2 Copyright (C) 2016-2019 Carl Hetherington <cth@carlh.net>
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software
16 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
20 #include "ssa_reader.h"
22 #include "sub_assert.h"
23 #include "raw_convert.h"
25 #include "compose.hpp"
26 #include <boost/algorithm/string.hpp>
27 #include <boost/bind.hpp>
35 using boost::optional;
36 using boost::function;
37 using namespace boost::algorithm;
40 /** @param s Subtitle string encoded in UTF-8 */
41 SSAReader::SSAReader (string s)
43 this->read (boost::bind(&get_line_string, &s));
46 /** @param f Subtitle file encoded in UTF-8 */
47 SSAReader::SSAReader (FILE* f)
49 this->read (boost::bind (&get_line_file, f));
55 /* There are both BGR and ABGR versions of these colours */
56 if ((s.length() != 8 && s.length() != 10) || s[0] != '&' || s[1] != 'H') {
57 throw SSAError(String::compose("Badly formatted colour tag %1", s));
60 /* XXX: ignoring alpha channel here; note that 00 is opaque and FF is transparent */
61 int const off = s.length() == 10 ? 4 : 2;
62 if (sscanf(s.c_str() + off, "%2x%2x%2x", &ib, &ig, &ir) < 3) {
63 throw SSAError(String::compose("Badly formatted colour tag %1", s));
65 return sub::Colour(ir / 255.0, ig / 255.0, ib / 255.0);
73 , primary_colour (255, 255, 255)
77 , horizontal_reference (HORIZONTAL_CENTRE_OF_SCREEN)
78 , vertical_reference (BOTTOM_OF_SCREEN)
82 Style (string format_line, string style_line)
84 , primary_colour (255, 255, 255)
88 , horizontal_reference (HORIZONTAL_CENTRE_OF_SCREEN)
89 , vertical_reference (BOTTOM_OF_SCREEN)
93 split (keys, format_line, boost::is_any_of (","));
95 split (style, style_line, boost::is_any_of (","));
97 SUB_ASSERT (!keys.empty());
98 SUB_ASSERT (!style.empty());
99 SUB_ASSERT (keys.size() == style.size());
101 for (size_t i = 0; i < style.size(); ++i) {
104 if (keys[i] == "Name") {
106 } else if (keys[i] == "Fontname") {
107 font_name = style[i];
108 } else if (keys[i] == "Fontsize") {
109 font_size = raw_convert<int> (style[i]);
110 } else if (keys[i] == "PrimaryColour") {
111 primary_colour = colour (style[i]);
112 } else if (keys[i] == "BackColour") {
113 back_colour = colour (style[i]);
114 } else if (keys[i] == "Bold") {
115 bold = style[i] == "-1";
116 } else if (keys[i] == "Italic") {
117 italic = style[i] == "-1";
118 } else if (keys[i] == "Underline") {
119 underline = style[i] == "-1";
120 } else if (keys[i] == "BorderStyle") {
121 if (style[i] == "1") {
124 } else if (keys[i] == "Alignment") {
125 /* These values from libass' source code */
126 switch ((raw_convert<int> (style[i]) - 1) % 3) {
128 horizontal_reference = LEFT_OF_SCREEN;
131 horizontal_reference = HORIZONTAL_CENTRE_OF_SCREEN;
134 horizontal_reference = RIGHT_OF_SCREEN;
137 switch (raw_convert<int> (style[i]) & 12) {
139 vertical_reference = TOP_OF_SCREEN;
142 vertical_reference = VERTICAL_CENTRE_OF_SCREEN;
145 vertical_reference = BOTTOM_OF_SCREEN;
148 } else if (keys[i] == "MarginV") {
149 vertical_margin = raw_convert<int> (style[i]);
155 optional<string> font_name;
157 Colour primary_colour;
158 /** outline colour */
159 optional<Colour> back_colour;
163 optional<Effect> effect;
164 HorizontalReference horizontal_reference;
165 VerticalReference vertical_reference;
169 Colour colour (string c) const
171 if (c.length() > 0 && c[0] == '&') {
172 /* &Hbbggrr or &Haabbggrr */
176 int i = raw_convert<int>(c);
178 ((i & 0x0000ff) >> 0) / 255.0,
179 ((i & 0x00ff00) >> 8) / 255.0,
180 ((i & 0xff0000) >> 16) / 255.0
187 SSAReader::parse_time (string t) const
190 split (bits, t, is_any_of (":."));
191 SUB_ASSERT (bits.size() == 4);
192 return Time::from_hms (
193 raw_convert<int> (bits[0]),
194 raw_convert<int> (bits[1]),
195 raw_convert<int> (bits[2]),
196 raw_convert<int> (bits[3]) * 10
201 SSAReader::parse_style (RawSubtitle& sub, string style, int play_res_x, int play_res_y)
203 if (style == "\\i1") {
205 } else if (style == "\\i0" || style == "\\i") {
207 } else if (style == "\\b1") {
209 } else if (style == "\\b0") {
211 } else if (style == "\\u1") {
212 sub.underline = true;
213 } else if (style == "\\u0") {
214 sub.underline = false;
215 } else if (style == "\\an1") {
216 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
217 sub.vertical_position.reference = sub::BOTTOM_OF_SCREEN;
218 } else if (style == "\\an2") {
219 sub.horizontal_position.reference = sub::HORIZONTAL_CENTRE_OF_SCREEN;
220 sub.vertical_position.reference = sub::BOTTOM_OF_SCREEN;
221 } else if (style == "\\an3") {
222 sub.horizontal_position.reference = sub::RIGHT_OF_SCREEN;
223 sub.vertical_position.reference = sub::BOTTOM_OF_SCREEN;
224 } else if (style == "\\an4") {
225 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
226 sub.vertical_position.reference = sub::VERTICAL_CENTRE_OF_SCREEN;
227 } else if (style == "\\an5") {
228 sub.horizontal_position.reference = sub::HORIZONTAL_CENTRE_OF_SCREEN;
229 sub.vertical_position.reference = sub::VERTICAL_CENTRE_OF_SCREEN;
230 } else if (style == "\\an6") {
231 sub.horizontal_position.reference = sub::RIGHT_OF_SCREEN;
232 sub.vertical_position.reference = sub::VERTICAL_CENTRE_OF_SCREEN;
233 } else if (style == "\\an7") {
234 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
235 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
236 } else if (style == "\\an8") {
237 sub.horizontal_position.reference = sub::HORIZONTAL_CENTRE_OF_SCREEN;
238 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
239 } else if (style == "\\an9") {
240 sub.horizontal_position.reference = sub::RIGHT_OF_SCREEN;
241 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
242 } else if (boost::starts_with(style, "\\pos")) {
244 boost::algorithm::split (bits, style, boost::is_any_of("(,"));
245 SUB_ASSERT (bits.size() == 3);
246 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
247 sub.horizontal_position.proportional = raw_convert<float>(bits[1]) / play_res_x;
248 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
249 sub.vertical_position.proportional = raw_convert<float>(bits[2]) / play_res_y;
250 } else if (boost::starts_with(style, "\\fs")) {
251 SUB_ASSERT (style.length() > 3);
252 sub.font_size.set_points (raw_convert<int>(style.substr(3)));
253 } else if (boost::starts_with(style, "\\c")) {
255 if (style.length() <= 2) {
256 throw SSAError(String::compose("Badly formatted colour tag %1", style));
258 sub.colour = h_colour (style.substr(2, style.length() - 3));
262 /** @param base RawSubtitle filled in with any required common values.
263 * @param line SSA line string (i.e. just the subtitle, possibly with embedded stuff)
264 * @return List of RawSubtitles to represent line with vertical reference TOP_OF_SUBTITLE.
267 SSAReader::parse_line (RawSubtitle base, string line, int play_res_x, int play_res_y)
275 vector<RawSubtitle> subs;
276 RawSubtitle current = base;
279 if (!current.vertical_position.reference) {
280 current.vertical_position.reference = BOTTOM_OF_SCREEN;
283 if (!current.vertical_position.proportional) {
284 current.vertical_position.proportional = 0;
287 /* We must have a font size, as there could be a margin specified
288 in pixels and in that case we must know how big the subtitle
289 lines are to work out the position on screen.
291 if (!current.font_size.points()) {
292 current.font_size.set_points (72);
295 /* Count the number of line breaks */
297 if (line.length() > 0) {
298 for (size_t i = 0; i < line.length() - 1; ++i) {
299 if (line[i] == '\\' && (line[i+1] == 'n' || line[i+1] == 'N')) {
305 /* Imagine that the screen is 792 points (i.e. 11 inches) high (as with DCP) */
306 double const line_size = current.font_size.proportional(792) * 1.2;
308 /* Tweak vertical_position accordingly */
309 switch (current.vertical_position.reference.get()) {
311 case TOP_OF_SUBTITLE:
314 case VERTICAL_CENTRE_OF_SCREEN:
315 current.vertical_position.proportional = current.vertical_position.proportional.get() - ((line_breaks + 1) * line_size) / 2;
317 case BOTTOM_OF_SCREEN:
318 current.vertical_position.proportional = current.vertical_position.proportional.get() + line_breaks * line_size;
322 for (size_t i = 0; i < line.length(); ++i) {
323 char const c = line[i];
328 } else if (c == '\\') {
330 } else if (c != '\r' && c != '\n') {
335 if (c == '}' || c == '\\') {
336 if (!current.text.empty ()) {
337 subs.push_back (current);
340 parse_style (current, style, play_res_x, play_res_y);
351 if (c == 'n' || c == 'N') {
352 if (!current.text.empty ()) {
353 subs.push_back (current);
356 /* Move down one line (1.2 times the font size) */
357 if (current.vertical_position.reference.get() == BOTTOM_OF_SCREEN) {
358 current.vertical_position.proportional = current.vertical_position.proportional.get() - line_size;
360 current.vertical_position.proportional = current.vertical_position.proportional.get() + line_size;
368 if (!current.text.empty ()) {
369 subs.push_back (current);
376 SSAReader::read (function<optional<string> ()> get_line)
384 int play_res_x = 288;
385 int play_res_y = 288;
386 map<string, Style> styles;
387 string style_format_line;
388 vector<string> event_format;
391 optional<string> line = get_line ();
397 remove_unicode_bom (line);
399 if (starts_with (*line, ";") || line->empty ()) {
403 if (starts_with (*line, "[")) {
404 /* Section heading */
405 if (line.get() == "[Script Info]") {
407 } else if (line.get() == "[V4 Styles]" || line.get() == "[V4+ Styles]") {
409 } else if (line.get() == "[Events]") {
415 size_t const colon = line->find (":");
416 SUB_ASSERT (colon != string::npos);
417 string const type = line->substr (0, colon);
418 string body = line->substr (colon + 1);
423 if (type == "PlayResX") {
424 play_res_x = raw_convert<int> (body);
425 } else if (type == "PlayResY") {
426 play_res_y = raw_convert<int> (body);
430 if (type == "Format") {
431 style_format_line = body;
432 } else if (type == "Style") {
433 SUB_ASSERT (!style_format_line.empty ());
434 Style s (style_format_line, body);
439 if (type == "Format") {
440 split (event_format, body, is_any_of (","));
441 for (auto& i: event_format) {
444 } else if (type == "Dialogue") {
445 SUB_ASSERT (!event_format.empty ());
446 vector<string> event;
447 split (event, body, is_any_of (","));
449 /* There may be commas in the subtitle part; reassemble any extra parts
450 from when we just split it.
452 while (event.size() > event_format.size()) {
453 string const ex = event.back ();
455 event.back() += "," + ex;
458 SUB_ASSERT (!event.empty());
459 SUB_ASSERT (event_format.size() == event.size());
463 for (size_t i = 0; i < event.size(); ++i) {
465 if (event_format[i] == "Start") {
466 sub.from = parse_time (event[i]);
467 } else if (event_format[i] == "End") {
468 sub.to = parse_time (event[i]);
469 } else if (event_format[i] == "Style") {
470 /* libass trims leading '*'s from style names, commenting that
471 "they seem to mean literally nothing". Go figure...
473 trim_left_if (event[i], boost::is_any_of ("*"));
474 SUB_ASSERT (styles.find(event[i]) != styles.end());
475 Style style = styles[event[i]];
476 sub.font = style.font_name;
477 sub.font_size = FontSize::from_points (style.font_size);
478 sub.colour = style.primary_colour;
479 sub.effect_colour = style.back_colour;
480 sub.bold = style.bold;
481 sub.italic = style.italic;
482 sub.underline = style.underline;
483 sub.effect = style.effect;
484 sub.horizontal_position.reference = style.horizontal_reference;
485 sub.vertical_position.reference = style.vertical_reference;
486 sub.vertical_position.proportional = float(style.vertical_margin) / play_res_y;
487 } else if (event_format[i] == "MarginV") {
488 sub.vertical_position.proportional = raw_convert<float>(event[i]) / play_res_y;
489 } else if (event_format[i] == "Text") {
490 for (auto j: parse_line (sub, event[i], play_res_x, play_res_y)) {