2 * Subtitle reader with format autodetection
4 * Copyright (c) 2001 laaz
5 * Some code cleanup & realloc() by A'rpi/ESP-team
7 * This file is part of MPlayer.
9 * MPlayer is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License as published by
11 * the Free Software Foundation; either version 2 of the License, or
12 * (at your option) any later version.
14 * MPlayer is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU General Public License for more details.
19 * You should have received a copy of the GNU General Public License along
20 * with MPlayer; if not, write to the Free Software Foundation, Inc.,
21 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
29 #include <sys/types.h>
34 #include "subreader.h"
36 #include "stream/stream.h"
37 #include "libavutil/common.h"
38 #include "libavutil/avstring.h"
44 #define ERR ((void *) -1)
51 #include <fribidi/fribidi.h>
52 char *fribidi_charset
= NULL
; ///character set that will be passed to FriBiDi
53 int flip_hebrew
= 1; ///flip subtitles using fribidi
54 int fribidi_flip_commas
= 0; ///flip comma when fribidi is used
57 /* Maximal length of line of a subtitle */
59 static float mpsub_position
=0;
60 static float mpsub_multiplier
=1.;
61 static int sub_slacktime
= 20000; //20 sec
63 int sub_no_text_pp
=0; // 1 => do not apply text post-processing
64 // like {\...} elimination in SSA format.
66 int sub_match_fuzziness
=0; // level of sub name matching fuzziness
68 /* Use the SUB_* constant defined in the header file */
69 int sub_format
=SUB_INVALID
;
72 Some subtitling formats, namely AQT and Subrip09, define the end of a
73 subtitle as the beginning of the following. Since currently we read one
74 subtitle at time, for these format we keep two global *subtitle,
75 previous_aqt_sub and previous_subrip09_sub, pointing to previous subtitle,
76 so we can change its end when we read current subtitle starting time.
77 When CONFIG_SORTSUB is defined, we use a single global unsigned long,
78 previous_sub_end, for both (and even future) formats, to store the end of
79 the previous sub: it is initialized to 0 in sub_read_file and eventually
80 modified by sub_read_aqt_line or sub_read_subrip09_line.
82 unsigned long previous_sub_end
;
85 static int eol(char p
) {
86 return p
=='\r' || p
=='\n' || p
=='\0';
89 /* Remove leading and trailing space */
90 static void trail_space(char *s
) {
92 while (isspace(s
[i
])) ++i
;
93 if (i
) strcpy(s
, s
+ i
);
95 while (i
> 0 && isspace(s
[i
])) s
[i
--] = '\0';
98 static char *stristr(const char *haystack
, const char *needle
) {
100 const char *p
= haystack
;
102 if (!(haystack
&& needle
)) return NULL
;
106 if (strncasecmp(p
, needle
, len
) == 0) return (char*)p
;
113 static void sami_add_line(subtitle
*current
, char *buffer
, char **pos
) {
117 if (*buffer
&& current
->lines
< SUB_MAX_TEXT
)
118 current
->text
[current
->lines
++] = strdup(buffer
);
122 static subtitle
*sub_read_line_sami(stream_t
* st
, subtitle
*current
, int utf16
) {
123 static char line
[LINE_LEN
+1];
124 static char *s
= NULL
, *slacktime_s
;
125 char text
[LINE_LEN
+1], *p
=NULL
, *q
;
128 current
->lines
= current
->start
= current
->end
= 0;
129 current
->alignment
= SUB_ALIGNMENT_BOTTOMCENTER
;
132 /* read the first line */
134 if (!(s
= stream_read_line(st
, line
, LINE_LEN
, utf16
))) return 0;
139 case 0: /* find "START=" or "Slacktime:" */
140 slacktime_s
= stristr (s
, "Slacktime:");
142 sub_slacktime
= strtol (slacktime_s
+10, NULL
, 0) / 10;
144 s
= stristr (s
, "Start=");
146 current
->start
= strtol (s
+ 6, &s
, 0) / 10;
148 for (; *s
!= '>' && *s
!= '\0'; s
++);
154 case 1: /* find (optional) "<P", skip other TAGs */
155 for (; *s
== ' ' || *s
== '\t'; s
++); /* strip blanks, if any */
156 if (*s
== '\0') break;
157 if (*s
!= '<') { state
= 3; p
= text
; continue; } /* not a TAG */
159 if (*s
== 'P' || *s
== 'p') { s
++; state
= 2; continue; } /* found '<P' */
160 for (; *s
!= '>' && *s
!= '\0'; s
++); /* skip remains of non-<P> TAG */
166 case 2: /* find ">" */
167 if ((s
= strchr (s
, '>'))) { s
++; state
= 3; p
= text
; continue; }
170 case 3: /* get all text until '<' appears */
171 if (*s
== '\0') break;
172 else if (!strncasecmp (s
, "<br>", 4)) {
173 sami_add_line(current
, text
, &p
);
176 else if ((*s
== '{') && !sub_no_text_pp
) { state
= 5; ++s
; continue; }
177 else if (*s
== '<') { state
= 4; }
178 else if (!strncasecmp (s
, " ", 6)) { *p
++ = ' '; s
+= 6; }
179 else if (*s
== '\t') { *p
++ = ' '; s
++; }
180 else if (*s
== '\r' || *s
== '\n') { s
++; }
183 /* skip duplicated space */
184 if (p
> text
+ 2) if (*(p
-1) == ' ' && *(p
-2) == ' ') p
--;
188 case 4: /* get current->end or skip <TAG> */
189 q
= stristr (s
, "Start=");
191 current
->end
= strtol (q
+ 6, &q
, 0) / 10 - 1;
192 *p
= '\0'; trail_space (text
);
194 current
->text
[current
->lines
++] = strdup (text
);
195 if (current
->lines
> 0) { state
= 99; break; }
199 if (s
) { s
++; state
= 3; continue; }
201 case 5: /* get rid of {...} text, but read the alignment code */
202 if ((*s
== '\\') && (*(s
+ 1) == 'a') && !sub_no_text_pp
) {
203 if (stristr(s
, "\\a1") != NULL
) {
204 current
->alignment
= SUB_ALIGNMENT_BOTTOMLEFT
;
207 if (stristr(s
, "\\a2") != NULL
) {
208 current
->alignment
= SUB_ALIGNMENT_BOTTOMCENTER
;
210 } else if (stristr(s
, "\\a3") != NULL
) {
211 current
->alignment
= SUB_ALIGNMENT_BOTTOMRIGHT
;
213 } else if ((stristr(s
, "\\a4") != NULL
) || (stristr(s
, "\\a5") != NULL
) || (stristr(s
, "\\a8") != NULL
)) {
214 current
->alignment
= SUB_ALIGNMENT_TOPLEFT
;
216 } else if (stristr(s
, "\\a6") != NULL
) {
217 current
->alignment
= SUB_ALIGNMENT_TOPCENTER
;
219 } else if (stristr(s
, "\\a7") != NULL
) {
220 current
->alignment
= SUB_ALIGNMENT_TOPRIGHT
;
222 } else if (stristr(s
, "\\a9") != NULL
) {
223 current
->alignment
= SUB_ALIGNMENT_MIDDLELEFT
;
225 } else if (stristr(s
, "\\a10") != NULL
) {
226 current
->alignment
= SUB_ALIGNMENT_MIDDLECENTER
;
228 } else if (stristr(s
, "\\a11") != NULL
) {
229 current
->alignment
= SUB_ALIGNMENT_MIDDLERIGHT
;
233 if (*s
== '}') state
= 3;
239 if (state
!= 99 && !(s
= stream_read_line (st
, line
, LINE_LEN
, utf16
))) {
240 if (current
->start
> 0) {
241 break; // if it is the last subtitle
247 } while (state
!= 99);
249 // For the last subtitle
250 if (current
->end
<= 0) {
251 current
->end
= current
->start
+ sub_slacktime
;
252 sami_add_line(current
, text
, &p
);
259 static char *sub_readtext(char *source
, char **dest
) {
263 // printf("src=%p dest=%p \n",source,dest);
265 while ( !eol(*p
) && *p
!= '|' ) {
269 *dest
= malloc (len
+1);
270 if (!dest
) {return ERR
;}
272 strncpy(*dest
, source
, len
);
275 while (*p
=='\r' || *p
=='\n' || *p
=='|') p
++;
277 if (*p
) return p
; // not-last text field
278 else return NULL
; // last text field
281 static subtitle
*sub_read_line_microdvd(stream_t
*st
,subtitle
*current
, int utf16
) {
282 char line
[LINE_LEN
+1];
283 char line2
[LINE_LEN
+1];
288 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
289 } while ((sscanf (line
,
291 &(current
->start
), line2
) < 2) &&
293 "{%ld}{%ld}%[^\r\n]",
294 &(current
->start
), &(current
->end
), line2
) < 3));
299 while ((next
=sub_readtext (next
, &(current
->text
[i
])))) {
300 if (current
->text
[i
]==ERR
) {return ERR
;}
302 if (i
>=SUB_MAX_TEXT
) { mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"Too many lines in a subtitle\n");current
->lines
=i
;return current
;}
309 static subtitle
*sub_read_line_mpl2(stream_t
*st
,subtitle
*current
, int utf16
) {
310 char line
[LINE_LEN
+1];
311 char line2
[LINE_LEN
+1];
316 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
317 } while ((sscanf (line
,
318 "[%ld][%ld]%[^\r\n]",
319 &(current
->start
), &(current
->end
), line2
) < 3));
320 current
->start
*= 10;
325 while ((next
=sub_readtext (next
, &(current
->text
[i
])))) {
326 if (current
->text
[i
]==ERR
) {return ERR
;}
328 if (i
>=SUB_MAX_TEXT
) { mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"Too many lines in a subtitle\n");current
->lines
=i
;return current
;}
335 static subtitle
*sub_read_line_subrip(stream_t
* st
, subtitle
*current
, int utf16
) {
336 char line
[LINE_LEN
+1];
337 int a1
,a2
,a3
,a4
,b1
,b2
,b3
,b4
;
338 char *p
=NULL
, *q
=NULL
;
342 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
343 if (sscanf (line
, "%d:%d:%d.%d,%d:%d:%d.%d",&a1
,&a2
,&a3
,&a4
,&b1
,&b2
,&b3
,&b4
) < 8) continue;
344 current
->start
= a1
*360000+a2
*6000+a3
*100+a4
;
345 current
->end
= b1
*360000+b2
*6000+b3
*100+b4
;
347 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
350 for (current
->lines
=1; current
->lines
< SUB_MAX_TEXT
; current
->lines
++) {
351 for (q
=p
,len
=0; *p
&& *p
!='\r' && *p
!='\n' && *p
!='|' && strncmp(p
,"[br]",4); p
++,len
++);
352 current
->text
[current
->lines
-1]=malloc (len
+1);
353 if (!current
->text
[current
->lines
-1]) return ERR
;
354 strncpy (current
->text
[current
->lines
-1], q
, len
);
355 current
->text
[current
->lines
-1][len
]='\0';
356 if (!*p
|| *p
=='\r' || *p
=='\n') break;
358 else while (*p
++!=']');
365 static subtitle
*sub_read_line_subviewer(stream_t
*st
,subtitle
*current
, int utf16
) {
366 char line
[LINE_LEN
+1];
367 int a1
,a2
,a3
,a4
,b1
,b2
,b3
,b4
;
371 while (!current
->text
[0]) {
372 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
373 if ((len
=sscanf (line
, "%d:%d:%d%[,.:]%d --> %d:%d:%d%[,.:]%d",&a1
,&a2
,&a3
,(char *)&i
,&a4
,&b1
,&b2
,&b3
,(char *)&i
,&b4
)) < 10)
375 current
->start
= a1
*360000+a2
*6000+a3
*100+a4
/10;
376 current
->end
= b1
*360000+b2
*6000+b3
*100+b4
/10;
377 for (i
=0; i
<SUB_MAX_TEXT
;) {
379 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) break;
381 for (p
=line
; *p
!='\n' && *p
!='\r' && *p
; p
++,len
++)
382 if (*p
!= ' ' && *p
!= '\t')
386 char *curptr
=current
->text
[i
]=malloc (len
+1);
387 if (!current
->text
[i
]) return ERR
;
388 //strncpy (current->text[i], line, len); current->text[i][len]='\0';
390 /* let's filter html tags ::atmos */
417 static subtitle
*sub_read_line_subviewer2(stream_t
*st
,subtitle
*current
, int utf16
) {
418 char line
[LINE_LEN
+1];
423 while (!current
->text
[0]) {
424 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
427 if ((len
=sscanf (line
, "{T %d:%d:%d:%d",&a1
,&a2
,&a3
,&a4
)) < 4)
429 current
->start
= a1
*360000+a2
*6000+a3
*100+a4
/10;
430 for (i
=0; i
<SUB_MAX_TEXT
;) {
431 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) break;
432 if (line
[0]=='}') break;
434 for (p
=line
; *p
!='\n' && *p
!='\r' && *p
; ++p
,++len
);
436 current
->text
[i
]=malloc (len
+1);
437 if (!current
->text
[i
]) return ERR
;
438 strncpy (current
->text
[i
], line
, len
); current
->text
[i
][len
]='\0';
450 static subtitle
*sub_read_line_vplayer(stream_t
*st
,subtitle
*current
, int utf16
) {
451 char line
[LINE_LEN
+1];
453 char *p
=NULL
, *next
,separator
;
456 while (!current
->text
[0]) {
457 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
458 if ((len
=sscanf (line
, "%d:%d:%d%c%n",&a1
,&a2
,&a3
,&separator
,&plen
)) < 4)
461 if (!(current
->start
= a1
*360000+a2
*6000+a3
*100))
465 // finds the body of the subtitle
472 printf("SUB: Skipping incorrect subtitle line!\n");
476 // by wodzu: hey! this time we know what length it has! what is
477 // that magic for? it can't deal with space instead of third
478 // colon! look, what simple it can be:
485 while ((next
=sub_readtext (next
, &(current
->text
[i
])))) {
486 if (current
->text
[i
]==ERR
) {return ERR
;}
488 if (i
>=SUB_MAX_TEXT
) { mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"Too many lines in a subtitle\n");current
->lines
=i
;return current
;}
496 static subtitle
*sub_read_line_rt(stream_t
*st
,subtitle
*current
, int utf16
) {
497 //TODO: This format uses quite rich (sub/super)set of xhtml
498 // I couldn't check it since DTD is not included.
499 // WARNING: full XML parses can be required for proper parsing
500 char line
[LINE_LEN
+1];
501 int a1
,a2
,a3
,a4
,b1
,b2
,b3
,b4
;
502 char *p
=NULL
,*next
=NULL
;
505 while (!current
->text
[0]) {
506 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
507 //TODO: it seems that format of time is not easily determined, it may be 1:12, 1:12.0 or 0:1:12.0
508 //to describe the same moment in time. Maybe there are even more formats in use.
509 //if ((len=sscanf (line, "<Time Begin=\"%d:%d:%d.%d\" End=\"%d:%d:%d.%d\"",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4)) < 8)
510 plen
=a1
=a2
=a3
=a4
=b1
=b2
=b3
=b4
=0;
512 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d.%d\"%*[^<]<clear/>%n",&a3
,&a4
,&b3
,&b4
,&plen
)) < 4) &&
513 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a3
,&a4
,&b2
,&b3
,&b4
,&plen
)) < 5) &&
514 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2
,&a3
,&b2
,&b3
,&plen
)) < 4) &&
515 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2
,&a3
,&b2
,&b3
,&b4
,&plen
)) < 5) &&
516 // ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&b2,&b3,&plen)) < 5) &&
517 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2
,&a3
,&a4
,&b2
,&b3
,&b4
,&plen
)) < 6) &&
518 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d:%d:%d.%d\" %*[Ee]nd=\"%d:%d:%d.%d\"%*[^<]<clear/>%n",&a1
,&a2
,&a3
,&a4
,&b1
,&b2
,&b3
,&b4
,&plen
)) < 8) &&
519 //now try it without end time
520 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d.%d\"%*[^<]<clear/>%n",&a3
,&a4
,&plen
)) < 2) &&
521 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d:%d\"%*[^<]<clear/>%n",&a2
,&a3
,&plen
)) < 2) &&
522 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2
,&a3
,&a4
,&plen
)) < 3) &&
523 ((len
=sscanf (line
, "<%*[tT]ime %*[bB]egin=\"%d:%d:%d.%d\"%*[^<]<clear/>%n",&a1
,&a2
,&a3
,&a4
,&plen
)) < 4)
526 current
->start
= a1
*360000+a2
*6000+a3
*100+a4
/10;
527 current
->end
= b1
*360000+b2
*6000+b3
*100+b4
/10;
528 if (b1
== 0 && b2
== 0 && b3
== 0 && b4
== 0)
529 current
->end
= current
->start
+200;
531 // TODO: I don't know what kind of convention is here for marking multiline subs, maybe <br/> like in xml?
532 next
= strstr(line
,"<clear/>");
533 if(next
&& strlen(next
)>8){
535 while ((next
=sub_readtext (next
, &(current
->text
[i
])))) {
536 if (current
->text
[i
]==ERR
) {return ERR
;}
538 if (i
>=SUB_MAX_TEXT
) { mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"Too many lines in a subtitle\n");current
->lines
=i
;return current
;}
546 static subtitle
*sub_read_line_ssa(stream_t
*st
,subtitle
*current
, int utf16
) {
548 * Sub Station Alpha v4 (and v2?) scripts have 9 commas before subtitle
549 * other Sub Station Alpha scripts have only 8 commas before subtitle
550 * Reading the "ScriptType:" field is not reliable since many scripts appear
553 * http://www.scriptclub.org is a good place to find more examples
554 * http://www.eswat.demon.co.uk is where the SSA specs can be found
557 static int max_comma
= 32; /* let's use 32 for the case that the */
558 /* amount of commas increase with newer SSA versions */
560 int hour1
, min1
, sec1
, hunsec1
,
561 hour2
, min2
, sec2
, hunsec2
, nothing
;
564 char line
[LINE_LEN
+1],
570 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) return NULL
;
571 } while (sscanf (line
, "Dialogue: Marked=%d,%d:%d:%d.%d,%d:%d:%d.%d"
572 "%[^\n\r]", ¬hing
,
573 &hour1
, &min1
, &sec1
, &hunsec1
,
574 &hour2
, &min2
, &sec2
, &hunsec2
,
577 sscanf (line
, "Dialogue: %d,%d:%d:%d.%d,%d:%d:%d.%d"
578 "%[^\n\r]", ¬hing
,
579 &hour1
, &min1
, &sec1
, &hunsec1
,
580 &hour2
, &min2
, &sec2
, &hunsec2
,
583 line2
=strchr(line3
, ',');
584 if (!line2
) return NULL
;
586 for (comma
= 4; comma
< max_comma
; comma
++)
589 if(!(tmp
=strchr(++tmp
, ','))) break;
590 if(*(++tmp
) == ' ') break;
591 /* a space after a comma means we're already in a sentence */
595 if(comma
< max_comma
)max_comma
= comma
;
596 /* eliminate the trailing comma */
597 if(*line2
== ',') line2
++;
599 current
->lines
=0;num
=0;
600 current
->start
= 360000*hour1
+ 6000*min1
+ 100*sec1
+ hunsec1
;
601 current
->end
= 360000*hour2
+ 6000*min2
+ 100*sec2
+ hunsec2
;
603 while (((tmp
=strstr(line2
, "\\n")) != NULL
) || ((tmp
=strstr(line2
, "\\N")) != NULL
) ){
604 current
->text
[num
]=malloc(tmp
-line2
+1);
605 strncpy (current
->text
[num
], line2
, tmp
-line2
);
606 current
->text
[num
][tmp
-line2
]='\0';
610 if (current
->lines
>= SUB_MAX_TEXT
) return current
;
613 current
->text
[num
]=strdup(line2
);
619 static void sub_pp_ssa(subtitle
*sub
) {
624 /* eliminate any text enclosed with {}, they are font and color settings */
625 so
=de
=sub
->text
[--l
];
627 if(*so
== '{' && so
[1]=='\\') {
628 for (start
=so
; *so
&& *so
!='}'; so
++);
629 if(*so
) so
++; else so
=start
;
641 * PJS subtitles reader.
642 * That's the "Phoenix Japanimation Society" format.
643 * I found some of them in http://www.scriptsclub.org/ (used for anime).
644 * The time is in tenths of second.
646 * by set, based on code by szabi (dunnowhat sub format ;-)
648 static subtitle
*sub_read_line_pjs(stream_t
*st
,subtitle
*current
, int utf16
) {
649 char line
[LINE_LEN
+1];
650 char text
[LINE_LEN
+1], *s
, *d
;
652 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
))
655 for (s
=line
; *s
&& isspace(*s
); s
++);
656 /* allow empty lines at the end of the file */
660 if (sscanf (s
, "%ld,%ld,", &(current
->start
),
661 &(current
->end
)) <2) {
664 /* the files I have are in tenths of second */
665 current
->start
*= 10;
667 /* walk to the beggining of the string */
668 for (; *s
; s
++) if (*s
==',') break;
670 for (s
++; *s
; s
++) if (*s
==',') break;
676 /* copy the string to the text buffer */
677 for (s
++, d
=text
; *s
&& *s
!='"'; s
++, d
++)
680 current
->text
[0] = strdup(text
);
686 static subtitle
*sub_read_line_mpsub(stream_t
*st
, subtitle
*current
, int utf16
) {
687 char line
[LINE_LEN
+1];
694 if (!stream_read_line(st
, line
, LINE_LEN
, utf16
)) return NULL
;
695 } while (sscanf (line
, "%f %f", &a
, &b
) !=2);
697 mpsub_position
+= a
*mpsub_multiplier
;
698 current
->start
=(int) mpsub_position
;
699 mpsub_position
+= b
*mpsub_multiplier
;
700 current
->end
=(int) mpsub_position
;
702 while (num
< SUB_MAX_TEXT
) {
703 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
)) {
704 if (num
== 0) return NULL
;
708 while (isspace(*p
)) p
++;
709 if (eol(*p
) && num
> 0) return current
;
710 if (eol(*p
)) return NULL
;
712 for (q
=p
; !eol(*q
); q
++);
715 current
->text
[num
]=strdup(p
);
716 // printf (">%s<\n",p);
717 current
->lines
= ++num
;
719 if (num
) return current
;
723 return NULL
; // we should have returned before if it's OK
726 #ifndef CONFIG_SORTSUB
727 //we don't need this if we use previous_sub_end
728 subtitle
*previous_aqt_sub
= NULL
;
731 static subtitle
*sub_read_line_aqt(stream_t
*st
,subtitle
*current
, int utf16
) {
732 char line
[LINE_LEN
+1];
737 // try to locate next subtitle
738 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
))
740 if (!(sscanf (line
, "-->> %ld", &(current
->start
)) <1))
744 #ifdef CONFIG_SORTSUB
745 previous_sub_end
= (current
->start
) ? current
->start
- 1 : 0;
747 if (previous_aqt_sub
!= NULL
)
748 previous_aqt_sub
->end
= current
->start
-1;
750 previous_aqt_sub
= current
;
753 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
))
756 sub_readtext((char *) &line
,¤t
->text
[0]);
758 current
->end
= current
->start
; // will be corrected by next subtitle
760 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
))
764 while ((next
=sub_readtext (next
, &(current
->text
[i
])))) {
765 if (current
->text
[i
]==ERR
) {return ERR
;}
767 if (i
>=SUB_MAX_TEXT
) { mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"Too many lines in a subtitle\n");current
->lines
=i
;return current
;}
771 if (!strlen(current
->text
[0]) && !strlen(current
->text
[1])) {
772 #ifdef CONFIG_SORTSUB
773 previous_sub_end
= 0;
775 // void subtitle -> end of previous marked and exit
776 previous_aqt_sub
= NULL
;
784 #ifndef CONFIG_SORTSUB
785 subtitle
*previous_subrip09_sub
= NULL
;
788 static subtitle
*sub_read_line_subrip09(stream_t
*st
,subtitle
*current
, int utf16
) {
789 char line
[LINE_LEN
+1];
795 // try to locate next subtitle
796 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
))
798 if (!((len
=sscanf (line
, "[%d:%d:%d]",&a1
,&a2
,&a3
)) < 3))
802 current
->start
= a1
*360000+a2
*6000+a3
*100;
804 #ifdef CONFIG_SORTSUB
805 previous_sub_end
= (current
->start
) ? current
->start
- 1 : 0;
807 if (previous_subrip09_sub
!= NULL
)
808 previous_subrip09_sub
->end
= current
->start
-1;
810 previous_subrip09_sub
= current
;
813 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
))
818 current
->text
[0]=""; // just to be sure that string is clear
820 while ((next
=sub_readtext (next
, &(current
->text
[i
])))) {
821 if (current
->text
[i
]==ERR
) {return ERR
;}
823 if (i
>=SUB_MAX_TEXT
) { mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"Too many lines in a subtitle\n");current
->lines
=i
;return current
;}
827 if (!strlen(current
->text
[0]) && (i
==0)) {
828 #ifdef CONFIG_SORTSUB
829 previous_sub_end
= 0;
831 // void subtitle -> end of previous marked and exit
832 previous_subrip09_sub
= NULL
;
840 static subtitle
*sub_read_line_jacosub(stream_t
* st
, subtitle
* current
, int utf16
)
842 char line1
[LINE_LEN
], line2
[LINE_LEN
], directive
[LINE_LEN
], *p
, *q
;
843 unsigned a1
, a2
, a3
, a4
, b1
, b2
, b3
, b4
, comment
= 0;
844 static unsigned jacoTimeres
= 30;
845 static int jacoShift
= 0;
847 memset(current
, 0, sizeof(subtitle
));
848 memset(line1
, 0, LINE_LEN
);
849 memset(line2
, 0, LINE_LEN
);
850 memset(directive
, 0, LINE_LEN
);
851 while (!current
->text
[0]) {
852 if (!stream_read_line(st
, line1
, LINE_LEN
, utf16
)) {
856 (line1
, "%u:%u:%u.%u %u:%u:%u.%u %[^\n\r]", &a1
, &a2
, &a3
, &a4
,
857 &b1
, &b2
, &b3
, &b4
, line2
) < 9) {
858 if (sscanf(line1
, "@%u @%u %[^\n\r]", &a4
, &b4
, line2
) < 3) {
859 if (line1
[0] == '#') {
860 int hours
= 0, minutes
= 0, seconds
, delta
, inverter
=
862 unsigned units
= jacoShift
;
863 switch (toupper(line1
[1])) {
865 if (isalpha(line1
[2])) {
870 if (sscanf(&line1
[delta
], "%d", &hours
)) {
875 if (sscanf(&line1
[delta
], "%*d:%d", &minutes
)) {
877 (&line1
[delta
], "%*d:%*d:%d",
879 sscanf(&line1
[delta
], "%*d:%*d:%*d.%d",
883 sscanf(&line1
[delta
], "%d:%d.%d",
884 &minutes
, &seconds
, &units
);
889 sscanf(&line1
[delta
], "%d.%d", &seconds
,
894 ((hours
* 3600 + minutes
* 60 +
895 seconds
) * jacoTimeres
+
900 if (isalpha(line1
[2])) {
905 sscanf(&line1
[delta
], "%u", &jacoTimeres
);
912 (unsigned long) ((a4
+ jacoShift
) * 100.0 /
915 (unsigned long) ((b4
+ jacoShift
) * 100.0 /
921 long) (((a1
* 3600 + a2
* 60 + a3
) * jacoTimeres
+ a4
+
922 jacoShift
) * 100.0 / jacoTimeres
);
925 long) (((b1
* 3600 + b2
* 60 + b3
) * jacoTimeres
+ b4
+
926 jacoShift
) * 100.0 / jacoTimeres
);
930 while ((*p
== ' ') || (*p
== '\t')) {
933 if (isalpha(*p
)||*p
== '[') {
936 if (sscanf(p
, "%s %[^\n\r]", directive
, line1
) < 2)
937 return (subtitle
*) ERR
;
938 jLength
= strlen(directive
);
939 for (cont
= 0; cont
< jLength
; ++cont
) {
940 if (isalpha(*(directive
+ cont
)))
941 *(directive
+ cont
) = toupper(*(directive
+ cont
));
943 if ((strstr(directive
, "RDB") != NULL
)
944 || (strstr(directive
, "RDC") != NULL
)
945 || (strstr(directive
, "RLB") != NULL
)
946 || (strstr(directive
, "RLG") != NULL
)) {
949 if (strstr(directive
, "JL") != NULL
) {
950 current
->alignment
= SUB_ALIGNMENT_BOTTOMLEFT
;
951 } else if (strstr(directive
, "JR") != NULL
) {
952 current
->alignment
= SUB_ALIGNMENT_BOTTOMRIGHT
;
954 current
->alignment
= SUB_ALIGNMENT_BOTTOMCENTER
;
956 strcpy(line2
, line1
);
959 for (q
= line1
; (!eol(*p
)) && (current
->lines
< SUB_MAX_TEXT
); ++p
) {
967 //the next line to get rid of a blank after the comment
968 if ((*(p
+ 1)) == ' ')
980 if ((*(p
+ 1) == ' ') || (*(p
+ 1) == '\t'))
988 if (*(p
+ 1) == 'n') {
991 current
->text
[current
->lines
++] = strdup(line1
);
995 if ((toupper(*(p
+ 1)) == 'C')
996 || (toupper(*(p
+ 1)) == 'F')) {
1000 if ((*(p
+ 1) == 'B') || (*(p
+ 1) == 'b') || (*(p
+ 1) == 'D') || //actually this means "insert current date here"
1001 (*(p
+ 1) == 'I') || (*(p
+ 1) == 'i') || (*(p
+ 1) == 'N') || (*(p
+ 1) == 'T') || //actually this means "insert current time here"
1002 (*(p
+ 1) == 'U') || (*(p
+ 1) == 'u')) {
1006 if ((*(p
+ 1) == '\\') ||
1007 (*(p
+ 1) == '~') || (*(p
+ 1) == '{')) {
1009 } else if (eol(*(p
+ 1))) {
1010 if (!stream_read_line(st
, directive
, LINE_LEN
, utf16
))
1012 trail_space(directive
);
1013 av_strlcat(line2
, directive
, LINE_LEN
);
1024 current
->text
[current
->lines
] = strdup(line1
);
1030 static int sub_autodetect (stream_t
* st
, int *uses_time
, int utf16
) {
1031 char line
[LINE_LEN
+1];
1036 if (!stream_read_line (st
, line
, LINE_LEN
, utf16
))
1039 if (sscanf (line
, "{%d}{%d}", &i
, &i
)==2)
1040 {*uses_time
=0;return SUB_MICRODVD
;}
1041 if (sscanf (line
, "{%d}{}", &i
)==1)
1042 {*uses_time
=0;return SUB_MICRODVD
;}
1043 if (sscanf (line
, "[%d][%d]", &i
, &i
)==2)
1044 {*uses_time
=1;return SUB_MPL2
;}
1045 if (sscanf (line
, "%d:%d:%d.%d,%d:%d:%d.%d", &i
, &i
, &i
, &i
, &i
, &i
, &i
, &i
)==8)
1046 {*uses_time
=1;return SUB_SUBRIP
;}
1047 if (sscanf (line
, "%d:%d:%d%[,.:]%d --> %d:%d:%d%[,.:]%d", &i
, &i
, &i
, (char *)&i
, &i
, &i
, &i
, &i
, (char *)&i
, &i
)==10)
1048 {*uses_time
=1;return SUB_SUBVIEWER
;}
1049 if (sscanf (line
, "{T %d:%d:%d:%d",&i
, &i
, &i
, &i
)==4)
1050 {*uses_time
=1;return SUB_SUBVIEWER2
;}
1051 if (strstr (line
, "<SAMI>"))
1052 {*uses_time
=1; return SUB_SAMI
;}
1053 if (sscanf(line
, "%d:%d:%d.%d %d:%d:%d.%d", &i
, &i
, &i
, &i
, &i
, &i
, &i
, &i
) == 8)
1054 {*uses_time
= 1; return SUB_JACOSUB
;}
1055 if (sscanf(line
, "@%d @%d", &i
, &i
) == 2)
1056 {*uses_time
= 1; return SUB_JACOSUB
;}
1057 if (sscanf (line
, "%d:%d:%d:", &i
, &i
, &i
)==3)
1058 {*uses_time
=1;return SUB_VPLAYER
;}
1059 if (sscanf (line
, "%d:%d:%d ", &i
, &i
, &i
)==3)
1060 {*uses_time
=1;return SUB_VPLAYER
;}
1061 if (!strncasecmp(line
, "<window", 7))
1062 {*uses_time
=1;return SUB_RT
;}
1063 if (!memcmp(line
, "Dialogue: Marked", 16))
1064 {*uses_time
=1; return SUB_SSA
;}
1065 if (!memcmp(line
, "Dialogue: ", 10))
1066 {*uses_time
=1; return SUB_SSA
;}
1067 if (sscanf (line
, "%d,%d,\"%c", &i
, &i
, (char *) &i
) == 3)
1068 {*uses_time
=1;return SUB_PJS
;}
1069 if (sscanf (line
, "FORMAT=%d", &i
) == 1)
1070 {*uses_time
=0; return SUB_MPSUB
;}
1071 if (!memcmp(line
, "FORMAT=TIME", 11))
1072 {*uses_time
=1; return SUB_MPSUB
;}
1073 if (strstr (line
, "-->>"))
1074 {*uses_time
=0; return SUB_AQTITLE
;}
1075 if (sscanf (line
, "[%d:%d:%d]", &i
, &i
, &i
)==3)
1076 {*uses_time
=1;return SUB_SUBRIP09
;}
1079 return SUB_INVALID
; // too many bad lines
1082 extern int sub_utf8
;
1083 int sub_utf8_prev
=0;
1085 extern float sub_delay
;
1086 extern float sub_fps
;
1089 static iconv_t icdsc
= (iconv_t
)(-1);
1091 void subcp_open (stream_t
*st
)
1093 char *tocp
= "UTF-8";
1096 const char *cp_tmp
= sub_cp
;
1098 char enca_lang
[3], enca_fallback
[100];
1099 if (sscanf(sub_cp
, "enca:%2s:%99s", enca_lang
, enca_fallback
) == 2
1100 || sscanf(sub_cp
, "ENCA:%2s:%99s", enca_lang
, enca_fallback
) == 2) {
1101 if (st
&& st
->flags
& MP_STREAM_SEEK
) {
1102 cp_tmp
= guess_cp(st
, enca_lang
, enca_fallback
);
1104 cp_tmp
= enca_fallback
;
1106 mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"SUB: enca failed, stream must be seekable.\n");
1110 if ((icdsc
= iconv_open (tocp
, cp_tmp
)) != (iconv_t
)(-1)){
1111 mp_msg(MSGT_SUBREADER
,MSGL_V
,"SUB: opened iconv descriptor.\n");
1114 mp_msg(MSGT_SUBREADER
,MSGL_ERR
,"SUB: error opening iconv descriptor.\n");
1118 void subcp_close (void)
1120 if (icdsc
!= (iconv_t
)(-1)){
1121 (void) iconv_close (icdsc
);
1122 icdsc
= (iconv_t
)(-1);
1123 mp_msg(MSGT_SUBREADER
,MSGL_V
,"SUB: closed iconv descriptor.\n");
1127 subtitle
* subcp_recode (subtitle
*sub
)
1130 size_t ileft
, oleft
;
1132 if(icdsc
== (iconv_t
)(-1)) return sub
;
1135 ip
= sub
->text
[--l
];
1139 if (!(ot
= malloc(oleft
+ 1))){
1140 mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"SUB: error allocating mem.\n");
1144 if (iconv(icdsc
, &ip
, &ileft
,
1145 &op
, &oleft
) == (size_t)(-1)) {
1146 mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"SUB: error recoding line.\n");
1150 // In some stateful encodings, we must clear the state to handle the last character
1151 if (iconv(icdsc
, NULL
, NULL
,
1152 &op
, &oleft
) == (size_t)(-1)) {
1153 mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"SUB: error recoding line, can't clear encoding state.\n");
1156 free (sub
->text
[l
]);
1163 #ifdef CONFIG_FRIBIDI
1165 * Do conversion necessary for right-to-left language support via fribidi.
1166 * @param sub subtitle to convert
1167 * @param sub_utf8 whether the subtitle is encoded in UTF-8
1168 * @param from first new subtitle, all lines before this are assumed to be already converted
1170 static subtitle
* sub_fribidi (subtitle
*sub
, int sub_utf8
, int from
)
1172 FriBidiChar logical
[LINE_LEN
+1], visual
[LINE_LEN
+1]; // Hopefully these two won't smash the stack
1173 char *ip
= NULL
, *op
= NULL
;
1174 size_t len
,orig_len
;
1177 fribidi_boolean log2vis
;
1180 fribidi_set_mirroring(1);
1181 fribidi_set_reorder_nsm(0);
1183 if( sub_utf8
== 0 ) {
1184 char_set_num
= fribidi_parse_charset (fribidi_charset
?fribidi_charset
:"ISO8859-8");
1186 char_set_num
= fribidi_parse_charset ("UTF-8");
1189 ip
= sub
->text
[--l
];
1190 orig_len
= len
= strlen( ip
); // We assume that we don't use full unicode, only UTF-8 or ISO8859-x
1191 if(len
> LINE_LEN
) {
1192 mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"SUB: sub->text is longer than LINE_LEN.\n");
1196 len
= fribidi_charset_to_unicode (char_set_num
, ip
, len
, logical
);
1197 #if FRIBIDI_INTERFACE_VERSION < 3
1198 FriBidiCharType base
= fribidi_flip_commas
?FRIBIDI_TYPE_ON
:FRIBIDI_TYPE_L
;
1200 FriBidiParType base
= fribidi_flip_commas
?FRIBIDI_TYPE_ON
:FRIBIDI_TYPE_L
;
1202 log2vis
= fribidi_log2vis (logical
, len
, &base
,
1204 visual
, NULL
, NULL
, NULL
);
1206 len
= fribidi_remove_bidi_marks (visual
, len
, NULL
, NULL
,
1208 if((op
= malloc((FFMAX(2*orig_len
,2*len
) + 1))) == NULL
) {
1209 mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"SUB: error allocating mem.\n");
1213 fribidi_unicode_to_charset ( char_set_num
, visual
, len
,op
);
1219 for (l
= sub
->lines
; l
;)
1220 free (sub
->text
[--l
]);
1228 static void adjust_subs_time(subtitle
* sub
, float subtime
, float fps
, int block
,
1229 int sub_num
, int sub_uses_time
) {
1233 unsigned long subfms
= (sub_uses_time
? 100 : fps
) * subtime
;
1234 unsigned long overlap
= (sub_uses_time
? 100 : fps
) / 5; // 0.2s
1238 if (sub
->end
<= sub
->start
){
1239 sub
->end
= sub
->start
+ subfms
;
1246 if ((sub
->end
> nextsub
->start
) && (sub
->end
<= nextsub
->start
+ overlap
)) {
1247 // these subtitles overlap for less than 0.2 seconds
1248 // and would result in very short overlapping subtitle
1249 // so let's fix the problem here, before overlapping code
1250 // get its hands on them
1251 unsigned delta
= sub
->end
- nextsub
->start
, half
= delta
/ 2;
1252 sub
->end
-= half
+ 1;
1253 nextsub
->start
+= delta
- half
;
1255 if (sub
->end
>= nextsub
->start
){
1256 sub
->end
= nextsub
->start
- 1;
1257 if (sub
->end
- sub
->start
> subfms
)
1258 sub
->end
= sub
->start
+ subfms
;
1265 * Movies are often converted from FILM (24 fps)
1266 * to PAL (25) by simply speeding it up, so we
1267 * to multiply the original timestmaps by
1268 * (Movie's FPS / Subtitle's (guessed) FPS)
1269 * so eg. for 23.98 fps movie and PAL time based
1270 * subtitles we say -subfps 25 and we're fine!
1273 /* timed sub fps correction ::atmos */
1274 /* the frame-based case is handled in mpcommon.c
1275 * where find_sub is called */
1276 if(sub_uses_time
&& sub_fps
) {
1277 sub
->start
*= sub_fps
/fps
;
1278 sub
->end
*= sub_fps
/fps
;
1284 if (n
) mp_msg(MSGT_SUBREADER
,MSGL_V
,"SUB: Adjusted %d subtitle(s).\n", n
);
1288 subtitle
* (*read
)(stream_t
*st
,subtitle
*dest
,int utf16
);
1289 void (*post
)(subtitle
*dest
);
1294 const char* guess_buffer_cp(unsigned char* buffer
, int buflen
, const char *preferred_language
, const char *fallback
)
1296 const char **languages
;
1298 EncaAnalyser analyser
;
1299 EncaEncoding encoding
;
1300 const char *detected_sub_cp
= NULL
;
1303 languages
= enca_get_languages(&langcnt
);
1304 mp_msg(MSGT_SUBREADER
, MSGL_V
, "ENCA supported languages: ");
1305 for (i
= 0; i
< langcnt
; i
++) {
1306 mp_msg(MSGT_SUBREADER
, MSGL_V
, "%s ", languages
[i
]);
1308 mp_msg(MSGT_SUBREADER
, MSGL_V
, "\n");
1310 for (i
= 0; i
< langcnt
; i
++) {
1311 if (strcasecmp(languages
[i
], preferred_language
) != 0) continue;
1312 analyser
= enca_analyser_alloc(languages
[i
]);
1313 encoding
= enca_analyse_const(analyser
, buffer
, buflen
);
1314 enca_analyser_free(analyser
);
1315 if (encoding
.charset
!= ENCA_CS_UNKNOWN
) {
1316 detected_sub_cp
= enca_charset_name(encoding
.charset
, ENCA_NAME_STYLE_ICONV
);
1323 if (!detected_sub_cp
) {
1324 detected_sub_cp
= fallback
;
1325 mp_msg(MSGT_SUBREADER
, MSGL_INFO
, "ENCA detection failed: fallback to %s\n", fallback
);
1327 mp_msg(MSGT_SUBREADER
, MSGL_INFO
, "ENCA detected charset: %s\n", detected_sub_cp
);
1330 return detected_sub_cp
;
1333 #define MAX_GUESS_BUFFER_SIZE (256*1024)
1334 const char* guess_cp(stream_t
*st
, const char *preferred_language
, const char *fallback
)
1337 unsigned char *buffer
;
1338 const char *detected_sub_cp
= NULL
;
1340 buffer
= malloc(MAX_GUESS_BUFFER_SIZE
);
1341 buflen
= stream_read(st
,buffer
, MAX_GUESS_BUFFER_SIZE
);
1343 detected_sub_cp
= guess_buffer_cp(buffer
, buflen
, preferred_language
, fallback
);
1349 return detected_sub_cp
;
1351 #undef MAX_GUESS_BUFFER_SIZE
1354 sub_data
* sub_read_file (char *filename
, float fps
) {
1357 int n_max
, n_first
, i
, j
, sub_first
, sub_orig
;
1358 subtitle
*first
, *second
, *sub
, *return_sub
, *alloced_sub
= NULL
;
1359 sub_data
*subt_data
;
1360 int uses_time
= 0, sub_num
= 0, sub_errs
= 0;
1361 static const struct subreader sr
[]=
1363 { sub_read_line_microdvd
, NULL
, "microdvd" },
1364 { sub_read_line_subrip
, NULL
, "subrip" },
1365 { sub_read_line_subviewer
, NULL
, "subviewer" },
1366 { sub_read_line_sami
, NULL
, "sami" },
1367 { sub_read_line_vplayer
, NULL
, "vplayer" },
1368 { sub_read_line_rt
, NULL
, "rt" },
1369 { sub_read_line_ssa
, sub_pp_ssa
, "ssa" },
1370 { sub_read_line_pjs
, NULL
, "pjs" },
1371 { sub_read_line_mpsub
, NULL
, "mpsub" },
1372 { sub_read_line_aqt
, NULL
, "aqt" },
1373 { sub_read_line_subviewer2
, NULL
, "subviewer 2.0" },
1374 { sub_read_line_subrip09
, NULL
, "subrip 0.9" },
1375 { sub_read_line_jacosub
, NULL
, "jacosub" },
1376 { sub_read_line_mpl2
, NULL
, "mpl2" }
1378 const struct subreader
*srp
;
1380 if(filename
==NULL
) return NULL
; //qnx segfault
1381 fd
=open_stream (filename
, NULL
, NULL
); if (!fd
) return NULL
;
1383 sub_format
= SUB_INVALID
;
1384 for (utf16
= 0; sub_format
== SUB_INVALID
&& utf16
< 3; utf16
++) {
1385 sub_format
=sub_autodetect (fd
, &uses_time
, utf16
);
1391 mpsub_multiplier
= (uses_time
? 100.0 : 1.0);
1392 if (sub_format
==SUB_INVALID
) {mp_msg(MSGT_SUBREADER
,MSGL_WARN
,"SUB: Could not determine file format\n");return NULL
;}
1394 mp_msg(MSGT_SUBREADER
, MSGL_V
, "SUB: Detected subtitle file format: %s\n", srp
->name
);
1397 sub_utf8_prev
=sub_utf8
;
1401 if ((l
=strlen(filename
))>4){
1402 char *exts
[] = {".utf", ".utf8", ".utf-8" };
1404 if (l
>= strlen(exts
[k
]) && !strcasecmp(filename
+(l
- strlen(exts
[k
])), exts
[k
])){
1409 if (k
<0) subcp_open(fd
);
1414 first
=malloc(n_max
*sizeof(subtitle
));
1418 sub_utf8
=sub_utf8_prev
;
1423 #ifdef CONFIG_SORTSUB
1425 sub
= malloc(sizeof(subtitle
));
1426 //This is to deal with those formats (AQT & Subrip) which define the end of a subtitle
1427 //as the beginning of the following
1428 previous_sub_end
= 0;
1433 first
=realloc(first
,n_max
*sizeof(subtitle
));
1435 #ifndef CONFIG_SORTSUB
1436 sub
= &first
[sub_num
];
1438 memset(sub
, '\0', sizeof(subtitle
));
1439 sub
=srp
->read(fd
,sub
,utf16
);
1440 if(!sub
) break; // EOF
1442 if ((sub
!=ERR
) && sub_utf8
== 2) sub
=subcp_recode(sub
);
1444 #ifdef CONFIG_FRIBIDI
1445 if (sub
!=ERR
) sub
=sub_fribidi(sub
,sub_utf8
,0);
1456 // Apply any post processing that needs recoding first
1457 if ((sub
!=ERR
) && !sub_no_text_pp
&& srp
->post
) srp
->post(sub
);
1458 #ifdef CONFIG_SORTSUB
1459 if(!sub_num
|| (first
[sub_num
- 1].start
<= sub
->start
)){
1460 first
[sub_num
].start
= sub
->start
;
1461 first
[sub_num
].end
= sub
->end
;
1462 first
[sub_num
].lines
= sub
->lines
;
1463 first
[sub_num
].alignment
= sub
->alignment
;
1464 for(i
= 0; i
< sub
->lines
; ++i
){
1465 first
[sub_num
].text
[i
] = sub
->text
[i
];
1467 if (previous_sub_end
){
1468 first
[sub_num
- 1].end
= previous_sub_end
;
1469 previous_sub_end
= 0;
1472 for(j
= sub_num
- 1; j
>= 0; --j
){
1473 first
[j
+ 1].start
= first
[j
].start
;
1474 first
[j
+ 1].end
= first
[j
].end
;
1475 first
[j
+ 1].lines
= first
[j
].lines
;
1476 first
[j
+ 1].alignment
= first
[j
].alignment
;
1477 for(i
= 0; i
< first
[j
].lines
; ++i
){
1478 first
[j
+ 1].text
[i
] = first
[j
].text
[i
];
1480 if(!j
|| (first
[j
- 1].start
<= sub
->start
)){
1481 first
[j
].start
= sub
->start
;
1482 first
[j
].end
= sub
->end
;
1483 first
[j
].lines
= sub
->lines
;
1484 first
[j
].alignment
= sub
->alignment
;
1485 for(i
= 0; i
< SUB_MAX_TEXT
; ++i
){
1486 first
[j
].text
[i
] = sub
->text
[i
];
1488 if (previous_sub_end
){
1489 first
[j
].end
= first
[j
- 1].end
;
1490 first
[j
- 1].end
= previous_sub_end
;
1491 previous_sub_end
= 0;
1498 if(sub
==ERR
) ++sub_errs
; else ++sub_num
; // Error vs. Valid
1508 // printf ("SUB: Subtitle format %s time.\n", uses_time?"uses":"doesn't use");
1509 mp_msg(MSGT_SUBREADER
, MSGL_V
,"SUB: Read %i subtitles, %i bad line(s).\n",
1517 // we do overlap if the user forced it (suboverlap_enable == 2) or
1518 // the user didn't forced no-overlapsub and the format is Jacosub or Ssa.
1519 // this is because usually overlapping subtitles are found in these formats,
1520 // while in others they are probably result of bad timing
1521 if ((suboverlap_enabled
== 2) ||
1522 ((suboverlap_enabled
) && ((sub_format
== SUB_JACOSUB
) || (sub_format
== SUB_SSA
)))) {
1523 adjust_subs_time(first
, 6.0, fps
, 0, sub_num
, uses_time
);/*~6 secs AST*/
1524 // here we manage overlapping subtitles
1529 // for each subtitle in first[] we deal with its 'block' of
1531 for (sub_first
= 0; sub_first
< n_first
; ++sub_first
) {
1532 unsigned long global_start
= first
[sub_first
].start
,
1533 global_end
= first
[sub_first
].end
, local_start
, local_end
;
1534 int lines_to_add
= first
[sub_first
].lines
, sub_to_add
= 0,
1535 **placeholder
= NULL
, higher_line
= 0, counter
, start_block_sub
= sub_num
;
1536 char real_block
= 1;
1538 // here we find the number of subtitles inside the 'block'
1539 // and its span interval. this works well only with sorted
1541 while ((sub_first
+ sub_to_add
+ 1 < n_first
) && (first
[sub_first
+ sub_to_add
+ 1].start
< global_end
)) {
1543 lines_to_add
+= first
[sub_first
+ sub_to_add
].lines
;
1544 if (first
[sub_first
+ sub_to_add
].start
< global_start
) {
1545 global_start
= first
[sub_first
+ sub_to_add
].start
;
1547 if (first
[sub_first
+ sub_to_add
].end
> global_end
) {
1548 global_end
= first
[sub_first
+ sub_to_add
].end
;
1552 /* Avoid n^2 memory use for the "placeholder" data structure
1553 * below with subtitles that have a huge number of
1554 * consecutive overlapping lines. */
1555 lines_to_add
= FFMIN(lines_to_add
, SUB_MAX_TEXT
);
1557 // we need a structure to keep trace of the screen lines
1558 // used by the subs, a 'placeholder'
1559 counter
= 2 * sub_to_add
+ 1; // the maximum number of subs derived
1560 // from a block of sub_to_add+1 subs
1561 placeholder
= malloc(sizeof(int *) * counter
);
1562 for (i
= 0; i
< counter
; ++i
) {
1563 placeholder
[i
] = malloc(sizeof(int) * lines_to_add
);
1564 for (j
= 0; j
< lines_to_add
; ++j
) {
1565 placeholder
[i
][j
] = -1;
1570 local_end
= global_start
- 1;
1574 // here we find the beginning and the end of a new
1575 // subtitle in the block
1576 local_start
= local_end
+ 1;
1577 local_end
= global_end
;
1578 for (j
= 0; j
<= sub_to_add
; ++j
) {
1579 if ((first
[sub_first
+ j
].start
- 1 > local_start
) && (first
[sub_first
+ j
].start
- 1 < local_end
)) {
1580 local_end
= first
[sub_first
+ j
].start
- 1;
1581 } else if ((first
[sub_first
+ j
].end
> local_start
) && (first
[sub_first
+ j
].end
< local_end
)) {
1582 local_end
= first
[sub_first
+ j
].end
;
1585 // here we allocate the screen lines to subs we must
1586 // display in current local_start-local_end interval.
1587 // if the subs were yet presents in the previous interval
1588 // they keep the same lines, otherside they get unused lines
1589 for (j
= 0; j
<= sub_to_add
; ++j
) {
1590 if ((first
[sub_first
+ j
].start
<= local_end
) && (first
[sub_first
+ j
].end
> local_start
)) {
1591 unsigned long sub_lines
= first
[sub_first
+ j
].lines
, fragment_length
= lines_to_add
+ 1,
1594 int fragment_position
= -1;
1596 // if this is not the first new sub of the block
1597 // we find if this sub was present in the previous
1600 for (i
= 0; i
< lines_to_add
; ++i
) {
1601 if (placeholder
[counter
- 1][i
] == sub_first
+ j
) {
1602 placeholder
[counter
][i
] = sub_first
+ j
;
1609 // we are looking for the shortest among all groups of
1610 // sequential blank lines whose length is greater than or
1611 // equal to sub_lines. we store in fragment_position the
1612 // position of the shortest group, in fragment_length its
1613 // length, and in tmp the length of the group currently
1615 for (i
= 0; i
< lines_to_add
; ++i
) {
1616 if (placeholder
[counter
][i
] == -1) {
1617 // placeholder[counter][i] is part of the current group
1621 if (tmp
== sub_lines
) {
1622 // current group's size fits exactly the one we
1623 // need, so we stop looking
1624 fragment_position
= i
- tmp
;
1628 if ((tmp
) && (tmp
> sub_lines
) && (tmp
< fragment_length
)) {
1629 // current group is the best we found till here,
1630 // but is still bigger than the one we are looking
1631 // for, so we keep on looking
1632 fragment_length
= tmp
;
1633 fragment_position
= i
- tmp
;
1636 // current group doesn't fit at all, so we forget it
1642 // last screen line is blank, a group ends with it
1643 if ((tmp
>= sub_lines
) && (tmp
< fragment_length
)) {
1644 fragment_position
= i
- tmp
;
1647 if (fragment_position
== -1) {
1648 // it was not possible to find free screen line(s) for a subtitle,
1649 // usually this means a bug in the code; however we do not overlap
1650 mp_msg(MSGT_SUBREADER
, MSGL_WARN
, "SUB: we could not find a suitable position for an overlapping subtitle\n");
1651 higher_line
= SUB_MAX_TEXT
+ 1;
1654 for (tmp
= 0; tmp
< sub_lines
; ++tmp
) {
1655 placeholder
[counter
][fragment_position
+ tmp
] = sub_first
+ j
;
1660 for (j
= higher_line
+ 1; j
< lines_to_add
; ++j
) {
1661 if (placeholder
[counter
][j
] != -1)
1666 if (higher_line
>= SUB_MAX_TEXT
) {
1667 // the 'block' has too much lines, so we don't overlap the
1669 second
= realloc(second
, (sub_num
+ sub_to_add
+ 1) * sizeof(subtitle
));
1670 for (j
= 0; j
<= sub_to_add
; ++j
) {
1672 memset(&second
[sub_num
+ j
], '\0', sizeof(subtitle
));
1673 second
[sub_num
+ j
].start
= first
[sub_first
+ j
].start
;
1674 second
[sub_num
+ j
].end
= first
[sub_first
+ j
].end
;
1675 second
[sub_num
+ j
].lines
= first
[sub_first
+ j
].lines
;
1676 second
[sub_num
+ j
].alignment
= first
[sub_first
+ j
].alignment
;
1677 for (ls
= 0; ls
< second
[sub_num
+ j
].lines
; ls
++) {
1678 second
[sub_num
+ j
].text
[ls
] = strdup(first
[sub_first
+ j
].text
[ls
]);
1681 sub_num
+= sub_to_add
+ 1;
1682 sub_first
+= sub_to_add
;
1687 // we read the placeholder structure and create the new
1689 second
= realloc(second
, (sub_num
+ 1) * sizeof(subtitle
));
1690 memset(&second
[sub_num
], '\0', sizeof(subtitle
));
1691 second
[sub_num
].start
= local_start
;
1692 second
[sub_num
].end
= local_end
;
1693 second
[sub_num
].alignment
= first
[sub_first
].alignment
;
1694 n_max
= (lines_to_add
< SUB_MAX_TEXT
) ? lines_to_add
: SUB_MAX_TEXT
;
1695 for (i
= 0, j
= 0; j
< n_max
; ++j
) {
1696 if (placeholder
[counter
][j
] != -1) {
1697 int lines
= first
[placeholder
[counter
][j
]].lines
;
1698 for (ls
= 0; ls
< lines
; ++ls
) {
1699 second
[sub_num
].text
[i
++] = strdup(first
[placeholder
[counter
][j
]].text
[ls
]);
1703 second
[sub_num
].text
[i
++] = strdup(" ");
1708 } while (local_end
< global_end
);
1710 for (i
= 0; i
< counter
; ++i
)
1711 second
[start_block_sub
+ i
].lines
= higher_line
+ 1;
1713 counter
= 2 * sub_to_add
+ 1;
1714 for (i
= 0; i
< counter
; ++i
) {
1715 free(placeholder
[i
]);
1718 sub_first
+= sub_to_add
;
1721 for (j
= sub_orig
- 1; j
>= 0; --j
) {
1722 for (i
= first
[j
].lines
- 1; i
>= 0; --i
) {
1723 free(first
[j
].text
[i
]);
1728 return_sub
= second
;
1729 } else { //if(suboverlap_enabled)
1730 adjust_subs_time(first
, 6.0, fps
, 1, sub_num
, uses_time
);/*~6 secs AST*/
1733 if (return_sub
== NULL
) return NULL
;
1734 subt_data
= malloc(sizeof(sub_data
));
1735 subt_data
->filename
= strdup(filename
);
1736 subt_data
->sub_uses_time
= uses_time
;
1737 subt_data
->sub_num
= sub_num
;
1738 subt_data
->sub_errs
= sub_errs
;
1739 subt_data
->subtitles
= return_sub
;
1744 char * strreplace( char * in
,char * what
,char * whereof
)
1749 if ( ( in
== NULL
)||( what
== NULL
)||( whereof
== NULL
)||( ( tmp
=strstr( in
,what
) ) == NULL
) ) return NULL
;
1750 for( i
=0;i
<strlen( whereof
);i
++ ) tmp
[i
]=whereof
[i
];
1751 if ( strlen( what
) > strlen( whereof
) ) tmp
[i
]=0;
1757 static void strcpy_trim(char *d
, char *s
)
1759 // skip leading whitespace
1760 while (*s
&& isspace(*s
)) {
1765 while (*s
&& !isspace(*s
)) {
1770 // trim excess whitespace
1771 while (*s
&& isspace(*s
)) {
1780 static void strcpy_strip_ext(char *d
, char *s
)
1782 char *tmp
= strrchr(s
,'.');
1787 strncpy(d
, s
, tmp
-s
);
1796 static void strcpy_get_ext(char *d
, char *s
)
1798 char *tmp
= strrchr(s
,'.');
1807 static int whiteonly(char *s
)
1810 if (!isspace(*s
)) return 0;
1816 typedef struct subfn
1822 static int compare_sub_priority(const void *a
, const void *b
)
1824 if (((const subfn
*)a
)->priority
> ((const subfn
*)b
)->priority
) {
1826 } else if (((const subfn
*)a
)->priority
< ((const subfn
*)b
)->priority
) {
1829 return strcoll(((const subfn
*)a
)->fname
, ((const subfn
*)b
)->fname
);
1833 char** sub_filenames(const char* path
, char *fname
)
1835 char *f_dir
, *f_fname
, *f_fname_noext
, *f_fname_trim
, *tmp
, *tmp_sub_id
;
1836 char *tmp_fname_noext
, *tmp_fname_trim
, *tmp_fname_ext
, *tmpresult
;
1838 int len
, pos
, found
, i
, j
;
1839 char * sub_exts
[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "rt", "txt", "ssa", "aqt", "jss", "js", "ass", NULL
};
1850 len
= (strlen(fname
) > 256 ? strlen(fname
) : 256)
1851 +(strlen(path
) > 256 ? strlen(path
) : 256)+2;
1853 f_dir
= malloc(len
);
1854 f_fname
= malloc(len
);
1855 f_fname_noext
= malloc(len
);
1856 f_fname_trim
= malloc(len
);
1858 tmp_fname_noext
= malloc(len
);
1859 tmp_fname_trim
= malloc(len
);
1860 tmp_fname_ext
= malloc(len
);
1862 tmpresult
= malloc(len
);
1864 result
= calloc(MAX_SUBTITLE_FILES
, sizeof(*result
));
1868 tmp
= strrchr(fname
,'/');
1870 if(!tmp
)tmp
= strrchr(fname
,'\\');
1871 if(!tmp
)tmp
= strrchr(fname
,':');
1874 // extract filename & dirname from fname
1876 strcpy(f_fname
, tmp
+1);
1878 strncpy(f_dir
, fname
, pos
+1);
1881 strcpy(f_fname
, fname
);
1882 strcpy(f_dir
, "./");
1885 strcpy_strip_ext(f_fname_noext
, f_fname
);
1886 strcpy_trim(f_fname_trim
, f_fname_noext
);
1888 /* The code using sub language here is broken - it assumes strict
1889 * "videoname languagename" syntax for the subtitle file, which is
1890 * very unlikely to match especially if language name uses "en,de"
1894 if (dvdsub_lang
&& !whiteonly(dvdsub_lang
)) {
1895 tmp_sub_id
= malloc(strlen(dvdsub_lang
)+1);
1896 strcpy_trim(tmp_sub_id
, dvdsub_lang
);
1901 // 1 = any subtitle file
1902 // 2 = any sub file containing movie name
1903 // 3 = sub file containing movie name and the lang extension
1904 for (j
= 0; j
<= 1; j
++) {
1905 d
= opendir(j
== 0 ? f_dir
: path
);
1907 while ((de
= readdir(d
))) {
1908 // retrieve various parts of the filename
1909 strcpy_strip_ext(tmp_fname_noext
, de
->d_name
);
1910 strcpy_get_ext(tmp_fname_ext
, de
->d_name
);
1911 strcpy_trim(tmp_fname_trim
, tmp_fname_noext
);
1913 // does it end with a subtitle extension?
1917 for (i
= ((sub_cp
&& strncasecmp(sub_cp
, "enca", 4) != 0) ? 3 : 0); sub_exts
[i
]; i
++) {
1919 for (i
= (sub_cp
? 3 : 0); sub_exts
[i
]; i
++) {
1922 for (i
= 0; sub_exts
[i
]; i
++) {
1924 if (strcasecmp(sub_exts
[i
], tmp_fname_ext
) == 0) {
1930 // we have a (likely) subtitle file
1933 if (!prio
&& tmp_sub_id
)
1935 sprintf(tmpresult
, "%s %s", f_fname_trim
, tmp_sub_id
);
1936 if (strcmp(tmp_fname_trim
, tmpresult
) == 0 && sub_match_fuzziness
>= 1) {
1937 // matches the movie name + lang extension
1941 if (!prio
&& strcmp(tmp_fname_trim
, f_fname_trim
) == 0) {
1942 // matches the movie name
1945 if (!prio
&& (tmp
= strstr(tmp_fname_trim
, f_fname_trim
)) && (sub_match_fuzziness
>= 1)) {
1946 // contains the movie name
1947 tmp
+= strlen(f_fname_trim
);
1948 if (tmp_sub_id
&& strstr(tmp
, tmp_sub_id
)) {
1949 // with sub_id specified prefer localized subtitles
1951 } else if ((tmp_sub_id
== NULL
) && whiteonly(tmp
)) {
1952 // without sub_id prefer "plain" name
1955 // with no localized subs found, try any else instead
1960 // doesn't contain the movie name
1961 // don't try in the mplayer subtitle directory
1962 if ((j
== 0) && (sub_match_fuzziness
>= 2)) {
1967 mp_msg(MSGT_SUBREADER
, MSGL_DBG2
, "Potential sub file: "
1968 "\"%s\" Priority: %d\n", de
->d_name
, prio
);
1972 if (i
<3){ // prefer UTF-8 coded
1976 sprintf(tmpresult
, "%s%s", j
== 0 ? f_dir
: path
, de
->d_name
);
1977 // fprintf(stderr, "%s priority %d\n", tmpresult, prio);
1978 if ((f
= fopen(tmpresult
, "rt"))) {
1980 result
[subcnt
].priority
= prio
;
1981 result
[subcnt
].fname
= strdup(tmpresult
);
1987 if (subcnt
>= MAX_SUBTITLE_FILES
) break;
1998 free(f_fname_noext
);
2001 free(tmp_fname_noext
);
2002 free(tmp_fname_trim
);
2003 free(tmp_fname_ext
);
2007 qsort(result
, subcnt
, sizeof(subfn
), compare_sub_priority
);
2009 result2
= calloc(subcnt
+ 1, sizeof(*result2
));
2011 for (i
= 0; i
< subcnt
; i
++) {
2012 result2
[i
] = result
[i
].fname
;
2014 result2
[subcnt
] = NULL
;
2021 void list_sub_file(sub_data
* subd
){
2023 subtitle
*subs
= subd
->subtitles
;
2025 for(j
=0; j
< subd
->sub_num
; j
++){
2026 subtitle
* egysub
=&subs
[j
];
2027 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"%i line%c (%li-%li)\n",
2029 (1==egysub
->lines
)?' ':'s',
2032 for (i
=0; i
<egysub
->lines
; i
++) {
2033 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"\t\t%d: %s%s", i
,egysub
->text
[i
], i
==egysub
->lines
-1?"":" \n ");
2035 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"\n");
2038 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"Subtitle format %s time.\n",
2039 subd
->sub_uses_time
? "uses":"doesn't use");
2040 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"Read %i subtitles, %i errors.\n", subd
->sub_num
, subd
->sub_errs
);
2043 void dump_srt(sub_data
* subd
, float fps
){
2049 subtitle
*subs
= subd
->subtitles
;
2051 if (!subd
->sub_uses_time
&& sub_fps
== 0)
2053 fd
=fopen("dumpsub.srt","w");
2056 perror("dump_srt: fopen");
2059 for(i
=0; i
< subd
->sub_num
; i
++)
2061 onesub
=subs
+i
; //=&subs[i];
2062 fprintf(fd
,"%d\n",i
+1);//line number
2065 if (!subd
->sub_uses_time
)
2066 temp
= temp
* 100 / sub_fps
;
2067 temp
-= sub_delay
* 100;
2068 h
=temp
/360000;temp
%=360000; //h =1*100*60*60
2069 m
=temp
/6000; temp
%=6000; //m =1*100*60
2070 s
=temp
/100; temp
%=100; //s =1*100
2071 ms
=temp
*10; //ms=1*10
2072 fprintf(fd
,"%02d:%02d:%02d,%03d --> ",h
,m
,s
,ms
);
2075 if (!subd
->sub_uses_time
)
2076 temp
= temp
* 100 / sub_fps
;
2077 temp
-= sub_delay
* 100;
2078 h
=temp
/360000;temp
%=360000;
2079 m
=temp
/6000; temp
%=6000;
2080 s
=temp
/100; temp
%=100;
2082 fprintf(fd
,"%02d:%02d:%02d,%03d\n",h
,m
,s
,ms
);
2084 for(j
=0;j
<onesub
->lines
;j
++)
2085 fprintf(fd
,"%s\n",onesub
->text
[j
]);
2090 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"SUB: Subtitles dumped in \'dumpsub.srt\'.\n");
2093 void dump_mpsub(sub_data
* subd
, float fps
){
2097 subtitle
*subs
= subd
->subtitles
;
2099 mpsub_position
= subd
->sub_uses_time
? (sub_delay
*100) : (sub_delay
*fps
);
2100 if (sub_fps
==0) sub_fps
=fps
;
2102 fd
=fopen ("dump.mpsub", "w");
2104 perror ("dump_mpsub: fopen");
2109 if (subd
->sub_uses_time
) fprintf (fd
,"FORMAT=TIME\n\n");
2110 else fprintf (fd
, "FORMAT=%5.2f\n\n", fps
);
2112 for(j
=0; j
< subd
->sub_num
; j
++){
2113 subtitle
* egysub
=&subs
[j
];
2114 if (subd
->sub_uses_time
) {
2115 a
=((egysub
->start
-mpsub_position
)/100.0);
2116 b
=((egysub
->end
-egysub
->start
)/100.0);
2117 if ( (float)((int)a
) == a
)
2118 fprintf (fd
, "%.0f",a
);
2120 fprintf (fd
, "%.2f",a
);
2122 if ( (float)((int)b
) == b
)
2123 fprintf (fd
, " %.0f\n",b
);
2125 fprintf (fd
, " %.2f\n",b
);
2127 fprintf (fd
, "%ld %ld\n", (long)((egysub
->start
*(fps
/sub_fps
))-((mpsub_position
*(fps
/sub_fps
)))),
2128 (long)(((egysub
->end
)-(egysub
->start
))*(fps
/sub_fps
)));
2131 mpsub_position
= egysub
->end
;
2132 for (i
=0; i
<egysub
->lines
; i
++) {
2133 fprintf (fd
, "%s\n",egysub
->text
[i
]);
2138 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"SUB: Subtitles dumped in \'dump.mpsub\'.\n");
2141 void dump_microdvd(sub_data
* subd
, float fps
) {
2144 subtitle
*subs
= subd
->subtitles
;
2147 fd
= fopen("dumpsub.sub", "w");
2149 perror("dumpsub.sub: fopen");
2152 delay
= sub_delay
* sub_fps
;
2153 for (i
= 0; i
< subd
->sub_num
; ++i
) {
2155 start
= subs
[i
].start
;
2157 if (subd
->sub_uses_time
) {
2158 start
= start
* sub_fps
/ 100 ;
2159 end
= end
* sub_fps
/ 100;
2162 start
= start
* sub_fps
/ fps
;
2163 end
= end
* sub_fps
/ fps
;
2167 fprintf(fd
, "{%d}{%d}", start
, end
);
2168 for (j
= 0; j
< subs
[i
].lines
; ++j
)
2169 fprintf(fd
, "%s%s", j
? "|" : "", subs
[i
].text
[j
]);
2173 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"SUB: Subtitles dumped in \'dumpsub.sub\'.\n");
2176 void dump_jacosub(sub_data
* subd
, float fps
) {
2182 subtitle
*subs
= subd
->subtitles
;
2184 if (!subd
->sub_uses_time
&& sub_fps
== 0)
2186 fd
=fopen("dumpsub.jss","w");
2189 perror("dump_jacosub: fopen");
2192 fprintf(fd
, "#TIMERES %d\n", (subd
->sub_uses_time
) ? 100 : (int)sub_fps
);
2193 for(i
=0; i
< subd
->sub_num
; i
++)
2195 onesub
=subs
+i
; //=&subs[i];
2198 if (!subd
->sub_uses_time
)
2199 temp
= temp
* 100 / sub_fps
;
2200 temp
-= sub_delay
* 100;
2201 h
=temp
/360000;temp
%=360000; //h =1*100*60*60
2202 m
=temp
/6000; temp
%=6000; //m =1*100*60
2203 s
=temp
/100; temp
%=100; //s =1*100
2205 fprintf(fd
,"%02d:%02d:%02d.%02d ",h
,m
,s
,cs
);
2208 if (!subd
->sub_uses_time
)
2209 temp
= temp
* 100 / sub_fps
;
2210 temp
-= sub_delay
* 100;
2211 h
=temp
/360000;temp
%=360000;
2212 m
=temp
/6000; temp
%=6000;
2213 s
=temp
/100; temp
%=100;
2215 fprintf(fd
,"%02d:%02d:%02d.%02d {~} ",h
,m
,s
,cs
);
2217 for(j
=0;j
<onesub
->lines
;j
++)
2218 fprintf(fd
,"%s%s",j
? "\\n" : "", onesub
->text
[j
]);
2223 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"SUB: Subtitles dumped in \'dumpsub.js\'.\n");
2226 void dump_sami(sub_data
* subd
, float fps
) {
2231 subtitle
*subs
= subd
->subtitles
;
2233 if (!subd
->sub_uses_time
&& sub_fps
== 0)
2235 fd
=fopen("dumpsub.smi","w");
2238 perror("dump_jacosub: fopen");
2241 fprintf(fd
, "<SAMI>\n"
2243 " <STYLE TYPE=\"Text/css\">\n"
2245 " P {margin-left: 29pt; margin-right: 29pt; font-size: 24pt; text-align: center; font-family: Tahoma; font-weight: bold; color: #FCDD03; background-color: #000000;}\n"
2246 " .SUBTTL {Name: 'Subtitles'; Lang: en-US; SAMIType: CC;}\n"
2251 for(i
=0; i
< subd
->sub_num
; i
++)
2253 onesub
=subs
+i
; //=&subs[i];
2256 if (!subd
->sub_uses_time
)
2257 temp
= temp
* 100 / sub_fps
;
2258 temp
-= sub_delay
* 100;
2259 fprintf(fd
,"\t<SYNC Start=%lu>\n"
2260 "\t <P>", temp
* 10);
2262 for(j
=0;j
<onesub
->lines
;j
++)
2263 fprintf(fd
,"%s%s",j
? "<br>" : "", onesub
->text
[j
]);
2268 if (!subd
->sub_uses_time
)
2269 temp
= temp
* 100 / sub_fps
;
2270 temp
-= sub_delay
* 100;
2271 fprintf(fd
,"\t<SYNC Start=%lu>\n"
2272 "\t <P> \n", temp
* 10);
2274 fprintf(fd
, "</BODY>\n"
2277 mp_msg(MSGT_SUBREADER
,MSGL_INFO
,"SUB: Subtitles dumped in \'dumpsub.smi\'.\n");
2280 void sub_free( sub_data
* subd
)
2284 if ( !subd
) return;
2286 for (i
= 0; i
< subd
->sub_num
; i
++)
2287 for (j
= 0; j
< subd
->subtitles
[i
].lines
; j
++)
2288 free( subd
->subtitles
[i
].text
[j
] );
2289 free( subd
->subtitles
);
2290 free( subd
->filename
);
2294 #define MAX_SUBLINE 512
2296 * \brief parse text and append it to subtitle in sub
2297 * \param sub subtitle struct to add text to
2298 * \param txt text to parse
2299 * \param len length of text in txt
2300 * \param endpts pts at which this subtitle text should be removed again
2302 * <> and {} are interpreted as comment delimiters, "\n", "\N", '\n', '\r'
2303 * and '\0' are interpreted as newlines, duplicate, leading and trailing
2304 * newlines are ignored.
2306 void sub_add_text(subtitle
*sub
, const char *txt
, int len
, double endpts
) {
2308 int double_newline
= 1; // ignore newlines at the beginning
2311 #ifdef CONFIG_FRIBIDI
2312 int orig_lines
= sub
->lines
;
2314 if (sub
->lines
>= SUB_MAX_TEXT
) return;
2316 buf
= malloc(MAX_SUBLINE
+ 1);
2317 sub
->text
[sub
->lines
] = buf
;
2318 sub
->endpts
[sub
->lines
] = endpts
;
2319 for (i
= 0; i
< len
&& pos
< MAX_SUBLINE
; i
++) {
2321 if (c
== '<') comment
|= 1;
2322 if (c
== '{') comment
|= 2;
2324 if (c
== '}') comment
&= ~2;
2325 if (c
== '>') comment
&= ~1;
2328 if (pos
== MAX_SUBLINE
- 1) {
2332 if (c
== '\\' && i
+ 1 < len
) {
2334 if (c
== 'n' || c
== 'N') c
= 0;
2336 if (c
== '\n' || c
== '\r') c
= 0;
2340 } else if (!double_newline
) {
2341 if (sub
->lines
>= SUB_MAX_TEXT
- 1) {
2342 mp_msg(MSGT_VO
, MSGL_WARN
, "Too many subtitle lines\n");
2349 buf
= malloc(MAX_SUBLINE
+ 1);
2350 sub
->text
[sub
->lines
] = buf
;
2351 sub
->endpts
[sub
->lines
] = endpts
;
2355 if (sub
->lines
< SUB_MAX_TEXT
&&
2356 strlen(sub
->text
[sub
->lines
]))
2358 #ifdef CONFIG_FRIBIDI
2359 sub
= sub_fribidi(sub
, sub_utf8
, orig_lines
);
2364 * \brief remove outdated subtitle lines.
2365 * \param sub subtitle struct to modify
2366 * \param pts current pts. All lines with endpts <= this will be removed.
2367 * Use MP_NOPTS_VALUE to remove all lines
2368 * \return 1 if sub was modified, 0 otherwise.
2370 int sub_clear_text(subtitle
*sub
, double pts
) {
2373 while (i
< sub
->lines
) {
2374 double endpts
= sub
->endpts
[i
];
2375 if (pts
== MP_NOPTS_VALUE
|| (endpts
!= MP_NOPTS_VALUE
&& pts
>= endpts
)) {
2378 for (j
= i
+ 1; j
< sub
->lines
; j
++) {
2379 sub
->text
[j
- 1] = sub
->text
[j
];
2380 sub
->endpts
[j
- 1] = sub
->endpts
[j
];