1 /*****************************************************************************
2 * subsdec.c : text subtitles decoder
3 *****************************************************************************
4 * Copyright (C) 2000-2001 VideoLAN
7 * Authors: Gildas Bazin <gbazin@videolan.org>
8 * Samuel Hocevar <sam@zoy.org>
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
23 *****************************************************************************/
25 /*****************************************************************************
27 *****************************************************************************/
30 #include <vlc/decoder.h>
33 #include "vlc_filter.h"
35 #if defined(HAVE_ICONV)
41 /*****************************************************************************
42 * decoder_sys_t : decoder descriptor
43 *****************************************************************************/
46 int i_align; /* Subtitles alignment on the vout */
48 #if defined(HAVE_ICONV)
49 iconv_t iconv_handle; /* handle to iconv instance */
54 /*****************************************************************************
56 *****************************************************************************/
57 static int OpenDecoder ( vlc_object_t * );
58 static void CloseDecoder ( vlc_object_t * );
60 static subpicture_t *DecodeBlock ( decoder_t *, block_t ** );
61 static subpicture_t *ParseText ( decoder_t *, block_t * );
62 static void StripTags ( char * );
64 #define DEFAULT_NAME "System Default"
66 /*****************************************************************************
68 *****************************************************************************/
69 #if defined(HAVE_ICONV)
70 static char *ppsz_encodings[] = { DEFAULT_NAME, "ASCII", "UTF-8", "",
71 "ISO-8859-1", "CP1252", "MacRoman", "MacIceland","ISO-8859-15", "",
72 "ISO-8859-2", "CP1250", "MacCentralEurope", "MacCroatian", "MacRomania", "",
73 "ISO-8859-5", "CP1251", "MacCyrillic", "MacUkraine", "KOI8-R", "KOI8-U", "KOI8-RU", "",
74 "ISO-8859-6", "CP1256", "MacArabic", "",
75 "ISO-8859-7", "CP1253", "MacGreek", "",
76 "ISO-8859-8", "CP1255", "MacHebrew", "",
77 "ISO-8859-9", "CP1254", "MacTurkish", "",
78 "ISO-8859-13", "CP1257", "",
79 "ISO-2022-JP", "ISO-2022-JP-1", "ISO-2022-JP-2", "EUC-JP", "SHIFT_JIS", "",
80 "ISO-2022-CN", "ISO-2022-CN-EXT", "EUC-CN", "EUC-TW", "BIG5", "BIG5-HKSCS", "",
81 "ISO-2022-KR", "EUC-KR", "",
82 "MacThai", "KOI8-T", "",
83 "ISO-8859-3", "ISO-8859-4", "ISO-8859-10", "ISO-8859-14", "ISO-8859-16", "",
84 "CP850", "CP862", "CP866", "CP874", "CP932", "CP949", "CP950", "CP1133", "CP1258", "",
86 "UTF-7", "UTF-16", "UTF-16BE", "UTF-16LE", "UTF-32", "UTF-32BE", "UTF-32LE",
87 "C99", "JAVA", "UCS-2", "UCS-2BE", "UCS-2LE", "UCS-4", "UCS-4BE", "UCS-4LE", "",
88 "HZ", "GBK", "GB18030", "JOHAB", "ARMSCII-8",
89 "Georgian-Academy", "Georgian-PS", "TIS-620", "MuleLao-1", "VISCII", "TCVN",
90 "HPROMAN8", "NEXTSTEP" };
93 static int pi_justification[] = { 0, 1, 2 };
94 static char *ppsz_justification_text[] = {N_("Center"),N_("Left"),N_("Right")};
96 #define ENCODING_TEXT N_("Subtitles text encoding")
97 #define ENCODING_LONGTEXT N_("Set the encoding used in text subtitles")
98 #define ALIGN_TEXT N_("Subtitles justification")
99 #define ALIGN_LONGTEXT N_("Set the justification of subtitles")
102 set_description( _("text subtitles decoder") );
103 set_capability( "decoder", 50 );
104 set_callbacks( OpenDecoder, CloseDecoder );
106 add_integer( "subsdec-align", 0, NULL, ALIGN_TEXT, ALIGN_LONGTEXT,
108 change_integer_list( pi_justification, ppsz_justification_text, 0 );
109 #if defined(HAVE_ICONV)
110 add_string( "subsdec-encoding", DEFAULT_NAME, NULL,
111 ENCODING_TEXT, ENCODING_LONGTEXT, VLC_FALSE );
112 change_string_list( ppsz_encodings, 0, 0 );
116 /*****************************************************************************
117 * OpenDecoder: probe the decoder and return score
118 *****************************************************************************
119 * Tries to launch a decoder and return score so that the interface is able
121 *****************************************************************************/
122 static int OpenDecoder( vlc_object_t *p_this )
124 decoder_t *p_dec = (decoder_t*)p_this;
125 decoder_sys_t *p_sys;
128 if( p_dec->fmt_in.i_codec != VLC_FOURCC('s','u','b','t') &&
129 p_dec->fmt_in.i_codec != VLC_FOURCC('s','s','a',' ') )
134 p_dec->pf_decode_sub = DecodeBlock;
136 /* Allocate the memory needed to store the decoder's structure */
137 if( ( p_dec->p_sys = p_sys =
138 (decoder_sys_t *)malloc(sizeof(decoder_sys_t)) ) == NULL )
140 msg_Err( p_dec, "out of memory" );
144 var_Create( p_dec, "subsdec-align", VLC_VAR_INTEGER | VLC_VAR_DOINHERIT );
145 var_Get( p_dec, "subsdec-align", &val );
146 p_sys->i_align = val.i_int;
148 #if defined(HAVE_ICONV)
149 if( p_dec->fmt_in.subs.psz_encoding && *p_dec->fmt_in.subs.psz_encoding )
151 msg_Dbg( p_dec, "using character encoding: %s",
152 p_dec->fmt_in.subs.psz_encoding );
153 p_sys->iconv_handle =
154 iconv_open( "UTF-8", p_dec->fmt_in.subs.psz_encoding );
158 var_Create( p_dec, "subsdec-encoding",
159 VLC_VAR_STRING | VLC_VAR_DOINHERIT );
160 var_Get( p_dec, "subsdec-encoding", &val );
161 if( !strcmp( val.psz_string, DEFAULT_NAME ) )
163 char *psz_charset =(char*)malloc( 100 );
164 vlc_current_charset( &psz_charset );
165 p_sys->iconv_handle = iconv_open( "UTF-8", psz_charset );
166 msg_Dbg( p_dec, "using character encoding: %s", psz_charset );
169 else if( val.psz_string )
171 msg_Dbg( p_dec, "using character encoding: %s", val.psz_string );
172 p_sys->iconv_handle = iconv_open( "UTF-8", val.psz_string );
175 if( p_sys->iconv_handle == (iconv_t)-1 )
177 msg_Warn( p_dec, "unable to do requested conversion" );
180 if( val.psz_string ) free( val.psz_string );
184 msg_Dbg( p_dec, "no iconv support available" );
190 /****************************************************************************
191 * DecodeBlock: the whole thing
192 ****************************************************************************
193 * This function must be fed with complete subtitles units.
194 ****************************************************************************/
195 static subpicture_t *DecodeBlock( decoder_t *p_dec, block_t **pp_block )
199 if( !pp_block || *pp_block == NULL ) return NULL;
201 p_spu = ParseText( p_dec, *pp_block );
203 block_Release( *pp_block );
209 /*****************************************************************************
210 * CloseDecoder: clean up the decoder
211 *****************************************************************************/
212 static void CloseDecoder( vlc_object_t *p_this )
214 decoder_t *p_dec = (decoder_t *)p_this;
215 decoder_sys_t *p_sys = p_dec->p_sys;
217 #if defined(HAVE_ICONV)
218 if( p_sys->iconv_handle != (iconv_t)-1 )
220 iconv_close( p_sys->iconv_handle );
227 /*****************************************************************************
228 * ParseText: parse an text subtitle packet and send it to the video output
229 *****************************************************************************/
230 static subpicture_t *ParseText( decoder_t *p_dec, block_t *p_block )
232 decoder_sys_t *p_sys = p_dec->p_sys;
233 subpicture_t *p_spu = 0;
235 int i_align_h, i_align_v;
238 /* We cannot display a subpicture with no date */
239 if( p_block->i_pts == 0 )
241 msg_Warn( p_dec, "subtitle without a date" );
245 /* Check validity of packet data */
246 if( p_block->i_buffer <= 1 || p_block->p_buffer[0] == '\0' )
248 msg_Warn( p_dec, "empty subtitle" );
252 /* Should be resiliant against bad subtitles */
253 psz_subtitle = strndup( p_block->p_buffer, p_block->i_buffer );
255 i_align_h = p_sys->i_align ? 20 : 0;
258 #if defined(HAVE_ICONV)
259 if( p_sys->iconv_handle != (iconv_t)-1 )
261 char *psz_new_subtitle;
262 char *psz_convert_buffer_out;
263 char *psz_convert_buffer_in;
264 size_t ret, inbytes_left, outbytes_left;
266 psz_new_subtitle = malloc( 6 * strlen( psz_subtitle ) );
267 psz_convert_buffer_out = psz_new_subtitle;
268 psz_convert_buffer_in = psz_subtitle;
269 inbytes_left = strlen( psz_subtitle );
270 outbytes_left = 6 * inbytes_left;
271 ret = iconv( p_sys->iconv_handle, &psz_convert_buffer_in,
272 &inbytes_left, &psz_convert_buffer_out, &outbytes_left );
273 *psz_convert_buffer_out = '\0';
277 msg_Warn( p_dec, "Failed to convert subtitle encoding, "
278 "dropping subtitle.\nTry setting a different "
279 "character-encoding for the subtitle." );
280 free( psz_subtitle );
285 free( psz_subtitle );
286 psz_subtitle = psz_new_subtitle;
291 if( p_dec->fmt_in.i_codec == VLC_FOURCC('s','s','a',' ') )
293 /* Decode SSA strings */
294 /* We expect: ReadOrder, Layer, Style, Name, MarginL, MarginR,
295 * MarginV, Effect, Text */
296 char *psz_new_subtitle;
297 char *psz_buffer_sub;
301 psz_buffer_sub = psz_subtitle;
305 while( i_comma < 8 &&
306 *psz_buffer_sub != '\0' )
308 if( *psz_buffer_sub == ',' )
314 psz_new_subtitle = malloc( strlen( psz_buffer_sub ) + 1);
316 while( psz_buffer_sub[0] != '\0' )
318 if( psz_buffer_sub[0] == '\\' && ( psz_buffer_sub[1] == 'n' ||
319 psz_buffer_sub[1] == 'N' ) )
321 psz_new_subtitle[i_text] = '\n';
325 else if( psz_buffer_sub[0] == '{' &&
326 psz_buffer_sub[1] == '\\' )
328 /* SSA control code */
329 while( psz_buffer_sub[0] != '\0' &&
330 psz_buffer_sub[0] != '}' )
338 psz_new_subtitle[i_text] = psz_buffer_sub[0];
343 psz_new_subtitle[i_text] = '\0';
344 free( psz_subtitle );
345 psz_subtitle = psz_new_subtitle;
350 StripTags( psz_subtitle );
352 p_spu = p_dec->pf_spu_buffer_new( p_dec );
355 msg_Warn( p_dec, "can't get spu buffer" );
356 free( psz_subtitle );
360 /* Create a new subpicture region */
361 memset( &fmt, 0, sizeof(video_format_t) );
362 fmt.i_chroma = VLC_FOURCC('T','E','X','T');
364 fmt.i_width = fmt.i_height = 0;
365 fmt.i_x_offset = fmt.i_y_offset = 0;
366 p_spu->p_region = p_spu->pf_create_region( VLC_OBJECT(p_dec), &fmt );
367 if( !p_spu->p_region )
369 msg_Err( p_dec, "cannot allocate SPU region" );
370 free( psz_subtitle );
371 p_dec->pf_spu_buffer_del( p_dec, p_spu );
375 p_spu->p_region->psz_text = psz_subtitle;
376 p_spu->i_start = p_block->i_pts;
377 p_spu->i_stop = p_block->i_pts + p_block->i_length;
378 p_spu->b_ephemer = (p_block->i_length == 0);
379 p_spu->b_absolute = VLC_FALSE;
381 p_spu->i_flags = OSD_ALIGN_BOTTOM | p_sys->i_align;
382 p_spu->i_x = i_align_h;
383 p_spu->i_y = i_align_v;
388 static void StripTags( char *psz_text )
390 int i_left_moves = 0;
391 vlc_bool_t b_inside_tag = VLC_FALSE;
393 int i_tag_start = -1;
394 while( psz_text[ i ] )
398 if( psz_text[ i ] == '<' )
400 b_inside_tag = VLC_TRUE;
403 psz_text[ i - i_left_moves ] = psz_text[ i ];
407 if( ( psz_text[ i ] == ' ' ) ||
408 ( psz_text[ i ] == '\t' ) ||
409 ( psz_text[ i ] == '\n' ) ||
410 ( psz_text[ i ] == '\r' ) )
412 b_inside_tag = VLC_FALSE;
415 else if( psz_text[ i ] == '>' )
417 i_left_moves += i - i_tag_start + 1;
419 b_inside_tag = VLC_FALSE;
423 psz_text[ i - i_left_moves ] = psz_text[ i ];
428 psz_text[ i - i_left_moves ] = '\0';