If Canon makes a full-frame mirrorless camera, should they forego some of the potential size benefits to maintain full EF-mount compatibility? And just how compatible would it be?

More than ever, rumors are circulating that Canon and Nikon are finally going to take mirrorless seriously by building full-frame mirrorless cameras. These rumors, which may well turn out to be nonsense, all seem to suggest that these cameras will be built around the companies’ respective DSLR mounts.

It’s an interesting quandary: develop a new, space-efficient mount or stick with your existing system?

I’m going to argue that the right answer is much clearer for a manufacturer than for the end consumer. And I think I can guess which option we’re likely to see.

Supporting your legacy

The benefits of supporting your legacy mount seem obvious: the manufacturer gets to keep selling their existing lenses and the consumer ends up with a huge range of lenses to choose from. Surely there’s no conflict there?

The advantage of all (or most) of an existing system’s lenses being compatible from day one seem overwhelming. Plenty of choice, the ability to sell mirrorless cameras to existing lens owners and no reputational damage. Everybody wins, right?

"There's a risk of building a system that prioritizes backward compatibility over maximizing performance"

The problem with building a mirrorless camera with a full-depth mount goes deeper (pun intended) than all of your mirrorless models being bigger than necessary. That said, even people who prefer larger cameras are usually referring to grip depth and spacing of controls, rather than demanding their camera has a big box of fresh air in the middle of it, for no functional reason.

A question of focus

No, the bigger issue is that most DSLR lenses aren’t designed for mirrorless. I’m not just talking about some designs being larger than necessary, I’m talking about the use of focus motors that are great for DSLR phase detection but that are woefully clunky when driven using contrast detection AF. Secondary sensor AF, as used and painstakingly optimized for DSLRs is very effective at telling the lens where it needs to move its focus elements to. The ring-type ultrasonic motors used in most high-end DSLR lenses are great at responding to such a command.

Contrast detection asks very different things of its lenses. Instead of racing to a particular point, they need to smoothly scan through their focus range then perform a series of back-and-forth movements to find perfect focus. The result tends to be more accurate but requires a lightweight focus element and a very differently type of focus motor.

The K-01 used a full depth Pentax K mount which gave instant access to lots of lenses. Unfortunately, none of them had really been designed with contrast detection in mind...

The alternative approach: on-sensor phase detection, is in its relative infancy. It may be able to make better use of existing lenses with ring-type motors, but it’s still not clear how well it can interpret significantly defocused scenes.

Also, at present, most on-sensor phase detection information is fed into what are more precisely described as 'Hybrid' AF systems: they get very close to focus using phase detection then perform a CDAF hunt to confirm the optimal position. Perhaps this will change, hopefully without the loss of the precision that mirrorless AF tends to excel at, leaving us just with the size disadvantage.

It's notable that, when it's trying to build fast-focusing lenses for its mirrorless E-mount, Sony doesn't tend to use ring-type focus motors. In the case of the 16-35mm F2.8 GM, it uses twin piezoelectric actuators.

Alternatively, of course, there's a risk of building a system that prioritizes backward compatibility over maximizing performance. It's noticeable, for instance, that Sony makes very little use of ring-type focus motors in the lenses its developing for the E-mount, despite having experience of using them for its DSLR A-mount.

‘It fits’ isn’t the same as ‘it’s good’

Either way, there's a risk that we'll be offered something that fits but doesn't necessarily work as well as it could.

As an enthusiast photographer with limited lens-buying resources, one of the things that has always irritated me is seeing camera companies produce high-end lenses for their full-frame customers and carefully marketing the idea that this benefits all their users, so they need not develop anything good for their APS-C users. It’s a situation that leaves APS-C users with poor choices and the arguably false impression that by buying these poorly-suited lenses, they’re making progress along an upgrade path (a fallacy that benefits the camera makers more than the photographers).

The decision to adopt a new mount or continue with a legacy one risks the same thing: the appearance of lots of choice when what you’re actually being offered is compromise, and a situation with limited incentive for the manufacturers to dedicate their efforts towards the needs of their mirrorless users. Instead they can produce a lovely picture of their mirrorless camera flanked with 30 years’ worth of lens development and watch as brand loyalists insist that ‘their’ system has the most lenses, regardless of performance.

And this wouldn’t necessarily only apply to existing lenses. Let’s say Manufacturer X needs to develop a new fast 70-200mm F2.8 and the focusing design that would work best for mirrorless turns out to be slower than the one that suits the company’s flagship sports DSLR, which version of the lens do you think we’d see?

Precedent

There is something of a precedent for this. Canon got its reputation badly burned by abandoning its FD mount – something it took some photographers a long time to forgive –whereas Nikon and Pentax pressed on with progressively trying to modernize their 1950s and 70s film mounts.

Taking the hard decision arguably left Canon in the better position: the long-term benefit was a wide-throated, all-electronic mount. With the introduction of its latest ‘E’ lenses, Nikon’s venerable F-mount has finally caught up: with autofocus and aperture operated by the lens, but with a complex series of compatibility issues cropping up along the way. And, while they are still using a somewhat restrictively narrow mount at the end of it, they've benefited from not having to burn their users on the way (though they arguably handed-off responsibility for understanding the complexities of the F-mount’s development to every user looking to buy lenses).

So what’s the alternative?

Olympus expressly made the E-M1 to provide continued support for its legacy system but also developed the 'PRO' range of high-end lenses to make full use of the capabilities of Micro Four Thirds cameras.

The other way of doing things is to develop a dedicated mount and dedicatedly support it. This is the approach that Olympus took with the development of Micro Four Thirds and, to an extent, which Canon has with its EF-M mount. Olympus, along with Panasonic, took the brave step of designing a mirrorless-optimized mount when they developed Micro Four Thirds, rather than trying to press on with Four Thirds. They then offered an adapter to use the older lenses and, with the E-M1 and E-M1 II, developed cameras expressly with the intention of maintaining support for the older, outgoing system. This meant existing customers didn’t get too badly burned and new Micro Four Thirds customers got an increasingly impressive range of native lenses designed for them.

It’ll be interesting to see if Sony takes any pointers from this, as they decide how to support both E and A mounts.

I hope to be proven wrong

Perhaps I’ll be proved wrong in the end. Maybe Canon’s EF-M/EF cross-compatibility will end up reducing the incentives to develop interesting lenses for M series owners, in the same way that I worry sharing a mount would. Equally, perhaps Canon’s Dual Pixel AF (and Nikon’s on-sensor PDAF experience gleaned from its 1-Series cameras) will mean that there ends up being no AF compromise to sharing a mount. It may partly be overcoming this challenge that has led to its camera taking so long to arrive. At which point, using the existing mount would just mean carrying around a camera that’s a little lumpier than it needs to be.

Time, I’m sure, will tell. But, in the meantime, don’t necessarily take at face-value any promises that backwards compatibility is an unalloyed user benefit.