Well, I've been wondering how feasible it would be to implement video transfer in such a way that the "cameras" don't know the "buffers" in advance. Haven't put pen to paper yet, but I believe 802.11b allows for up to 11Mb/s (roughly like 10BaseT). So theoretcially, compressed video (never mind pictures!) could be transferred in several tens of seconds, provided there are not too many devices contending for that bandwidth (anyone know how contention issues are resolved? I assume just like shared medium 10BaseT...).
If there's a crowd situation, and enough "buffers" are around, it should be relatively easy to implement an auto-transfer protocol whereby a buffer senses that a camera's nearby and initiates the upload. Within a few minutes of transfer, practically any MPEG video shot during the rally/whatever should be uploadable, but if not hopefully another buffer walks within range later and another transfer is attempted. A sort of best-effort.
Of course, if back-packable 802.11b switch/routers become cheap enough (I suspect they'll be within a few hundred bucks within a year or two), some interesting things become possible. Then if this group of walking routers is dense enough, any video originating from within the march can theoretically find its way to all buffers, or to any sink/uplink that any of the switches is near, but this obviously has lots of "but only ifs" involved. Oh, and of course once the video is uploaded it might be possible to get rid of it quasi-automatically, so that a buffer may never know what/if has passed through their knapsack.
As for powering issues, I'm not a PC guy, but I'd hope that the fact that the laptop's video is not needed might be a boost.
In any event, not knowing who is an uplink, who is a buffer is very nice....I wonder if there's some other way to accomplish this.
From: "Charles Lucas" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Subject: Confiscation of Sensitive Video
Date: Mon, 28 Oct 2002 22:44:30 -0800
I recently had a discussion about this with some local folks...
We dreamed up a scheme something like this:
N number of people with broadcast cameras.
N or less people with receiving "buffer" backpacks.
A few basestations, which could be housed in a local building, or in a
parked vehicle with some sort of uplink.
The camera people would roam around shooting video, with no memory buffer,
broadcasting on set channels/frequencies.
The backpackers would be running around with receivers, recording from one
or more cameras. When the backpack buffer/HD is nearing full, the
backpacker would move within ranger of a basestation, dump their buffer,
then move back towards the cameras.
The basestation would then upload the video to some site on the net, which
would re-distribute to multiple locations, ensuring the survival of the
captured video.
Goal:
To capture and distribute video in such a fashion that confiscation of the
camera and other equipment will not result in confiscation of captured
video.
Considerations:
All the hardware, especially the cameras, should be as small and cheap as
possible, with the caveat that we need to capture decent quality video. The
cameras are the most susceptible to being confiscated, and should
essentially be treated as disposable.
Cell phones are the most reliable method of data transfer in cities, but
have limited bandwidth. Per-minute charges would make this option
impractical pretty quickly as well.
Un-boosted unidirectional 802.11 has a limited range (about 100 feet, or one
city block). The hardware tends to be expensive, and generally requires a
full computer to operate. Probably too expensive for "disposable" cameras,
although a solution could theoretically be worked out.
The best option for now would seem to be cheap cameras, either of the X10
variety, or something cobbled together by hand that broadcast on standard
analog frequencies. The cameras could each broadcast on a different
frequency, and the backpackers could have multiple recievers, with more than
one backpacker set to receive each camera, for redundancy.
The backpacks could receive a few analog radio channels (one per camera),
convert the stream to a compressed video format, and buffer on a hard drive.
When in range of a designated 802.11 receiver, dump the video as quickly as
possible.
Inter-unit communication is a difficulty. With the receivers having only
about a one-block range, the cameras have to have some way of telling the
backpackers their location, or notifying them if something interesting is
happening some distance away. Portable radios would be useable, but would
be open to interception and require visible landmarks. If a practical way
could be found to have cheap 802.11 between the cameras and backpacks, the
cameras could broadcast their GPS coordinates, allowing backpackers to keep
within range.
Power is a major concern, as batteries are neither light, nor cheap. I have
no feeling at this point for how long you could run a camera and a radio
transmitter, or a backpack receiver/transmitter off of what type(s) of
batteries. This could be a real limiting factor.
The weak points are the basestations. If they can be taken out, the
operation falls apart.
Plausible deniability isn't really possible with this setup. Everyone
pretty much knows who all is involved, and has to keep in contact with each
other in order to capture video optimally. I would see this setup being
used by above-board organizations, such as Indymedia or similar groups. It
would mainly be useful for video capture in large urban protests, like the
ones surrounding the Seattle WTO conference.
Any more thoughts/suggestions around these areas?
_________________________________________________________________
Choose an Internet access plan right for you -- try MSN! http://resourcecenter.msn.com/access/plans/default.asp
