Ad
Ad
Ad
Pages: « 1 2 [3] 4 5 ... 8 »   Bottom of Page
Print
Author Topic: Open Source CMOS Medium Format Camera/Back - An initial exploration  (Read 11677 times)
Ken R
Sr. Member
****
Offline Offline

Posts: 460


WWW
« Reply #40 on: February 26, 2014, 11:00:06 AM »
ReplyReply

My opinion on the subject:

#1 The new camera / back MUST have Live View and a very nice screen on board. It should be a self-contained unit. No cables needed for power or anything but still have wifi for external control if desired and USB3 for tethering. (If its a hassle to use it people will just grab a A7R or D800E)
#2 The new camera / back MUST work with a wide range of lenses. Yes, from the excellent Canon TS-E II lenses to Rodenstock HR-W's. That should limit sensor size to about 54mmx40mm approx. but the does not rule out a smaller sensor like the one in the IQ250 since the Canon TS-E lenses are plenty wide enough to satisfy the wide angle needs of just about any user. This is HUGE, the A7R is already tapping into an extremely large number of Canon lens owners and doing well. It is a BIG market.
#3 The new camera / back MUST cost less than $10,000
#4 The new camera / back MUST have a sensor with at least 50MP and perform well in low light / long exposures and reasonably high iso. (anything less and people will just grab a A7R or D800E like they do now)

So, I am thinking that the new product should be a digital back like unit with a shutter unit in front (as short as possible) and with interchangeable lens mounts or adapters so that almost any lens can be used whether they have an integrated shutter or not. An optional external viewfinder would also be awesome (no need to design a new one just make the ones already available work) with viewing modes that aid manual focusing when off the tripod.

Think of it as a much more refined, compact and integrated (and affordable) FPS unit or Hartblei B1 camera that can be used handheld also if desired.
Logged
torger
Sr. Member
****
Offline Offline

Posts: 1503


« Reply #41 on: February 26, 2014, 11:32:10 AM »
ReplyReply

My opinion on the subject:

The sum of #1 to #4 seems like an IQ260 with CMOS. I don't think we could compete on sole rational image quality and feature set, ie make better products than Phase One and price them 1/4 of the price. Making a fantastic all-around consumer product I don't feel likely. The image quality of the 135 systems won't stop with A7r or D800. 50 megapixel sensors are coming and the Otus lens line is expanding.

I think it's more reasonable to think that a back made in a project like this would look more like backs did 15-20 years ago. Like the 60x60mm tether-only Dicomed Bigshot, those were made with standard components by very small companies. The products look ugly home-brew style, had limited feature set but they could produce images with larger sensor area than anyone else. You'd have to focus on a niche within the niche.

I see two opportunities. 1) make a large sensor back for 6x6 and 6x7 cameras. Phase One and Hassy won't go there. You're in your own niche, and with the Hy6 you don't only need to rely on legacy cameras. 2) make a back aimed at budget tech cam landscape photography, to attract amateurs of similar mindset as those that play with 4x5" film today. It's not only about having the highest resolution and widest DR, some like the "old-school" feel of these cameras and just need a decent back they can afford.

Somewhat going back to the time a digital back was nothing more than a drop-in replacement for a film magazine. If it would lack display, letting it look like a film magazine would be quite cool actually. I think no display could work if you still have onboard power and storage and could get decent integration with a phone/tablet.

The big gain with no display is however to go tethered all the way have no storage or raw format processing and draw power from cable, just send the raw signal from the sensor through the cable and solve the rest in software on the computer.
« Last Edit: February 26, 2014, 11:40:21 AM by torger » Logged
Kumar
Sr. Member
****
Offline Offline

Posts: 655


WWW
« Reply #42 on: February 26, 2014, 11:55:00 AM »
ReplyReply

Is putting two (or more) Sony A7 sensors side-by-side and stitching the resulting images feasible? There may be small differences in image quality between the two sensors, and other problems, but is it doable at all?

Kumar
Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 556


« Reply #43 on: February 26, 2014, 12:11:56 PM »
ReplyReply

Is putting two (or more) Sony A7 sensors side-by-side and stitching the resulting images feasible? There may be small differences in image quality between the two sensors, and other problems, but is it doable at all?

This has been tried by Minolta on the RD-3000 and RD-175 cameras, which used 2 and 3 CCDs at the time to increase resolution. In a nutshell: it was not a good idea. The optics needed in front of the CCD limited the system considerably.
Logged
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2509


There is no rule! No - wait ...


« Reply #44 on: February 26, 2014, 01:13:28 PM »
ReplyReply

I'd like to have a digital back working on a view camera with:
- large sensor area (~60*60 mm or more)
- 24-36 MP would be enough, more wouldn't hurt - with a 60 mm sensor and 36 MP this
   would be 10µ pixels - lots of room for great full well capacity and awesome low ISO ...
- Long exposures up to many many minutes (60?)
- great color filtering and awesome profiles
- great angular response without need of any LCC stuff.
- Also working with old MF systems, like Hassy V, Contax, Mamiya RZ ...
- great screen with live view and focus peeking at any place of the image / live view interface to a tablet (large screen !!!)
- True RAW histogram
- Interface to view camera parameters, like speed, f-stop, lens, shift and tilt data ...
   (maybe just an interface to enter this data manually in a quick and easy way)

Cheers
~Chris
Logged

eronald
Sr. Member
****
Offline Offline

Posts: 4035



« Reply #45 on: February 26, 2014, 02:52:20 PM »
ReplyReply


The big gain with no display is however to go tethered all the way have no storage or raw format processing and draw power from cable, just send the raw signal from the sensor through the cable and solve the rest in software on the computer.

Yes, this is why I believe the project would be feasible. Digital signals can be transmitted across a cable.

The sensor would be in a metal cartridge with a heatsink on one side and a cover/IR filter on the other.
Suitable external clips on the sensor cartridge would allow the adaptation of the desired camera mount plate and there would be  easily accessible vernier adjustment of the sensor plane.

A local connector would mate the sensor cartridge with a local electronics pod, and then a tough external connector would mount outside whatever back housing is used to tether to the computer. IO wires from the local pod could go to the camera's sync port and other camera connectors.

With such a modular concept, one can design and prototype pieces fairly independently. An updated electronics pod might later contain a full processor and a wifi interface. Users could do field upgrades.

Edmund
« Last Edit: February 26, 2014, 03:21:10 PM by eronald » Logged
MichaelEzra
Sr. Member
****
Offline Offline

Posts: 656



WWW
« Reply #46 on: February 26, 2014, 03:38:34 PM »
ReplyReply

It would be most flexible/customizable to standardize on a smartphone-based control/interface, similar in concept to Sony's QX10 and QX100.

We might also want RawTherapee compatibility:)

The housing can be designed to be elegant and 3D printed!
« Last Edit: February 26, 2014, 03:56:39 PM by MichaelEzra » Logged

Joe Towner
Full Member
***
Offline Offline

Posts: 107


« Reply #47 on: February 26, 2014, 04:49:37 PM »
ReplyReply

I had a long list of things to consider but timed-out and didn't save it prior to posting...

- Working within current shipping chips, it figures safe to assume it's either Dalsa or Sony.  I would want to see what pricing is like for engineering/dev board samples of both would cost.  Toss in if there is an ability to dev against a 35mm cmos from Sony and have that all move over to the 44x33 sensor (cheaper initial testing).

- Who's the target user?  It can't be everything to everyone, so what is the ideal?  For form factor, something Alpa-ish would be nice and simple, where we can do 'lens boards' to offer up larger compatibility.

- If we dump the screen and it's associated electronics, making it a tethered specific creation, would it drop the costs significantly?

- Can we do a v1 that's with 35mm chips cheap enough to prove it to do a larger product?
Logged
eronald
Sr. Member
****
Offline Offline

Posts: 4035



« Reply #48 on: February 26, 2014, 06:13:31 PM »
ReplyReply

Joe,


I believe when firms sell you and SDK they will sell you "engineering grade" chips which have defects but are cheap and useful for testing. I don't think the electronics of the SDK is really different in price depending on the size of the sensor within the same family.

The electronic design toolkit specific to this project of this is not costly IMHO - less than an MF lens cap Smiley what is a nuisance to start up  is the mechanical design necessary to make the "sensor cartridge" and get the thermal issues and cover glass right. Although it is quite possible that one of the industrial camera companies already has a housing that would be suitable, if not I am sure they could make one up for a reasonable price.

The motivational side here is that anyone with an SDK gets images from day 1, provided he can project an image on the sensor, and from the day we have a mechanical mount, anyone with the camera for that mount can start doing studio files with a hacked version of something like dcraw, although presumably the calibration will be off, the workflow painful, and the setup very iffy. The idea of using a digital CMOS chip allows us to move this design to a toughened version, while in the CCD world analog electronic noise is a really hard issue.

I expect the sensor cartridge would initially mount on a Hasselblad V mount, possibly even obtained from A Hasselblad film back or clone. This would then mount on a Hassy or an Alpa.

Of course Eric will be welcome to help supply Rollei plates Smiley


Edmund
« Last Edit: February 26, 2014, 07:05:51 PM by eronald » Logged
alan_b
Jr. Member
**
Offline Offline

Posts: 74


WWW
« Reply #49 on: February 26, 2014, 06:21:48 PM »
ReplyReply

Thinking as a potential customer, I start with use cases and product types.  It seems like some shortcuts could be found by modifying existing (forthcoming) products.

1. Small, inexpensive CMOS back (35mm size):  Modify an A7R for easier use on a view/tech camera (maybe not necessary with forthcoming adapter from Arca ––> M2)

2. Live View MFDB at less price than the Phase/HB offerings: Modify the new Pentax

3. Large-sensor (6x6 or larger) MFDB or camera insert: This seems like the ideal target for a ground-up design, also the most likely to be viable when the user has to accept compromises like tethered-only, no live view.

Edit: grammar & clarity
« Last Edit: February 26, 2014, 06:30:46 PM by alan_b » Logged
EricWHiss
Sr. Member
****
Offline Offline

Posts: 2412



WWW
« Reply #50 on: February 26, 2014, 07:09:52 PM »
ReplyReply

Would love to see 5 axis sensor image stabilization like in the Oly 4/3rds cameras in MFDB.  Woah! That would be amazing!
Logged

Authorized Rolleiflex Dealer:
Find product information, download user manuals, or purchase online - Rolleiflex USA
synn
Sr. Member
****
Offline Offline

Posts: 399



WWW
« Reply #51 on: February 26, 2014, 08:05:29 PM »
ReplyReply

I am no scientist, but considering the scientists have no qualms offering advise to artists here, maybe the reverse is also welcome! Smiley

Here are a few questions I have regarding this whole project:

1) It's not impossible to get a lab prototype. But is that the extent of this project? Are you looking at manufacturing, certification and everything? Who's gonna take charge of that?

2) No matter how cheap the test kits are (If it was actually as cheap as an MF lens cap, I'd wire USD 20 right now to Sony and get one. Why not?), there's still a significant investment involved. You're looking at 5 figures for sure (The sensor is only one part of the equation). There are enough complaints on this forum about spending that sorts of amount on MF vs spending a lot less on 35mm. What exactly is different about this that will make people donate significant amounts of money into a backyard project; sums they are reluctant to invest in functional, off the rack camera gear?

3) How exactly will the responsibilities be delegated? Members here reside in various geographic locations. Would all the hardware be at one person's residence and the rest provide their contributions via the internet? I can't imagine some fragile electronics being sent across the world a few times during the course of development.

4) Marketing question: Who exactly is the target audience for this product?

 - Is it only academia?
 - Is it a 35mm format shooter who can spend maybe 20% more than his current top of the line body?
 - Is it a current day MF shooter who will find it justifiable to discard his existing MF kit for this?
 - For any of the above, would the project team provide a level of support that is at least on par with the user's existing platforms to make this switch justifiable?

5) Continued from the above, here's a question from an MF shooter.
What will this project offer me, a Credo 40 shooter who mostly does portraits to switch over and buy one of these? Is the promise "Everything my current camera does, only better and at a cheaper price"?

Logged

my portfolio: www.sandeepmurali.com
eronald
Sr. Member
****
Offline Offline

Posts: 4035



« Reply #52 on: February 26, 2014, 09:13:33 PM »
ReplyReply

Dear Synn,

 In the US quite a few enthusiasts build their own planes from kits. This gives them access to a hobby and a tool which they could not otherwise afford. In the same way, I am proposing defining a DIY Medium format kit which can then be used by enthusiasts while tethered to a computer.

 Quite clearly such a kit is not for you: it won't take images quickly and reliably, will need to be rebooted occasionally, will not work with c1 or Lightroom, but will work with RPP, dcraw and friends, will not provide a composite Jpeg histogram, but will give simple channel by channel Raw histograms. It will not have Douge+Steve support or 24 hour replacement policies, and it will break if your beautiful models sneeze near it.

 This kit is a geeks' delight. It will be adaptable to almost anything as the sensor cartridge will be fairly robust and remountable, spare interface connections will be provided if possible so people can drive eg. a color filter wheel from the computer, or fire several flash units in sequence, and every setting the sensor has will be exposed. Users will be free to order and deploy monochrome or multispectral sensors if such are available.

 As a kit such an object will not have, nor does it need certification; however I do not intend it to be offered anywhere except the US, Europe and Japan, until I can request an opinion on export controls.  This version is strictly envisaged as being used by people who are comfortable with computers and lab equipment, and who are either hobbyists, or professional photographers who see a need for exceptionally great control and are familiar with purpose-built devices.

 I hope this answers your questions - clearly you are a busy commercial photographer, and better served by reliable and supported equipment such as is marketed by Phase One, Hasselblad and Sinar.

Edmund



I am no scientist, but considering the scientists have no qualms offering advise to artists here, maybe the reverse is also welcome! Smiley

Here are a few questions I have regarding this whole project:

1) It's not impossible to get a lab prototype. But is that the extent of this project? Are you looking at manufacturing, certification and everything? Who's gonna take charge of that?

2) No matter how cheap the test kits are (If it was actually as cheap as an MF lens cap, I'd wire USD 20 right now to Sony and get one. Why not?), there's still a significant investment involved. You're looking at 5 figures for sure (The sensor is only one part of the equation). There are enough complaints on this forum about spending that sorts of amount on MF vs spending a lot less on 35mm. What exactly is different about this that will make people donate significant amounts of money into a backyard project; sums they are reluctant to invest in functional, off the rack camera gear?

3) How exactly will the responsibilities be delegated? Members here reside in various geographic locations. Would all the hardware be at one person's residence and the rest provide their contributions via the internet? I can't imagine some fragile electronics being sent across the world a few times during the course of development.

4) Marketing question: Who exactly is the target audience for this product?

 - Is it only academia?
 - Is it a 35mm format shooter who can spend maybe 20% more than his current top of the line body?
 - Is it a current day MF shooter who will find it justifiable to discard his existing MF kit for this?
 - For any of the above, would the project team provide a level of support that is at least on par with the user's existing platforms to make this switch justifiable?

5) Continued from the above, here's a question from an MF shooter.
What will this project offer me, a Credo 40 shooter who mostly does portraits to switch over and buy one of these? Is the promise "Everything my current camera does, only better and at a cheaper price"?


« Last Edit: February 26, 2014, 09:38:38 PM by eronald » Logged
synn
Sr. Member
****
Offline Offline

Posts: 399



WWW
« Reply #53 on: February 26, 2014, 09:18:03 PM »
ReplyReply

Dear Synn,

 In the US quite a few enthusiasts build their own planes from kits. This gives them access to a hobby and a tool which they could not otherwise afford. In the same way, I am proposing defining a DIY Medium format kit which can then be used by enthusiasts while tethered to a computer.

 Quite clearly such a kit is not for you: it won't take images quickly and reliably, will need to be rebooted occasionally, will not work with c1 or Lightroom, but will work with RPP, dcraw and friends, will not provide a Jpeg histogram, but will give Raw histograms. It will not have Douge+Steve support or 24 hour replacement policies, and it will break if your beautiful models sneeze near it.

 This kit is a geeks' delight. It will be adaptable to almost anything as the sensor cartridge will be fairly robust and remountable, spare interface connections will be provided if possible so people can drive eg. a color filter wheel from the computer, or fire several flash units in sequence, and every setting the sensor has will be exposed. Users will be free to order and deploy monochrome or multispectral sensors if such are available.

 As a kit such an object does not have, nor does it need certification. This version is strictly envisaged as being used by people who are comfortable with computers and lab equipment, and who are either hobbyists, or professional photographers who see a need for exceptionally great control and are familiar with purpose-built devices.

 I hope this answers your questions - clearly you are a busy commercial photographer, and better served by reliable and supported equipment such as is marketed by Phase One, Hasselblad and Sinar.

Edmund




Thanks for clarifying that this is a hobby kit and nothing more. Your earlier response to Doug etc. led me to believe that there are commercial aspirations for this project.

Doug,

 I would expect that if something like this takes off there would be some specialist vendors like you who would qualify the hardware and distribute it. 

Edmund


My bad, clearly.
Logged

my portfolio: www.sandeepmurali.com
eronald
Sr. Member
****
Offline Offline

Posts: 4035



« Reply #54 on: February 26, 2014, 09:25:52 PM »
ReplyReply

Thanks for clarifying that this is a hobby kit and nothing more. Your earlier response to Doug etc. led me to believe that there are commercial aspirations for this project.

My bad, clearly.
Dear Synn,

Doug or Steve will be able to distribute and even originate kit parts if they wish to do so - some of their museum or academic customers may need a community-supported  open source kit camera for very special applications. You will doubtless recall that Masters and Johnson designed a special purpose camera for their research. I hope I have sufficiently satisfied your curiosity. Smiley

Edmund
« Last Edit: February 26, 2014, 09:42:07 PM by eronald » Logged
synn
Sr. Member
****
Offline Offline

Posts: 399



WWW
« Reply #55 on: February 26, 2014, 09:40:56 PM »
ReplyReply

Doug or Steve will be able to distribute kit parts if they wish to do so - some of their museum or academic customers may need a community-supported  kit camera for very special applications. You will doubtless recall that Masters and Johnson designed a special purpose camera for their research. I hope I have sufficiently satisfied your curiosity. Smiley

Edmund

Hopefully, this would be the last edit to your post and I would not have to re-think my reply after I post it.  Cheesy

Yes indeed, my curiosity has been satisfied. As you've very clearly explained, there's nothing of particular interest to a working photographer in this project, so all I can contribute is my sincere well wishes for you and your team. I hope you'll come up with something that will teach some guys in Denmark, Sweden and Japan a thing or two about medium format imaging!

Logged

my portfolio: www.sandeepmurali.com
eronald
Sr. Member
****
Offline Offline

Posts: 4035



« Reply #56 on: February 26, 2014, 09:51:11 PM »
ReplyReply

I think most people would be very surprised to hear of any modern photographic technology being invented or perfected in France, although the people here have a great respect for art.

This project basically consists of a wire between a sensor and a computer, so it really isn't rocket science.

The big CCD backs employ very sophisticated analog engineering and signal processing to keep noise under control, and hardening such designs in an electronically noisy environment is difficult, an expertise which the designers of remote sensing devices such as Kodak and Dalsa have historically possessed but have not always been allowed to bring to the open market.

The arrival of digital-output CMOS technology appears to be disruptive; Sony's decision to resell their sensors is admirable.

Edmund

Hopefully, this would be the last edit to your post and I would not have to re-think my reply after I post it.  Cheesy

Yes indeed, my curiosity has been satisfied. As you've very clearly explained, there's nothing of particular interest to a working photographer in this project, so all I can contribute is my sincere well wishes for you and your team. I hope you'll come up with something that will teach some guys in Denmark, Sweden and Japan a thing or two about medium format imaging!


« Last Edit: February 26, 2014, 10:44:04 PM by eronald » Logged
torger
Sr. Member
****
Offline Offline

Posts: 1503


« Reply #57 on: February 27, 2014, 12:57:34 AM »
ReplyReply

Yes, this is why I believe the project would be feasible. Digital signals can be transmitted across a cable.

The sensor would be in a metal cartridge with a heatsink on one side and a cover/IR filter on the other.

The digital back would look quite similar to Sinar 54m I guess. Sinar did have stand-alone backs for a while but now today go with the tethered-only concept again, I'm guessing that's because their customer base shoot tethered anyway, and tethered-only backs have the advantage of minimizing hardware design, and making software design easier.

This type of project could prove to be a good alternative to a Sinar back actually, it fits about the same niche. Multishot will be lacking though. If you'd want that the hardware design gets more complicated and I guess you'd need a patent license.

« Last Edit: February 27, 2014, 01:02:54 AM by torger » Logged
Kumar
Sr. Member
****
Offline Offline

Posts: 655


WWW
« Reply #58 on: February 27, 2014, 03:19:46 AM »
ReplyReply

Sinar had the CyberKit, a portable dedicated computer with a screen, much like today's tablets. It could save upto 1600 images, and could be used for review, focus-check, etc. In short, it was a specialized computer for Sinarbacks. Megavision also had an OQO computer that went along with their 22mp backs. Richard Chang showed it to me in 2006, and I thought it was a great concept. Unfortunately, at that time, everyone wanted self-contained backs, and these ideas didn't last long. Perhaps it's time to revive them?

Kumar
Logged

haplo602
Newbie
*
Offline Offline

Posts: 22


« Reply #59 on: February 27, 2014, 03:36:23 AM »
ReplyReply

Edmund, have a look at Dalsa web and the pictures of their dev kit. I think it partly covers the modularity you proposed.

Anyway one can also start with a CCD just to prototype some parts of the design (i.e. bus/interface between chip module and processing part).

I do agree that CMOS ships are much easier to work with as they put out digital data already, not need for A/Ds and complicated timing schemes. The only problem is that cmos chips readily available are very small (even 35mm chips are a problem to get).
Logged
Pages: « 1 2 [3] 4 5 ... 8 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad