Actually, through further testing we've found that it's not. What we've found is jitter is secondary to packet loss when classifying a call based on QoS definitions. This has caused calls to be mis-classified as NA even though the match a jitter value but not a packet loss when serarate definitions exist for the same call quality type using either PL or Jitter.
Further, after obtaining additional information on CAR and the IP Phones, I was informed by TAC that the jitter reported by the phones is suppose to conform to RFC 1889 specification for jitter calcualtion. Since RFC 1889 specifies an average mean deviation across all RTP packets, there should always be reported into CAR a value >0 for jitter, unless of course the jitter was 0 throughout the entire call, highly unlikely. However, I have monitored calls with varying amounts of jitter yet still have seen CAR report a jitter of 0 for that call. The issue with CAR and jitter reporting is that CAR only get the last jitter score as reported by the IP phone into the Call Diagnostic Records. From our testing the reported jitter value is not an average mean deviation as defined by the RFC, therefore it is not a valid measurement against the health of the entire call as it is the last reported jitter before call termination and not representative of the entire call.
