Re: Difference between digital phone and VOIP

["Robert L Bass" <RobertLBass@xxxxxxxxxxx>]

> This is because VoIP needs a constant stream of
> data to carry the audio data...

Well, sort of.  By definition, packetization implies
a series of bursts -- not a steady stream.  However,
in an ideal situation the steady series of bursts seems
like a constant stream.

 The voice data is chopped into "packets", each
> packet holding maybe 50-100ms worth of audio...

ISTR it's more like 10-30 mSecs of audio per packet.

> If some of those packets get lost, the system tries
> to "guess" what should have been there...

There are two commonly used methods of dealing with
packet loss. One is to insert noise which sounds like
human speach over the phone.  This tricks the brain
into hearing it as steady sound.

Another technique is called redundancy.  It is exactly
what it sounds like.  Packets are sent several times
in case one gets lost.  If a few packets are lost the
system still has another packet to read.

A more advanced method is called FEC.  This
technique includes partial audio data from one packet
in subsequent packets.  When a packet is lost, a
mathematical algorithm calculates what the missing
data was from information in the adjoining packets.
This might be what some refer to as "guessing" since
the algorithm is not perfect.  For human speach it's
more than adequate.

For critical alarm data we need 100% acuracy. With
various means of dealing with packet loss, it's possible
to be accurate enough to approximate POTS.  IMO
the real issue for most is down time.  No Internet =
no reporting.

> and that actually works pretty good. The human
> brain also has a tendency to fill in the gaps so
> most people don't even hear a few lost packets.
> A control panel modem is not as forgiving.

Agreed.

> You can get a feeling for how many "hops" you
> have between yourself and a destination by
> opening a command prompt and typing:
> tracert www.google.com (or whatever destination).

There are two sides to that coin.  Since Google.com
is the busiest site in the galaxy I ran tracert andd then
pinged it.  It took 9 hops and 89 mSec to get from my
PC to Google.  That's less than the number of switches
I'd have hit and about 135 times faster than alarm
dialup (avg. 12 seconds).  If I were transmitting a
couple of rounds of alarm data I'd have finished the
entire transaction before the alarm panel finished
dialing a phone number.  Of course, the comparison
isn't entirely fair since we're discussing VoIP which
still uses the telco analog network.

> Like someone else wrote in this thread, the voice
> streams share the same bandwidth as all other data
> on the Internet.

That is true.  The reality is that in a short time *all*
voice comms will be via Internet.  The question is how
to deal with it, not whether it is a good thing.

> Thus it may appear as the bundled VoIP works
> better than for example Vonage.

I'll let you know after I've had more time to test it.  I
just installed VIOS for some of my lines plus TV and
Internet.

> So the bottom line is that VoIP is designed for
> VOICE. Not the bleeps and burps a control panel
> spits out when it communicates.

That is 100% true.  This is not lost on the alarm
manufacturing industry.  As VoIP becomes more
uniquitous, expect to see more and more alarm
systems with Internet connectivity built-in or at
least offered as an add-on.

> A panel designed for networking, bypassing the
> VoIP modem, will work a whole lot better than
> one using VoIP as a POTS replacement...

Agreed.  However, there are many millions of
existing systems which currently use POTS and
will soon be sitting behind VoIP.  I expect we'll also
see plenty of Alarm-Internet "dialers" offered to
fill the coming market niche.  This would be
a great opportunity for that Loxxon character.  :^)

--

Regards,
Robert L Bass

=============================>
Bass Home Electronics
941-925-8650
4883 Fallcrest Circle
Sarasota · Florida · 34233
http://www.bassburglaralarms.com
=============================>