]> git.sesse.net Git - ffmpeg/blob - libavfilter/opencl/convolution.cl
Merge commit '434b44cd6fb4bb9a2bf2bb29ef55ce1a315314b8'
[ffmpeg] / libavfilter / opencl / convolution.cl
1 /*
2  * Copyright (c) 2018 Danil Iashchenko
3  *
4  * This file is part of FFmpeg.
5  *
6  * FFmpeg is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU Lesser General Public
8  * License as published by the Free Software Foundation; either
9  * version 2.1 of the License, or (at your option) any later version.
10  *
11  * FFmpeg is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
14  * Lesser General Public License for more details.
15  *
16  * You should have received a copy of the GNU Lesser General Public
17  * License along with FFmpeg; if not, write to the Free Software
18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19  */
20
21 __kernel void convolution_global(__write_only image2d_t dst,
22                                  __read_only  image2d_t src,
23                                  int coef_matrix_dim,
24                                  __constant float *coef_matrix,
25                                  float div,
26                                  float bias)
27 {
28     const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
29                                CLK_ADDRESS_CLAMP_TO_EDGE   |
30                                CLK_FILTER_NEAREST);
31
32     const int half_matrix_dim = (coef_matrix_dim / 2);
33     int2 loc = (int2)(get_global_id(0), get_global_id(1));
34     float4 convPix = (float4)(0.0f, 0.0f, 0.0f, 0.0f);
35
36     for (int conv_i = -half_matrix_dim; conv_i <= half_matrix_dim; conv_i++) {
37         for (int conv_j = -half_matrix_dim; conv_j <= half_matrix_dim; conv_j++) {
38             float4 px = read_imagef(src, sampler, loc + (int2)(conv_j, conv_i));
39             convPix += px * coef_matrix[(conv_i + half_matrix_dim) * coef_matrix_dim +
40                                         (conv_j + half_matrix_dim)];
41         }
42      }
43      float4 dstPix = convPix * div + bias;
44      write_imagef(dst, loc, dstPix);
45 }