We consider the retrieval of data from a time-domain spectral hole-burning (SHB) memory system. A new iterative log-likelihood (ILL) algorithm is used to reliably detect corrupted retrieved data signals. It is a blockwise technique that takes advantage of the known SHB system characteristics to mitigate time-varying intersymbol interference and detector shot noise. We present bit-error-rate results obtained with the ILL algorithm and five other typical methods (i.e., precompensator, simple threshold, adaptive threshold, a simple Wiener filter, and an adaptive Wiener filter). Results show that the ILL algorithm outperforms all five techniques and hence offers improved SHB storage capacity. In a SHB system with typical material parameters, we find that ILL offers a storage capacity gain of 197% as compared with simple thresholding.
© 2001 Optical Society of America
Original Manuscript: June 7, 2000
Revised Manuscript: December 5, 2000
Published: April 10, 2001
Lilin Zhang and Mark A. Neifeld, "Blockwise data detection for spectral hole-burning memories," Appl. Opt. 40, 1832-1842 (2001)