What a Difference a Decade Makes

(commentary)

I have no easy way of proving the assertion I’m about to make, as I don’t still own the older cameras I’m going to mention. Still, I trust my eyes, and I’ve spent a fair amount of time looking deeply at various tests made over the years that both I and others I trust did. With that said…

Today’s 20mp 1” sensor cameras are performing at a level somewhat better than the 10mp DX/APS DSLRs of 10 years ago.

Specifically, I’ve been considering the Nikon J5 versus the Nikon D80, but I also have a couple of other recent 1” sensor cameras, including the even more tricked out sensor of the Sony RX100IV, plus I still have my tests from the D200 and an 8mp Canon Rebel of that era to look at, too. 

Here’s the thing, though. If you just pixel peep at 100%, you might not make the proper comparison. And if you don’t view images with at least 50% size when in Lightroom or Photoshop, you’ll miss it too (because Adobe does some rendering tricks to keep screen updates nearly instantaneous). You really have to compare apples to apples, and the best way to do that is to print at 11x14” from the older and newer cameras at various ISO values. Why 11x14”? Well, that keeps us near 300 dpi for the older 10mp cameras, so modest pixel level crud should stay invisible without looking at the paper with a loupe.

The old 10mp cameras start to fall down at either ISO 1600 or 3200 with such a test (assuming you exposed and post processed well). I’m a tough grader, so I wrote years ago that my D80 was failing me starting at about ISO 800. But I’m comfortable with what I wrote in the previous sentence, too: most of you probably would have tended to pick a max of ISO 1600 or ISO 3200 for a 14” print using the full 10mp of those older cameras.

Yeah, that last bit got italicized because that’s not how most of you use your cameras. I’ve watched many photographers over the years, and they all seem to crop a lot. More than anyone wants to admit, and I think, more than they should. Which is one of the things that was putting pressure on “more pixels” and “better high ISO” simultaneously. (More on that in a bit.)

Yes, despite my saying don’t do it, I crop most of my shots, too. Though not in the way most people crop: the filmmaking-trained boy in me likes 16:9 (actually, it likes 2:1 better, but all of our best output devices are 16:9 at the moment). Thus, when I post photos here, you’ll often see them cropped at the top and bottom because that’s how I shot them. The snarling lioness shot that was on the site during my break is a good example: I used all the horizontal pixels, but I had shot the image thinking 16:9 in the first place and cropped the top and bottom to get the final image you see here:

INT Africa Botswana Aug-2014 D7100 2344.jpg

As usual, I’m straying a bit, but this crop thing is also important to our discussion of old 10mp DX/APS versus new 20mp 1”: I feel that you can crop them proportionally about the same. Indeed, that was one of the things that prompted this not-quite-analytical article in the first place: I happened to be fiddling around with an old D200 image that wasn’t shot for 16:9 and thus ended up getting cropped a bit in ways I hadn’t considered when I shot it, and it got me thinking about how far I was able to push the D80/D200 in terms of both ISO and crop, especially simultaneously. And there was a J5 with it’s 1” sensor sitting on my desk…so I shot a couple of images with it that would require a similar crop to what I had just made with the output from an older, bigger sensor. My conclusion? Well, you saw it above: 2015 CX is a lot like 2006 DX.

There are a lot of ways to go with this little conclusion of mine. Let’s try one you might not come up with: why did Nikon make the Nikon 1 line of cameras?

I used to keep a full chart of sensor-driven gains over time. There were plenty of things I graphed, starting with Signal-to-Noise ratio. Kind of interestingly, the data I was tracking was remarkably prescient at predicting gains, much like Moore’s Law does was with semiconductors in general. Imaging sensors don’t tend to gain directly from Moore’s Law, so they map differently for progress over time. Yet the trend was visible in the 90’s and still visible here in the 21st century: image quality gains are persistent and predictable. We’ve had a couple of blips where gains happened slightly faster than predicted, and others where a generation of gains were slightly behind the moving average, but pretty much everything I was graphing formed a line that trended upwards on a regular slope. 

Nikon had digital sensor groups back into the 1980’s. They had to have been keeping the same sort of charts. And depending upon how confident they were in them, at some point those charts would have said “you can do the same thing or better today with a smaller sensor than you did with a bigger one X years ago.” 

I have to stop and apologize for a moment. Well, not exactly apologize as, well…okay I don’t know what this is, but it’s not exactly an apology. Still, I have to somewhat take back some words I wrote long ago in order if I’m going to go further with my argument today.

When Olympus entered the DSLR fray with 4/3 (the precursor to m4/3) I wrote about them bringing a knife to a gun fight. Those words were much publicized by others and became somewhat viral. I’m not going to take back those words, because I think they were true. However, I need to apologize for not crediting Olympus for being too early to a knife fight ;~). Yeah, the camera world is that tricky. 

The problem for all of the camera companies other than Canon and Nikon is that those two companies—already almost a duopoly in film SLRs by the end of the 20th century—hit the ground running really hard with DSLRs. You have to give credit where credit is due: in order, Kodak, Sony, Fujifilm, Olympus, and others all saw that digital was coming and began poking around at it. Canon and Nikon, did too, but the moment Nikon thought they had the idea of a DSLR truly figured out, they dropped their partnerships with Fujifilm and Kodak and ran with their own ideas at full speed. Canon immediately followed because they didn’t want to lose their number one spot in interchangeable lens cameras to Nikon. 

By the time the others got fully into the rush, Canikon had established DX/APS as the entry point. Kodak understood that and tried the opposite approach, being early to move to full frame. Fujifilm fully understood the Canikon rush and went with APS, as did Minolta and Pentax. Olympus, on the other hand, jumped the gun on going smaller. A 4/3 sensor at the time clearly gave up image quality against the bigger sensors, and it was going to take some time for that growth line of sensor quality to get them to where entry level DSLRs from the big companies were already performing. But it very well may have been that Olympus saw what I’m writing about today and just decided to take the future path first. Unfortunately, they suffered mightily because it was too little, too early. 

I need to apologize for something else, too. Nikon started calling their DSLRs and lenses for them DX very early on. I didn’t make enough of that distinction at the time. I think it’s clear in retrospect—especially now that we have FX and CX—that Nikon knew where they were driving. I originally took the DX designation as a marketing method of trying to get around the crop factor thing, but now looking back on it, I don’t think it was. Or more to the point, Nikon knew they were likely going to create FX before they actually launched DX in 2009. 

d80-vs-j5.jpg

Which brings me to CX. Nikon also knows what I wrote about here at the start of my article: long term we can get excellent image quality out of smaller sensors. Image quality that will match what your previous DSLR might have been achieving. And if you want to drive costs out of a camera product, one thing you’ve definitely got to figure out how to do is to stop using so much silicon for the most expensive part. Moreover, you need to make the product more convenient for the masses, which means smaller and lighter.

Like Olympus with 4/3, Nikon got to 1” too early, though. I didn’t repeat the “knife to a gun fight” remark about the Nikon 1 when it was introduced because, well, at that point Nikon was showing up with a knife, a gun, and a bazooka. Kind of difficult to fault them for covering all kinds of competitive combat. 

The “more pixels” and “better high ISO” thing isn’t going to go away. Think of it as an infinite elevator being built. Some get off at lower floors because that’s all they need. Some get off on a higher floor, walk around, somehow get dissatisfied, and get back on the elevator and try a higher floor. Some just go to highest floor currently completed, and then the minute new floors open above, they feel compelled to re-instate their “on the highest floor” claim, so get back on the elevator and press whatever new button that’s appeared for a higher floor. 

Thing is, this becomes a self-perpetuating cycle of decline. Virtually everyone is going to eventually find a floor they’re satisfied with and just stay there. After all, traveling up the elevator to a “better” floor costs them money. People don’t have infinite money, and they’ve got plenty of other such tempting elevators to explore (house, car, boat, education, other hobbies, etc.). Worse still, more and more of those floors are being occupied with smartphones, leaving only the highest floors for the camera companies to work. 

Put another way: we don’t all need to be on the top floor, nor are you ever going to find us all there. To roll this back into where I started this article: if you enjoyed the 2006 floor (D80) and just need to replace a dead or dying camera, you can pretty much do so with a Nikon 1. You’ll end up with a smaller, lighter camera, too! 

It’s not surprising that I’m writing this article while sitting in room on a high floor in a pyramid (Luxor Las Vegas). Fewer and fewer rooms the higher you go up, and fewer and fewer people going up that high. The “critical mass” of humanity in the building is down low. 

This is exactly what’s happened in the camera industry. It’s a pyramid, and the smartphone crowd is filling more and more of the lower floors. But few really need to go to the top, and the pyramid is somewhat slow to expand upwards. 

Let’s say that the 10th floor was the D70, the 11th the D80, and so on. That means that the pocketable Sony RX100IV is also the 11th floor, and the Nikon 1 cameras live around that floor, too.

Now think about all the folk that still have a D70 or D80, which will eventually get dropped, break, or get disused because it’s just a bit too large and cumbersome to carry around all the time. A Nikon V4 might be perfect for those folks. If it were designed correctly. 

And that’s exactly where we have this very strange battle going on that the Japanese are losing. In the lower floors (smartphones) the design mantra is “sharing convenience.” In the high floors (e.g. D810, Sony A7rII, 5DS, etc.), the design mantra is mostly “traditional camera with best possible technology.” 

But for those who were already on those middle floors (e.g. D80 user) or someone who is targeting those middle floors, do you see what the design needs to do? Yep, “sharing convenience in traditional camera with good technology.” Is that what the Japanese are producing? No. Instead they keep trying to reinvent the camera interface (scene and effects modes, iAutoAnything, masking the photographic decisions with new interfaces. Not. What. The. User. Wants. 

So I’m stuck with this: my current 1” cameras are as good in image quality or better than my decade-old cameras. But they weren’t designed for sharing convenience nor were they really designed as traditional cameras (thankfully the m4/3 Panasonic LX100 is a floor above, or I’d have an exception to my statement to deal with ;~). Who do these cameras really appeal to? 

Message to camera companies: stop designing cameras for customers that don’t exist. You have two basic customers to target: (1) smartphone users who want better image quality and more capability with the same sharing convenience; and (2) traditional camera users who either want smaller/lighter with the same capabilities or better image quality and more capability, or both, but recognizably the type of camera interface they’re used to. Fishing for other customers has proven futile and the few wins there have been extremely short term.

If you made it this far, consider which floor you’re on:

  • On a low floor: just keep cycling through new smartphones as they become more capable. Sensor: small as they can make it.
  • On a middle floor: you can update, downsize, and be happy, except maybe for the photographic interface, which might turn out to be something you don’t like. Sensor: probably in the 1/2.3", 1”, m4/3 range, and maybe into the DX/APS range.
  • On a top floor: only the best traditional camera will do, and it needs a state of the art sensor of some sort. That tends to limit you to Canikony’s best offering at any given time. Sensor: FX or exceptional DX/APS.


text and images © 2017 Thom Hogan
portions Copyright 1999-2016 Thom Hogan-- All Rights Reserved
Follow us on Twitter: @bythom, hashtags #bythom, #dslrbodies