vf_overlay.c 14 KB
Newer Older
1 2 3 4 5
/*
 * Copyright (c) 2010 Stefano Sabatini
 * Copyright (c) 2010 Baptiste Coudurier
 * Copyright (c) 2007 Bobby Bingham
 *
6
 * This file is part of Libav.
7
 *
8
 * Libav is free software; you can redistribute it and/or
9 10 11 12
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
13
 * Libav is distributed in the hope that it will be useful,
14 15 16 17 18
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
19
 * License along with Libav; if not, write to the Free Software
20 21 22 23 24 25 26 27 28
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/**
 * @file
 * overlay one video on top of another
 */

#include "avfilter.h"
29
#include "formats.h"
30
#include "libavutil/common.h"
31 32
#include "libavutil/eval.h"
#include "libavutil/avstring.h"
33
#include "libavutil/avassert.h"
34
#include "libavutil/pixdesc.h"
35
#include "libavutil/imgutils.h"
36
#include "libavutil/mathematics.h"
37
#include "libavutil/opt.h"
38
#include "internal.h"
39
#include "video.h"
40

41
static const char *const var_names[] = {
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
    "E",
    "PHI",
    "PI",
    "main_w",    "W", ///< width  of the main    video
    "main_h",    "H", ///< height of the main    video
    "overlay_w", "w", ///< width  of the overlay video
    "overlay_h", "h", ///< height of the overlay video
    NULL
};

enum var_name {
    VAR_E,
    VAR_PHI,
    VAR_PI,
    VAR_MAIN_W,    VAR_MW,
    VAR_MAIN_H,    VAR_MH,
    VAR_OVERLAY_W, VAR_OW,
    VAR_OVERLAY_H, VAR_OH,
    VAR_VARS_NB
};

63 64 65 66 67 68 69 70 71 72
enum EOFAction {
    EOF_ACTION_REPEAT,
    EOF_ACTION_ENDALL,
    EOF_ACTION_PASS
};

static const char *eof_action_str[] = {
    "repeat", "endall", "pass"
};

73 74 75
#define MAIN    0
#define OVERLAY 1

76
typedef struct OverlayContext {
77
    const AVClass *class;
78 79 80 81 82
    int x, y;                   ///< position of overlayed picture

    int max_plane_step[4];      ///< steps per pixel for each plane
    int hsub, vsub;             ///< chroma subsampling values

83
    char *x_expr, *y_expr;
84

85 86
    enum EOFAction eof_action;  ///< action to take on EOF from source

87 88
    AVFrame *main;
    AVFrame *over_prev, *over_next;
89 90 91 92
} OverlayContext;

static av_cold void uninit(AVFilterContext *ctx)
{
93
    OverlayContext *s = ctx->priv;
94

95 96 97
    av_frame_free(&s->main);
    av_frame_free(&s->over_prev);
    av_frame_free(&s->over_next);
98 99 100 101
}

static int query_formats(AVFilterContext *ctx)
{
102 103
    const enum AVPixelFormat inout_pix_fmts[] = { AV_PIX_FMT_YUV420P,  AV_PIX_FMT_NONE };
    const enum AVPixelFormat blend_pix_fmts[] = { AV_PIX_FMT_YUVA420P, AV_PIX_FMT_NONE };
104 105
    AVFilterFormats *inout_formats = ff_make_format_list(inout_pix_fmts);
    AVFilterFormats *blend_formats = ff_make_format_list(blend_pix_fmts);
106

107 108 109
    ff_formats_ref(inout_formats, &ctx->inputs [MAIN   ]->out_formats);
    ff_formats_ref(blend_formats, &ctx->inputs [OVERLAY]->out_formats);
    ff_formats_ref(inout_formats, &ctx->outputs[MAIN   ]->in_formats );
110 111 112 113 114 115

    return 0;
}

static int config_input_main(AVFilterLink *inlink)
{
116
    OverlayContext *s = inlink->dst->priv;
117
    const AVPixFmtDescriptor *pix_desc = av_pix_fmt_desc_get(inlink->format);
118

119 120 121
    av_image_fill_max_pixsteps(s->max_plane_step, NULL, pix_desc);
    s->hsub = pix_desc->log2_chroma_w;
    s->vsub = pix_desc->log2_chroma_h;
122 123 124 125 126 127 128

    return 0;
}

static int config_input_overlay(AVFilterLink *inlink)
{
    AVFilterContext *ctx  = inlink->dst;
129
    OverlayContext  *s = inlink->dst->priv;
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144
    char *expr;
    double var_values[VAR_VARS_NB], res;
    int ret;

    /* Finish the configuration by evaluating the expressions
       now when both inputs are configured. */
    var_values[VAR_E  ] = M_E;
    var_values[VAR_PHI] = M_PHI;
    var_values[VAR_PI ] = M_PI;

    var_values[VAR_MAIN_W   ] = var_values[VAR_MW] = ctx->inputs[MAIN   ]->w;
    var_values[VAR_MAIN_H   ] = var_values[VAR_MH] = ctx->inputs[MAIN   ]->h;
    var_values[VAR_OVERLAY_W] = var_values[VAR_OW] = ctx->inputs[OVERLAY]->w;
    var_values[VAR_OVERLAY_H] = var_values[VAR_OH] = ctx->inputs[OVERLAY]->h;

145
    if ((ret = av_expr_parse_and_eval(&res, (expr = s->x_expr), var_names, var_values,
146 147
                                      NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
        goto fail;
148 149
    s->x = res;
    if ((ret = av_expr_parse_and_eval(&res, (expr = s->y_expr), var_names, var_values,
150 151
                                      NULL, NULL, NULL, NULL, NULL, 0, ctx)))
        goto fail;
152
    s->y = res;
153
    /* x may depend on y */
154
    if ((ret = av_expr_parse_and_eval(&res, (expr = s->x_expr), var_names, var_values,
155 156
                                      NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
        goto fail;
157
    s->x = res;
158

159
    av_log(ctx, AV_LOG_VERBOSE,
160
           "main w:%d h:%d fmt:%s overlay x:%d y:%d w:%d h:%d fmt:%s eof_action:%s\n",
161
           ctx->inputs[MAIN]->w, ctx->inputs[MAIN]->h,
162
           av_get_pix_fmt_name(ctx->inputs[MAIN]->format),
163
           s->x, s->y,
164
           ctx->inputs[OVERLAY]->w, ctx->inputs[OVERLAY]->h,
165 166
           av_get_pix_fmt_name(ctx->inputs[OVERLAY]->format),
           eof_action_str[s->eof_action]);
167

168 169 170
    if (s->x < 0 || s->y < 0 ||
        s->x + var_values[VAR_OVERLAY_W] > var_values[VAR_MAIN_W] ||
        s->y + var_values[VAR_OVERLAY_H] > var_values[VAR_MAIN_H]) {
171 172
        av_log(ctx, AV_LOG_ERROR,
               "Overlay area (%d,%d)<->(%d,%d) not within the main area (0,0)<->(%d,%d) or zero-sized\n",
173 174 175
               s->x, s->y,
               (int)(s->x + var_values[VAR_OVERLAY_W]),
               (int)(s->y + var_values[VAR_OVERLAY_H]),
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192
               (int)var_values[VAR_MAIN_W], (int)var_values[VAR_MAIN_H]);
        return AVERROR(EINVAL);
    }
    return 0;

fail:
    av_log(NULL, AV_LOG_ERROR,
           "Error when evaluating the expression '%s'\n", expr);
    return ret;
}

static int config_output(AVFilterLink *outlink)
{
    AVFilterContext *ctx = outlink->src;

    outlink->w = ctx->inputs[MAIN]->w;
    outlink->h = ctx->inputs[MAIN]->h;
193
    outlink->time_base = ctx->inputs[MAIN]->time_base;
194 195 196 197

    return 0;
}

198
static void blend_frame(AVFilterContext *ctx,
199
                        AVFrame *dst, AVFrame *src,
200
                        int x, int y)
201
{
202
    OverlayContext *s = ctx->priv;
203 204
    int i, j, k;
    int width, height;
205
    int overlay_end_y = y + src->height;
206 207
    int end_y, start_y;

208 209
    width = FFMIN(dst->width - x, src->width);
    end_y = FFMIN(dst->height, overlay_end_y);
210
    start_y = FFMAX(y, 0);
211 212
    height = end_y - start_y;

213
    if (dst->format == AV_PIX_FMT_BGR24 || dst->format == AV_PIX_FMT_RGB24) {
214 215
        uint8_t *dp = dst->data[0] + x * 3 + start_y * dst->linesize[0];
        uint8_t *sp = src->data[0];
216 217
        int b = dst->format == AV_PIX_FMT_BGR24 ? 2 : 0;
        int r = dst->format == AV_PIX_FMT_BGR24 ? 0 : 2;
218 219
        if (y < 0)
            sp += -y * src->linesize[0];
220 221 222 223 224 225 226 227 228 229 230 231 232 233
        for (i = 0; i < height; i++) {
            uint8_t *d = dp, *s = sp;
            for (j = 0; j < width; j++) {
                d[r] = (d[r] * (0xff - s[3]) + s[0] * s[3] + 128) >> 8;
                d[1] = (d[1] * (0xff - s[3]) + s[1] * s[3] + 128) >> 8;
                d[b] = (d[b] * (0xff - s[3]) + s[2] * s[3] + 128) >> 8;
                d += 3;
                s += 4;
            }
            dp += dst->linesize[0];
            sp += src->linesize[0];
        }
    } else {
        for (i = 0; i < 3; i++) {
234 235
            int hsub = i ? s->hsub : 0;
            int vsub = i ? s->vsub : 0;
236 237 238 239 240 241
            uint8_t *dp = dst->data[i] + (x >> hsub) +
                (start_y >> vsub) * dst->linesize[i];
            uint8_t *sp = src->data[i];
            uint8_t *ap = src->data[3];
            int wp = FFALIGN(width, 1<<hsub) >> hsub;
            int hp = FFALIGN(height, 1<<vsub) >> vsub;
242 243 244
            if (y < 0) {
                sp += ((-y) >> vsub) * src->linesize[i];
                ap += -y * src->linesize[3];
245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
            }
            for (j = 0; j < hp; j++) {
                uint8_t *d = dp, *s = sp, *a = ap;
                for (k = 0; k < wp; k++) {
                    // average alpha for color components, improve quality
                    int alpha_v, alpha_h, alpha;
                    if (hsub && vsub && j+1 < hp && k+1 < wp) {
                        alpha = (a[0] + a[src->linesize[3]] +
                                 a[1] + a[src->linesize[3]+1]) >> 2;
                    } else if (hsub || vsub) {
                        alpha_h = hsub && k+1 < wp ?
                            (a[0] + a[1]) >> 1 : a[0];
                        alpha_v = vsub && j+1 < hp ?
                            (a[0] + a[src->linesize[3]]) >> 1 : a[0];
                        alpha = (alpha_v + alpha_h) >> 1;
                    } else
                        alpha = a[0];
                    *d = (*d * (0xff - alpha) + *s++ * alpha + 128) >> 8;
                    d++;
                    a += 1 << hsub;
                }
                dp += dst->linesize[i];
                sp += src->linesize[i];
                ap += (1 << vsub) * src->linesize[3];
            }
        }
    }
}

274
static int filter_frame_main(AVFilterLink *inlink, AVFrame *frame)
275
{
276 277 278
    OverlayContext *s = inlink->dst->priv;

    av_assert0(!s->main);
279
    s->main         = frame;
280

281 282
    return 0;
}
283

284
static int filter_frame_overlay(AVFilterLink *inlink, AVFrame *frame)
285
{
286 287 288
    OverlayContext *s = inlink->dst->priv;

    av_assert0(!s->over_next);
289
    s->over_next    = frame;
290

291 292
    return 0;
}
293

294 295 296 297
static int output_frame(AVFilterContext *ctx)
{
    OverlayContext *s = ctx->priv;
    AVFilterLink *outlink = ctx->outputs[0];
298
    int ret = ff_filter_frame(outlink, s->main);
299 300 301 302 303 304 305 306
    s->main = NULL;

    return ret;
}

static int handle_overlay_eof(AVFilterContext *ctx)
{
    OverlayContext *s = ctx->priv;
307 308
    /* Repeat previous frame on secondary input */
    if (s->over_prev && s->eof_action == EOF_ACTION_REPEAT)
309
        blend_frame(ctx, s->main, s->over_prev, s->x, s->y);
310 311 312
    /* End both streams */
    else if (s->eof_action == EOF_ACTION_ENDALL)
        return AVERROR_EOF;
313 314 315 316
    return output_frame(ctx);
}

static int request_frame(AVFilterLink *outlink)
317
{
318 319 320 321 322 323 324 325 326 327 328 329
    AVFilterContext *ctx = outlink->src;
    OverlayContext    *s = ctx->priv;
    AVRational tb_main = ctx->inputs[MAIN]->time_base;
    AVRational tb_over = ctx->inputs[OVERLAY]->time_base;
    int ret = 0;

    /* get a frame on the main input */
    if (!s->main) {
        ret = ff_request_frame(ctx->inputs[MAIN]);
        if (ret < 0)
            return ret;
    }
330

331
    /* get a new frame on the overlay input, on EOF check setting 'eof_action' */
332 333 334 335 336 337 338 339 340 341 342
    if (!s->over_next) {
        ret = ff_request_frame(ctx->inputs[OVERLAY]);
        if (ret == AVERROR_EOF)
           return handle_overlay_eof(ctx);
        else if (ret < 0)
            return ret;
    }

    while (s->main->pts != AV_NOPTS_VALUE &&
           s->over_next->pts != AV_NOPTS_VALUE &&
           av_compare_ts(s->over_next->pts, tb_over, s->main->pts, tb_main) < 0) {
343 344
        av_frame_free(&s->over_prev);
        FFSWAP(AVFrame*, s->over_prev, s->over_next);
345 346 347 348 349 350 351

        ret = ff_request_frame(ctx->inputs[OVERLAY]);
        if (ret == AVERROR_EOF)
            return handle_overlay_eof(ctx);
        else if (ret < 0)
            return ret;
    }
352

353 354 355 356
    if (s->main->pts == AV_NOPTS_VALUE ||
        s->over_next->pts == AV_NOPTS_VALUE ||
        !av_compare_ts(s->over_next->pts, tb_over, s->main->pts, tb_main)) {
        blend_frame(ctx, s->main, s->over_next, s->x, s->y);
357 358
        av_frame_free(&s->over_prev);
        FFSWAP(AVFrame*, s->over_prev, s->over_next);
359 360 361 362 363
    } else if (s->over_prev) {
        blend_frame(ctx, s->main, s->over_prev, s->x, s->y);
    }

    return output_frame(ctx);
364 365
}

366 367 368 369 370 371 372
#define OFFSET(x) offsetof(OverlayContext, x)
#define FLAGS AV_OPT_FLAG_VIDEO_PARAM
static const AVOption options[] = {
    { "x", "Horizontal position of the left edge of the overlaid video on the "
        "main video.",          OFFSET(x_expr), AV_OPT_TYPE_STRING, { .str = "0" }, .flags = FLAGS },
    { "y", "Vertical position of the top edge of the overlaid video on the "
        "main video.",          OFFSET(y_expr), AV_OPT_TYPE_STRING, { .str = "0" }, .flags = FLAGS },
373 374 375 376 377 378
    { "eof_action", "Action to take when encountering EOF from secondary input ",
        OFFSET(eof_action), AV_OPT_TYPE_INT, { .i64 = EOF_ACTION_REPEAT },
        EOF_ACTION_REPEAT, EOF_ACTION_PASS, .flags = FLAGS, "eof_action" },
        { "repeat", "Repeat the previous frame.",   0, AV_OPT_TYPE_CONST, { .i64 = EOF_ACTION_REPEAT }, .flags = FLAGS, "eof_action" },
        { "endall", "End both streams.",            0, AV_OPT_TYPE_CONST, { .i64 = EOF_ACTION_ENDALL }, .flags = FLAGS, "eof_action" },
        { "pass",   "Pass through the main input.", 0, AV_OPT_TYPE_CONST, { .i64 = EOF_ACTION_PASS },   .flags = FLAGS, "eof_action" },
379 380 381 382 383 384 385 386 387 388
    { NULL },
};

static const AVClass overlay_class = {
    .class_name = "overlay",
    .item_name  = av_default_item_name,
    .option     = options,
    .version    = LIBAVUTIL_VERSION_INT,
};

389 390 391 392 393
static const AVFilterPad avfilter_vf_overlay_inputs[] = {
    {
        .name         = "main",
        .type         = AVMEDIA_TYPE_VIDEO,
        .config_props = config_input_main,
394
        .filter_frame = filter_frame_main,
395
        .needs_writable = 1,
396 397 398 399 400 401
        .needs_fifo   = 1,
    },
    {
        .name         = "overlay",
        .type         = AVMEDIA_TYPE_VIDEO,
        .config_props = config_input_overlay,
402
        .filter_frame = filter_frame_overlay,
403 404 405 406 407 408 409 410 411 412 413 414 415 416 417
        .needs_fifo   = 1,
    },
    { NULL }
};

static const AVFilterPad avfilter_vf_overlay_outputs[] = {
    {
        .name          = "default",
        .type          = AVMEDIA_TYPE_VIDEO,
        .config_props  = config_output,
        .request_frame = request_frame,
    },
    { NULL }
};

418
AVFilter ff_vf_overlay = {
419 420 421 422 423 424
    .name      = "overlay",
    .description = NULL_IF_CONFIG_SMALL("Overlay a video source on top of the input."),

    .uninit    = uninit,

    .priv_size = sizeof(OverlayContext),
425
    .priv_class = &overlay_class,
426 427 428

    .query_formats = query_formats,

429 430
    .inputs    = avfilter_vf_overlay_inputs,
    .outputs   = avfilter_vf_overlay_outputs,
431
};