Introducing stable oversampling patterns for DFT error recovery codes in bursty erasure channels

Discrete Fourier transform (DFT) error recovery codes like Reed Solomon codes in the finite field are widely used in the communication systems. But the DFT codes in the real and complex fields of numbers are very unstable in the burst of error situations. This paper proposes to change the oversampling patterns in the encoder from the consecutive zero padding to sparse zero padding. It will also prove that the DFT codes based on this type of oversampling are decodable and stable in the case of burst of errors. The simulation results also show the robustness of these codes against quantization error and burst of errors.