Open Camera Project

[joseph_sanders]

So we can build software via. open source. Could we build a camera system? How

much of the cost of a camera is attributable to marketing and development? I

know the initial gut reaction is "no". Cameras are complex and there is a

marginal cost to every unit whereas software has no marginal cost per unit.

Lets just set that aside for a moment. Lets answer the question: What would an

open source camera look like?

 

Here is (one of) my answer(s):

 

1) square format (not necessarily 6x6 though. possibly Nikon/canon mount.

a. User selectable orientation horizontal, vertical or square.. rotate able

back?)

2) separate body/lens/back... possibly separate finders and winder/motordrive.

3) diamond (square) pattern AF sensors. One sensor at each rule of thirds

intersection, one sensor in the center, sensors centered vertically and

horizontally just outside the box created by the rule of thirds sensors.

4) lens based focus motors

5)

6) USB port for firmware updates. (open source firmware obviously)

7) open flash protocol.

 

 

What if Tamron donated the designs to the bronica sqai or RF as starting point?

 

What if foveon was willing to donate some engineering in an effort to create a

market for their sensors?

 

In my little sci-fi world the open camera project would aggregate orders in

blocks of 1000 or 5000 or whatever made sense for demand. Non refundable

deposits would be taken to hold a spot in the production run. Full payment

before production is ordered from manufacturing sub-contractor. Abandoned

deposits would lower cost for the rest of the buyers.

 

I know this seems crazy... but I think it could be the next step in economic

evolution. It not really socialist.. it exists in a market but it is very

decentralized and design is "owned" by no single person or entity. Look at what

Lego had done with their kit design contests. Look at the system they developed

to enable the sharing of designs among their customers.

 

 

-Joe

 

If you think this isn't even worth discusing, put your money where your mouth is

and don't post.

 

If you think it impossible, you are probabally right, tell me why.

 

Lots of things are impossible, but still worth discussing.

[doug grosjean]

Your concept of an open-architecture camera already exists, to some degree.

 

Google on "Lookaround" camera. It's a homebuilt pano camera. You buy the book on how to build it, but you don't buy the camera itself.

 

Great concept. And I think it would work for some other basic landscape or large format cameras.

[ben conover]

Hi Joseph, interesting!

 

Fuji nearly got it right 30 years ago. I'd go for a re-make of the Fuji g690bl. I was actually thinking of this last night, until 3am.

 

"big camera" 690 format 120 film rangefinder from solid titanium, with light meter, and lenses from 50mm to 250mm. Price would be at least $5,000. There would also be a digital back for this mutha, that would be an extra heap o dosh.

 

Cheers.

[klix]

<i>So we can build software via. open source. Could we build a camera system?</i>

<p><p>

When you say camera system, are you talking about the software (or firmware) for a camera system? If so, then wouldn't that depend what will work with the hardware platform? And if that's the case, why would Nikon, Canon, etc... allow that?

<p><p>

Or, when you say "could we build camera system," do you mean the hardware? If h/w, then you're comparing apples to oranges when comparing open source SOFTWARE to camera HARDWARE. And the answer is still NO -- UNLESS, you have money and resources to start your own company to compete with Nikon, Canon, etc... Then you can do whatever you want.

<p><p>

Come to think of it, given an infinite amount of resources -- time, people, money -- you can pretty much do whatever you want.

[wigwam jones]

Stone soup!

[pico_digoliardi]

Great Think Tank work comes from just such provocative challenges. Let's go for it! Very cool idea! Even if nothing comes off it as an actual product, we can possibly spin off into good ideas nonetheless. That's how it works.

 

Pretty busy at the moment, so just an idea: Modular. That is, of components that can be swapped up to better ideas. Like perhaps a back designed after Hasselblad's. Why, a person could revert to film if something went bad, or if he just wanted to, then go back.

 

A smaller, say 40mm x 40mm sensor would fit the bill. The 'back' could contain slotted modules for hardware updates.

 

Truly, in software there is a movement towards more of a central 'thematic' design with plugin objects. Start simple, plug in more as you need it, and unload modules when you find something better or just want to toss it. A Open Source camera could be the same thing.

 

Now, let's for a moment consider prescedents - there are none that I know of, but we do have considerable modularity in view cameras. Dunno if that's at all applicable here.

 

Gotta run. Got 40 web pages to rip out in 1/2 an hour. :(

[doug grosjean]

Pico is on the right path.

 

You pick out some really common existing architecture that has the features you need.

 

You buy used lenses, backs, etc off Ebay.

 

You end up just needing to produce a body.

 

I'm not a camera designer (though I am a mechanical designer), but building an open-source all-manual camera would be relatively easy.

 

Digital? With sensors, and packing the electronic guts from some existing common units into the space you have - *that* would be the tricky part.

[remco-jan.woldhuis]

In theory it should be possible. The practical difference with open source software is the

collaboration between the developers and manufacturers. Software can be written

everywhere in the world by virtual teams, and can be distributed free over the Internet. For

open source camera's you need physical channels, which is infinitely more costly and

complex.

[joe tarrant]

<P>

The core things you probably need are a light-tight box, a viewfinder, a (fixed?) prime

lens, shutter, sensor, storage, a power source, and input/output (ie, a USB port). This

model assumes manual lenses, manual controls and a zen attitude on the part of the user.

You're going back to basic photography: think Holga, Rolleiflex, unmetered Leicas. And

don't worry about how many megapixels: take whatever you can get to start with. And the

less in the way of electronics, EVFs, LCDs, etc, the less power the thing will use.

 

<P>

If you've never used anything like this sort of camera, let me reassure you. I only started

using fully manual cameras three years ago, but getting a good photo from 1930s-style

cameras isn't that hard. Do a search for Holga or Lubitel or TLR here or on Flickr and you'll

see how people just learn from doing. The easiest solution would be to think of the

camera as a 1930s camera, plus a sensor, storage, power and input/output. Nothing else.

A 21st Century Barnack Leica.

 

<P>

Thinking about it, if someone like Cosina created the electronic guts of this camera and

then sold it to anyone who wanted to fit it in their own body, there could be a whole

cottage industry out there for anyone who wanted to start a little camera company. Kind of

like the way that Intel sells motherboards to anybody who wants to make a PC. You could

make cameras from biscuit tins, wooden boxes, little bits of plastic, small moulds. You

could have plain ones, leather ones, big red ones. You could create repro classics, like

1930s Rolleis. Whatever you liked.

[doug grosjean]

Joe T. wrote:

 

"If you've never used anything like this sort of camera, let me reassure you. I only started using fully manual cameras three years ago, but getting a good photo from 1930s-style cameras isn't that hard. Do a search for Holga or Lubitel or TLR here or on Flickr and you'll see how people just learn from doing."

 

Seconded. Been doing the same for about 2 years myself: Ciro-Flex, Rollei TLR, Zeiss folding Ikonta, etc.

 

 

Joe T. wrote:

 

"The easiest solution would be to think of the camera as a 1930s camera, plus a sensor, storage, power and input/output. Nothing else."

 

Yes, exactly.

 

I wouldn't want to donate a Rollei to the excercise, but I have Ciro-Flex that would be hilarious to convert to digital. But then it's a conversion, and not an open source camera.

 

Doug Grosjean

[wbowman]

Generally and open source project is in reaction to a particular need that is not being

addressed by a particluar commercial vendor. For example, open source software projects

are often sustained by contributions from developers who are addressing a specific need

that someone like Microsoft is unable to meet for a number of reasons. Another is to

distribute R&D costs across a large group of interested people. Yet another is to break a

hold that a specific company has on some piece of intellectual property for a process.

 

Like any good project, the first thing to determine is what problem is in need of a solution.

From the initial post and the subsequent responses I am unclear on what photography

needs are not being met by current market providers.

 

I think this is an interesting idea and deserves discussion. Just let me know what I'm

missing with my current systems so I can contribute my thoughts the design.

 

I guess I want a Canon FTb with f-stop and shutter speed vible in the finder. Everything

else has limited value to me.

[peter_yee]

It is possible, you look at the example of large format cameras. There, the film holder for 4x5, 8x10, 11x14 (digital SLR equivalent would be the sensor and the associated electronics) has an ANSI standard. Manufacturers of film holder can make a holder as cheap or expensive as they want as long as the film holder meets the specified standard. The digital equivalent would be the sensor will fit within a certain size. The support electronics will have a specified battery power, input/output ports, specified data structure etc.

 

The lens mount sizes for large format and the operating shutters are basically standardized (or similar enough so simple adaption is possible), so a newer lens from Fuji, Nikon, Rodenstock, Schneider will work as well as older lens or even ancient lens such as Kodak, Zeiss, etc, will work, on most new or more old models of large format.

For the digital SLR equivalent, you probably go with an existing lens mount whose patent or copyright have expired (Such as the M42 mount, Nikon manual F mount, Canon FD mount (?). That way a whole range of lens will be available. Sadly, no autofocus and probably tied up in patents.

 

In the large format camera, the camera body is then left to the imagination of the manufacturer as long as the lens and film holder fits.

 

As a result there is incredible number of large format camera makers, ranging from ones who only sells you drawings, or complete do it yourself kits, to very expensive high end cameras manufactures.

[conrad_fischer1]

Some musings on the subject: warning the following thoughts could be self-contradictory

 

I would guess that the eventual product of any such exercise would be pretty big. In my mind, each component, the body, lens, and film/sensor back would all have to be self contained and self supporting. That would include a separate power source for each item.

 

Another thought is to create a standard interface for each component and perhaps add a power source as a component. The standard interface would be a bear to design in order to be all things to all people. Camera companies presently make a lot of compromises in order to economize in dollars and size.

 

Certain technological advances would be precluded as the camera companies would not be likely to relinquish their patent rights. That said this could also be a great test bed for proposed technologies, just be prepared for a lot of ideas that may be less than practical.

 

Well enough of that, I am now confusing myself.

[joseph_sanders]

Thank you all for posting. I�m superized at the response. I really want to write a 3 page response to each poster but I can�t or I�ll die. So I�m going try to break things down in brief orderly fashion. Like my original post comments (generally) fall along two line of thought:

 

What would be cool in a camera.

 

And

 

It will work IF (fill in the blank)

 

I�ve realized that for the project to work it will have to move large numbers. To move large numbers is will have to have broad appeal and serve a segment of the market better the current offerings (thank you Wesley) Which brings us to:

 

What would be cool in a camera? There are lots of answers to this question. I�m big fan of the 6x9 120 cameras mentioned by. Ben. And the Bronica 645 RF. But I�m not sure that these type of cameras will have mass appeal.. not because they are not great cameras but because they have limited flexibility. I�ve been eyeing those Fuji 6x9�s with changeable lenses all summer� but with interests in many different types of photography I�ve come to the conclusion that the 6x9 RF is going to have be a second or third system for me. I think to achieve critical mass we have to rethink the original question. We should be asking:

 

If you had to live with ONE camera system for the rest of your days on this planet.. and you could influence the design of it what would it look like.

 

I personally think that anything based on a small sensor (or film) size will fail because the market had done a great job of meeting those needs. Canon, Nikon, Minolta, Sony all are making great products in that arena. However the gap between Nikon/Cannon/Sony and Hasselblad is MASSIVE. I think that is the proverbial hole in the market. Not to be dismissive, but I think the any older 35mm based cameras won�t sell. People can just buy a used one.

 

I�m convinced modular is the way to go. The body, lenses and accessories could be used across several generations of higher and higher quality digital backs or mated to film backs. I�m intrigued by using an available mount Nikon/Canon? My guess is that there would be some patent trouble though. I�ll bet sigma would go for as it would mean selling more of their lenses. Perhaps a unique lens mount with adapters.

 

I�m also intrigued by the idea of convincing Tamron to kick down the guts of their Bronica cameras.

 

 

How about a sensor that would capture the entire light circle of a 35mm (or 645) format lens. (40mm in dia?) With a digital back you could choose you orientation on the fly horz/vert/square. No flash brackets!!

 

Ok.. enough from me about what I think would be cool.. on to: IS IT POSSIBLE? The biggest obstacle to possible is will people buy it. As I read Wesley Bowman�s post for the first time my knee jerk reaction was �no he is wrong� there was not a real looming commercial reason for Linux at its first iteration therefore there doesn�t necessarily need to be burning hole in the market for success. But upon further reflection I see the point: Why would people buy the Open Camera as opposed to a solution on the market already? That is the big question. If the Open Camera is only as good or flexible as the other solutions on the market the only carrot is price. Which brings us too two other thoughts: 1) How much of the price of a camera is attributable to marketing, profit and development? If the Open Camera can be %95 as good as Hassy H2 and cost %40 as much folks will buy it. 2) What will the open camera do that an H2 won�t?

Either the Open Camera has to be way cheaper or way better or some compelling mix there of. Hopefully we are addressing the �way better� first section of this post (what is the most flexible camera system)

There are those that think it just out right impossible.. though I haven�t heard the reasons. For the sake of discussion I�ll reject that out of hand. This guy is trying to design a freaking Helicopter from an entirely new direction:

http://www.unicopter.com/UniCopter.html#top

He may fail.. but I don�t think its impossible.

We would need CAD software. We would need to be able to test the camera�s mechanicals inside a computer 10,000 times before machining a single part. I know there are software�s that can do this�. But are they cheap enough to spread copies around to 1000+ volunteer developers.

 

http://www.emachineshop.com/

 

Has their freebie CAD software.

 

I�m thinking injection molded Liquid Metal ( www.lqmt.com ) for the main material (Doug G. please weigh in here!) Is injection molding cheaper then casting or machining?

 

 

 

How many components can be bought off the shelf? If we go with an atypical format� probably not many.

 

Thanks again for all your posts.

 

JOE

[doug grosjean]

<<< We would need CAD software. >>>

 

There's a wide range of options there. I don't know what's available, low-end, that can do 3D modeling.

 

Stuff I use at work is about $20k per seat (ProE). And getting a qualified user costs, too.

 

 

<<< We would need to be able to test the camera?s mechanicals inside a computer 10,000 times before machining a single part. I know there are software?s that can do this?. >>>

 

Not neccesarily correct.

 

FEA (finite element analysis) can analyze strains on a single part, if you can define the stress / strain pretty accuratetly ahead of time. But it can't replicate actual use.

 

Much of actual product testing is still done via accelerated testing of prototype and production units. For instance, a transmission, you might rig up an engine and a robotic arm to shift through the gears constantly (24/7) for 3 months, maybe one total rowing through the gears per minute, accumulating 129,600 shifts in that time.

 

For a camera, you'd probably rig something to cock and fire the shutter repeatedly at 70F, for some number of cycles (and I don't know what the standard is in cameras, but I suspect there is one - and current makers may use a tougher standard of their own). Then same thing in a cold room at -30F, and then the same in HOT conditions like maybe 130F.

 

 

 

<<< But are they cheap enough to spread copies around to 1000+ volunteer developers. >>>

 

That would be a trick. Even at $2 / per body, somebody has to front the costs of $2k for the bodies.

 

And my belief based on experience is that molding those bodies would be nowhere near $2. You'd have to design it, then submit that to the Emachine people for a quote, and go from there.

 

IIRC, the Emachine people are not cheap.

 

 

<<< I?m thinking injection molded....>>>>

 

Who pays for the mold? I'd figure that could run tens of thousands easily.

 

At best, your 1,000 volunteers are suddenly needing to spend $5 here and $10 there for mold design / mold cutting. Plus the price of the part producted.

 

 

<<< Liquid Metal ( www.lqmt.com ) for the main material (Doug G. please weigh in here!) >>>

 

This isn't my field, so I can't reply directly. But Argus C3 Bricks were basically a hunk of plastic. A person would do well to study one. My gut feel is that it would be rather high-$, because most plastics come from oil.

 

 

<<< Is injection molding cheaper then casting or machining? >>>>

 

Depends on how many you make.

 

If you make 10 of something, machining will win, because just making the mold(s) will take more time than machining the part from scratch.

 

If 100, probably still machining would win.

 

1,000? I don't know.

 

10,000? Now maybe a mold / die is cheaper than machining. Depends on the complexity, and is strictly case-by-case.

 

You'd probably want to buy as many parts as possible off the shelf, and simply produce the best light-tight box (maybe with bolt-on lens mounts) that you could, with the modular concept others have suggested.

[matt_young5]

I work on the Open Camera project for industrial digital video, similiar to the PC architecture. So far the vendors I work with have agreed to:

 

A data bus, PCI Express

 

A focal plane sloat, slot one. This PCI slot brings the optical path out through the front plate and connects with standard optical connect.

 

Web Based Enterprise Management: essentially the camera is a thin web server with interpetable C.

 

Open archit3ecture streams based on MPEG7 ideas.