1 /*****************************************************************************
2 * subsdec.c : text subtitles decoder
3 *****************************************************************************
4 * Copyright (C) 2000-2001 VideoLAN
7 * Authors: Gildas Bazin <gbazin@netcourrier.com>
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>
34 #if defined(HAVE_ICONV)
40 /*****************************************************************************
41 * decoder_sys_t : decoder descriptor
42 *****************************************************************************/
45 int i_align; /* Subtitles alignment on the vout */
46 int i_subpic_channel; /* Subpic channel for subtitles */
48 vout_thread_t *p_vout; /* last vout used */
50 #if defined(HAVE_ICONV)
51 iconv_t iconv_handle; /* handle to iconv instance */
55 /*****************************************************************************
57 *****************************************************************************/
58 static int OpenDecoder ( vlc_object_t * );
59 static void CloseDecoder ( vlc_object_t * );
61 static void DecodeBlock ( decoder_t *, block_t ** );
63 static void ParseText ( decoder_t *, block_t *, vout_thread_t * );
64 static void StripTags ( char * );
66 #define DEFAULT_NAME "System Default"
68 /*****************************************************************************
70 *****************************************************************************/
71 #if defined(HAVE_ICONV)
72 static char *ppsz_encodings[] = { DEFAULT_NAME, "ASCII", "UTF-8", "",
73 "ISO-8859-1", "CP1252", "MacRoman", "MacIceland","ISO-8859-15", "",
74 "ISO-8859-2", "CP1250", "MacCentralEurope", "MacCroatian", "MacRomania", "",
75 "ISO-8859-5", "CP1251", "MacCyrillic", "MacUkraine", "KOI8-R", "KOI8-U", "KOI8-RU", "",
76 "ISO-8859-6", "CP1256", "MacArabic", "",
77 "ISO-8859-7", "CP1253", "MacGreek", "",
78 "ISO-8859-8", "CP1255", "MacHebrew", "",
79 "ISO-8859-9", "CP1254", "MacTurkish", "",
80 "ISO-8859-13", "CP1257", "",
81 "ISO-2022-JP", "ISO-2022-JP-1", "ISO-2022-JP-2", "EUC-JP", "SHIFT_JIS", "",
82 "ISO-2022-CN", "ISO-2022-CN-EXT", "EUC-CN", "EUC-TW", "BIG5", "BIG5-HKSCS", "",
83 "ISO-2022-KR", "EUC-KR", "",
84 "MacThai", "KOI8-T", "",
85 "ISO-8859-3", "ISO-8859-4", "ISO-8859-10", "ISO-8859-14", "ISO-8859-16", "",
86 "CP850", "CP862", "CP866", "CP874", "CP932", "CP949", "CP950", "CP1133", "CP1258", "",
88 "UTF-7", "UTF-16", "UTF-16BE", "UTF-16LE", "UTF-32", "UTF-32BE", "UTF-32LE",
89 "C99", "JAVA", "UCS-2", "UCS-2BE", "UCS-2LE", "UCS-4", "UCS-4BE", "UCS-4LE", "",
90 "HZ", "GBK", "GB18030", "JOHAB", "ARMSCII-8",
91 "Georgian-Academy", "Georgian-PS", "TIS-620", "MuleLao-1", "VISCII", "TCVN",
92 "HPROMAN8", "NEXTSTEP" };
95 static int pi_justification[] = { 0, 1, 2 };
96 static char *ppsz_justification_text[] = {N_("Center"),N_("Left"),N_("Right")};
98 #define ENCODING_TEXT N_("Subtitles text encoding")
99 #define ENCODING_LONGTEXT N_("Set the encoding used in text subtitles")
100 #define ALIGN_TEXT N_("Subtitles justification")
101 #define ALIGN_LONGTEXT N_("Set the justification of subtitles")
104 set_description( _("text subtitles decoder") );
105 set_capability( "decoder", 50 );
106 set_callbacks( OpenDecoder, CloseDecoder );
108 add_integer( "subsdec-align", 0, NULL, ALIGN_TEXT, ALIGN_LONGTEXT,
110 change_integer_list( pi_justification, ppsz_justification_text, 0 );
111 #if defined(HAVE_ICONV)
112 add_string( "subsdec-encoding", DEFAULT_NAME, NULL,
113 ENCODING_TEXT, ENCODING_LONGTEXT, VLC_FALSE );
114 change_string_list( ppsz_encodings, 0, 0 );
118 /*****************************************************************************
119 * OpenDecoder: probe the decoder and return score
120 *****************************************************************************
121 * Tries to launch a decoder and return score so that the interface is able
123 *****************************************************************************/
124 static int OpenDecoder( vlc_object_t *p_this )
126 decoder_t *p_dec = (decoder_t*)p_this;
127 decoder_sys_t *p_sys;
130 if( p_dec->fmt_in.i_codec != VLC_FOURCC('s','u','b','t') &&
131 p_dec->fmt_in.i_codec != VLC_FOURCC('s','s','a',' ') )
136 p_dec->pf_decode_sub = DecodeBlock;
138 /* Allocate the memory needed to store the decoder's structure */
139 if( ( p_dec->p_sys = p_sys =
140 (decoder_sys_t *)malloc(sizeof(decoder_sys_t)) ) == NULL )
142 msg_Err( p_dec, "out of memory" );
146 var_Create( p_dec, "subsdec-align", VLC_VAR_INTEGER | VLC_VAR_DOINHERIT );
147 var_Get( p_dec, "subsdec-align", &val );
148 p_sys->i_align = val.i_int;
150 #if defined(HAVE_ICONV)
151 if( p_dec->fmt_in.subs.psz_encoding && *p_dec->fmt_in.subs.psz_encoding )
153 msg_Dbg( p_dec, "using character encoding: %s",
154 p_dec->fmt_in.subs.psz_encoding );
155 p_sys->iconv_handle = iconv_open( "UTF-8",
156 p_dec->fmt_in.subs.psz_encoding );
160 var_Create( p_dec, "subsdec-encoding",
161 VLC_VAR_STRING | VLC_VAR_DOINHERIT );
162 var_Get( p_dec, "subsdec-encoding", &val );
163 if( !strcmp( val.psz_string, DEFAULT_NAME ) )
165 char *psz_charset =(char*)malloc( 100 );
166 vlc_current_charset( &psz_charset );
167 p_sys->iconv_handle = iconv_open( "UTF-8", psz_charset );
168 msg_Dbg( p_dec, "using character encoding: %s", psz_charset );
171 else if( val.psz_string )
173 msg_Dbg( p_dec, "using character encoding: %s", val.psz_string );
174 p_sys->iconv_handle = iconv_open( "UTF-8", val.psz_string );
177 if( p_sys->iconv_handle == (iconv_t)-1 )
179 msg_Warn( p_dec, "unable to do requested conversion" );
182 if( val.psz_string ) free( val.psz_string );
186 msg_Dbg( p_dec, "no iconv support available" );
189 p_dec->p_sys->p_vout = NULL;
194 /****************************************************************************
195 * DecodeBlock: the whole thing
196 ****************************************************************************
197 * This function must be fed with complete subtitles units.
198 ****************************************************************************/
199 static void DecodeBlock( decoder_t *p_dec, block_t **pp_block )
201 vout_thread_t *p_vout;
203 if( !pp_block || *pp_block == NULL )
208 /* Here we are dealing with text subtitles */
209 p_vout = vlc_object_find( p_dec, VLC_OBJECT_VOUT, FIND_ANYWHERE );
212 if( p_dec->p_sys->p_vout != p_vout )
214 p_dec->p_sys->i_subpic_channel = vout_RegisterOSDChannel( p_vout );
216 ParseText( p_dec, *pp_block, p_vout );
217 vlc_object_release( p_vout );
221 msg_Warn( p_dec, "couldn't find a video output, trashing subtitle" );
223 p_dec->p_sys->p_vout = p_vout;
225 block_Release( *pp_block );
229 /*****************************************************************************
230 * CloseDecoder: clean up the decoder
231 *****************************************************************************/
232 static void CloseDecoder( vlc_object_t *p_this )
234 decoder_t *p_dec = (decoder_t *)p_this;
235 decoder_sys_t *p_sys = p_dec->p_sys;
236 vout_thread_t *p_vout;
238 p_vout = vlc_object_find( p_dec, VLC_OBJECT_VOUT, FIND_ANYWHERE );
239 if( p_vout != NULL && p_vout->p_subpicture != NULL )
241 subpicture_t *p_subpic;
244 for( i_subpic = 0; i_subpic < VOUT_MAX_SUBPICTURES; i_subpic++ )
246 p_subpic = &p_vout->p_subpicture[i_subpic];
248 if( p_subpic != NULL &&
249 ( p_subpic->i_status == RESERVED_SUBPICTURE
250 || p_subpic->i_status == READY_SUBPICTURE ) )
252 vout_DestroySubPicture( p_vout, p_subpic );
256 if( p_vout ) vlc_object_release( p_vout );
258 #if defined(HAVE_ICONV)
259 if( p_sys->iconv_handle != (iconv_t)-1 )
261 iconv_close( p_sys->iconv_handle );
268 /*****************************************************************************
269 * ParseText: parse an text subtitle packet and send it to the video output
270 *****************************************************************************/
271 static void ParseText( decoder_t *p_dec, block_t *p_block,
272 vout_thread_t *p_vout )
274 decoder_sys_t *p_sys = p_dec->p_sys;
276 int i_align_h, i_align_v;
278 /* We cannot display a subpicture with no date */
279 if( p_block->i_pts == 0 )
281 msg_Warn( p_dec, "subtitle without a date" );
285 /* Check validity of packet data */
286 if( p_block->i_buffer <= 1 || p_block->p_buffer[0] == '\0' )
288 msg_Warn( p_dec, "empty subtitle" );
292 /* Should be resiliant against bad subtitles */
293 psz_subtitle = strndup( p_block->p_buffer, p_block->i_buffer );
295 i_align_h = p_sys->i_align ? 20 : 0;
298 #if defined(HAVE_ICONV)
299 if( p_sys->iconv_handle != (iconv_t)-1 )
301 char *psz_new_subtitle;
302 char *psz_convert_buffer_out;
303 char *psz_convert_buffer_in;
304 size_t ret, inbytes_left, outbytes_left;
306 psz_new_subtitle = malloc( 6 * strlen( psz_subtitle ) );
307 psz_convert_buffer_out = psz_new_subtitle;
308 psz_convert_buffer_in = psz_subtitle;
309 inbytes_left = strlen( psz_subtitle );
310 outbytes_left = 6 * inbytes_left;
311 ret = iconv( p_sys->iconv_handle, &psz_convert_buffer_in,
312 &inbytes_left, &psz_convert_buffer_out, &outbytes_left );
313 *psz_convert_buffer_out = '\0';
317 msg_Warn( p_dec, "Failed to convert subtitle encoding, dropping subtitle.\nTry setting a different character-encoding for the subtitle." );
318 free( psz_subtitle );
323 free( psz_subtitle );
324 psz_subtitle = psz_new_subtitle;
329 if( p_dec->fmt_in.i_codec == VLC_FOURCC('s','s','a',' ') )
331 /* Decode SSA strings */
332 /* We expect: ReadOrder, Layer, Style, Name, MarginL, MarginR, MarginV, Effect, Text */
333 char *psz_new_subtitle;
334 char *psz_buffer_sub;
338 psz_buffer_sub = psz_subtitle;
342 while( i_comma < 8 &&
343 *psz_buffer_sub != '\0' )
345 if( *psz_buffer_sub == ',' )
351 psz_new_subtitle = malloc( strlen( psz_buffer_sub ) + 1);
353 while( psz_buffer_sub[0] != '\0' )
355 if( psz_buffer_sub[0] == '\\' && ( psz_buffer_sub[1] =='n' || psz_buffer_sub[1] =='N' ) )
357 psz_new_subtitle[i_text] = '\n';
361 else if( psz_buffer_sub[0] == '{' && psz_buffer_sub[1] == '\\' )
363 /* SSA control code */
364 while( psz_buffer_sub[0] != '\0' && psz_buffer_sub[0] != '}' )
372 psz_new_subtitle[i_text] = psz_buffer_sub[0];
377 psz_new_subtitle[i_text] = '\0';
378 free( psz_subtitle );
379 psz_subtitle = psz_new_subtitle;
383 StripTags( psz_subtitle );
384 vout_ShowTextAbsolute( p_vout, p_sys->i_subpic_channel, psz_subtitle, NULL,
385 OSD_ALIGN_BOTTOM | p_sys->i_align, i_align_h,
386 i_align_v, p_block->i_pts,
387 p_block->i_length ? p_block->i_pts + p_block->i_length : 0 );
389 free( psz_subtitle );
392 static void StripTags( char *psz_text )
394 int i_left_moves = 0;
395 vlc_bool_t b_inside_tag = VLC_FALSE;
397 int i_tag_start = -1;
398 while( psz_text[ i ] )
402 if( psz_text[ i ] == '<' )
404 b_inside_tag = VLC_TRUE;
407 psz_text[ i - i_left_moves ] = psz_text[ i ];
411 if( ( psz_text[ i ] == ' ' ) ||
412 ( psz_text[ i ] == '\t' ) ||
413 ( psz_text[ i ] == '\n' ) ||
414 ( psz_text[ i ] == '\r' ) )
416 b_inside_tag = VLC_FALSE;
419 else if( psz_text[ i ] == '>' )
421 i_left_moves += i - i_tag_start + 1;
423 b_inside_tag = VLC_FALSE;
427 psz_text[ i - i_left_moves ] = psz_text[ i ];
432 psz_text[ i - i_left_moves ] = '\0';