Written by Ctein, TOP Technical Editor
Happy holidays to one and all! As long-time readers know I'm inclined to write an X-mas column with the emphasis on the “X.” That is, the essential but unknown part of the equation. The stuff we don't know is as interesting as the stuff we do.
(I just like this photo. A whole lot.)
This year, I'm going in a slightly different direction. The answers are fundamentally unknown, because they are speculations about the future. I can make well-educated guesses. My crystal ball is, though, unreliable.
The inspiration for this column is a thread I read not too long back on another site I frequent which was discussing whether the new 20-megapixel Micro 4/3 cameras were a waste of time. I'm not going to tell you which site or what thread, because within a page it had degenerated into the unfortunately-common ignorant blather that pollutes such discussions. I was mentally composing a rebuttal when I realized it would make a nice X-mas column. I will say one thing about that other thread: I think it should be against the law for people to invoke “laws of physics” unless they, in fact, hold degrees in physics! You can be pretty sure that anyone who says that, and doesn't, is going to be wrong.
All right, then. How far might image quality be pushed, without resorting to nonstandard technologies? Let's just stick with the usual Bayer array cameras for the moment and get to the more exotic later. I'll be talking about Micro 4/3 because that's what I currently use and care about. We could have this conversation about any format, and a lot of what I say will apply across formats.
One last thing before I get started: in the interests of holiday harmony, please suppress the impulse to write comments in the vein of “I don't need X; I don't see why anyone would need X.” Take my word for it that someone out there needs X, even if it's not you. In the interests of peace on Earth, goodwill towards bloggers, let's not go there, okay? Thank you!
Pixel count seems to be what everyone goes at first. How much can that get pushed before we really do run up against physical limits where we're sacrificing quality for quantity? My best estimate, at the moment, is that Micro 4/3 sensors can pick up another two stops in sensitivity before they start to approach real physical limits. Maybe three, but I'm not sure of that.
In practice, that means you could ultimately quadruple the pixel count of the OM-D cameras to 64 megapixels with no loss of sensitivity or increase in noise compared to today's 16-megapixel models (actually, the perceived noise would be a little better, because it would be finer). Another way to parse it would be that if you keep the pixel count where it is currently, you can expect to see half the noise at any given ISO, or a 4X speed increase for the same noise level.
Or you could split the difference, at 32 megapixels and twice the ISO, with modestly lower noise.
We've got a way to go in terms of pixel counts before we are forced into unpleasant tradeoffs. Understand, the engineering and technology has to advance to allow this, but there's nothing in the laws of physics that prevents that. It's just steady progress.
Well then, how about exposure range? Currently, we are running around 12 stops and change. That's pretty good! Exposure range isn't directly a function of pixel count, but there are some connections. Again, my best guess would be that along with upping the pixel count, you can probably squeeze another two stops of exposure range out of the conventional technologies. Call it 14 stops.
The bit depth? That's really a function of the readout electronics. We are nowhere near close to theoretical limits. It's what's cost-effective to make. You want true, clean 14-bit output? You'll get it. Sixteen bit? Hang in there.
Okay, so pretty damned impressive sensors are in our future. Can we really make use of them? Are lenses good enough? Oh yeah, really, they are! Too long to discuss here, but you can read this column and the links back from there and learn more about this subject than you ever really wanted to know.
I can tell you for a fact that the 45mm and 75mm Olympus lenses will hold up just fine, corner-to-corner, when a 32-megapixel camera comes along. I can't swear to it for a 64-megapixel camera; I can't reliably extrapolate my existing test data that far. It wouldn't surprise me. In any case, we are future-proofed for at least a factor of two improvement in resolution, and better lenses come down the pike every day.
Of course, sane people don't demand corner-to-corner pixel-level sharpness as a criterion for a useful camera system. I'm just saying that even if you do, I've got to proofs of existence in my kit, and I know there are others out there.
To infinity...and beyond!
What happens if we don't restrict ourselves to the current conventional camera technology? In that case, there are another two stops of sensitivity and efficiency waiting for you before we hit physical limits. Those Bayer arrays we use today are only about 25% efficient, at best. Those colored filters toss out out three-fourths of the light before it even gets to the sensor. At some point we're going to have commercially feasible detectors that are truly panchromatic: they'll detect photons of any color and measure their energy as they detect them. That stuff exists in the lab; it's not particularly useful for commercial sensors. (No, Foveon sensors are not the answer. They're inefficient, and the other technical problems that they have aren't going to be solved with the level of industrial investment and interest out there. They'll go on as a niche product but, basically, they're the Wankel rotary engines in the world of digital cameras.)
As for exposure range, the sky is the limit. There already exist logarithmic-response sensors in industrial cameras that have over a 20-stop exposure range. There's a whole bunch of reasons why they're not likely to appear in conventional cameras anytime soon, not the least of which is that you need a huge bit depth in the readout and conversion electronics to make them useful for pictorial photography. That's not a physics limit, that's just my opinion that I don't think it's the development path we'll see. But, we might!
Some clever folks at MIT have demonstrated different approach. They've built a system where, when a pixel gets saturated with light, it dumps its charge and starts filling up again. Meanwhile, a counter records how many times a pixel dumps its charge during the course of the single exposure. Hence, instead of each pixel being limited to a charge range of 0.0 to 0.9999, it can, in effect, catch a range from 0.0 to N.9999, where N can be a fairly large number. Three or four stops additional range is no problem.
Further down the pike, it's going to get a lot more interesting. The past year has seen the development of electronics that can operate at optical frequencies. Currently they are working in the infrared, but I have no doubt it will push down into the visible within a year or so. That's a monstrous game-changer; it's like going from crystal radio sets to high-frequency vacuum tubes that let you build REAL radio electronics. The same thing's going to happen in the optical part of the electromagnetic spectrum. It'll be unimaginable.
That's your really, really big X for the holidays.
Now, go do whatever it is you do on December 25 and have fun doing it! That is an order!
UPDATE 12:26 p.m., December 25, 2015: I just gotta share this! My Xmas gift from my Other Significant Other, Laura.
Paula and I both agree this wins the Gift Of The Day (and there were some strong contenders).
©2015 by Ctein, all rights reserved
Links in this post may be to our affiliates; sales through affiliate links may benefit this site.
(To see all the comments, click on the "Comments" link below.)
Featured Comments from:
John Camp: "One thing that Ctein somewhat brushed past is the issue of 'human use.' Possibly the most revolutionary and fast-developing device of our time is the computer—but in terms of mass/commercial/consumer computing, there really hasn't been much in the way of advancement in the past five years, because the very large majority of people don't need and won't use those advancements. In fact, we seem to be going somewhat backwards, with the widespread adaption of less capable machines, that happen to have the added capability of making phone calls and taking low-res photos. There has been some advancement in peripherals, like displays and linking cables (Thunderbolt) and flash drive speed, but those improvements really couldn't be classified as essential for most people. For 99% of users (the number pulled from between my buttocks), a 2010 screen provides about as much function as a 2015 screen. The potential for much faster products is there, but 99% of computer users don't need it, so they won't buy it unless the cost is low enough and they're replacing a computer that no longer serves their purposes.
"I think the same is true of cameras and printers—the high-end cameras now serve almost all purposes that cameras need to serve for 99% even of high-end users. Sure, there'll be 1% who want that OM 64-MP sensor...but it's the 99% who'll have to finance that research and development, and I have some reason to think that they won't be interested in doing that. In other words, the stuff we have now is good enough for 99% of photo purposes, and I think that 99% will increase to 99.1, 99.2 and so on because display is shifting more and more to digital, and useful digital displays don't require extreme resolution. Will we get the 64-MP OM? Maybe, but if so, I think it'll be a side effect of other advances, with the other advances (whatever they are) providing the financing for the spinoff."
Ctein replies: John, I think you're mostly correct, but...the effect of horsepower races and prestige and bragging rights on product development is not ignorable. I thought digital printer improvements would pretty much grind to a halt when they equaled (and surpassed) chromogenic prints by pretty much every meaningful measure. Seeing as chromogenics satisfied 99.999% (conservatively) of photographic needs. That happened a good decade ago. There's no real consumer-driven need for printers to have gotten better (folks like me and Charlie Cramer are an insignificant market share). That hasn't stopped steady improvements—it's a horsepower race. Similarly, the market for 120 format cameras and films collapsed before the turn of the century. Film and camera makers continued to churn out new medium-format product because it was considered the prestige market.
Wayne Fox: "I enjoyed the article.
"Regarding the idea of most people not needing it (I assume we are discussing buyers of photo gear which means they are after more than their very capable smart phone can deliver), I have a slightly different perspective. To me it’s about most captures not needing it—but then some do. As one who produces prints for photographers, at least 20–30 large prints or canvas wraps a month leave my store which suffer from serious quality issues because someone wants a bigger print than the detail in the image can pleasantly render. Many small prints also suffer because they have been cropped and not enough data is left to render the image at the desired size.
"So to me it's not 'if' we need need, it, but 'when' we need it. When I'm out shooting, most of the things I shoot could have been easy handled by a less capable device, but then a few times a year I get an image that looks best printed large (which happens to be what 'floats my boat' so to speak). I've had discussions with many customers who say they are completely happy with their current gear, because they never print larger than 13x19, then they show up and want a 40x60. We never know when we are going to capture that special image or need to do some serious cropping. Ironically this means our gear is good enough for most of what we shoot, but perhaps isn’t good enough for our best work."
"As far as the computer analogy, I agree for most needs were met by computers a few generations ago. But I’m not sure we are using less capable computing devices, it's more about a new class of device supplying what we need in a more convenient form factor. A current iPad Pro has more computing and graphics power than most computers of just a few years ago, so the device is still continuing to get more powerful, but the form factor has changed the game. To compare that to cameras, perhaps the 99% number regarding photography works if you consider everyone who takes pictures (which is most of the human population).
"But just as computers can be broken down based on needs, there are niche markets of people who need more computing power. And fortunately there are enough photographers who want more than a cell phone can deliver so we have a niche market size large enough to appeal to several companies to supply those needs. Once we narrow our field to this group I don’t think the 99% number is accurate (again because I think at that point we are talking about captures which need it, not people).
"I don't know whether a game changing new form factor is in store for photographers (wouldn't surprise me) but as the equipment becomes more capable through progression as outlined by Ctein, many who don't think they need it will find on occasion they wish they had it. The good news is if you don’t want or can’t afford to be on the bleeding edge in gear and don’t rush out and buy the latest and greatest, at least there are enough photographers who do to keep driving the technology forward. This means eventually almost everyone will find occasion to be using better gear than they have today, and will find on some occasions, some captures benefit."