| 1 | /* |
| 2 | * Copyright (c) 2011 Baptiste Coudurier |
| 3 | * Copyright (c) 2011 Stefano Sabatini |
| 4 | * Copyright (c) 2012 Clément Bœsch |
| 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 | /** |
| 24 | * @file |
| 25 | * Libass subtitles burning filter. |
| 26 | * |
| 27 | * @see{http://www.matroska.org/technical/specs/subtitles/ssa.html} |
| 28 | */ |
| 29 | |
| 30 | #include <ass/ass.h> |
| 31 | |
| 32 | #include "config.h" |
| 33 | #if CONFIG_SUBTITLES_FILTER |
| 34 | # include "libavcodec/avcodec.h" |
| 35 | # include "libavformat/avformat.h" |
| 36 | #endif |
| 37 | #include "libavutil/avstring.h" |
| 38 | #include "libavutil/imgutils.h" |
| 39 | #include "libavutil/opt.h" |
| 40 | #include "libavutil/parseutils.h" |
| 41 | #include "drawutils.h" |
| 42 | #include "avfilter.h" |
| 43 | #include "internal.h" |
| 44 | #include "formats.h" |
| 45 | #include "video.h" |
| 46 | |
| 47 | typedef struct { |
| 48 | const AVClass *class; |
| 49 | ASS_Library *library; |
| 50 | ASS_Renderer *renderer; |
| 51 | ASS_Track *track; |
| 52 | char *filename; |
| 53 | char *charenc; |
| 54 | int stream_index; |
| 55 | uint8_t rgba_map[4]; |
| 56 | int pix_step[4]; ///< steps per pixel for each plane of the main output |
| 57 | int original_w, original_h; |
| 58 | int shaping; |
| 59 | FFDrawContext draw; |
| 60 | } AssContext; |
| 61 | |
| 62 | #define OFFSET(x) offsetof(AssContext, x) |
| 63 | #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM |
| 64 | |
| 65 | #define COMMON_OPTIONS \ |
| 66 | {"filename", "set the filename of file to read", OFFSET(filename), AV_OPT_TYPE_STRING, {.str = NULL}, CHAR_MIN, CHAR_MAX, FLAGS }, \ |
| 67 | {"f", "set the filename of file to read", OFFSET(filename), AV_OPT_TYPE_STRING, {.str = NULL}, CHAR_MIN, CHAR_MAX, FLAGS }, \ |
| 68 | {"original_size", "set the size of the original video (used to scale fonts)", OFFSET(original_w), AV_OPT_TYPE_IMAGE_SIZE, {.str = NULL}, CHAR_MIN, CHAR_MAX, FLAGS }, \ |
| 69 | |
| 70 | /* libass supports a log level ranging from 0 to 7 */ |
| 71 | static const int ass_libavfilter_log_level_map[] = { |
| 72 | [0] = AV_LOG_FATAL, /* MSGL_FATAL */ |
| 73 | [1] = AV_LOG_ERROR, /* MSGL_ERR */ |
| 74 | [2] = AV_LOG_WARNING, /* MSGL_WARN */ |
| 75 | [3] = AV_LOG_WARNING, /* <undefined> */ |
| 76 | [4] = AV_LOG_INFO, /* MSGL_INFO */ |
| 77 | [5] = AV_LOG_INFO, /* <undefined> */ |
| 78 | [6] = AV_LOG_VERBOSE, /* MSGL_V */ |
| 79 | [7] = AV_LOG_DEBUG, /* MSGL_DBG2 */ |
| 80 | }; |
| 81 | |
| 82 | static void ass_log(int ass_level, const char *fmt, va_list args, void *ctx) |
| 83 | { |
| 84 | const int ass_level_clip = av_clip(ass_level, 0, |
| 85 | FF_ARRAY_ELEMS(ass_libavfilter_log_level_map) - 1); |
| 86 | const int level = ass_libavfilter_log_level_map[ass_level_clip]; |
| 87 | |
| 88 | av_vlog(ctx, level, fmt, args); |
| 89 | av_log(ctx, level, "\n"); |
| 90 | } |
| 91 | |
| 92 | static av_cold int init(AVFilterContext *ctx) |
| 93 | { |
| 94 | AssContext *ass = ctx->priv; |
| 95 | |
| 96 | if (!ass->filename) { |
| 97 | av_log(ctx, AV_LOG_ERROR, "No filename provided!\n"); |
| 98 | return AVERROR(EINVAL); |
| 99 | } |
| 100 | |
| 101 | ass->library = ass_library_init(); |
| 102 | if (!ass->library) { |
| 103 | av_log(ctx, AV_LOG_ERROR, "Could not initialize libass.\n"); |
| 104 | return AVERROR(EINVAL); |
| 105 | } |
| 106 | ass_set_message_cb(ass->library, ass_log, ctx); |
| 107 | |
| 108 | ass->renderer = ass_renderer_init(ass->library); |
| 109 | if (!ass->renderer) { |
| 110 | av_log(ctx, AV_LOG_ERROR, "Could not initialize libass renderer.\n"); |
| 111 | return AVERROR(EINVAL); |
| 112 | } |
| 113 | |
| 114 | return 0; |
| 115 | } |
| 116 | |
| 117 | static av_cold void uninit(AVFilterContext *ctx) |
| 118 | { |
| 119 | AssContext *ass = ctx->priv; |
| 120 | |
| 121 | if (ass->track) |
| 122 | ass_free_track(ass->track); |
| 123 | if (ass->renderer) |
| 124 | ass_renderer_done(ass->renderer); |
| 125 | if (ass->library) |
| 126 | ass_library_done(ass->library); |
| 127 | } |
| 128 | |
| 129 | static int query_formats(AVFilterContext *ctx) |
| 130 | { |
| 131 | ff_set_common_formats(ctx, ff_draw_supported_pixel_formats(0)); |
| 132 | return 0; |
| 133 | } |
| 134 | |
| 135 | static int config_input(AVFilterLink *inlink) |
| 136 | { |
| 137 | AssContext *ass = inlink->dst->priv; |
| 138 | |
| 139 | ff_draw_init(&ass->draw, inlink->format, 0); |
| 140 | |
| 141 | ass_set_frame_size (ass->renderer, inlink->w, inlink->h); |
| 142 | if (ass->original_w && ass->original_h) |
| 143 | ass_set_aspect_ratio(ass->renderer, (double)inlink->w / inlink->h, |
| 144 | (double)ass->original_w / ass->original_h); |
| 145 | if (ass->shaping != -1) |
| 146 | ass_set_shaper(ass->renderer, ass->shaping); |
| 147 | |
| 148 | return 0; |
| 149 | } |
| 150 | |
| 151 | /* libass stores an RGBA color in the format RRGGBBTT, where TT is the transparency level */ |
| 152 | #define AR(c) ( (c)>>24) |
| 153 | #define AG(c) (((c)>>16)&0xFF) |
| 154 | #define AB(c) (((c)>>8) &0xFF) |
| 155 | #define AA(c) ((0xFF-c) &0xFF) |
| 156 | |
| 157 | static void overlay_ass_image(AssContext *ass, AVFrame *picref, |
| 158 | const ASS_Image *image) |
| 159 | { |
| 160 | for (; image; image = image->next) { |
| 161 | uint8_t rgba_color[] = {AR(image->color), AG(image->color), AB(image->color), AA(image->color)}; |
| 162 | FFDrawColor color; |
| 163 | ff_draw_color(&ass->draw, &color, rgba_color); |
| 164 | ff_blend_mask(&ass->draw, &color, |
| 165 | picref->data, picref->linesize, |
| 166 | picref->width, picref->height, |
| 167 | image->bitmap, image->stride, image->w, image->h, |
| 168 | 3, 0, image->dst_x, image->dst_y); |
| 169 | } |
| 170 | } |
| 171 | |
| 172 | static int filter_frame(AVFilterLink *inlink, AVFrame *picref) |
| 173 | { |
| 174 | AVFilterContext *ctx = inlink->dst; |
| 175 | AVFilterLink *outlink = ctx->outputs[0]; |
| 176 | AssContext *ass = ctx->priv; |
| 177 | int detect_change = 0; |
| 178 | double time_ms = picref->pts * av_q2d(inlink->time_base) * 1000; |
| 179 | ASS_Image *image = ass_render_frame(ass->renderer, ass->track, |
| 180 | time_ms, &detect_change); |
| 181 | |
| 182 | if (detect_change) |
| 183 | av_log(ctx, AV_LOG_DEBUG, "Change happened at time ms:%f\n", time_ms); |
| 184 | |
| 185 | overlay_ass_image(ass, picref, image); |
| 186 | |
| 187 | return ff_filter_frame(outlink, picref); |
| 188 | } |
| 189 | |
| 190 | static const AVFilterPad ass_inputs[] = { |
| 191 | { |
| 192 | .name = "default", |
| 193 | .type = AVMEDIA_TYPE_VIDEO, |
| 194 | .filter_frame = filter_frame, |
| 195 | .config_props = config_input, |
| 196 | .needs_writable = 1, |
| 197 | }, |
| 198 | { NULL } |
| 199 | }; |
| 200 | |
| 201 | static const AVFilterPad ass_outputs[] = { |
| 202 | { |
| 203 | .name = "default", |
| 204 | .type = AVMEDIA_TYPE_VIDEO, |
| 205 | }, |
| 206 | { NULL } |
| 207 | }; |
| 208 | |
| 209 | #if CONFIG_ASS_FILTER |
| 210 | |
| 211 | static const AVOption ass_options[] = { |
| 212 | COMMON_OPTIONS |
| 213 | {"shaping", "set shaping engine", OFFSET(shaping), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 1, FLAGS, "shaping_mode"}, |
| 214 | {"auto", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = -1}, INT_MIN, INT_MAX, FLAGS, "shaping_mode"}, |
| 215 | {"simple", "simple shaping", 0, AV_OPT_TYPE_CONST, {.i64 = ASS_SHAPING_SIMPLE}, INT_MIN, INT_MAX, FLAGS, "shaping_mode"}, |
| 216 | {"complex", "complex shaping", 0, AV_OPT_TYPE_CONST, {.i64 = ASS_SHAPING_COMPLEX}, INT_MIN, INT_MAX, FLAGS, "shaping_mode"}, |
| 217 | {NULL}, |
| 218 | }; |
| 219 | |
| 220 | AVFILTER_DEFINE_CLASS(ass); |
| 221 | |
| 222 | static av_cold int init_ass(AVFilterContext *ctx) |
| 223 | { |
| 224 | AssContext *ass = ctx->priv; |
| 225 | int ret = init(ctx); |
| 226 | |
| 227 | if (ret < 0) |
| 228 | return ret; |
| 229 | |
| 230 | /* Initialize fonts */ |
| 231 | ass_set_fonts(ass->renderer, NULL, NULL, 1, NULL, 1); |
| 232 | |
| 233 | ass->track = ass_read_file(ass->library, ass->filename, NULL); |
| 234 | if (!ass->track) { |
| 235 | av_log(ctx, AV_LOG_ERROR, |
| 236 | "Could not create a libass track when reading file '%s'\n", |
| 237 | ass->filename); |
| 238 | return AVERROR(EINVAL); |
| 239 | } |
| 240 | return 0; |
| 241 | } |
| 242 | |
| 243 | AVFilter ff_vf_ass = { |
| 244 | .name = "ass", |
| 245 | .description = NULL_IF_CONFIG_SMALL("Render ASS subtitles onto input video using the libass library."), |
| 246 | .priv_size = sizeof(AssContext), |
| 247 | .init = init_ass, |
| 248 | .uninit = uninit, |
| 249 | .query_formats = query_formats, |
| 250 | .inputs = ass_inputs, |
| 251 | .outputs = ass_outputs, |
| 252 | .priv_class = &ass_class, |
| 253 | }; |
| 254 | #endif |
| 255 | |
| 256 | #if CONFIG_SUBTITLES_FILTER |
| 257 | |
| 258 | static const AVOption subtitles_options[] = { |
| 259 | COMMON_OPTIONS |
| 260 | {"charenc", "set input character encoding", OFFSET(charenc), AV_OPT_TYPE_STRING, {.str = NULL}, CHAR_MIN, CHAR_MAX, FLAGS}, |
| 261 | {"stream_index", "set stream index", OFFSET(stream_index), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS}, |
| 262 | {"si", "set stream index", OFFSET(stream_index), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, FLAGS}, |
| 263 | {NULL}, |
| 264 | }; |
| 265 | |
| 266 | static const char * const font_mimetypes[] = { |
| 267 | "application/x-truetype-font", |
| 268 | "application/vnd.ms-opentype", |
| 269 | "application/x-font-ttf", |
| 270 | NULL |
| 271 | }; |
| 272 | |
| 273 | static int attachment_is_font(AVStream * st) |
| 274 | { |
| 275 | const AVDictionaryEntry *tag = NULL; |
| 276 | int n; |
| 277 | |
| 278 | tag = av_dict_get(st->metadata, "mimetype", NULL, AV_DICT_MATCH_CASE); |
| 279 | |
| 280 | if (tag) { |
| 281 | for (n = 0; font_mimetypes[n]; n++) { |
| 282 | if (av_strcasecmp(font_mimetypes[n], tag->value) == 0) |
| 283 | return 1; |
| 284 | } |
| 285 | } |
| 286 | return 0; |
| 287 | } |
| 288 | |
| 289 | AVFILTER_DEFINE_CLASS(subtitles); |
| 290 | |
| 291 | static av_cold int init_subtitles(AVFilterContext *ctx) |
| 292 | { |
| 293 | int j, ret, sid; |
| 294 | int k = 0; |
| 295 | AVDictionary *codec_opts = NULL; |
| 296 | AVFormatContext *fmt = NULL; |
| 297 | AVCodecContext *dec_ctx = NULL; |
| 298 | AVCodec *dec = NULL; |
| 299 | const AVCodecDescriptor *dec_desc; |
| 300 | AVStream *st; |
| 301 | AVPacket pkt; |
| 302 | AssContext *ass = ctx->priv; |
| 303 | |
| 304 | /* Init libass */ |
| 305 | ret = init(ctx); |
| 306 | if (ret < 0) |
| 307 | return ret; |
| 308 | ass->track = ass_new_track(ass->library); |
| 309 | if (!ass->track) { |
| 310 | av_log(ctx, AV_LOG_ERROR, "Could not create a libass track\n"); |
| 311 | return AVERROR(EINVAL); |
| 312 | } |
| 313 | |
| 314 | /* Open subtitles file */ |
| 315 | ret = avformat_open_input(&fmt, ass->filename, NULL, NULL); |
| 316 | if (ret < 0) { |
| 317 | av_log(ctx, AV_LOG_ERROR, "Unable to open %s\n", ass->filename); |
| 318 | goto end; |
| 319 | } |
| 320 | ret = avformat_find_stream_info(fmt, NULL); |
| 321 | if (ret < 0) |
| 322 | goto end; |
| 323 | |
| 324 | /* Locate subtitles stream */ |
| 325 | if (ass->stream_index < 0) |
| 326 | ret = av_find_best_stream(fmt, AVMEDIA_TYPE_SUBTITLE, -1, -1, NULL, 0); |
| 327 | else { |
| 328 | ret = -1; |
| 329 | if (ass->stream_index < fmt->nb_streams) { |
| 330 | for (j = 0; j < fmt->nb_streams; j++) { |
| 331 | if (fmt->streams[j]->codec->codec_type == AVMEDIA_TYPE_SUBTITLE) { |
| 332 | if (ass->stream_index == k) { |
| 333 | ret = j; |
| 334 | break; |
| 335 | } |
| 336 | k++; |
| 337 | } |
| 338 | } |
| 339 | } |
| 340 | } |
| 341 | |
| 342 | if (ret < 0) { |
| 343 | av_log(ctx, AV_LOG_ERROR, "Unable to locate subtitle stream in %s\n", |
| 344 | ass->filename); |
| 345 | goto end; |
| 346 | } |
| 347 | sid = ret; |
| 348 | st = fmt->streams[sid]; |
| 349 | |
| 350 | /* Load attached fonts */ |
| 351 | for (j = 0; j < fmt->nb_streams; j++) { |
| 352 | AVStream *st = fmt->streams[j]; |
| 353 | if (st->codec->codec_type == AVMEDIA_TYPE_ATTACHMENT && |
| 354 | attachment_is_font(st)) { |
| 355 | const AVDictionaryEntry *tag = NULL; |
| 356 | tag = av_dict_get(st->metadata, "filename", NULL, |
| 357 | AV_DICT_MATCH_CASE); |
| 358 | |
| 359 | if (tag) { |
| 360 | av_log(ctx, AV_LOG_DEBUG, "Loading attached font: %s\n", |
| 361 | tag->value); |
| 362 | ass_add_font(ass->library, tag->value, |
| 363 | st->codec->extradata, |
| 364 | st->codec->extradata_size); |
| 365 | } else { |
| 366 | av_log(ctx, AV_LOG_WARNING, |
| 367 | "Font attachment has no filename, ignored.\n"); |
| 368 | } |
| 369 | } |
| 370 | } |
| 371 | |
| 372 | /* Initialize fonts */ |
| 373 | ass_set_fonts(ass->renderer, NULL, NULL, 1, NULL, 1); |
| 374 | |
| 375 | /* Open decoder */ |
| 376 | dec_ctx = st->codec; |
| 377 | dec = avcodec_find_decoder(dec_ctx->codec_id); |
| 378 | if (!dec) { |
| 379 | av_log(ctx, AV_LOG_ERROR, "Failed to find subtitle codec %s\n", |
| 380 | avcodec_get_name(dec_ctx->codec_id)); |
| 381 | return AVERROR(EINVAL); |
| 382 | } |
| 383 | dec_desc = avcodec_descriptor_get(dec_ctx->codec_id); |
| 384 | if (dec_desc && !(dec_desc->props & AV_CODEC_PROP_TEXT_SUB)) { |
| 385 | av_log(ctx, AV_LOG_ERROR, |
| 386 | "Only text based subtitles are currently supported\n"); |
| 387 | return AVERROR_PATCHWELCOME; |
| 388 | } |
| 389 | if (ass->charenc) |
| 390 | av_dict_set(&codec_opts, "sub_charenc", ass->charenc, 0); |
| 391 | ret = avcodec_open2(dec_ctx, dec, &codec_opts); |
| 392 | if (ret < 0) |
| 393 | goto end; |
| 394 | |
| 395 | /* Decode subtitles and push them into the renderer (libass) */ |
| 396 | if (dec_ctx->subtitle_header) |
| 397 | ass_process_codec_private(ass->track, |
| 398 | dec_ctx->subtitle_header, |
| 399 | dec_ctx->subtitle_header_size); |
| 400 | av_init_packet(&pkt); |
| 401 | pkt.data = NULL; |
| 402 | pkt.size = 0; |
| 403 | while (av_read_frame(fmt, &pkt) >= 0) { |
| 404 | int i, got_subtitle; |
| 405 | AVSubtitle sub = {0}; |
| 406 | |
| 407 | if (pkt.stream_index == sid) { |
| 408 | ret = avcodec_decode_subtitle2(dec_ctx, &sub, &got_subtitle, &pkt); |
| 409 | if (ret < 0) { |
| 410 | av_log(ctx, AV_LOG_WARNING, "Error decoding: %s (ignored)\n", |
| 411 | av_err2str(ret)); |
| 412 | } else if (got_subtitle) { |
| 413 | for (i = 0; i < sub.num_rects; i++) { |
| 414 | char *ass_line = sub.rects[i]->ass; |
| 415 | if (!ass_line) |
| 416 | break; |
| 417 | ass_process_data(ass->track, ass_line, strlen(ass_line)); |
| 418 | } |
| 419 | } |
| 420 | } |
| 421 | av_free_packet(&pkt); |
| 422 | avsubtitle_free(&sub); |
| 423 | } |
| 424 | |
| 425 | end: |
| 426 | av_dict_free(&codec_opts); |
| 427 | if (dec_ctx) |
| 428 | avcodec_close(dec_ctx); |
| 429 | if (fmt) |
| 430 | avformat_close_input(&fmt); |
| 431 | return ret; |
| 432 | } |
| 433 | |
| 434 | AVFilter ff_vf_subtitles = { |
| 435 | .name = "subtitles", |
| 436 | .description = NULL_IF_CONFIG_SMALL("Render text subtitles onto input video using the libass library."), |
| 437 | .priv_size = sizeof(AssContext), |
| 438 | .init = init_subtitles, |
| 439 | .uninit = uninit, |
| 440 | .query_formats = query_formats, |
| 441 | .inputs = ass_inputs, |
| 442 | .outputs = ass_outputs, |
| 443 | .priv_class = &subtitles_class, |
| 444 | }; |
| 445 | #endif |