Re: [hackers-il] VoIP the SIP way and NAT traversal
- Oleg Goldshmidt wrote:
>>>The 4 seconds round trip time is most likely caused by your internetThta's the compromise I'm talking about.
>>>connection going via a satelite link
>>I rather think it's the jitter/latency compromise that the clients
>What compromise are you talking about? If anything, network latency
>and jitter (and packet loss) are likely to be correlated. In fact,
>there is probably significant jitter and out-of-order arrivals that
>get translated into extra latency to keep speech intelligible. I doubt
>there are any working systems that wait for 2 seconds to reorderNo, 2 second is the time the packets need to travel one way PLUS the
>because that's way above what is considered bearable (anything above
>250 ms one way - including packetization and re-assembly - is
>considered unusable for practical purposes).
reordering latency. That being said, I had long distance calls with
more-or-less the same roundtrip delay, out of "regular" voice lines.
They are more than usable. After some time you just get used to it
("some time" defined as 10 minutes give or take).
>>Okay I'm nuts. Kill me and be done with it.Well, actually most phone calls I had from India from a cellphone in
>You are. 4 sec RTT is not a duplex phone conversation as we know
>it. But if your girlfriend likes the quality of conversation you
>manage to maintain in these adverse conditions, doesn't kill you, and
>is not done with it, who are we to judge?... ;-)
India to a cellphone in Israel have a latency close to that, and like I
said they're usable.