libvips/libvips/colour/im_LabQ2Lab.c

134 lines
3.1 KiB
C

/* @(#) LabQ2Lab convert Lab 32bit packed format to float Lab
@(#) uses imb_LabQ2Lab
Copyright Kirk Martinez 2/5/1993
Modified: 16/6/93
* 7/6/93 JC
* - adapted for partial v2
* 16/11/94 JC
* - adapted to new im_wrap_oneonebuf() function.
* 9/2/95 JC
* - new im_wrapone function
* 22/5/95 JC
* - changed char to unsigned char for RS/6000
* - small tidies and speed-ups
* 4/9/97 JC
* - L* = 100.0 now handled correctly
*/
/*
This file is part of VIPS.
VIPS is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif /*HAVE_CONFIG_H*/
#include <vips/intl.h>
#include <stdio.h>
#include <vips/vips.h>
#ifdef WITH_DMALLOC
#include <dmalloc.h>
#endif /*WITH_DMALLOC*/
/* imb_LabQ2Lab: CONVERT n pels from packed 32bit Lab to float values
* in a buffer
* ARGS: PEL *inp pointer to first byte of Lab32 buffer
* float *outbuf destination buffer
* int n number of pels to process
* (C) K.Martinez 2/5/93
*/
void
imb_LabQ2Lab( PEL *inp, float *outbuf, int n )
{
signed char *b; /* to read input bytes */
int l;
int lsbs; /* for lsbs byte */
int c; /* counter */
float *out;
/* Read input with a signed pointer to get signed ab easily.
*/
b = (signed char *) inp;
out = outbuf;
for( c = 0; c < n; c++ ) {
/* Get extra bits.
*/
lsbs = ((unsigned char *) b)[3];
/* Build L.
*/
l = ((unsigned char *)b)[0];
l = (l << 2) | (lsbs >> 6);
out[0] = (float) l * (100.0 / 1023.0);
/* Build a.
*/
l = (b[1] << 3) | ((lsbs >> 3) & 0x7);
out[1] = (float) l * 0.125;
/* And b.
*/
l = (b[2] << 3) | (lsbs & 0x7);
out[2] = (float) l * 0.125;
b += 4;
out += 3;
}
}
/* unpack a Lab32 ie 10,11,11 Lab image into float image
this is a wrapper around the buffer processing function
(C) K.Martinez 2/5/93
*/
int
im_LabQ2Lab( IMAGE *labim, IMAGE *outim )
{
/* check for coded Lab type
*/
if( labim->Coding != IM_CODING_LABQ ) {
im_errormsg( "im_LabQ2Lab: not a packed Lab image" );
return( -1 );
}
/* set up output image
*/
if( im_cp_desc( outim, labim ) )
return( -1 );
outim->Bands = 3;
outim->Type = IM_TYPE_LAB;
outim->BandFmt = IM_BANDFMT_FLOAT;
outim->Bbits = 32;
outim->Coding = IM_CODING_NONE;
if( im_wrapone( labim, outim,
(im_wrapone_fn) imb_LabQ2Lab, NULL, NULL ) )
return( -1 );
return( 0 );
}