According to the Forum, in order to maximize use of bandwidth, it is possible to carry multiple voice samples in a single frame, further minimizing overhead.