Now that Nikon has rolled the EXPEED7 generation from top of the line (Z9) to the bottom (Z50II), one of the questions I keep getting is about the next generation cameras.
Many expect EXPEED8 popping up at the top of the lineup again in something that might be called a Z9II. I’d more likely expect an EXPEED7 Extended, as in a companion chip that performs a subset of actions, likely all AI-related. We’ve seen Sony do this recently, and I believe this is the correct approach for the volume-constrained dedicated camera market; you can’t keep completely redesigning your complex SoC (system on chip) to smaller process and newer core/IP as well as new functions constantly, it’s just too expensive to do that for the small volume of units (likely ~3m units lifetime for Nikon now, which would be done over four years minimum at current volumes). But you can augment that by offloading some of the work to a simpler, new companion chip in the top camera that returns more R&D money due to its lower design cost and the fact it is in a product with a high retail price.
I believe Nikon experimented with that idea using the dual EXPEED6 chips of the Z6II/Z7II cameras, though because that was basically putting two full SoCs together it wasn’t particularly efficient at adding capability plus doubled the SoC cost for the cameras. The Z8 and Z9 already have that unique dual stream coming off their image sensor, so this seems exactly where you might be able to derive a benefit with an auxiliary chip: you even further split that stream so that two chips are looking at it, perhaps with one doing the viewfinder work, the other doing the focus work.
Now that we have the Canon R1 and Sony A1 II available, the Nikon Z9 is the elder statesman of the top-of-the-line pro cameras and we should be comparing potential Z9II additions to those cameras. It's illustrative therefore to compile a list of the major things these two newer cameras have that the Z9 doesn't:
- Cross type AF detection (R1)
- Eye control focus (R1)
- 40 fps (R1)
- Main subject priority (R1) or stickier subject recognition (A1 II)
- Person registry (R1)
- Stills while recording video (R1)
- 9.4m dot EVF (R1 and A1 II)
- Hot shoe accessory electronics (R1 and A1 II)
- UVC streaming directly (R1)
- Auto framing (A1 II)
- Tilt/Fully Articulating LCD (A1 II)
- 1/400 flash sync (A1 II)
- Content authentication (A1 II, R1?)
With the above in mind, it's time to talk about the fruit. As in what might be low-hanging fruit for Nikon to harvest, and what is the tougher fruit to get to and pick? So let’s discuss what might be in a Z9II by which type of fruit it might contain.
Low-hanging fruit:
- 5m dot or higher EVF. The natural thing would be to move to the Z6III viewfinder. That’s very low-hanging fruit. Going higher in dots or nits would still be relatively low-hanging, but requires a bit of extra work and may require more use of EXPEED’s bandwidth during composition.
- CFe 4.0 support. We have faster cards now, but the camera needs card slots that support them. This might not be as low-hanging a fruit as it first seems—at least not if you want to use all that speed—because EXPEED itself needs to support the extra PCIe lanes, and I don’t know if EXPEED7 does so directly or not.
- Fill in the missing features. Surprisingly, a Z9 after all the big firmware updates still doesn’t have HEIF, Pixel shift shooting, Nikon Imaging Cloud, and a few other things that appeared first on later, lesser cameras.
- Fix the customization. I outlined how the way to save and change configurations on the camera should work over six years ago. We still are using separate Banks and a single Save menu settings capability that is no longer anywhere close to state-of-the-art for a top end camera. A configuration-save rethink is mostly reprogramming the menu system to support it. Very low-hanging fruit, you just need some laborers to do the picking.
- Improve the existing features. Another pass on the machine learning for subject detection could improve focus, plus we should easily get things like stills-while-recording-video, and UVC direct streaming.
- Any kind of raw pre-release capture. It doesn’t matter if it’s 15 fps High efficiency, or even 10 fps Lossless compressed. While neither of those are optimal, they’re 100% better than we’ve got.
- Content authentication. Nikon was first to demonstrate this (on the original Z9), but it appears they’ll be last to ship it. This is fruit that probably will fall right off the tree if you look at it hard enough.
Let me stop there for a moment. Would the above things be enough to sell a Z9II? I believe so. Basically you’d get a followup model without a ton of extra R&D investment. Every one of the things I just mentioned are things that would improve my ability to do the level of work in sport and wildlife I’m currently achieving, because they relieve pressure points. I’m pretty sure that most pros would feel the same way. (Given the low-key reaction to the Sony A1 II, I’d expect a Z9II defined by low-hanging fruit to get the same response.)
Okay, but what about other fruit?
Hard-to-reach fruit:
- Faster electronic shutter. The need for this starts with the group using flash: it would enable higher flash sync. However, in general, less rolling shutter simply tidies up the performance of the camera overall, particularly when used for video. The Z9 is good, the Z9II could be better. The Sony A9 Mark III sets the bar with its global shutter, though that comes at a cost we probably don't want to pay. Can Nikon move closer to that bar with a refresh of the image sensor to increase internal bandwidth? It’s about the right time period where such internal speed increases would be expected.
- Faster frame rates. Related to the former, as both require more bandwidth in the underlying electronics. The Z9 is already 8K/60P (though in raw). Can you really get another 2x push inside the sensor without breaking things? And just how much benefit is 30 fps over 20 fps for stills? Both things are worth pursuing, but it's a cost/benefit problem right now as the fruit might not quite be ripe.
- Better dynamic range. For every photon we do convert into an electron, we currently throw away at least one, and more typically closer to two. Increasing sensor efficiency would definitely have direct image quality implications, but nobody’s moving the efficiency bar these days with current sensor technologies, which implies you'd need a completely new sensor design. That's fruit sitting at the very top of the tree, plus it might not be ripe yet.
- Direct storage. Why not let an SSD connect directly to the camera? At the simplest level (image transfer), you could do that today, though we probably need to be concerned about the power implications. I’m not sure that EXPEED is currently set up for anything more than what it's currently doing in pushing to CFe 2.0, so any higher speed use—recording raw video direct to SSD, for example—probably gets us into heavier R&D efforts.
- Better direct communication. I had a video camera that supported a cellular slide-in accessory eight years ago. Today I have a portable video switcher that supports multiple, banded-together cell streams today. What I don’t have is a still camera that truly (directly) supports any form of direct communication (FTP via Ethernet or USB is indirect). I can run a cable to my phone and set up the phone to push to a single FTP point (using NX MobileAir), but that’s slower and more cumbersome than it need be (again, it’s also indirect, not direct). The ultimate, of course, would be to have a cellular option in (or an accessory for) the camera, with all the VoIP type options (including streaming) as well as direct push to anywhere. This fruit has been ripe for some time, but apparently no one wants to climb the ladder to pick it.
The problem with almost all that tougher to pick fruit is that it requires replacing current parts with newer, better performing parts. That is not just added R&D cost, but it also takes time—sensor bandwidth increases on a regular basis, and can’t really be forced to happen faster—as well as significantly more testing to insure that it works as expected.
Also: the more forbidden the fruit you’re reaching for, the more likely you fail to reach it. You can say “we’re going to implement Technology X.” However, as you try to do that you may discover you can’t get it to the point where it’s worth deploying yet. Either the technology isn’t quite there yet, or the cost of it is too high.
That said, a Z9II with both the low-hanging and harder-to-reach fruit in it would be a formidable beast, and further cement that camera’s abilities and reputation.
Finally, we have fruit that are not yet ripe:
- Global shutter. As we saw with the Sony A9 Mark III, pushing all the way to a fully global shutter has a detrimental impact on dynamic range. No one has yet shown that you can do otherwise. To me, this limits the camera’s function, whereas a Z9II needs to again be a do-all, be-all camera.
- Quad pixel focus. Canon will probably get to this first, mostly because they’ve been exploring this longer. But I’ll use a weaker standard here: any dual axis focus capability. Right now, Nikon (as well as Fujifilm and Sony) are single axis with gaps between rows. Nikon has numerous patents in this area, but I’m not sure we’ll see anything in the next generation. The patents feel more like describing unripe fruit than pointing to something about to be picked.
Studying what Nikon’s done and the patents for what they’re likely to do, I assess the above two items to not be ready to pick. I’m guessing that those are at least another generation of camera away.
Yes, I know some of you are screaming about other things you want. Most of those likely would be EXPEED8 types of things, so that also puts them into a future full generation update. If Nikon could add an EXPEED8 bundled with more capability to the Z9II, great, but what I'm looking at now is a Z9 original that's several years old that's still pretty much at the top of class. It really just needs some low-hanging fruit picked to stay at the top of the class, and any higher level fruit is just sweetener. My guess is that EXPEED8 is not in any Z9II, partly because of the cost considerations. Nikon really needs to milk EXPEED7 as much as possible at the moment to recover R&D costs.
I suppose we could consider a 24mp Z9h instead of a Z9II (h was Nikon's designation for high speed camera in the digital era). The reduction in pixel count should allow 40 fps and 1/400 flash sync to be possible with EXPEED7. I'm not sure there's much demand for such a product, though.
Sony moved first (A1 at the start of 2021). Canon moved next with a mid-2021 R3. Nikon moved last with the Z9 (at the end of 2021). Both Canon and Sony have now already iterated, yet here we are with the Z9 holding its own simply via firmware updates. Some low-hanging fruit could be done simply with version 6.0 firmware, or Nikon can make a small number of technical and hardware changes (EVF change, for example) and create a Z9II. I'd be happy with either.
