1 /* GStreamer AVI GAB2 subtitle parser

     2  * Copyright (C) <2007> Thijs Vermeir <thijsvermeir@gmail.com>

     3  * Copyright (C) <2007> Tim-Philipp Müller <tim centricular net>

     4  *

     5  * This library is free software; you can redistribute it and/or

     6  * modify it under the terms of the GNU Library General Public

     7  * License as published by the Free Software Foundation; either

     8  * version 2 of the License, or (at your option) any later version.

     9  *

    10  * This library is distributed in the hope that it will be useful,

    11  * but WITHOUT ANY WARRANTY; without even the implied warranty of

    12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

    13  * Library General Public License for more details.

    14  *

    15  * You should have received a copy of the GNU Library General Public

    16  * License along with this library; if not, write to the

    17  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,

    18  * Boston, MA 02111-1307, USA.

    19  */

    20 

    21 /**

    22  * SECTION:element-avisubtitle

    23  *

    24  * <refsect2>

    25  * <para>

    26  * Parses the subtitle stream from an avi file.

    27  * </para>

    28  * <title>Example launch line</title>

    29  * <para>

    30  * <programlisting>

    31  * gst-launch filesrc location=subtitle.avi ! avidemux name=demux ! queue ! avisubtitle ! subparse ! textoverlay name=overlay ! ffmpegcolorspace ! autovideosink demux. ! queue ! decodebin ! overlay.

    32  * </programlisting>

    33  * This plays an avi file with a video and subtitle stream.

    34  * </para>

    35  * </refsect2>

    36  *

    37  * Last reviewed on 2008-02-01

    38  */

    39 

    40 /* example of a subtitle chunk in an avi file

    41  * 00000000: 47 41 42 32 00 02 00 10 00 00 00 45 00 6e 00 67  GAB2.......E.n.g

    42  * 00000010: 00 6c 00 69 00 73 00 68 00 00 00 04 00 8e 00 00  .l.i.s.h........

    43  * 00000020: 00 ef bb bf 31 0d 0a 30 30 3a 30 30 3a 30 30 2c  ....1..00:00:00,

    44  * 00000030: 31 30 30 20 2d 2d 3e 20 30 30 3a 30 30 3a 30 32  100 --> 00:00:02

    45  * 00000040: 2c 30 30 30 0d 0a 3c 62 3e 41 6e 20 55 54 46 38  ,000..<b>An UTF8

    46  * 00000050: 20 53 75 62 74 69 74 6c 65 20 77 69 74 68 20 42   Subtitle with B

    47  * 00000060: 4f 4d 3c 2f 62 3e 0d 0a 0d 0a 32 0d 0a 30 30 3a  OM</b>....2..00:

    48  * 00000070: 30 30 3a 30 32 2c 31 30 30 20 2d 2d 3e 20 30 30  00:02,100 --> 00

    49  * 00000080: 3a 30 30 3a 30 34 2c 30 30 30 0d 0a 53 6f 6d 65  :00:04,000..Some

    50  * 00000090: 74 68 69 6e 67 20 6e 6f 6e 41 53 43 49 49 20 2d  thing nonASCII -

    51  * 000000a0: 20 c2 b5 c3 b6 c3 a4 c3 bc c3 9f 0d 0a 0d 0a      ..............

    52  */

    53 

    54 #ifdef HAVE_CONFIG_H

    55 #include "config.h"

    56 #endif

    57 

    58 #include <string.h>

    59 

    60 #include "gstavisubtitle.h"

    61 

    62 GST_DEBUG_CATEGORY_STATIC (avisubtitle_debug);

    63 #define GST_CAT_DEFAULT avisubtitle_debug

    64 

    65 static GstStaticPadTemplate sink_template = GST_STATIC_PAD_TEMPLATE ("sink",

    66     GST_PAD_SINK,

    67     GST_PAD_ALWAYS,

    68     GST_STATIC_CAPS ("application/x-subtitle-avi")

    69     );

    70 

    71 static GstStaticPadTemplate src_template = GST_STATIC_PAD_TEMPLATE ("src",

    72     GST_PAD_SRC,

    73     GST_PAD_ALWAYS,

    74     GST_STATIC_CAPS ("application/x-subtitle")

    75     );

    76 

    77 static void gst_avi_subtitle_title_tag (GstAviSubtitle * sub, gchar * title);

    78 static GstFlowReturn gst_avi_subtitle_chain (GstPad * pad, GstBuffer * buffer);

    79 static GstStateChangeReturn gst_avi_subtitle_change_state (GstElement * element,

    80     GstStateChange transition);

    81 static gboolean gst_avi_subtitle_send_event (GstElement * element,

    82     GstEvent * event);

    83 

    84 GST_BOILERPLATE (GstAviSubtitle, gst_avi_subtitle, GstElement,

    85     GST_TYPE_ELEMENT);

    86 

    87 #define IS_BOM_UTF8(data)     ((GST_READ_UINT32_BE(data) >> 8) == 0xEFBBBF)

    88 #define IS_BOM_UTF16_BE(data) (GST_READ_UINT16_BE(data) == 0xFEFF)

    89 #define IS_BOM_UTF16_LE(data) (GST_READ_UINT16_LE(data) == 0xFEFF)

    90 #define IS_BOM_UTF32_BE(data) (GST_READ_UINT32_BE(data) == 0xFEFF)

    91 #define IS_BOM_UTF32_LE(data) (GST_READ_UINT32_LE(data) == 0xFEFF)

    92 

    93 static GstBuffer *

    94 gst_avi_subtitle_extract_file (GstAviSubtitle * sub, GstBuffer * buffer,

    95     guint offset, guint len)

    96 {

    97   const gchar *input_enc = NULL;

    98   GstBuffer *ret = NULL;

    99   gchar *data;

   100 

   101   data = (gchar *) GST_BUFFER_DATA (buffer) + offset;

   102 

   103   if (len >= (3 + 1) && IS_BOM_UTF8 (data) &&

   104       g_utf8_validate (data + 3, len - 3, NULL)) {

   105     ret = gst_buffer_create_sub (buffer, offset + 3, len - 3);

   106   } else if (len >= 2 && IS_BOM_UTF16_BE (data)) {

   107     input_enc = "UTF-16BE";

   108     data += 2;

   109     len -= 2;

   110   } else if (len >= 2 && IS_BOM_UTF16_LE (data)) {

   111     input_enc = "UTF-16LE";

   112     data += 2;

   113     len -= 2;

   114   } else if (len >= 4 && IS_BOM_UTF32_BE (data)) {

   115     input_enc = "UTF-32BE";

   116     data += 4;

   117     len -= 4;

   118   } else if (len >= 4 && IS_BOM_UTF32_LE (data)) {

   119     input_enc = "UTF-32LE";

   120     data += 4;

   121     len -= 4;

   122   } else if (g_utf8_validate (data, len, NULL)) {

   123     /* not specified, check if it's UTF-8 */

   124     ret = gst_buffer_create_sub (buffer, offset, len);

   125   } else {

   126     /* we could fall back to gst_tag_freeform_to_utf8() here */

   127     GST_WARNING_OBJECT (sub, "unspecified encoding, and not UTF-8");

   128     return NULL;

   129   }

   130 

   131   g_return_val_if_fail (ret != NULL || input_enc != NULL, NULL);

   132 

   133   if (input_enc) {

   134     GError *err = NULL;

   135     gchar *utf8;

   136 

   137     GST_DEBUG_OBJECT (sub, "converting subtitles from %s to UTF-8", input_enc);

   138     utf8 = g_convert (data, len, "UTF-8", input_enc, NULL, NULL, &err);

   139 

   140     if (err != NULL) {

   141       GST_WARNING_OBJECT (sub, "conversion to UTF-8 failed : %s", err->message);

   142       g_error_free (err);

   143       return NULL;

   144     }

   145 

   146     ret = gst_buffer_new ();

   147     GST_BUFFER_DATA (ret) = (guint8 *) utf8;

   148     GST_BUFFER_MALLOCDATA (ret) = (guint8 *) utf8;

   149     GST_BUFFER_SIZE (ret) = strlen (utf8);

   150     GST_BUFFER_OFFSET (ret) = 0;

   151   }

   152 

   153   GST_BUFFER_CAPS (ret) = gst_caps_new_simple ("application/x-subtitle", NULL);

   154   return ret;

   155 }

   156 

   157 /**

   158  * gst_avi_subtitle_title_tag:

   159  * @sub: subtitle element

   160  * @title: the title of this subtitle stream

   161  *

   162  * Send an event to the srcpad of the @sub element with the title

   163  * of the subtitle stream as a GST_TAG_TITLE

   164  */

   165 static void

   166 gst_avi_subtitle_title_tag (GstAviSubtitle * sub, gchar * title)

   167 {

   168   GstTagList *temp_list = gst_tag_list_new ();

   169 

   170   gst_tag_list_add (temp_list, GST_TAG_MERGE_APPEND, GST_TAG_TITLE, title,

   171       NULL);

   172   gst_pad_push_event (sub->src, gst_event_new_tag (temp_list));

   173 }

   174 

   175 static GstFlowReturn

   176 gst_avi_subtitle_parse_gab2_chunk (GstAviSubtitle * sub, GstBuffer * buf)

   177 {

   178   const guint8 *data;

   179   gchar *name_utf8;

   180   guint name_length;

   181   guint file_length;

   182   guint size;

   183 

   184   data = GST_BUFFER_DATA (buf);

   185   size = GST_BUFFER_SIZE (buf);

   186 

   187   /* check the magic word "GAB2\0", and the next word must be 2 */

   188   if (size < 12 || memcmp (data, "GAB2\0\2\0", 5 + 2) != 0)

   189     goto wrong_magic_word;

   190 

   191   /* read 'name' of subtitle */

   192   name_length = GST_READ_UINT32_LE (data + 5 + 2);

   193   GST_LOG_OBJECT (sub, "length of name: %u", name_length);

   194   if (size <= 17 + name_length)

   195     goto wrong_name_length;

   196 

   197   name_utf8 = g_convert ((gchar *) data + 11, name_length, "UTF-8", "UTF-16LE",

   198       NULL, NULL, NULL);

   199 

   200   if (name_utf8) {

   201     GST_LOG_OBJECT (sub, "subtitle name: %s", name_utf8);

   202     gst_avi_subtitle_title_tag (sub, name_utf8);

   203     g_free (name_utf8);

   204   }

   205 

   206   /* next word must be 4 */

   207   if (GST_READ_UINT16_LE (data + 11 + name_length) != 0x4)

   208     goto wrong_fixed_word_2;

   209 

   210   file_length = GST_READ_UINT32_LE (data + 13 + name_length);

   211   GST_LOG_OBJECT (sub, "length srt/ssa file: %u", file_length);

   212 

   213   if (size < (17 + name_length + file_length))

   214     goto wrong_total_length;

   215 

   216   /* store this, so we can send it again after a seek; note that we shouldn't

   217    * assume all the remaining data in the chunk is subtitle data, there may

   218    * be padding at the end for some reason, so only parse file_length bytes */

   219   sub->subfile =

   220       gst_avi_subtitle_extract_file (sub, buf, 17 + name_length, file_length);

   221 

   222   if (sub->subfile == NULL)

   223     goto extract_failed;

   224 

   225   return GST_FLOW_OK;

   226 

   227   /* ERRORS */

   228 wrong_magic_word:

   229   {

   230     GST_ELEMENT_ERROR (sub, STREAM, DECODE, (NULL), ("Wrong magic word"));

   231     return GST_FLOW_ERROR;

   232   }

   233 wrong_name_length:

   234   {

   235     GST_ELEMENT_ERROR (sub, STREAM, DECODE, (NULL),

   236         ("name doesn't fit in buffer (%d < %d)", size, 17 + name_length));

   237     return GST_FLOW_ERROR;

   238   }

   239 wrong_fixed_word_2:

   240   {

   241     GST_ELEMENT_ERROR (sub, STREAM, DECODE, (NULL),

   242         ("wrong fixed word: expected %u, got %u", 4,

   243             GST_READ_UINT16_LE (data + 11 + name_length)));

   244     return GST_FLOW_ERROR;

   245   }

   246 wrong_total_length:

   247   {

   248     GST_ELEMENT_ERROR (sub, STREAM, DECODE, (NULL),

   249         ("buffer size is wrong: need %d bytes, have %d bytes",

   250             17 + name_length + file_length, size));

   251     return GST_FLOW_ERROR;

   252   }

   253 extract_failed:

   254   {

   255     GST_ELEMENT_ERROR (sub, STREAM, DECODE, (NULL),

   256         ("could not extract subtitles"));

   257     return GST_FLOW_ERROR;

   258   }

   259 }

   260 

   261 static GstFlowReturn

   262 gst_avi_subtitle_chain (GstPad * pad, GstBuffer * buffer)

   263 {

   264   GstAviSubtitle *sub = GST_AVI_SUBTITLE (GST_PAD_PARENT (pad));

   265   GstFlowReturn ret;

   266 

   267   if (sub->subfile != NULL) {

   268     GST_WARNING_OBJECT (sub, "Got more buffers than expected, dropping");

   269     ret = GST_FLOW_UNEXPECTED;

   270     goto done;

   271   }

   272 

   273   /* we expect exactly one buffer with the whole srt/ssa file in it */

   274   ret = gst_avi_subtitle_parse_gab2_chunk (sub, buffer);

   275   if (ret != GST_FLOW_OK)

   276     goto done;

   277 

   278   /* now push the subtitle data downstream */

   279   ret = gst_pad_push (sub->src, gst_buffer_ref (sub->subfile));

   280 

   281 done:

   282 

   283   gst_buffer_unref (buffer);

   284   return ret;

   285 }

   286 

   287 static gboolean

   288 gst_avi_subtitle_send_event (GstElement * element, GstEvent * event)

   289 {

   290   GstAviSubtitle *avisubtitle = GST_AVI_SUBTITLE (element);

   291   gboolean ret = FALSE;

   292 

   293   if (GST_EVENT_TYPE (event) == GST_EVENT_SEEK) {

   294     if (avisubtitle->subfile) {

   295       if (gst_pad_push (avisubtitle->src,

   296               gst_buffer_ref (avisubtitle->subfile)) == GST_FLOW_OK)

   297         ret = TRUE;

   298     }

   299   }

   300   gst_event_unref (event);

   301   return ret;

   302 }

   303 

   304 static void

   305 gst_avi_subtitle_base_init (gpointer klass)

   306 {

   307   GstElementClass *element_class = GST_ELEMENT_CLASS (klass);

   308 

   309   GST_DEBUG_CATEGORY_INIT (avisubtitle_debug, "avisubtitle", 0,

   310       "parse avi subtitle stream");

   311 

   312   gst_element_class_add_pad_template (element_class,

   313       gst_static_pad_template_get (&sink_template));

   314   gst_element_class_add_pad_template (element_class,

   315       gst_static_pad_template_get (&src_template));

   316 

   317   gst_element_class_set_details_simple (element_class,

   318       "Avi subtitle parser", "Codec/Parser/Subtitle",

   319       "Parse avi subtitle stream", "Thijs Vermeir <thijsvermeir@gmail.com>");

   320 }

   321 

   322 static void

   323 gst_avi_subtitle_class_init (GstAviSubtitleClass * klass)

   324 {

   325   GstElementClass *gstelement_class = (GstElementClass *) klass;

   326 

   327   gstelement_class->change_state =

   328       GST_DEBUG_FUNCPTR (gst_avi_subtitle_change_state);

   329   gstelement_class->send_event =

   330       GST_DEBUG_FUNCPTR (gst_avi_subtitle_send_event);

   331 }

   332 

   333 static void

   334 gst_avi_subtitle_init (GstAviSubtitle * self, GstAviSubtitleClass * klass)

   335 {

   336   self->src = gst_pad_new_from_static_template (&src_template, "src");

   337   gst_element_add_pad (GST_ELEMENT (self), self->src);

   338 

   339   self->sink = gst_pad_new_from_static_template (&sink_template, "sink");

   340   gst_pad_set_chain_function (self->sink,

   341       GST_DEBUG_FUNCPTR (gst_avi_subtitle_chain));

   342   gst_element_add_pad (GST_ELEMENT (self), self->sink);

   343 

   344   self->subfile = NULL;

   345 }

   346 

   347 static GstStateChangeReturn

   348 gst_avi_subtitle_change_state (GstElement * element, GstStateChange transition)

   349 {

   350   GstStateChangeReturn ret;

   351   GstAviSubtitle *sub = GST_AVI_SUBTITLE (element);

   352 

   353   switch (transition) {

   354     case GST_STATE_CHANGE_NULL_TO_READY:

   355     case GST_STATE_CHANGE_READY_TO_PAUSED:

   356     default:

   357       break;

   358   }

   359 

   360   ret = GST_ELEMENT_CLASS (parent_class)->change_state (element, transition);

   361   if (ret == GST_STATE_CHANGE_FAILURE)

   362     return ret;

   363 

   364   switch (transition) {

   365     case GST_STATE_CHANGE_PAUSED_TO_READY:

   366       if (sub->subfile) {

   367         gst_buffer_unref (sub->subfile);

   368         sub->subfile = NULL;

   369       }

   370       break;

   371     default:

   372       break;

   373   }

   374 

   375   return ret;

   376 }