Hello, I am working with NSpeex, a .NET implementation of Speex. But my question is not implementation-specific. My app is receiving the speex narrowband (160 bytes frames and 40 bytes subframes) encoded audio in chunks of 20 bytes. The problem is. This is a continuous stream of data, and currently I don't know where a frame begins or ends. The first byte I send to the decoder can be from beginning of the frame (which would be optimal), or from the second half of a subframe (worst case). I don't know which will be. Is it possible to decode this audio in such conditions? Thanks in advance Jader Dias