| |
PA2AGA > TCPDIG 14.02.98 01:44l 137 Lines 5526 Bytes #-9590 (0) @ EU
BID : TCP_98_18B
Read: GUEST
Subj: TCP-Group Digest 98/18B
Path: DB0AAB<DB0PV<DB0WGS<DB0RGB<DB0MAK<DB0ERF<DB0HSK<PI8DRS<DB0PKE<DB0SM<
PI8DAZ<PI8GCB<PI8HGL<PI8VAD<PI8VNW
Sent: 980213/2033Z @:PI8VNW.#ZH2.NLD.EU #:30330 [HvHolland] FBB7.00f $:TCP_98_1
From: PA2AGA@PI8VNW.#ZH2.NLD.EU
To : TCPDIG@EU
Received: from pa2aga by pi1hvh with SMTP
id AA2372 ; Fri, 13 Feb 98 19:58:32 UTC
Received: from pa2aga by pa2aga (NET/Mac 2.3.65/7.1) with SMTP
id AA00006579 ; Fri, 13 Feb 98 20:33:01 MET
Date: Fri, 13 Feb 98 20:30:42 MET
Message-Id: <tcp_98_18B>
From: pa2aga
To: tcp_broadcast@pa2aga
Subject: TCP-Group Digest 98/18B
X-BBS-Msg-Type: B
connector's hood; very compact. The Libretto I have is the low-end
486-dx100 version, and it ran well. I have a question for Alan: You say a
486 cannot run the soundcard modem, do you know how much more CPU it
requires than the baycom driver?
I know the Baycom driver uses quite a bit of CPU, and thus have decided to
try out my same net setup with a TNC instead of the Baycom for a direct
comparison. I'll let y'all know what happens....
Erik
N2NRP
---
Erik M. Hall - N2NRP
University Of Central Florida, CREOL
Center for Research and Education in Optics and Lasers
Orlando, FL
(407)823-6942
------------------------------
Date: Thu, 12 Feb 1998 17:16:13 +0000 (GMT)
From: Alan Cox <alan@cymru.net>
Subject: wearable telephone gateway
> 486-dx100 version, and it ran well. I have a question for Alan: You say a
> 486 cannot run the soundcard modem, do you know how much more CPU it
> requires than the baycom driver?
The 486DX100 probably can handle it - the PC110 is a 486SLC and has no FPU
which sort of makes it a nohoper unless someone wants to write an integer
filter set for the system..
> I know the Baycom driver uses quite a bit of CPU, and thus have decided to
> try out my same net setup with a TNC instead of the Baycom for a direct
> comparison. I'll let y'all know what happens....
Baycom burns CPU - but on the other hand its cheap and if the CPU is free
why not use it. That CPU is so cheap as part of the cost of a PC nowdays
------------------------------
Date: Thu, 12 Feb 1998 14:38:05 EST
From: CAPAjt@aol.com
Subject: wearable telephone gateway
Since you seem to have advanced knowledge of these systems, Let me ask you is
it possible to have the screen/monitor on the Libretto closed yet still input
data? (I'm trying to avoid having to remove the screen in a wearable
application)
Thanks
------------------------------
Date: Fri, 13 Feb 1998 08:24:52 +1100
From: Terry Dawson <terry@perf.no.itg.telstra.com.au>
Subject: wearable telephone gateway
Dr. Terrance Boult wrote:
> >>I have a Libretto, great Linux machine, but unfortunately it doesn't
> >>have any audio input, only output, probably the most annoying feature.
>
> Did not do it myself (yet), but I recall a demo at the wearables conference
> where someone (from MIT media I believe) went in and just soldered an input
connection
> to the sound-blaster chip. Input support is there, just no access.
Is that so? I haven't opened it up yet, so I haven't looked. I'd wondered
whether that was the case or not. If it is then that is excellent and I'll
do just that.
I can't understand why they wouldn't have access .. the cost of a jack ?
Terry
------------------------------
Date: Thu, 12 Feb 98 22:26:17 UT
From: "Dan Newell" <DanNewell@classic.msn.com>
Subject: wearable telephone gateway
FYI - I've connected my ViA to a metricom modem and used Net2Phone to place
calls. The results are pretty amusing - the people I called could not
understand a thing I was saying and reported back that calls sounded like they
were coins shaking on Mars. The audio I received back was spotty and at times
15-20 seconds delayed. This was my experience after simply bolting things
together and trying it out without any tuning or attempts to identify the
bottlenecks.
A couple caveats and notes then. I chose Net2Phone because some of the other
internet phone product/providers (such as Vocaltec) with IP to PSTN (public
switched telephony networks) gateways had business models which provided
direct pipelines to my credit card in the event security was compromised.
Net2Phone allowed me to limit the damage by prepaying an amount and calls are
stopped/prevented when this is exhausted. Not that I was that worried about
the other products but at least one ITSP (internet telephony service provider)
went under because of fraud among other reasons - I just didn't want to be one
of those pioneers with an arrow in my back. When Net2Phone is used on my LAN
connection where I have a 1.6mb radio link out to the rest of the universe,
Net2Phone calls to their service department and people's phones provided
decent service so there's something about the Metricom data rate and/or
latency at work. Someone else's mail about Metricom modem's ratcheting down
response time when unused recently may be to blame. Will have to check on
this.
Unfortunately, QOS (quality of service) is not so good with the internet and
it's kind of important for telephony. Too many links or delays in the loop
really play havoc. It's one thing for me to know I need to revert to
walkie-talkie mode ("Meet you at Bob's...over.") but the receptionist at the
dentist is going to be confused if they get a call like this from me. Telco's
all have psychometric data on what is and isn't acceptable. Just be aware that
even a few tenths of a second delay starts to become noticable and intrusive
and simply connecting the IP/voice dots together doesn't guarantee something
useful. What often makes sense for computer data (gathering big buffers and
sweeping data all at once improves efficiency) isn't so hot for voice which is
To be continued in digest: tcp_98_18C
Read previous mail | Read next mail
| |
