Back in 2009 I argued that the choice of sensor format should be made based on the value it provides us as photographers and not simply by its size. In the intervening years, I’ve suffered slowly from sensor size creep, going from APS-C, to APS-H, and finally to full frame. While, each of theses steps were driven by my photography, though even I have to admit there’s still a little voice nagging at me – “why I’d ever want to go back to a smaller sensor?”
In a fantastical world where money, size, and weight are never considerations, the sensor size question is trivially answerable. Physics says that a bigger sensor will always produce better images simply because it collects more light. All else not being a factor, a 600 mm f/4 on a full frame sensor is going to produce better images than a 400 mm f/4 on a 1.5x APS-C sensor.
However, in the real world, size, weight, and cost are some of the most important factors.
When I was preparing for my trip to Alaska last year, I knew I was going to be shooting at long distances, pretty much everywhere. That’s just the reality of being on a big ship, combined with the realities of having to make the best of whatever other opportunities presented themselves along the way. In gearing up for this, I seriously started thinking about crop sensor bodies again.
Like most things in photography, everything is a balancing act. Shutter speed versus aperture versus ISO; motion blur versus depth of field versus noise. Or in this case, frame rate, resolution, and reach.
Going with something like the 7D mark II would have given my 100–400 the reach of a 160–640, and the AF system would allow me to stick a 1.4x teleconverter on there and still retain an AF point. And it did all of this with an action shooting 10 FPS.
As an aside, there’s alway someone who loves to point out that focal lengths don’t change when you change the size of the sensor. They are of course right, but as I’ve argued before, focal lengths as used by photographers are a proxy for angle of view, and angle of view is defined by the frame size.
Of course, what I’m calling reach is more complicated than just frame size. Pixel pitch has to be considered too. If I had a full frame camera with the same pixel pitch, like say a 5DS instead of a 7D mark II, I could crop in post for the same effect. However, while the pixel pitch is the same, the number of pixels aren’t, and as a result the 5DS doesn’t shoot at 10 FPS, only 5. Like I’ve said, there is always a trade off.
Now, with Nikon’s return to the pro APS-C camera space with the D500, Nikon users have the same kind of top tier few compromises camera as Canon uses have enjoyed.
In many ways, the biggest limiting factor now is no longer sensor or the camera. It’s the lenses and physics.
Sure you can get a comparatively inexpensive 100–400 or 150–600 and get tremendous reach with a 20MP APS-C camera. But what makes these lenses inexpensive is that they have relatively slow apertures.
Canon’s 100–400 and Nikon’s 80–400 are f/4.5–5.6 lens, Nikon’s 200–500 is a fixed f/5.6, and the Sigma/Tamron 150–500s are f/5–6.3. A 7D mark II is diffraction limited at f/6.3, and a Nikon D500 just shortly after that. Admittedly it’s not like diffraction driven softening is a hard wall, and since the bayer pattern limits the sensors resolution to less than the actual pixel pitch there’s a tiny bit of breathing room. But not much.
If you’re shooting either of those 20 MP crop sensors, with an f/6.3 lens, you basically have no room to stop down. Any gains you might have gotten in sharpness form stopping down are going to be taken away by diffraction.
On the other hand, this is where the larger lower resolutions sensors shine.
A 22.3 MP 5D mark III isn’t diffraction limited unit f/9 — and practically speaking f/10 or f/11 before you begin to start seeing it. An f/5.6 or 6.3 zoom has 2 stops of room to improve the lens’s sharpness before you start running into diffraction.
And of course this is also where the really expensive lenses come back into the equation. A $10,000 400 mm f/2.8, not only starts out as a considerably sharper lens than these slower zooms, but it can be stopped down 2 stops, even on a 24MP APS-C camera before you hit diffraction limiting behavior.
And thus you have the predicament for the crop camera.
On one hand, the smaller denser sensors allow you to get much more reach out of a shorter focal length lens while still maintaining high frame rates. On the other, you’re almost immediately going to be seeing diffraction effects, as a result you need either a fast lens or one that’s sharp when wide open; neither of which make for inexpensive lenses.
And thus we kind of come full circle. With a $3000 5D mark III, you can use an $1000–1800 lens with some margin to stop down. With a $1800 7D mark II, you need a much more expensive lens to “match” the image quality.
Maybe in the end, the real argument here is that we’ve reached the end of the point where it makes sense to push sensor densities any higher, at least for crop cameras. Or maybe it just means that we should push further. Going well into the diffraction limited region means you can rely on diffraction to band limit the image and therefore you don’t need to have an anti-aliasing filter at all. But that’s a thought for another time.