Line | Branch | Exec | Source |
---|---|---|---|
1 | /* | ||
2 | * TTML subtitle encoder | ||
3 | * Copyright (c) 2020 24i | ||
4 | * | ||
5 | * This file is part of FFmpeg. | ||
6 | * | ||
7 | * FFmpeg is free software; you can redistribute it and/or | ||
8 | * modify it under the terms of the GNU Lesser General Public | ||
9 | * License as published by the Free Software Foundation; either | ||
10 | * version 2.1 of the License, or (at your option) any later version. | ||
11 | * | ||
12 | * FFmpeg is distributed in the hope that it will be useful, | ||
13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
15 | * Lesser General Public License for more details. | ||
16 | * | ||
17 | * You should have received a copy of the GNU Lesser General Public | ||
18 | * License along with FFmpeg; if not, write to the Free Software | ||
19 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||
20 | */ | ||
21 | |||
22 | /** | ||
23 | * @file | ||
24 | * TTML subtitle encoder | ||
25 | * @see https://www.w3.org/TR/ttml1/ | ||
26 | * @see https://www.w3.org/TR/ttml2/ | ||
27 | * @see https://www.w3.org/TR/ttml-imsc/rec | ||
28 | */ | ||
29 | |||
30 | #include "avcodec.h" | ||
31 | #include "codec_internal.h" | ||
32 | #include "libavutil/bprint.h" | ||
33 | #include "libavutil/internal.h" | ||
34 | #include "libavutil/mem.h" | ||
35 | #include "ass_split.h" | ||
36 | #include "ttmlenc.h" | ||
37 | |||
38 | typedef struct { | ||
39 | AVCodecContext *avctx; | ||
40 | ASSSplitContext *ass_ctx; | ||
41 | AVBPrint buffer; | ||
42 | } TTMLContext; | ||
43 | |||
44 | 306 | static void ttml_text_cb(void *priv, const char *text, int len) | |
45 | { | ||
46 | 306 | TTMLContext *s = priv; | |
47 | AVBPrint cur_line; | ||
48 | 306 | AVBPrint *buffer = &s->buffer; | |
49 | |||
50 | 306 | av_bprint_init(&cur_line, len, AV_BPRINT_SIZE_UNLIMITED); | |
51 | |||
52 | 306 | av_bprint_append_data(&cur_line, text, len); | |
53 |
1/2✗ Branch 1 not taken.
✓ Branch 2 taken 306 times.
|
306 | if (!av_bprint_is_complete(&cur_line)) { |
54 | ✗ | av_log(s->avctx, AV_LOG_ERROR, | |
55 | "Failed to move the current subtitle dialog to AVBPrint!\n"); | ||
56 | ✗ | av_bprint_finalize(&cur_line, NULL); | |
57 | ✗ | return; | |
58 | } | ||
59 | |||
60 | |||
61 | 306 | av_bprint_escape(buffer, cur_line.str, NULL, AV_ESCAPE_MODE_XML, | |
62 | 0); | ||
63 | |||
64 | 306 | av_bprint_finalize(&cur_line, NULL); | |
65 | } | ||
66 | |||
67 | 195 | static void ttml_new_line_cb(void *priv, int forced) | |
68 | { | ||
69 | 195 | TTMLContext *s = priv; | |
70 | |||
71 | 195 | av_bprintf(&s->buffer, "<br/>"); | |
72 | 195 | } | |
73 | |||
74 | static const ASSCodesCallbacks ttml_callbacks = { | ||
75 | .text = ttml_text_cb, | ||
76 | .new_line = ttml_new_line_cb, | ||
77 | }; | ||
78 | |||
79 | 111 | static int ttml_encode_frame(AVCodecContext *avctx, uint8_t *buf, | |
80 | int bufsize, const AVSubtitle *sub) | ||
81 | { | ||
82 | 111 | TTMLContext *s = avctx->priv_data; | |
83 | ASSDialog *dialog; | ||
84 | int i; | ||
85 | |||
86 | 111 | av_bprint_init_for_buffer(&s->buffer, buf, bufsize); | |
87 | |||
88 |
2/2✓ Branch 0 taken 111 times.
✓ Branch 1 taken 111 times.
|
222 | for (i=0; i<sub->num_rects; i++) { |
89 | 111 | const char *ass = sub->rects[i]->ass; | |
90 | int ret; | ||
91 | |||
92 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 111 times.
|
111 | if (sub->rects[i]->type != SUBTITLE_ASS) { |
93 | ✗ | av_log(avctx, AV_LOG_ERROR, "Only SUBTITLE_ASS type supported.\n"); | |
94 | ✗ | return AVERROR(EINVAL); | |
95 | } | ||
96 | |||
97 | 111 | dialog = ff_ass_split_dialog(s->ass_ctx, ass); | |
98 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 111 times.
|
111 | if (!dialog) |
99 | ✗ | return AVERROR(ENOMEM); | |
100 | |||
101 |
1/2✓ Branch 0 taken 111 times.
✗ Branch 1 not taken.
|
111 | if (dialog->style) { |
102 | 111 | av_bprintf(&s->buffer, "<span region=\""); | |
103 | 111 | av_bprint_escape(&s->buffer, dialog->style, NULL, | |
104 | AV_ESCAPE_MODE_XML, | ||
105 | AV_ESCAPE_FLAG_XML_DOUBLE_QUOTES); | ||
106 | 111 | av_bprintf(&s->buffer, "\">"); | |
107 | } | ||
108 | |||
109 | 111 | ret = ff_ass_split_override_codes(&ttml_callbacks, s, dialog->text); | |
110 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 111 times.
|
111 | if (ret < 0) { |
111 | ✗ | int log_level = (ret != AVERROR_INVALIDDATA || | |
112 | ✗ | avctx->err_recognition & AV_EF_EXPLODE) ? | |
113 | ✗ | AV_LOG_ERROR : AV_LOG_WARNING; | |
114 | ✗ | av_log(avctx, log_level, | |
115 | "Splitting received ASS dialog text %s failed: %s\n", | ||
116 | ✗ | dialog->text, | |
117 | ✗ | av_err2str(ret)); | |
118 | |||
119 | ✗ | if (log_level == AV_LOG_ERROR) { | |
120 | ✗ | ff_ass_free_dialog(&dialog); | |
121 | ✗ | return ret; | |
122 | } | ||
123 | } | ||
124 | |||
125 |
1/2✓ Branch 0 taken 111 times.
✗ Branch 1 not taken.
|
111 | if (dialog->style) |
126 | 111 | av_bprintf(&s->buffer, "</span>"); | |
127 | |||
128 | 111 | ff_ass_free_dialog(&dialog); | |
129 | } | ||
130 | |||
131 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 111 times.
|
111 | if (!s->buffer.len) |
132 | ✗ | return 0; | |
133 |
1/2✗ Branch 1 not taken.
✓ Branch 2 taken 111 times.
|
111 | if (!av_bprint_is_complete(&s->buffer)) { |
134 | ✗ | av_log(avctx, AV_LOG_ERROR, "Buffer too small for TTML event.\n"); | |
135 | ✗ | return AVERROR_BUFFER_TOO_SMALL; | |
136 | } | ||
137 | |||
138 | 111 | return s->buffer.len; | |
139 | } | ||
140 | |||
141 | 3 | static av_cold int ttml_encode_close(AVCodecContext *avctx) | |
142 | { | ||
143 | 3 | TTMLContext *s = avctx->priv_data; | |
144 | |||
145 | 3 | ff_ass_split_free(s->ass_ctx); | |
146 | |||
147 | 3 | return 0; | |
148 | } | ||
149 | |||
150 | 3 | static const char *ttml_get_display_alignment(int alignment) | |
151 | { | ||
152 |
1/4✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
|
3 | switch (alignment) { |
153 | 3 | case 1: | |
154 | case 2: | ||
155 | case 3: | ||
156 | 3 | return "after"; | |
157 | ✗ | case 4: | |
158 | case 5: | ||
159 | case 6: | ||
160 | ✗ | return "center"; | |
161 | ✗ | case 7: | |
162 | case 8: | ||
163 | case 9: | ||
164 | ✗ | return "before"; | |
165 | ✗ | default: | |
166 | ✗ | return NULL; | |
167 | } | ||
168 | } | ||
169 | |||
170 | 3 | static const char *ttml_get_text_alignment(int alignment) | |
171 | { | ||
172 |
1/4✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
|
3 | switch (alignment) { |
173 | ✗ | case 1: | |
174 | case 4: | ||
175 | case 7: | ||
176 | ✗ | return "left"; | |
177 | 3 | case 2: | |
178 | case 5: | ||
179 | case 8: | ||
180 | 3 | return "center"; | |
181 | ✗ | case 3: | |
182 | case 6: | ||
183 | case 9: | ||
184 | ✗ | return "right"; | |
185 | ✗ | default: | |
186 | ✗ | return NULL; | |
187 | } | ||
188 | } | ||
189 | |||
190 | 3 | static void ttml_get_origin(ASSScriptInfo script_info, ASSStyle style, | |
191 | int *origin_left, int *origin_top) | ||
192 | { | ||
193 | 3 | *origin_left = av_rescale(style.margin_l, 100, script_info.play_res_x); | |
194 | 3 | *origin_top = | |
195 | 3 | av_rescale((style.alignment >= 7) ? style.margin_v : 0, | |
196 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
|
3 | 100, script_info.play_res_y); |
197 | 3 | } | |
198 | |||
199 | 3 | static void ttml_get_extent(ASSScriptInfo script_info, ASSStyle style, | |
200 | int *width, int *height) | ||
201 | { | ||
202 | 3 | *width = av_rescale(script_info.play_res_x - style.margin_r, | |
203 | 3 | 100, script_info.play_res_x); | |
204 | 9 | *height = av_rescale((style.alignment <= 3) ? | |
205 | 3 | script_info.play_res_y - style.margin_v : | |
206 | ✗ | script_info.play_res_y, | |
207 |
1/2✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
|
3 | 100, script_info.play_res_y); |
208 | 3 | } | |
209 | |||
210 | 3 | static int ttml_write_region(AVCodecContext *avctx, AVBPrint *buf, | |
211 | ASSScriptInfo script_info, ASSStyle style) | ||
212 | { | ||
213 | 3 | const char *display_alignment = NULL; | |
214 | 3 | const char *text_alignment = NULL; | |
215 | 3 | int origin_left = 0; | |
216 | 3 | int origin_top = 0; | |
217 | 3 | int width = 0; | |
218 | 3 | int height = 0; | |
219 | |||
220 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
|
3 | if (!style.name) { |
221 | ✗ | av_log(avctx, AV_LOG_ERROR, "Subtitle style name not set!\n"); | |
222 | ✗ | return AVERROR_INVALIDDATA; | |
223 | } | ||
224 | |||
225 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
|
3 | if (style.font_size < 0) { |
226 | ✗ | av_log(avctx, AV_LOG_ERROR, "Invalid font size for TTML: %d!\n", | |
227 | style.font_size); | ||
228 | ✗ | return AVERROR_INVALIDDATA; | |
229 | } | ||
230 | |||
231 |
3/6✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
✓ Branch 2 taken 3 times.
✗ Branch 3 not taken.
✗ Branch 4 not taken.
✓ Branch 5 taken 3 times.
|
3 | if (style.margin_l < 0 || style.margin_r < 0 || style.margin_v < 0) { |
232 | ✗ | av_log(avctx, AV_LOG_ERROR, | |
233 | "One or more negative margin values in subtitle style: " | ||
234 | "left: %d, right: %d, vertical: %d!\n", | ||
235 | style.margin_l, style.margin_r, style.margin_v); | ||
236 | ✗ | return AVERROR_INVALIDDATA; | |
237 | } | ||
238 | |||
239 | 3 | display_alignment = ttml_get_display_alignment(style.alignment); | |
240 | 3 | text_alignment = ttml_get_text_alignment(style.alignment); | |
241 |
2/4✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
✗ Branch 2 not taken.
✓ Branch 3 taken 3 times.
|
3 | if (!display_alignment || !text_alignment) { |
242 | ✗ | av_log(avctx, AV_LOG_ERROR, | |
243 | "Failed to convert ASS style alignment %d of style %s to " | ||
244 | "TTML display and text alignment!\n", | ||
245 | style.alignment, | ||
246 | style.name); | ||
247 | ✗ | return AVERROR_INVALIDDATA; | |
248 | } | ||
249 | |||
250 | 3 | ttml_get_origin(script_info, style, &origin_left, &origin_top); | |
251 | 3 | ttml_get_extent(script_info, style, &width, &height); | |
252 | |||
253 | 3 | av_bprintf(buf, " <region xml:id=\""); | |
254 | 3 | av_bprint_escape(buf, style.name, NULL, AV_ESCAPE_MODE_XML, | |
255 | AV_ESCAPE_FLAG_XML_DOUBLE_QUOTES); | ||
256 | 3 | av_bprintf(buf, "\"\n"); | |
257 | |||
258 | 3 | av_bprintf(buf, " tts:origin=\"%d%% %d%%\"\n", | |
259 | origin_left, origin_top); | ||
260 | 3 | av_bprintf(buf, " tts:extent=\"%d%% %d%%\"\n", | |
261 | width, height); | ||
262 | |||
263 | 3 | av_bprintf(buf, " tts:displayAlign=\""); | |
264 | 3 | av_bprint_escape(buf, display_alignment, NULL, AV_ESCAPE_MODE_XML, | |
265 | AV_ESCAPE_FLAG_XML_DOUBLE_QUOTES); | ||
266 | 3 | av_bprintf(buf, "\"\n"); | |
267 | |||
268 | 3 | av_bprintf(buf, " tts:textAlign=\""); | |
269 | 3 | av_bprint_escape(buf, text_alignment, NULL, AV_ESCAPE_MODE_XML, | |
270 | AV_ESCAPE_FLAG_XML_DOUBLE_QUOTES); | ||
271 | 3 | av_bprintf(buf, "\"\n"); | |
272 | |||
273 | // if we set cell resolution to our script reference resolution, | ||
274 | // then a single line is a single "point" on our canvas. Thus, by setting | ||
275 | // our font size to font size in cells, we should gain a similar enough | ||
276 | // scale without resorting to explicit pixel based font sizing, which is | ||
277 | // frowned upon in the TTML community. | ||
278 | 3 | av_bprintf(buf, " tts:fontSize=\"%dc\"\n", | |
279 | style.font_size); | ||
280 | |||
281 |
1/2✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
|
3 | if (style.font_name) { |
282 | 3 | av_bprintf(buf, " tts:fontFamily=\""); | |
283 | 3 | av_bprint_escape(buf, style.font_name, NULL, AV_ESCAPE_MODE_XML, | |
284 | AV_ESCAPE_FLAG_XML_DOUBLE_QUOTES); | ||
285 | 3 | av_bprintf(buf, "\"\n"); | |
286 | } | ||
287 | |||
288 | 3 | av_bprintf(buf, " tts:overflow=\"visible\" />\n"); | |
289 | |||
290 | 3 | return 0; | |
291 | } | ||
292 | |||
293 | 3 | static int ttml_write_header_content(AVCodecContext *avctx) | |
294 | { | ||
295 | 3 | TTMLContext *s = avctx->priv_data; | |
296 | 3 | ASS *ass = (ASS *)s->ass_ctx; | |
297 | 3 | ASSScriptInfo script_info = ass->script_info; | |
298 | 3 | const size_t base_extradata_size = TTMLENC_EXTRADATA_SIGNATURE_SIZE + 1 + | |
299 | AV_INPUT_BUFFER_PADDING_SIZE; | ||
300 | 3 | size_t additional_extradata_size = 0; | |
301 | int ret; | ||
302 | |||
303 |
2/4✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
✗ Branch 2 not taken.
✓ Branch 3 taken 3 times.
|
3 | if (script_info.play_res_x <= 0 || script_info.play_res_y <= 0) { |
304 | ✗ | av_log(avctx, AV_LOG_ERROR, | |
305 | "Invalid subtitle reference resolution %dx%d!\n", | ||
306 | script_info.play_res_x, script_info.play_res_y); | ||
307 | ✗ | return AVERROR_INVALIDDATA; | |
308 | } | ||
309 | |||
310 | 3 | av_bprint_init(&s->buffer, 0, INT_MAX - base_extradata_size); | |
311 | |||
312 | // write the first string in extradata, attributes in the base "tt" element. | ||
313 | 3 | av_bprintf(&s->buffer, TTML_DEFAULT_NAMESPACING); | |
314 | // the cell resolution is in character cells, so not exactly 1:1 against | ||
315 | // a pixel based resolution, but as the tts:extent in the root | ||
316 | // "tt" element is frowned upon (and disallowed in the EBU-TT profile), | ||
317 | // we mimic the reference resolution by setting it as the cell resolution. | ||
318 | 3 | av_bprintf(&s->buffer, " ttp:cellResolution=\"%d %d\"\n", | |
319 | script_info.play_res_x, script_info.play_res_y); | ||
320 | 3 | av_bprint_chars(&s->buffer, '\0', 1); | |
321 | |||
322 | // write the second string in extradata, head element containing the styles | ||
323 | 3 | av_bprintf(&s->buffer, " <head>\n"); | |
324 | 3 | av_bprintf(&s->buffer, " <layout>\n"); | |
325 | |||
326 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
6 | for (int i = 0; i < ass->styles_count; i++) { |
327 | 3 | ret = ttml_write_region(avctx, &s->buffer, script_info, | |
328 | 3 | ass->styles[i]); | |
329 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
|
3 | if (ret < 0) |
330 | ✗ | goto fail; | |
331 | } | ||
332 | |||
333 | 3 | av_bprintf(&s->buffer, " </layout>\n"); | |
334 | 3 | av_bprintf(&s->buffer, " </head>\n"); | |
335 | 3 | av_bprint_chars(&s->buffer, '\0', 1); | |
336 | |||
337 |
1/2✗ Branch 1 not taken.
✓ Branch 2 taken 3 times.
|
3 | if (!av_bprint_is_complete(&s->buffer)) { |
338 | ✗ | ret = AVERROR(ENOMEM); | |
339 | ✗ | goto fail; | |
340 | } | ||
341 | |||
342 | 3 | additional_extradata_size = s->buffer.len; | |
343 | |||
344 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
|
3 | if (!(avctx->extradata = |
345 | 3 | av_mallocz(base_extradata_size + additional_extradata_size))) { | |
346 | ✗ | ret = AVERROR(ENOMEM); | |
347 | ✗ | goto fail; | |
348 | } | ||
349 | |||
350 | 3 | avctx->extradata_size = | |
351 | 3 | TTMLENC_EXTRADATA_SIGNATURE_SIZE + additional_extradata_size; | |
352 | 3 | memcpy(avctx->extradata, TTMLENC_EXTRADATA_SIGNATURE, | |
353 | TTMLENC_EXTRADATA_SIGNATURE_SIZE); | ||
354 | |||
355 | 3 | memcpy(avctx->extradata + TTMLENC_EXTRADATA_SIGNATURE_SIZE, | |
356 | 3 | s->buffer.str, additional_extradata_size); | |
357 | |||
358 | 3 | ret = 0; | |
359 | 3 | fail: | |
360 | 3 | av_bprint_finalize(&s->buffer, NULL); | |
361 | |||
362 | 3 | return ret; | |
363 | } | ||
364 | |||
365 | 3 | static av_cold int ttml_encode_init(AVCodecContext *avctx) | |
366 | { | ||
367 | 3 | TTMLContext *s = avctx->priv_data; | |
368 | 3 | int ret = AVERROR_BUG; | |
369 | 3 | s->avctx = avctx; | |
370 | |||
371 |
1/2✗ Branch 1 not taken.
✓ Branch 2 taken 3 times.
|
3 | if (!(s->ass_ctx = ff_ass_split(avctx->subtitle_header))) { |
372 | ✗ | return AVERROR_INVALIDDATA; | |
373 | } | ||
374 | |||
375 |
1/2✗ Branch 1 not taken.
✓ Branch 2 taken 3 times.
|
3 | if ((ret = ttml_write_header_content(avctx)) < 0) { |
376 | ✗ | return ret; | |
377 | } | ||
378 | |||
379 | 3 | return 0; | |
380 | } | ||
381 | |||
382 | const FFCodec ff_ttml_encoder = { | ||
383 | .p.name = "ttml", | ||
384 | CODEC_LONG_NAME("TTML subtitle"), | ||
385 | .p.type = AVMEDIA_TYPE_SUBTITLE, | ||
386 | .p.id = AV_CODEC_ID_TTML, | ||
387 | .priv_data_size = sizeof(TTMLContext), | ||
388 | .init = ttml_encode_init, | ||
389 | FF_CODEC_ENCODE_SUB_CB(ttml_encode_frame), | ||
390 | .close = ttml_encode_close, | ||
391 | .caps_internal = FF_CODEC_CAP_INIT_CLEANUP, | ||
392 | }; | ||
393 |