Love, err, rumor is in the air, Sony has their new popular A7 III making inroads, Nikon just officially confirmed their full frame mirrorless development with expected launch on or before spring of 2019, and rumor sites are reporting Canon is actively testing multiple prototypes in the field for their own full frame mirrorless model/s.

So a couple things worth bringing up, first of which, why bother?

(A conversation about mirrorless optimized mounts vs legacy mounts)

Some of the original benefits touted for mirrorless were size and weight savings resulting from the elimination of the mirror box and re-engineering of the optics to exploit the shorter flange distance. Although true, and my former favorite platform of choice, the EOS M, is an excellent example of those benefits with normal aperture optics using finished plastic body and optics and STM motors to maximize the weight and size savings coupled, with the loss of the mirror and lenses like the EF-M 22mm f/2 STM and EF-M 11-22 f/4-5.6 IS STM which exploit the flange distance benefits (which wide angle and fast-er wide to normal focal primes can), it (mirrorless) does deliver to the promise on that front.

But, things change in relevance when you move to a full frame front, which has more complete lens lineups (Sony E, Canon EF and Nikon F mounts) including faster equivalent apertures and/or longer or wider focal offerings where the glass itself compromises a larger ratio of size/weight of the lens and conversely, you need a larger grip on the body it’s attached to in order to maintain balance / handle this size and weight accordingly. Thus the body has to remain so big even if the mirror box isn’t there. Simply put, just my opinion, the crop body EOS M for example makes sense if you want an ILC and want to save on size and weight. ~50g weight reduction on a crop body with and 50-200g reduction on normal aperture crop lenses and a good solid inch in flange distance is a big deal on a crop. But, the same 100g, maybe 200g savings when you’ve got an f/2.8 (or faster) optic attached or 400mm capable lens that’s already big and heavy, that 100g or inch is small potatoes as it was big and heavy before, and still big and heavy after on a full frame.

Numbers don’t lie, so I’ll make 2 brief average examples (trust me I’ve run multiple scenarios with differing optics and cameras, these two examples represent the state of things) as it took me running the numbers for me to make the jump to the EOS M5 over the SL1 series (as I considered switching to the SL1 right before the EOS M5 launched), and likewise for me to mentally settle the debate as to the relevance of mount differences on a full frame format.

First up, Canon’s smallest modern crop DSLR body vs their native mirrorless crop body with each’s own native superzoom attached.

Rebel SL2 - 453g in weight, approximately 48.28 cubic inches in volume

EF-S 18-135 f/3.5-5.6 IS USM - 515g in weight, approximately 27.56 cubic inches in volume

Combined, we have 968g in weight and 75.84 cu inches in volume for the smallest modern Rebel and modern native crop superzoom, pretty good for a DSLR and capable superzoom.

Next up, the Canon mirrorless equivalent, sorta, the M5 is higher up in the food chain, but I’m doing this intentionally instead of say an M100, which would be the smallest M, and also wouldn’t represent this scenario accurately as the FF’s I’m comparing to in a moment have EVF/OVFs…

Canon EOS M5 - 427g; 38.13 cubic inches

EF-M 18-150 f/3.5-6.3 IS STM - 300g; 15.38 cubic inches

Combined we have 727g in weight and 53.51 cubic inches

To put that in perspective the native mirrorless crop system (EOS M + EF-M) occupies 25% less weight and 29% less volume then its smallest Rebel counterpart with equivalent native crop glass, that’s a notable savings in size and weight indeed. It’s also notable in use as I can attest to having owned it.

Now let’s turn attention to non-plastic fantastic builds, full frame optics and bodies, some weather sealing and no-compromise fast optics.

In this case, by comparing a native full frame mirrorless format, albeit not Canon (Sony), against a Canon DSLR competitor you can get a good feel as to what to expect in size and weight savings if Canon did do a mount optimized for mirrorless full frame from scratch as the same rules apply...

Sony A7 III - 650g / 54.56 cubic inches

Sony FE 24-70 f/2.8 GM - 886 g / 50.02 cubic inches

Combined 1536g / 104.58 cubic inches

Canon EOS 6D II - 765g / 72.63 cubic inches

EF 24-70 f/2.8L II USM - 805g / 42.42 cubic inches

Combined 1570g / 115.05 cubic inches

But in this case a mere 2.2% weight savings and a 9.1% volume savings is realized when comparing full frame DSLR against a full frame mirrorless with native mount. This is a completely different ball park here with the added glass makeup of full frame optics, weather sealing, non-plastic builds, and larger grips is largely diluting the proportional benefits of mirror box loss and optical optimization on the full frame professional body and optics.

Moreover, you can rerun this particular scenario with telephoto options, or ultra wide angle f/2.8 lenses and the benefits shrink further still as now you’re talking even larger glass makeup of the total weight and size.

Arguably where mirrorless benefits still matter in optics is with normal aperture normal primes and normal aperture wide angles; as I mentioned before where that flange distance permits “better” optical formulas. Thus it would seem to make some sense that perhaps Canon would want to consider some kind of hybrid mount solution for these scenarios at first glance (Update: See my post on the ultra wide angle FF mirrorless lens patent, they have considered it LINK). However comma, the great irony is Canon already has this covered via the EF-M 22 f/2 STM and EF-M 11-22 f/4-5.6 IS STM, the crown jewel lenses for the aforementioned EOS M series... Almost like Canon saw that coming as those convert to a 35mm f/3.2 and 17-35mm f/6.4-9 respectively, and, do so with an even smaller body then a FF would permit, while also remaining economical and those (EOS M) bodies double for crop duty.

Hmmmm, smells of someone saw this one coming (the fork in the road deciding what to do about a native FF mirrorless mount) between the EF-M mount being just big enough if they wanted to do a hybrid solution, they could, but, they also covered their bases on the other end creating the most optimal optics for the native crop format they were going to produce anyways for real size/weight benefits (not to mention cheaper production costs, which is important to entry level offerings where cost is king).

I bring this up as there’s been some speculation that the EF-M mount is just large enough to adapt EF optics to with the imaging circle’s size. True, I believe it is, and Canon may well have future-proofed the mount for a full frame mirrorless future in mind in case they decided to go that route later on. That said, when you start looking at say any of the A7’s with a Sony G attached (like I just did), they’re very in line with Canon 5D and 6D bodies with L’s attached, both in size and weight. Thus I feel the subject of mounts actually is one of the least important matters here after careful consideration; whether Canon makes it one way (new native FF mirrorless mount) or the other (rip the mirror out of the box leaving empty space but maintaining the 44mm flange for EF compatibility), the promise of size and weight savings of mirrorless is too small to be noteworthy in the full frame weather sealed professional environment vs the crop, entry level, non-sealed environment.

Simply put, the weight and size savings of going to a new mount, is too small for most traditional full frame professional optics (and bodies); you can’t ditch your bag, it still weighs 1500g in my examples be it 1536g or 1570g, and you can squeeze maybe an extra nifty fifty in your bag at the end of the day in volume savings, all for going through the whole exercise of designing and producing new optics.

You can see where this is going as Canon directing their efforts in that direction, instead of say making a new EF 50mm f/1.4 USM update or say EF 24-70 f/2.8L IS USM, the answer is pretty obvious from professional shooters pragmatism and thus Canon management’s decision making viewpoint is clear (if I were in their shoes, anyhow).

Now where this would make sense is if you’re starting from scratch, you don’t have a load of EF or F mount optics, or, those optics need a redesign anyways. In that case this becomes a “freebie” to do it in the most optimal way. The fact is though from a Canon point of view, us Canon folk (I can speak for myself here), have beef with select lenses, like the ones above, but otherwise, no redesign is needed largely. But, to play devil’s advocate, if mirrorless is the way of the future, why not change the mounts now, and use an adapter, and make new lenses mirrorless-optimized going forward, say that 50mm f/1.4, which btw, would benefit from being mirrorless native, so? Canon might.

But to rebut that argument, making native mirrorless mount optimized glass and producing an excellent adapter does have a downside, existing DSLRs can’t use it. Oops. Yeah, there’s the gotcha, that’s a big one for Canon whose bread and butter is still, DSLRs...

Thus that’s not to say Canon won’t say do a special full frame mirrorless with a native mount and select lenses, they may, and give it an slick adapter (say a modular adapter), but it’s pretty clear in my eyes the mainstream Canon full frame mirrorless body that’s rumored late this year (maybe early next) will be EF mount after the dust settles, or, even if they do a new mount, they won’t be redesigning all their glass until mirrorless becomes the dominant player in their own sandbox and instead will push an adapter in some shape or form for all but optics that benefit from mirrorless. They’re just not going to redesign and reproduce that much glass for a tiny, tiny benefit. If it were a larger benefit, absolutely, but that’s simply not the case; you can’t change physics that you need so much glass at the end of the day for whatever f/2.8 monster lens you have, and you should have a somewhat capable grip on the other end of it.

In any event, I’m perfectly happy either way as I have no DSLR or EF glass at the moment, easy for me to say...

So if it’s not about weight and size savings, why again does Canon care to bother moving away from a full frame DSLR? That’s part of the dilemma here, and arguably why Canon hasn’t engaged the subject sooner, instead choosing to continue to do all the technological legwork necessary for a full frame mirrorless success via their crop mirrorless EOS M series which stands to replace their existing entry level Rebel (KISS) DSLRs someday and keeping the full frame mirrorless on the back burner in their labs till they feel the time has come to displace their bread and butter 1D, 5D and 6D DSLRs. However, there are serious benefits other than size and weight and hence it does make sense to release a full frame mirrorless option eventually, particularly as the competition steps up to the plate like Sony has been of late.

 

Benefits of Mirrorless:

 

Silent Shutter / Electronic Shutter: No mirror, no need to make noise. Handy for low profile events like weddings, libraries, etc. The EOS M50 did adopt this, however is intentionally limited to Auto mode only. Higher end offerings like the M5 II later this year and upcoming FF mirrorless should adopt it’s availability in PASM.

No Auto Focus Micro Adjustment (AFMA): Right now you have to calibrate each lens for precise AF to prevent any calibration drift on a DSLR. Mirrorless eliminates this entirely as focusing is done on sensor now. This is very important for shallow depth of field lenses (think portrait shooting), or, for peak sharpness which many folks have noted the A7 III is much sharper then say a 5DIV shot next to it; this is a large contributor of why the A7 III is sharper yet has less megapixels then the 5D IV.

Face Detection / Eye-AF: Takes all of the hassle out of shallow depth of field portraiture focusing. Allows you to either focus (and hold that focus) on a face or an eye (useful for f/1.2 situations). Presently available on the EOS M50 in single shot AF; however, that may (probably) be an intentionally gimped feature due to the EOS M50’s market position as it has no problem doing the former with tracking, thus why eye-AF is limited is probably a design limitation on the M50, not a technical limitation (just like the silent shutter being auto only). To play devil’s advocate for a second, this could be a compute constraint, but, even if so, there’s a (very) good chance the full frame mirrorless offering we’re discussing may be DIGIC8+ or Dual DIGIC 8 thus not a compute issue for it (the upcoming full frame Canon mirrorless) even though it’s possibly a compute constraint on the smaller, lower end EOS M50.

Touch and Drag AF: Very novel way of quickly focusing using a touchscreen while using the EVF, other than a traditional AF joystick, which Canon’s particular flavor is class leading I might add.

Full frame AF coverage potential: No more diamond AF coverage. The EOS M50 does 80%x80% coverage across the frame on all lenses and 88%x100% coverage for select lenses (more on that in a bit). The former is given, the latter, will be lens dependent IE we may only see 88%x100% coverage on select lenses, probably newer and/or more popular ones based off what we know about the M50 in this regard. Update: See the link here for what is presently supported

Focus peaking: That EVF (electronic viewfinder) can overlay focus with colors so you can readily gauge critical focus for things like macro photography or manual focus only lenses with ease.

Live exposure: Live preview (from the electronic viewfinder) of exposure vs traditional post-review (on the cameras LCD, or computer later) of exposure. No need to have a series of shots out of exposure and find out later anymore that you should’ve bumped your exposure comp / changed your metering spot.

IBIS (in body image stabilization): IS for all lenses via stabilization on sensor. Think your nifty fifty or 35mm f/1.4L has IS now. Handy for low-light situations with static objects and no tripod, or, adding additional stops of IS in conjunction with existing in-lens IS. There is a caveat here, Canon hasn’t implemented this yet on any of their EOS M’s and thus isn’t guaranteed to do this anytime soon. It would require a special sensor meant for a mirrorless, which flies opposite of Canon’s present logic of reusing existing parts across models. Also, Canon has quite a robust lineup of stabilized lenses, including many stabilized primes now (100mm f/2.8L, 85mm f/1.4L, 35mm f/2, 28mm f/2.8, 24mm f/2.8) so they’re not going to be as quick here to run to IBIS due to abundance of in-lens IS options already available in the same way they haven’t been quick to run to mirrorless either because of DSLR availability. That said, if Canon was going to do IBIS, they’d do it first here on a full frame pro body that has a sensor optimized for mirrorless use (as I expect they may optimize the sensor in other areas too, IE newer DPAF, or light absorption units which I’ll get to later). I give this feature 33% probability as such in the upcoming model. Canon may or may not get to this feature immediately, but it is a benefit of mirrorless so they may eventually get around to it.

Enhanced IS: Recently Canon has been rolling out an additional 1/2 stop of IS for specific lenses, the G7X II was the first, followed by the G1X III, and now EF-M 15-45, EF-M 55-200 and EF-M 18-150 lenses. The premise is the sensor data itself is used as an additional point of reference for the IS gyro motors to make in-lens IS corrections more accurate based off my understanding. Canon may be opting to go this route in lieu of IBIS for the time being, or foregoing IBIS entirely for a time should they decide they wish to standardize their sensors across mirrorless and mirrored models until popularity / adoption of mirrorless exceeds their own DSLRs. I’d say this is the route Canon will be going for some time instead of IBIS, if I had to guess, just like they’ll probably go EF mount. I also think it’s very smart for a variety of other technical reasons I won’t get into on this article. Just like the wider AF coverage support, this may be very lens specific in support. Right now the trend of the EOS M50 is 2014 or newer optics we know can support this feature. A few noteworthy lenses to almost certainly get this support; EF 24-105 f/4L IS II USM, EF 24-105 f/3.5-5.6 IS STM, EF 100-400 f/4.5-5.6L IS II USM, EF 70-300 f/4-5.6 IS II USM and EF 85mm f/1.4L IS USM. They’ll probably be more, but I don’t have enough data at this junction to make an educated guess with any real accuracy otherwise.

Faster FPS: DSLRs are limited to how fast the mechanical shutter and mirror can operate, as such, mirrorless cameras can greatly exceed traditional frames per second through electronic shutter. However, this is very dependent upon where Canon places the upcoming mirrorless in their profile and what sensor they choose to use, or, make from scratch. If say, they repurpose the 1DX II sensor, which they could as it has a fast readout with 16FPS in live view (same as mirrorless) and, full 4k readout at 60FPS, they could permit all the above, but, it would compete with the 1DX II to a degree potentially at a lower price point, somewhat anyways (we’ll get to the somewhat, that is low light AF sensitivity is an on-sensor AF weakness) so that’s a big question if Canon will permit 16FPS or artificially limit it to less than that should they use that particular sensor (I think they’ll use a newly developed sensor btw, more on that in just a bit).

New sensor?

There are another set of “benefits” specific to the upcoming Canon Full Frame mirrorless offering; but these aren’t mirrorless specific. Rather, there is a good chance Canon may choose to make a sensor just for it, be it for IBIS reasons, or more likely, in a very Canon-esque fashion, use one the new upcoming FF sensors from another upcoming DSLR model which could include: more MP, newer DPAF, improved ADC, a newer CFA and maybe if we’re really lucky, BSI finally. I think, in all likelihood, they’ll make a successor sensor to the 1DX II or 5DSR, or both, and launch it here first (via two models, hence the rumor about multiple models in testing, and maybe they’re testing a special model with special mount too). If the latter (re-use of an upcoming DSLR refresh sensor), it probably won’t get IBIS though as that’s not suitable for DSLR implementation as IBIS to my knowledge (could be wrong) is a very mirrorless affair with constant readout dependency (mirror would get in the way) and those electromagnets are an appendix in a DSLR. For the purpose of discussion, I’ll highlight what I think is coming, that is the upcoming 1DX II and 5DSR successor sensors as I don’t think IBIS is coming this round, could be wrong though, but Canon’s historically conservative in advancements, and with their existing IS lens lineup, it’s (arguably) less necessary.

The 5DSR was launched June of 2015, it’s just about due for the 4 year refresh mark in 2019, which late 2018 / early 2019 is the expected launch of the rumored full frame Canon mirrorless so it’s within striking range here, that is Canon already is working on this sensor anyways. The reason Canon would go this route is to capitalize on one area the mirrorless is going to be very good at, portraits (you need an audience to buy this camera at the end of the day). There’s various rumors circulating the web on this one (the 5DS/R update) so I won’t bother on specifics, but this is logical option due to the timing of the release cycle of the 5DS/R and the capitalizing on no AFMA with the Face/Eye AF feature sets and no mirror shock (a very big deal on the 5DS/R) in a high megapixel environment. Just look at who’s moving to the A7R III and A7 III (portrait shooters / wedding shooters), it’s this bunch that’s ground Canon needs to cover with their own mirrorless full frame offering at the end of the day, not the pro sports shooters, no offense (more on that later).

Logical option B, the 1DX II was launched February of 2016, so this is a bit early (IE 2020 refresh for the 1DX III per it’s 4 lifecycle), but considering the full frame mirrorless is by definition bullish to Canon, Canon could be a bit bullish here too and go for this a bit earlier than planned. Its sensor is suitable for a variety of reasons, first foremost is the 1DX series is designed for folks that want higher FPS, and don’t need higher megapixel counts. But, as such, these lower megapixel (if you count 20MP as low) sensors have faster readouts, which is utterly essential for mirrorless success. It impacts your refresh rate on your EVF/LCD, DPAF readout rates (essential if you want your first mirrorless pro offering to offer close to DSLR performance), FPS, 4K potential (as that’s a very readout intensive ordeal) and electronic shutter response (IE rolling shutter, or lack thereof). Also, Canon doesn’t quite have the silicon fabrication prowess of Sony as Canon maintains their own fabs (which they did buy Toshibas I might add) something which is incredibly costly to upgrade and you need a lot of volume to fuel (hence Sony’s advantage here making everyone and their mom’s smartphone sensor on their fabs). Thus, Canon will be in a better position here between already having a sensor suitable (they could just use the 1DX II sensor frankly, even though we’re talking the 1DX III sensor here), and their DIGIC processor itself needs to be able to keep up with the dataflow coming off it, another who’s fab is that coming off of… (which I’ll touch on DIGIC8 in a bit btw). Also, if you note, Sony’s first “serious” mirrorless was the A9, it went this route of lower megapixel offering and faster readout for their first serious attempt.

Either way, my guess is it’ll either be a very high MP count (50MP+), or a lower one (20-24MP) for target audience reasons and technology reasons respectively; I don’t expect this thing (the first FF mirrorless out the gate) to be a classic 5D or 6D competitor per say. Said competitor may come after though when their refresh cycles are closer for reusing their (5D/6D) sensors. Canon doesn’t want to upset their own paradigm too.

So while we’re on sensors, a few important notes on the state of Canon sensors…

Canon’s ADC tech is still new to them and thus they’re still refining it; we’ll probably see a newer ADC IE additional DR on any newer chips, probably iterative, but better; I don’t expect quantum leaps though.

More megapixels: Be it the 5DS/R or 1DX II updates, more is likely from both where the competition is going (the A9 does 24MP to the 1DX II’s 20), and there’s rumors of the 5DS/R successor being more than 50MP.

Newer DPAF: There’s patents out there from Canon regarding hexagon search multi pixel AF akin to an x264 or HEVC (high efficiency video coding) algorithm. The wagons are circling here in Canon’s R&D, IMO. This is a very cool development on an engineering level that really amounts to improved AF accuracy regarding varying patterns (like trying to shoot a line of trees, we’ll get there a bit later) but also could potentially assist DPAF further in low light situations too as it could (potentially) increase the search threshold if needed to establish an AF lock. Side note, some of Nikon’s mirrorless AF patents mirror this technique, except stagger the pixels used a bit, but still use the same hexagon-like search pattern thus it isn’t patent infringement, but Nikon clearly has the same idea. Once again, the wagons are circling this one in both camps. Don’t be too surprised if we get QPAF (quad pixel) or HPAF (hexa pixel) soon.

Updated CFA (color filtering array): We saw Canon go cooler/more muted in color rendering with the 1DX II/T6i/M3/80D/5DIV, but, recently turn back with the 6DII, something I hope they’ll continue reversing course on as I prefer the more punchy warmth of my former 5DIII over recent Canon models I’ve handled.

BSI anyone? We can only hope for some additional ISO and readout improvement, but honestly I’m not holding my breath on this one; just mentioning it because Canon can, not that they will. Everything above this is more probable. Canon maintains their own fabs, they’ll do this whenever they (can) do this; this could be the new 4K (which I’m getting to that in a bit), that is Canon may be the last one to finally implement BSI, someday, I hope.

4K (and 2018)

The recent EOS M50 is a breakthrough for Canon, not in that it’s 4K video is very good, it’s not between being a crop on a crop with jello rolling shutter and no DPAF in 4K (which Canon admits not doing the latter because it’s an entry level offering), but the big potatoes here is that it does 4K at all, and, does it not in MJPEG like the 1DX II and 5DIV, but in x264, the standard everyone else uses. It means, Canon, like virtually every smartphone manufacturer, has paid their dues for a dedicated x264 encoding chip/tech/license which last I heard runs somewhere around $30-40 per chip/license; Canon felt this was the year to finally pull that trigger and spend a couple bucks per device to bring mainstream 4K to them going forward, that’s a big deal. These tiny dedicated encoder chips/licensed tech are akin to a GPU, they are designed to do one thing, and do it (energy and thermally) efficient, encode x264 video. Thus the issue of 4K is now almost entirely can the sensor do the readout, which the jello rolling shutter of the M50 indicates the now aging 80D sensor isn’t quite up to the challenge. Considering the 1DX II chip doesn’t have this issue, Canon’s full frame mirrorless, depending which sensor they do, may have some pretty sweet 4K, or if they do a higher megapixel IE 5DSR II gone mirrorless, might not (but it’ll have some form of 4K though either way). In any event, this is the year for 4K and Canon as mirrorless has been associated with 4K between the M43’s and Sony offerings sporting it for some time so to roll out a high end mirrorless without it, in 2018 or 2019, you’re funny. I think Canon’s going to play ball here finally as the EOS M50 is an omen to on many fronts, this being one of them.

DIGIC8

The DIGIC8 in the EOS M50 shows us a couple things (and is coming to the full frame mirrorless and any other new higher end Canon going forward). 1. It’s exponentially more powerful going from being able to support 49 to 143 AF points, with the same 80D sensor as it’s big brother, the M5. As I’ve said in a former lifetime, the DPAF firehouse just gives too much data for DIGIC5, 6 and 7 to handle, and its successor probably will be able to do better, but still probably won’t make full use of it. DIGIC8 is fulfilling that prediction. 2.DLO in camera; this is a big deal as DLO is hugely CPU intensive, if DIGIC8 can do it (it can), Canon is possibly using newer fabrication technology on their DIGIC8, maybe even outsourced to Samsung or someone else, but could be a Toshiba thing. What about if Canon does a dual DIGIC8 in the full frame mirrorless or even just a DIGIC8+ (a higher clocked DIGIC8)? Even more AF points, eye AF in AI servo guaranteed, lots of FPS, huge buffer maybe even more lower light sensitivity. The DIGIC8 is a big deal, it truly is hard to overestimate what a new full frame sensor coupled with one, or two of them in tandem could do. Canon’s got critical mass here to do a serious mirrorless offering with having a backend overhaul that’s been sorely needed for some time.

Downsides of mirrorless:

Now, all isn’t peaches and cream with a full frame mirrorless offering from Canon; there are a few areas mirrorless does not do better, and some of these areas are improving as they evolve and continue to evolve, but will “hit the wall” eventually coming close, but never quite matching their DSLR counterparts. It’s a tradeoff you have to weigh and some segments of photography may continue (low light AF, think those sports pros again, particularly indoor / night sports) to utilize a traditional DSLR for the foreseeable future because of it, but some of these lines may blur (startup time, battery life) as to not matter with future iterations of mirrorless offerings.

Battery life: There’s no way to skin this cat differently, at the end of the day, you’re running your sensor constantly, your EVF/LCD constantly, your IS either in camera or in lens constantly and your cameras CPU at a higher rate to process all this data. More power draw, pure and simple. Now, as chip technology evolves, your sensor, LCD/EVF and CPU will draw less juice through things like fabrication process evolution (die shrinks) and LCD/LED tech. But mirrorless will always draw more power than a DSLR that has an OVF as it’s always drawing more current then it’s mirrored counterpart would’ve with the same technology, that is the upcoming full frame Canon mirrorless for example, may not have the same battery life as say the 5DIV, but a few revisions from now it might, but, the newer 5D at that time will still have better battery life then it’s mirrorless equivalent at that time, if that makes sense. Battery grips here can be your friend, but that’s not small or light, either (but we established it’s not about small and light, right?). Likewise carrying extra batteries is a good idea.

Startup time: Once again, your EVF/LCD and sensor have to fire up, and, your CPU and its larger OS with it before you can shoot vs the OVF of the DSLR and barebones OS/preboot OS are just going to win here. Once again, as technology improves, the gap will close, but the DSLR will always win. Now this gap may get so small it doesn’t matter in the future, but it may be a bit before that line blurs, but eventually it will as we’re down to 1 second now on some models; it won’t be long before we’re down to a half a second, and faster than that sometime thereafter. But a DSLR for example, is 0.1 seconds. How fast, is fast enough?

Low light AF: DSLRs have a dedicated phase detection autofocus sensor that unlike your imaging sensor, has VERY BIG PIXELs that soak up light with no care to resolution etc. Pure accuracy and is there something changing the phase or not at the end of the day to establish AF lock, so like an A7S (big pixel thinking) it can see in the dark. Thus it’s no wonder that even the most modern top of the line Sony (or even formerly Samsung) still can’t match a good DSLR in low light. Now, this is one area where DPAF has an advantage over its competitors as it is drawing more reference data off the sensor to make a more accurate judgment in low light as it’s every other pixel (DPAF) computing phase detection vs staggered stripes of pixels dedicated to AF (hybrid phase detect ala everyone not Canon or Samsung) has less reference data in which to establish an AF lock. Thus DSLRs will always be better here too, but, DPAF (or whatever update Canon brings to DPAF in the future) will be better than the competition unless the competition gets more creative, which could happen, but hasn’t yet, and doesn’t look to anytime soon either.