blob: dd45d6f60bab2c3283033e1d7f284a2c068f9a6c
1 | /* |
2 | * Copyright (C) 2013 Wei Gao <weigao@multicorewareinc.com> |
3 | * Copyright (C) 2013 Lenny Wang |
4 | * |
5 | * |
6 | * This file is part of FFmpeg. |
7 | * |
8 | * FFmpeg is free software; you can redistribute it and/or |
9 | * modify it under the terms of the GNU Lesser General Public |
10 | * License as published by the Free Software Foundation; either |
11 | * version 2.1 of the License, or (at your option) any later version. |
12 | * |
13 | * FFmpeg is distributed in the hope that it will be useful, |
14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
16 | * Lesser General Public License for more details. |
17 | * |
18 | * You should have received a copy of the GNU Lesser General Public |
19 | * License along with FFmpeg; if not, write to the Free Software |
20 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
21 | */ |
22 | |
23 | #ifndef AVFILTER_DESHAKE_OPENCL_KERNEL_H |
24 | #define AVFILTER_DESHAKE_OPENCL_KERNEL_H |
25 | |
26 | #include "libavutil/opencl.h" |
27 | |
28 | const char *ff_kernel_deshake_opencl = AV_OPENCL_KERNEL( |
29 | inline unsigned char pixel(global const unsigned char *src, int x, int y, |
30 | int w, int h,int stride, unsigned char def) |
31 | { |
32 | return (x < 0 || y < 0 || x >= w || y >= h) ? def : src[x + y * stride]; |
33 | } |
34 | |
35 | unsigned char interpolate_nearest(float x, float y, global const unsigned char *src, |
36 | int width, int height, int stride, unsigned char def) |
37 | { |
38 | return pixel(src, (int)(x + 0.5f), (int)(y + 0.5f), width, height, stride, def); |
39 | } |
40 | |
41 | unsigned char interpolate_bilinear(float x, float y, global const unsigned char *src, |
42 | int width, int height, int stride, unsigned char def) |
43 | { |
44 | int x_c, x_f, y_c, y_f; |
45 | int v1, v2, v3, v4; |
46 | x_f = (int)x; |
47 | y_f = (int)y; |
48 | x_c = x_f + 1; |
49 | y_c = y_f + 1; |
50 | |
51 | if (x_f < -1 || x_f > width || y_f < -1 || y_f > height) { |
52 | return def; |
53 | } else { |
54 | v4 = pixel(src, x_f, y_f, width, height, stride, def); |
55 | v2 = pixel(src, x_c, y_f, width, height, stride, def); |
56 | v3 = pixel(src, x_f, y_c, width, height, stride, def); |
57 | v1 = pixel(src, x_c, y_c, width, height, stride, def); |
58 | return (v1*(x - x_f)*(y - y_f) + v2*((x - x_f)*(y_c - y)) + |
59 | v3*(x_c - x)*(y - y_f) + v4*((x_c - x)*(y_c - y))); |
60 | } |
61 | } |
62 | |
63 | unsigned char interpolate_biquadratic(float x, float y, global const unsigned char *src, |
64 | int width, int height, int stride, unsigned char def) |
65 | { |
66 | int x_c, x_f, y_c, y_f; |
67 | unsigned char v1, v2, v3, v4; |
68 | float f1, f2, f3, f4; |
69 | x_f = (int)x; |
70 | y_f = (int)y; |
71 | x_c = x_f + 1; |
72 | y_c = y_f + 1; |
73 | |
74 | if (x_f < - 1 || x_f > width || y_f < -1 || y_f > height) |
75 | return def; |
76 | else { |
77 | v4 = pixel(src, x_f, y_f, width, height, stride, def); |
78 | v2 = pixel(src, x_c, y_f, width, height, stride, def); |
79 | v3 = pixel(src, x_f, y_c, width, height, stride, def); |
80 | v1 = pixel(src, x_c, y_c, width, height, stride, def); |
81 | |
82 | f1 = 1 - sqrt((x_c - x) * (y_c - y)); |
83 | f2 = 1 - sqrt((x_c - x) * (y - y_f)); |
84 | f3 = 1 - sqrt((x - x_f) * (y_c - y)); |
85 | f4 = 1 - sqrt((x - x_f) * (y - y_f)); |
86 | return (v1 * f1 + v2 * f2 + v3 * f3 + v4 * f4) / (f1 + f2 + f3 + f4); |
87 | } |
88 | } |
89 | |
90 | inline const float clipf(float a, float amin, float amax) |
91 | { |
92 | if (a < amin) return amin; |
93 | else if (a > amax) return amax; |
94 | else return a; |
95 | } |
96 | |
97 | inline int mirror(int v, int m) |
98 | { |
99 | while ((unsigned)v > (unsigned)m) { |
100 | v = -v; |
101 | if (v < 0) |
102 | v += 2 * m; |
103 | } |
104 | return v; |
105 | } |
106 | |
107 | kernel void avfilter_transform_luma(global unsigned char *src, |
108 | global unsigned char *dst, |
109 | float4 matrix, |
110 | int interpolate, |
111 | int fill, |
112 | int src_stride_lu, |
113 | int dst_stride_lu, |
114 | int height, |
115 | int width) |
116 | { |
117 | int x = get_global_id(0); |
118 | int y = get_global_id(1); |
119 | int idx_dst = y * dst_stride_lu + x; |
120 | unsigned char def = 0; |
121 | float x_s = x * matrix.x + y * matrix.y + matrix.z; |
122 | float y_s = x * (-matrix.y) + y * matrix.x + matrix.w; |
123 | |
124 | if (x < width && y < height) { |
125 | switch (fill) { |
126 | case 0: //FILL_BLANK |
127 | def = 0; |
128 | break; |
129 | case 1: //FILL_ORIGINAL |
130 | def = src[y*src_stride_lu + x]; |
131 | break; |
132 | case 2: //FILL_CLAMP |
133 | y_s = clipf(y_s, 0, height - 1); |
134 | x_s = clipf(x_s, 0, width - 1); |
135 | def = src[(int)y_s * src_stride_lu + (int)x_s]; |
136 | break; |
137 | case 3: //FILL_MIRROR |
138 | y_s = mirror(y_s, height - 1); |
139 | x_s = mirror(x_s, width - 1); |
140 | def = src[(int)y_s * src_stride_lu + (int)x_s]; |
141 | break; |
142 | } |
143 | switch (interpolate) { |
144 | case 0: //INTERPOLATE_NEAREST |
145 | dst[idx_dst] = interpolate_nearest(x_s, y_s, src, width, height, src_stride_lu, def); |
146 | break; |
147 | case 1: //INTERPOLATE_BILINEAR |
148 | dst[idx_dst] = interpolate_bilinear(x_s, y_s, src, width, height, src_stride_lu, def); |
149 | break; |
150 | case 2: //INTERPOLATE_BIQUADRATIC |
151 | dst[idx_dst] = interpolate_biquadratic(x_s, y_s, src, width, height, src_stride_lu, def); |
152 | break; |
153 | default: |
154 | return; |
155 | } |
156 | } |
157 | } |
158 | |
159 | kernel void avfilter_transform_chroma(global unsigned char *src, |
160 | global unsigned char *dst, |
161 | float4 matrix, |
162 | int interpolate, |
163 | int fill, |
164 | int src_stride_lu, |
165 | int dst_stride_lu, |
166 | int src_stride_ch, |
167 | int dst_stride_ch, |
168 | int height, |
169 | int width, |
170 | int ch, |
171 | int cw) |
172 | { |
173 | |
174 | int x = get_global_id(0); |
175 | int y = get_global_id(1); |
176 | int pad_ch = get_global_size(1)>>1; |
177 | global unsigned char *dst_u = dst + height * dst_stride_lu; |
178 | global unsigned char *src_u = src + height * src_stride_lu; |
179 | global unsigned char *dst_v = dst_u + ch * dst_stride_ch; |
180 | global unsigned char *src_v = src_u + ch * src_stride_ch; |
181 | src = y < pad_ch ? src_u : src_v; |
182 | dst = y < pad_ch ? dst_u : dst_v; |
183 | y = select(y - pad_ch, y, y < pad_ch); |
184 | float x_s = x * matrix.x + y * matrix.y + matrix.z; |
185 | float y_s = x * (-matrix.y) + y * matrix.x + matrix.w; |
186 | int idx_dst = y * dst_stride_ch + x; |
187 | unsigned char def; |
188 | |
189 | if (x < cw && y < ch) { |
190 | switch (fill) { |
191 | case 0: //FILL_BLANK |
192 | def = 0; |
193 | break; |
194 | case 1: //FILL_ORIGINAL |
195 | def = src[y*src_stride_ch + x]; |
196 | break; |
197 | case 2: //FILL_CLAMP |
198 | y_s = clipf(y_s, 0, ch - 1); |
199 | x_s = clipf(x_s, 0, cw - 1); |
200 | def = src[(int)y_s * src_stride_ch + (int)x_s]; |
201 | break; |
202 | case 3: //FILL_MIRROR |
203 | y_s = mirror(y_s, ch - 1); |
204 | x_s = mirror(x_s, cw - 1); |
205 | def = src[(int)y_s * src_stride_ch + (int)x_s]; |
206 | break; |
207 | } |
208 | switch (interpolate) { |
209 | case 0: //INTERPOLATE_NEAREST |
210 | dst[idx_dst] = interpolate_nearest(x_s, y_s, src, cw, ch, src_stride_ch, def); |
211 | break; |
212 | case 1: //INTERPOLATE_BILINEAR |
213 | dst[idx_dst] = interpolate_bilinear(x_s, y_s, src, cw, ch, src_stride_ch, def); |
214 | break; |
215 | case 2: //INTERPOLATE_BIQUADRATIC |
216 | dst[idx_dst] = interpolate_biquadratic(x_s, y_s, src, cw, ch, src_stride_ch, def); |
217 | break; |
218 | default: |
219 | return; |
220 | } |
221 | } |
222 | } |
223 | ); |
224 | |
225 | #endif /* AVFILTER_DESHAKE_OPENCL_KERNEL_H */ |
226 |