2 * This file is part of FFmpeg.
4 * FFmpeg is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public
6 * License as published by the Free Software Foundation; either
7 * version 2.1 of the License, or (at your option) any later version.
9 * FFmpeg is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with FFmpeg; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 #include "libavutil/common.h"
20 #include "libavutil/imgutils.h"
21 #include "libavutil/mem.h"
22 #include "libavutil/opt.h"
23 #include "libavutil/pixdesc.h"
28 #include "opencl_source.h"
31 #define MAX_DIAMETER 23
33 typedef struct UnsharpOpenCLContext {
34 OpenCLFilterContext ocf;
38 cl_command_queue command_queue;
51 float blur_x[MAX_DIAMETER];
52 float blur_y[MAX_DIAMETER];
63 } UnsharpOpenCLContext;
66 static int unsharp_opencl_init(AVFilterContext *avctx)
68 UnsharpOpenCLContext *ctx = avctx->priv;
72 err = ff_opencl_filter_load_program(avctx, &ff_opencl_source_unsharp, 1);
76 ctx->command_queue = clCreateCommandQueue(ctx->ocf.hwctx->context,
77 ctx->ocf.hwctx->device_id,
79 if (!ctx->command_queue) {
80 av_log(avctx, AV_LOG_ERROR, "Failed to create OpenCL "
81 "command queue: %d.\n", cle);
86 // Use global kernel if mask size will be too big for the local store..
87 ctx->global = (ctx->luma_size_x > 17.0f ||
88 ctx->luma_size_y > 17.0f ||
89 ctx->chroma_size_x > 17.0f ||
90 ctx->chroma_size_y > 17.0f);
92 ctx->kernel = clCreateKernel(ctx->ocf.program,
93 ctx->global ? "unsharp_global"
94 : "unsharp_local", &cle);
96 av_log(avctx, AV_LOG_ERROR, "Failed to create kernel: %d.\n", cle);
101 ctx->initialised = 1;
105 if (ctx->command_queue)
106 clReleaseCommandQueue(ctx->command_queue);
108 clReleaseKernel(ctx->kernel);
112 static int unsharp_opencl_make_filter_params(AVFilterContext *avctx)
114 UnsharpOpenCLContext *ctx = avctx->priv;
115 const AVPixFmtDescriptor *desc;
121 float diam_x, diam_y, amount;
122 int err, p, x, y, size_x, size_y;
124 desc = av_pix_fmt_desc_get(ctx->ocf.output_format);
127 for (p = 0; p < desc->nb_components; p++)
128 ctx->nb_planes = FFMAX(ctx->nb_planes, desc->comp[p].plane + 1);
130 for (p = 0; p < ctx->nb_planes; p++) {
131 if (p == 0 || (desc->flags & AV_PIX_FMT_FLAG_RGB)) {
132 diam_x = ctx->luma_size_x;
133 diam_y = ctx->luma_size_y;
134 amount = ctx->luma_amount;
136 diam_x = ctx->chroma_size_x;
137 diam_y = ctx->chroma_size_y;
138 amount = ctx->chroma_amount;
140 size_x = (int)ceil(diam_x) | 1;
141 size_y = (int)ceil(diam_y) | 1;
142 matrix_bytes = size_x * size_y * sizeof(float);
144 matrix = av_malloc(matrix_bytes);
146 err = AVERROR(ENOMEM);
151 for (x = 0; x < size_x; x++) {
152 double dx = (double)(x - size_x / 2) / diam_x;
153 sum += ctx->plane[p].blur_x[x] = exp(-16.0 * (dx * dx));
155 for (x = 0; x < size_x; x++)
156 ctx->plane[p].blur_x[x] /= sum;
159 for (y = 0; y < size_y; y++) {
160 double dy = (double)(y - size_y / 2) / diam_y;
161 sum += ctx->plane[p].blur_y[y] = exp(-16.0 * (dy * dy));
163 for (y = 0; y < size_y; y++)
164 ctx->plane[p].blur_y[y] /= sum;
166 for (y = 0; y < size_y; y++) {
167 for (x = 0; x < size_x; x++) {
168 val = ctx->plane[p].blur_x[x] * ctx->plane[p].blur_y[y];
169 matrix[y * size_x + x] = val;
174 buffer = clCreateBuffer(ctx->ocf.hwctx->context,
176 CL_MEM_COPY_HOST_PTR |
177 CL_MEM_HOST_NO_ACCESS,
178 matrix_bytes, matrix, &cle);
180 av_log(avctx, AV_LOG_ERROR, "Failed to create matrix buffer: "
185 ctx->plane[p].matrix = buffer;
187 buffer = clCreateBuffer(ctx->ocf.hwctx->context,
189 CL_MEM_COPY_HOST_PTR |
190 CL_MEM_HOST_NO_ACCESS,
191 sizeof(ctx->plane[p].blur_x),
192 ctx->plane[p].blur_x, &cle);
194 av_log(avctx, AV_LOG_ERROR, "Failed to create x-coef buffer: "
199 ctx->plane[p].coef_x = buffer;
201 buffer = clCreateBuffer(ctx->ocf.hwctx->context,
203 CL_MEM_COPY_HOST_PTR |
204 CL_MEM_HOST_NO_ACCESS,
205 sizeof(ctx->plane[p].blur_y),
206 ctx->plane[p].blur_y, &cle);
208 av_log(avctx, AV_LOG_ERROR, "Failed to create y-coef buffer: "
213 ctx->plane[p].coef_y = buffer;
218 ctx->plane[p].size_x = size_x;
219 ctx->plane[p].size_y = size_y;
220 ctx->plane[p].amount = amount;
229 static int unsharp_opencl_filter_frame(AVFilterLink *inlink, AVFrame *input)
231 AVFilterContext *avctx = inlink->dst;
232 AVFilterLink *outlink = avctx->outputs[0];
233 UnsharpOpenCLContext *ctx = avctx->priv;
234 AVFrame *output = NULL;
236 size_t global_work[2];
237 size_t local_work[2];
241 av_log(ctx, AV_LOG_DEBUG, "Filter input: %s, %ux%u (%"PRId64").\n",
242 av_get_pix_fmt_name(input->format),
243 input->width, input->height, input->pts);
245 if (!input->hw_frames_ctx)
246 return AVERROR(EINVAL);
248 if (!ctx->initialised) {
249 err = unsharp_opencl_init(avctx);
253 err = unsharp_opencl_make_filter_params(avctx);
258 output = ff_get_video_buffer(outlink, outlink->w, outlink->h);
260 err = AVERROR(ENOMEM);
264 for (p = 0; p < FF_ARRAY_ELEMS(output->data); p++) {
265 src = (cl_mem) input->data[p];
266 dst = (cl_mem)output->data[p];
271 cle = clSetKernelArg(ctx->kernel, 0, sizeof(cl_mem), &dst);
272 if (cle != CL_SUCCESS) {
273 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
274 "destination image argument: %d.\n", cle);
277 cle = clSetKernelArg(ctx->kernel, 1, sizeof(cl_mem), &src);
278 if (cle != CL_SUCCESS) {
279 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
280 "source image argument: %d.\n", cle);
283 cle = clSetKernelArg(ctx->kernel, 2, sizeof(cl_int), &ctx->plane[p].size_x);
284 if (cle != CL_SUCCESS) {
285 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
286 "matrix size argument: %d.\n", cle);
289 cle = clSetKernelArg(ctx->kernel, 3, sizeof(cl_int), &ctx->plane[p].size_y);
290 if (cle != CL_SUCCESS) {
291 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
292 "matrix size argument: %d.\n", cle);
295 cle = clSetKernelArg(ctx->kernel, 4, sizeof(cl_float), &ctx->plane[p].amount);
296 if (cle != CL_SUCCESS) {
297 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
298 "amount argument: %d.\n", cle);
302 cle = clSetKernelArg(ctx->kernel, 5, sizeof(cl_mem), &ctx->plane[p].matrix);
303 if (cle != CL_SUCCESS) {
304 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
305 "matrix argument: %d.\n", cle);
309 cle = clSetKernelArg(ctx->kernel, 5, sizeof(cl_mem), &ctx->plane[p].coef_x);
310 if (cle != CL_SUCCESS) {
311 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
312 "x-coef argument: %d.\n", cle);
315 cle = clSetKernelArg(ctx->kernel, 6, sizeof(cl_mem), &ctx->plane[p].coef_y);
316 if (cle != CL_SUCCESS) {
317 av_log(avctx, AV_LOG_ERROR, "Failed to set kernel "
318 "y-coef argument: %d.\n", cle);
323 err = ff_opencl_filter_work_size_from_image(avctx, global_work, output, p,
324 ctx->global ? 0 : 16);
331 av_log(avctx, AV_LOG_DEBUG, "Run kernel on plane %d "
332 "(%"SIZE_SPECIFIER"x%"SIZE_SPECIFIER").\n",
333 p, global_work[0], global_work[1]);
335 cle = clEnqueueNDRangeKernel(ctx->command_queue, ctx->kernel, 2, NULL,
336 global_work, ctx->global ? NULL : local_work,
338 if (cle != CL_SUCCESS) {
339 av_log(avctx, AV_LOG_ERROR, "Failed to enqueue kernel: %d.\n",
346 cle = clFinish(ctx->command_queue);
347 if (cle != CL_SUCCESS) {
348 av_log(avctx, AV_LOG_ERROR, "Failed to finish command queue: %d.\n",
354 err = av_frame_copy_props(output, input);
358 av_frame_free(&input);
360 av_log(ctx, AV_LOG_DEBUG, "Filter output: %s, %ux%u (%"PRId64").\n",
361 av_get_pix_fmt_name(output->format),
362 output->width, output->height, output->pts);
364 return ff_filter_frame(outlink, output);
367 clFinish(ctx->command_queue);
368 av_frame_free(&input);
369 av_frame_free(&output);
373 static av_cold void unsharp_opencl_uninit(AVFilterContext *avctx)
375 UnsharpOpenCLContext *ctx = avctx->priv;
379 for (i = 0; i < ctx->nb_planes; i++) {
380 if (ctx->plane[i].matrix)
381 clReleaseMemObject(ctx->plane[i].matrix);
382 if (ctx->plane[i].coef_x)
383 clReleaseMemObject(ctx->plane[i].coef_x);
384 if (ctx->plane[i].coef_y)
385 clReleaseMemObject(ctx->plane[i].coef_y);
389 cle = clReleaseKernel(ctx->kernel);
390 if (cle != CL_SUCCESS)
391 av_log(avctx, AV_LOG_ERROR, "Failed to release "
392 "kernel: %d.\n", cle);
395 if (ctx->command_queue) {
396 cle = clReleaseCommandQueue(ctx->command_queue);
397 if (cle != CL_SUCCESS)
398 av_log(avctx, AV_LOG_ERROR, "Failed to release "
399 "command queue: %d.\n", cle);
402 ff_opencl_filter_uninit(avctx);
405 #define OFFSET(x) offsetof(UnsharpOpenCLContext, x)
406 #define FLAGS (AV_OPT_FLAG_FILTERING_PARAM | AV_OPT_FLAG_VIDEO_PARAM)
407 static const AVOption unsharp_opencl_options[] = {
408 { "luma_msize_x", "Set luma mask horizontal diameter (pixels)",
409 OFFSET(luma_size_x), AV_OPT_TYPE_FLOAT,
410 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
411 { "lx", "Set luma mask horizontal diameter (pixels)",
412 OFFSET(luma_size_x), AV_OPT_TYPE_FLOAT,
413 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
414 { "luma_msize_y", "Set luma mask vertical diameter (pixels)",
415 OFFSET(luma_size_y), AV_OPT_TYPE_FLOAT,
416 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
417 { "ly", "Set luma mask vertical diameter (pixels)",
418 OFFSET(luma_size_y), AV_OPT_TYPE_FLOAT,
419 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
420 { "luma_amount", "Set luma amount (multiplier)",
421 OFFSET(luma_amount), AV_OPT_TYPE_FLOAT,
422 { .dbl = 1.0 }, -10, 10, FLAGS },
423 { "la", "Set luma amount (multiplier)",
424 OFFSET(luma_amount), AV_OPT_TYPE_FLOAT,
425 { .dbl = 1.0 }, -10, 10, FLAGS },
427 { "chroma_msize_x", "Set chroma mask horizontal diameter (pixels after subsampling)",
428 OFFSET(chroma_size_x), AV_OPT_TYPE_FLOAT,
429 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
430 { "cx", "Set chroma mask horizontal diameter (pixels after subsampling)",
431 OFFSET(chroma_size_x), AV_OPT_TYPE_FLOAT,
432 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
433 { "chroma_msize_y", "Set chroma mask vertical diameter (pixels after subsampling)",
434 OFFSET(chroma_size_y), AV_OPT_TYPE_FLOAT,
435 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
436 { "cy", "Set chroma mask vertical diameter (pixels after subsampling)",
437 OFFSET(chroma_size_y), AV_OPT_TYPE_FLOAT,
438 { .dbl = 5.0 }, 1, MAX_DIAMETER, FLAGS },
439 { "chroma_amount", "Set chroma amount (multiplier)",
440 OFFSET(chroma_amount), AV_OPT_TYPE_FLOAT,
441 { .dbl = 0.0 }, -10, 10, FLAGS },
442 { "ca", "Set chroma amount (multiplier)",
443 OFFSET(chroma_amount), AV_OPT_TYPE_FLOAT,
444 { .dbl = 0.0 }, -10, 10, FLAGS },
449 AVFILTER_DEFINE_CLASS(unsharp_opencl);
451 static const AVFilterPad unsharp_opencl_inputs[] = {
454 .type = AVMEDIA_TYPE_VIDEO,
455 .filter_frame = &unsharp_opencl_filter_frame,
456 .config_props = &ff_opencl_filter_config_input,
461 static const AVFilterPad unsharp_opencl_outputs[] = {
464 .type = AVMEDIA_TYPE_VIDEO,
465 .config_props = &ff_opencl_filter_config_output,
470 AVFilter ff_vf_unsharp_opencl = {
471 .name = "unsharp_opencl",
472 .description = NULL_IF_CONFIG_SMALL("Apply unsharp mask to input video"),
473 .priv_size = sizeof(UnsharpOpenCLContext),
474 .priv_class = &unsharp_opencl_class,
475 .init = &ff_opencl_filter_init,
476 .uninit = &unsharp_opencl_uninit,
477 .query_formats = &ff_opencl_filter_query_formats,
478 .inputs = unsharp_opencl_inputs,
479 .outputs = unsharp_opencl_outputs,
480 .flags_internal = FF_FILTER_FLAG_HWFRAME_AWARE,