Two Dimensional Equalizer Using Parallel FIR Filters for Data Storage

This paper presents an efficient two dimensional equalizer architecture for data storage. Using parallel FIR structures with modulo arithmetic, the equalizer complexity is significantly reduced, while still operating with low latency. The architecture uses a combination of number theoretic transforms (NTTs) which perform better than conventional interpolation structures with decimation factors that are powers of 2 and 3. A key contribution to the reduction in complexity is the reuse of transform blocks.