Re: NTLK Re: MP3 decoding chip?

Date: Tue Jul 18 2000 - 11:36:55 CDT

On Tue, 18 Jul 2000, Stainless Steel Rat wrote:

> * <> on Mon, 17 Jul 2000
> | Huh? Why would you claim that? WaveLAN does signalling at 11Mbps. MP3s
> | are around 128kbps usually. That's around a factor of 100x slower than
> | you need. I know that there are inefficiences and overhead, but they
> | aren't that huge.
> You're not going to get 11Mbps out of your MP2x00. It just isn't fast
> enough, and it will suck power like nobody's business (which defeats the
> purpose). On top of that, when Ethernet traffic exceeds ~50% of maximum

You originally stated that WaveLAN wouldn't be fast enough. I haven't
tested the Newton's performance, though that could easily be done via

Regarding battery consumption, I've talked to other people using handheld
devices and they've found the battery life with Lucent cards to be quite
good. But I don't walk around all day. I want this for when I'm sitting
down doing something, and there's always a power outlet nearby. (there's
also almost always an ethernet jack, but less cables is better)

> bandwidth, things grind to a halt (there are ways of preventing this, but
> 802.11 isn't switched).

> | I have personally spoken to people who have played MP3s over a large scale
> | production WaveLAN system that only runs at 2Mbps.
> Define "large scale", please. Large scale to me means switched Gigabit
> fibre connecting HP N-class servers and 4000-class ALPHAservers and the
> like.

I believe that Carnegie Mellon University has one of, if not the largest
installation of WaveLAN base units. I don't have exact figures for the
number, but the entire campus has continuous coverage from one side to the
other. There are currently 33 buildings with coverage, and due to various
factors related to range, there is often at least one base unit per
floor. There are enough base stations that one research group has some
code that triangulates your position based on signal strength from
multiple base stations. Larger classrooms have their own units too.

I generally consider on the order of 5000 to 10,000 computers, spread over
a large area including ADSL users, WaveLAN, and normal ethernet, to be
"large scale".

(Our campus does have some strange features. It's not that big a
campus. But at 1/2 mile or so, there are radio black spots using a 25W
repeater and a 5W HT (handheld radio))

> | I don't know why you believe it doesn't have the bandwidth, but I can
> | assure you that it does...
> If you're not doing anything else, sure.

Since MP3 playback would be a hundredth of the signalling rate, please
cite a source or some benchmarks or figures to justify this claim?

When I watched somebody playing mp3s, there were, in that room, about 5
other people doing network-heavy traffic. Remote connections to lots of
other places, web browsing (fast browsing, not slow :), the usual
stuff. None of them noticed a slowdown.

I've been in a room with ~50 people all using the same base station and
they seemed quite happy with the performance.

Really, the problem here is the Newton's performance. And I have seen
little measurement done. I know that it can handle 230.4kbps continuous
double buffered recording and playback over the serial port; I've done
that myself. Theoretically, it's possible. Practically, we'll see :)


NewtonTalk brought to you by:

EVOTE.COM -- the ESPN of politics on the Internet! All the players, all the news, and the hottest analysis and features (plus 'toons!) anywhere.... visit today!

Need Subscribe/Unsubscribe info?

Visit the NewtonTalk section at

This archive was generated by hypermail 2b29 : Tue Aug 01 2000 - 06:50:09 CDT