Latest or all posts or last 15, 30, 90 or 180 days.
Welcome to diglloyd.com
In-depth review coverage is by subscription.
Also by Lloyd: MacPerformanceGuide.com and WindInMyFace.com
First-time visitor
Speed To Create, Capacity To Dream
Storage Wishlist…

EVF Mirrorless vs DLSR

With pro-grade DSLRs, we at least enjoy a quality view through the optical viewfinder; good solid reasons make the OVF still relevant, at least for some shooters some of the time. That said, I’d use an optional hot-shoe area EVF most of the time on my Canon 5DS R or Nikon D810, if such as EVF option existed (but I’m not a sports/action shooter).

I got to thinking about APS-C DSLRs: invariably, the optical viewfinders on APS-C DSLRs are miserable porthole affairs with poor magnification, poor eyepoint, etc. Some may be decent, none are excellent, and most suck. Worse, they are total garbage for manual focus: it is impossible to achieve critical focus for the tiny photosites of an APS-C DSLR, and the focusing screens are very poor for manual focus, being designed solely for the AF system.

Sony A6300
Nikon D3300

And so, CaNikon moving to EVF on APS-C DSLRs makes a lot of sense: size would shrink, lens compatibility would remain—the closest thing to mirrorless CaNikon could offer with minimal effort in a short time frame. If the high end demands convetional AF, so be it, but I don’t see how the vast bulk of APS-C DSLR cookie-cutter models need an OVF for the vast majority of usrs.

So why do CaNikon sit on their collective hands? Cost is likely an issue, but it would net against the removal of the optical viewfinder. Patents and sourcing of EVFs might be another issue.

David S writes:

I think that the thing that's holding CaNikon back from producing DSLRs with EVFs is more about the type of autofocus system that is possible with a DSLR with an OVF. The autofocus sensors used in DSLRs can inherently have greater depth separation between the AF sensors which allows for faster AF, especially with long lenses. In the case of Canon's higher-end DSLRs, 3 AF sensors are used. Each sensor is set at a different focusing distance: one is at the same distance as the focal plane of the sensor, one set in front of that distance and one behind. Such a degree of AF depth perception cannot yet be achieved on the image sensor itself. The speed of AF performance that CaNikon DSLRs deliver with long lenses is the reason why they are the cameras of choice for almost all wildlife and sports photographers. They offer unbeatable AF performance for action photography.

Strides have been made to improve the speed of Image sensor based AF systems. Some image sensors now use phase detection (depth perception) to improve AF speed. Canon now includes it's image sensor based, phase detecting, Dual-Pixel AF technology on several of its DSLR models including the new 1DX Mk II. However, it is still not fast enough to replace the AF system that the 1DX and other DSLRs use in 'reflex' mode. The main purpose of Dual-Pixel AF in DSLRs is therefore faster video autofocus (faster than contrast only AF systems) with no or minimal focus hunting.

Perhaps one day sensor based phase detection will be fast enough to match or surpass 'reflex' AF. But until then, CaNikon will likely continue to produce DSLRs with OVFs.

DIGLLOYD: I suspect that 99% of APS-C DSLR shooters have the kit lens on the camera. It may be less skewed at the high end, but I’d still argue that most APS-C DSLRs are used by consumers with the kit lens and little more. If so, the idea that AF is designed for long lens shooting would be an anti-optimization.

I confess to ignorance about this market on several counts, but I would make several counterpoints that are hard to ignore:

  • I do not know whether the A6300, Sony’s “world’s fastest and most tenacious autofocus” holds up in the real world for sports/action (side issue: is the lens selection even viable?).
  • What fraction of the APS-C DSLR market is sports/action and moving wildlife? Saying that every shooter on a nature trip to Costa Rica has a Canon 7D II is meaningless in market terms (such a group is rarefied regardless of camera!) Less than 1%, 2%, 5%, 20%? I'm guessing 1% at best in context of the total APS-C DSLR market, but maybe I’m way off? I expect that the fraction is notably higher for a very few specific camera models which are essentially pro bodies for what David S refers to—and thus essentially out of scope for this entire discussion.
  • Frame rates with APS-C DSLRs are stuck at ~10 fps. What about 60 fps at 24 megapixels (the high end for APS-C)?. Real burst mode is physically impossible with mirror up/down. Thus mirrorless offers the only hope for high frame rates where it really matters—and with no mirror slap. Add in an Eye AF for animal eyes—that’s a wowee combination.
  • Sony Eye AF works better than anything I’ve ever tried for portraits. Nature shooters with long teles might find such a feature a 'killer' one since animals aren’t always flying or running; many opportunities exist with relatively static subjects (even for a second or two), where critically sharp focus on the eye makes or breaks the shot. This is awkward with a DSLR (placing a high accuracy sensor right on the eye, then often recomposing at least slightly). An Eye AF feature for animal eyes holds big potential for eliminating one source of error.
  • The Sony A7R II had a 100% sucess rate in my test on a static subject, and Nikon’s best camera (D5) had a 10% success rate on the same (D5 success rate was higher but still miserable on a different static subject). I find it hard to believe that a camera like the D5 will magically exhibit superiority when it can’t even focus properly on a static subject. I *do* understand that it *may* work better with long teles and in tracking moving subjects. But the damn things can’t even track the face of a cross-country runner coming at the camera—very high failure rate when I’ve tried to do so. Can Sony do it with Eye AF? Dunno.
  • Mirrorless AF is getting better by leaps and bounds, with rapid advances in sensor technology and CPU speed offering enormous potential. My points are not about “here and now” but about the next few years. There is huge potential to see more and more sophisticated focusing algorithms made possible by special sensors and real-time image processing. It hardly matters that right now Sony model ZZZZ does or does not currently beat CaNikon model YYYY tracking flying egrets or whatever. For now, the issue is probably lenses anyway!

Given my recent study of outdoors real-world D5 autofocus and Sony A7R II autofocus (the D5 putting in a pathetic performance), theories on “depth perception” do not ring true for Nikon’s D5, at least not for accuracy and precision, which is what matters for a sharp image, and consistently sharp results for any situation with relatively static subject matter. OTOH, I expect that the D5 performs exceedingly well with long telephoto lenses and AF tracking. But AF tracking aside, I’d argue that any camera should do well with the razor thin DoF and relatively high micro contrast of super teles and similar; focus pops in and out by tiny margins. Those multiple sensors may allow the camera to predict subject movement better, but I doubt they have anything do with with obtaining critically sharp focus—or the D5 would not have made a mess of things in my tests.

I say, retire the old and tired AF technology starting at the low end and build up from there. I think there is a strong possibility that Sony will take the lead in all areas of performance before long, particularly if a higher-end A9 appears. That includes being able to analyze the entire sensor in real time, not just a few AF points. All that’s required is a fast CPU. For the A6300, Sony claims “world’s fastest and most tenacious autofocus system with coverage density that no separate autofocus module can match”. Claims and reality are two different things, but consider what might be possible in a high-end Sony.

And so I’m not buying the argument that real-time image analysis of the entire sensor along with dedicated imaging points is going to remain inferior to CaNikon’s best: these things are subject to Moore’s Law (computer chips) and the advances in just a few years compress what took decades for CaNikon to develop with their AF systems.

Jason W writes:

People aren't obsessed with keeping their OVF's, they just want the camera to feel a certain way in their hands. The biggest resistance I hear to mirrorless is form factor. That goes for professional and consumers.

Make it look like a DSLR, handle like a DSLR, and make it cost the same or less, and it will sell. I work at a regional sports network and the Sony A7 series was universally rejected entirely on this basis before any other criteria was evaluated. The placement of the video record button was enough to stop all inquiry, and nobody liked how it handled.

A Canon Rebel with the same form factor, EF mount, EVF and lower overall cost for the kit would sell big, in my opinion.

DIGLLOYD: I agree and here’s an example: take my D810, turn it into a D810e with EVF and “I’m done”.

As for the video record button, it’s a confusing nuisance: I handed the A7R II to someone to take a picture of my daughter and me—the camera would not shoot and why? It was busy recording a video due to the idiotic placment of the video button. I had forgotten to deprogram that “land mine”. The #1 flaw (many flaws) of Sony are lousy haptics, lousy ergonomics. I work around them, but for some users they are show stoppers. And so the “love/hate” thing applies to Nikon and to Sony but for very different reasons!

John G writes:

Loved and was fascinated by your discussion/debate re EVF Mirrorless vs DLSR. Your points were thoughtfully and convincingly made.

They also happen to mirror (no pun intended) my own experiences with the SL/A7Rii vs D810/D4S. (Forget about MF autofocus—it never even enters the race.)

For most types of photography, the DSLR simply cannot compete with the autofocus capabilities of the best EVFs. But like you, I don’t shoot wildlife or sports. And as you point out, it’s hard to comprehend that these anachronistic systems would suddenly morph into superiority.

DIGLLOYD: I expect there will be a place for phase-detect AF for some years to come, but probably only as an adjunct. The first role of the AF system is a tight group of bullet holes in the target (precision*) and all right in the center of the bullseye (accuracy*). Let’s get that failure rate from autofocus error down to 0.1%.

* 'Accuracy' and 'precision' used in the scientific sense; see my discussion on the Autofocus AF-S Accuracy and Precision: Lens Align Target page.

ThunderBay 4 - The Speed To Create. The Capacity To Dream.
Storage Wishlist…
B&H Deal ZoneDeals by Brand/Category/Savings
Deals expire in 14 min unless noted. Certain deals may last longer.

diglloyd Inc. | FTC Disclosure | PRIVACY POLICY | Trademarks | Terms of Use
Contact | About Lloyd Chambers | Consulting | Photo Tours
RSS Feeds | Twitter
Copyright © 2008-2017 diglloyd Inc, all rights reserved.