Efficient Quality-Based Playout Buffer Algorithm

TL;DR

This paper introduces a low-complexity, quality-based playout buffer algorithm for VoIP that optimizes speech quality by modeling network delay with a Pareto distribution and simplifying the ITU-T E-Model.

Contribution

It presents a novel, computationally efficient playout buffer algorithm that maximizes speech quality using a closed-form solution based on a Pareto delay model and a simplified E-Model.

Findings

Outperforms existing algorithms in speech quality.

Reduces computational complexity significantly.

Effective modeling of network delay with Pareto distribution.

Abstract

Playout buffers are used in VoIP systems to compensate for network delay jitter by making a trade-off between delay and loss. In this work we propose a playout buffer algorithm that makes the trade-off based on maximization of conversational speech quality, aiming to keep the computational complexity lowest possible. We model the network delay using a Pareto distribution and show that it is a good compromise between providing an appropriate fit to the network delay characteristics and yielding a low arithmetical complexity. We use the ITU-T E-Model as the quality model and simplify its delay impairment function. The proposed playout buffer algorithm finds the optimum playout delay using a closed-form solution that minimizes the sum of the simplified delay impairment factor and the loss-dependent equipment impairment factor of the E-model. The simulation results show that our proposed algorithm outperforms existing state-of-the-art algorithms with a reduced complexity for a quality-based algorithm.