Here's the moment where you pretend that a breakthrough in a given industry would just revolutionize the way you do work, yet you know -- deep down in your heart -- that you'd never take advantage. Okay, so maybe you would, but your friend wouldn't. At any rate, a gaggle of boffins at Stanford have set out to "reinvent digital photography" with the advent of the open-source digital camera. The idea here is to give programmers the power to conjure up new software to teach old cameras new tricks, with the hope being to eliminate software limits that currently exist. In fact, a prototype shooter has already been developed, with the Frankencamera hinting at a future where owners can download apps to their devices and continuously improve its performance and add to its abilities. The actual science behind the concept is stupendously in-depth, so if you're thinking of holding off on that new Nikon or Canon in 2034, you might want to give the read link a look for a little more encouragement.
I think they're promoting it incorrectly. From the video, I still don't see what this will let me do that I can't do already. All they want to do is move some post processing into the camera. That's already starting to be done to death by the manufacturers. I prefer to preserve the initial images and perform my post processing externally.
Where they do have potential is in the possibility of modularizing camera components. If an open standard is well defined for the firmware that brings us one step closer to a modular and multi-company options.
Not just camera components - it would be great to be able to fix camera interfaces, and at least some of that could be done without adding any extra knobs, dials, screens, LCD readouts, but just using the existing ones better. (Of course 'better' is my opinion and other people might have different ones, that's sort of the point.)
For example, if I had (a) the capability, and (b) the time to rewrite the firmware for my [Canon] DSLR, I'd do the following:
1) Make the overexposure highlights accurate to RAW data, not JPEG. 2) Make it display on auto-review (as an inset or allow toggle between) a small 1:1 section of image around focus point so you can see if the shot was sharp. 3) Declutter the interface by removing all settings that affect JPEG only (I only shoot raw). 4) Make it detect when the shutter speed I'm about to use is probably too slow (according to lens focal length) and flash the shutter speed indicator so I actually notice it.
BTW, the reason that pro level cameras don't have a flash is because they are too under powered to an adequate job of lighting for professional purposes, you are not able to bounce the light which leads to hot spots, and it sucks way too much of the battery.
I remember my old Kodak DC290 2MP camera with Digita OS. It was a great idea, which allowed developers to make apps for the camera. I remember one digital distance calculator which was amazingly precise, and of course MAME which was actually ported to it. I still don't know why such tech never moved forward....
Anyone remember the Digital scripting language on a couple of cameras 10+ years ago? Everything old is new again... I ran Doom on my old Minolta EX1500...
This looks nice, i always wanted to get more access to how my camera works. Making the camera do what you want while taking a picture is surely better than post-processing to get the desired effect.
OMG! People you are missing the point! What has it been 4~5 years since we heard of the 90.000 microfilm-lenses. http://www.newscientist.com/article/mg18825255.000 It's great to hear that they are making some progress, although he only mentioned that they could put it in between the camera and the lens, not that they had made a final product :( BTW I thought that 4/3 cameras were open to manufactures...
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
I've heard rumors that this camera is a first step in the "computational camera" direction -- namely towards cameras that have more advanced physical features and are more amenable to hacking
to me, the camera (like a scanner for example) is used to capture, for a lack of a better word, raw data. the issue i see with some of the software processing on the camera is that it is more processing power that will be required on the camera especially if you shoot RAW. after i load it onto the computer, i can use whatever program i want to edit the photo. at an event, the last thing i want to do is to attempt to do post processing on the spot and later learn it wasn't what i wanted and lose the originals.
if this were an idea for more P&S style cameras, then i can see it being useful since at this level, most people don't consider using products like adobe and such. for example, quadcamera, an iphone app would be useful to have on a P&S ready to go. it's a fun app.
Yes, there was the Digita OS that ran on a number of different cameras back in the 90s (Kodak, HP, Minolta, Pentax). It included a scripting language and an SDK that allowed you to do just about anything on the camera. There were games, photo editors, custom slide show and image capture utilities, MP3 players and a user-accessible, simple scripting language that allowed you to control most of the basic features of the camera with just a text file. Interestingly enough, this technology is rumored to have come out of Apple as a follow on to Apple's QuickTake digital cameras.
It would be nice to shoot a camera as close to its hardware limitations as possible instead of being limited or handicapped to lower specs based on how much it costed just so they could make extra money giving us less.
on the same side it would also be nice to get real specs on hardware, instead of that whole fuzzy math associated with MP and interpolation BS.
Just give us hardware that works, that what was so great about old school film, if you didnt get a good shot, for the most part, it was because your settings or composition was bad. sure film choice and iso an such made a difference, however a great photo could be gotten with a childs camera, if you have a shitty digital cam, you really need optimal conditions, optimal lighting, optimal composition and maybe youll get a good shot, however most likley if you have a small shitty sensor with a plastic lens, youll most likley get a pixelated, grainy, photo with wiered colors, grays for blacks, fringing, and all the other associated bad junk you get.
This sounds fantastic. I believe that software is what is holding back cameras the most currently. And the camera manufacturers are very slow at adding new features.
The kodak dc290 had an operating system everyday people wrote code for. Along with a bunch of useful little photography apps, people wrote game emulators, calculators and all sorts of stuff for it. I'd sit on the train playing doom on the camera. Sure, it got a little hot, but doom... on a camera!
The pictures soon will be better than what you see with the naked eye. Its already there don't you think! I take a photo look at it later and say to myself I don't remember it being that beautiful.
Now that we've thrown 'em off the trail, use the form below to get in touch with the people at Engadget. Please fill in all of the required fields because they're required.
posted September 6th 2009 1:10PM
Good Job Eddy! He gave me a tour of this stuff in person in the lab. It was pretty cool.
I think they're promoting it incorrectly. From the video, I still don't see what this will let me do that I can't do already. All they want to do is move some post processing into the camera. That's already starting to be done to death by the manufacturers. I prefer to preserve the initial images and perform my post processing externally.
Where they do have potential is in the possibility of modularizing camera components. If an open standard is well defined for the firmware that brings us one step closer to a modular and multi-company options.
Not just camera components - it would be great to be able to fix camera interfaces, and at least some of that could be done without adding any extra knobs, dials, screens, LCD readouts, but just using the existing ones better. (Of course 'better' is my opinion and other people might have different ones, that's sort of the point.)
For example, if I had (a) the capability, and (b) the time to rewrite the firmware for my [Canon] DSLR, I'd do the following:
1) Make the overexposure highlights accurate to RAW data, not JPEG.
2) Make it display on auto-review (as an inset or allow toggle between) a small 1:1 section of image around focus point so you can see if the shot was sharp.
3) Declutter the interface by removing all settings that affect JPEG only (I only shoot raw).
4) Make it detect when the shutter speed I'm about to use is probably too slow (according to lens focal length) and flash the shutter speed indicator so I actually notice it.
etc etc
Errrr.....not too compact is it?
BTW, the reason that pro level cameras don't have a flash is because they are too under powered to an adequate job of lighting for professional purposes, you are not able to bounce the light which leads to hot spots, and it sucks way too much of the battery.
Does this allow the rise of the iCamera, where Apple will provide software updates?
Only for a small fee...
I still want the Panasonic GH1
great... now make an open source ebook reader already!
I remember my old Kodak DC290 2MP camera with Digita OS. It was a great idea, which allowed developers to make apps for the camera. I remember one digital distance calculator which was amazingly precise, and of course MAME which was actually ported to it. I still don't know why such tech never moved forward....
Anyone remember the Digital scripting language on a couple of cameras 10+ years ago? Everything old is new again... I ran Doom on my old Minolta EX1500...
Interesting concept...
This looks nice, i always wanted to get more access to how my camera works. Making the camera do what you want while taking a picture is surely better than post-processing to get the desired effect.
I use 'P' mode, thankyouverymuch
OMG! People you are missing the point!
What has it been 4~5 years since we heard of the 90.000 microfilm-lenses.
http://www.newscientist.com/article/mg18825255.000
It's great to hear that they are making some progress, although he only mentioned that they could put it in between the camera and the lens, not that they had made a final product :(
BTW I thought that 4/3 cameras were open to manufactures...
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
I've heard rumors that this camera is a first step in the "computational camera" direction -- namely towards cameras that have more advanced physical features and are more amenable to hacking
cool..they're demonstrating that cam on a pic from Zurich in Switzerland
to me, the camera (like a scanner for example) is used to capture, for a lack of a better word, raw data. the issue i see with some of the software processing on the camera is that it is more processing power that will be required on the camera especially if you shoot RAW. after i load it onto the computer, i can use whatever program i want to edit the photo. at an event, the last thing i want to do is to attempt to do post processing on the spot and later learn it wasn't what i wanted and lose the originals.
if this were an idea for more P&S style cameras, then i can see it being useful since at this level, most people don't consider using products like adobe and such. for example, quadcamera, an iphone app would be useful to have on a P&S ready to go. it's a fun app.
Yes, there was the Digita OS that ran on a number of different cameras back in the 90s (Kodak, HP, Minolta, Pentax). It included a scripting language and an SDK that allowed you to do just about anything on the camera. There were games, photo editors, custom slide show and image capture utilities, MP3 players and a user-accessible, simple scripting language that allowed you to control most of the basic features of the camera with just a text file. Interestingly enough, this technology is rumored to have come out of Apple as a follow on to Apple's QuickTake digital cameras.
It would be nice to shoot a camera as close to its hardware limitations as possible instead of being limited or handicapped to lower specs based on how much it costed just so they could make extra money giving us less.
on the same side it would also be nice to get real specs on hardware, instead of that whole fuzzy math associated with MP and interpolation BS.
Just give us hardware that works, that what was so great about old school film, if you didnt get a good shot, for the most part, it was because your settings or composition was bad. sure film choice and iso an such made a difference, however a great photo could be gotten with a childs camera, if you have a shitty digital cam, you really need optimal conditions, optimal lighting, optimal composition and maybe youll get a good shot, however most likley if you have a small shitty sensor with a plastic lens, youll most likley get a pixelated, grainy, photo with wiered colors, grays for blacks, fringing, and all the other associated bad junk you get.
This makes me giddy about the future, and also makes me wish I had a fleet of computer scientists at my disposal.
Oh my. That Frankencamera is the size of God's ass.
http://blogs.elphel.com/2009/09/elphel-and-stanford-frankencamera/
This sounds fantastic.
I believe that software is what is holding back cameras the most currently. And the camera manufacturers are very slow at adding new features.
I doubt many people would hold on to their old camera though auto. The current trend are dSLRs.
The kodak dc290 had an operating system everyday people wrote code for. Along with a bunch of useful little photography apps, people wrote game emulators, calculators and all sorts of stuff for it. I'd sit on the train playing doom on the camera. Sure, it got a little hot, but doom... on a camera!
The pictures soon will be better than what you see with the naked eye. Its already there don't you think! I take a photo look at it later and say to myself I don't remember it being that beautiful.
I can't wait to see what this program or photo's will look like.