The FlexCode Speech and Audio Coding Approach

Inhalt:
The increasing heterogeneity of communication networks and the variability in user requirements form a challenge for source coding algorithms. To address this challenge, the aim of the Flex-Code source coding approach is to create a speech and audio coder that can adapt instantaneously to network and user requirements. The approach is based on a statistical description of the source, a model of perception, and analytic relations between bit rate, distortion, packet-loss rate, and the reconstruction-point distribution density of the quantizers. The method facilitates analytic optimization of the source coder for given network and input signal conditions and user requirements at any time. Preliminary results confirm that the flexibility in rate and robustness can be obtained without a loss in performance compared to state-of-the-art coders optimized for particular conditions.