"People who say they don't need eye AF are doing their clients a disservice."
Apparently I've been doing my clients a disservice for over 40 years. Who knew? Thing is, even in the manual focus days we found ways through practice, et.al., to get eyes in focus. The folk that are running around saying that a camera is a failure because it doesn't have an automatic feature are basically indicting themselves: they think the automatic feature now gives them parity with those who have been doing the job for years. The same was said for automatic exposure, automatic focusing, and a host of other automatic features.
Don't get me wrong. The addition of a feature (e.g. optional Eye Detect AF) is always welcome. It gives me another option to use. But in order to use it effectively, you need to make sure you know how it really works. My experience with the Sony A7/A9 cameras, for example, is that this feature can be helpful at times, and it can produce problems at others.
One real issue with Eye AF is that it often doesn't focus on the eye. Oh, it centers the focus sensors on the eye area, but typically the eyebrows and especially strong eyelashes tend to be the focus point that is actually chosen, not the eye itself. With fast lenses, particularly telephoto ones, the DOF is shallow enough to see a degradation of the acuity of the iris and pupil.
This is one of those "close enough for me" kinds of things that you have to be careful about. It very well may be that close enough is good enough for you. But frankly, you don't need a US$2000+ camera if you're into "good enough" shooting. And on gear that expensive I demand and expect features that will allow me to nail focus exactly where I want it to be. On mirrorless cameras, one of my issues still is that manual focus override by wire is not nearly as easy to get right as it is with the DSLRs.
"You need a dual slot camera to be safe."
Hmm. Where does that stop? Three slots? Four slots? Two slots with offsite backup via Wi-Fi? ;~)
It's only been recently that we've gotten two slot cameras. Quite frankly, the mismatched two slot cameras (XQD+SD, CF+SD, or mismatched SD tech ala Sony) I think do as much harm as good. They tend to be like having a throttle limiter if you use the second slot for Backup (and that's the only shooting option that truly makes sense for dual slots in the arguments people are making for why they need two slots).
Personally, I'd rather have performance and reliability, which means a single XQD over a mismatched dual slot anything. Sure, Sony recently introduced "tough" SD cards, which should help improve SD card reliability, but we still have the mismatched slot speeds to deal with and the way the Sony cameras perform (or fail to perform) when the buffer fills and needs emptying using the backup option.
Again, an option is nice to have. But it would really be nice to have the option optimized (matching slots). I shoot with two cards in my D5 (matching slots, though I don't use Backup, I use Overflow). I don't shoot with two cards in my D500, D850, or A7Rm3 (mismatching slots). Your mileage may differ, but you don't see a lot of pros running around saying that they've lost images due to total card failure that they couldn’t recover from.
Indeed, I wrote an article about card failures not too long ago. Short version: too many people think cards are "forever" and just want to use that original SD card they bought 10 years ago and have been using in their old camera forever without doing any integrity checks on it. Nope. This is like thinking you drive on the same tires on your car forever. ALL cards have basically the same write limits. Like all media, they can develop low level formatting or sector errors. Use them long and often enough and they will fail. Everyone should be regularly formatting, rotating, and retiring cards. That and buying quality cards in the first place is your best protection against losing images. There are no "bargain" XQD cards; they're all quality cards. There are "bargain" SD cards, and you have to wonder what shortcuts they're taking in sourcing and testing the chips.
"More focus points is better than fewer."
For a really slow shooter willing to take the time to be precise where they put focus, probably true. Landscape photographers are in second heaven with their ability to pinpoint where focus is achieved on mirrorless cameras. In reality, though, if you're shooting any moving subjects, the more points you have when you're trying to force the focus to a specific area—as opposed to an all-auto mode—the more likely you'll still be moving the focus position when you should have been taking the shot (I often take my focus positioning choices down to 11 or half on my Nikon cameras for that reason).
So we have two things here. If you're manually controlling the focus point, you need a very quick way of moving it. Quickness is often better than Precision in this case. If I can just get the focus to the head of the flying bird or the running player, I'll be all right (particularly as I often give myself a bit of DOF to fail in). If your subject is static, the more points you can move to is obviously useful, though this can make you work more slowly.
The automatic modes—noticing a trend here?—are where things get crazy. If the Canon system is really evaluating 5000+ points while the Nikon is only looking at 200+ points (Z6), you have to wonder whether all those extra points actually are doing the right thing. The Canon might move the focus more discretely than the Nikon, obviously, but sometimes you don't want to be seeing the trees instead of the forest.
Which brings me to this: it's all about shooting and figuring out what the tool does, then optimizing your use of it. You always want to do that, no matter what the specs say.
One final point here: I was amused to get into a discussion with someone last week who was touting an all automatic focus feature and how it was better and always worked. I asked them if they'd ride in an all-automated vehicle with no user controls. Nope. And the reason why? It might not always work. Hmm. So automated cameras always work right and automated cars don't? (Yes, I know that there's a different scale of risk here.)
Software engineers develop algorithms for automatic performance that try to deal with as many of the common cases as possible. It's when you get out of the common cases that things start to go haywire with automated routines. The more you shoot, the more you're likely to encounter uncommon cases where having a human in charge is preferred to automation. Here's your new mantra: Automate, but verify.
"On-Sensor Image Stabilization (IBIS) is a must."
Another automated feature statement?
Image stabilization is situational in terms of its help, but yes, it can be very useful in those situations. Where the stabilization is done isn't necessarily as important as anyone thinks it is. In the lens means you're putting more complexity, cost, and fail points in the lens, but potentially putting the rotation at the optical center, which is probably where you want it for telephoto lenses. IBIS means you're putting more complexity and heat where you don't want it, plus more fail points in the camera, but every lens is then stabilized.
A lot of the Sony full frame users don't realize that their cameras can be capturing about 1% less than a true full frame. Why? To allow for the motion of the sensor during IBIS there's a very small bit of crop (48 pixels on the long axis to be exact). Is this crop a problem? No, but if it's there and they don't know about it, what else don't the users know?
See my article about image stabilization.
What I see as far more important than where the stabilization is done is in how easy it is to enable and disable. If the function control is buried in menus, then people just turn it on and use it all the time, even in situations where they probably shouldn't (generally high shutter speed use or where they are truly concerned about bokeh [yes, IS tends to impact bokeh; that's a long story for another day]).
Since we keep talking about automated modes in this article, I need to put a straw man proposal out there: it's fine and useful to have automated modes, as long as there is a clear and easy to use ability to manually turn them on and off (or in some cases, modify their behavior). We don't want all automated, and we don't want all manual. We want a perfect blend where we can let the camera do the heavy lifting when it's appropriate, but immediately take control back when it is not. I'd argue that this—quick manual override or adjustment—is far more important than how well the feature does or doesn't work in the first place.
"Brand X (usually Sony) is the technology leader."
This is usually written by someone who is just reading marketing statements and doesn't actually know a lot about the underlying technology and where it came from, particularly things that might be buried deep in the ISPs (BIONZ, DIGIC, EXPEED), or in the sensor (BSI, Dual Gain, Dual Pixel, etc.). People making this kind of statement usually don't know the actual origins of the technology. Image sensor stabilization, for instance, was invented and patented by Hewlett Packard, back when they made cameras just after the turn of the century. It got to Japan via HP's manufacturing partner, Pentax. Pentax has long been a technology provider to other camera makers. Doh!
Leads don't stay leads long, either. Canon and Sony, in particular, are big elephants with deep R&D and product marketing teams that are constantly evaluating what each other is doing, and looking for technologies invented elsewhere that might be useful to them in the future. But the R&D budgets at Fujifilm, Nikon, Olympus, and Panasonic aren't to be scoffed at, either.
Sure, it's great to be first with a feature or performance spec, but those things don't actually last all that long as a product advantage.
Finally, the tech marketer in me makes me want to force everyone to write feature/benefit statements for all those technology bits being trumpeting. Okay, so Camera A has X feature. What's that mean in terms of a true user benefit?
Which leads me to this: are you buying a camera for its specs or for its usage?
Now that we have four players all producing multiple full frame mirrorless cameras, I can see those different cameras appealing to different people for different reasons. Rather than arguing over which specification is better, maybe it would be more useful to examine the customer need and figure out which camera is better for that customer.
There's room enough in the market for all three major players (and six total if you count the L mount). Personally, I don't care if Canon, Nikon, or Sony turns out to sell more than the others. I think they'll all see plenty of customers, and I'll continue to do what I always do: evaluate each tool for the problems I have to solve.
To give just one example, Sony isn't really the choice at the moment if you need tilt/shift lenses. Both Canon and Nikon have a full line of tilt/shift lenses that work pretty much exactly the same on one their new mirrorless cameras as they do on their DSLRs. So if I'm into architectural shooting, for instance, the Nikon Z7 with the 19mm f/4 PC-E mounted on the FTZ adapter seems like a pretty good choice, as would a D850 DSLR. (Yes, I know I can get an adapter to mount a Canon tilt/shift lens on a Sony. The key words in this paragraph were "pretty much exactly.")
On the other hand, the A7m3 with one of the Samyang small prime near-pancakes (24mm, 35mm, and now 45mm) seems like a really good choice for silent, unobtrusive street shooting compared to my big DSLRs.
Right tool for the job. That's the discussion you should be looking for, not whose marketing brochure has more "technology wins" in it.
"Canon and Nikon can't match Sony's relentless launch pace."
Hmm. Have you actually looked at what that "pace" is? About two years+ between generations of an individual Sony mirrorless camera. Pretty much the same as Nikon did with the D8xx series ;~). What Sony did that caused this "faster" myth to spread was very nicely position their mirrorless launches so it seemed they were announcing something new every quarter, sometimes every month. By positioning A7, A9, A6xxx, and RX launches very carefully, and by emphasizing the new technology in each of those, Sony always had a new technology message that hits again and again. Realistically, though, it was a multi-year transition from DSLR-type cameras to mirrorless that was happening.
That seemed particularly different compared to Nikon, who mostly went into silent running mode after the big and loud D500/D5 launch in early 2016, with really only the D850 introduction in 2017 being played up. Coolpix and other minor updates like the D7500 didn't attract much media attention, and weren't really pushed heavily by Nikon marketing.
Canon was more active in the recent period than Nikon, but still not with a lot of dramatically new technology; they were more about making sure their current technology was in everything you could buy (e.g. the dual-pixel, 24mp, APS-C sensor is now in compact cameras, mirrorless cameras, and DSLRs).
Chalk this one up to Sony's good marketing and messaging. But if you're expecting that we won't see the same two year+ iterations on the Canon R and Nikon Z systems, I'd guess that you’ll be wrong. Moreover, these are completely new platforms: we're going to see additional models over time and more technology added, too, much like Sony rolled the three A7 models (and A9).
Put another way, in the coming couple of years, it's probably going to look like Sony is doing updates, while Canon and Nikon are releasing entirely new product, about the opposite of the last few years.
Particularly when it comes to optics, watch for Canon and Nikon to also start making bigger splashes than Sony (after all, they're deeply rooted optics companies while Sony is more deeply rooted as an electronics company). Both the RF and Z mount open up some new design territory that I’d guess Sony will have a tougher time directly matching in the FE mount. Indeed, Sony's already been heavy on the marketing defense about lenses, trying to dull the Canon/Nikon messaging before it has time to set in.
That said, I'm very happy with the Sony lens offerings now (the first few lenses were problematic, but Sony's got the rhythm now). The Sony G and GM lenses, in particular, have all been very good. As good as I really need right now if I'm to be honest. I don't think Sony has to be defensive at all.
Summing Up
The Internet tends to get all tied up with specific new features and technologies. Marketing messaging is all about FOMO (Fear of Missing Out). I'd say that you should spend much more time looking at your actual uses, and making sure that the products you buy enhance those particular uses. In other words, that the features/technology provide a benefit that you actually will gain from.
Sometimes new technology does align with use cases. One thing that I think the camera companies discovered by accident is that most people were focusing-and-reframing, but counting on the camera to do the heavy lifting.
For instance, the new full time tracking modes actually play into that, though that's not how they were designed to work. In other words, I see lots of Sony users keeping the focus sensor in the center, framing the center on what they want to keep in focus, starting the tracking (e.g. AF-ON button) and reframing, believing that the camera will still follow the thing they initially focused on. It does, so the user is happy.
I'd point out that focus-and-reframe means the chance for missing a moment in time, though. I'd rather be able to specify the thing to focus on by pointing to it while holding the framing constant, which is how the tracking methods were all designed to work. (This is why I prefer thumb-on-LCD to move the focus cursor rather than thumb stick or Direction pad, by the way. I can do that faster and more reliably if implemented well.)
So how's all this help you today? After all, we're currently in a period where we don't expect a lot of new product to be available (fall will be the next big product drop time).
Price.
We're in a period where the camera makers are unloading inventory, particularly of older product. Buying a generation behind (or at the tail end of a generation) can produce large discounts, but still net you a very usable product. But only if you don't get caught up in the hyping of features and technologies. It very well may be that a discounted older camera might let you happily shoot for several years without reducing your ability to create great images.
Back when I was in college, I use to have a button that said Question Authority. These days, that button would probably read Question Technology. The promoters of technology today have a vested interest in selling it to you, whether you need it or not. Only you can determine whether you actually need it.
