1 /*****************************************************************************
2 * telx.c : Minimalistic Teletext subtitles decoder
3 *****************************************************************************
4 * Copyright (C) 2007 Vincent Penne
5 * Some code converted from ProjectX java dvb decoder (c) 2001-2005 by dvb.matt
8 * This program is free software; you can redistribute it and/or modify it
9 * under the terms of the GNU Lesser General Public License as published by
10 * the Free Software Foundation; either version 2.1 of the License, or
11 * (at your option) any later version.
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public License
19 * along with this program; if not, write to the Free Software Foundation,
20 * Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
21 *****************************************************************************/
22 /*****************************************************************************
24 * information on teletext format can be found here :
25 * http://pdc.ro.nu/teletext.html
27 *****************************************************************************/
34 #include <vlc_common.h>
35 #include <vlc_plugin.h>
38 #include <vlc_codec.h>
40 /* #define TELX_DEBUG */
42 # define dbg( a ) msg_Dbg a
47 /*****************************************************************************
49 *****************************************************************************/
50 static int Open ( vlc_object_t
* );
51 static void Close( vlc_object_t
* );
52 static int Decode( decoder_t
*, block_t
* );
54 #define OVERRIDE_PAGE_TEXT N_("Override page")
55 #define OVERRIDE_PAGE_LONGTEXT N_("Override the indicated page, try this if " \
56 "your subtitles don't appear (-1 = autodetect from TS, " \
57 "0 = autodetect from teletext, " \
58 ">0 = actual page number, usually 888 or 889).")
60 #define IGNORE_SUB_FLAG_TEXT N_("Ignore subtitle flag")
61 #define IGNORE_SUB_FLAG_LONGTEXT N_("Ignore the subtitle flag, try this if " \
62 "your subtitles don't appear.")
64 #define FRENCH_WORKAROUND_TEXT N_("Workaround for France")
65 #define FRENCH_WORKAROUND_LONGTEXT N_("Some French channels do not flag " \
66 "their subtitling pages correctly due to a historical " \
67 "interpretation mistake. Try using this wrong interpretation if " \
68 "your subtitles don't appear.")
71 set_description( N_("Teletext subtitles decoder") )
72 set_shortname( "Teletext" )
73 set_capability( "spu decoder", 50 )
74 set_category( CAT_INPUT
)
75 set_subcategory( SUBCAT_INPUT_SCODEC
)
76 set_callbacks( Open
, Close
)
78 add_integer( "telx-override-page", -1,
79 OVERRIDE_PAGE_TEXT
, OVERRIDE_PAGE_LONGTEXT
, true )
80 add_bool( "telx-ignore-subtitle-flag", false,
81 IGNORE_SUB_FLAG_TEXT
, IGNORE_SUB_FLAG_LONGTEXT
, true )
82 add_bool( "telx-french-workaround", false,
83 FRENCH_WORKAROUND_TEXT
, FRENCH_WORKAROUND_LONGTEXT
, true )
87 /****************************************************************************
89 ****************************************************************************/
94 bool b_is_subtitle
[9];
95 char ppsz_lines
[32][128];
96 char psz_prev_text
[512];
100 const uint16_t * pi_active_national_set
[9];
101 int i_wanted_page
, i_wanted_magazine
;
102 bool b_ignore_sub_flag
;
105 /****************************************************************************
107 ****************************************************************************/
110 * My doc only mentions 13 national characters, but experiments show there
111 * are more, in france for example I already found two more (0x9 and 0xb).
113 * Conversion is in this order :
115 * 0x23 0x24 0x40 0x5b 0x5c 0x5d 0x5e 0x5f 0x60 0x7b 0x7c 0x7d 0x7e
116 * (these are the standard ones)
117 * 0x08 0x09 0x0a 0x0b 0x0c 0x0d (apparently a control character) 0x0e 0x0f
120 static const uint16_t ppi_national_subsets
[][20] =
122 { 0x00a3, 0x0024, 0x0040, 0x00ab, 0x00bd, 0x00bb, 0x005e, 0x0023,
123 0x002d, 0x00bc, 0x00a6, 0x00be, 0x00f7 }, /* english ,000 */
125 { 0x00e9, 0x00ef, 0x00e0, 0x00eb, 0x00ea, 0x00f9, 0x00ee, 0x0023,
126 0x00e8, 0x00e2, 0x00f4, 0x00fb, 0x00e7, 0, 0x00eb, 0, 0x00ef }, /* french ,001 */
128 { 0x0023, 0x00a4, 0x00c9, 0x00c4, 0x00d6, 0x00c5, 0x00dc, 0x005f,
129 0x00e9, 0x00e4, 0x00f6, 0x00e5, 0x00fc }, /* swedish,finnish,hungarian ,010 */
131 { 0x0023, 0x016f, 0x010d, 0x0165, 0x017e, 0x00fd, 0x00ed, 0x0159,
132 0x00e9, 0x00e1, 0x011b, 0x00fa, 0x0161 }, /* czech,slovak ,011 */
134 { 0x0023, 0x0024, 0x00a7, 0x00c4, 0x00d6, 0x00dc, 0x005e, 0x005f,
135 0x00b0, 0x00e4, 0x00f6, 0x00fc, 0x00df }, /* german ,100 */
137 { 0x00e7, 0x0024, 0x00a1, 0x00e1, 0x00e9, 0x00ed, 0x00f3, 0x00fa,
138 0x00bf, 0x00fc, 0x00f1, 0x00e8, 0x00e0 }, /* portuguese,spanish ,101 */
140 { 0x00a3, 0x0024, 0x00e9, 0x00b0, 0x00e7, 0x00bb, 0x005e, 0x0023,
141 0x00f9, 0x00e0, 0x00f2, 0x00e8, 0x00ec }, /* italian ,110 */
143 { 0x0023, 0x00a4, 0x0162, 0x00c2, 0x015e, 0x0102, 0x00ce, 0x0131,
144 0x0163, 0x00e2, 0x015f, 0x0103, 0x00ee }, /* rumanian ,111 */
146 /* I have these tables too, but I don't know how they can be triggered */
147 { 0x0023, 0x0024, 0x0160, 0x0117, 0x0119, 0x017d, 0x010d, 0x016b,
148 0x0161, 0x0105, 0x0173, 0x017e, 0x012f }, /* lettish,lithuanian ,1000 */
150 { 0x0023, 0x0144, 0x0105, 0x005a, 0x015a, 0x0141, 0x0107, 0x00f3,
151 0x0119, 0x017c, 0x015b, 0x0142, 0x017a }, /* polish, 1001 */
153 { 0x0023, 0x00cb, 0x010c, 0x0106, 0x017d, 0x0110, 0x0160, 0x00eb,
154 0x010d, 0x0107, 0x017e, 0x0111, 0x0161 }, /* serbian,croatian,slovenian, 1010 */
156 { 0x0023, 0x00f5, 0x0160, 0x00c4, 0x00d6, 0x017e, 0x00dc, 0x00d5,
157 0x0161, 0x00e4, 0x00f6, 0x017e, 0x00fc }, /* estonian ,1011 */
159 { 0x0054, 0x011f, 0x0130, 0x015e, 0x00d6, 0x00c7, 0x00dc, 0x011e,
160 0x0131, 0x015f, 0x00f6, 0x00e7, 0x00fc }, /* turkish ,1100 */
164 /*****************************************************************************
165 * Open: probe the decoder and return score
166 *****************************************************************************
167 * Tries to launch a decoder and return score so that the interface is able
169 *****************************************************************************/
170 static int Open( vlc_object_t
*p_this
)
172 decoder_t
*p_dec
= (decoder_t
*) p_this
;
173 decoder_sys_t
*p_sys
= NULL
;
177 if( p_dec
->fmt_in
.i_codec
!= VLC_CODEC_TELETEXT
)
182 p_dec
->pf_decode
= Decode
;
183 p_sys
= p_dec
->p_sys
= calloc( 1, sizeof(*p_sys
) );
186 p_dec
->fmt_out
.i_codec
= 0;
189 for ( int i
= 0; i
< 9; i
++ )
190 p_sys
->pi_active_national_set
[i
] = ppi_national_subsets
[1];
192 i_val
= var_CreateGetInteger( p_dec
, "telx-override-page" );
193 if( i_val
== -1 && p_dec
->fmt_in
.subs
.teletext
.i_magazine
!= -1 &&
194 ( p_dec
->fmt_in
.subs
.teletext
.i_magazine
!= 1 ||
195 p_dec
->fmt_in
.subs
.teletext
.i_page
!= 0 ) ) /* ignore if TS demux wants page 100 (unlikely to be sub) */
198 p_sys
->i_wanted_magazine
= p_dec
->fmt_in
.subs
.teletext
.i_magazine
;
199 p_sys
->i_wanted_page
= p_dec
->fmt_in
.subs
.teletext
.i_page
;
201 b_val
= var_CreateGetBool( p_dec
, "telx-french-workaround" );
202 if( p_sys
->i_wanted_page
< 100 &&
203 (b_val
|| (p_sys
->i_wanted_page
% 16) >= 10))
205 /* See http://www.nada.kth.se/~ragge/vdr/ttxtsubs/TROUBLESHOOTING.txt
206 * paragraph about French channels - they mix up decimal and
208 p_sys
->i_wanted_page
= (p_sys
->i_wanted_page
/ 10) * 16 +
209 (p_sys
->i_wanted_page
% 10);
212 else if( i_val
<= 0 )
214 p_sys
->i_wanted_magazine
= -1;
215 p_sys
->i_wanted_page
= -1;
219 p_sys
->i_wanted_magazine
= i_val
/ 100;
220 p_sys
->i_wanted_page
= (((i_val
% 100) / 10) << 4)
221 |((i_val
% 100) % 10);
223 p_sys
->b_ignore_sub_flag
= var_CreateGetBool( p_dec
,
224 "telx-ignore-subtitle-flag" );
226 msg_Dbg( p_dec
, "starting telx on magazine %d page %02x flag %d",
227 p_sys
->i_wanted_magazine
, p_sys
->i_wanted_page
,
228 p_sys
->b_ignore_sub_flag
);
237 /* return VLC_EGENERIC; */
240 /*****************************************************************************
242 *****************************************************************************/
243 static void Close( vlc_object_t
*p_this
)
245 decoder_t
*p_dec
= (decoder_t
*) p_this
;
246 decoder_sys_t
*p_sys
= p_dec
->p_sys
;
251 /**************************
252 * change bits endianness *
253 **************************/
254 static uint8_t bytereverse( int n
)
256 n
= (((n
>> 1) & 0x55) | ((n
<< 1) & 0xaa));
257 n
= (((n
>> 2) & 0x33) | ((n
<< 2) & 0xcc));
258 n
= (((n
>> 4) & 0x0f) | ((n
<< 4) & 0xf0));
262 static int hamming_8_4( int a
)
298 return -1; // decoding error , not yet corrected
303 // this is not a general function, but it's enough for what we do here
304 // the result buffer need to be at least 4 bytes long
305 static void to_utf8( char * res
, uint16_t ch
)
311 res
[0] = (ch
>> 12) | 0xE0;
312 res
[1] = ((ch
>> 6) & 0x3F) | 0x80;
313 res
[2] = (ch
& 0x3F) | 0x80;
318 res
[0] = (ch
>> 6) | 0xC0;
319 res
[1] = (ch
& 0x3F) | 0x80;
330 static void decode_string( char * res
, int res_len
,
331 decoder_sys_t
*p_sys
, int magazine
,
332 uint8_t * packet
, int len
)
337 for ( int i
= 0; i
< len
; i
++ )
339 int in
= bytereverse( packet
[i
] ) & 0x7f;
345 /* special national characters */
347 out
= p_sys
->pi_active_national_set
[magazine
][0];
350 out
= p_sys
->pi_active_national_set
[magazine
][1];
353 out
= p_sys
->pi_active_national_set
[magazine
][2];
356 out
= p_sys
->pi_active_national_set
[magazine
][3];
359 out
= p_sys
->pi_active_national_set
[magazine
][4];
362 out
= p_sys
->pi_active_national_set
[magazine
][5];
365 out
= p_sys
->pi_active_national_set
[magazine
][6];
368 out
= p_sys
->pi_active_national_set
[magazine
][7];
371 out
= p_sys
->pi_active_national_set
[magazine
][8];
374 out
= p_sys
->pi_active_national_set
[magazine
][9];
377 out
= p_sys
->pi_active_national_set
[magazine
][10];
380 out
= p_sys
->pi_active_national_set
[magazine
][11];
383 out
= p_sys
->pi_active_national_set
[magazine
][12];
386 /* some special control characters (empirical) */
388 /* apparently this starts a sequence that ends with 0xb 0xb */
389 while ( i
+ 1 < len
&& (bytereverse( packet
[i
+1] ) & 0x7f) != 0x0b )
396 /* non documented national range 0x08 - 0x0f */
397 if ( in
>= 0x08 && in
<= 0x0f )
399 out
= p_sys
->pi_active_national_set
[magazine
][13 + in
- 8];
404 if ( in
> 32 && in
< 0x7f )
408 /* handle undefined national characters */
412 /* convert to utf-8 */
413 to_utf8( utf8
, out
);
415 if ( pt
+ l
< res
+ res_len
- 1 )
427 /*****************************************************************************
429 *****************************************************************************/
430 static int Decode( decoder_t
*p_dec
, block_t
*p_block
)
432 decoder_sys_t
*p_sys
= p_dec
->p_sys
;
433 subpicture_t
*p_spu
= NULL
;
438 int i_wanted_magazine
= i_conf_wanted_page
/ 100;
439 int i_wanted_page
= 0x10 * ((i_conf_wanted_page
% 100) / 10)
440 | (i_conf_wanted_page
% 10);
442 bool b_update
= false;
443 char psz_text
[512], *pt
= psz_text
;
447 if( p_block
== NULL
) /* No Drain */
448 return VLCDEC_SUCCESS
;
450 dbg((p_dec
, "start of telx packet with header %2x\n",
451 * (uint8_t *) p_block
->p_buffer
));
452 len
= p_block
->i_buffer
;
453 for ( int offset
= 1; offset
+ 46 <= len
; offset
+= 46 )
455 uint8_t * packet
= (uint8_t *) p_block
->p_buffer
+offset
;
456 // int vbi = ((0x20 & packet[2]) != 0 ? 0 : 313) + (0x1F & packet[2]);
458 // dbg((p_dec, "vbi %d header %02x %02x %02x\n", vbi, packet[0], packet[1], packet[2]));
459 if ( packet
[0] == 0xFF ) continue;
461 /* if (packet[1] != 0x2C) { */
462 /* printf("wrong header\n"); */
467 int mpag
= (hamming_8_4( packet
[4] ) << 4) | hamming_8_4( packet
[5] );
472 dbg((p_dec
, "mpag hamming error\n"));
476 row
= 0xFF & bytereverse(mpag
);
477 magazine
= (7 & row
) == 0 ? 8 : (7 & row
);
480 if ( p_sys
->i_wanted_page
!= -1
481 && magazine
!= p_sys
->i_wanted_magazine
)
486 /* row 0 : flags and header line */
489 for ( int a
= 0; a
< 6; a
++ )
491 flag
|= (0xF & (bytereverse( hamming_8_4(packet
[8 + a
]) ) >> 4))
495 /* if (!p_sys->b_ignore_sub_flag && !(1 & flag>>15)) */
498 p_sys
->i_page
[magazine
] = (0xF0 & bytereverse( hamming_8_4(packet
[7]) )) |
499 (0xF & (bytereverse( hamming_8_4(packet
[6]) ) >> 4) );
501 decode_string( psz_line
, sizeof(psz_line
), p_sys
, magazine
,
502 packet
+ 14, 40 - 14 );
504 dbg((p_dec
, "mag %d flags %x page %x character set %d subtitles %d", magazine
, flag
,
505 p_sys
->i_page
[magazine
],
506 7 & flag
>>21, 1 & flag
>>15, psz_line
));
508 p_sys
->pi_active_national_set
[magazine
] =
509 ppi_national_subsets
[7 & (flag
>> 21)];
511 p_sys
->b_is_subtitle
[magazine
] = p_sys
->b_ignore_sub_flag
512 || ( (1 & (flag
>> 15))
513 && (1 & (flag
>>16)) );
515 dbg(( p_dec
, "FLAGS%s%s%s%s%s%s%s mag_ser %d",
516 (1 & (flag
>>14))? " news" : "",
517 (1 & (flag
>>15))? " subtitle" : "",
518 (1 & (flag
>>7))? " erase" : "",
519 (1 & (flag
>>16))? " suppressed_head" : "",
520 (1 & (flag
>>17))? " update" : "",
521 (1 & (flag
>>18))? " interrupt" : "",
522 (1 & (flag
>>19))? " inhibit" : "",
525 if ( (p_sys
->i_wanted_page
!= -1
526 && p_sys
->i_page
[magazine
] != p_sys
->i_wanted_page
)
527 || !p_sys
->b_is_subtitle
[magazine
] )
530 p_sys
->b_erase
[magazine
] = (1 & (flag
>> 7));
532 dbg((p_dec
, "%ld --> %ld\n", (long int) p_block
->i_pts
, (long int)(p_sys
->prev_pts
+1500000)));
534 * we ignore the erase flag if it happens less than 1.5 seconds
535 * before last caption
536 * TODO make this time configurable
537 * UPDATE the kludge seems to be no more necessary
538 * so it's commented out*/
539 if ( /*p_block->i_pts > p_sys->prev_pts + 1500000 && */
540 p_sys
->b_erase
[magazine
] )
542 dbg((p_dec
, "ERASE !\n"));
544 p_sys
->b_erase
[magazine
] = 0;
545 for ( int i
= 1; i
< 32; i
++ )
547 if ( !p_sys
->ppsz_lines
[i
][0] ) continue;
548 /* b_update = true; */
549 p_sys
->ppsz_lines
[i
][0] = 0;
553 /* replace the row if it's different */
554 if ( strcmp(psz_line
, p_sys
->ppsz_lines
[row
]) )
556 strncpy( p_sys
->ppsz_lines
[row
], psz_line
,
557 sizeof(p_sys
->ppsz_lines
[row
]) - 1);
566 /* row 1-23 : normal lines */
568 if ( (p_sys
->i_wanted_page
!= -1
569 && p_sys
->i_page
[magazine
] != p_sys
->i_wanted_page
)
570 || !p_sys
->b_is_subtitle
[magazine
]
571 || (p_sys
->i_wanted_page
== -1
572 && p_sys
->i_page
[magazine
] > 0x99) )
575 decode_string( psz_line
, sizeof(psz_line
), p_sys
, magazine
,
579 /* remove starting spaces */
580 while ( *t
== 32 ) t
++;
582 /* remove trailing spaces */
583 for ( i
= strlen(t
) - 1; i
>= 0 && t
[i
] == 32; i
-- );
586 /* replace the row if it's different */
587 if ( strcmp( t
, p_sys
->ppsz_lines
[row
] ) )
589 strncpy( p_sys
->ppsz_lines
[row
], t
,
590 sizeof(p_sys
->ppsz_lines
[row
]) - 1 );
595 p_sys
->prev_pts
= p_block
->i_pts
;
597 dbg((p_dec
, "%d %d : ", magazine
, row
));
598 dbg((p_dec
, "%s\n", t
));
604 for ( i
= 0; i
< 40; i
++ )
606 int in
= bytereverse(packet
[6 + i
]) & 0x7f;
607 sprintf(dbg
+ strlen(dbg
), "%02x ", in
);
609 dbg((p_dec
, "%s\n", dbg
));
611 for ( i
= 0; i
< 40; i
++ )
613 decode_string( psz_line
, sizeof(psz_line
), p_sys
, magazine
,
615 sprintf( dbg
+ strlen(dbg
), "%s ", psz_line
);
617 dbg((p_dec
, "%s\n", dbg
));
621 else if ( row
== 25 )
623 /* row 25 : alternate header line */
624 if ( (p_sys
->i_wanted_page
!= -1
625 && p_sys
->i_page
[magazine
] != p_sys
->i_wanted_page
)
626 || !p_sys
->b_is_subtitle
[magazine
] )
629 decode_string( psz_line
, sizeof(psz_line
), p_sys
, magazine
,
632 /* replace the row if it's different */
633 if ( strcmp( psz_line
, p_sys
->ppsz_lines
[0] ) )
635 strncpy( p_sys
->ppsz_lines
[0], psz_line
,
636 sizeof(p_sys
->ppsz_lines
[0]) - 1 );
637 /* b_update = true; */
640 /* else if (row == 26) { */
641 /* // row 26 : TV listings */
643 /* dbg((p_dec, "%d %d : %s\n", magazine, row, decode_string(p_sys, magazine, packet+6, 40))); */
650 for ( int i
= 1; i
< 24; i
++ )
652 size_t l
= strlen( p_sys
->ppsz_lines
[i
] );
654 if ( l
> sizeof(psz_text
) - total
- 1 )
655 l
= sizeof(psz_text
) - total
- 1;
659 memcpy( pt
, p_sys
->ppsz_lines
[i
], l
);
662 if ( sizeof(psz_text
) - total
- 1 > 0 )
671 if ( !strcmp(psz_text
, p_sys
->psz_prev_text
) )
674 dbg((p_dec
, "UPDATE TELETEXT PICTURE\n"));
676 assert( sizeof(p_sys
->psz_prev_text
) >= sizeof(psz_text
) );
677 strcpy( p_sys
->psz_prev_text
, psz_text
);
679 /* Create the subpicture unit */
680 p_spu
= decoder_NewSubpicture( p_dec
, NULL
);
683 msg_Warn( p_dec
, "can't get spu buffer" );
687 /* Create a new subpicture region */
688 video_format_Init(&fmt
, VLC_CODEC_TEXT
);
689 fmt
.i_width
= fmt
.i_height
= 0;
690 fmt
.i_x_offset
= fmt
.i_y_offset
= 0;
691 p_spu
->p_region
= subpicture_region_New( &fmt
);
692 if( p_spu
->p_region
== NULL
)
694 msg_Err( p_dec
, "cannot allocate SPU region" );
698 /* Normal text subs, easy markup */
699 p_spu
->p_region
->i_align
= SUBPICTURE_ALIGN_BOTTOM
| p_sys
->i_align
;
700 p_spu
->p_region
->i_x
= p_sys
->i_align
? 20 : 0;
701 p_spu
->p_region
->i_y
= 10;
702 p_spu
->p_region
->p_text
= text_segment_New(psz_text
);
704 p_spu
->i_start
= p_block
->i_pts
;
705 p_spu
->i_stop
= p_block
->i_pts
+ p_block
->i_length
;
706 p_spu
->b_ephemer
= (p_block
->i_length
== 0);
707 p_spu
->b_absolute
= false;
708 dbg((p_dec
, "%ld --> %ld\n", (long int) p_block
->i_pts
/100000, (long int)p_block
->i_length
/100000));
710 block_Release( p_block
);
712 decoder_QueueSub( p_dec
, p_spu
);
713 return VLCDEC_SUCCESS
;
718 subpicture_Delete( p_spu
);
722 block_Release( p_block
);
723 return VLCDEC_SUCCESS
;