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
29 *****************************************************************************/
34 /* This is an ugly test so that this source compile for both 0.8.6 and current
35 * trunk version. Considere this hack to be temporary, but I need it so that
36 * I can maintain my patch for both current svn and 0.8.6 in an easy way. */
43 #include <vlc/decoder.h>
45 /* #include "vlc_es.h" */
46 /* #include "vlc_block.h" */
47 /* #include "vlc_video.h" */
48 /* #include "vlc_spu.h" */
52 #include "vlc_codec.h"
55 /* #define TELX_DEBUG */
58 # define dbg( a ) msg_Dbg a
63 /*****************************************************************************
65 *****************************************************************************/
66 static int telx_conf_cb ( vlc_object_t *, /* variable's object */
67 char const *, /* variable name */
68 vlc_value_t, /* old value */
69 vlc_value_t, /* new value */
70 void * ); /* callback data */
72 /*****************************************************************************
74 *****************************************************************************/
75 static int Open ( vlc_object_t * );
76 static void Close( vlc_object_t * );
77 static subpicture_t *Decode( decoder_t *, block_t ** );
80 #define PAGE_TEXT N_("Teletext page")
81 #define PAGE_LONGTEXT N_("Set displayed teletext page for subtitles, 0 for all pages, 888 should be a standard value. Just leave it to zero if your stream has only one language for subtitles.")
83 #define IGNORE_SUB_FLAG_TEXT N_("Ignore subtitle flag")
84 #define IGNORE_SUB_FLAG_LONGTEXT N_("Ignore the subtitle flag, try this if your subtitles don't appear.")
87 # define TELX_CFG_PREFIX "telx-"
88 set_description( _("Teletext subtitles decoder") );
89 set_capability( "decoder", 50 );
90 set_category( CAT_INPUT );
91 set_subcategory( SUBCAT_INPUT_SCODEC );
92 set_callbacks( Open, Close );
94 add_integer( "telx-page", 0, telx_conf_cb, PAGE_TEXT, PAGE_LONGTEXT,
96 add_bool( "telx-ignore-subtitle-flag", 0, telx_conf_cb,
97 IGNORE_SUB_FLAG_TEXT, IGNORE_SUB_FLAG_LONGTEXT, VLC_TRUE );
101 /****************************************************************************
103 ****************************************************************************/
108 vlc_bool_t b_is_subtitle[9];
109 char ppsz_lines[32][128];
110 char psz_prev_text[512];
113 vlc_bool_t b_erase[9];
114 uint16_t * pi_active_national_set[9];
117 /****************************************************************************
119 ****************************************************************************/
121 static int i_conf_wanted_page = 0; /* default 0 = all pages */
122 static vlc_bool_t b_ignore_sub_flag = 0;
125 * My doc only mentions 13 national characters, but experiments show there
126 * are more, in france for example I already found two more (0x9 and 0xb).
128 * Conversion is in this order :
130 * 0x23 0x24 0x40 0x5b 0x5c 0x5d 0x5e 0x5f 0x60 0x7b 0x7c 0x7d 0x7e
131 * (these are the standard ones)
132 * 0x08 0x09 0x0a 0x0b 0x0c 0x0d (apparently a control character) 0x0e 0x0f
135 static uint16_t ppi_national_subsets[][20] =
137 { 0x00a3, 0x0024, 0x0040, 0x00ab, 0x00bd, 0x00bb, 0x005e, 0x0023,
138 0x002d, 0x00bc, 0x00a6, 0x00be, 0x00f7 }, /* english ,000 */
140 { 0x00e9, 0x00ef, 0x00e0, 0x00eb, 0x00ea, 0x00f9, 0x00ee, 0x0023,
141 0x00e8, 0x00e2, 0x00f4, 0x00fb, 0x00e7, 0, 0x00eb, 0, 0x00ef }, /* french ,001 */
143 { 0x0023, 0x00a4, 0x00c9, 0x00c4, 0x00d6, 0x00c5, 0x00dc, 0x005f,
144 0x00e9, 0x00e4, 0x00f6, 0x00e5, 0x00fc }, /* swedish,finnish,hungarian ,010 */
146 { 0x0023, 0x016f, 0x010d, 0x0165, 0x017e, 0x00fd, 0x00ed, 0x0159,
147 0x00e9, 0x00e1, 0x0115, 0x00fa, 0x0161 }, /* czech,slovak ,011 */
149 { 0x0023, 0x0024, 0x00a7, 0x00c4, 0x00d6, 0x00dc, 0x005e, 0x005f,
150 0x00b0, 0x00e4, 0x00f6, 0x00fc, 0x00df }, /* german ,100 */
152 { 0x00e7, 0x0024, 0x00a1, 0x00e1, 0x00e9, 0x00ed, 0x00f3, 0x00fa,
153 0x00bf, 0x00fc, 0x00f1, 0x00e8, 0x00e0 }, /* portuguese,spanish ,101 */
155 { 0x00a3, 0x0024, 0x00e9, 0x00b0, 0x00e7, 0x00bb, 0x005e, 0x0023,
156 0x00f9, 0x00e0, 0x00f2, 0x00e8, 0x00ec }, /* italian ,110 */
158 { 0x0023, 0x00a4, 0x0162, 0x00c2, 0x015e, 0x0102, 0x00ce, 0x0131,
159 0x0163, 0x00e2, 0x015f, 0x0103, 0x00ee }, /* rumanian ,111 */
161 { 0x0023, 0x0024, 0x0160, 0x0117, 0x0119, 0x017d, 0x010d, 0x016b,
162 0x0161, 0x0105, 0x0173, 0x017e, 0x012f }, /* lettish,lithuanian ,1000 */
164 { 0x0023, 0x0144, 0x0105, 0x005a, 0x015a, 0x0141, 0x0107, 0x00f3,
165 0x0119, 0x017c, 0x015b, 0x0142, 0x017a }, /* polish, 1001 */
167 { 0x0023, 0x00cb, 0x010c, 0x0106, 0x017d, 0x0110, 0x0160, 0x00eb,
168 0x010d, 0x0107, 0x017e, 0x0111, 0x0161 }, /* serbian,croatian,slovenian, 1010 */
170 { 0x0023, 0x00f5, 0x0160, 0x00c4, 0x00d6, 0x017e, 0x00dc, 0x00d5,
171 0x0161, 0x00e4, 0x00f6, 0x017e, 0x00fc }, /* estonian ,1011 */
173 { 0x0054, 0x011f, 0x0130, 0x015e, 0x00d6, 0x00c7, 0x00dc, 0x011e,
174 0x0131, 0x015f, 0x00f6, 0x00e7, 0x00fc }, /* turkish ,1100 */
178 /*****************************************************************************
179 * Open: probe the decoder and return score
180 *****************************************************************************
181 * Tries to launch a decoder and return score so that the interface is able
183 *****************************************************************************/
184 static int Open( vlc_object_t *p_this )
186 decoder_t *p_dec = (decoder_t *) p_this;
187 decoder_sys_t *p_sys = NULL;
191 if( p_dec->fmt_in.i_codec != VLC_FOURCC('t','e','l','x'))
196 p_dec->pf_decode_sub = Decode;
197 p_sys = p_dec->p_sys = malloc( sizeof(decoder_sys_t) );
200 msg_Err( p_dec, "out of memory" );
204 memset( p_sys, 0, sizeof(decoder_sys_t) );
207 for ( i = 0; i < 9; i++ )
208 p_sys->pi_active_national_set[i] = ppi_national_subsets[1];
210 var_Create( p_dec, "telx-page", VLC_VAR_INTEGER | VLC_VAR_DOINHERIT );
211 var_Get( p_dec, "telx-page", &val );
212 i_conf_wanted_page = val.i_int;
214 var_Create( p_dec, "telx-ignore-subtitle-flag",
215 VLC_VAR_BOOL | VLC_VAR_DOINHERIT );
216 var_Get( p_dec, "telx-ignore-subtitle-flag", &val );
217 b_ignore_sub_flag = val.b_bool;
226 /* return VLC_EGENERIC; */
229 /*****************************************************************************
231 *****************************************************************************/
232 static int telx_conf_cb ( vlc_object_t * obj, /* variable's object */
233 char const * name, /* variable name */
234 vlc_value_t oldv, /* old value */
235 vlc_value_t newv, /* new value */
236 void * data) /* callback data */
238 if ( !strcmp(name, "telx-page") )
240 i_conf_wanted_page = newv.i_int;
241 dbg((obj, "display teletext page changed to %d\n", i_conf_wanted_page));
243 else if ( !strcmp(name, "telx-ignore-subtitle-flag") )
245 b_ignore_sub_flag = newv.b_bool;
246 dbg((obj, "ignore sub flag changed to %d\n", (int) b_ignore_sub_flag));
253 /*****************************************************************************
255 *****************************************************************************/
256 static void Close( vlc_object_t *p_this )
258 decoder_t *p_dec = (decoder_t*) p_this;
259 decoder_sys_t *p_sys = p_dec->p_sys;
264 /**************************
265 * change bits endianness *
266 **************************/
267 static uint8_t bytereverse( int n )
269 n = (((n >> 1) & 0x55) | ((n << 1) & 0xaa));
270 n = (((n >> 2) & 0x33) | ((n << 2) & 0xcc));
271 n = (((n >> 4) & 0x0f) | ((n << 4) & 0xf0));
275 static int hamming_8_4( int a )
311 return -1; // decoding error , not yet corrected
316 // this is not a general function, but it's enough for what we do here
317 static void to_utf8( char * res, uint16_t ch )
323 res[0] = (ch >> 12) | 0xE0;
324 res[1] = ((ch >> 6) & 0x3F) | 0x80;
325 res[2] = (ch & 0x3F) | 0x80;
330 res[0] = (ch >> 6) | 0xC0;
331 res[1] = (ch & 0x3F) | 0x80;
342 static void decode_string( char * res, int res_len,
343 decoder_sys_t *p_sys, int magazine,
344 uint8_t * packet, int len )
350 for ( i = 0; i < len; i++ )
352 int in = bytereverse( packet[i] ) & 0x7f;
358 /* special national characters */
360 out = p_sys->pi_active_national_set[magazine][0];
363 out = p_sys->pi_active_national_set[magazine][1];
366 out = p_sys->pi_active_national_set[magazine][2];
369 out = p_sys->pi_active_national_set[magazine][3];
372 out = p_sys->pi_active_national_set[magazine][4];
375 out = p_sys->pi_active_national_set[magazine][5];
378 out = p_sys->pi_active_national_set[magazine][6];
381 out = p_sys->pi_active_national_set[magazine][7];
384 out = p_sys->pi_active_national_set[magazine][8];
387 out = p_sys->pi_active_national_set[magazine][9];
390 out = p_sys->pi_active_national_set[magazine][10];
393 out = p_sys->pi_active_national_set[magazine][11];
396 out = p_sys->pi_active_national_set[magazine][12];
399 /* some special control characters (empirical) */
401 /* apparently this starts a sequence that ends with 0xb 0xb */
402 while ( i + 1 < len && (bytereverse( packet[i+1] ) & 0x7f) != 0x0b )
409 /* non documented national range 0x08 - 0x0f */
410 if ( in >= 0x08 && in <= 0x0f )
412 out = p_sys->pi_active_national_set[magazine][13 + in - 8];
417 if ( in > 32 && in < 0x7f )
421 /* handle undefined national characters */
425 /* convert to utf-8 */
426 to_utf8( utf8, out );
428 if ( pt + l < res + res_len - 1 )
440 /*****************************************************************************
442 *****************************************************************************/
443 static subpicture_t *Decode( decoder_t *p_dec, block_t **pp_block )
445 decoder_sys_t *p_sys = p_dec->p_sys;
447 subpicture_t *p_spu = NULL;
451 int i_wanted_magazine = i_conf_wanted_page / 100;
452 int i_wanted_page = 0x10 * ((i_conf_wanted_page % 100) / 10)
453 | (i_conf_wanted_page % 10);
454 vlc_bool_t b_update = VLC_FALSE;
455 char psz_text[512], *pt = psz_text;
459 if( pp_block == NULL || *pp_block == NULL ) return NULL;
463 len = p_block->i_buffer;
464 for ( offset = 1; offset + 46 <= len; offset += 46 )
466 uint8_t * packet = (uint8_t *) p_block->p_buffer+offset;
468 /* dbg((p_dec, "header %02x %02x %02x\n", packet[0], packet[1], packet[2])); */
469 /* if (packet[1] != 0x2C) { */
470 /* printf("wrong header\n"); */
475 int mpag = (hamming_8_4( packet[4] ) << 4) | hamming_8_4( packet[5] );
480 dbg((p_dec, "mpag hamming error\n"));
484 row = 0xFF & bytereverse(mpag);
485 magazine = (7 & row) == 0 ? 8 : (7 & row);
488 if ( i_conf_wanted_page && magazine != i_wanted_magazine ) continue;
492 /* row 0 : flags and header line */
497 for ( a = 0; a < 6; a++ )
499 flag |= (0xF & (bytereverse( hamming_8_4(packet[8 + a]) ) >> 4))
503 dbg((p_dec, "mag %d flags %x page %x character set %d subtitles %d\n%s\n", magazine, flag,
505 (0xF0 & bytereverse(hamming_8_4(packet[7]) )) |
506 (0xF & bytereverse(hamming_8_4(packet[6]) )>>4 ),
507 7 & flag>>21, 1 & flag>>15, t));
509 /* if (!b_ignore_sub_flag && !(1 & flag>>15)) */
512 p_sys->i_page = (0xF0 & bytereverse( hamming_8_4(packet[7]) )) |
513 (0xF & (bytereverse( hamming_8_4(packet[6]) ) >> 4) );
515 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
516 packet + 14, 40 - 14 );
518 p_sys->pi_active_national_set[magazine] =
519 ppi_national_subsets[7 & (flag >> 21)];
521 p_sys->b_is_subtitle[magazine] = b_ignore_sub_flag
522 || (1 & (flag >> 15));
524 if ( (i_conf_wanted_page && p_sys->i_page != i_wanted_page)
525 || !p_sys->b_is_subtitle[magazine] )
528 p_sys->b_erase[magazine] = (1 & (flag >> 7));
530 dbg((p_dec, "%ld --> %ld\n", (long int) p_block->i_pts, (long int)(p_sys->prev_pts+1500000)));
532 * we ignore the erase flag if it happens less than 1.5 seconds
533 * before last caption
534 *TODO make this time configurable */
535 if ( p_block->i_pts > p_sys->prev_pts + 1500000 &&
536 p_sys->b_erase[magazine] )
540 dbg((p_dec, "ERASE !\n"));
542 p_sys->b_erase[magazine] = 0;
543 for ( i = 1; i < 32; i++ )
545 if ( !p_sys->ppsz_lines[i][0] ) continue;
546 /* b_update = VLC_TRUE; */
547 p_sys->ppsz_lines[i][0] = 0;
551 /* replace the row if it's different */
552 if ( strcmp(psz_line, p_sys->ppsz_lines[row]) )
554 strncpy( p_sys->ppsz_lines[row], psz_line,
555 sizeof(p_sys->ppsz_lines[row]) - 1);
564 /* row 1-23 : normal lines */
566 if ( (i_conf_wanted_page && p_sys->i_page != i_wanted_page)
567 || !p_sys->b_is_subtitle[magazine] )
570 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
574 /* remove starting spaces */
575 while ( *t == 32 ) t++;
577 /* remove trailing spaces */
578 for ( i = strlen(t) - 1; i >= 0 && t[i] == 32; i-- );
581 /* replace the row if it's different */
582 if ( strcmp( t, p_sys->ppsz_lines[row] ) )
584 strncpy( p_sys->ppsz_lines[row], t,
585 sizeof(p_sys->ppsz_lines[row]) - 1 );
590 p_sys->prev_pts = p_block->i_pts;
592 dbg((p_dec, "%d %d : ", magazine, row));
593 dbg((p_dec, "%s\n", t));
599 for ( i = 0; i < 40; i++ )
601 int in = bytereverse(packet[6 + i]) & 0x7f;
602 sprintf(dbg + strlen(dbg), "%02x ", in);
604 dbg((p_dec, "%s\n", dbg));
606 for ( i = 0; i < 40; i++ )
608 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
610 sprintf( dbg + strlen(dbg), "%s ", psz_line );
612 dbg((p_dec, "%s\n", dbg));
617 else if ( row == 25 )
619 /* row 25 : alternate header line */
620 if ( (i_conf_wanted_page && p_sys->i_page != i_wanted_page)
621 || !p_sys->b_is_subtitle[magazine] )
624 decode_string( psz_line, sizeof(psz_line), p_sys, magazine,
627 /* replace the row if it's different */
628 if ( strcmp( psz_line, p_sys->ppsz_lines[0] ) )
630 strncpy( p_sys->ppsz_lines[0], psz_line,
631 sizeof(p_sys->ppsz_lines[0]) - 1 );
632 /* b_update = VLC_TRUE; */
635 /* else if (row == 26) { */
636 /* // row 26 : TV listings */
638 /* dbg((p_dec, "%d %d : %s\n", magazine, row, decode_string(p_sys, magazine, packet+6, 40))); */
645 for ( i = 1; i < 24; i++ )
647 size_t l = strlen( p_sys->ppsz_lines[i] );
649 if ( l > sizeof(psz_text) - total - 1 )
650 l = sizeof(psz_text) - total - 1;
654 memcpy( pt, p_sys->ppsz_lines[i], l );
657 if ( sizeof(psz_text) - total - 1 > 0 )
666 if ( !strcmp(psz_text, p_sys->psz_prev_text) )
669 dbg((p_dec, "UPDATE TELETEXT PICTURE\n"));
671 assert( sizeof(p_sys->psz_prev_text) >= sizeof(psz_text) );
672 strcpy( p_sys->psz_prev_text, psz_text );
674 /* Create the subpicture unit */
675 p_spu = p_dec->pf_spu_buffer_new( p_dec );
678 msg_Warn( p_dec, "can't get spu buffer" );
682 p_spu->b_pausable = VLC_TRUE;
684 /* Create a new subpicture region */
685 memset( &fmt, 0, sizeof(video_format_t) );
686 fmt.i_chroma = VLC_FOURCC('T','E','X','T');
688 fmt.i_width = fmt.i_height = 0;
689 fmt.i_x_offset = fmt.i_y_offset = 0;
690 p_spu->p_region = p_spu->pf_create_region( VLC_OBJECT(p_dec), &fmt );
691 if( p_spu->p_region == NULL )
693 msg_Err( p_dec, "cannot allocate SPU region" );
697 /* Normal text subs, easy markup */
698 p_spu->i_flags = SUBPICTURE_ALIGN_BOTTOM | p_sys->i_align;
699 p_spu->i_x = p_sys->i_align ? 20 : 0;
702 p_spu->p_region->psz_text = strdup(psz_text);
703 p_spu->i_start = p_block->i_pts;
704 p_spu->i_stop = p_block->i_pts + p_block->i_length;
705 p_spu->b_ephemer = (p_block->i_length == 0);
706 p_spu->b_absolute = VLC_FALSE;
707 dbg((p_dec, "%ld --> %ld\n", (long int) p_block->i_pts/100000, (long int)p_block->i_length/100000));
709 block_Release( p_block );
715 p_dec->pf_spu_buffer_del( p_dec, p_spu );
719 block_Release( p_block );