Large Format Digital.

[Nikon 2]

Will there ever be a digital camera made for large format…?

 

[koraks]

There have been.

One angle are the scanning backs that were sort of, a little, popular at some point. They were basically a flatbed scanner mounted to an LF camera, conceptually.

The other one is the LargeSense development: http://largesense.com/ This one took several years to come to fruition and I'm not sure if it's actually a marketable product today. But it's the only 'real' digital LF camera that captures an entire frame all at once.

I don't think there will be much more along these lines, given the affordability, accessibility, image quality and flexibility of existing digital medium format solutions. They fill the niche that view cameras occupied in terms of product photography etc. insofar as that niche didn't actually dry up to begin with due to the advent of digitally-generated imagery etc.

 

[Helge]

There is of course all the experiments with projecting on a viewing screen and photographing that with a digital camera.
Surprisingly nice results, with some of the advantages of LF like DoF and perspective control, but of course none of the resolution advantage.

 

[Nikon 2]

Very surprised at the achievements in digital…!

 

[Dustin McAmera]

There are one-offs like this, being made for astronomy:

Rubin Observatory

NSF–DOE Vera C. Rubin Observatory will revolutionize the way we explore the cosmos

rubinobservatory.org

It will use 200 sixteen-MP CCD chips in an array to make a 3200 MP sensor.

 

[Nikon 2]

There are one-offs like this, being made for astronomy:

Rubin Observatory

NSF–DOE Vera C. Rubin Observatory will revolutionize the way we explore the cosmos

rubinobservatory.org

It will use 200 sixteen-MP CCD chips in an array to make a 3200 MP sensor.

Wow!
CCD sensor…!

 

[Dustin McAmera]

I think (that is, I think I recall reading...) they are still the thing to use for some technical uses; that recollection may itself be out of date. But also, they have been designing and building this thing for some years. I guess designers have to put blinkers on at some point and say 'no more improvements to this component'.

 

[Nikon 2]

There have been.

One angle are the scanning backs that were sort of, a little, popular at some point. They were basically a flatbed scanner mounted to an LF camera, conceptually.

The other one is the LargeSense development: http://largesense.com/ This one took several years to come to fruition and I'm not sure if it's actually a marketable product today. But it's the only 'real' digital LF camera that captures an entire frame all at once.

I don't think there will be much more along these lines, given the affordability, accessibility, image quality and flexibility of existing digital medium format solutions. They fill the niche that view cameras occupied in terms of product photography etc. insofar as that niche didn't actually dry up to begin with due to the advent of digitally-generated imagery etc.

I don’t see the logic of anyone spending over $100,000 for a LF digital system that falls way short of the resolution of a LF film camera…!

 

[Chan Tran]

The Sinar P3 has all the controls of a view camera but it's not large format. It's barely fit in the medium format as for size. The sensor is only 48x36mm.

 

[Chan Tran]

With the new mirrorless digital camera,( these have very shallow flange distance vs the DSLR) you can build or buy a shifting back to take multiple frames then stitch them. Use a regular view camera.

 

[Chan Tran]

But I doubt that they will ever make a digital imaging sensor in 4x5 size or larger.

 

[koraks]

I don’t see the logic of anyone spending over $100,000 for a LF digital system that falls way short of the resolution of a LF film camera…!

Resolution is a fickle concept.

Having said that, I think you've just voiced the very reason why we don't see widespread market adoption of this largesense camera. It fills a niche that basically doesn't exist.

Of course, there are other reasons why people moved to digital at some point. Convenience and time to publication were, and remain important factors. The first generation of digital cameras was pretty abysmal, quality-wise, and yet, we saw photojournalists in particular embracing the technology at an early stage, when absolute quality of film was still far supreme. Getting your images into the newspaper (and, increasingly, onto the web) was and remains an important reason. Of course, these people virtually never work with large format, and hence, large format digital is kind of a non-starter. It turns out to be pretty much a contradictio in terminis.

 

[Pieter12]

The one time I saw an advantage to a somewhat large format digital camera was 20 years ago for a food shoot. The fact that we were making a shot that incorporated pouring liquid and steam, it was very reassuring to be able to leave the shoot knowing it was successful, rather than counting on Polaroids—which may show a good pour, etc, but not guarantee the film shot would be identical, or waiting for a rush-processed sheet of film. I don’t recall what the equipment was, though, nor the size of the sensor in the back that was attached to the view camera.

 

[Chan Tran]

I don't think they will make large sensor because it's easier to make small sensor with high pixel count than to make large sensor. But then you would need much higher resolution lenses to do that.

 

[Helge]

But I doubt that they will ever make a digital imaging sensor in 4x5 size or larger.

It will have to be based on a different technology than CMOS.
Perhaps like a mix between a photo copier and a camera tube scanned out with a laser after exposure. Only way to really get some worthwhile resolution.

 

[koraks]

Perhaps like a mix between a photo copier and a camera tube scanned out with a laser after exposure.

That might be one way. But I think if anyone would seriously consider it today, they'd do a variant on LCD manufacturing. It has patterning and layering steps that would be suitable for fairly large pixels on a fairly large surface. The complexity of translating this to a CCD or CMOS application would evidently be higher, with several consecutive deposition and patterning steps to build a layer stack on a substrate. But most of the tooling, chemistry and process control could remain the same. BTW, CMOS on existing Si-based lithography would just as well be an option. Either in one go, or more practically by constructing an array of smaller sensors to make up a bigger one.

Either way, it would of course result in an extraordinarily expensive sensor given the R&D involved, and only in military or high-budget scientific contexts it might make sense to build a handful of such sensors. There's a couple of firms doing stuff like this, but it's the kind of technology that the general public never gets to hear about.

 

[Helge]

That might be one way. But I think if anyone would seriously consider it today, they'd do a variant on LCD manufacturing. It has patterning and layering steps that would be suitable for fairly large pixels on a fairly large surface. The complexity of translating this to a CCD or CMOS application would evidently be higher, with several consecutive deposition and patterning steps to build a layer stack on a substrate. But most of the tooling, chemistry and process control could remain the same. BTW, CMOS on existing Si-based lithography would just as well be an option. Either in one go, or more practically by constructing an array of smaller sensors to make up a bigger one.

Either way, it would of course result in an extraordinarily expensive sensor given the R&D involved, and only in military or high-budget scientific contexts it might make sense to build a handful of such sensors. There's a couple of firms doing stuff like this, but it's the kind of technology that the general public never gets to hear about.

I have no idea about the possible feature size of LCD fabbing. I’d guess it’s not anywhere near the level of CMOS sensor photo lithography through.
The limit of 4K for smaller commercial panels is probably not dictated solely by addressing and update speed limitations.
Some kind of thin film tech would probably be possible though. Question is as always development costs and manufacturability.

I know for a fact that very large CCD and probably CMOS sensors exists for government use. Astronomi and military cases are obvious.
If you can eat very low yields, anything is possible.
Throwing away dinner plate size wafers because of a tiny error gets extremely expensive real fast.

In the case of astronomy and large sensors quantum efficiency is more important than feature size, so there resolution on those large sensors is not maxed out.
So who knows how high a resolution it’s possible to eke out.

I was talking about realistic manufacture and good cost benefit ratio. That is resolution competitive with or just approaching film.
The existence of companies such as my home country’s Phase One shows that there is a market for high end digital still.

 

[koraks]

I’d guess it’s not anywhere near the level of CMOS sensor photo lithography through.

The advantage of an LF sensor is that the pixels can be pretty big. Bix pixels also means that on-die components are allowed to be fairly big. I'm quite sure that fairly large margins between pixels also won't hurt as much as on smaller sensors.
And since it's basically a photolithography process, it's feasible to do this at let's say the 90nm node or thereabouts, which is very well-established technology. In fact, Canon at some point didn't bother extending their photolithography developments in competition with ASML because their primary application was LCD manufacturing and a CD of a few hundred nm was adequate for that. Of course they're now hitting the news with their contact imprint lithography solution, which really isn't new (it was around in the 1970s AFAIK!)

I know for a fact that very large CCD and probably CMOS sensors exists for government use.

Yes, there certainly are. In fact, the present range of MF camera sensors stem from these developments. Once upon a time, those were considered 'large' as well.

I was talking about realistic manufacture and good cost benefit ratio.

Volume is tiny. There's no real money in it apart from the oddball military or science application. Story ends there.

 

[OAPOli]

Large format digital sensors (8x10" or larger) are common in medical imaging. Monochrome, with pixels 50 to 300 microns. It's likely they could be repurposed for photography?

 

[Nikon 2]

Is there a reason those LF sensors favor CCD…?

 

[Helge]

Is there a reason those LF sensors favor CCD…?

My guess would be it’s pretty simple, scalable and well known process, bucket brigade and readout on the side.
The timing and capacitance of CMOS on a large piece of wafer could be a problem.
CCD comes with only few downsides. The main reason for CMOS sensors is economy. Only later did slightly better amplification come into the picture (pic).

 

[Nikon 2]

I don't think they will make large sensor because it's easier to make small sensor with high pixel count than to make large sensor. But then you would need much higher resolution lenses to do that.

And very expensive ones also…!