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>
28 #include <boost/foreach.hpp>
37 using boost::optional;
38 using boost::function;
39 using namespace boost::algorithm;
42 /** @param s Subtitle string encoded in UTF-8 */
43 SSAReader::SSAReader (string s)
45 this->read (boost::bind(&get_line_string, &s));
48 /** @param f Subtitle file encoded in UTF-8 */
49 SSAReader::SSAReader (FILE* f)
51 this->read (boost::bind (&get_line_file, f));
57 /* There are both BGR and ABGR versions of these colours */
58 if ((s.length() != 8 && s.length() != 10) || s[0] != '&' || s[1] != 'H') {
59 throw SSAError(String::compose("Badly formatted colour tag %1", s));
62 /* XXX: ignoring alpha channel here; note that 00 is opaque and FF is transparent */
63 int const off = s.length() == 10 ? 4 : 2;
64 if (sscanf(s.c_str() + off, "%2x%2x%2x", &ib, &ig, &ir) < 3) {
65 throw SSAError(String::compose("Badly formatted colour tag %1", s));
67 return sub::Colour(ir / 255.0, ig / 255.0, ib / 255.0);
75 , primary_colour (255, 255, 255)
79 , horizontal_reference (HORIZONTAL_CENTRE_OF_SCREEN)
80 , vertical_reference (BOTTOM_OF_SCREEN)
84 Style (string format_line, string style_line)
86 , primary_colour (255, 255, 255)
90 , horizontal_reference (HORIZONTAL_CENTRE_OF_SCREEN)
91 , vertical_reference (BOTTOM_OF_SCREEN)
95 split (keys, format_line, boost::is_any_of (","));
97 split (style, style_line, boost::is_any_of (","));
99 SUB_ASSERT (!keys.empty());
100 SUB_ASSERT (!style.empty());
101 SUB_ASSERT (keys.size() == style.size());
103 for (size_t i = 0; i < style.size(); ++i) {
106 if (keys[i] == "Name") {
108 } else if (keys[i] == "Fontname") {
109 font_name = style[i];
110 } else if (keys[i] == "Fontsize") {
111 font_size = raw_convert<int> (style[i]);
112 } else if (keys[i] == "PrimaryColour") {
113 primary_colour = colour (style[i]);
114 } else if (keys[i] == "BackColour") {
115 back_colour = colour (style[i]);
116 } else if (keys[i] == "Bold") {
117 bold = style[i] == "-1";
118 } else if (keys[i] == "Italic") {
119 italic = style[i] == "-1";
120 } else if (keys[i] == "Underline") {
121 underline = style[i] == "-1";
122 } else if (keys[i] == "BorderStyle") {
123 if (style[i] == "1") {
126 } else if (keys[i] == "Alignment") {
127 /* These values from libass' source code */
128 switch ((raw_convert<int> (style[i]) - 1) % 3) {
130 horizontal_reference = LEFT_OF_SCREEN;
133 horizontal_reference = HORIZONTAL_CENTRE_OF_SCREEN;
136 horizontal_reference = RIGHT_OF_SCREEN;
139 switch (raw_convert<int> (style[i]) & 12) {
141 vertical_reference = TOP_OF_SCREEN;
144 vertical_reference = VERTICAL_CENTRE_OF_SCREEN;
147 vertical_reference = BOTTOM_OF_SCREEN;
150 } else if (keys[i] == "MarginV") {
151 vertical_margin = raw_convert<int> (style[i]);
157 optional<string> font_name;
159 Colour primary_colour;
160 /** outline colour */
161 optional<Colour> back_colour;
165 optional<Effect> effect;
166 HorizontalReference horizontal_reference;
167 VerticalReference vertical_reference;
171 Colour colour (string c) const
173 if (c.length() > 0 && c[0] == '&') {
174 /* &Hbbggrr or &Haabbggrr */
178 int i = raw_convert<int>(c);
180 ((i & 0x0000ff) >> 0) / 255.0,
181 ((i & 0x00ff00) >> 8) / 255.0,
182 ((i & 0xff0000) >> 16) / 255.0
189 SSAReader::parse_time (string t) const
192 split (bits, t, is_any_of (":."));
193 SUB_ASSERT (bits.size() == 4);
194 return Time::from_hms (
195 raw_convert<int> (bits[0]),
196 raw_convert<int> (bits[1]),
197 raw_convert<int> (bits[2]),
198 raw_convert<int> (bits[3]) * 10
203 SSAReader::parse_style (RawSubtitle& sub, string style, int play_res_x, int play_res_y)
205 if (style == "\\i1") {
207 } else if (style == "\\i0" || style == "\\i") {
209 } else if (style == "\\b1") {
211 } else if (style == "\\b0") {
213 } else if (style == "\\u1") {
214 sub.underline = true;
215 } else if (style == "\\u0") {
216 sub.underline = false;
217 } else if (style == "\\an1") {
218 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
219 sub.vertical_position.reference = sub::BOTTOM_OF_SCREEN;
220 } else if (style == "\\an2") {
221 sub.horizontal_position.reference = sub::HORIZONTAL_CENTRE_OF_SCREEN;
222 sub.vertical_position.reference = sub::BOTTOM_OF_SCREEN;
223 } else if (style == "\\an3") {
224 sub.horizontal_position.reference = sub::RIGHT_OF_SCREEN;
225 sub.vertical_position.reference = sub::BOTTOM_OF_SCREEN;
226 } else if (style == "\\an4") {
227 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
228 sub.vertical_position.reference = sub::VERTICAL_CENTRE_OF_SCREEN;
229 } else if (style == "\\an5") {
230 sub.horizontal_position.reference = sub::HORIZONTAL_CENTRE_OF_SCREEN;
231 sub.vertical_position.reference = sub::VERTICAL_CENTRE_OF_SCREEN;
232 } else if (style == "\\an6") {
233 sub.horizontal_position.reference = sub::RIGHT_OF_SCREEN;
234 sub.vertical_position.reference = sub::VERTICAL_CENTRE_OF_SCREEN;
235 } else if (style == "\\an7") {
236 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
237 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
238 } else if (style == "\\an8") {
239 sub.horizontal_position.reference = sub::HORIZONTAL_CENTRE_OF_SCREEN;
240 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
241 } else if (style == "\\an9") {
242 sub.horizontal_position.reference = sub::RIGHT_OF_SCREEN;
243 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
244 } else if (boost::starts_with(style, "\\pos")) {
246 boost::algorithm::split (bits, style, boost::is_any_of("(,"));
247 SUB_ASSERT (bits.size() == 3);
248 sub.horizontal_position.reference = sub::LEFT_OF_SCREEN;
249 sub.horizontal_position.proportional = raw_convert<float>(bits[1]) / play_res_x;
250 sub.vertical_position.reference = sub::TOP_OF_SCREEN;
251 sub.vertical_position.proportional = raw_convert<float>(bits[2]) / play_res_y;
252 } else if (boost::starts_with(style, "\\fs")) {
253 SUB_ASSERT (style.length() > 3);
254 sub.font_size.set_points (raw_convert<int>(style.substr(3)));
255 } else if (boost::starts_with(style, "\\c")) {
257 if (style.length() <= 2) {
258 throw SSAError(String::compose("Badly formatted colour tag %1", style));
260 sub.colour = h_colour (style.substr(2, style.length() - 3));
264 /** @param base RawSubtitle filled in with any required common values.
265 * @param line SSA line string (i.e. just the subtitle, possibly with embedded stuff)
266 * @return List of RawSubtitles to represent line with vertical reference TOP_OF_SUBTITLE.
269 SSAReader::parse_line (RawSubtitle base, string line, int play_res_x, int play_res_y)
277 list<RawSubtitle> subs;
278 RawSubtitle current = base;
281 if (!current.vertical_position.reference) {
282 current.vertical_position.reference = BOTTOM_OF_SCREEN;
285 if (!current.vertical_position.proportional) {
286 current.vertical_position.proportional = 0;
289 /* We must have a font size, as there could be a margin specified
290 in pixels and in that case we must know how big the subtitle
291 lines are to work out the position on screen.
293 if (!current.font_size.points()) {
294 current.font_size.set_points (72);
297 /* Count the number of line breaks */
299 if (line.length() > 0) {
300 for (size_t i = 0; i < line.length() - 1; ++i) {
301 if (line[i] == '\\' && (line[i+1] == 'n' || line[i+1] == 'N')) {
307 /* Imagine that the screen is 792 points (i.e. 11 inches) high (as with DCP) */
308 double const line_size = current.font_size.proportional(792) * 1.2;
310 /* Tweak vertical_position accordingly */
311 switch (current.vertical_position.reference.get()) {
313 case TOP_OF_SUBTITLE:
316 case VERTICAL_CENTRE_OF_SCREEN:
317 current.vertical_position.proportional = current.vertical_position.proportional.get() - ((line_breaks + 1) * line_size) / 2;
319 case BOTTOM_OF_SCREEN:
320 current.vertical_position.proportional = current.vertical_position.proportional.get() + line_breaks * line_size;
324 for (size_t i = 0; i < line.length(); ++i) {
325 char const c = line[i];
330 } else if (c == '\\') {
332 } else if (c != '\r' && c != '\n') {
337 if (c == '}' || c == '\\') {
338 if (!current.text.empty ()) {
339 subs.push_back (current);
342 parse_style (current, style, play_res_x, play_res_y);
353 if (c == 'n' || c == 'N') {
354 if (!current.text.empty ()) {
355 subs.push_back (current);
358 /* Move down one line (1.2 times the font size) */
359 if (current.vertical_position.reference.get() == BOTTOM_OF_SCREEN) {
360 current.vertical_position.proportional = current.vertical_position.proportional.get() - line_size;
362 current.vertical_position.proportional = current.vertical_position.proportional.get() + line_size;
370 if (!current.text.empty ()) {
371 subs.push_back (current);
378 SSAReader::read (function<optional<string> ()> get_line)
386 int play_res_x = 288;
387 int play_res_y = 288;
388 map<string, Style> styles;
389 string style_format_line;
390 vector<string> event_format;
393 optional<string> line = get_line ();
399 remove_unicode_bom (line);
401 if (starts_with (*line, ";") || line->empty ()) {
405 if (starts_with (*line, "[")) {
406 /* Section heading */
407 if (line.get() == "[Script Info]") {
409 } else if (line.get() == "[V4 Styles]" || line.get() == "[V4+ Styles]") {
411 } else if (line.get() == "[Events]") {
417 size_t const colon = line->find (":");
418 SUB_ASSERT (colon != string::npos);
419 string const type = line->substr (0, colon);
420 string body = line->substr (colon + 1);
425 if (type == "PlayResX") {
426 play_res_x = raw_convert<int> (body);
427 } else if (type == "PlayResY") {
428 play_res_y = raw_convert<int> (body);
432 if (type == "Format") {
433 style_format_line = body;
434 } else if (type == "Style") {
435 SUB_ASSERT (!style_format_line.empty ());
436 Style s (style_format_line, body);
441 if (type == "Format") {
442 split (event_format, body, is_any_of (","));
443 BOOST_FOREACH (string& i, event_format) {
446 } else if (type == "Dialogue") {
447 SUB_ASSERT (!event_format.empty ());
448 vector<string> event;
449 split (event, body, is_any_of (","));
451 /* There may be commas in the subtitle part; reassemble any extra parts
452 from when we just split it.
454 while (event.size() > event_format.size()) {
455 string const ex = event.back ();
457 event.back() += "," + ex;
460 SUB_ASSERT (!event.empty());
461 SUB_ASSERT (event_format.size() == event.size());
465 for (size_t i = 0; i < event.size(); ++i) {
467 if (event_format[i] == "Start") {
468 sub.from = parse_time (event[i]);
469 } else if (event_format[i] == "End") {
470 sub.to = parse_time (event[i]);
471 } else if (event_format[i] == "Style") {
472 /* libass trims leading '*'s from style names, commenting that
473 "they seem to mean literally nothing". Go figure...
475 trim_left_if (event[i], boost::is_any_of ("*"));
476 SUB_ASSERT (styles.find(event[i]) != styles.end());
477 Style style = styles[event[i]];
478 sub.font = style.font_name;
479 sub.font_size = FontSize::from_points (style.font_size);
480 sub.colour = style.primary_colour;
481 sub.effect_colour = style.back_colour;
482 sub.bold = style.bold;
483 sub.italic = style.italic;
484 sub.underline = style.underline;
485 sub.effect = style.effect;
486 sub.horizontal_position.reference = style.horizontal_reference;
487 sub.vertical_position.reference = style.vertical_reference;
488 sub.vertical_position.proportional = float(style.vertical_margin) / play_res_y;
489 } else if (event_format[i] == "MarginV") {
490 sub.vertical_position.proportional = raw_convert<float>(event[i]) / play_res_y;
491 } else if (event_format[i] == "Text") {
492 BOOST_FOREACH (sub::RawSubtitle j, parse_line (sub, event[i], play_res_x, play_res_y)) {