typefinding: fix AC-3 typefinding so that it actually checks for a second frame

Fix typo that made the AC-3 typefinder not actually check for a
second frame, but rather compare the sync point found to itself,
which resulted in the AC-3 typefinder reporting an overly optimistic
MAXIMUM or VERY_LIKELY probability when it found a possible frame
sync.
This commit is contained in:
Tim-Philipp Müller 2010-06-05 16:58:50 +01:00
parent 13f6829497
commit 14d14a9143

View file

@ -1217,8 +1217,10 @@ ac3_type_find (GstTypeFind * tf, gpointer unused)
guint frame_size;
frame_size = ac3_frmsizecod_tbl[frmsizecod].frm_size[fscod];
GST_LOG ("possible frame sync at offset %" G_GUINT64_FORMAT ", size=%u",
c.offset, frame_size);
if (data_scan_ctx_ensure_data (tf, &c_next, (frame_size * 2) + 5)) {
data_scan_ctx_advance (tf, &c, frame_size * 2);
data_scan_ctx_advance (tf, &c_next, frame_size * 2);
if (c_next.data[0] == 0x0b && c_next.data[1] == 0x77) {
guint fscod2 = (c_next.data[4] >> 6) & 0x03;
@ -1227,6 +1229,7 @@ ac3_type_find (GstTypeFind * tf, gpointer unused)
if (fscod == fscod2 && frmsizecod == frmsizecod2) {
GstTypeFindProbability prob;
GST_LOG ("found second frame, looks good");
if (c.offset == 0)
prob = GST_TYPE_FIND_MAXIMUM;
else
@ -1235,6 +1238,8 @@ ac3_type_find (GstTypeFind * tf, gpointer unused)
gst_type_find_suggest (tf, prob, AC3_CAPS);
return;
}
} else {
GST_LOG ("no second frame found, false sync");
}
}
}