111-Megapixel CCD Chip Ships

georgewilliamherbert writes "EETimes is reporting that Dalsa has shipped a record-breaking 111-megapixel CCD image sensor to customer Semiconductor Technology Associates. The chip was paid for by a U.S. Navy SBIR project. At four inches across, a bit big for camera phones, but the 10560x10560 format will probably get professional digital camera users drooling."

That's a big sensor.

[Anonymous Coward]

Here's a pic of the sensor itself: http://www.dalsa.com/shared/content/images/STA1600 _1_1200w.jpg [dalsa.com]. (Too bad there aren't any pics from the sensor...)

my calculations...

[laxcat]

Someone correct me if I'm wrong but that's just shy of a 3 by 3 foot image at 300 dpi !

Re:Film

[pla]

Well sure sounds like that'll BLOW AWAY 35mm film and definitely be about comprable to 4x5 film.

ISO100 film has a grain size of approximately 5 microns, which corresponds to a resolution of 36MP. Standard 4k scanning (12.5MP) captures all the detail in anthing short of the pro-est of the pro, and 8k scanning (54MP) all but guarantees that even future advances in scanner technology won't have the ability to extract any further detail from a 35mm negative.

You would need godlike optics, bright light, and a perfectly still subject and camera to come anywhere near that 36MP with ISO100 35mm film, but it represents a sort of upper limit at that speed. 4x5in film therefore has an effective resolution (at something comparable to ISO100) of 500MP.


So, this can effectively replace 35mm film in terms of resolution. It falls a bit short of replacing truly professional-quality film, however. But then, how often do you need to print out your personal pics at literally bilboard size?

Re:Consumer version already available, kinda

[jpatters]

4x5 film doesn't come in rolls, it comes in sheets that you load into a holder, one to a side. You have to load the film in complete darkness, and hope that the holders won't leak. When taking the picture, you focus with a groundglass that is situated where the film will be, then close the lens, insert the holder into the camera, and pull out the dark-slide, and then take your exposure, and you should be taking lots of notes. Because there is so much manual labor that you have to do for each exposure, there is a whole different mindset to Large Format Photography, you will go out and expect to take a half dozen exposures, while the digital camera encourages the practice of just shooting anything and everything, and then sifting through the thousand or so exposures for the good ones.

The owner of a camera shop near where I live once had the opportunity to use a Large Format Polaroid camera [polaroid.com], which exposes Polaroid fim that is 20 by 24 inches. He described it this way: "Take your megapixels and shove them up your ass!"

Re:Film

[Anonymous Coward]

wrong, big time. I've seen this camera in action. it kicks ass.

The biggest problem they're going to have in digital cinema is the fact that the beautiful babe actress
actually has a more than noticable moustache. She's going to need a lot of time to be ready for her closeup, Mr DeMille.

http://www.dalsa.com/dc/index.asp [dalsa.com]

Re:Consumer version already available, kinda

[wolenczak]

The concern on X-Ray radiation is not the radiation the patient gets, but the one the radiologist is exposed to on a daily basis

Re:Film

[moosesocks]

no. the size of the sensor has nothing to do with noise. the grandparent poster was correct -- it's the DENSITY of the sensor that affects the amount noise you get.

DSLRs have the advantage, not because their sensors are necessarily larger, but because the pixels aren't packed so tightly together. You could hypothetically use the same processes they use to make those tiny 8MP compact-camera CCDs to make an APS-C sized CCD for a DSLR. You'd have tons of (hypothetical) resolution, but the noise would make it useless, and it'd be painfully expensive.

Film still wins in this arena. We're only approaching the point where huge large-format sensors like this one can challenge high-quality 35mm film. We need to get to the point where we can match 35mm resolution in a 35mm (or more likely, APS-C) sensor.

We're pretty close to conquering the noise issue, and even compact cameras are beginning to perform well at ISO 800 and 1600, whilst full-frame 35mm DSLRs can produce virtually noiseless prints at ISO 1600 and 3200. Improving dynamic range and pixel-density will be the next big technological hurdle to leap over.

Personally, the idea of a high dynamic range CCD excites me. Imagine the possibilities.....

Re:Film

[ryusen]

given the same number of pixels, larger sensor = lower density. if you used the same process to make a sensor liek you described, you just get more pixels, rather than better pixels. i guess i should not have assumed i mean for the same number of pixels a bigger sensor will produce a cleaner image, but i thought it was aobvious. also, if you really want to get technical, it's not even the density, but the size of the individual pixels that makes the difference.

I do think it's neat to have this kind of technology, but i would be much more excited about a CCD or CMOS sensor that can capture 11 stops of light or some other type of technology that would get around the problem of high contrast scenes, than more megapixels. i just feel the mega pixel race has gone the way of the MHz race, in CPUs. sure it's bigger, badder, better, but after a certain point it won't make a noticeable difference on the end result, for most people.

Re:Film

[dfghjk]

While the Canon lenses you referred to may be some of the finest ever produced for 35mm format, they aren't the last word in what's possible when larger sensors are employed. Canon's wides are inferior to Nikon's as well so pushing the limit there isn't really saying anything. I owned a 1DS2 until recently BTW.

Otherwise, you're completely correct about the dynamic range of modern digital SLR's. I can't stand hearing the same uninformed comments made about the superiority of film over digital. 35mm film was surpassed quite a while ago.

Re:Consumer version already available, kinda

[mkuki]

I totally agree with this.I have both an Nikon d70 and a Shen-Hao HZX (4x5 inch camera) and you wouldn't be able to pry the 4x5 from my hands.I rarely use it compared to the D70, but when I do, the images are astounding (you can blow up one image to about 4x5 feet at 300dpi).

I think that digital and film both have their place and it is pointless to try and replace one with the other.

One other thing, the whole loading the darkslides manually can be done using a loading tent or even by using a Fuji Quickloader and forgoing the whole headache (though if you want a film that does not come in the Fuji Quickloader packs you are back to using film holders.

Oh yeah, another thing, there are a million ways to screw up with large format until you get it down.

Re:Consumer version already available, kinda

[jpatters]

In 10 years cheaper digital cameras will exceed the quality of large format photography.

I would dispute that assumption. Due to the limitations of lenses that exist in the physical world, you cannot simply make the pixels on a CCD arbitrarily small so that you can have more of them. Even if you could conquer the noise problems that go along with the small pixels in the consumer grade 6MP and 8MP sensors, which are much smaller than the sensors that you find in the more expensive DSLRs, you would run into the limit of the lens before you had enough pixels to rival 4x5 LF film, let alone 8x10. The only way to make a digital sensor match the resolution of 4x5 LF film is to make the sensor nearly as big as the 4x5 sheet of film, which is what they have done here. The problem is, it is wicked expensive. I mean, they made one of these things, it probably cost millions of dollars all told, and even if they were to mass produce them, it would cost hundreds of thousands of dollars per part. (Just a guess)

People are used to tech products like this getting vastly better and vastly cheaper over time, and the reason this happens is, to over simplify, that we develop the ability to manufacture them with smaller and fewer chips. The fact is, CCDs have already reached the point that the resolution is more bound by the lens then by the number of pixels on the sensor. They really aren't going to get much higher resolution, at least, not meaningfully. One thing I do think will continue to improve is sensitivity, I would definitely like to have a digital camera capable of decent noise performance at ISO 12800, that would be fantastic. (And that would be something that would revolutionize available-light photography.)

Large Format film will be resolution king for the foreseeable future. You simply cannot project an image with as much resolution as you can get from a scan of an 8x10 inch sheet of film onto something that is 18x24 millimeters, within the physical constraints of this universe. With 18x24 millimeters being the practical limit for how big you can make a sensor that will go in an affordable, consumer grade digital camera, I simply don't see your prediction coming true.

Re:Film

[Aqua OS X]

Yup, viewing distance is important.

If you're doing something for a small print piece you want a high DPI (ie 300). If it's a poster you can use a lower DPI. If it's a Billboard you can use a significantly lower DPI.

I'm a graphic designer and I recently committed the industry's cardinal sin the other day... I had a comp printed at Kinkos. I was printing a fairly large bus shelter poster that was 150dpi. The newb behind the counter had the audacity to bitch about DPI, even though (I would imagine) it was fairly obvious that I did this for a living.

If you're developing something large in Photoshop you do -not- want to play around in 300 DPI. People read those things from a few feet away and, I don't care if you have a new dual-core dual g5, you do -not- want to wait for a 30x40in bitmap to rotate on a multilayered 300 DPI document.

That said, high res photography is important. You may only want to highlight a small piece from a large image, and you can't do that unless you have good source material.

Re:Film

[Cowclops]

Oh yes, I was aware of that after I did the math, though it is interesting that 1080p is really "plenty" for video. I realize that while I calculated the detail we can discern to be about 1.3 arcminutes, I think the actual number is more like .3 (as in about a third of an arcminute, not a little more than one)

Re:Film

[darthdavid]

Yes I can believe it wasn't designed by someone who knew anything at all about optics. We have a blind spot wich the brain fills in. This is because part of optic nerve is in front of the retina. This would be like putting the wiring for a ccd between it and the lens in a camera. As such, any 'intelligent' designer is a total junkslut who had to ride the short bus to diety school or we evloved, where some ancestor of ours had the miswired eye and kept passing it on because it was "good enough" which is the driving pricipal behind evolution. Hell, you can even see it in squids/octopuses, they need to see in dim light and so the mis-wired eye wasn't "good enough" so they evolved an eye that was properly wired and could see better in the conditions they faced. If there was a designer this makes then they are obviously a posterchild for post-natal abortion because it'd be about the stupidest thing I've ever heard of to design a proper eye for cephalopods and then, remember by christian mythology (you're not fooling anyone with the abstract "designer" you know) animals were made first, go and design a flawed eye for people.

Re:Film

[arminw]

.....Yes I can believe it wasn't designed by someone who knew anything at all about optics. We have a blind spot wich the brain fills in. This is because part of optic nerve is in front of the retina......

Repeating someone's ignorant statements isn't exactly the most intelligent thing to do.

Someone who does know about eye design is the ophthalmologist Dr George Marshall, who said:
'The idea that the eye is wired backward comes from a lack of knowledge of eye function and anatomy.'
He explained that the nerves could not go behind the eye, because that space is reserved for the choroid, which provides the rich blood supply needed for the very metabolically active retinal pigment epithelium (RPE). This is necessary to regenerate the photoreceptors, and to absorb excess heat. So it is necessary for the nerves to go in front instead. This doesn't spoil vision, because the nerves are virtually transparent because of their small size and also having about the same refractive index as the surrounding vitreous humour. In fact, what limits the eye's resolution is the diffraction of light waves at the pupil (proportional to the wavelength and inversely proportional to the pupil's size), so alleged improvements of the retina would make no difference.

Any explanation of how the form and function of a structure, such as the eye, came to be, without the application of an intelligence, takes far more FAITH than I can muster. I can see that your faith in blind chance, or whatever other mechanism, is greater than mine in God. I congratulate you in your great faith. I pray that someday you will see fit to redirect this faith to the one true God who loves you.