oximedia-gpu 0.1.8

GPU compute pipeline using WGPU for OxiMedia - cross-platform acceleration
Documentation 
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
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184

// Image scaling shaders with bilinear and bicubic interpolation

@group(0) @binding(0) var<storage, read> input_image: array<u32>;
@group(0) @binding(1) var<storage, read_write> output_image: array<u32>;
@group(0) @binding(2) var<uniform> params: ScaleParams;

struct ScaleParams {
    src_width: u32,
    src_height: u32,
    dst_width: u32,
    dst_height: u32,
    src_stride: u32,
    dst_stride: u32,
    filter_type: u32, // 0=nearest, 1=bilinear, 2=bicubic
    padding: u32,
}

fn unpack_rgba(packed: u32) -> vec4<f32> {
    let r = f32((packed >> 24u) & 0xFFu) / 255.0;
    let g = f32((packed >> 16u) & 0xFFu) / 255.0;
    let b = f32((packed >> 8u) & 0xFFu) / 255.0;
    let a = f32(packed & 0xFFu) / 255.0;
    return vec4<f32>(r, g, b, a);
}

fn pack_rgba(color: vec4<f32>) -> u32 {
    let r = u32(clamp(color.r * 255.0, 0.0, 255.0));
    let g = u32(clamp(color.g * 255.0, 0.0, 255.0));
    let b = u32(clamp(color.b * 255.0, 0.0, 255.0));
    let a = u32(clamp(color.a * 255.0, 0.0, 255.0));
    return (r << 24u) | (g << 16u) | (b << 8u) | a;
}

fn sample_image(x: i32, y: i32) -> vec4<f32> {
    let cx = clamp(x, 0, i32(params.src_width) - 1);
    let cy = clamp(y, 0, i32(params.src_height) - 1);
    let idx = u32(cy) * params.src_stride + u32(cx);
    return unpack_rgba(input_image[idx]);
}

// Nearest neighbor sampling
fn sample_nearest(u: f32, v: f32) -> vec4<f32> {
    let x = i32(u);
    let y = i32(v);
    return sample_image(x, y);
}

// Bilinear interpolation
fn sample_bilinear(u: f32, v: f32) -> vec4<f32> {
    let x = u - 0.5;
    let y = v - 0.5;

    let x0 = i32(floor(x));
    let y0 = i32(floor(y));
    let x1 = x0 + 1;
    let y1 = y0 + 1;

    let fx = fract(x);
    let fy = fract(y);

    let c00 = sample_image(x0, y0);
    let c10 = sample_image(x1, y0);
    let c01 = sample_image(x0, y1);
    let c11 = sample_image(x1, y1);

    let c0 = mix(c00, c10, fx);
    let c1 = mix(c01, c11, fx);

    return mix(c0, c1, fy);
}

// Cubic interpolation kernel (Catmull-Rom)
fn cubic_weight(t: f32) -> f32 {
    let a = -0.5;
    let t2 = t * t;
    let t3 = t2 * t;

    if (t < 1.0) {
        return (a + 2.0) * t3 - (a + 3.0) * t2 + 1.0;
    } else if (t < 2.0) {
        return a * t3 - 5.0 * a * t2 + 8.0 * a * t - 4.0 * a;
    }
    return 0.0;
}

// Bicubic interpolation
fn sample_bicubic(u: f32, v: f32) -> vec4<f32> {
    let x = u - 0.5;
    let y = v - 0.5;

    let x0 = i32(floor(x));
    let y0 = i32(floor(y));

    let fx = fract(x);
    let fy = fract(y);

    var result = vec4<f32>(0.0, 0.0, 0.0, 0.0);
    var weight_sum = 0.0;

    for (var dy = -1; dy <= 2; dy = dy + 1) {
        for (var dx = -1; dx <= 2; dx = dx + 1) {
            let sample_x = x0 + dx;
            let sample_y = y0 + dy;

            let wx = cubic_weight(abs(f32(dx) - fx));
            let wy = cubic_weight(abs(f32(dy) - fy));
            let w = wx * wy;

            result += sample_image(sample_x, sample_y) * w;
            weight_sum += w;
        }
    }

    if (weight_sum > 0.0) {
        result /= weight_sum;
    }

    return result;
}

@compute @workgroup_size(16, 16, 1)
fn scale_main(@builtin(global_invocation_id) global_id: vec3<u32>) {
    let x = global_id.x;
    let y = global_id.y;

    if (x >= params.dst_width || y >= params.dst_height) {
        return;
    }

    // Calculate source coordinates
    let u = (f32(x) + 0.5) * f32(params.src_width) / f32(params.dst_width);
    let v = (f32(y) + 0.5) * f32(params.src_height) / f32(params.dst_height);

    var color: vec4<f32>;

    if (params.filter_type == 0u) {
        color = sample_nearest(u, v);
    } else if (params.filter_type == 1u) {
        color = sample_bilinear(u, v);
    } else {
        color = sample_bicubic(u, v);
    }

    let dst_idx = y * params.dst_stride + x;
    output_image[dst_idx] = pack_rgba(color);
}

// Downscale with area averaging (box filter)
@compute @workgroup_size(16, 16, 1)
fn downscale_area(@builtin(global_invocation_id) global_id: vec3<u32>) {
    let x = global_id.x;
    let y = global_id.y;

    if (x >= params.dst_width || y >= params.dst_height) {
        return;
    }

    let scale_x = f32(params.src_width) / f32(params.dst_width);
    let scale_y = f32(params.src_height) / f32(params.dst_height);

    let src_x_start = f32(x) * scale_x;
    let src_y_start = f32(y) * scale_y;
    let src_x_end = src_x_start + scale_x;
    let src_y_end = src_y_start + scale_y;

    let x_start = i32(floor(src_x_start));
    let y_start = i32(floor(src_y_start));
    let x_end = i32(ceil(src_x_end));
    let y_end = i32(ceil(src_y_end));

    var sum = vec4<f32>(0.0, 0.0, 0.0, 0.0);
    var count = 0.0;

    for (var py = y_start; py < y_end; py = py + 1) {
        for (var px = x_start; px < x_end; px = px + 1) {
            sum += sample_image(px, py);
            count += 1.0;
        }
    }

    let color = sum / count;
    let dst_idx = y * params.dst_stride + x;
    output_image[dst_idx] = pack_rgba(color);
}