I was catching up on some of my photonics journals last week and they reminded me that I still haven't answered for myself the question of just how fast digital cameras can get. What is the ultimate limiting ISO?
You know there has to be one. Simple logic. I mean, when you reach the point where the shutter is open just long enough to let one photon get through, you can be pretty confident that you're not going to be able to form an image.
Unfortunately that simple logic doesn't really tell you very much. There are a zillion comments online to the effect that the "laws of physics" limit how fast the camera can get, but it's pretty clear to me that the people saying that don't have any idea what that limit is. Statements can be both true and irrelevant. Much like pointing out that the laws of physics limit how fast a rocket can travel. No argument there, but rocket designers aren't bumping their heads on that speed-of-light brick wall just yet.
Where is ISO's "speed of light" limit? I don't know, but I can tell you of some improvements that push sensor speeds way beyond anything you've imagined. All are either proven technologies or such simple extensions of existing technology that we can be entirely confident they can be implemented. That doesn't mean they will be, but they do constitute a proof of existence.
The current speed champ, the Nikon D3S, provides very decent quality up to ISO 6400, and is usable up to 102K. Not that that would ever be my choice, but if it's the only way I can photograph The Second Coming or First Contact, it'll do. How do we make it even faster with no loss in image quality?
First, install a back-side-illuminated sensor (unless I misread the specs, the D3s isn't using one). BSI sensors are about twice as efficient as conventional ones at gathering light. Large pixels don't benefit as much as small ones. For the D3S I'd be conservative and call it a 2/3 stop gain (for micron sized pixels, it's distinctly more than one stop).
BSI sensors haven't made big inroads yet, although they're in a few cameras. I suspect a combination of cost and the fact that smaller-sensor cameras need them more desperately.
Next, Kodak has this nifty modified Bayer filter array that gains you another stop of efficiency. That's three-year-old technology; not certain why we haven't seen it in a camera yet. Maybe I'm not looking in the right places, maybe Kodak's being a pain about licensing, dunno. Whatever, it's not a physical issue, so far as I know.
The Kodak array isn't the theoretically best one. It's probably possible to squeeze another half to full stop of efficiency out before you've reached the limit. I'll be conservative and say half a stop more.
The new hot area of development is sensitizers for silicon. Silicon detectors, at their very best, run about 50% efficiency. In addition, they are much more sensitive to long wavelengths of light than short ones. That's particularly bad because most often you're using high ISOs under indoor lighting that is generous in the long wavelengths and weak in the short. That's why the red channel in your high ISO indoor photos looks pretty clean, the green channel looks mediocre, and the blue channel looks horrible.
Anyway, an assortment of companies are developing layers and coatings that can be applied to the surface of the sensor that catch nearly all the photons and transfer their energy to the silicon. Some of these technologies are very close to rollout, if they're not already in production for the minuscule cameras that go into cell phones.
Combine BSI sensors, improved filters, and sensitizers and there's a 10X performance improvement. Let me emphasize that this is real improvement, not funny games being played with more amplification and noise reduction. What the Nikon D3S can do at ISO 6400, these technologies would let it deliver at 64,000. And when the flying saucers (or Jesus) descend onto your front lawn, you can crank that dial up to "11" for an unbelievable extended speed of 1,000,000.
Honest to God, I have no idea what you do with a speed of 1,000,000. Make handheld photographs by starlight? Just sayin'—it can be done, and we haven't exhausted the bag of technical tricks so don't assume that even this represents that performance brick wall.
Realistically, these technologies will appear in small-sensor cameras sooner, because they have more need for them. I don't think too many people will be unhappy when digicams offer the image quality of today's full-frame cameras. Trust me, they are going to get there.
Meanwhile, the skies still the limit, folks, and we still don't now how high is up.
TOP's weekly column by Ctein (pronounced "kuh-TINE," and yes, that's his entire legal name) appears on Wednesdays.
Note: Links in this post may be to our affiliates; sales through affiliate links may benefit this site. More...
Original contents copyright 2011 by Michael C. Johnston and/or the bylined author. All Rights Reserved.