diff --git a/gst/isomp4/qtdemux.c b/gst/isomp4/qtdemux.c
index 7df6259706..f2ea73b6c4 100644
--- a/gst/isomp4/qtdemux.c
+++ b/gst/isomp4/qtdemux.c
@@ -10493,6 +10493,11 @@ gst_qtdemux_handle_esds (GstQTDemux * qtdemux, QtDemuxStream * stream,
       caps = gst_caps_new_simple ("audio/x-ac3",
           "framed", G_TYPE_BOOLEAN, TRUE, NULL);
       break;
+    case 0xA9:                 /* AC3 */
+      codec_name = "DTS audio";
+      caps = gst_caps_new_simple ("audio/x-dts",
+          "framed", G_TYPE_BOOLEAN, TRUE, NULL);
+      break;
     case 0xE1:                 /* QCELP */
       /* QCELP, the codec_data is a riff tag (little endian) with
        * more info (http://ftp.3gpp2.org/TSGC/Working/2003/2003-05-SanDiego/TSG-C-2003-05-San%20Diego/WG1/SWG12/C12-20030512-006%20=%20C12-20030217-015_Draft_Baseline%20Text%20of%20FFMS_R2.doc). */
@@ -11033,12 +11038,27 @@ qtdemux_audio_caps (GstQTDemux * qtdemux, QtDemuxStream * stream,
           "framed", G_TYPE_BOOLEAN, TRUE, NULL);
       stream->sampled = TRUE;
       break;
+    case GST_MAKE_FOURCC ('s', 'a', 'c', '3'): // Nero Recode
     case GST_MAKE_FOURCC ('a', 'c', '-', '3'):
       _codec ("AC-3 audio");
       caps = gst_caps_new_simple ("audio/x-ac3",
           "framed", G_TYPE_BOOLEAN, TRUE, NULL);
       stream->sampled = TRUE;
       break;
+    case GST_MAKE_FOURCC ('d', 't', 's', 'c'):
+    case GST_MAKE_FOURCC ('D', 'T', 'S', ' '):
+      _codec ("DTS audio");
+      caps = gst_caps_new_simple ("audio/x-dts",
+          "framed", G_TYPE_BOOLEAN, TRUE, NULL);
+      stream->sampled = TRUE;
+      break;
+    case GST_MAKE_FOURCC ('d', 't', 's', 'h'): // DTS-HD
+    case GST_MAKE_FOURCC ('d', 't', 's', 'l'): // DTS-HD Lossless
+      _codec ("DTS-HD audio");
+      caps = gst_caps_new_simple ("audio/x-dts",
+          "framed", G_TYPE_BOOLEAN, TRUE, NULL);
+      stream->sampled = TRUE;
+      break;
     case GST_MAKE_FOURCC ('M', 'A', 'C', '3'):
       _codec ("MACE-3");
       caps = gst_caps_new_simple ("audio/x-mace",