Revisiting
"Need For DV Specialization"

One Year Later

Daniel Joseph Pezely

Original: 22 March 2002
Annotated: 7 May 2003

www.play.org


History repeats itself in strange ways. Understanding this, I'd like to make a prediction. In the near future, DV camera makers will expand their offerings into more than just the cost-versus-performance spectrum.

We'll start seeing cameras with pluggable modules for customization.

While we have yet to see swappable codecs, there are camers with multiple encoding standards built-in: MPEG4 via ethernet, DVCAM via disc, DV to tape.

We'll start seeing single-chip cameras encoding to HD-- not that the single chip would necessarily be anything better than the current low-end miniDV chips.

At NAB2003, both Panasonic and JVC had HD camcorder for the US$4,000 range.

We'll start seeing high-end cameras with QuickTime, MPEG or even loss-less codecs.

MPEG already mentioned; the digital film 2k and 4k resolution cameras from Thomson Grass Valley and others are considered loss-less for their resolution.

I've been a researcher and consultant in the software development business for twelve years. Before that, I was an engineering CAD consultant-- computer aided design.

Now, I consider myself part of the film community. The early part of 2002, I was producing a documentary. I made some commercial videos and have been following the technology over the past couple of years.

These experiences have revealed something to me: There are parallels between the DV world and my past careers.

In the early days of CAD, vendors attempted making their products to be all things to all people. One package was supposed to work for architects, mechanical and civil engineers, artists and graphics designers, etc.

By the mid-1980's, they realized enough to separate the artists from the engineers because their respective needs went in different directions.

We see this in the DV software market today: FinalCut Pro caters to storytellers reaching for a new medium while the current version of Toaster aims to make the career video engineer feel at home.

Eventually the CAD vendors figured out enough to offer a base package that could be customized. A cottage industry blossomed offering extensions.

And in the software side of the video world, we have similar possibilities with the various editing, effects and compositing packages.

Such as the niche that Media 100 has staked out for editor-compositor types with 844/X

It took less than a decade for the CAD industry to accommodate your needs however unusual.

Again in the world of movie making, we see this being played out with the various effects tools: AfterEffects for those probably coming from a PhotoShop background versus inferno/flame for the ILMs of the world and discreet combustion for Post houses in between.

For the world of CAD and 3D modeling, those early days are over. The various elements are being merged again. The software architecture is probably similar inside whether you're using 3ds MAX as an animator or AutoCAD as an architect. (The same company, AutoDesk, owns both even though one is in California and the other, Montréal.)

But there's a bigger picture here.

In 3D animation and modeling, for example, it doesn't really matter today whether you're using Maya, 3ds MAX, SoftImage or LightWave. (Blender has some catching up yet, but they'll get there, and hey-- it's free!)

Today, it's more a matter of taste. If you prefer the command and control aesthetics of Maya over MAX, fine. If SoftImage's organization and layout melds with your mind appropriately, fine. Or if you've used LightWave since the early days of the Amgia, you're probably considered a goddess or god by now anyway, so stick with it.

They all pretty much do the same thing. It's taken a while, but it's there for the 3D crowd.

After the industry has reached maturity, the differentiation becomes the user interface. (But even that will probably be configurable after another version or two in the realm of 3D.)

But digital video has yet to reach that maturity.

After all, DV started a little later than CAD and 3D, even though we saw the early HD prototypes in 1986.

That..., and there's the fact that software is better able to have shorter generations and lifespans.

The manufacturers of DV gear need to get clued in to this.

Why is it that the makers of digital video technology are seemingly trying for the one-size-fits-all approach?

Yes, there is a spectrum of camera technologies but only varying in cost-performance terms.

Hasn't anyone noticed that film comes in various speeds suitable for different situations such as low light or high shutter speed? The same camera accommodates the different types of film. The major specialization with film cameras now has to do with mechanical noise and insanely higher frame rates for effects.

Yet why do we accept that the same DV technology will handle all situations?

Why not a cheap Digi-Beta with a permanent lens?

How about a 2K resolution version of the GL1?--Sony, give Cannon a run for the money!

News reporters and field photographers need one type of camera, and studios can benefit from others. Things like battery life are almost insignificant on a sound stage.

JVC was testing the waters with a new HD box camera at NAB 2003: yet unnamed, it mounts on a tripod, accepts a throttle but is without any storage. There's just a couple of output jacks on the back. There seemed to be so much attention from the public on that little camera that it's a safe bet it will ship by autumn 2003.

Rather than debate which DV camera lends itself for the best film transfer, I'd rather see a class of cameras optimized for that and others suited for end-to-end DV.

Once again, NAB 2003 showed a number of 2K and 4K resolution cameras. They came from all over the world. The 4K models may have been just prototypes, but now that principle film production has been using the Kodak VISION2 film that is nearly grainless, 2K is suddenly insufficient.

Why not a camera that encodes directly to QuickTime or MPEG? For most filmmakers, they're on a set-- a controlled environment with an electric supply-- so arguments about processing power and batter life are moot.

Done. And streams via ethernet.

The Atari game generation grew up using plug-in modules to play different games. Your Web browser uses plug-ins to show QuickTime, Flash and PDF.

So why not plug-in modules to your camera to achieve different features? We already change lenses on everything from pro-sumer, on up. Give me a plug-in codec. And perhaps a cartridge for changing aspect ratios, so we can junk that software-based pixel squashing.

Calling all DV gear vendors: please think outside the consumer electronics box!

Heck, the current version of Video Toaster can run on any home computer sold today, and it performs real-time effects without additional hardware so you really do have full studio functionality.

There's no excuse for all cameras being exclusively NTSC, PAL or 24p. That should be switchable as easily as changing lenses. That should be what divides pro-sumer versus high-end.

Done.

And there's still a market for gear with all fixed features: tourists and extended family.

Some people say that specialization can be bad, that it leads to fragmenting the standards and that it makes more confusion.

The point is to understand that there is a time and place for everything. Until the technology and standards mature sufficiently, specialization works like natural selection. Some of the progeny will become extinct, the rest will lead to gear we actually find useful.


Copyright © 2002-03 Daniel Joseph Pezely