On Sat, 29 Dec 2018 10:41:10 -0800, Jeff Liebermann wrote:
> That's actually very good, especially with an RF link.
Hi Jeff,
Thanks for your advice, as I'm also in the Santa Cruz mountains (other side
of the hill from you) where WISP is the only thing in town (although Comcast
threatens to bring up cable some day, which would put the small-guy WISPs
like Dave & Brett at Surfnet, Loren at Hilltop, Mike at Ridge, and Bob at
Etheric out of business in a heartbeat - all of whom I presume you know
well).
I'm on 5GHz with a 30dBi Rocketdish with a straight shot, mountain to
mountain, of about 25 miles by road, but only a couple of miles (maybe two
and a half to three miles?) air-to-air (which is what counts).
The Ooma technical folks ran a probe, after trying to talk me into hooking
the "modem" (I never tell them it's a transceiver because that just confuses
them) directly to the Ooma box, where my Ooma box is hanging off the router.
The telephone base is hanging off the Ooma box, and then I use hand helds
around the house. The problem is mostly on the handhelds, but I can't
imagine that they're causing the 0.25% packet loss that Ooma tech support
measured.
> Interference
> from co-channel users usually produces some packet loss. Try a
> continuous ping test to your WISP's router or access point (so that
> you're only testing the wireless path). For Windoze, something like:
> ping -t ip_address_of_WISP
This is a good idea. I need to log it though, so I'm running a
ping -t to an internal hop that I found using a tracert.
Is something like that what you are suggesting?
C:\> ping -t WISP_AP_IP >> ping.log
I'll check that out, as if I find missing packets, that would explain where
the problem lies.
> The "PureVoice" feature may also be involved:
> <
https://www.voip-info.org/ooma-telo/>
> To combat the packet loss that some VoIP users experience
> as garbled or interrupted voice signals, Ooma Telo’s PureVoice
> HD also incorporates adaptive redundancy — the Ooma Telo VoIP
> home phone system detects packet loss and issues duplicate
> packets to cover the gap.
Hmmmmm.... I'm not sure if I can tell that is kicking in or not,
nor what to do about it if it does kick in.
> That can be packet loss, but my guess(tm) is that it's jitter or
> packets lots in the Asterisk switch.
I'm not sure what an "Asterisk" switch is, where googling,
<
https://en.wikipedia.org/wiki/Asterisk_(PBX)>
"Asterisk supports several standard voice over IP protocols".
I guess it's part of the VOIP protocol that Ooma Telo uses...?
> It's not. Ooma does not tell you the end to end (POTS to your phone)
> packet loss. It only displays the packet loss between their servers
> and your Omma device. It does not show anything happening between the
> POTS line and the Omma servers, which can product garble, without
> showing any packet loss.
It happens on almost all calls, so, I'd "think" it's on my side.
(But that's why I ask for debugging help.)
>>Ooma suggested a new cordless phone set.
>
> You old and new cordless phone does not do packetized data and
> therefore would not affect the packet loss. However, if the RF link
> in the cordless phone is defective or there is interference on the
> cordless phone frequency, then you would get garble from the cordless
> phone. Try testing the cordless phone at some other location with a
> POTS line, or temporarily replacing the cordless phone with a wired
> POTS phone.
It _does_ seem to be better (less garbled) when I use the wired handset
which is directly connected to the Ooma device. That diagnostic, alone,
might indicate it's the phones.
But do older (maybe 5 to 10 years?) Panasonic Costco phones cause garbling
in and of themselves? And even so, as you said, they wouldn't cause 0.25%
packet loss (they said the jitter was only 1ms where 20ms would be a
problem, as I recall).
> I would say something about the included wireless handset that comes
> with some Ooma base units, but since you didn't see fit to provide the
> model you're using, I won't bother.
My bad. I apologize.
It's a Panasonic KXTG6671 base plus a few Panasonic PNLC1017 cordless
charger units spread about the home. It was a Costco thing, which, in
reality, I never did like so I'm looking for an excuse to replace it.
Looking at your next post, I first tried this:
<
http://ooma.speedtestcustom.com>
Which reported 3ms jitter, which was more than Ooma had reported
<
http://www.bild.me/bild.php?file=3773901ooma01.jpg>
Pressing the "Again" button reported a 15ms jitter, which is huge
<
http://www.bild.me/bild.php?file=4418379ooma02.jpg>
And, one more time, in sequence, gave me a 2 ms jitter:
<
http://www.bild.me/bild.php?file=4255229ooma03.jpg>
Go figure.
The 2ms is ok, but the 15 ms is at the limit, or nearly so.
I also tried this nice suggestion of yours...
<
https://www.onsip.com/blog/what-your-voip-test-results-mean>
Which seemed, by the GUI, to be EXACTLY the same as the Ooma test,
only, for some odd reason, it picked New York to test against, where it
came up with a 4ms jitter, even as it went across the country:
<
http://www.bild.me/bild.php?file=8371168ooma04.jpg>
The second in the sequence came up with 4ms jitter:
<
http://www.bild.me/bild.php?file=2265190ooma05.jpg>
And yet, the third, came up whoppingly high with 98ms jitter!
<
http://www.bild.me/bild.php?file=8194174ooma06.jpg>
How is _that_ for lack of consistency!
The Sourceforge site says it's "designed to test your current Internet
connection speed for Latency/Ping, Jitter, Download Speed, Upload Speed,
Buffer Bloat, and Packet Loss", which seems like a good test for me!
<
https://sourceforge.net/speedtest/?source=voip-info>
Wow, those are detailed results, where the jitter was 4ms and the packet
loss was a whoppingly high 4% as shown in the screenshot below.
<
http://www.bild.me/bild.php?file=3699166ooma07.jpg>
Surprisingly, even with a 4% packet loss, the quality metric was 4.1 out of
5, which seems higher than it should be with such high packet losses:
<
http://www.bild.me/bild.php?file=9173143ooma08.jpg>
And, just as surprisingly, they gave VOIP a checkmark in the summary:
<
http://www.bild.me/bild.php?file=2095808ooma09.jpg>
Looking at that last suggestion, it seems to be an EXACT copy of the
Sourceforge site where it came up with 4ms jitter & 0% packet loss:
<
https://www.voipreview.org/speedtest>
But this doesn't show the same level of detail as did Sourceforge:
<
http://www.bild.me/bild.php?file=2921509ooma10.jpg>
Hey Jeff! Now that's interesting. Very interesting.
I normally do a "*82" or a "*67" but I didn't know about the others.
The first thing I tried was "*#*#001" which reported "240828".
Kewl.
Then I made a phone call using: *82*96-1-408-123-4567 which had decent call
quality. I'll keep doing this "*96" stuff, which might be the cat's meow.
Thanks.