Time-Scale Modification of Audio Signals Using Enhanced WSOLA With Management of Transients
IEEE Transactions on Audio, Speech, and Language Processing
A survey of packet loss recovery techniques for streaming audio
IEEE Network: The Magazine of Global Internetworking
Hi-index | 0.00 |
Waveform Similarity OverLap-and-Add technique (WSOLA) is widely used in Packet Loss Concealment (PLC) of real time voice communications. In this paper, based on our enhanced WSOLA technique, an algorithm for PLC of voice signals is proposed. One of the disadvantages of the standard WSOLA technique is its lack of efficient amplitude controls, which might lead to a significant mismatch between the original voice signal and the restored signal. In the proposed algorithm, we introduce a gain into the standard WSOLA technique which could help to adjust the amplitude of the voice segments excised from the input signal. The modified voice segments are applied in the OverLap-and-Add stage instead of the previous ones. Our study shows that by the introduction of a gain into the standard WSOLA technique, the quality of the restored voice signal is probably to be higher than that of the standard WSOLA technique.