A methodology for speech codec FEC design and optimisation based upon the ITU-T Rec. P.861 objective speech quality measure

This paper describes a methodology for designing and optimizing a forward error correction (FEC) scheme which has been employed in the standardization of the ETSI TETRA 2 speech codec. The method involves using a modified version of ITU-T Rec. P.861 [1] in order to measure the peak distortions between the error-free and channel impaired decoded speech samples averaged over approximately 2 minutes. This objective measure provides bit priority ordering which is close to that obtained by conventional subjective methods and has permitted a degree of FEC optimization which would otherwise be time consuming and prohibitively costly.