FLOAT_TO_EXP(3) 3 (November 10, 1987) FLOAT_TO_EXP(3)
NAME
float_to_exp - Convert floating point values into
"exponential" pixels.
SYNOPSIS
#include <rle.h>
float_to_exp( count, floats, pixels )
int count;
float * floats;
rle_pixel * pixels;
DESCRIPTION
The function float_to_exp converts count floating point
numbers (pointed to by floats) into count+1 bytes (pointed
to by pixels) using an "exponential" format. This format
generates count pixels as eight bit "mantissa" values, and
another byte containing a common exponent for all of the
data values. This format has a wider dynamic range of
values with little extra overhead. The inverse mapping is
float expnt, flt_val;
rle_pixel exponent, val;
expnt = ldexp( 1/256.0, (int)exponent - 127 );
flt_val = (float)val * expnt;
Files containing exponential data may be converted into
displayable images using the unexp(1) command. Unexp should
be used before using any tools that perform arithmetic on
pixel values, or displaying the image. Unexp expects files
containing exponential data to have an "exponential_data"
picture comment.
SEE ALSO
unexp(1), rle_putcom(3), librle(3), RLE(5).
AUTHOR
John W. Peterson, based on code by Spencer Thomas.
University of Utah
Page 1 (printed 12/1/98)
