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
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 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 General Public License for more details.
18 * You should have received a copy of the GNU General Public License
19 * along with this program; if not, write to the Free Software
20 * Foundation, 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>
39 #include "vlc_codec.h"
41 /* #define TELX_DEBUG */
43 # define dbg( a ) msg_Dbg a
48 /*****************************************************************************
50 *****************************************************************************/
51 static int Open ( vlc_object_t * );
52 static void Close( vlc_object_t * );
53 static subpicture_t *Decode( decoder_t *, block_t ** );
55 #define OVERRIDE_PAGE_TEXT N_("Override page")
56 #define OVERRIDE_PAGE_LONGTEXT N_("Override the indicated page, try this if " \
57 "your subtitles don't appear (-1 = autodetect from TS, " \
58 "0 = autodetect from teletext, " \
59 ">0 = actual page number, usually 888 or 889).")
61 #define IGNORE_SUB_FLAG_TEXT N_("Ignore subtitle flag")
62 #define IGNORE_SUB_FLAG_LONGTEXT N_("Ignore the subtitle flag, try this if " \
63 "your subtitles don't appear.")
65 #define FRENCH_WORKAROUND_TEXT N_("Workaround for France")
66 #define FRENCH_WORKAROUND_LONGTEXT N_("Some French channels do not flag " \
67 "their subtitling pages correctly due to a historical " \
68 "interpretation mistake. Try using this wrong interpretation if " \
69 "your subtitles don't appear.")
72 set_description( N_("Teletext subtitles decoder") )
73 set_shortname( "Teletext" )
74 set_capability( "decoder", 50 )
75 set_category( CAT_INPUT )
76 set_subcategory( SUBCAT_INPUT_SCODEC )
77 set_callbacks( Open, Close )
79 add_integer( "telx-override-page", -1, NULL,
80 OVERRIDE_PAGE_TEXT, OVERRIDE_PAGE_LONGTEXT, true )
81 add_bool( "telx-ignore-subtitle-flag", 0, NULL,
82 IGNORE_SUB_FLAG_TEXT, IGNORE_SUB_FLAG_LONGTEXT, true )
83 add_bool( "telx-french-workaround", 0, NULL,
84 FRENCH_WORKAROUND_TEXT, FRENCH_WORKAROUND_LONGTEXT, true )
88 /****************************************************************************
90 ****************************************************************************/
95 bool b_is_subtitle[9];
96 char ppsz_lines[32][128];
97 char psz_prev_text[512];
101 const uint16_t * pi_active_national_set[9];
102 int i_wanted_page, i_wanted_magazine;
103 bool b_ignore_sub_flag;
106 /****************************************************************************
108 ****************************************************************************/
111 * My doc only mentions 13 national characters, but experiments show there
112 * are more, in france for example I already found two more (0x9 and 0xb).
114 * Conversion is in this order :
116 * 0x23 0x24 0x40 0x5b 0x5c 0x5d 0x5e 0x5f 0x60 0x7b 0x7c 0x7d 0x7e
117 * (these are the standard ones)
118 * 0x08 0x09 0x0a 0x0b 0x0c 0x0d (apparently a control character) 0x0e 0x0f
121 static const uint16_t ppi_national_subsets[][20] =
123 { 0x00a3, 0x0024, 0x0040, 0x00ab, 0x00bd, 0x00bb, 0x005e, 0x0023,
124 0x002d, 0x00bc, 0x00a6, 0x00be, 0x00f7 }, /* english ,000 */
126 { 0x00e9, 0x00ef, 0x00e0, 0x00eb, 0x00ea, 0x00f9, 0x00ee, 0x0023,
127 0x00e8, 0x00e2, 0x00f4, 0x00fb, 0x00e7, 0, 0x00eb, 0, 0x00ef }, /* french ,001 */
129 { 0x0023, 0x00a4, 0x00c9, 0x00c4, 0x00d6, 0x00c5, 0x00dc, 0x005f,
130 0x00e9, 0x00e4, 0x00f6, 0x00e5, 0x00fc }, /* swedish,finnish,hungarian ,010 */
132 { 0x0023, 0x016f, 0x010d, 0x0165, 0x017e, 0x00fd, 0x00ed, 0x0159,
133 0x00e9, 0x00e1, 0x011b, 0x00fa, 0x0161 }, /* czech,slovak ,011 */
135 { 0x0023, 0x0024, 0x00a7, 0x00c4, 0x00d6, 0x00dc, 0x005e, 0x005f,
136 0x00b0, 0x00e4, 0x00f6, 0x00fc, 0x00df }, /* german ,100 */
138 { 0x00e7, 0x0024, 0x00a1, 0x00e1, 0x00e9, 0x00ed, 0x00f3, 0x00fa,
139 0x00bf, 0x00fc, 0x00f1, 0x00e8, 0x00e0 }, /* portuguese,spanish ,101 */
141 { 0x00a3, 0x0024, 0x00e9, 0x00b0, 0x00e7, 0x00bb, 0x005e, 0x0023,
142 0x00f9, 0x00e0, 0x00f2, 0x00e8, 0x00ec }, /* italian ,110 */
144 { 0x0023, 0x00a4, 0x0162, 0x00c2, 0x015e, 0x0102, 0x00ce, 0x0131,
145 0x0163, 0x00e2, 0x015f, 0x0103, 0x00ee }, /* rumanian ,111 */
147 /* I have these tables too, but I don't know how they can be triggered */
148 { 0x0023, 0x0024, 0x0160, 0x0117, 0x0119, 0x017d, 0x010d, 0x016b,
149 0x0161, 0x0105, 0x0173, 0x017e, 0x012f }, /* lettish,lithuanian ,1000 */
151 { 0x0023, 0x0144, 0x0105, 0x005a, 0x015a, 0x0141, 0x0107, 0x00f3,
152 0x0119, 0x017c, 0x015b, 0x0142, 0x017a }, /* polish, 1001 */
154 { 0x0023, 0x00cb, 0x010c, 0x0106, 0x017d, 0x0110, 0x0160, 0x00eb,
155 0x010d, 0x0107, 0x017e, 0x0111, 0x0161 }, /* serbian,croatian,slovenian, 1010 */
157 { 0x0023, 0x00f5, 0x0160, 0x00c4, 0x00d6, 0x017e, 0x00dc, 0x00d5,
158 0x0161, 0x00e4, 0x00f6, 0x017e, 0x00fc }, /* estonian ,1011 */
160 { 0x0054, 0x011f, 0x0130, 0x015e, 0x00d6, 0x00c7, 0x00dc, 0x011e,
161 0x0131, 0x015f, 0x00f6, 0x00e7, 0x00fc }, /* turkish ,1100 */
165 /*****************************************************************************
166 * Open: probe the decoder and return score
167 *****************************************************************************
168 * Tries to launch a decoder and return score so that the interface is able
170 *****************************************************************************/
171 static int Open( vlc_object_t *p_this )
173 decoder_t *p_dec = (decoder_t *) p_this;
174 decoder_sys_t *p_sys = NULL;
178 if( p_dec->fmt_in.i_codec != VLC_FOURCC('t','e','l','x'))
183 p_dec->pf_decode_sub = Decode;
184 p_sys = p_dec->p_sys = calloc( 1, sizeof(*p_sys) );
187 p_dec->fmt_out.i_cat = SPU_ES;
188 p_dec->fmt_out.i_codec = 0;
191 for ( i = 0; i < 9; i++ )
192 p_sys->pi_active_national_set[i] = ppi_national_subsets[1];
194 var_Create( p_dec, "telx-override-page",
195 VLC_VAR_INTEGER | VLC_VAR_DOINHERIT );
196 var_Get( p_dec, "telx-override-page", &val );
197 if( val.i_int == -1 &&
198 p_dec->fmt_in.subs.teletext.i_magazine != -1 &&
199 ( p_dec->fmt_in.subs.teletext.i_magazine != 1 ||
200 p_dec->fmt_in.subs.teletext.i_page != 0 ) ) /* ignore if TS demux wants page 100 (unlikely to be sub) */
202 p_sys->i_wanted_magazine = p_dec->fmt_in.subs.teletext.i_magazine;
203 p_sys->i_wanted_page = p_dec->fmt_in.subs.teletext.i_page;
205 var_Create( p_dec, "telx-french-workaround",
206 VLC_VAR_BOOL | VLC_VAR_DOINHERIT );
207 var_Get( p_dec, "telx-french-workaround", &val );
208 if( p_sys->i_wanted_page < 100 &&
209 (val.b_bool || (p_sys->i_wanted_page % 16) >= 10))
211 /* See http://www.nada.kth.se/~ragge/vdr/ttxtsubs/TROUBLESHOOTING.txt
212 * paragraph about French channels - they mix up decimal and
214 p_sys->i_wanted_page = (p_sys->i_wanted_page / 10) * 16 +
215 (p_sys->i_wanted_page % 10);
218 else if( val.i_int <= 0 )
220 p_sys->i_wanted_magazine = -1;
221 p_sys->i_wanted_page = -1;
225 p_sys->i_wanted_magazine = val.i_int / 100;
226 p_sys->i_wanted_page = (((val.i_int % 100) / 10) << 4)
227 | ((val.i_int % 100) % 10);
229 var_Create( p_dec, "telx-ignore-subtitle-flag",
230 VLC_VAR_BOOL | VLC_VAR_DOINHERIT );
231 var_Get( p_dec, "telx-ignore-subtitle-flag", &val );
232 p_sys->b_ignore_sub_flag = val.b_bool;
234 msg_Dbg( p_dec, "starting telx on magazine %d page %02x flag %d",
235 p_sys->i_wanted_magazine, p_sys->i_wanted_page,
236 p_sys->b_ignore_sub_flag );
245 /* return VLC_EGENERIC; */
248 /*****************************************************************************
250 *****************************************************************************/
251 static void Close( vlc_object_t *p_this )
253 decoder_t *p_dec = (decoder_t*) p_this;
254 decoder_sys_t *p_sys = p_dec->p_sys;
259 /**************************
260 * change bits endianness *
261 **************************/
262 static uint8_t bytereverse( int n )
264 n = (((n >> 1) & 0x55) | ((n << 1) & 0xaa));
265 n = (((n >> 2) & 0x33) | ((n << 2) & 0xcc));
266 n = (((n >> 4) & 0x0f) | ((n << 4) & 0xf0));
270 static int hamming_8_4( int a )
306 return -1; // decoding error , not yet corrected
311 // this is not a general function, but it's enough for what we do here
312 // the result buffer need to be at least 4 bytes long
313 static void to_utf8( char * res, uint16_t ch )
319 res[0] = (ch >> 12) | 0xE0;
320 res[1] = ((ch >> 6) & 0x3F) | 0x80;
321 res[2] = (ch & 0x3F) | 0x80;
326 res[0] = (ch >> 6) | 0xC0;
327 res[1] = (ch & 0x3F) | 0x80;
338 static void decode_string( char * res, int res_len,
339 decoder_sys_t *p_sys, int magazine,
340 uint8_t * packet, int len )
346 for ( i = 0; i < len; i++ )
348 int in = bytereverse( packet[i] ) & 0x7f;
354 /* special national characters */
356 out = p_sys->pi_active_national_set[magazine][0];
359 out = p_sys->pi_active_national_set[magazine][1];
362 out = p_sys->pi_active_national_set[magazine][2];
365 out = p_sys->pi_active_national_set[magazine][3];
368 out = p_sys->pi_active_national_set[magazine][4];
371 out = p_sys->pi_active_national_set[magazine][5];
374 out = p_sys->pi_active_national_set[magazine][6];
377 out = p_sys->pi_active_national_set[magazine][7];
380 out = p_sys->pi_active_national_set[magazine][8];
383 out = p_sys->pi_active_national_set[magazine][9];
386 out = p_sys->pi_active_national_set[magazine][10];
389 out = p_sys->pi_active_national_set[magazine][11];
392 out = p_sys->pi_active_national_set[magazine][12];
395 /* some special control characters (empirical) */
397 /* apparently this starts a sequence that ends with 0xb 0xb */
398 while ( i + 1 < len && (bytereverse( packet[i+1] ) & 0x7f) != 0x0b )
405 /* non documented national range 0x08 - 0x0f */
406 if ( in >= 0x08 && in <= 0x0f )
408 out = p_sys->pi_active_national_set[magazine][13 + in - 8];
413 if ( in > 32 && in < 0x7f )
417 /* handle undefined national characters */
421 /* convert to utf-8 */
422 to_utf8( utf8, out );
424 if ( pt + l < res + res_len - 1 )
436 /*****************************************************************************
438 *****************************************************************************/
439 static subpicture_t *Decode( decoder_t *p_dec, block_t **pp_block )
441 decoder_sys_t *p_sys = p_dec->p_sys;
443 subpicture_t *p_spu = NULL;
448 int i_wanted_magazine = i_conf_wanted_page / 100;
449 int i_wanted_page = 0x10 * ((i_conf_wanted_page % 100) / 10)
450 | (i_conf_wanted_page % 10);
452 bool b_update = false;
453 char psz_text[512], *pt = psz_text;
457 if( pp_block == NULL || *pp_block == NULL )
462 dbg((p_dec, "start of telx packet with header %2x\n",
463 * (uint8_t *) p_block->p_buffer));
464 len = p_block->i_buffer;
465 for ( offset = 1; offset + 46 <= len; offset += 46 )
467 uint8_t * packet = (uint8_t *) p_block->p_buffer+offset;
468 // int vbi = ((0x20 & packet[2]) != 0 ? 0 : 313) + (0x1F & packet[2]);
470 // dbg((p_dec, "vbi %d header %02x %02x %02x\n", vbi, packet[0], packet[1], packet[2]));
471 if ( packet[0] == 0xFF ) continue;
473 /* if (packet[1] != 0x2C) { */
474 /* printf("wrong header\n"); */
479 int mpag = (hamming_8_4( packet[4] ) << 4) | hamming_8_4( packet[5] );
484 dbg((p_dec, "mpag hamming error\n"));
488 row = 0xFF & bytereverse(mpag);
489 magazine = (7 & row) == 0 ? 8 : (7 & row);
492 if ( p_sys->i_wanted_page != -1
493 && magazine != p_sys->i_wanted_magazine )
498 /* row 0 : flags and header line */
502 for ( a = 0; a < 6; a++ )
504 flag |= (0xF & (bytereverse( hamming_8_4(packet[8 + a]) ) >> 4))
508 /* if (!p_sys->b_ignore_sub_flag && !(1 & flag>>15)) */
511 p_sys->i_page[magazine] = (0xF0 & bytereverse( hamming_8_4(packet[7]) )) |
512 (0xF & (bytereverse( hamming_8_4(packet[6]) ) >> 4) );
514 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
515 packet + 14, 40 - 14 );
517 dbg((p_dec, "mag %d flags %x page %x character set %d subtitles %d", magazine, flag,
518 p_sys->i_page[magazine],
519 7 & flag>>21, 1 & flag>>15, psz_line));
521 p_sys->pi_active_national_set[magazine] =
522 ppi_national_subsets[7 & (flag >> 21)];
524 p_sys->b_is_subtitle[magazine] = p_sys->b_ignore_sub_flag
525 || ( (1 & (flag >> 15))
526 && (1 & (flag>>16)) );
528 dbg(( p_dec, "FLAGS%s%s%s%s%s%s%s mag_ser %d",
529 (1 & (flag>>14))? " news" : "",
530 (1 & (flag>>15))? " subtitle" : "",
531 (1 & (flag>>7))? " erase" : "",
532 (1 & (flag>>16))? " suppressed_head" : "",
533 (1 & (flag>>17))? " update" : "",
534 (1 & (flag>>18))? " interrupt" : "",
535 (1 & (flag>>19))? " inhibit" : "",
538 if ( (p_sys->i_wanted_page != -1
539 && p_sys->i_page[magazine] != p_sys->i_wanted_page)
540 || !p_sys->b_is_subtitle[magazine] )
543 p_sys->b_erase[magazine] = (1 & (flag >> 7));
545 dbg((p_dec, "%ld --> %ld\n", (long int) p_block->i_pts, (long int)(p_sys->prev_pts+1500000)));
547 * we ignore the erase flag if it happens less than 1.5 seconds
548 * before last caption
549 * TODO make this time configurable
550 * UPDATE the kludge seems to be no more necessary
551 * so it's commented out*/
552 if ( /*p_block->i_pts > p_sys->prev_pts + 1500000 && */
553 p_sys->b_erase[magazine] )
557 dbg((p_dec, "ERASE !\n"));
559 p_sys->b_erase[magazine] = 0;
560 for ( i = 1; i < 32; i++ )
562 if ( !p_sys->ppsz_lines[i][0] ) continue;
563 /* b_update = true; */
564 p_sys->ppsz_lines[i][0] = 0;
568 /* replace the row if it's different */
569 if ( strcmp(psz_line, p_sys->ppsz_lines[row]) )
571 strncpy( p_sys->ppsz_lines[row], psz_line,
572 sizeof(p_sys->ppsz_lines[row]) - 1);
581 /* row 1-23 : normal lines */
583 if ( (p_sys->i_wanted_page != -1
584 && p_sys->i_page[magazine] != p_sys->i_wanted_page)
585 || !p_sys->b_is_subtitle[magazine]
586 || (p_sys->i_wanted_page == -1
587 && p_sys->i_page[magazine] > 0x99) )
590 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
594 /* remove starting spaces */
595 while ( *t == 32 ) t++;
597 /* remove trailing spaces */
598 for ( i = strlen(t) - 1; i >= 0 && t[i] == 32; i-- );
601 /* replace the row if it's different */
602 if ( strcmp( t, p_sys->ppsz_lines[row] ) )
604 strncpy( p_sys->ppsz_lines[row], t,
605 sizeof(p_sys->ppsz_lines[row]) - 1 );
610 p_sys->prev_pts = p_block->i_pts;
612 dbg((p_dec, "%d %d : ", magazine, row));
613 dbg((p_dec, "%s\n", t));
619 for ( i = 0; i < 40; i++ )
621 int in = bytereverse(packet[6 + i]) & 0x7f;
622 sprintf(dbg + strlen(dbg), "%02x ", in);
624 dbg((p_dec, "%s\n", dbg));
626 for ( i = 0; i < 40; i++ )
628 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
630 sprintf( dbg + strlen(dbg), "%s ", psz_line );
632 dbg((p_dec, "%s\n", dbg));
636 else if ( row == 25 )
638 /* row 25 : alternate header line */
639 if ( (p_sys->i_wanted_page != -1
640 && p_sys->i_page[magazine] != p_sys->i_wanted_page)
641 || !p_sys->b_is_subtitle[magazine] )
644 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
647 /* replace the row if it's different */
648 if ( strcmp( psz_line, p_sys->ppsz_lines[0] ) )
650 strncpy( p_sys->ppsz_lines[0], psz_line,
651 sizeof(p_sys->ppsz_lines[0]) - 1 );
652 /* b_update = true; */
655 /* else if (row == 26) { */
656 /* // row 26 : TV listings */
658 /* dbg((p_dec, "%d %d : %s\n", magazine, row, decode_string(p_sys, magazine, packet+6, 40))); */
665 for ( i = 1; i < 24; i++ )
667 size_t l = strlen( p_sys->ppsz_lines[i] );
669 if ( l > sizeof(psz_text) - total - 1 )
670 l = sizeof(psz_text) - total - 1;
674 memcpy( pt, p_sys->ppsz_lines[i], l );
677 if ( sizeof(psz_text) - total - 1 > 0 )
686 if ( !strcmp(psz_text, p_sys->psz_prev_text) )
689 dbg((p_dec, "UPDATE TELETEXT PICTURE\n"));
691 assert( sizeof(p_sys->psz_prev_text) >= sizeof(psz_text) );
692 strcpy( p_sys->psz_prev_text, psz_text );
694 /* Create the subpicture unit */
695 p_spu = decoder_NewSubpicture( p_dec );
698 msg_Warn( p_dec, "can't get spu buffer" );
702 /* Create a new subpicture region */
703 memset( &fmt, 0, sizeof(video_format_t) );
704 fmt.i_chroma = VLC_FOURCC('T','E','X','T');
706 fmt.i_width = fmt.i_height = 0;
707 fmt.i_x_offset = fmt.i_y_offset = 0;
708 p_spu->p_region = subpicture_region_New( &fmt );
709 if( p_spu->p_region == NULL )
711 msg_Err( p_dec, "cannot allocate SPU region" );
715 /* Normal text subs, easy markup */
716 p_spu->p_region->i_align = SUBPICTURE_ALIGN_BOTTOM | p_sys->i_align;
717 p_spu->p_region->i_x = p_sys->i_align ? 20 : 0;
718 p_spu->p_region->i_y = 10;
719 p_spu->p_region->psz_text = strdup(psz_text);
721 p_spu->i_start = p_block->i_pts;
722 p_spu->i_stop = p_block->i_pts + p_block->i_length;
723 p_spu->b_ephemer = (p_block->i_length == 0);
724 p_spu->b_absolute = false;
725 dbg((p_dec, "%ld --> %ld\n", (long int) p_block->i_pts/100000, (long int)p_block->i_length/100000));
727 block_Release( p_block );
733 decoder_DeleteSubpicture( p_dec, p_spu );
737 block_Release( p_block );