You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

902 lines
35KB

  1. /*
  2. * Copyright (c) 2011 Stefano Sabatini
  3. * Copyright (c) 2010 S.N. Hemanth Meenakshisundaram
  4. * Copyright (c) 2003 Gustavo Sverzut Barbieri <gsbarbieri@yahoo.com.br>
  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. * @file
  24. * drawtext filter, based on the original vhook/drawtext.c
  25. * filter by Gustavo Sverzut Barbieri
  26. */
  27. #include <sys/time.h>
  28. #include <time.h>
  29. #include "config.h"
  30. #include "libavutil/avstring.h"
  31. #include "libavutil/colorspace.h"
  32. #include "libavutil/file.h"
  33. #include "libavutil/eval.h"
  34. #include "libavutil/opt.h"
  35. #include "libavutil/random_seed.h"
  36. #include "libavutil/parseutils.h"
  37. #include "libavutil/pixdesc.h"
  38. #include "libavutil/timecode.h"
  39. #include "libavutil/tree.h"
  40. #include "libavutil/lfg.h"
  41. #include "avfilter.h"
  42. #include "drawutils.h"
  43. #undef time
  44. #include <ft2build.h>
  45. #include <freetype/config/ftheader.h>
  46. #include FT_FREETYPE_H
  47. #include FT_GLYPH_H
  48. static const char *const var_names[] = {
  49. "main_w", "w", "W", ///< width of the input video
  50. "main_h", "h", "H", ///< height of the input video
  51. "tw", "text_w", ///< width of the rendered text
  52. "th", "text_h", ///< height of the rendered text
  53. "max_glyph_w", ///< max glyph width
  54. "max_glyph_h", ///< max glyph height
  55. "max_glyph_a", "ascent", ///< max glyph ascent
  56. "max_glyph_d", "descent", ///< min glyph descent
  57. "line_h", "lh", ///< line height, same as max_glyph_h
  58. "sar",
  59. "dar",
  60. "hsub",
  61. "vsub",
  62. "x",
  63. "y",
  64. "n", ///< number of frame
  65. "t", ///< timestamp expressed in seconds
  66. NULL
  67. };
  68. static const char *const fun2_names[] = {
  69. "rand"
  70. };
  71. static double drand(void *opaque, double min, double max)
  72. {
  73. return min + (max-min) / UINT_MAX * av_lfg_get(opaque);
  74. }
  75. typedef double (*eval_func2)(void *, double a, double b);
  76. static const eval_func2 fun2[] = {
  77. drand,
  78. NULL
  79. };
  80. enum var_name {
  81. VAR_MAIN_W, VAR_w, VAR_W,
  82. VAR_MAIN_H, VAR_h, VAR_H,
  83. VAR_TW, VAR_TEXT_W,
  84. VAR_TH, VAR_TEXT_H,
  85. VAR_MAX_GLYPH_W,
  86. VAR_MAX_GLYPH_H,
  87. VAR_MAX_GLYPH_A, VAR_ASCENT,
  88. VAR_MAX_GLYPH_D, VAR_DESCENT,
  89. VAR_LINE_H, VAR_LH,
  90. VAR_SAR,
  91. VAR_DAR,
  92. VAR_HSUB,
  93. VAR_VSUB,
  94. VAR_X,
  95. VAR_Y,
  96. VAR_N,
  97. VAR_T,
  98. VAR_VARS_NB
  99. };
  100. typedef struct {
  101. const AVClass *class;
  102. int reinit; ///< tells if the filter is being reinited
  103. uint8_t *fontfile; ///< font to be used
  104. uint8_t *text; ///< text to be drawn
  105. uint8_t *expanded_text; ///< used to contain the strftime()-expanded text
  106. size_t expanded_text_size; ///< size in bytes of the expanded_text buffer
  107. int ft_load_flags; ///< flags used for loading fonts, see FT_LOAD_*
  108. FT_Vector *positions; ///< positions for each element in the text
  109. size_t nb_positions; ///< number of elements of positions array
  110. char *textfile; ///< file with text to be drawn
  111. int x; ///< x position to start drawing text
  112. int y; ///< y position to start drawing text
  113. int max_glyph_w; ///< max glyph width
  114. int max_glyph_h; ///< max glyph heigth
  115. int shadowx, shadowy;
  116. unsigned int fontsize; ///< font size to use
  117. char *fontcolor_string; ///< font color as string
  118. char *boxcolor_string; ///< box color as string
  119. char *shadowcolor_string; ///< shadow color as string
  120. uint8_t fontcolor[4]; ///< foreground color
  121. uint8_t boxcolor[4]; ///< background color
  122. uint8_t shadowcolor[4]; ///< shadow color
  123. uint8_t fontcolor_rgba[4]; ///< foreground color in RGBA
  124. uint8_t boxcolor_rgba[4]; ///< background color in RGBA
  125. uint8_t shadowcolor_rgba[4]; ///< shadow color in RGBA
  126. short int draw_box; ///< draw box around text - true or false
  127. int use_kerning; ///< font kerning is used - true/false
  128. int tabsize; ///< tab size
  129. int fix_bounds; ///< do we let it go out of frame bounds - t/f
  130. FT_Library library; ///< freetype font library handle
  131. FT_Face face; ///< freetype font face handle
  132. struct AVTreeNode *glyphs; ///< rendered glyphs, stored using the UTF-32 char code
  133. int hsub, vsub; ///< chroma subsampling values
  134. int is_packed_rgb;
  135. int pixel_step[4]; ///< distance in bytes between the component of each pixel
  136. uint8_t rgba_map[4]; ///< map RGBA offsets to the positions in the packed RGBA format
  137. uint8_t *box_line[4]; ///< line used for filling the box background
  138. char *x_expr; ///< expression for x position
  139. char *y_expr; ///< expression for y position
  140. AVExpr *x_pexpr, *y_pexpr; ///< parsed expressions for x and y
  141. int64_t basetime; ///< base pts time in the real world for display
  142. double var_values[VAR_VARS_NB];
  143. char *d_expr;
  144. AVExpr *d_pexpr;
  145. int draw; ///< set to zero to prevent drawing
  146. AVLFG prng; ///< random
  147. char *tc_opt_string; ///< specified timecode option string
  148. AVRational tc_rate; ///< frame rate for timecode
  149. AVTimecode tc; ///< timecode context
  150. int frame_id;
  151. } DrawTextContext;
  152. #define OFFSET(x) offsetof(DrawTextContext, x)
  153. static const AVOption drawtext_options[]= {
  154. {"fontfile", "set font file", OFFSET(fontfile), AV_OPT_TYPE_STRING, {.str=NULL}, CHAR_MIN, CHAR_MAX },
  155. {"text", "set text", OFFSET(text), AV_OPT_TYPE_STRING, {.str=NULL}, CHAR_MIN, CHAR_MAX },
  156. {"textfile", "set text file", OFFSET(textfile), AV_OPT_TYPE_STRING, {.str=NULL}, CHAR_MIN, CHAR_MAX },
  157. {"fontcolor", "set foreground color", OFFSET(fontcolor_string), AV_OPT_TYPE_STRING, {.str="black"}, CHAR_MIN, CHAR_MAX },
  158. {"boxcolor", "set box color", OFFSET(boxcolor_string), AV_OPT_TYPE_STRING, {.str="white"}, CHAR_MIN, CHAR_MAX },
  159. {"shadowcolor", "set shadow color", OFFSET(shadowcolor_string), AV_OPT_TYPE_STRING, {.str="black"}, CHAR_MIN, CHAR_MAX },
  160. {"box", "set box", OFFSET(draw_box), AV_OPT_TYPE_INT, {.dbl=0}, 0, 1 },
  161. {"fontsize", "set font size", OFFSET(fontsize), AV_OPT_TYPE_INT, {.dbl=16}, 1, INT_MAX },
  162. {"x", "set x expression", OFFSET(x_expr), AV_OPT_TYPE_STRING, {.str="0"}, CHAR_MIN, CHAR_MAX },
  163. {"y", "set y expression", OFFSET(y_expr), AV_OPT_TYPE_STRING, {.str="0"}, CHAR_MIN, CHAR_MAX },
  164. {"shadowx", "set x", OFFSET(shadowx), AV_OPT_TYPE_INT, {.dbl=0}, INT_MIN, INT_MAX },
  165. {"shadowy", "set y", OFFSET(shadowy), AV_OPT_TYPE_INT, {.dbl=0}, INT_MIN, INT_MAX },
  166. {"tabsize", "set tab size", OFFSET(tabsize), AV_OPT_TYPE_INT, {.dbl=4}, 0, INT_MAX },
  167. {"basetime", "set base time", OFFSET(basetime), AV_OPT_TYPE_INT64, {.dbl=AV_NOPTS_VALUE}, INT64_MIN, INT64_MAX },
  168. {"draw", "if false do not draw", OFFSET(d_expr), AV_OPT_TYPE_STRING, {.str="1"}, CHAR_MIN, CHAR_MAX },
  169. {"timecode", "set initial timecode", OFFSET(tc_opt_string), AV_OPT_TYPE_STRING, {.str=NULL}, CHAR_MIN, CHAR_MAX },
  170. {"r", "set rate (timecode only)", OFFSET(tc_rate), AV_OPT_TYPE_RATIONAL, {.dbl=0}, 0, INT_MAX },
  171. {"rate", "set rate (timecode only)", OFFSET(tc_rate), AV_OPT_TYPE_RATIONAL, {.dbl=0}, 0, INT_MAX },
  172. {"fix_bounds", "if true, check and fix text coords to avoid clipping",
  173. OFFSET(fix_bounds), AV_OPT_TYPE_INT, {.dbl=1}, 0, 1 },
  174. /* FT_LOAD_* flags */
  175. {"ft_load_flags", "set font loading flags for libfreetype", OFFSET(ft_load_flags), AV_OPT_TYPE_FLAGS, {.dbl=FT_LOAD_DEFAULT|FT_LOAD_RENDER}, 0, INT_MAX, 0, "ft_load_flags" },
  176. {"default", "set default", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_DEFAULT}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  177. {"no_scale", "set no_scale", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_NO_SCALE}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  178. {"no_hinting", "set no_hinting", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_NO_HINTING}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  179. {"render", "set render", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_RENDER}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  180. {"no_bitmap", "set no_bitmap", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_NO_BITMAP}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  181. {"vertical_layout", "set vertical_layout", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_VERTICAL_LAYOUT}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  182. {"force_autohint", "set force_autohint", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_FORCE_AUTOHINT}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  183. {"crop_bitmap", "set crop_bitmap", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_CROP_BITMAP}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  184. {"pedantic", "set pedantic", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_PEDANTIC}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  185. {"ignore_global_advance_width", "set ignore_global_advance_width", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_IGNORE_GLOBAL_ADVANCE_WIDTH}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  186. {"no_recurse", "set no_recurse", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_NO_RECURSE}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  187. {"ignore_transform", "set ignore_transform", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_IGNORE_TRANSFORM}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  188. {"monochrome", "set monochrome", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_MONOCHROME}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  189. {"linear_design", "set linear_design", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_LINEAR_DESIGN}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  190. {"no_autohint", "set no_autohint", 0, AV_OPT_TYPE_CONST, {.dbl=FT_LOAD_NO_AUTOHINT}, INT_MIN, INT_MAX, 0, "ft_load_flags" },
  191. {NULL},
  192. };
  193. static const char *drawtext_get_name(void *ctx)
  194. {
  195. return "drawtext";
  196. }
  197. static const AVClass drawtext_class = {
  198. "DrawTextContext",
  199. drawtext_get_name,
  200. drawtext_options
  201. };
  202. #undef __FTERRORS_H__
  203. #define FT_ERROR_START_LIST {
  204. #define FT_ERRORDEF(e, v, s) { (e), (s) },
  205. #define FT_ERROR_END_LIST { 0, NULL } };
  206. struct ft_error
  207. {
  208. int err;
  209. const char *err_msg;
  210. } static ft_errors[] =
  211. #include FT_ERRORS_H
  212. #define FT_ERRMSG(e) ft_errors[e].err_msg
  213. typedef struct {
  214. FT_Glyph *glyph;
  215. uint32_t code;
  216. FT_Bitmap bitmap; ///< array holding bitmaps of font
  217. FT_BBox bbox;
  218. int advance;
  219. int bitmap_left;
  220. int bitmap_top;
  221. } Glyph;
  222. static int glyph_cmp(void *key, const void *b)
  223. {
  224. const Glyph *a = key, *bb = b;
  225. int64_t diff = (int64_t)a->code - (int64_t)bb->code;
  226. return diff > 0 ? 1 : diff < 0 ? -1 : 0;
  227. }
  228. /**
  229. * Load glyphs corresponding to the UTF-32 codepoint code.
  230. */
  231. static int load_glyph(AVFilterContext *ctx, Glyph **glyph_ptr, uint32_t code)
  232. {
  233. DrawTextContext *dtext = ctx->priv;
  234. Glyph *glyph;
  235. struct AVTreeNode *node = NULL;
  236. int ret;
  237. /* load glyph into dtext->face->glyph */
  238. if (FT_Load_Char(dtext->face, code, dtext->ft_load_flags))
  239. return AVERROR(EINVAL);
  240. /* save glyph */
  241. if (!(glyph = av_mallocz(sizeof(*glyph))) ||
  242. !(glyph->glyph = av_mallocz(sizeof(*glyph->glyph)))) {
  243. ret = AVERROR(ENOMEM);
  244. goto error;
  245. }
  246. glyph->code = code;
  247. if (FT_Get_Glyph(dtext->face->glyph, glyph->glyph)) {
  248. ret = AVERROR(EINVAL);
  249. goto error;
  250. }
  251. glyph->bitmap = dtext->face->glyph->bitmap;
  252. glyph->bitmap_left = dtext->face->glyph->bitmap_left;
  253. glyph->bitmap_top = dtext->face->glyph->bitmap_top;
  254. glyph->advance = dtext->face->glyph->advance.x >> 6;
  255. /* measure text height to calculate text_height (or the maximum text height) */
  256. FT_Glyph_Get_CBox(*glyph->glyph, ft_glyph_bbox_pixels, &glyph->bbox);
  257. /* cache the newly created glyph */
  258. if (!(node = av_mallocz(av_tree_node_size))) {
  259. ret = AVERROR(ENOMEM);
  260. goto error;
  261. }
  262. av_tree_insert(&dtext->glyphs, glyph, glyph_cmp, &node);
  263. if (glyph_ptr)
  264. *glyph_ptr = glyph;
  265. return 0;
  266. error:
  267. if (glyph)
  268. av_freep(&glyph->glyph);
  269. av_freep(&glyph);
  270. av_freep(&node);
  271. return ret;
  272. }
  273. static av_cold int init(AVFilterContext *ctx, const char *args, void *opaque)
  274. {
  275. int err;
  276. DrawTextContext *dtext = ctx->priv;
  277. Glyph *glyph;
  278. dtext->class = &drawtext_class;
  279. av_opt_set_defaults(dtext);
  280. if ((err = (av_set_options_string(dtext, args, "=", ":"))) < 0) {
  281. av_log(ctx, AV_LOG_ERROR, "Error parsing options string: '%s'\n", args);
  282. return err;
  283. }
  284. if (!dtext->fontfile) {
  285. av_log(ctx, AV_LOG_ERROR, "No font filename provided\n");
  286. return AVERROR(EINVAL);
  287. }
  288. if (dtext->textfile) {
  289. uint8_t *textbuf;
  290. size_t textbuf_size;
  291. if (dtext->text) {
  292. av_log(ctx, AV_LOG_ERROR,
  293. "Both text and text file provided. Please provide only one\n");
  294. return AVERROR(EINVAL);
  295. }
  296. if ((err = av_file_map(dtext->textfile, &textbuf, &textbuf_size, 0, ctx)) < 0) {
  297. av_log(ctx, AV_LOG_ERROR,
  298. "The text file '%s' could not be read or is empty\n",
  299. dtext->textfile);
  300. return err;
  301. }
  302. if (!(dtext->text = av_malloc(textbuf_size+1)))
  303. return AVERROR(ENOMEM);
  304. memcpy(dtext->text, textbuf, textbuf_size);
  305. dtext->text[textbuf_size] = 0;
  306. av_file_unmap(textbuf, textbuf_size);
  307. }
  308. if (dtext->tc_opt_string) {
  309. int ret = av_timecode_init_from_string(&dtext->tc, dtext->tc_rate,
  310. dtext->tc_opt_string, ctx);
  311. if (ret < 0)
  312. return ret;
  313. if (!dtext->text)
  314. dtext->text = av_strdup("");
  315. }
  316. if (!dtext->text) {
  317. av_log(ctx, AV_LOG_ERROR,
  318. "Either text, a valid file or a timecode must be provided\n");
  319. return AVERROR(EINVAL);
  320. }
  321. if ((err = av_parse_color(dtext->fontcolor_rgba, dtext->fontcolor_string, -1, ctx))) {
  322. av_log(ctx, AV_LOG_ERROR,
  323. "Invalid font color '%s'\n", dtext->fontcolor_string);
  324. return err;
  325. }
  326. if ((err = av_parse_color(dtext->boxcolor_rgba, dtext->boxcolor_string, -1, ctx))) {
  327. av_log(ctx, AV_LOG_ERROR,
  328. "Invalid box color '%s'\n", dtext->boxcolor_string);
  329. return err;
  330. }
  331. if ((err = av_parse_color(dtext->shadowcolor_rgba, dtext->shadowcolor_string, -1, ctx))) {
  332. av_log(ctx, AV_LOG_ERROR,
  333. "Invalid shadow color '%s'\n", dtext->shadowcolor_string);
  334. return err;
  335. }
  336. if ((err = FT_Init_FreeType(&(dtext->library)))) {
  337. av_log(ctx, AV_LOG_ERROR,
  338. "Could not load FreeType: %s\n", FT_ERRMSG(err));
  339. return AVERROR(EINVAL);
  340. }
  341. /* load the face, and set up the encoding, which is by default UTF-8 */
  342. if ((err = FT_New_Face(dtext->library, dtext->fontfile, 0, &dtext->face))) {
  343. av_log(ctx, AV_LOG_ERROR, "Could not load fontface from file '%s': %s\n",
  344. dtext->fontfile, FT_ERRMSG(err));
  345. return AVERROR(EINVAL);
  346. }
  347. if ((err = FT_Set_Pixel_Sizes(dtext->face, 0, dtext->fontsize))) {
  348. av_log(ctx, AV_LOG_ERROR, "Could not set font size to %d pixels: %s\n",
  349. dtext->fontsize, FT_ERRMSG(err));
  350. return AVERROR(EINVAL);
  351. }
  352. dtext->use_kerning = FT_HAS_KERNING(dtext->face);
  353. /* load the fallback glyph with code 0 */
  354. load_glyph(ctx, NULL, 0);
  355. /* set the tabsize in pixels */
  356. if ((err = load_glyph(ctx, &glyph, ' ')) < 0) {
  357. av_log(ctx, AV_LOG_ERROR, "Could not set tabsize.\n");
  358. return err;
  359. }
  360. dtext->tabsize *= glyph->advance;
  361. return 0;
  362. }
  363. static int query_formats(AVFilterContext *ctx)
  364. {
  365. static const enum PixelFormat pix_fmts[] = {
  366. PIX_FMT_ARGB, PIX_FMT_RGBA,
  367. PIX_FMT_ABGR, PIX_FMT_BGRA,
  368. PIX_FMT_RGB24, PIX_FMT_BGR24,
  369. PIX_FMT_YUV420P, PIX_FMT_YUV444P,
  370. PIX_FMT_YUV422P, PIX_FMT_YUV411P,
  371. PIX_FMT_YUV410P, PIX_FMT_YUV440P,
  372. PIX_FMT_NONE
  373. };
  374. avfilter_set_common_pixel_formats(ctx, avfilter_make_format_list(pix_fmts));
  375. return 0;
  376. }
  377. static int glyph_enu_free(void *opaque, void *elem)
  378. {
  379. av_free(elem);
  380. return 0;
  381. }
  382. static av_cold void uninit(AVFilterContext *ctx)
  383. {
  384. DrawTextContext *dtext = ctx->priv;
  385. int i;
  386. av_expr_free(dtext->x_pexpr); dtext->x_pexpr = NULL;
  387. av_expr_free(dtext->y_pexpr); dtext->y_pexpr = NULL;
  388. av_freep(&dtext->boxcolor_string);
  389. av_freep(&dtext->expanded_text);
  390. av_freep(&dtext->fontcolor_string);
  391. av_freep(&dtext->fontfile);
  392. av_freep(&dtext->shadowcolor_string);
  393. av_freep(&dtext->text);
  394. av_freep(&dtext->x_expr);
  395. av_freep(&dtext->y_expr);
  396. av_freep(&dtext->positions);
  397. dtext->nb_positions = 0;
  398. av_tree_enumerate(dtext->glyphs, NULL, NULL, glyph_enu_free);
  399. av_tree_destroy(dtext->glyphs);
  400. dtext->glyphs = NULL;
  401. FT_Done_Face(dtext->face);
  402. FT_Done_FreeType(dtext->library);
  403. for (i = 0; i < 4; i++) {
  404. av_freep(&dtext->box_line[i]);
  405. dtext->pixel_step[i] = 0;
  406. }
  407. }
  408. static inline int is_newline(uint32_t c)
  409. {
  410. return c == '\n' || c == '\r' || c == '\f' || c == '\v';
  411. }
  412. static int config_input(AVFilterLink *inlink)
  413. {
  414. AVFilterContext *ctx = inlink->dst;
  415. DrawTextContext *dtext = ctx->priv;
  416. const AVPixFmtDescriptor *pix_desc = &av_pix_fmt_descriptors[inlink->format];
  417. int ret;
  418. dtext->hsub = pix_desc->log2_chroma_w;
  419. dtext->vsub = pix_desc->log2_chroma_h;
  420. if ((ret =
  421. ff_fill_line_with_color(dtext->box_line, dtext->pixel_step,
  422. inlink->w, dtext->boxcolor,
  423. inlink->format, dtext->boxcolor_rgba,
  424. &dtext->is_packed_rgb, dtext->rgba_map)) < 0)
  425. return ret;
  426. if (!dtext->is_packed_rgb) {
  427. uint8_t *rgba = dtext->fontcolor_rgba;
  428. dtext->fontcolor[0] = RGB_TO_Y_CCIR(rgba[0], rgba[1], rgba[2]);
  429. dtext->fontcolor[1] = RGB_TO_U_CCIR(rgba[0], rgba[1], rgba[2], 0);
  430. dtext->fontcolor[2] = RGB_TO_V_CCIR(rgba[0], rgba[1], rgba[2], 0);
  431. dtext->fontcolor[3] = rgba[3];
  432. rgba = dtext->shadowcolor_rgba;
  433. dtext->shadowcolor[0] = RGB_TO_Y_CCIR(rgba[0], rgba[1], rgba[2]);
  434. dtext->shadowcolor[1] = RGB_TO_U_CCIR(rgba[0], rgba[1], rgba[2], 0);
  435. dtext->shadowcolor[2] = RGB_TO_V_CCIR(rgba[0], rgba[1], rgba[2], 0);
  436. dtext->shadowcolor[3] = rgba[3];
  437. }
  438. dtext->var_values[VAR_w] = dtext->var_values[VAR_W] = dtext->var_values[VAR_MAIN_W] = inlink->w;
  439. dtext->var_values[VAR_h] = dtext->var_values[VAR_H] = dtext->var_values[VAR_MAIN_H] = inlink->h;
  440. dtext->var_values[VAR_SAR] = inlink->sample_aspect_ratio.num ? av_q2d(inlink->sample_aspect_ratio) : 1;
  441. dtext->var_values[VAR_DAR] = (double)inlink->w / inlink->h * dtext->var_values[VAR_SAR];
  442. dtext->var_values[VAR_HSUB] = 1<<pix_desc->log2_chroma_w;
  443. dtext->var_values[VAR_VSUB] = 1<<pix_desc->log2_chroma_h;
  444. dtext->var_values[VAR_X] = NAN;
  445. dtext->var_values[VAR_Y] = NAN;
  446. if (!dtext->reinit)
  447. dtext->var_values[VAR_N] = 0;
  448. dtext->var_values[VAR_T] = NAN;
  449. av_lfg_init(&dtext->prng, av_get_random_seed());
  450. if ((ret = av_expr_parse(&dtext->x_pexpr, dtext->x_expr, var_names,
  451. NULL, NULL, fun2_names, fun2, 0, ctx)) < 0 ||
  452. (ret = av_expr_parse(&dtext->y_pexpr, dtext->y_expr, var_names,
  453. NULL, NULL, fun2_names, fun2, 0, ctx)) < 0 ||
  454. (ret = av_expr_parse(&dtext->d_pexpr, dtext->d_expr, var_names,
  455. NULL, NULL, fun2_names, fun2, 0, ctx)) < 0)
  456. return AVERROR(EINVAL);
  457. return 0;
  458. }
  459. static int command(AVFilterContext *ctx, const char *cmd, const char *arg, char *res, int res_len, int flags)
  460. {
  461. DrawTextContext *dtext = ctx->priv;
  462. if (!strcmp(cmd, "reinit")) {
  463. int ret;
  464. uninit(ctx);
  465. dtext->reinit = 1;
  466. if ((ret = init(ctx, arg, NULL)) < 0)
  467. return ret;
  468. return config_input(ctx->inputs[0]);
  469. }
  470. return AVERROR(ENOSYS);
  471. }
  472. #define GET_BITMAP_VAL(r, c) \
  473. bitmap->pixel_mode == FT_PIXEL_MODE_MONO ? \
  474. (bitmap->buffer[(r) * bitmap->pitch + ((c)>>3)] & (0x80 >> ((c)&7))) * 255 : \
  475. bitmap->buffer[(r) * bitmap->pitch + (c)]
  476. #define SET_PIXEL_YUV(picref, yuva_color, val, x, y, hsub, vsub) { \
  477. luma_pos = ((x) ) + ((y) ) * picref->linesize[0]; \
  478. alpha = yuva_color[3] * (val) * 129; \
  479. picref->data[0][luma_pos] = (alpha * yuva_color[0] + (255*255*129 - alpha) * picref->data[0][luma_pos] ) >> 23; \
  480. if (((x) & ((1<<(hsub)) - 1)) == 0 && ((y) & ((1<<(vsub)) - 1)) == 0) {\
  481. chroma_pos1 = ((x) >> (hsub)) + ((y) >> (vsub)) * picref->linesize[1]; \
  482. chroma_pos2 = ((x) >> (hsub)) + ((y) >> (vsub)) * picref->linesize[2]; \
  483. picref->data[1][chroma_pos1] = (alpha * yuva_color[1] + (255*255*129 - alpha) * picref->data[1][chroma_pos1]) >> 23; \
  484. picref->data[2][chroma_pos2] = (alpha * yuva_color[2] + (255*255*129 - alpha) * picref->data[2][chroma_pos2]) >> 23; \
  485. }\
  486. }
  487. static inline int draw_glyph_yuv(AVFilterBufferRef *picref, FT_Bitmap *bitmap,
  488. int x, int y, int width, int height,
  489. const uint8_t yuva_color[4], int hsub, int vsub)
  490. {
  491. int r, c, alpha;
  492. unsigned int luma_pos, chroma_pos1, chroma_pos2;
  493. uint8_t src_val;
  494. for (r = 0; r < bitmap->rows && r+y < height; r++) {
  495. for (c = 0; c < bitmap->width && c+x < width; c++) {
  496. if (c+x < 0 || r+y < 0)
  497. continue;
  498. /* get intensity value in the glyph bitmap (source) */
  499. src_val = GET_BITMAP_VAL(r, c);
  500. if (!src_val)
  501. continue;
  502. SET_PIXEL_YUV(picref, yuva_color, src_val, c+x, y+r, hsub, vsub);
  503. }
  504. }
  505. return 0;
  506. }
  507. #define SET_PIXEL_RGB(picref, rgba_color, val, x, y, pixel_step, r_off, g_off, b_off, a_off) { \
  508. p = picref->data[0] + (x) * pixel_step + ((y) * picref->linesize[0]); \
  509. alpha = rgba_color[3] * (val) * 129; \
  510. *(p+r_off) = (alpha * rgba_color[0] + (255*255*129 - alpha) * *(p+r_off)) >> 23; \
  511. *(p+g_off) = (alpha * rgba_color[1] + (255*255*129 - alpha) * *(p+g_off)) >> 23; \
  512. *(p+b_off) = (alpha * rgba_color[2] + (255*255*129 - alpha) * *(p+b_off)) >> 23; \
  513. }
  514. static inline int draw_glyph_rgb(AVFilterBufferRef *picref, FT_Bitmap *bitmap,
  515. int x, int y, int width, int height, int pixel_step,
  516. const uint8_t rgba_color[4], const uint8_t rgba_map[4])
  517. {
  518. int r, c, alpha;
  519. uint8_t *p;
  520. uint8_t src_val;
  521. for (r = 0; r < bitmap->rows && r+y < height; r++) {
  522. for (c = 0; c < bitmap->width && c+x < width; c++) {
  523. if (c+x < 0 || r+y < 0)
  524. continue;
  525. /* get intensity value in the glyph bitmap (source) */
  526. src_val = GET_BITMAP_VAL(r, c);
  527. if (!src_val)
  528. continue;
  529. SET_PIXEL_RGB(picref, rgba_color, src_val, c+x, y+r, pixel_step,
  530. rgba_map[0], rgba_map[1], rgba_map[2], rgba_map[3]);
  531. }
  532. }
  533. return 0;
  534. }
  535. static inline void drawbox(AVFilterBufferRef *picref, int x, int y,
  536. int width, int height,
  537. uint8_t *line[4], int pixel_step[4], uint8_t color[4],
  538. int hsub, int vsub, int is_rgba_packed, uint8_t rgba_map[4])
  539. {
  540. int i, j, alpha;
  541. if (color[3] != 0xFF) {
  542. if (is_rgba_packed) {
  543. uint8_t *p;
  544. for (j = 0; j < height; j++)
  545. for (i = 0; i < width; i++)
  546. SET_PIXEL_RGB(picref, color, 255, i+x, y+j, pixel_step[0],
  547. rgba_map[0], rgba_map[1], rgba_map[2], rgba_map[3]);
  548. } else {
  549. unsigned int luma_pos, chroma_pos1, chroma_pos2;
  550. for (j = 0; j < height; j++)
  551. for (i = 0; i < width; i++)
  552. SET_PIXEL_YUV(picref, color, 255, i+x, y+j, hsub, vsub);
  553. }
  554. } else {
  555. ff_draw_rectangle(picref->data, picref->linesize,
  556. line, pixel_step, hsub, vsub,
  557. x, y, width, height);
  558. }
  559. }
  560. static int draw_glyphs(DrawTextContext *dtext, AVFilterBufferRef *picref,
  561. int width, int height, const uint8_t rgbcolor[4], const uint8_t yuvcolor[4], int x, int y)
  562. {
  563. char *text = dtext->expanded_text;
  564. uint32_t code = 0;
  565. int i, x1, y1;
  566. uint8_t *p;
  567. Glyph *glyph = NULL;
  568. for (i = 0, p = text; *p; i++) {
  569. Glyph dummy = { 0 };
  570. GET_UTF8(code, *p++, continue;);
  571. /* skip new line chars, just go to new line */
  572. if (code == '\n' || code == '\r' || code == '\t')
  573. continue;
  574. dummy.code = code;
  575. glyph = av_tree_find(dtext->glyphs, &dummy, (void *)glyph_cmp, NULL);
  576. if (glyph->bitmap.pixel_mode != FT_PIXEL_MODE_MONO &&
  577. glyph->bitmap.pixel_mode != FT_PIXEL_MODE_GRAY)
  578. return AVERROR(EINVAL);
  579. x1 = dtext->positions[i].x+dtext->x+x;
  580. y1 = dtext->positions[i].y+dtext->y+y;
  581. if (dtext->is_packed_rgb) {
  582. draw_glyph_rgb(picref, &glyph->bitmap,
  583. x1, y1, width, height,
  584. dtext->pixel_step[0], rgbcolor, dtext->rgba_map);
  585. } else {
  586. draw_glyph_yuv(picref, &glyph->bitmap,
  587. x1, y1, width, height,
  588. yuvcolor, dtext->hsub, dtext->vsub);
  589. }
  590. }
  591. return 0;
  592. }
  593. static int draw_text(AVFilterContext *ctx, AVFilterBufferRef *picref,
  594. int width, int height)
  595. {
  596. DrawTextContext *dtext = ctx->priv;
  597. uint32_t code = 0, prev_code = 0;
  598. int x = 0, y = 0, i = 0, ret;
  599. int max_text_line_w = 0, len;
  600. int box_w, box_h;
  601. char *text = dtext->text;
  602. uint8_t *p;
  603. int y_min = 32000, y_max = -32000;
  604. int x_min = 32000, x_max = -32000;
  605. FT_Vector delta;
  606. Glyph *glyph = NULL, *prev_glyph = NULL;
  607. Glyph dummy = { 0 };
  608. time_t now = time(0);
  609. struct tm ltime;
  610. uint8_t *buf = dtext->expanded_text;
  611. int buf_size = dtext->expanded_text_size;
  612. if(dtext->basetime != AV_NOPTS_VALUE)
  613. now= picref->pts*av_q2d(ctx->inputs[0]->time_base) + dtext->basetime/1000000;
  614. if (!buf) {
  615. buf_size = 2*strlen(dtext->text)+1;
  616. buf = av_malloc(buf_size);
  617. }
  618. #if HAVE_LOCALTIME_R
  619. localtime_r(&now, &ltime);
  620. #else
  621. if(strchr(dtext->text, '%'))
  622. ltime= *localtime(&now);
  623. #endif
  624. do {
  625. *buf = 1;
  626. if (strftime(buf, buf_size, dtext->text, &ltime) != 0 || *buf == 0)
  627. break;
  628. buf_size *= 2;
  629. } while ((buf = av_realloc(buf, buf_size)));
  630. if (dtext->tc_opt_string) {
  631. char tcbuf[AV_TIMECODE_STR_SIZE];
  632. av_timecode_make_string(&dtext->tc, tcbuf, dtext->frame_id++);
  633. buf = av_asprintf("%s%s", dtext->text, tcbuf);
  634. }
  635. if (!buf)
  636. return AVERROR(ENOMEM);
  637. text = dtext->expanded_text = buf;
  638. dtext->expanded_text_size = buf_size;
  639. if ((len = strlen(text)) > dtext->nb_positions) {
  640. if (!(dtext->positions =
  641. av_realloc(dtext->positions, len*sizeof(*dtext->positions))))
  642. return AVERROR(ENOMEM);
  643. dtext->nb_positions = len;
  644. }
  645. x = 0;
  646. y = 0;
  647. /* load and cache glyphs */
  648. for (i = 0, p = text; *p; i++) {
  649. GET_UTF8(code, *p++, continue;);
  650. /* get glyph */
  651. dummy.code = code;
  652. glyph = av_tree_find(dtext->glyphs, &dummy, glyph_cmp, NULL);
  653. if (!glyph) {
  654. load_glyph(ctx, &glyph, code);
  655. }
  656. y_min = FFMIN(glyph->bbox.yMin, y_min);
  657. y_max = FFMAX(glyph->bbox.yMax, y_max);
  658. x_min = FFMIN(glyph->bbox.xMin, x_min);
  659. x_max = FFMAX(glyph->bbox.xMax, x_max);
  660. }
  661. dtext->max_glyph_h = y_max - y_min;
  662. dtext->max_glyph_w = x_max - x_min;
  663. /* compute and save position for each glyph */
  664. glyph = NULL;
  665. for (i = 0, p = text; *p; i++) {
  666. GET_UTF8(code, *p++, continue;);
  667. /* skip the \n in the sequence \r\n */
  668. if (prev_code == '\r' && code == '\n')
  669. continue;
  670. prev_code = code;
  671. if (is_newline(code)) {
  672. max_text_line_w = FFMAX(max_text_line_w, x);
  673. y += dtext->max_glyph_h;
  674. x = 0;
  675. continue;
  676. }
  677. /* get glyph */
  678. prev_glyph = glyph;
  679. dummy.code = code;
  680. glyph = av_tree_find(dtext->glyphs, &dummy, glyph_cmp, NULL);
  681. /* kerning */
  682. if (dtext->use_kerning && prev_glyph && glyph->code) {
  683. FT_Get_Kerning(dtext->face, prev_glyph->code, glyph->code,
  684. ft_kerning_default, &delta);
  685. x += delta.x >> 6;
  686. }
  687. /* save position */
  688. dtext->positions[i].x = x + glyph->bitmap_left;
  689. dtext->positions[i].y = y - glyph->bitmap_top + y_max;
  690. if (code == '\t') x = (x / dtext->tabsize + 1)*dtext->tabsize;
  691. else x += glyph->advance;
  692. }
  693. max_text_line_w = FFMAX(x, max_text_line_w);
  694. dtext->var_values[VAR_TW] = dtext->var_values[VAR_TEXT_W] = max_text_line_w;
  695. dtext->var_values[VAR_TH] = dtext->var_values[VAR_TEXT_H] = y + dtext->max_glyph_h;
  696. dtext->var_values[VAR_MAX_GLYPH_W] = dtext->max_glyph_w;
  697. dtext->var_values[VAR_MAX_GLYPH_H] = dtext->max_glyph_h;
  698. dtext->var_values[VAR_MAX_GLYPH_A] = dtext->var_values[VAR_ASCENT ] = y_max;
  699. dtext->var_values[VAR_MAX_GLYPH_D] = dtext->var_values[VAR_DESCENT] = y_min;
  700. dtext->var_values[VAR_LINE_H] = dtext->var_values[VAR_LH] = dtext->max_glyph_h;
  701. dtext->x = dtext->var_values[VAR_X] = av_expr_eval(dtext->x_pexpr, dtext->var_values, &dtext->prng);
  702. dtext->y = dtext->var_values[VAR_Y] = av_expr_eval(dtext->y_pexpr, dtext->var_values, &dtext->prng);
  703. dtext->x = dtext->var_values[VAR_X] = av_expr_eval(dtext->x_pexpr, dtext->var_values, &dtext->prng);
  704. dtext->draw = av_expr_eval(dtext->d_pexpr, dtext->var_values, &dtext->prng);
  705. if(!dtext->draw)
  706. return 0;
  707. dtext->x &= ~((1 << dtext->hsub) - 1);
  708. dtext->y &= ~((1 << dtext->vsub) - 1);
  709. box_w = FFMIN(width - 1 , max_text_line_w);
  710. box_h = FFMIN(height - 1, y + dtext->max_glyph_h);
  711. /* draw box */
  712. if (dtext->draw_box)
  713. drawbox(picref, dtext->x, dtext->y, box_w, box_h,
  714. dtext->box_line, dtext->pixel_step, dtext->is_packed_rgb ? dtext->boxcolor_rgba : dtext->boxcolor,
  715. dtext->hsub, dtext->vsub, dtext->is_packed_rgb, dtext->rgba_map);
  716. if (dtext->shadowx || dtext->shadowy) {
  717. if ((ret = draw_glyphs(dtext, picref, width, height, dtext->shadowcolor_rgba,
  718. dtext->shadowcolor, dtext->shadowx, dtext->shadowy)) < 0)
  719. return ret;
  720. }
  721. if ((ret = draw_glyphs(dtext, picref, width, height, dtext->fontcolor_rgba,
  722. dtext->fontcolor, 0, 0)) < 0)
  723. return ret;
  724. return 0;
  725. }
  726. static void null_draw_slice(AVFilterLink *link, int y, int h, int slice_dir) { }
  727. static void end_frame(AVFilterLink *inlink)
  728. {
  729. AVFilterContext *ctx = inlink->dst;
  730. AVFilterLink *outlink = ctx->outputs[0];
  731. DrawTextContext *dtext = ctx->priv;
  732. AVFilterBufferRef *picref = inlink->cur_buf;
  733. dtext->var_values[VAR_T] = picref->pts == AV_NOPTS_VALUE ?
  734. NAN : picref->pts * av_q2d(inlink->time_base);
  735. draw_text(ctx, picref, picref->video->w, picref->video->h);
  736. av_log(ctx, AV_LOG_DEBUG, "n:%d t:%f text_w:%d text_h:%d x:%d y:%d\n",
  737. (int)dtext->var_values[VAR_N], dtext->var_values[VAR_T],
  738. (int)dtext->var_values[VAR_TEXT_W], (int)dtext->var_values[VAR_TEXT_H],
  739. dtext->x, dtext->y);
  740. dtext->var_values[VAR_N] += 1.0;
  741. avfilter_draw_slice(outlink, 0, picref->video->h, 1);
  742. avfilter_end_frame(outlink);
  743. }
  744. AVFilter avfilter_vf_drawtext = {
  745. .name = "drawtext",
  746. .description = NULL_IF_CONFIG_SMALL("Draw text on top of video frames using libfreetype library."),
  747. .priv_size = sizeof(DrawTextContext),
  748. .init = init,
  749. .uninit = uninit,
  750. .query_formats = query_formats,
  751. .inputs = (const AVFilterPad[]) {{ .name = "default",
  752. .type = AVMEDIA_TYPE_VIDEO,
  753. .get_video_buffer = avfilter_null_get_video_buffer,
  754. .start_frame = avfilter_null_start_frame,
  755. .draw_slice = null_draw_slice,
  756. .end_frame = end_frame,
  757. .config_props = config_input,
  758. .min_perms = AV_PERM_WRITE |
  759. AV_PERM_READ,
  760. .rej_perms = AV_PERM_PRESERVE },
  761. { .name = NULL}},
  762. .outputs = (const AVFilterPad[]) {{ .name = "default",
  763. .type = AVMEDIA_TYPE_VIDEO, },
  764. { .name = NULL}},
  765. .process_command = command,
  766. };