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"
37 /*****************************************************************************
38 * decoder_sys_t : decoder descriptor
39 *****************************************************************************/
42 int i_align; /* Subtitles alignment on the vout */
43 vlc_iconv_t iconv_handle; /* handle to iconv instance */
46 /*****************************************************************************
48 *****************************************************************************/
49 static int OpenDecoder ( vlc_object_t * );
50 static void CloseDecoder ( vlc_object_t * );
52 static subpicture_t *DecodeBlock ( decoder_t *, block_t ** );
53 static subpicture_t *ParseText ( decoder_t *, block_t * );
54 static void StripTags ( char * );
56 #define DEFAULT_NAME "System Default"
58 /*****************************************************************************
60 *****************************************************************************/
61 static char *ppsz_encodings[] = { DEFAULT_NAME, "ASCII", "UTF-8", "",
62 "ISO-8859-1", "CP1252", "MacRoman", "MacIceland","ISO-8859-15", "",
63 "ISO-8859-2", "CP1250", "MacCentralEurope", "MacCroatian", "MacRomania", "",
64 "ISO-8859-5", "CP1251", "MacCyrillic", "MacUkraine", "KOI8-R", "KOI8-U", "KOI8-RU", "",
65 "ISO-8859-6", "CP1256", "MacArabic", "",
66 "ISO-8859-7", "CP1253", "MacGreek", "",
67 "ISO-8859-8", "CP1255", "MacHebrew", "",
68 "ISO-8859-9", "CP1254", "MacTurkish", "",
69 "ISO-8859-13", "CP1257", "",
70 "ISO-2022-JP", "ISO-2022-JP-1", "ISO-2022-JP-2", "EUC-JP", "SHIFT_JIS", "",
71 "ISO-2022-CN", "ISO-2022-CN-EXT", "EUC-CN", "EUC-TW", "BIG5", "BIG5-HKSCS", "",
72 "ISO-2022-KR", "EUC-KR", "",
73 "MacThai", "KOI8-T", "",
74 "ISO-8859-3", "ISO-8859-4", "ISO-8859-10", "ISO-8859-14", "ISO-8859-16", "",
75 "CP850", "CP862", "CP866", "CP874", "CP932", "CP949", "CP950", "CP1133", "CP1258", "",
77 "UTF-7", "UTF-16", "UTF-16BE", "UTF-16LE", "UTF-32", "UTF-32BE", "UTF-32LE",
78 "C99", "JAVA", "UCS-2", "UCS-2BE", "UCS-2LE", "UCS-4", "UCS-4BE", "UCS-4LE", "",
79 "HZ", "GBK", "GB18030", "JOHAB", "ARMSCII-8",
80 "Georgian-Academy", "Georgian-PS", "TIS-620", "MuleLao-1", "VISCII", "TCVN",
81 "HPROMAN8", "NEXTSTEP" };
83 static int pi_justification[] = { 0, 1, 2 };
84 static char *ppsz_justification_text[] = {N_("Center"),N_("Left"),N_("Right")};
86 #define ENCODING_TEXT N_("Subtitles text encoding")
87 #define ENCODING_LONGTEXT N_("Set the encoding used in text subtitles")
88 #define ALIGN_TEXT N_("Subtitles justification")
89 #define ALIGN_LONGTEXT N_("Set the justification of subtitles")
92 set_description( _("text subtitles decoder") );
93 set_capability( "decoder", 50 );
94 set_callbacks( OpenDecoder, CloseDecoder );
95 set_category( CAT_INPUT );
96 set_subcategory( SUBCAT_INPUT_SCODEC );
98 add_integer( "subsdec-align", 0, NULL, ALIGN_TEXT, ALIGN_LONGTEXT,
100 change_integer_list( pi_justification, ppsz_justification_text, 0 );
101 add_string( "subsdec-encoding", DEFAULT_NAME, NULL,
102 ENCODING_TEXT, ENCODING_LONGTEXT, VLC_FALSE );
103 change_string_list( ppsz_encodings, 0, 0 );
106 /*****************************************************************************
107 * OpenDecoder: probe the decoder and return score
108 *****************************************************************************
109 * Tries to launch a decoder and return score so that the interface is able
111 *****************************************************************************/
112 static int OpenDecoder( vlc_object_t *p_this )
114 decoder_t *p_dec = (decoder_t*)p_this;
115 decoder_sys_t *p_sys;
118 if( p_dec->fmt_in.i_codec != VLC_FOURCC('s','u','b','t') &&
119 p_dec->fmt_in.i_codec != VLC_FOURCC('s','s','a',' ') )
124 p_dec->pf_decode_sub = DecodeBlock;
126 /* Allocate the memory needed to store the decoder's structure */
127 if( ( p_dec->p_sys = p_sys =
128 (decoder_sys_t *)malloc(sizeof(decoder_sys_t)) ) == NULL )
130 msg_Err( p_dec, "out of memory" );
134 var_Create( p_dec, "subsdec-align", VLC_VAR_INTEGER | VLC_VAR_DOINHERIT );
135 var_Get( p_dec, "subsdec-align", &val );
136 p_sys->i_align = val.i_int;
138 if( p_dec->fmt_in.subs.psz_encoding && *p_dec->fmt_in.subs.psz_encoding )
140 msg_Dbg( p_dec, "using character encoding: %s",
141 p_dec->fmt_in.subs.psz_encoding );
142 p_sys->iconv_handle =
143 vlc_iconv_open( "UTF-8", p_dec->fmt_in.subs.psz_encoding );
147 var_Create( p_dec, "subsdec-encoding",
148 VLC_VAR_STRING | VLC_VAR_DOINHERIT );
149 var_Get( p_dec, "subsdec-encoding", &val );
150 if( !strcmp( val.psz_string, DEFAULT_NAME ) )
152 char *psz_charset =(char*)malloc( 100 );
153 vlc_current_charset( &psz_charset );
154 p_sys->iconv_handle = vlc_iconv_open( "UTF-8", psz_charset );
155 msg_Dbg( p_dec, "using character encoding: %s", psz_charset );
158 else if( val.psz_string )
160 msg_Dbg( p_dec, "using character encoding: %s", val.psz_string );
161 p_sys->iconv_handle = vlc_iconv_open( "UTF-8", val.psz_string );
164 if( p_sys->iconv_handle == (vlc_iconv_t)-1 )
166 msg_Warn( p_dec, "unable to do requested conversion" );
169 if( val.psz_string ) free( val.psz_string );
175 /****************************************************************************
176 * DecodeBlock: the whole thing
177 ****************************************************************************
178 * This function must be fed with complete subtitles units.
179 ****************************************************************************/
180 static subpicture_t *DecodeBlock( decoder_t *p_dec, block_t **pp_block )
184 if( !pp_block || *pp_block == NULL ) return NULL;
186 p_spu = ParseText( p_dec, *pp_block );
188 block_Release( *pp_block );
194 /*****************************************************************************
195 * CloseDecoder: clean up the decoder
196 *****************************************************************************/
197 static void CloseDecoder( vlc_object_t *p_this )
199 decoder_t *p_dec = (decoder_t *)p_this;
200 decoder_sys_t *p_sys = p_dec->p_sys;
202 if( p_sys->iconv_handle != (vlc_iconv_t)-1 )
204 vlc_iconv_close( p_sys->iconv_handle );
210 /*****************************************************************************
211 * ParseText: parse an text subtitle packet and send it to the video output
212 *****************************************************************************/
213 static subpicture_t *ParseText( decoder_t *p_dec, block_t *p_block )
215 decoder_sys_t *p_sys = p_dec->p_sys;
216 subpicture_t *p_spu = 0;
218 int i_align_h, i_align_v;
221 /* We cannot display a subpicture with no date */
222 if( p_block->i_pts == 0 )
224 msg_Warn( p_dec, "subtitle without a date" );
228 /* Check validity of packet data */
229 if( p_block->i_buffer <= 1 || p_block->p_buffer[0] == '\0' )
231 msg_Warn( p_dec, "empty subtitle" );
235 /* Should be resiliant against bad subtitles */
236 psz_subtitle = strndup( p_block->p_buffer, p_block->i_buffer );
238 i_align_h = p_sys->i_align ? 20 : 0;
241 if( p_sys->iconv_handle != (vlc_iconv_t)-1 )
243 char *psz_new_subtitle;
244 char *psz_convert_buffer_out;
245 char *psz_convert_buffer_in;
246 size_t ret, inbytes_left, outbytes_left;
248 psz_new_subtitle = malloc( 6 * strlen( psz_subtitle ) );
249 psz_convert_buffer_out = psz_new_subtitle;
250 psz_convert_buffer_in = psz_subtitle;
251 inbytes_left = strlen( psz_subtitle );
252 outbytes_left = 6 * inbytes_left;
253 ret = vlc_iconv( p_sys->iconv_handle, &psz_convert_buffer_in,
254 &inbytes_left, &psz_convert_buffer_out,
256 *psz_convert_buffer_out = '\0';
260 msg_Warn( p_dec, "Failed to convert subtitle encoding, "
261 "dropping subtitle.\nTry setting a different "
262 "character-encoding for the subtitle." );
263 free( psz_subtitle );
268 free( psz_subtitle );
269 psz_subtitle = psz_new_subtitle;
273 if( p_dec->fmt_in.i_codec == VLC_FOURCC('s','s','a',' ') )
275 /* Decode SSA strings */
276 /* We expect: ReadOrder, Layer, Style, Name, MarginL, MarginR,
277 * MarginV, Effect, Text */
278 char *psz_new_subtitle;
279 char *psz_buffer_sub;
283 psz_buffer_sub = psz_subtitle;
287 while( i_comma < 8 &&
288 *psz_buffer_sub != '\0' )
290 if( *psz_buffer_sub == ',' )
296 psz_new_subtitle = malloc( strlen( psz_buffer_sub ) + 1);
298 while( psz_buffer_sub[0] != '\0' )
300 if( psz_buffer_sub[0] == '\\' && ( psz_buffer_sub[1] == 'n' ||
301 psz_buffer_sub[1] == 'N' ) )
303 psz_new_subtitle[i_text] = '\n';
307 else if( psz_buffer_sub[0] == '{' &&
308 psz_buffer_sub[1] == '\\' )
310 /* SSA control code */
311 while( psz_buffer_sub[0] != '\0' &&
312 psz_buffer_sub[0] != '}' )
320 psz_new_subtitle[i_text] = psz_buffer_sub[0];
325 psz_new_subtitle[i_text] = '\0';
326 free( psz_subtitle );
327 psz_subtitle = psz_new_subtitle;
332 StripTags( psz_subtitle );
334 p_spu = p_dec->pf_spu_buffer_new( p_dec );
337 msg_Warn( p_dec, "can't get spu buffer" );
338 free( psz_subtitle );
342 /* Create a new subpicture region */
343 memset( &fmt, 0, sizeof(video_format_t) );
344 fmt.i_chroma = VLC_FOURCC('T','E','X','T');
346 fmt.i_width = fmt.i_height = 0;
347 fmt.i_x_offset = fmt.i_y_offset = 0;
348 p_spu->p_region = p_spu->pf_create_region( VLC_OBJECT(p_dec), &fmt );
349 if( !p_spu->p_region )
351 msg_Err( p_dec, "cannot allocate SPU region" );
352 free( psz_subtitle );
353 p_dec->pf_spu_buffer_del( p_dec, p_spu );
357 p_spu->p_region->psz_text = psz_subtitle;
358 p_spu->i_start = p_block->i_pts;
359 p_spu->i_stop = p_block->i_pts + p_block->i_length;
360 p_spu->b_ephemer = (p_block->i_length == 0);
361 p_spu->b_absolute = VLC_FALSE;
363 p_spu->i_flags = OSD_ALIGN_BOTTOM | p_sys->i_align;
364 p_spu->i_x = i_align_h;
365 p_spu->i_y = i_align_v;
370 static void StripTags( char *psz_text )
372 int i_left_moves = 0;
373 vlc_bool_t b_inside_tag = VLC_FALSE;
375 int i_tag_start = -1;
376 while( psz_text[ i ] )
380 if( psz_text[ i ] == '<' )
382 b_inside_tag = VLC_TRUE;
385 psz_text[ i - i_left_moves ] = psz_text[ i ];
389 if( ( psz_text[ i ] == ' ' ) ||
390 ( psz_text[ i ] == '\t' ) ||
391 ( psz_text[ i ] == '\n' ) ||
392 ( psz_text[ i ] == '\r' ) )
394 b_inside_tag = VLC_FALSE;
397 else if( psz_text[ i ] == '>' )
399 i_left_moves += i - i_tag_start + 1;
401 b_inside_tag = VLC_FALSE;
405 psz_text[ i - i_left_moves ] = psz_text[ i ];
410 psz_text[ i - i_left_moves ] = '\0';