New Piggybacking Algorithm on G.722.2 VoIP Codec with Multiple Frame Sizes

In this paper, we present the design of a new piggybacking algorithm implemented on the G.722.2 VoIP codec. In a piggybacking algorithm, multiple speech frames that include those transmitted in the past are encapsulated in a single packet. Because redundant copies of each frame are transmitted to the receiver, the receiver can recover frames lost when one or more packets are lost or arrive late in transmission. In this paper, we have developed a codec that removes further redundancies in the multiple frames encapsulated in piggybacking by sending only one set of Linear Prediction (LP) coefficients for all subframes encapsulated. We create multiple versions of the codec, each working with a different frame size. Our new codec can encode the multiple frames with little degradation in Perceptual Evaluation Speech Quality (PESQ), while having substantially more bit savings. The performance of the new algorithm is evaluated against the original method of piggybacking over random losses, as well as that using packet traces collected over PlanetLab.