/mcpthe gateway, or if you are Telneted into a gateway, make sure you have issued the terminal monitor command from the exec level. Note: This command is not available on the AS5x00/AS5x50 platforms. In this output, the value for Rx Delay Est (ms) is 71. This is the current jitter buffer value. A value for the high water mark and low water mark is deduced on this. An average initial value for the high water mark is 70 msec, while that for the low water mark is 60 msec. Once an initial value is set, the gateway keeps track of any early packets or late packets received. As is seen in the output here, the prediction concealment drops are close to 250 ms, while the silence concealment are 30 ms. There is always a higher value for prediction concealment since silence concealment is only a worse case scenario of Prediction concealment. For every Prediction concealment drop, there is an increase in the buffer overflow discard. If you see buffer discard, it does not necessarily mean that you see an increase in the high water mark. The high water mark is the upper limit of the jitter buffer. It changes only if a trend is observed. In other words, there should be a continuous flow of late packets. This results in an increase of the jitter buffer. In the output here, such a trend is present. Therefore, the high water mark is increased from 70 msec to 161 msec. If this value is not changed (and if you still see 14 late packets), it implies that these are sporadic late packets, not forming a trend. From the output of the show call active voice command, look out for lost packets. For every lost packet, you see two packets that are out of sequence. This is seen on the Rx Non−Seq Pkts output. Since it is not a positive value, it is concluded that there has not been any packet losses either. 3640−6# DSP VOICE TX STATISTICS Tx Vox/Fax Pkts: 195, Tx Sig Pkts: 0, Tx Comfort Pkts: 10 Tx Dur(ms): 192070, Tx Vox Dur(ms): 388, Tx Fax Dur(ms): 0 DSP VOICE RX STATISTICS Rx Vox/Fax Pkts: 9604, Rx Signal Pkts: 0, Rx Comfort Pkts: 0 Rx Dur(ms): 192070, Rx Vox Dur(ms): 191560, Rx Fax Dur(ms): 0 Rx Non−seq Pkts: 0, Rx Bad Hdr Pkts: 0 Rx Early Pkts: 0, Rx Late Pkts: 14 DSP VOICE VP_DELAY STATISTICS Clk Offset(ms): 0, Rx Delay Est(ms): 71 Rx Delay Lo Water Mark(ms): 60, Rx Delay Hi Water Mark(ms): 161 DSP VOICE VP_ERROR STATISTICS Predict Conceal(ms): 250, Interpolate Conceal(ms): 0 Silence Conceal(ms): 30, Retroact Mem Update(ms): 0 Buf Overflow Discard(ms): 500, Talkspurt Endpoint Detect Err: 0 DSP LEVELS TDM Bus Levels(dBm0): Rx −49.9 from PBX/Phone, Tx −41.7 to PBX/Phone TDM ACOM Levels(dBm0): +2.0, TDM ERL Level(dBm0): +11.1 TDM Bgd Levels(dBm0): −58.9, with activity being voice DSP VOICE ERROR STATISTICS Rx Pkt Drops(Invalid Header): 0, Tx Pkt Drops(HPI SAM Overflow): 0 Observe the Tx Comfort Pkts and Rx Comfort Pkts. As from the example outputs, it is concluded that the phone connected to this router mostly keeps quiet since you have lots of Tx Comfort Pkts. At the same time, you have zero Rx Comfort Pkts, which means that the other end continuously speaks. Compare the output here with the previous command output. There is an increased number of Rx Late Pkts (from 14 to 26). However, there is no increment in the high water mark value. This indicates that the 12 packets are sporadically delayed. The Buffer Overflow discard is increased to 910 msecs. However, since there is no trend observed, the high water mark is not increased. In the output here, you have a Rx Early Pkts: 3. This means that a packet arrives much before it is expected. As seen from the output here, the Jitter buffer has stretched itself to accommodate for any more early packets by reducing the low water mark from 60 to 51.