Publications at NRL

Search by Title


Search by Author


Conference Paper


An Ultimate Encoding Layer for Layered Real-Time Speech Streams over Multi-hop Wireless Networks


Future integrated networks are expected to offer packetized voice and other multimedia conferencing services to mobile users over wireless links. Wireless networks cannot easily support multimedia applications due to the media high probability, burstiness, persistence of errors, and delay jitter suffered by packets. This is not only due to queuing but due to link-layer (MAC) protocols as well. Mobility results in even higher loss rates and jitter for the higher layers. Adding real-time, reliability and multicast constrains to ad-hoc, multihop or cellular configurations makes the problem even more difficult. Providing for good end user perception in such environments cannot be done by applying traditional methods such as scheduling and reservations as in wired fixed networks. The high variance of the media response has to be handled by managing (monitoring and adapting to) QoS. A finer and more aggressive layering can improve perception in some applications intended for use in these or in hybrid systems. A lower layer can be defined as the minimal set of information that is needed to keep the application alive under extremely adverse network conditions. In this paper, using a real testbed and our audio on demand application ([1], [2], [3]) we propose one more possibility for the lower rate layer which is useful in speech streaming over such networks. A text transcription can be generated from the audio stream using a speech recognition engine at the sender side. The text traverses the path easier (very low bit rate) and more reliably (e.g. using redundancy). The data can be displayed at the receiver side in a caption window or, more importantly, the speech can be reproduced using the transcription with a text-to-speech synthesizer. Our experiments show that, by adapting to QoS and using our transcription scheme, end user perception can be greatly enhanced, and meaningful communication can be sustained even at most adverse network conditions.

Paper: PDF file of paper

Information & Date

2nd Annucal UCSD Conference on Wireless Communications, SD, CA, March. 1999


Tsuwei Chen
Yuri Romanenko
Mario Gerla