On the Fair Comparison of FEC Schemes

Two new metrics for the high-level comparison of Forward Error Correction (FEC) schemes are proposed and applied to the study of Turbo and LDPC codes. These new metrics retain simplicity and allow us to compare FEC schemes by factoring in their relative complexity. Furthermore, for the case of Turbo and LDPC codes, we determine the number $cal I$ of LDPC iterations that correspond to an iteration performed by a Turbo decoder under the constraint that an equal number of elementary operations is performed at the receiver. As a practical example, we apply these metrics to the comparison of the FEC schemes debated in the ITU-T G.hn Working Group (WG) confirming the technical validity of the selection of the LDPC-based proposal over the Turbo Code one.