Fant rotates or
scales an image by an arbitrary amount. It does this by using pixel integration
(if the image size is reduced) or pixel interpolation if the image size
is increased. Because it works with subpixel precision, aliasing artifacts
are not introduced (hah! see BUGS). Fant uses a two-pass sampling technique
to perform the transformation. If infile is "-" or absent, input is read
from the standard input.
Amount to rotate image by, a real
number from 0 to 45 degrees (positive numbers rotate clockwise). Use rleflip(1)
first to rotate an image by larger amounts.
-b blur_factor
Control the amount
of blurring in the output image. If the blur factor is greater than one,
image blurring will increase. If the blur factor is smaller than one, image
blurring will decrease but aliasing artifacts may be visible.
-o outfile
Specifies where to place the resulting image. The default is to write to
stdout. If outfile is "-", the output will be written to the standard output
stream.
-p xoff yoff
Specifies where the origin of the image is - the image
is rotated or scaled about this point. If no origin is specified, the center
of the image is used.
-s xscale yscale
The amount (in real numbers) to scale
an image by. This is often useful for correcting the aspect of an image
for display on a frame buffer with non square pixels. For this use, the
origin should be specified as 0, 0 (see -p above). If an image is only scaled
in Y and no rotation is performed, fant only uses one sampling pass over
the image, cutting the computation time in half.
-S xsize ysize
An alternate
method of specifying the scale factors. xsize and ysize give the desired
output image size.
The -S option can not be used in combination with -a,
-p, or -s.
Fant
uses a rather poor anti-aliasing filter (a triangle filter). This is usually
good enough but will exhibit noticeable aliasing artifacts on nasty input
images.