HOME GUIDE OPERATIONS DOCS ERRORS FORMATS INSTALL NEW TIPS WEB SITES

FQ NP - Filter - Quick, No Padding

(4/14/95)

PURPOSE

Applies Fourier filters to 2-D or 3-D real images. Images NEED NOT have power-of-two dimensions (see notes). No padding is applied.

SEE ALSO

FQ [Filter - quick]
FF [Fourier filter]
FL [Fourier list]
FP [Fourier interpolation]
FT [Fourier transformation]

USAGE

.OPERATION: FQ NP

.INPUT FILE: PIC001
[Enter name of input file containing real picture.]

.OUTPUT FILE: FIC001
[Enter name of output file containing filtered real picture.]

1: low-pass, 2: high-pass, 3: Gauss low-pass, 4: Gauss high-pass 5: Fermi low-pass, 6: Fermi high-pass, 7: Buter low-pass, 8: Buter high-pass

.Filter type (1-8): 7
[Enter filter option code.]

Option "1" - Low-pass truncation. Filter is "top-hat" function that truncates Fourier transform at spatial frequency RAD.

Option "2" - High-pass truncation. Filter is inverse "top-hat" function that passes Fourier transform beyond spatial frequency radius RAD.

Option "3" - Gaussian low-pass. Filter is Gaussian function EXP(-f**2/(2.*RAD**2)), where F is the frequency.

Option "4" - Gaussian high-pass. Filter is complement of Gaussian function: 1.0-EXP(-F**2/(2.*RAD**2)).

Option "5" - Fermi low-pass. Filter is 1/(1+EXP[(F-RAD)/T]) which negotiates between "top-hat" and Gaussian characteristics, depending on the value of T (see below).

Option "6" - Fermi high-pass. As low-pass, but T is replaced by -T.

Option "7" - Butterworth low-pass. Filter is 1/(SQRT(1+F/RAD)**(2*ORDER)) In the actual implementation of the filter the parameters, RAD and ORDER are calculated from the pass-band frequency and cut-off frequency specified by the user. See NOTE 5 below.

Option "8" - Butterworth high-pass. Filter is 1-(1/(SQRT(1+F/RAD)**(2*ORDER)))]

For options 1-6:

.FILTER RADIUS: 0.12
[Enter filter function radius in frequency units. They are of the range 0.0<=f<=0.5. For an explanation of this units, see FF PLOT.]

For options 7-8:

.PASS-BAND AND STOP-BAND FREQUENCY: 0.1,0.2
[Enter filter function radii in frequency units.

For the Fermi filters [(5) and (6)], the temperature parameter T will be solicited:

.TEMPERATURE (0=CUTOFF): .3
[Roughly within this reciprocal distance (in terms of frequency units), the filter falls off.]

NOTES

  1. FQ NP has similar functions to FF except that:

  2. it accepts only real space images (2- & 3-D) and leaves the filtered image real,

  3. it runs only in core, so it is much faster than the FF operation,

  4. The FILTER RADIUS can be given either in absolute units or pixel units. If answer is >1.0 it is treated as given in pixel units.

  5. In the Butterworth filter the ORDER determines the filter fall off and RAD corresponds to the cut-off radius. In the program RAD and ORDER are calculated from the parameters specified by the user using following equations:
    RAD = fp/((eps)**(2/ORDER))
    ORDER = [2*log(eps/sqrt(a**2-1))]/[log(fp/fs)]
    where fp and fs are the pass-band and stop-band frequencies and parameters "eps" and "a" are given by 0.882 and 10.624 resp. Note that fp and fs deviate from 1.0 and 0.0 by about 0.2 and 0.09 respectively (for a low-pass filter.) For reference see 'FF PLOT'.

  6. Implemented by: Paul Penczek.

SUBROUTINES

FOUR1A, FQ_Q, FQ3_P

CALLER

FOUR1

© Copyright Notice /       Enquiries: spider@wadsworth.org