I have raised this issue before but nobody had any thoughts (or expressed any) but it is still driving me crazy after 3 months and keeping me from using AAH as my main phone system.
I have two SIP phones (one phone and one ATA). I have BV service.
Randomly during conversations the person I am talking to starts complaining about my voice breaking up or sounding like I am under water (sometimes it happens in the other direction but usually I can hear just fine).
It is completely random. E.g. this morning I spent 20 minutes on the phone with somebody and it was just fine. A few minutes later somebody else made a call and had to hang up because the other person complained. (The other person is ALWAYS on the PSTN). Sometimes the problem shows up about 30 minutes and sometimes in the first 5.
I recently upgraded to FIOS to improve upstream bandwidth from 384K to 2M (I regularly measure 1.6-1.7M). Delay and jitter to broadvoice are reasonable (12.5 to 15ms). THere is only one call in progress through the PBX at any time (right now).
I am not sure even what to post. If anybody has any thoughts, suggestions or questions … please post them.
well it sounds like you have enought upstream bw - however it sounds like you still may need to do someting about QoS - do you have anything in place? (Check out the dlink DI-102 as one good soho example).
I’ve heard a lot of negative about BV although have not used them direclty. The ping and jitter to them may be misleading. If they are redirecting you (and just doing the sip signaling) then you need to check those locations. You would need to turn on sip debug and then scan the log to see where the actual rtp streams are traveling to and it is probably different for outbound and inbound calls.
You may be seeing some of both.
Thanks … I will look into these things …
I really doubt if it is a QoS issue (at least not on my uplink). This was on Sunday morning … neither my wife nor I was even sitting at our computers … other than random background stuff like receiving mail, there was absolutely nothing else happening on the network, locally.
What codec’s are you using? I had BV for a year and gave them up due to deteriorating quality and impossibly long waits for help. But, they are a lot better if you’re close to one of their proxy servers. (I’m up in the NW USA).
But, the underwater sound could be from one of the codecs. Just as a test, try only using ulaw or gsm. gsm probably isn’t supported by your ATA, but if its an option it compresses well without the underwater sound. U-law will always be supported and may give you nicer sound if you have the bandwidth.
i was under the impression that BV only supports ulaw … so that is what I am using…
So … It looks like p_lindheimer was on to something.
I turned on RTP Debug and got the address of the media gateway and did a ping to that… It was atrocious!!!
PING 22.214.171.124 (126.96.36.199): 56 data bytes
64 bytes from 188.8.131.52: icmp_seq=0 ttl=49 time=61.888 ms
64 bytes from 184.108.40.206: icmp_seq=1 ttl=49 time=16.796 ms
64 bytes from 220.127.116.11: icmp_seq=2 ttl=49 time=171.647 ms
64 bytes from 18.104.22.168: icmp_seq=3 ttl=49 time=248.953 ms
64 bytes from 22.214.171.124: icmp_seq=4 ttl=49 time=118.884 ms
64 bytes from 126.96.36.199: icmp_seq=5 ttl=49 time=101.162 ms
64 bytes from 188.8.131.52: icmp_seq=6 ttl=49 time=415.395 ms
64 bytes from 184.108.40.206: icmp_seq=7 ttl=49 time=35.276 ms
64 bytes from 220.127.116.11: icmp_seq=8 ttl=49 time=72.547 ms
64 bytes from 18.104.22.168: icmp_seq=9 ttl=49 time=379.881 ms
64 bytes from 22.214.171.124: icmp_seq=10 ttl=49 time=392.322 ms
Over 300ms of jitter!!! Over 400ms max delay!
I could trace the route as far as the boundary to the broadvoice network and the jitter and delay to that point was reasonable so it is clearly in their network.
So I was mislead because pinging the signaling gateway (the SIP Proxy) gave reasonable results …
I switched to their bos proxy which had decent results and made a call and the media gateway now ranges between 8 and 12ms … So I will try that for a while and see if the problem disappears (fingers crossed).
Thanks for pointing this out to me…
I’ve similar issues that am trying to pin down. Can you provide feedback to the proposed solution?