V\->\T
:W*V+X*U\->\V:T\->\U
:End
:End
:Return
\STOP82\
\START82\
\COMMENT=Program file dated 08/13/96, 16:52
\NAME=FFTDEMO
\FILE=FFTDEMO.82P
:Input "ENTER N ",N
:Input "ENTER FREQ1 ",F
:Input "ENTER FREQ2 ",G
:Input "ENTER NOISE ",D
:2\pi\/N\->\R
:seq(I*R,I,1,N,1)\->\\L3\
:sin (F*\L3\)+sin (G*\L3\)\->\\L2\
:cos (F*\L3\)+cos (G*\L3\)\->\\L1\
:seq(D*rand-D/2,I,1,N,1)\->\\L3\
:\L1\+\L3\\->\\L1\
:\L2\+\L3\\->\\L2\
:0*\L1\\->\\L2\
:\L1\\->\\L5\
:prgmFFT
:\sqrt\(\L1\\^2\+\L2\\^2\)\->\\L4\
:max(\L5\)\->\H
:min(\L5\)\->\G
:3/(H-G)*\L5\+((.5*H-3.5*G)/(H-G))\->\\L5\
:max(\L4\)\->\H
:min(\L4\)\->\G
:3.5/(H-G)*\L4\+((.5*G-4*H)/(H-G))\->\\L4\
:Param
:1\->\Tmin
:N\->\Tmax
:1\->\Tstep
:0\->\Xmin
:N-1\->\Xmax
:4\->\Xscl
:\(-)\4\->\Ymin
:4\->\Ymax
:1\->\Yscl
:"T-1\->\\X1t\
:"\L5\(T)\->\\Y1t\
:"T-1\->\\X2t\
:"\L4\(T)\->\\Y2t\
:DispGraph
\STOP82\
----begin uue----
begin 644 FFT.82G
M*BI423@R*BH:"@!'<F]U<&5D(&9I;&4@9&%T960@,#8O,#DO.3,L(#$P.C`X
M``#.0`86571``PL`G`$%1D94``````"<`9H!L1"^3H.^,G`Z-1$$33\Q!$H_
MTTDK,2M.<3$1/\Y);$H_SS]=`!!*$014/UT`$$D1!%T`$$H1/U0$70`021$_
M70$02A$$5#]=`1!)$01=`1!*$3]4!%T!$$D1/]0_3H,R!$L_T4MK2C]*<4L$
M2C]+@S($2S_4/TIP2P1*/]0_TTDK,2M.*S(1/TEP,010/UT`$%`1!%4_70$0
M4!$$5C]=`!!)$7%5!%T`$%`1/UT!$$D1<58$70$04!$_70`021%P501=`!!)
M$3]=`1!)$7!6!%T!$$D1/]0_,@1%/]-,*S(K31$_1013/C)%!$4_,015/C`$
M5C\R@JR#101!/\1!!%<_PD$$6#_32BLQ*U,1/]-)*THK3BM%$3])<%,$4#]5
M@ET`$%`1<5:"70$04!$$0C]5@ET!$%`1<%:"70`04!$$0S]=`!!)$7%"!%T`
M$%`1/UT!$$D1<4,$70$04!$_70`021%P0@1=`!!)$3]=`1!)$7!#!%T!$$D1
M/]0_5X)5<5B"5@14/U>"5G!8@E4$5CY4!%4_U#_4/]4+`(8!!49&5$1%34\`
MA@&$`=PJ14Y415(I3BDJ*TX_W"I%3E1%4BE&4D51,2DJ*T8_W"I%3E1%4BE&
M4D51,BDJ*T<_W"I%3E1%4BE.3TE312DJ*T0_,JR#3@12/R-)@E(K22LQ*TXK
M,1$$70(_PA!&@ET"$7#"$$>"70(1!%T!/\001H)=`A%PQ!!'@ET"$01=`#\C
M1(*K<42#,BM)*S$K3BLQ$01=`C]=`'!=`@1=`#]=`7!=`@1=`3\P@ET`!%T!
M/UT`!%T$/U]&1E0_O!!=``UP70$-$01=`S\97001!$@_&ET$$01'/S.#$$AQ
M1Q&"701P$!`Z-8)(<3,Z-8)'$8,02'%'$1$$700_&5T#$01(/QI=`Q$$1S\S
M.C6#$$AQ1Q&"70-P$!`Z-8)'<32"2!&#$$AQ1Q$1!%T#/W<_,01C#C].!&,/
M/S$$8R(_,`1C"C].<3$$8PL_-`1C`C^P-`1C##\T!&,-/S$$8P,_*E1Q,01>
B(#\J70005!$$7B$_*E1Q,01>(C\J70,05!$$7B,_WS]HQ@``
`
end
----end uue----