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
162
163
164
165
166
|
#include <inttypes.h>
#include <math.h>
#include <stdint.h>
#include <stdlib.h>
#include "til.h"
#include "til_fb.h"
#include "til_module_context.h"
/* Copyright (C) 2017 Vito Caputo <vcaputo@pengaru.com> */
/* Normalize plasma size at 2*8K resolution, simply assume it's always being sampled
* smaller than this and ignore handling <1 fractional scaling factors.
*/
#define PLASMA_WIDTH 15360
#define PLASMA_HEIGHT 8640
#define FIXED_TRIG_LUT_SIZE 4096 /* size of the cos/sin look-up tables */
#define FIXED_BITS 9 /* fractional bits */
#define FIXED_EXP (1 << FIXED_BITS) /* 2^FIXED_BITS */
#define FIXED_MASK (FIXED_EXP - 1) /* fractional part mask */
#define FIXED_COS(_rad) costab[(_rad) & (FIXED_TRIG_LUT_SIZE-1)]
#define FIXED_SIN(_rad) sintab[(_rad) & (FIXED_TRIG_LUT_SIZE-1)]
#define FIXED_MULT(_a, _b) (((_a) * (_b)) >> FIXED_BITS)
#define FIXED_NEW(_i) ((_i) << FIXED_BITS)
#define FIXED_TO_INT(_f) ((_f) >> FIXED_BITS)
typedef struct color_t {
int r, g, b;
} color_t;
static int32_t costab[FIXED_TRIG_LUT_SIZE], sintab[FIXED_TRIG_LUT_SIZE];
typedef struct plasma_context_t {
til_module_context_t til_module_context;
unsigned rr;
} plasma_context_t;
static inline uint32_t color2pixel(color_t *color)
{
return (FIXED_TO_INT(color->r) << 16) | (FIXED_TO_INT(color->g) << 8) | FIXED_TO_INT(color->b);
}
static void init_plasma(int32_t *costab, int32_t *sintab)
{
/* Generate fixed-point cos & sin LUTs. */
for (int i = 0; i < FIXED_TRIG_LUT_SIZE; i++) {
costab[i] = ((cos((double)2*M_PI*i/FIXED_TRIG_LUT_SIZE))*FIXED_EXP);
sintab[i] = ((sin((double)2*M_PI*i/FIXED_TRIG_LUT_SIZE))*FIXED_EXP);
}
}
static til_module_context_t * plasma_create_context(const til_module_t *module, til_stream_t *stream, unsigned seed, unsigned ticks, unsigned n_cpus, til_setup_t *setup)
{
static int initialized;
plasma_context_t *ctxt;
if (!initialized) {
initialized = 1;
init_plasma(costab, sintab);
}
ctxt = til_module_context_new(module, sizeof(plasma_context_t), stream, seed, ticks, n_cpus, setup);
if (!ctxt)
return NULL;
ctxt->rr = rand_r(&seed);
return &ctxt->til_module_context;
}
/* Prepare a frame for concurrent drawing of fragment using multiple fragments */
static void plasma_prepare_frame(til_module_context_t *context, til_stream_t *stream, unsigned ticks, til_fb_fragment_t **fragment_ptr, til_frame_plan_t *res_frame_plan)
{
plasma_context_t *ctxt = (plasma_context_t *)context;
*res_frame_plan = (til_frame_plan_t){ .fragmenter = til_fragmenter_slice_per_cpu_x16 };
ctxt->rr += 3;
}
/* Draw a plasma effect */
static void plasma_render_fragment(til_module_context_t *context, til_stream_t *stream, unsigned ticks, unsigned cpu, til_fb_fragment_t **fragment_ptr)
{
plasma_context_t *ctxt = (plasma_context_t *)context;
til_fb_fragment_t *fragment = *fragment_ptr;
int xstep = PLASMA_WIDTH / fragment->frame_width;
int ystep = PLASMA_HEIGHT / fragment->frame_height;
unsigned width = fragment->width * xstep, height = fragment->height * ystep;
int fw2 = FIXED_NEW(width / 2), fh2 = FIXED_NEW(height / 2);
int x, y, cx, cy, dx2, dy2;
uint32_t *buf = fragment->buf;
color_t c = { .r = 0, .g = 0, .b = 0 }, cscale;
unsigned rr2, rr6, rr8, rr16, rr20, rr12;
rr2 = ctxt->rr * 2;
rr6 = ctxt->rr * 6;
rr8 = ctxt->rr * 8;
rr16 = ctxt->rr * 16;
rr20 = ctxt->rr * 20;
rr12 = ctxt->rr * 12;
/* vary the color channel intensities */
cscale.r = FIXED_MULT(FIXED_COS(ctxt->rr / 2), FIXED_NEW(64)) + FIXED_NEW(64);
cscale.g = FIXED_MULT(FIXED_COS(ctxt->rr / 5), FIXED_NEW(64)) + FIXED_NEW(64);
cscale.b = FIXED_MULT(FIXED_COS(ctxt->rr / 7), FIXED_NEW(64)) + FIXED_NEW(64);
cx = FIXED_TO_INT(FIXED_MULT(FIXED_COS(ctxt->rr), fw2) + fw2);
cy = FIXED_TO_INT(FIXED_MULT(FIXED_SIN(rr2), fh2) + fh2);
for (y = fragment->y * ystep; y < fragment->y * ystep + height; y += ystep) {
int y2 = y << 1;
int y4 = y << 2;
dy2 = cy - y;
dy2 *= dy2;
for (x = fragment->x * xstep; x < fragment->x * xstep + width; x += xstep, buf++) {
int v;
int hyp;
dx2 = cx - x;
dx2 *= dx2;
hyp = (dx2 + dy2) >> 13; /* XXX: technically this should be a sqrt(), but >> 10 is a whole lot faster. */
#define S 4
v = FIXED_MULT( ((FIXED_COS(rr8 + ((hyp * 5) >> S))) +
(FIXED_SIN(-rr16 + ((x << 2) >> S))) +
(FIXED_COS(rr20 + (y4 >> S)))),
FIXED_EXP / 3); /* XXX: note these '/ 3' get optimized out. */
c.r = FIXED_MULT(v, cscale.r) + cscale.r;
v = FIXED_MULT( ((FIXED_COS(rr12 + ((hyp << 2) >> S))) +
(FIXED_COS(rr6 + ((x << 1) >> S))) +
(FIXED_SIN(rr16 + (y2 >> S)))),
FIXED_EXP / 3);
c.g = FIXED_MULT(v, cscale.g) + cscale.g;
v = FIXED_MULT( ((FIXED_SIN(rr6 + ((hyp * 6) >> S))) +
(FIXED_COS(-rr12 + ((x * 5) >> S))) +
(FIXED_SIN(-rr6 + (y2 >> S)))),
FIXED_EXP / 3);
c.b = FIXED_MULT(v, cscale.b) + cscale.b;
*buf = color2pixel(&c);
}
buf += fragment->stride;
}
}
til_module_t plasma_module = {
.create_context = plasma_create_context,
.prepare_frame = plasma_prepare_frame,
.render_fragment = plasma_render_fragment,
.name = "plasma",
.description = "Oldskool plasma effect (threaded)",
.author = "Vito Caputo <vcaputo@pengaru.com>",
};
|
