mirror of
https://github.com/kasmtech/KasmVNC.git
synced 2024-11-25 09:33:50 +01:00
126 lines
3.7 KiB
C++
126 lines
3.7 KiB
C++
|
/* Copyright (C) 2006 TightVNC Team. All Rights Reserved.
|
||
|
*
|
||
|
* This is free software; you can redistribute it and/or modify
|
||
|
* it under the terms of the GNU General Public License as published by
|
||
|
* the Free Software Foundation; either version 2 of the License, or
|
||
|
* (at your option) any later version.
|
||
|
*
|
||
|
* This software 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 General Public License for more details.
|
||
|
*
|
||
|
* You should have received a copy of the GNU General Public License
|
||
|
* along with this software; if not, write to the Free Software
|
||
|
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
|
||
|
* USA.
|
||
|
*/
|
||
|
|
||
|
#include <string.h>
|
||
|
#include <assert.h>
|
||
|
#include <math.h>
|
||
|
|
||
|
#include <rfb/Rect.h>
|
||
|
#include <rfb/ScaleFilters.h>
|
||
|
|
||
|
using namespace rfb;
|
||
|
|
||
|
//
|
||
|
// -=- 1-D filters functions
|
||
|
//
|
||
|
|
||
|
// Nearest neighbor filter function
|
||
|
double nearest_neighbor(double x) {
|
||
|
if (x < -0.5) return 0.0;
|
||
|
if (x < 0.5) return 1.0;
|
||
|
return 0.0;
|
||
|
}
|
||
|
|
||
|
// Linear filter function
|
||
|
double linear(double x) {
|
||
|
if (x < -1.0) return 0.0;
|
||
|
if (x < 0.0) return 1.0+x;
|
||
|
if (x < 1.0) return 1.0-x;
|
||
|
return 0.0;
|
||
|
}
|
||
|
|
||
|
// Cubic filter functions
|
||
|
double cubic(double x) {
|
||
|
double t;
|
||
|
if (x < -2.0) return 0.0;
|
||
|
if (x < -1.0) {t = 2.0+x; return t*t*t/6.0;}
|
||
|
if (x < 0.0) return (4.0+x*x*(-6.0+x*-3.0))/6.0;
|
||
|
if (x < 1.0) return (4.0+x*x*(-6.0+x*3.0))/6.0;
|
||
|
if (x < 2.0) {t = 2.0-x; return t*t*t/6.0;}
|
||
|
return 0.0;
|
||
|
}
|
||
|
|
||
|
|
||
|
//
|
||
|
// -=- ScaleFilters class
|
||
|
//
|
||
|
|
||
|
SFilter &ScaleFilters::operator[](unsigned int filter_id) {
|
||
|
assert(filter_id <= scaleFilterMaxNumber);
|
||
|
return filters[filter_id];
|
||
|
}
|
||
|
|
||
|
int ScaleFilters::getFilterIdByName(char *filterName) {
|
||
|
for (unsigned int i = 0; i <= scaleFilterMaxNumber; i++) {
|
||
|
if (strcasecmp(filters[i].name, filterName) == 0) return i;
|
||
|
}
|
||
|
return -1;
|
||
|
}
|
||
|
|
||
|
void ScaleFilters::initFilters() {
|
||
|
filters[scaleFilterNearestNeighbor] = create("Nearest neighbor", 0.5, nearest_neighbor);
|
||
|
filters[scaleFilterBilinear] = create("Bilinear", 1, linear);
|
||
|
filters[scaleFilterBicubic] = create("Bicubic", 2, cubic);
|
||
|
}
|
||
|
|
||
|
SFilter ScaleFilters::create(const char *name_, double radius_, filter_func func_) {
|
||
|
SFilter filter;
|
||
|
strncpy(filter.name, name_, sizeof(filter.name)-1);
|
||
|
filter.name[sizeof(filter.name)-1] = '\0';
|
||
|
filter.radius = radius_;
|
||
|
filter.func = func_;
|
||
|
return filter;
|
||
|
}
|
||
|
|
||
|
void ScaleFilters::makeWeightTabs(int filter_id, int src_x, int dst_x, SFilterWeightTab **pWeightTabs) {
|
||
|
double sxc;
|
||
|
double offset = 0.5;
|
||
|
double ratio = (double)dst_x / src_x;
|
||
|
double sourceScale = __rfbmax(1.0, 1.0/ratio);
|
||
|
double sourceRadius = __rfbmax(0.5, sourceScale * filters[filter_id].radius);
|
||
|
double sum, nc;
|
||
|
int i, ci;
|
||
|
|
||
|
SFilter sFilter = filters[filter_id];
|
||
|
|
||
|
*pWeightTabs = new SFilterWeightTab[dst_x];
|
||
|
SFilterWeightTab *weightTabs = *pWeightTabs;
|
||
|
|
||
|
// Make the weight tab for the each dest x position
|
||
|
for (int x = 0; x < dst_x; x++) {
|
||
|
sxc = (double(x)+offset) / ratio;
|
||
|
|
||
|
// Calculate the scale filter interval, [i0, i1)
|
||
|
int i0 = int(__rfbmax(sxc-sourceRadius+0.5, 0));
|
||
|
int i1 = int(__rfbmin(sxc+sourceRadius+0.5, src_x));
|
||
|
|
||
|
weightTabs[x].i0 = i0; weightTabs[x].i1 = i1;
|
||
|
weightTabs[x].weight = new short[i1-i0];
|
||
|
|
||
|
// Calculate coeff to normalize the filter weights
|
||
|
for (sum = 0, i = i0; i < i1; i++) sum += sFilter.func((double(i)-sxc+0.5)/sourceScale);
|
||
|
if (sum == 0.) nc = (double)(WEIGHT_OF_ONE); else nc = (double)(WEIGHT_OF_ONE)/sum;
|
||
|
|
||
|
|
||
|
// Calculate the weight coeffs on the scale filter interval
|
||
|
for (ci = 0, i = i0; i < i1; i++) {
|
||
|
weightTabs[x].weight[ci++] = (short)floor((sFilter.func((double(i)-sxc+0.5)/sourceScale) * nc) + 0.5);
|
||
|
}
|
||
|
}
|
||
|
}
|