Fixes #94113

Original commit message from CVS:
Fixes #94113

1 files changed, 156 insertions, 62 deletions

diff --git a/gst/mpegaudioparse/gstmp3types.c b/gst/mpegaudioparse/gstmp3types.c
index 6ea3d6d1..30a4a7ff 100644
--- a/gst/mpegaudioparse/gstmp3types.c
+++ b/gst/mpegaudioparse/gstmp3types.c
@@ -22,9 +22,11 @@
 #include <string.h> /* memcmp */
 
 static GstCaps* mp3_type_find(GstBuffer *buf, gpointer private);
+static GstCaps* mp3_type_find_stream(GstBuffer *buf, gpointer private);
 
 static GstTypeDefinition mp3type_definitions[] = {
   { "mp3types_audio/x-mp3", "audio/x-mp3", ".mp3 .mp2 .mp1 .mpga", mp3_type_find },
+  { "mp3types_stream_audio/x-mp3", "audio/x-mp3", ".mp3 .mp2 .mp1 .mpga", mp3_type_find_stream },
   { NULL, NULL, NULL, NULL },
 };
 
@@ -32,95 +34,187 @@ static GstCaps*
 mp3_type_find(GstBuffer *buf, gpointer private) 
 {
   guint8 *data;
-  gint size, offset;
+  gint size;
   guint32 head;
   GstCaps *caps;
 
   data = GST_BUFFER_DATA (buf);
   size = GST_BUFFER_SIZE (buf);
 
-  GST_DEBUG (0,"mp3typefind: typefinding on buffer of size %d", size);
+  GST_DEBUG (0,"mp3typefind: typefind");
 
   /* gracefully ripped from libid3 */
   if (size >= 3 &&
       data[0] == 'T' && data[1] == 'A' && data[2] == 'G') {
     /* ID V1 tags */
     data += 128;
+    size -= 128;
 
     GST_DEBUG (0, "mp3typefind: detected ID3 Tag V1");
-  } else {
-    if (size >= 10 &&
+  } else if (size >= 10 &&
         (data[0] == 'I' && data[1] == 'D' && data[2] == '3') &&
         data[3] < 0xff && data[4] < 0xff &&
-        data[6] < 0x80 && data[7] < 0x80 && data[8] < 0x80 && data[9] < 0x80) {
-      guint32 skip = 0;
-
-      skip = (skip << 7) | (data[6] & 0x7f);
-      skip = (skip << 7) | (data[7] & 0x7f);
-      skip = (skip << 7) | (data[8] & 0x7f);
-      skip = (skip << 7) | (data[9] & 0x7f);
-
-      if (data[0] == 'I') {
-	/* ID3V2 */
-	/* footer present? */
-	if (data[5] & 0x10)
-          skip += 10;
+        data[6] < 0x80 && data[7] < 0x80 && data[8] < 0x80 && data[9] < 0x80)
+  {
+    guint32 skip = 0;
+
+    skip = (skip << 7) | (data[6] & 0x7f);
+    skip = (skip << 7) | (data[7] & 0x7f);
+    skip = (skip << 7) | (data[8] & 0x7f);
+    skip = (skip << 7) | (data[9] & 0x7f);
+
+    /* include size of header */
+    skip += 10;
+    /* footer present? (only available since version 4) */
+    if (data[3] > 3 && (data[5] & 0x10))
+      skip += 10;
+
+    GST_DEBUG (0, "mp3typefind: detected ID3 Tag V2 with %u bytes", skip);
+    size -= skip;
+    data += skip;      
+  }
 
-	skip += 10;
-      }
+  if (size < 4)
+    return NULL;
+  
+  /* now with the right postion, do typefinding */
+  head = GUINT32_FROM_BE(*((guint32 *)data));
+  if ((head & 0xffe00000) != 0xffe00000)
+    return NULL;
+  if (!((head >> 17) & 3))
+    return NULL;
+  if (((head >> 12) & 0xf) == 0xf)
+    return NULL;
+  if (!((head >> 12) & 0xf))
+    return NULL;
+  if (((head >> 10) & 0x3) == 0x3)
+    return NULL;
 
-      GST_DEBUG (0, "mp3typefind: detected ID3 Tag V2 with %u bytes", skip);
+  caps = gst_caps_new ("mp3_type_find", "audio/x-mp3", NULL);
 
-      /* we currently accept a valid ID3 tag as an mp3 as some ID3 tags have invalid
-       * offsets so the next check might fail */
-      goto success;
-    }
+  return caps;
+}
+static guint mp3types_bitrates[2][3][16] =
+{ { {0, 32, 64, 96, 128, 160, 192, 224, 256, 288, 320, 352, 384, 416, 448, },
+    {0, 32, 48, 56,  64,  80,  96, 112, 128, 160, 192, 224, 256, 320, 384, },
+    {0, 32, 40, 48,  56,  64,  80,  96, 112, 128, 160, 192, 224, 256, 320, } },
+  { {0, 32, 48, 56,  64,  80,  96, 112, 128, 144, 160, 176, 192, 224, 256, },
+    {0,  8, 16, 24,  32,  40,  48,  56,  64,  80,  96, 112, 128, 144, 160, },
+    {0,  8, 16, 24,  32,  40,  48,  56,  64,  80,  96, 112, 128, 144, 160, } },
+};
+static guint mp3types_freqs[3][3] =
+{ {44100, 48000, 32000},
+  {22050, 24000, 16000}, 
+  {11025, 12000,  8000}};
+static inline guint
+mp3_type_frame_length_from_header (guint32 header)
+{
+  guint length;
+  gulong samplerate, bitrate, layer, version;
+
+  /* we don't need extension, mode, copyright, original or emphasis for the frame length */
+  header >>= 9;
+  /* padding */
+  length = header & 0x1;
+  header >>= 1;
+  /* sampling frequency */
+  samplerate = header & 0x3;
+  if (samplerate == 3)
+    return 0;
+  header >>= 2;
+  /* bitrate index */
+  bitrate = header & 0xF;
+  if (bitrate == 15 || bitrate == 0)
+    return 0;
+  /* ignore error correction, too */
+  header >>= 5;
+  /* layer */
+  layer = 4 - (header & 0x3);
+  if (layer == 4)
+    return 0;
+  header >>= 2;
+  /* version */
+  version = header & 0x3;
+  if (version == 1)
+    return 0;
+  /* lookup */
+  bitrate = mp3types_bitrates[version == 3 ? 0 : 1][layer - 1][bitrate];
+  samplerate = mp3types_freqs[version > 0 ? version - 1 : 0][samplerate];
+  /* calculating */
+  if (layer == 1) {
+    length = ((12000 * bitrate / samplerate) + length) * 4;
+  } else {
+    length += ((layer == 3 && version == 0) ? 144000 : 72000) * bitrate / samplerate;
   }
 
-  /* if we don't have enough bytes left to check for a header, die */
-  if (size < 4)
-    return NULL;
+  GST_DEBUG (0, "Calculated mad frame length of %u bytes", length);
+  GST_DEBUG (0, "samplerate = %lu - bitrate = %lu - layer = %lu - version = %lu", samplerate, bitrate, layer, version);
+  return length;
 
-  /* disabled because on a random buffer of 4096 bytes,
-   * this has a chance of 35% or so for false positives.
-   * -> rather not */
-#if 0
-  /* Now search through the entire rest of the buffer for a valid header.
-   * This is a very rough initial cut, it should be checking the length of
-   * the frame and verifying that there is indeed a second frame.  The mp3
-   * header could very well be a false trigger otherwise.
-   */
-  for (offset=0; offset < (size-4); offset++) {
-    head = GUINT32_FROM_BE(*((gulong *)data));
-
-/*  old negative checks
-    if ((head & 0xffe00000) != 0xffe00000)
-      return NULL;
-    if (!((head >> 17) & 3))
-      return NULL;
-    if (((head >> 12) & 0xf) == 0xf)
-      return NULL;
-    if (!((head >> 12) & 0xf))
-      return NULL;
-    if (((head >> 10) & 0x3) == 0x3)
-      return NULL;
-*/
-
-    if ( ((head & 0xffe00000) == 0xffe00000) &&
-         (((head >> 17) & 0x3) != 0x0) &&
-         (((head >> 12) & 0xf) != 0xf) &&
-         (((head >> 12) & 0xf) != 0x0) &&
-         (((head >> 10) & 0x3) != 0x3) )
-      goto success;
+}
+/* increase this value when this function finds too many false positives */
+/**
+ * The chance that random data is identified as a valid mp3 header is 63 / 2^18
+ * (0.024%) per try. This makes the function for calculating false positives
+ *   1 - (1 - ((63 / 2 ^18) ^ GST_MP3_TYPEFIND_MIN_HEADERS)) ^ buffersize)
+ * This has the following probabilities of false positives:
+ * bufsize	          MIN_HEADERS
+ * (bytes)	1	2	3	4
+ * 4096		62.6%	 0.02%	 0%	 0%
+ * 16384	98%	 0.09%	 0%	 0%
+ * 1 MiB       100%	 5.88%	 0%	 0%
+ * 1 GiB       100%    100%	 1.44%   0%
+ * 1 TiB       100%    100%    100%      0.35%
+ * This means that the current choice (3 headers by most of the time 4096 byte
+ * buffers is pretty safe for now.
+ * It is however important to note that in a worst case example a buffer of size
+ *   1440 * GST_MP3_TYPEFIND_MIN_HEADERS + 3
+ * bytes is needed to reliable find the mp3 stream in a buffer when scanning
+ * starts at a random position. This is currently (4323 bytes) slightly above 
+ * the default buffer size. But you rarely hit the worst case - average mp3 
+ * frames are in the 500 bytes range.
+ */
+#define GST_MP3_TYPEFIND_MIN_HEADERS 3
+static GstCaps* 
+mp3_type_find_stream (GstBuffer *buf, gpointer private)
+{
+  guint8 *data;
+  guint size;
+  guint32 head;
+  
+  data = GST_BUFFER_DATA (buf);
+  size = GST_BUFFER_SIZE (buf);
+  
+  while (size >= 4) {
+    head = GUINT32_FROM_BE(*((guint32 *)data));
+    if ((head & 0xffe00000) == 0xffe00000) {
+      guint pos = 0;
+      guint length;
+      guint found = 0; /* number of valid headers found */
+      do {
+        if ((length = mp3_type_frame_length_from_header (head))) {
+	  pos += length;
+          found++;
+	  if (pos + 4 >= size) {
+	    if (found >= GST_MP3_TYPEFIND_MIN_HEADERS) 
+	      goto success;
+	  }
+          head = GUINT32_FROM_BE(*((guint32 *) &(data[pos])));  
+          if ((head & 0xffe00000) != 0xffe00000)
+	    break;
+	} else {
+	  break;
+	}
+      } while (TRUE);
+    }
+    data++;
+    size--;
   }
-#endif
 
   return NULL;
 
 success:
-  caps = gst_caps_new ("mp3_type_find", "audio/x-mp3", NULL);
-
-  return caps;
+  return gst_caps_new ("mp3_type_find", "audio/x-mp3", NULL);
 }
 
 static gboolean