parser should not be needed anymore, merge cases and set frame size for amr too

Originally committed as revision 12956 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavformat/mov.c

@@ -925,9 +925,8 @@ static int mov_read_stsd(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom)
         st->codec->channels= 1; /* really needed */
         break;
     case CODEC_ID_QCELP:
-        st->codec->frame_size = sc->samples_per_frame;
-        st->need_parsing= AVSTREAM_PARSE_FULL;
     case CODEC_ID_AMR_NB:
+        st->codec->frame_size = sc->samples_per_frame;
         st->codec->sample_rate= 8000;
         st->codec->channels= 1; /* really needed */
         break;