vf_overlay.c 13.3 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 "internal.h"
38
#include "video.h"
39

40
static const char *const var_names[] = {
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
    "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
};

#define MAIN    0
#define OVERLAY 1

typedef struct {
    int x, y;                   ///< position of overlayed picture

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

    char x_expr[256], y_expr[256];
72 73 74

    AVFilterBufferRef *main;
    AVFilterBufferRef *over_prev, *over_next;
75 76
} OverlayContext;

77
static av_cold int init(AVFilterContext *ctx, const char *args)
78 79 80 81 82 83 84 85 86 87 88 89 90 91
{
    OverlayContext *over = ctx->priv;

    av_strlcpy(over->x_expr, "0", sizeof(over->x_expr));
    av_strlcpy(over->y_expr, "0", sizeof(over->y_expr));

    if (args)
        sscanf(args, "%255[^:]:%255[^:]", over->x_expr, over->y_expr);

    return 0;
}

static av_cold void uninit(AVFilterContext *ctx)
{
92
    OverlayContext *s = ctx->priv;
93

94 95 96
    avfilter_unref_bufferp(&s->main);
    avfilter_unref_bufferp(&s->over_prev);
    avfilter_unref_bufferp(&s->over_next);
97 98 99 100
}

static int query_formats(AVFilterContext *ctx)
{
101 102
    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 };
103 104
    AVFilterFormats *inout_formats = ff_make_format_list(inout_pix_fmts);
    AVFilterFormats *blend_formats = ff_make_format_list(blend_pix_fmts);
105

106 107 108
    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 );
109 110 111 112 113 114 115

    return 0;
}

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

    av_image_fill_max_pixsteps(over->max_plane_step, NULL, pix_desc);
    over->hsub = pix_desc->log2_chroma_w;
    over->vsub = pix_desc->log2_chroma_h;

    return 0;
}

static int config_input_overlay(AVFilterLink *inlink)
{
    AVFilterContext *ctx  = inlink->dst;
    OverlayContext  *over = inlink->dst->priv;
    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;

    if ((ret = av_expr_parse_and_eval(&res, (expr = over->x_expr), var_names, var_values,
                                      NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
        goto fail;
    over->x = res;
    if ((ret = av_expr_parse_and_eval(&res, (expr = over->y_expr), var_names, var_values,
                                      NULL, NULL, NULL, NULL, NULL, 0, ctx)))
        goto fail;
    over->y = res;
    /* x may depend on y */
    if ((ret = av_expr_parse_and_eval(&res, (expr = over->x_expr), var_names, var_values,
                                      NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
        goto fail;
    over->x = res;

158
    av_log(ctx, AV_LOG_VERBOSE,
159 160
           "main w:%d h:%d fmt:%s overlay x:%d y:%d w:%d h:%d fmt:%s\n",
           ctx->inputs[MAIN]->w, ctx->inputs[MAIN]->h,
161
           av_get_pix_fmt_name(ctx->inputs[MAIN]->format),
162 163
           over->x, over->y,
           ctx->inputs[OVERLAY]->w, ctx->inputs[OVERLAY]->h,
164
           av_get_pix_fmt_name(ctx->inputs[OVERLAY]->format));
165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190

    if (over->x < 0 || over->y < 0 ||
        over->x + var_values[VAR_OVERLAY_W] > var_values[VAR_MAIN_W] ||
        over->y + var_values[VAR_OVERLAY_H] > var_values[VAR_MAIN_H]) {
        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",
               over->x, over->y,
               (int)(over->x + var_values[VAR_OVERLAY_W]),
               (int)(over->y + var_values[VAR_OVERLAY_H]),
               (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;
191
    outlink->time_base = ctx->inputs[MAIN]->time_base;
192 193 194 195

    return 0;
}

196
static void blend_frame(AVFilterContext *ctx,
197
                        AVFilterBufferRef *dst, AVFilterBufferRef *src,
198
                        int x, int y)
199 200 201 202
{
    OverlayContext *over = ctx->priv;
    int i, j, k;
    int width, height;
203
    int overlay_end_y = y + src->video->h;
204 205
    int end_y, start_y;

206 207 208
    width = FFMIN(dst->video->w - x, src->video->w);
    end_y = FFMIN(dst->video->h, overlay_end_y);
    start_y = FFMAX(y, 0);
209 210
    height = end_y - start_y;

211
    if (dst->format == AV_PIX_FMT_BGR24 || dst->format == AV_PIX_FMT_RGB24) {
212 213
        uint8_t *dp = dst->data[0] + x * 3 + start_y * dst->linesize[0];
        uint8_t *sp = src->data[0];
214 215
        int b = dst->format == AV_PIX_FMT_BGR24 ? 2 : 0;
        int r = dst->format == AV_PIX_FMT_BGR24 ? 0 : 2;
216 217
        if (y < 0)
            sp += -y * src->linesize[0];
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
        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++) {
            int hsub = i ? over->hsub : 0;
            int vsub = i ? over->vsub : 0;
            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;
240 241 242
            if (y < 0) {
                sp += ((-y) >> vsub) * src->linesize[i];
                ap += -y * src->linesize[3];
243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271
            }
            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];
            }
        }
    }
}

272
static int null_start_frame(AVFilterLink *inlink, AVFilterBufferRef *buf)
273
{
274
    return 0;
275 276
}

277
static int null_draw_slice(AVFilterLink *inlink, int y, int h, int slice_dir)
278
{
279
    return 0;
280 281
}

282
static int end_frame_main(AVFilterLink *inlink)
283
{
284 285 286 287 288 289
    OverlayContext *s = inlink->dst->priv;

    av_assert0(!s->main);
    s->main         = inlink->cur_buf;
    inlink->cur_buf = NULL;

290 291
    return 0;
}
292

293
static int end_frame_overlay(AVFilterLink *inlink)
294
{
295 296 297 298 299 300
    OverlayContext *s = inlink->dst->priv;

    av_assert0(!s->over_next);
    s->over_next    = inlink->cur_buf;
    inlink->cur_buf = NULL;

301 302
    return 0;
}
303

304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326
static int output_frame(AVFilterContext *ctx)
{
    OverlayContext *s = ctx->priv;
    AVFilterLink *outlink = ctx->outputs[0];
    int ret = ff_start_frame(outlink, s->main);
    if (ret >= 0)
        ret = ff_draw_slice(outlink, 0, outlink->h, 1);
    if (ret >= 0)
        ret = ff_end_frame(outlink);
    s->main = NULL;

    return ret;
}

static int handle_overlay_eof(AVFilterContext *ctx)
{
    OverlayContext *s = ctx->priv;
    if (s->over_prev)
        blend_frame(ctx, s->main, s->over_prev, s->x, s->y);
    return output_frame(ctx);
}

static int request_frame(AVFilterLink *outlink)
327
{
328 329 330 331 332 333 334 335 336 337 338 339
    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;
    }
340

341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
    /* get a new frame on the overlay input, on EOF
     * reuse previous */
    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) {
        avfilter_unref_bufferp(&s->over_prev);
        FFSWAP(AVFilterBufferRef*, s->over_prev, 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;
    }
363

364 365 366 367 368 369 370 371 372 373 374
    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);
        avfilter_unref_bufferp(&s->over_prev);
        FFSWAP(AVFilterBufferRef*, s->over_prev, s->over_next);
    } else if (s->over_prev) {
        blend_frame(ctx, s->main, s->over_prev, s->x, s->y);
    }

    return output_frame(ctx);
375 376
}

377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
static const AVFilterPad avfilter_vf_overlay_inputs[] = {
    {
        .name         = "main",
        .type         = AVMEDIA_TYPE_VIDEO,
        .start_frame  = null_start_frame,
        .config_props = config_input_main,
        .draw_slice   = null_draw_slice,
        .end_frame    = end_frame_main,
        .min_perms    = AV_PERM_READ,
        .rej_perms    = AV_PERM_REUSE2 | AV_PERM_PRESERVE,
        .needs_fifo   = 1,
    },
    {
        .name         = "overlay",
        .type         = AVMEDIA_TYPE_VIDEO,
        .start_frame  = null_start_frame,
        .config_props = config_input_overlay,
        .draw_slice   = null_draw_slice,
        .end_frame    = end_frame_overlay,
        .min_perms    = AV_PERM_READ,
        .rej_perms    = AV_PERM_REUSE2,
        .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 }
};

413 414 415 416 417 418 419 420 421 422 423
AVFilter avfilter_vf_overlay = {
    .name      = "overlay",
    .description = NULL_IF_CONFIG_SMALL("Overlay a video source on top of the input."),

    .init      = init,
    .uninit    = uninit,

    .priv_size = sizeof(OverlayContext),

    .query_formats = query_formats,

424 425
    .inputs    = avfilter_vf_overlay_inputs,
    .outputs   = avfilter_vf_overlay_outputs,
426
};