In the annals of photography, it’s likely that 2017 will be recognized as the year in which the dominance in the digital photography camera market shifted from DSLR to mirrorless designs. That’s not to say that 2017 marks the demise of the DSLR, but instead, the first year in which it’s no longer possible to point to a number of key SLR strengths—most notably, speed and accuracy—and say that only DSLRs can offer the highest level of performance in these areas.
It has taken a long time to get to this point, one in which mirrorless camera can outpace their DSLR counterparts, and along the way, manufacturers, industry analysts and customers pointed to a number of ways in which the DSLR outperformed mirrorless systems. Along the way, advances in technology have eliminated each of these performance issues, moving mirrorless cameras ever closer to feature parity with the best DSLR cameras on the market, which they now surpass.
In a 2015 survey of the state of cameras for this magazine called The DSLR Is Still King, I opined “for the DSLR manufacturers, there isn’t yet a reason to panic…,” because a camera like the then-new Sony a7R II “falls short of pro-level DSLRs (and some advanced enthusiast models) in some key ways, and most of the rest of the field of mirrorless cameras lags behind that system in terms of performance and power because they’re aimed at the lower end of the market.”
It’s two years later, and now it’s time for them to panic.
How Did We Get Here?
In 2009, Olympus really kicked off the mirrorless “revolution” with the PEN E-P1. I put revolution in quotation marks because, at the time, no one really knew they needed or wanted to have a revolution in camera design. The E-P1 was charming—it looked like an old film camera pulled into the digital age—and it was certainly small, thanks to the elimination of the mirror, but it was by no means a challenge to the DSLR. With no viewfinder, no flash and a stripped-down number of features, the E-P1 was designed to be inconspicuous but wasn’t intended to mark the start of a new era in photography.
In fact, the internals of the PEN E-P1 weren’t all that new. Existing compact digital cameras by were already capturing images without a mirror, though without a fixed lens and with a smaller sensor than the Olympus-designed Micro Four Thirds sensor, both of which were found in the E-P1. The smaller Micro Four Thirds sensor and the lack of the mirror box allowed the system’s lenses to be smaller and lighter than full-frame designs.
To step back a bit, it’s important to know the purpose of the mirror, which is at the heart of the DSLR and which has been removed to make a mirrorless system. In a DSLR, light enters through the lens and is bounced via a system of mirrors and prisms (also called a “reflex”) to an optical viewfinder. The same lens is used for framing the subject as is used for capturing the image, with the mirror flipping up and out of the way to expose an image.
Previous twin-lens reflex cameras used a single lens-and-prism reflex to focus and a second lens to capture an image. These systems were faster and lighter than large-format cameras, but not as small or light as an SLR, so the new technology made them go extinct pretty quickly.
The mirror in an SLR serves two functions. The first is to bounce light up to the viewfinder and, in modern cameras, to bounce light to a secondary autofocus sensor. When the mirror is down, you can see through the lens and the camera can focus. When the mirror goes up, there’s a short period of blackout in both viewing and in the camera’s focusing.
Since the mirror takes up a good portion of the inside of a camera, removing it allows for much smaller systems. It also allows for smaller lenses, because the distance from the back of the lens to the sensor (called flange distance) determines how big a lens must be to focus properly. Smaller flange distances (achieved by removing the mirror) mean smaller lenses.
The original mirrorless cameras were revolutionary in terms of size and weight, but their diminutive designs required numerous compromises, starting with the way that images were composed. Without the mirror and viewfinder of a DSLR, photographers need to compose their images with the rear LCD screen or with an Electronic Viewfinder (EVF). Early EVFs were dreadful, with low resolution, grainy images, and a lag between subject movement and the display on the EVF. In the tradition of “you never have a second chance to make a first impression,” many would-be mirrorless shooters were so turned off by the EVF experience that they spread the word about the experience.
Early mirrorless cameras lagged behind the DSLR in a number of other key areas as well. The first mirrorless systems were built around early iterations of the Olympus Micro Four Thirds platform, so even amateur DSLRs had faster focus speeds, better image quality and better ISO performance. While the advantages (in some situations) of having a portable, capable camera were undeniable, many people wrote off mirrorless cameras as being underperforming.
The DSLR-shooting world could easily list areas in which the conventional technology outperformed mirrorless, and it would take almost a decade for each point to be addressed.
Better, Faster, Stronger (Smaller)
One of my coworkers came up with a deft analogy for the difference between DSLRs and mirrorless cameras: Think of a tollbooth as representing the path of light through a camera—light enters one side and exits the other side. In traditional tollbooths, a barrier interrupts the flow of cars so that the toll attendant can take money from the driver (think of that as focusing). In a modern tollbooth, though, the toll is read by a tag E-ZPass, without the car needing to slow down.
The traditional tollbooth in this parallel represents the DSLR, while the E-ZPass system is analogous to mirrorless. Both systems achieve the same result, but they have different tools to do so and at different speeds. When these cashless systems were first installed, many tollbooths made the cars come to a complete stop so that the early, slow sensors could read the tags. At first, E-ZPass was no faster than a conventional tollbooth.
As the technology improved, the speed limits were increased; first to 15 mph, then 25 and now, in some areas, the tollbooths are being replaced with archways with tag sensors on them, allowing cars to pass through at highway speeds. While it took years for the technology to sufficiently advance to the point that tollbooths could be removed entirely, the result is an elimination of bottlenecks and improved traffic flow.
With the limited performance of the original iterations of the mirrorless technology from Olympus, Panasonic, Sony and Fujifilm (plus companies no longer in the market, such as Samsung), many saw mirrorless cameras as a nice thing for enthusiasts to use but far short of what a professional would use—and they were right.
Photographers, especially pros, would point to the key shortcomings of mirrorless: focus speed, frame rates, operational speed, image quality, ISO sensitivity, EVF quality, startup speed and blackout times (the period when an EVF goes black between capturing frames). DSLR shooters and the manufacturers rightly pointed out that no pro would use the new technology until it matched or exceeded that of a DSLR.
And then, of course, mirrorless cameras began to check off each one of these boxes. The APS-C-based Fujifilm X100 (mirrorless but with a fixed lens) had excellent image quality. The Sony a7 brought a full-frame sensor into play. Olympus ratcheted up its image quality while also ramping up AF speed and capture rates, and so on.
When I wrote my previous analysis of the DSLR and mirrorless markets, the Sony a7R II had just arrived. This camera introduced an incredibly high-resolution sensor in a camera a fraction of the size of a comparable DSLR. Sony’s sensor division created a new in-house sensor using a technology called Backside Illumination, which allowed the a7R II to produce high-resolution images but without the high-ISO noise that the increase in resolution would normally cause.
The a7R II marked the first time a full-frame mirrorless system exceeded the resolution of a full-frame DSLR, and it would retain that title until the Nikon D850 arrived in 2017. While it had powerful imaging capabilities, the a7R II wasn’t fast. The startup time was terrible, focus speeds were mediocre (compared to top-end DSLRs), frame rates low, the battery life comical, and the buffer filled too quickly. But the images were the best in class, and the focus tracking could follow both faces and eyes; the camera could output broadcast-quality 4K video.
In October 2015, Leica entered the mirrorless fray with its Leica SL, a camera with an understated design and relatively slim form factor that could capture 11 frames per second. When Nikon and Canon, respectively, released their top-end DSLRs in 2016, their frame rates, too, cracked the 10-frames-per-second mark, an impressive speed for DSLRs, but Leica had already reached that mark with a mirrorless camera.
Several months later, the Olympus OM-D E-M1 Mark II arrived, with a mechanical shutter rated at 10 fps with full AF, 18 fps with locked focus, an electronic shutter rated at 18 fps (AF-C), and a blazing 60 fps with locked focus. As with the Sony a7R II, the OM-D E-M1 Mark II had chipped away at one of the key limitations of mirrorless cameras, and like the Sony camera, it fell short of pro DSLRs in a number of other areas. But the battle lines were shifting, and for the first time it became clear that SLRs technology has advanced nearly as far as it can, while mirrorless systems are just getting started.
Those following the market pointed out that the small size of mirrorless systems would likely be the limiting factor in their development, that to add features and performance, mirrorless cameras, if only the pro ones, would need to get bigger. The a7R II only had room for a minuscule, low-capacity battery and a single SD slot, for example. There was no way Sony could both increase the performance and address these physical limitations while keeping the body the same size. If mirrorless systems would have to get larger in order to beat the performance of DSLRs in all of the key areas, they would lose much of their appeal.
Les Jeux Sont Faits
What pundits didn’t count on was the speed of miniaturization (plus, the aggressive R&D campaign undertaken by Sony to compete with Nikon and Canon) and how that engineering would overcome the physical limitations of the small camera body. When the Sony a9 arrived in April 2017, breaking past the performance of pro DSLRs in a body that was only marginally larger in size compared to the a7R II, it became clear that mirrorless cameras won’t necessarily have to grow in size to radically increase performance.
In fact, medium-format mirrorless cameras from Hasselblad and Fujifilm arrived, both of them in packages much smaller than any traditional medium-format camera. The Fujifilm GFX-50S (which I reviewed in the July/August 2017 issue) packs a sensor much larger than full-frame into an SLR-style body, with surprisingly fast performance.
As mirrorless systems evolved, each of the limitations fell away. EVFs got better, AF improved radically, resolution increased, ISO performance skyrocketed and so on. The last remaining argument in the battle between mirrorless and DSLR was that full-frame mirrorless cameras are too slow to be used for something like sports or news gathering.
The Sony a9 was clearly designed to counter this argument; it’s as if the engineers picked up the top-end DSLRs and said, “Let’s beat this in every specification, and keep the body about the same size.” The Sony a9 is only marginally larger than its a7 predecessors but outpaces pro DSLRs at every step.
The camera can capture at 20 fps with full autofocus and auto exposure (while DSLRs at their fastest shoot with focus locked on the first frame) and uses an electronic shutter for silent operation. There’s no blackout between frames, which means there’s no period where the image in the EVF goes black while the camera focuses. It also has Sony’s in-body 5-Axis stabilization, for reduced vibration effects—the pro DSLRs only offer in-lens stabilization on certain lens models.
With 693 phase-detect points across 93 percent of the frame, the AF system in the a9 vastly exceeds that in top-end DSLRs, which have around 200 points covering much less of the frame. The a9 can perform eye-detect focus with each frame (even at 20 fps) and sports dual SD slots, and a higher capacity battery that’s finally on par with a pro DSLR.
That would likely have been enough to shift the balance of power in the market, but a few more skirmishes erupted in time for the PhotoPlus Expo trade show. First, Nikon released the Nikon D850, a superb camera that’s the best DSLR to arrive in years and one of the best cameras Nikon has ever built. It has a 45.7-megapixel sensor and can capture 7 fps unless it’s connected to the external battery pack, at which point it can capture 9 fps. It has a silent shooting mode and can capture 4K video, albeit with manual focus.
A month later, the morning before PhotoPlus Expo, Sony announced the a7R III, which features the same 42-megapixel sensor of the a7R II with, essentially, all other features pulled from the a9. The specs read as if, once again, Sony’s engineers took a Nikon camera and tried to exceed it. The a7R III captures at 10 fps, has three times as many AF points, has a one-stop better ISO performance, better 4K video specs and so on.
We now have full-frame mirrorless cameras outpacing the performance of pro DSLRs, Micro Four Thirds cameras doubling the speed of pro DSLRs and grabbing 6K video stills, and medium-format mirrorless cameras that fit into the size of SLR bodies.
The Aftermath
SLRs have physical limitations due to the mechanical parts, and, after decades of evolution, there are only marginal gains to be had there. None of this is to say that the DSLR is dead—it’s not, and not by a long shot—nor is it to say Nikon and Canon are going to suddenly vanish from the camera market. Today’s DSLRs are tremendously powerful. I’d even say they’re the best cameras that have existed in the history of photography. A DSLR won’t suddenly become obsolete just because a new mirrorless camera comes out.
It also doesn’t mean that Nikon and Canon won’t get into the mirrorless market, which would certainly shake things up, and occasional comments from Japanese executives for both companies hint at expansion into mirrorless.
Every day that goes by, though, further the advances by the key players in the mirrorless space, which means that any system from Nikon and Canon has to be that much more advanced in order to compete. The next few years will determine the outcome of the mirrorless and DSLR conflagration. However it shakes out, photographers can only benefit from the increased competition and associated innovations that the battle for dominance creates.
In my opinion, the detracting quality of EVIL (a term I saw someplace some time back – Electronic Viewfinder Interchangeable Lens) cameras today is the EVF (electric view finder). This has to do with two aspects: lag time and resolution.
While EVIL cameras have certainly improved in the former department, there is still a barely noticeable time lag between the action of the subject and the instant that action is visible in the EVF. Now this is understandable, given the reality that the camera’s computer must de mosaic the raw data coming off the chip and present it to the EVF’s driver. It’s like there’s a tiny TV studio in there… go to an electronics store and wave your hand in front of a camera and note the slight lag before you see that wave on the attached monitor. In comparison, a DSLR has photons bouncing off a few reflective surfaces at the speed of light.
On the resolution side, again, a DSLR has multi bronto-pixel resolution (or however many photons are being bounced to your eye). An EVF is fine for most purposes, and certainly better in some – like if you have an 8 stop ND filter on the lens, where you cannot even see through it on a DSLR but the EVF acts as if it’s not there. But even with focus peeking it doesn’t come close to an OVF. For that matter, the effect of a CPL is all but impossible to discern in an EVF.
But those things are merely tech problems that will, no doubt, be addressed in time. Additionally, since the light path in an EVIL camera is no longer dependent on reflections, radical departure in design is possible. For instance, the entire notion of having to have the lens and the Viewfinder in the same box. While physics dictates that requirement for OVFs, there’s no reason the lens/chip unto can’t be separated from the EVF, which could lead to interesting new designs.