/mcpthe Total Voice Playout Duration. The Total Voice Playout Duration can be derived with the addition of the gap fill durations to the OnTimeRvPlayout duration. If the proportion of on time voice playout time is high then the call is likely to be healthy. Packets dropped or delayed too long in the packet network can cause voice quality issues. On receipt of packets that are delayed for so long that they cannot be used, or when packets are dropped in the network and not received at all, an IP phone or voice gateway attempts to reconstruct the voice stream as best it can by the prediction of the voice signal. Repeatedly issue the show call active voice command on an IOS gateway in order to provide visibility into this issue: LatePackets—The number of packets that arrive outside the de-jitter buffer playback delay period. These packets are discarded. LostPackets—The number of packets that never arrive at the receiving IP phone or gateway. GapFillWithPrediction—The amount of packet prediction in a call. Divide this number by the packet sample time in order to determine the number of packets affected. GapFillWithSilence—The amount of silence insertion in the call. Note: The show port voice active command on a Catalyst gateway gives you an indication of jitter for a call (Hi/Low water playout delay) though it does not differentiate between predictive and silence insertion. Synthetic voice A small amount of predictive insertion is undetectable to the human ear. However, a large amount probably causes a garbled quality in the voice that can be described as synthetic or robotic voice. Choppy voice If packets are dropped or arrive late, then it is not possible for the receiving codec decoder to predict the voice signal. In this case, the signal is replaced with silence inserted into speech. In addition, if delay is variable (jitter), packets that arrive late but within the playout delay period of the receiving de-jitter buffer, are played out but can cause an underrun of the de-jitter buffer. An underrun occurs when there are no packets left held in the buffer and the speech is delayed when the buffer waits for the next packet to arrive. Audible gap in speech can result. A small amount of silence insertion or jitter is undetectable to the human ear. However, a large amount probably causes a quality in the voice that can be described as choppy voice or broken voice. Note: If the network delay is variable enough, it is likely that the resulting sound of the speech is both synthetic and choppy.