src/parsec/disk-image/parsec/parsec-benchmark/pkgs/apps/vips/src/man/im_add.3

.TH ADDITION 3 "24 April 1991"
.SH NAME
im_add, im_gadd, im_gaddim, im_gfadd \- add two images
.SH SYNOPSIS
#include <vips/vips.h>

int im_add(in1, in2, out)
.br
IMAGE *in1, *in2, *out;

int im_gadd(a, in1, b, in2, c, out)
.br
double a, b, c;
.br
IMAGE *in1, *in2, *out;

int im_gaddim(a, in1, b, in2, c, out)
.br
double a, b, c;
.br
IMAGE *in1, *in2, *out;

int im_gfadd(a, in1, b, in2, c, out)
.br
double a, b, c;
.br
IMAGE *in1, *in2, *out;

.SH DESCRIPTION
These functions operate on two images held by image descriptors in1 and in2
and write the result to the image descriptor out.  Input images in1 and in2
should have the same channels and the same size; however they can be of
different types.  Only the history of the image descriptor pointed by in1 is
copied to out.

.B im_add(3) 

For two integer images, add the two images and write the output as

 in1    - uchar   char   ushort  short  uint  int 
 -------|-----------------------------------------
 in2    |
 uchar  | ushort  short  ushort  short  uint  int
 char   | short   short  short   short  int   int
 ushort | ushort  short  ushort  short  uint  int
 short  | short   short  short   short  int   int
 uint   | uint    int    uint    int    uint  int
 int    | int     int    int     int    int   int

If one or more of the images is a floating point type, the output is FMTFLOAT,
unless one or more of the inputs is FMTDOUBLE, in which case the output is
also FMTDOUBLE.

If one or more of the images is a complex type, the output is FMTCOMPLEX,
unless one or more of the inputs is FMTDPCOMPLEX, in which case the output is
also FMTDPCOMPLEX.

.B im_gadd(3) 
performs generalised addition of two images by calling 
.B im_gaddim(3)
and 
.B im_gfadd(3).
These are very old and tired things, and should not be used.

Input should be non complex.  Output depends on input according to function
called.  The result at each point is: a * pel1 + b * pel2 + c, properly
rounded if necessary.  Pel1 and pel2 are the
corresponding pixels from in1 and in2 respectively.

im_gaddim() performs generalised addition of in1 and in2, on the condition
they are neither float nor double nor complex.  The format of the resultant
image is given by the table:

 in1    - uchar   char   ushort  short  uint  int 
 -------|-----------------------------------------
 in2    |
 uchar  | ushort  short  ushort  short  uint  int
 char   | short   short  short   short  int   int
 ushort | ushort  short  ushort  short  uint  int
 short  | short   short  short   short  int   int
 uint   | uint    int    uint    int    uint  int
 int    | int     int    int     int    int   int

The result at each point is: a * pel1 + b * pel2 + c, properly rounded. Pel1
and pel2 are the corresponding pixels from in1 and in2 respectively.

.B im_gfadd(3) 
adds the non-complex images pointed by in1 and in2.  Result is
float except if one (or both) inputs is double.  In the latter case the result
is double.  The result at each point is: a * pel1 + b * pel2 + c.  Pel1 and
pel2 are the corresponding pixels from in1 and in2 respectively.

.SH BUGS
None of the functions checks the result for over/underflow.
.SH RETURN VALUE
All functions return 0 on success and -1 on error.
.SH SEE ALSO
im_subtract(3), im_lintra(3), im_multiply(3).
.SH AUTHOR
N. Dessipris \- 22/04/1991
.br
J. Cupitt, im_add(), \- 21/7/93