PFILT(1)						      General Commands Manual							  PFILT(1)

NAME

       pfilt - filter a RADIANCE picture

SYNOPSIS

       pfilt [ options ] [ file ]

DESCRIPTION

       Pfilt performs anti-aliasing and scaling on a RADIANCE picture.	The program makes two passes on the picture file in order to set the expo-
       sure to the correct average value.  If no file is given, the standard input is read.

       -x res	 Set the output x resolution to res.  This must be less than or equal to the x dimension of the target device.	If res is given as
		 a  slash followed by a real number, the input resolution is divided by this number to get the output resolution.  By default, the
		 output resolution is the same as the input.

       -y res	 Set the output y resolution to res, similar to the specification of the x resolution above.

       -p rat	 Set the pixel aspect ratio to rat.  Either the x or the y resolution will be reduced so that the pixels have this ratio  for  the
		 specified picture.  If rat is zero, then the x and y resolutions will adhere to the given maxima.  Zero is the default.

       -c	 Pixel aspect ratio is being corrected, so do not write PIXASPECT variable to output file.

       -e exp	 Adjust  the exposure.	If exp is preceded by a '+' or '-', the exposure is interpreted in f-stops (ie. the power of two).  Other-
		 wise, exp is interpreted as a straight multiplier.  The individual primaries can be changed using -er,  -eg  and  -eb.   Multiple
		 exposure options have a cumulative effect.

       -t lamp	 Color-balance	the  image as if it were illuminated by fixtures of the given type.  The specification must match a pattern listed
		 in the lamp lookup table (see the -f option below).

       -f lampdat
		 Use the specified lamp lookup table rather than the default (lamp.tab).

       -1	 Use only one pass on the file.  This allows the exposure to be controlled absolutely, without any averaging.  Note that a  single
		 pass is much quicker and should be used whenever the desired exposure is known and star patterns are not required.

       -2	 Use two passes on the input.  This is the default.

       -b	 Use box filtering (default).  Box filtering averages the input pixels corresponding to each separate output pixel.

       -r rad	 Use Gaussian filtering with a radius of rad relative to the output pixel size.  This option with a radius around 0.6 and a reduc-
		 tion in image width and height of 2 or 3 produces the highest quality pictures.  A radius greater than 0.7 results in a defocused
		 picture.

       -m frac	 Limit the influence of any given input pixel to frac of any given output pixel.  This option may be used to mitigate the problems
		 associated with inadequate image sampling, at the expense of a slightly blurred image.  The fraction given  should  not  be  less
		 than  the  output  picture dimensions over the input picture dimensions (x_o*y_o/x_i/y_i), or blurring will occur over the entire
		 image.  This option implies the -r option for Gaussian filtering, which defaults to a radius of 0.6.

       -h lvl	 Set intensity considered ``hot'' to lvl.  This is the level above which areas of the image will begin to exhibit star diffraction
		 patterns (see below).	The default is 100 watts/sr/m2.

       -n N	 Set  the  number  of  points  on star patterns to N.  A value of zero turns star patterns off.  The default is 0.  (Note that two
		 passes are required for star patterns.)

       -s val	 Set the spread for star patterns to val.  This is the value a star pattern will have at the edge of the image.   The  default	is
		 .0001.

       -a	 Average hot spots as well.  By default, the areas of the picture above the hot level are not used in setting the exposure.

ENVIRONMENT

       RAYPATH	      directories to search for lamp lookup table

FILES

       /tmp/rt??????

AUTHOR

       Greg Ward

SEE ALSO

       getinfo(1), ies2rad(1), pcompos(1), pflip(1), pinterp(1), pvalue(1), protate(1), rad(1), rpict(1), ximage(1)

RADIANCE
							      11/8/96								  PFILT(1)