User`s manual

SIP User's Manual 8. Telephony Capabilities
Version 5.0 215 October 2006
8.6 Dynamic Jitter Buffer Operation
Voice frames are transmitted at a fixed rate. If the frames arrive at the other end at the
same rate, voice quality is perceived as good. In many cases, however, some frames can
arrive slightly faster or slower than the other frames. This is called jitter (delay variation),
and degrades the perceived voice quality. To minimize this problem, the gateway uses a
jitter buffer. The jitter buffer collects voice packets, stores them and sends them to the
voice processor in evenly spaced intervals.
The gateway uses a dynamic jitter buffer that can be configured using two parameters:
 Minimum delay, ‘DJBufMinDelay’ (0 msec to 150 msec). Defines the starting jitter
capacity of the buffer. For example, at 0 msec, there is no buffering at the start. At the
default level of 10 msec, the gateway always buffers incoming packets by at least 10
msec worth of voice frames.
 Optimization Factor, ‘DJBufOptFactor’ (0 to 12, 13). Defines how the jitter buffer tracks
to changing network conditions. When set at its maximum value of 12, the dynamic
buffer aggressively tracks changes in delay (based on packet loss statistics) to
increase the size of the buffer and doesn’t decays back down. This results in the best
packet error performance, but at the cost of extra delay. At the minimum value of 0,
the buffer tracks delays only to compensate for clock drift and quickly decays back to
the minimum level. This optimizes the delay performance but at the expense of a
higher error rate.
The default settings of 10 msec Minimum delay and 10 Optimization Factor should provide
a good compromise between delay and error rate. The jitter buffer ‘holds’ incoming packets
for 10 msec before making them available for decoding into voice. The coder polls frames
from the buffer at regular intervals to produce continuous speech. As long as delays in the
network do not change (jitter) by more than 10 msec from one packet to the next, there is
always a sample in the buffer for the coder to use. If there is more than 10 msec of delay at
any time during the call, the packet arrives too late. The coder tries to access a frame and
is not able to find one. The coder must produce a voice sample even if a frame is not
available. It therefore compensates for the missing packet by adding a Bad-Frame-
Interpolation (BFI) packet. This loss is then flagged as the buffer being too small. The
dynamic algorithm then causes the size of the buffer to increase for the next voice session.
The size of the buffer may decrease again if the gateway notices that the buffer is not filling
up as much as expected. At no time does the buffer decrease to less than the minimum
size configured by the Minimum delay parameter.
Special Optimization Factor Value: 13
One of the purposes of the Jitter Buffer mechanism is to compensate for clock drift. If the
two sides of the VoIP call are not synchronized to the same clock source, one RTP source
generates packets at a lower rate, causing under-runs at the remote Jitter Buffer. In normal
operation (optimization factor 0 to 12), the Jitter Buffer mechanism detects and
compensates for the clock drift by occasionally dropping a voice packet or by adding a BFI
packet.
Fax and modem devices are sensitive to small packet losses or to added BFI packets.
Therefore to achieve better performance during modem and fax calls, the Optimization
Factor should be set to 13. In this special mode the clock drift correction is performed less
frequently - only when the Jitter Buffer is completely empty or completely full. When such
condition occurs, the correction is performed by dropping several voice packets
simultaneously or by adding several BFI packets simultaneously, so that the Jitter Buffer
returns to its normal condition.