A packet loss concealment method using recursive linear prediction

We proposed and evaluated a speech packet loss concealment method, which predicts lost segment from speech included in both packets before and after the lost packet. The lost segments are predicted by recursively using linear prediction both in the forward direction from the packet preceding the loss, as well as in the backward direction from the packet succeeding the lost segment. The predicted sample in each direction is smoothed to obtain the final interpolated signal. The adjacent segments are also smoothed to significantly reduce the discontinuity between the interpolated signals. Subjective quality of the proposed method showed higher scores than the packet loss concealment algorithm described in the ITU standard G.711 Appendix I, with MOS rating exceeding 2.4, even at an extremely high packet loss rate of 30%.