The list of established players in the imaging field is a long one. Nikon, Canon, Panasonic, Leica, Olympus, Pentax, Kodak… it goes on. For decades they’ve been fine-tuning their devices, and they continue fight fiercely over every market and price point.
Certainly this has produced some excellent devices: DSLRs today offer an unprecedented value for the amateur (or pro) photographer, but I can’t shake the feeling that all the big guys are spinning their wheels. After all, there are precious few real innovations in cameras these days — Casio and Fujifilm spring to mind with their innovative use of the sensor, but by and large, even the top-tier devices don’t really do anything different from the ancient one-megapixel point-and-shoots of the late 20th century.
Researchers at Stanford want to change that. Although they certainly don’t plan on toppling the powers that be (in fact, they’re funded by them), they’re tired of cameras falling under either the highly-specialized or highly-generalized categories. After all, it’s all just data, right? Why not make the camera a versatile platform with a real OS, an open hardware standard, and — hell, why not — an app store?
Okay, that’s really pushing it. But the idea is sound. Cameras these days have the potential to crunch and store data, connect to other devices (or the net) via wi-fi, and even have high-resolution touchscreen displays. Sure, they’re set up for a streamlined workflow of converting sensor data into RAW or JPEG files or adding scene adjustments, but that’s a matter of software.
The “Frankencamera” created by Mark Levoy’s lab uses essentially off-the-shelf parts: a TI system on a chip, Canon EF lenses, a generic LCD screen, and a Nokia N95’s photo sensor of all things. It’s ugly, bulky, and fragile, but that’s the way prototypes tend to be. Even the mighty iPhone used to be raw PCB, rudely wired to a dev rig — Apple didn’t just fill a plastic shell with sex.
The setup they’ve created (it runs Linux) lets them control and program each aspect of the camera with precision. With a standard API they could release it into the wild with a few prototypes and hackers would be able to go to town on it. How about a double curtain exposure at two aperture settings to make a weird hybrid bokeh? Or connecting the camera to an RGB LED flash, which could evaluate the scene and tint the flash to best illuminate it? I’m just making this stuff up (and have no way to follow through, damn it all), but a skilled hacker and photographer could think hard about it and make a truly new way for a camera to function.
In a perfect world, that is. The truth is that, with companies like Kodak and Adobe mothering the project, any real advancements would be likely be integrated piecemeal into existing camera designs, or licensed and commercialized. Although they plan to make sub-$1000 units available to researchers and hackers, I wouldn’t expect a consumer Frankencamera to grace the shelves of Best Buy any time soon.
It is, however, a fun idea to play with. The camera really is enough of a specialized platform that it could be one of the few devices that doesn’t succumb to convergence over the next decade (I don’t share TechCrunch’s optimism about phone cameras). Every other device we interact with is getting smarter, so why not the camera? Sure, you’ve got smile shutter and 100 scene modes and all those features we don’t want or need, but that’s not smart. It’s convenient, and barely that. And you’ve got HD movie mode now — great, but while admittedly the resolution is slightly higher now, my 2MP Sony was doing movies in 2001.
A smart camera (a device we’re sure to see soon) will be connected, possibly even social, and will be materially upgradeable like a smartphone or PC. More capability, more usability, more connectivity — the way the phone went from a way to call home from the car to a complete media and communications platform. I don’t think the scale of the business opportunities created by that particular market can be overstated in this forum, and I’m sure our readers can appreciate that a similarly improved camera might lead to quite a harvest by imaging and design businesses. RED has shown that there is room for new players in the game if you come in heavy, though of course they were lucky enough to have a visionary benefactor.
And remember, just because it’s smart doesn’t mean you’ll have to be smart to use it. Just look at the iPhone. Well — that didn’t come out right, but you know what I mean. It’s no coincidence that the current generation of cameras resembles the dying breed of phones Jobs spoke so truly of when he unveiled the iPhone: buttons all over, inflexible and static in design. That’s the enemy. Unlike the old phones, I believe those cameras will have a place in the new order of things, but their descendant, this new fantasy device I am so excited about, will be a thing of beauty.
Just (please) don’t call it Camera 2.0.
[via PhysOrg]
Good on them. We need more open standards and technologies when it comes to imaging (unlike the iPhone!).
Yes, although the iPhone was an important reference point for this particular idea, it’s really a poor comparison in many other ways.
Yeah looking fwd to see such a change. Lets see if it get adopted well
I sat in on a presentation in 2006 (http://www.stanford.edu/class/ee380/Abstracts/060118.html) where Andy Rubin (who heads up Google’s Android Platform) described just such a device, and showed a prototype. This was apparently started in 2004.
BTW, the camera prototype was called Android. :)
Ironically, it was a lesson in a business idea that failed to make it, perhaps because in consumer electronics devices, endusers don’t care as much about uber flexibility to install boatloads of stuff as they do about the simplicity of point-and-shoot.
Yeah, the business case doesn’t seem strong enough. Cameras already do advanced on-board processing (e.g., Canon does a fantastic job), and the “fixes” like brightening the shadows are really creative decisions, at least at the pro level: pro photos have deep shadows because the photographer wanted the shadows.
Nice article – you really get where we’re coming from!
Ray: I was always disappointed that Android failed in its initial incarnation. Fortunately in academia we don’t have to care so much if something is marketable or not ;)
- Andrew (The guy in the green shirt in the picture)
Hey, man. Thanks for dropping by. Hope the project continues to rock for you guys.
For now, can a company just introduce a great DSLR with great HD video? Companies have been doing the video with still or still with video, thing for years, but only recently has the technology reached a high level of quality. And even the venerable Canon 5D Mark II is good with stills, but the video could still use some improvements.
Who will be the first company to offer the perfect combination of great video and great stills? Maybe the next Olympus pro-model will do both, and best the 5D in price too?
Jason! The Panasonic GH1. Does it all, well.
True, but the GH1 leaves a little to be desired as far as still image quality. I am really hoping the new Oly will hit the sweet spot. Plus I already have some of the phenomenal Zuiko glass.
“Who will be the first company to offer the perfect combination of great video and great still”
Perfect combination? Heck, no two competing companies will define that the same way (unless they’re about to merge).
iphone is just like a box, a close box.
I like open things.
The more we get, the more we share.
One suggestion for all of these electronic manufacturers to take a step forward and waterproof their goods or would that eat into their replacement cycle too much? i.e. $800 dslr ruined because a drop of water hit it.
Rich, I owned a Pentax K10D, splashproof dustproof DSLR. It was outstanding and proved it’s worth the day I bought it – I was caught in a rainstorm in London (happens!) – got some great ‘wet city shots’ too. Despite both camera and I getting a soaking, the pictures were great and the camera was unharmed – as it was designed to survive both British and sub Saharan weather. The subsequent Pentax K20D and new K7 are equally robust. K7 does video too.
Splish splosh, snap snap.
I’m glad to see any type of open hardware; however, there is nothing intrinsically new about this hardware device. Essentially it is a commercial CCD or CMOS imager hooked up to an embedded Linux machine (such as a Beagle Board).
Something that would be much more exciting would be an open hardware initiative to build “computational cameras” — such as ones that employ assorted pixel masks for high dynamic range or actuated imaging elements to produce flexible depth of field.
If you’re into the idea of “computational cameras”, there was a really cool plenary talk at a conference earlier this year — check out some coverage at Hizook.com
http://www.hizook.com/blog/2009/06/26/computational-cameras-exploiting-megapixels-and-computers-redefine-modern-camera
Travis,
That sort of stuff is one of the target applications – while lots of people have partially hacked up cameras to add interesting stuff in the lens aperture (coded aperture) or in front of the image sensor (plenoptic camera) to name a few, it’s a slow process, and the existing camera SDKs make using such a thing very slow, and tethered to a laptop. And the result is rarely portable in any way.
So we’re building this camera and API in part so that experiments on novel hardware could be done much easier, and the result will hopefully be far more usable and portable than before.
-Eddy (one of the other graduate students on the project)
Eddy,
Thanks for chiming in. I hope the project really takes off, and I’m also hopeful that it will ultimately lead to a budding computational camera community with open hardware. I was certainly not trying to belittle the contribution; it seems like a step in the right direction.
Oh, no worries. The thing really is just a pile of off-the-shelf parts, intentionally so (cheaper, easier to reproduce, much bigger community). We’d actually use the BeagleBoard itself if we could (it doesn’t break out the camera interface pins from the OMAP3 chip, unfortunately).
The current hardware is mostly a platform for developing an API on – we’re hoping there’ll be a wide range of devices that at least some subset of the API will run on.
you know what’s cool?
film cameras.
I’m a fan of technologies and such,
but I really don’t want to be around the day my digital camera decides to kill somebody.
simplicity takes great pictures too.
with the cameras these days you don’t even have to have skill to take a beautiful picture. rather sad in some respects
because hardwork is dying!
You mean phones, cameras, and PDAs and laptops are all sort of veering towards one homgenised device!? no way? when did this start happening? Oh wait… phones with cameras… cameras with wifi… pdas with everything… phones with internet… wait a second…
“oh no I was wrong, it was Earth all along… you’ve finally made a monkey out of meeeeeeeeeeeee!”