I wondered if CPRI's quantization procedure for the downlink takes into account a power variation margin and thus needs extra quantization resolution bits for that. Does it simply acts on normalized amplitude signals after power information is sent by protocol? Is the uplink quantization procedure similar?