Amongst the many possible voice functions that can be implemented, voice coding has carried the most promise of finding early applications due to potential use in the existing telecommunications networks. However, commercial realisation of this opportunity has been somewhat retarded, partly due to the problem of coding voice down to low data rates not being addressed adequately. Whereas this has not been purely a matter of suitable algorithms being unavailable, the cost to performance necessary for many obvious consumer level applications has not been achieved. This paper discusses a Voice Coder for an application which may become one of the early successes in the mass market.