-
Notifications
You must be signed in to change notification settings - Fork 0
/
noise.py
94 lines (72 loc) · 2.09 KB
/
noise.py
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
import taichi as ti
import numpy as np
import taichi_glsl as ts
from PIL import Image
import cv2
ti.init(arch=ti.gpu)
nx = 256
ny = 256
blk_txt = ti.field(dtype=ti.f32, shape=(nx, ny))
pic = ti.field(dtype=ti.f32, shape=(nx, ny))
@ti.func
def smooth_lerp(vl, vr, frac):
# frac: [0.0, 1.0]
return vl + ts.scalar.smoothstep(frac) * (vr - vl)
@ti.func
def noise(i, j):
return ts.fract(ti.sin(i * 12.9898 + j * 78.233) * 43758.5453)
@ti.func
def interp_noise(i: ti.f32, j: ti.f32):
fract_p_i = ts.fract(i)
fract_p_j = ts.fract(j)
int_p_i = ti.floor(i)
int_p_j = ti.floor(j)
v1 = noise(int_p_i, int_p_j)
v2 = noise(int_p_i + 1, int_p_j)
v3 = noise(int_p_i, int_p_j + 1)
v4 = noise(int_p_i + 1, int_p_j + 1)
return smooth_lerp(smooth_lerp(v1, v2, fract_p_i),
smooth_lerp(v3, v4, fract_p_i),
fract_p_j)
@ti.func
def fbm(x: ti.f32, y: ti.f32):
persistence = 0.5
total = 0.0
for i in range(10):
freq = 2 ** i
amp = persistence ** i
total += interp_noise(x * freq, y * freq) * amp
return ts.scalar.clamp(total * 0.5)
@ti.kernel
def init():
for i, j in ti.ndrange(nx, ny):
pic[i, j] = noise(i * 0.05, j * 0.05)
# pic[i, j] = ts.scalar.smoothstep(fbm(ti.cast(i * 0.05, ti.f32), ti.cast(j * 0.05, ti.f32)), 0.2, 0.8)
@ti.kernel
def avg_lum() -> ti.f32:
avg_lum = 0.0
for i, j in pic:
avg_lum += pic[i, j] / (nx * ny)
return avg_lum
@ti.kernel
def test_bilerp1() -> ti.f32:
return ts.sampling.bilerp(pic, ts.vec2(1.5, 1.0))
@ti.kernel
def test_bilerp2() -> ti.f32:
return ts.sampling.bilerp(pic, ts.vec2(1.0, 1.0))
@ti.kernel
def test_bilerp3() -> ti.f32:
return ts.sampling.bilerp(pic, ts.vec2(2.0, 1.0))
if __name__ == '__main__':
init()
print(test_bilerp1())
print(test_bilerp2())
print(test_bilerp3())
# # print(avg_lum())
# gui = ti.GUI('lbm solver', (nx, ny))
# pic_np = pic.to_numpy()
# while True:
# gui.set_image(pic_np)
# gui.show()
# print(type(numpydata))
# print(numpydata.shape)