1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
|
/* This file is part of the Spring engine (GPL v2 or later), see LICENSE.html */
#include "System/float3.h"
#include "System/myMath.h"
#include <stdlib.h>
#include <time.h>
#define BOOST_TEST_MODULE Ellipsoid
#include <boost/test/unit_test.hpp>
static inline float3 randfloat3()
{
float3 r(rand() + 1.0f, rand() + 1.0f, rand() + 1.0f);
//Disallow insane ellipsoids
float minValue = std::fmax(r.x, std::max(r.y, r.z)) * 0.02;
r.x = std::max(r.x, minValue);
r.y = std::max(r.y, minValue);
r.z = std::max(r.z, minValue);
return r;
}
static inline float getdistSq(float x, float y, float z, float a, float b, float c, float theta, float phi) {
const float cost = cos(theta);
const float sint = sin(theta);
const float sinp = sin(phi);
const float cosp = cos(phi);
const float fx = a * cosp * cost - x;
const float fy = b * cosp * sint - y;
const float fz = c * sinp - z;
return fx * fx + fy * fy + fz * fz;
}
#define TEST_RUNS 100000
#define MAX_ITERATIONS 20
#define THRESHOLD 0.001
#define MAX_FAILS 5 //allow some fails
//We Fail if after half the iterations we have error > 5%
#define FAIL_THRESHOLD 0.05f
BOOST_AUTO_TEST_CASE( Ellipsoid )
{
srand( time(NULL) );
unsigned failCount = 0;
float relError[MAX_ITERATIONS + 1] = {};
float maxError[MAX_ITERATIONS + 1] = {};
float tempdist[MAX_ITERATIONS + 1] = {};
float errorsq[MAX_ITERATIONS + 1] = {};
unsigned finalIteration[MAX_ITERATIONS + 1] = {};
for (int j = 0; j < TEST_RUNS; ++j) {
float3 halfScales = randfloat3();
float3 pv = randfloat3();
const float& a = halfScales.x;
const float& b = halfScales.y;
const float& c = halfScales.z;
const float& x = pv.x;
const float& y = pv.y;
const float& z = pv.z;
float weightedLengthSq = (x * x) / (a * a) + (y * y) / (b * b) + (z * z) / (c * c);
while (weightedLengthSq <= 1.0f) {
halfScales = randfloat3();
pv = randfloat3();
weightedLengthSq = (x * x) / (a * a) + (y * y) / (b * b) + (z * z) / (c * c);
}
const float a2 = a * a;
const float b2 = b * b;
const float c2 = c * c;
const float x2 = x * x;
const float y2 = y * y;
const float a2b2 = a2 - b2;
const float xa = x * a;
const float yb = y * b;
const float zc = z * c;
//Initial guess
float theta = math::atan2(a * y, b * x);
float phi = math::atan2(z, c * math::sqrt(x2 / a2 + y2 / b2));
//Iterations
for (int i = 0; true; ) {
const float cost = math::cos(theta);
const float sint = math::sin(theta);
const float sinp = math::sin(phi);
const float cosp = math::cos(phi);
const float sin2t = sint * sint;
const float xacost_ybsint = xa * cost + yb * sint;
const float xasint_ybcost = xa * sint - yb * cost;
const float a2b2costsint = a2b2 * cost * sint;
const float a2cos2t_b2sin2t_c2 = a2 * cost * cost + b2 * sin2t - c2;
const float d1 = a2b2costsint * cosp - xasint_ybcost;
const float d2 = a2cos2t_b2sin2t_c2 * sinp * cosp - sinp * xacost_ybsint + zc * cosp;
{
const float fx = a * cosp * cost - x;
const float fy = b * cosp * sint - y;
const float fz = c * sinp - z;
tempdist[i] = math::sqrt(fx * fx + fy * fy + fz * fz);
if (i > 1 && (std::abs(tempdist[i] - tempdist[i - 1]) < THRESHOLD * tempdist[i]))
++finalIteration[i];
if (i++ == MAX_ITERATIONS)
break;
}
//Derivative matrix
const float a11 = a2b2 * (1 - 2 * sin2t) * cosp - xacost_ybsint;
const float a12 = -a2b2costsint * sinp;
const float a21 = 2 * a12 * cosp + sinp * xasint_ybcost;
const float a22 = a2cos2t_b2sin2t_c2 * (1 - 2 * sinp * sinp) - cosp * xacost_ybsint - zc;
const float invDet = 1.0f / (a11 * a22 - a21 * a12);
theta += (a12 * d2 - a22 * d1) * invDet;
theta = Clamp(theta, 0.0f, math::HALFPI);
phi += (a21 * d1 - a11 * d2) * invDet;
phi = Clamp(phi, 0.0f, math::HALFPI);
}
bool failed = false;
for (int i = 0; i < MAX_ITERATIONS; ++i) {
//Relative error for every iteration
const float tempError = std::abs(tempdist[i] - tempdist[MAX_ITERATIONS]) / tempdist[MAX_ITERATIONS];
relError[i] += tempError;
maxError[i] = std::max(tempError, maxError[i]);
if (i > MAX_ITERATIONS / 2 && tempError > FAIL_THRESHOLD && !failed) {
failed = true;
++failCount;
printf("x: %f, y: %f, z: %f, a: %f, b: %f, c: %f\n", x,y,z,a,b,c);
}
errorsq[i] += tempError * tempError;
}
}
for (int i = 0; i < MAX_ITERATIONS; ++i) {
const float meanError = relError[i] / TEST_RUNS;
const float devError = math::sqrt(errorsq[i] / TEST_RUNS - meanError * meanError);
const float pct = 100.0f * (1.0f - float(finalIteration[i]) / TEST_RUNS);
printf("Iteration %d:\n\tError: (Mean: %f, Dev: %f, Max: %f)\n\tPercent remaining: %.3f%%\n", i, meanError, devError, maxError[i], pct);
}
BOOST_CHECK_MESSAGE(failCount < MAX_FAILS, "Inaccurate ellipsoid distance approximation!");
}
|
