Optimal Retention Delay in the Receiver of a Digital Voice Network
Original Publication Date: 1985-Jan-01
Included in the Prior Art Database: 2005-Feb-18
At the receiving end of a digital voice packet transmission network, voice synthesized signals have to be re-synchronized due to packet transmission delay jitter. To avoid these distortions, instead of being fed into the synthesizer as they arrive, the packets are fed into a buffer register. They are then fetched out of the buffer sequencially after a fixed length retention delay greater than the maximum transmission jitter expected. This retention delay is reset after each long speech pause (e.g., greater than a predetermined length value). The transmission delay jitter is thus absorbed by the long speech pauses, while effective speech signal synthesis is not affected. This method may however penalize the system by unduly affecting slow packets. A solution is proposed here to avoid these drawbacks.