There has been few Lync call quality issues. The issue seems to be limited to PSTN calls (inbound and outbound) via SIP Trunk.
On investigation, it is noticed that Lync logs are showing Applied bandwidth Limit as 46 Kbps. In general the Applied bandwidth limit value is 151 Kbps for most of the good PSTN calls.
According to Microsoft, the Applied Bandwidth Value is basically the maximum bandwidth the send stream can take barring limits imposed by the bandwidth estimate. However, from Lync side there are no limits or Call Admission Control enabled for bandwidth estimate limit and I'm not sure how Lync calculates this Applied Bandwidth Value based on the combination network parameters. From Network side, there is enough bandwidth available for real time class but Lync still go for 46 Kbps and then select the x-msRTA codec. Due to low applied bandwidth, call quality is poor and intermittently dropping the calls as well.
From logs, packet loss, jitter and RTT are low (well below Microsofts recommended values) but Lync still gets 46 Kbps and reflecting poor Conversational MOS value (below 3).
This issues has been raised with SIP supplier and they confirmed no issues (no packet loss, no high jitter, no high latency) at their end.
Network team has investigated LAN and WAN and found no abnormality. As per network utilization report, offices are not even reaching to 50% real time class utilization but still Lync is causing poor call quality. QoS and Tagging confirmed - no issues.
Why the applied bandwidth is showing as 46 Kbps even though there is enough bandwidth available for real time class. What are the network parameters behind the calculation of Applied Bandwidth Limit and how does it impact on PSTN call quality?
Cheers,
H.