Speaking of VC equipment... 2008-08-21
At Google, a majority of our conference rooms are equipped with some
pretty top of the line VC equipment. Someone once told me that all new
conference rooms had to be VC equipped, but maybe that's just a rumor.
I do know, that once you start using VC, audio conferences seem like the
stone age. I looked into personal VC equipment last Christmas when my
newborn son made it clear I wasn't going to be traveling home for
Christmas for the first time in my life. The professional stuff is
outrageously expensive, making the dual projectors seem like a cheap
part of the standard big room VC equipment at work. Even a "TV Top"
setup runs to $1500 to $2500.
This struck me as insane. Practically everyone I know has a large flat
screen TV and broadband. Webcams are $100 and PCs are $400. The latest
Webcames from Logitech claim "HD" quality, and I can tell from
experience tha their RightLight technology works amazingly well. A consumer
level device should be easily under $500, possibly half that with work.
I only found one consumer level device at a reasonable price point, the
DVC-1000. It was a piece of crap, partially because its a little
old at this point. It didn't support upnp, so I had to manually open
wholes in my firewall... and then help my Dad do the same on his end.
Then, the picture was crap. It desparately needed the RightLight style
auto-contrast fixing software that my Logitech Webcam has. Even after
that, the picture quality was poor and the frames per second was near
useless. Instead, we used Skype video chat and just hooked PCs up to
the TVs. That ended up not working so well either, though we've
routinely used Skype for video chat other times without issue. Skype
mostly just works, though its interface is too focused on audio... time
to make the switch, Skype.
Going back to the conference room equipment, it has one major failing:
the software sucks. They get the video and audio quality right, but two
things are major fails: the addressbook and the layout choices. The
default interface is to dial a number... right. Instead, try to use the
addressbook. We have 40+ offices, thousands of conference rooms and
people's desktop computers... and it present it all as a very slow
alphabetical list. No hierarchy. You can prefix search, sorta. You
can bring up a search box which does substring search... except random
strings can't be searched for. It should take an engineer a week to fix
this.... The other major issue is layout. You can have multiple
locations called in, plus locations can project a separate screen
(usually a computer). And one quirk of current VC, you really kind of
need to see yourself, to make sure you're on camera, or that the group
of you is on camera. With the equipment we have, you can keep hitting
the layout button to shuffle all of these things on screen, but it never
does what you want. I don't need to see myself twice (one local, one
echo), and I certainly don't need my picture to be the largest. In some
modes, it tries to make the currently talking location the largest, but
often it fails to do that. It has no concept of room size or anything,
so often a single person location is as large or larger than a location
with 20+ people. White boards really don't work, since either their
"off screen" to one side or the other, or they're at the far end... and
you either zoom in on that and ignore all the people, or you see the
people and can't see the whiteboard. And then someone taking minutes
decides to project... and you lose half your screen real-estate to
something you don't care about, and you can't tell the VC equipment to
minimize or hide anything.
All of these are fixable, though some are harder than others. The
hardest is that everything should just work, as easy as the telephone,
at least. Some things, like the white board and large conference rooms
probably require multiple cameras, possibly even cameras which
automatically focus/zoom in on the speaker. If you've ever seen a
broadcast conference... or awards show, what you basically want is
multiple camera angles and intelligent cameras, but all automatic, no
one working all of that. The AV crew for our larger "all hands" style
multiple location conferences is easily 5-10, what we need is software
intelligent enough to give us a close approximation. And for the prices
of this equipment, that's what I'd expect.
But personally, all I need is a box that has an HDMI output, a good
wide-angle camera with intelligent assist for contrast, that can auto-scale
picture/audio quality based on connection speed, and a good intelligent
mic with echo cancellation, etc... for about $250-300. I even debated
started a company just to do it...