> > Hello all, > > I was looking at liblame for streaming data and had a few > questions about its usage. I would like to encode a stream > and will be sending the mp3 data over UDP and using liblame > on the other side to do the decoding. Is the mp3 format > able to handle the possibility of packet loss or can mp3 > reference previously received frames? Packet duplication > isn't a problem as the program checks for that. >
The frames are dependent, so you will have dropout (or noise) if you loose some of them. If you loose frame N, you wont be able to fully decode either frame N-1 or frame N+1, and in very rare occasiaons frames N+2 or even higher. But even if the frames were not dependent, you would always have some kind of glitch in the audio if you lost a frame? But I dont know if the decoder in liblame is up to the task of recovering after loosing a frame. We have added some "resyncing" capability to mpglib used in LAME, but it is not great. > Also, is there a way to determine how large an mp3 frame > is? Does the size of an mp3 frame vary with bitrate or is > it constant. The reason I ask is I would like to send > complete frames rather than partial frames for the packet > loss reason. > The formula is given in our the tech-FAQ, see the link at www.mp3dev.org Mark _______________________________________________ mp3encoder mailing list [EMAIL PROTECTED] http://minnie.tuhs.org/mailman/listinfo/mp3encoder
