With Canon and Nikon having introduced serious, full frame mirrorless cameras, the question now becomes can DSLRs continue to get better? And if so, would they still be relevant?
Truth be told, there are benefits both to shooting with optical viewfinders (DSLRs) and EVF viewfinders (mirrorless). Each has advantages that the other doesn't. That's one of the reasons why I've had my feet in both camps as a photographer since 2009.
I'd tend to say that the most likely scenario is that both types of cameras will figure out how to get as many of the benefits of the other as they can. If that happens, then, yes, DSLRs can still improve and stay relevant for awhile (eventually the parts situation favors mirrorless).
- DSLR
- (A1) view of a scene is always present in viewfinder, regardless of on/off status of camera
- (A2) view of a scene is instantaneous to reality
- (A3) focus system generates more precise phase detect data, has ability to put large number of cross sensors in an area without impacting image data
- (A4) uses little or no power for view through viewfinder
- (A5) viewfinder image doesn't impact night vision (particularly if camera recognizes low light and uses only red LED information displays)
- Mirrorless
- (B1) view of a scene can show exposure, white balance, and JPEG style impacts
- (B2) view of a scene can correct for optical distortions (e.g. linear distortion and vignetting)
- (B3) scene view can have no viewfinder blackout during continuous shooting, ala the Sony A9
- (B4) viewfinder can overlay real time histogram and/or zebra display; ditto for focus peaking confirmation
Every one of the things I just listed for the two different systems is desirable to a sophisticated photographer, particularly one that crosses shooting genres.
If I were in charge of designing future DSLRs, I'd look at seeing how many of the things in the mirrorless advantages I could pick off.
#B4 is certainly within the realm of being done with current viewfinder technologies; we already have crude overlay layering in the optical viewfinder, so we'd just have to make it a less crude and add capability. The other B# advantages would almost certainly require a hybrid viewfinder approach, though: the DSLR would have to have a mirror-up mode with a pop-in EVF display in the viewfinder (another mechanical mechanism, unfortunately, which adds cost and complexity).
So let's examine the other side: can you make mirrorless cameras that get the DSLR type advantages. #A2, for instance, can be (nearly perfectly) done by using genlock type technologies. Samsung did this with the NX1, and Sony is doing something similar on the A9. I'm convinced that this problem is one that will eventually be solved on all mirrorless cameras once the price of doing so drops. #A3 is something that Canon's dual pixel system starts to help with, and I've seen patents of other ideas that tackle the problem of focus sensor density in particular. Again, this is something that is on the verge of working its way through the mirrorless camera market. #A4 is a little trickier, but display power consumption is a problem everyone is working on, so it, too, will happen over time.
I'm not entirely sure you can do a lot with #A5. Sure, you can detect ambient light and run the EVF differently in low light to try to preserve night vision, but do you want to give up #B1 when you do so? I don't think so. Likewise, I don't know how you do #A1 without adding an optical finder to the display. (You'll notice that I didn't write about the quality of the view; the latest EVFs are quite good. A lot of folk comment that they didn't feel that they are all that different than looking at an optical view; at least if you don't gum them up with too much information display, have them set on their highest refresh setting, and aren't using JPEG settings that punch the hell out of contrast and color.)
Are there other areas outside of the viewfinder where a DSLR and mirrorless camera differ significantly? Sure. One simple one is in heat propagation: you're not running the sensor constantly. Another hidden one is in RTOS (real time operating system). The huge data stream coming off the mirrorless cameras mean that there's potentially a greater amount and complexity of data to process, meaning faster/better CPU needs and a lot of careful fine tuning of how the camera responds to requests for change. The more balls you juggle, the more likely that you drop one. DSLRs were juggling more balls than film SLRs, and as such had a period where they had to employ more and more horsepower and greater and greater attention to how everything timed against each other. (Note, for example, that Nikon's Z cameras don't continue to process exposure changes at their maximum frame rate. A lot of mirrorless cameras have some such limitation when you really press them.)
But I judge none of those other DSLR/mirrorless differences to be as important as the viewfinder difference. After all, you're constantly looking through the viewfinder, so any differences there are easily noticed.
So, while the answer to the question in the headline is yes, the question is for how long and will that be enough to keep them viable? In the short term, yes. In the long term, no. The more the mirrorless cameras start working on that #A list, the shorter the reign of the DSLR.
