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
|
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// Copyright (c) 2016, Intel Corporation
// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
// documentation files (the "Software"), to deal in the Software without restriction, including without limitation
// the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to
// permit persons to whom the Software is furnished to do so, subject to the following conditions:
// The above copyright notice and this permission notice shall be included in all copies or substantial portions of
// the Software.
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO
// THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
// TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
// SOFTWARE.
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// File changes (yyyy-mm-dd)
// 2016-09-07: filip.strugar@intel.com: first commit
// 2020-12-05: clayjohn: convert to Vulkan and Godot
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
#[compute]
#version 450
#VERSION_DEFINES
layout(local_size_x = 8, local_size_y = 8, local_size_z = 1) in;
layout(rgba8, set = 0, binding = 0) uniform restrict writeonly image2D dest_image;
layout(set = 1, binding = 0) uniform sampler2DArray source_texture;
layout(push_constant, binding = 1, std430) uniform Params {
float inv_sharpness;
uint size_modifier;
vec2 pixel_size;
}
params;
vec4 unpack_edges(float p_packed_val) {
uint packed_val = uint(p_packed_val * 255.5);
vec4 edgesLRTB;
edgesLRTB.x = float((packed_val >> 6) & 0x03) / 3.0;
edgesLRTB.y = float((packed_val >> 4) & 0x03) / 3.0;
edgesLRTB.z = float((packed_val >> 2) & 0x03) / 3.0;
edgesLRTB.w = float((packed_val >> 0) & 0x03) / 3.0;
return clamp(edgesLRTB + params.inv_sharpness, 0.0, 1.0);
}
void main() {
ivec2 ssC = ivec2(gl_GlobalInvocationID.xy);
if (any(greaterThanEqual(ssC, ivec2(1.0 / params.pixel_size)))) { //too large, do nothing
return;
}
#ifdef MODE_SMART
float ao;
uvec2 pix_pos = uvec2(gl_GlobalInvocationID.xy);
vec2 uv = (gl_GlobalInvocationID.xy + vec2(0.5)) * params.pixel_size;
// calculate index in the four deinterleaved source array texture
int mx = int(pix_pos.x % 2);
int my = int(pix_pos.y % 2);
int index_center = mx + my * 2; // center index
int index_horizontal = (1 - mx) + my * 2; // neighbouring, horizontal
int index_vertical = mx + (1 - my) * 2; // neighbouring, vertical
int index_diagonal = (1 - mx) + (1 - my) * 2; // diagonal
vec2 center_val = texelFetch(source_texture, ivec3(pix_pos / uvec2(params.size_modifier), index_center), 0).xy;
ao = center_val.x;
vec4 edgesLRTB = unpack_edges(center_val.y);
// convert index shifts to sampling offsets
float fmx = float(mx);
float fmy = float(my);
// in case of an edge, push sampling offsets away from the edge (towards pixel center)
float fmxe = (edgesLRTB.y - edgesLRTB.x);
float fmye = (edgesLRTB.w - edgesLRTB.z);
// calculate final sampling offsets and sample using bilinear filter
vec2 uv_horizontal = (gl_GlobalInvocationID.xy + vec2(0.5) + vec2(fmx + fmxe - 0.5, 0.5 - fmy)) * params.pixel_size;
float ao_horizontal = textureLod(source_texture, vec3(uv_horizontal, index_horizontal), 0.0).x;
vec2 uv_vertical = (gl_GlobalInvocationID.xy + vec2(0.5) + vec2(0.5 - fmx, fmy - 0.5 + fmye)) * params.pixel_size;
float ao_vertical = textureLod(source_texture, vec3(uv_vertical, index_vertical), 0.0).x;
vec2 uv_diagonal = (gl_GlobalInvocationID.xy + vec2(0.5) + vec2(fmx - 0.5 + fmxe, fmy - 0.5 + fmye)) * params.pixel_size;
float ao_diagonal = textureLod(source_texture, vec3(uv_diagonal, index_diagonal), 0.0).x;
// reduce weight for samples near edge - if the edge is on both sides, weight goes to 0
vec4 blendWeights;
blendWeights.x = 1.0;
blendWeights.y = (edgesLRTB.x + edgesLRTB.y) * 0.5;
blendWeights.z = (edgesLRTB.z + edgesLRTB.w) * 0.5;
blendWeights.w = (blendWeights.y + blendWeights.z) * 0.5;
// calculate weighted average
float blendWeightsSum = dot(blendWeights, vec4(1.0, 1.0, 1.0, 1.0));
ao = dot(vec4(ao, ao_horizontal, ao_vertical, ao_diagonal), blendWeights) / blendWeightsSum;
imageStore(dest_image, ivec2(gl_GlobalInvocationID.xy), vec4(ao));
#else // !MODE_SMART
vec2 uv = (gl_GlobalInvocationID.xy + vec2(0.5)) * params.pixel_size;
#ifdef MODE_HALF
float a = textureLod(source_texture, vec3(uv, 0), 0.0).x;
float d = textureLod(source_texture, vec3(uv, 3), 0.0).x;
float avg = (a + d) * 0.5;
#else
float a = textureLod(source_texture, vec3(uv, 0), 0.0).x;
float b = textureLod(source_texture, vec3(uv, 1), 0.0).x;
float c = textureLod(source_texture, vec3(uv, 2), 0.0).x;
float d = textureLod(source_texture, vec3(uv, 3), 0.0).x;
float avg = (a + b + c + d) * 0.25;
#endif
imageStore(dest_image, ivec2(gl_GlobalInvocationID.xy), vec4(avg));
#endif
}
|