I shot the Nikon Z8! (but I did not shoot the deputy)

I shot the Nikon Z8!  (but I did not shoot the deputy) by RC Jenkins:

My Z8 finally arrived!  I’m writing this in real-time over a few weeks as I mess with the camera, so I have no idea what I’m going to write.  Weeksauce, bro.  This will be all over the place, and I will probably go back and add asynchronous edits that make no sense in context.  And I’m guessing it will mostly be a comparison to my Z6 that I’ve had since it launched almost 5 years ago.  And everyone likes sprinkles, so I might even sprinkle some Nikon DSLR comparisons in as well.

Here it goes.  Dear diary…

Why?

I have no idea why I bought this camera.  I didn’t actually want it–what I really wanted was a Z6iii.  And when it comes out next year, I’ll probably regret buying this Z8 instead of waiting.  So please, feel sooo bad for me.

I’ve really liked my Z6 since the day I got it back in 2018.  It’s a great camera with versatility that works great for a wide spectrum of use cases.  I’ve personally been all over the place in how I use it.

But there have been some things that I found annoying or limiting.  It’s ok when it comes to following action like sports or wildlife, but could be better.  Autofocus has been good but sometimes less consistent than I’d like–and tracking has been annoying enough to use that I don’t use it.  Video is good but could be better when it comes to formats, detail, and malleability (mainly due to bit depths & bitrates)–and I don’t want an external recorder.  Buttons are OK but I found myself menu or i-menu diving a decent amount.  24MP is…actually that part is great for me.  I don’t actually want 45MP because of lossless file sizes and other speed- and storage- and workflow- related implications.  I find 24MP / 6K to be an excellent blend of resolution, speed, and storage.  But I’m not going to complain about perfectly oversampled 4K-from-8K video resolution either.

So because we know what we have with the Z8, and the Z6iii looks like it’s not coming out until next year (and we have no idea if Nikon will lean more consumer or pro with it), I just bit the bullet.  Nikon won.  They got twice as much of my money (and a year earlier) with the Z8.

Let’s see if I regret it.

(Seriously, I have no idea yet).

Body

The first thing I see is the body.  I immediately notice the size and weight.  It is…not bad at all, coming from a Z6.  It’s slightly heavier?  Maybe?  I can’t really tell, honestly.  By Z6 battery is charged, so as I swap it out, I realize the battery feels heavier than the difference in weight between the two.

I remember when I got my Z6, I thought it was so small compared to my DSLRs; but I don’t feel the Z8 is too big in the opposite way.  I added a base plate / L to my Z6 because I always swap between handheld, tripod, and gimbal; and I sometimes set my camera down on available flat surfaces, so I hate the little square tripod…squares.  And I don’t like my pinky dangling.  I like the height of my Z6 after adding the plate.

Luckily, I also bought a Z8 base plate / L bracket.  And oh, I should take a picture of this so everyone can see.  And ooo the closest camera to me to take a picture of these two is my Nikon 1 V3, which is also great for comparisons.  It also shoots 20FPS and ultimately helped develop the Z8.

The Z8 has a good grip height–actually quite similar between the two now.  The difference in viewfinder height is more major–but as long as it’s not too tall for my bag, I’m not bothered.  I am so glad that the base plate is thinner than my Z6 version; and also that the base plate doesn’t poke at my fingers like the Z6 version.  We can’t see it here, but the camera’s grip is also noticeably deeper (and/or the camera is thicker).  This is good.  This feels good.  Real good.  I like the Z6 height & weight better for travel and packing.  But this feels better when I hold it.

Now, I’m pushing buttons.  A few minor things:

• The rear dial feels cheaper and less clicky.  I like my Z6 better.
• The AF-On button is also softer and doesn’t push in as much.  I like my Z6 better.
• The front buttons are round and small again.  I remember this from my DSLR.  I got used to the large Z6 front rectangular buttons.  Though I would sometimes push them by mistake.  So we’ll see how this goes.
• The play button has replaced the drive mode button; and it’s not recessed enough.  I’m going to have to get used to this.
• One subtle ergonomics trick that Nikon uses is that each button feels different than its neighbors, so you can tell just by feel which one you’re pushing:  One button might be slightly recessed, its neighbor is slightly raised and with a slight plastic ridge around it; one might be concave vs convex; neighboring buttons might both be raised, with a plastic divider between them, etc.  This is one of those things you learn from muscle memory, even without noticing it.  My Z6 initially seems like it does this better than my Z8.
• The front dial clicks nicely, but it feels…more hollow? Maybe?  It’s different.
• I don’t like where they moved the exposure compensation button.  The old spot on the Z6 was much better and less awkward to get to for my index finger.  Also, it’s more recessed on the Z6–again, easier to tell apart from feel.  More than once, by feel, I’ve confused the ISO and EC button now.  Why did they move it?

A reminder: on my DSLRs, the EC button would change EC with the rear dial and metering mode with the front dial, which makes sense since they are related functions.  On my Z6, they inexplicably changed it so both dials do the same thing:  EC.  This is a waste of controls and redundant.  (Nikon:  are you reading this?)

The Z8 carries on this EC button redundancy rather than function.  This essentially means I will never ever use highlight-weighted metering ever again, just like I never did on my Z6.  I’ll just use EC or swap to manual ISO in those situations.  Not a huge deal, but slightly annoying.

As I’m just feeling around the camera body, I keep swapping back to my Z6 to compare.  I notice it’s getting harder and harder to tell which is which.  They both feel familiar.

The Z8 body feels more comfortable.  But the Z6 buttons and dials feel better–and it just feels like it will be more portable.  But the Z8 has more buttons.

I would give the Z8 body a win over the Z6 overall, due to the extra buttons and grip feel.

And now, after using it after a few weeks as my primary camera for a few shoots (alongside a Z6), I like the Z8 body even more in practice.

Durability

Durability is important to me, because I don’t want to worry about my camera.  I tapped the Z8 like a watermelon a few times.  It doesn’t feel like it will break.  Test complete.

It feels just as well built as my Z6, which itself is very well built and has lasted me 5 years so far (and still feels new). Both cameras feel much more solid than my old short-lived D810 or my D750 that lasted 10 years.  My D750 is lighter, but that’s because some of it’s internals became externals and it also decided to slowly take its rubber suit off over the years:

(That was a joke for humor purposes).

My cameras usually get abused.  I will abuse this Z8 too.

In fact, during a shoot a few days ago, I’ve already somehow cracked the screen protector I put over the top LCD (the actual LCD glass & camera are fine).  I thought I’d start the abuse early.  But I’m still not worried about it at all, and there’s no damage to the camera whatsoever.  This camera feels well built.

So in conclusion, the camera is durable enough to last at least 2 weeks.

Menus & Controls

I load up my custom picture controls, which I might need to recalibrate for the Z8 raws–though at first glance, things look fine.  (If that statement doesn’t make sense:  different cameras have different raw characteristics–and picture controls provide a path to converting between the raws and rendering, using 16-bit precision between the two.  Starting with a different raw and taking the same path can sometimes result in a different render).

Then, I do my usual menu options.  The menus and options are much deeper and better than the Z6.  This is important because useability is vital to performance.

First, I need to change to mode M.  Mode button instead of that dial I never use!  Missed ya, buddy.  Or did I?  I don’t really change modes that often.  Occasionally S for the weekend sports or wildlife…but only if it’s really sunny, but otherwise still M and auto-ISO.  A for portraits or products…though not unless they’re moving and again in harsh lighting–otherwise M.  Do I need this dedicated…?  Hmmm…

I think that over time, as technology has evolved, I’ve changed and I’ve learned about what I adjust most on my cameras.  Time to explore buttons.

I’m very big on aligning intuition, dexterity, and frequency of use when I assign buttons.  For example, on my Z6, I used Fn2 for AF mode, because I changed AF all the time and could hold it with my ring finger while easily using 3 other fingers:  my middle finger to change the AF area on the front dial, thumb to change AF mode on the rear dial, and my index finger on the shutter button.  But I couldn’t do this with Fn1 (middle finger), where I’d have to move my shutter finger to do the front dial and then move it back to the shutter button.  So I’d assign Fn1 to settings that I don’t need the front dial for.  On my Z6, Fn1 was set to magnify.

But autofocus already has a button.  White balance already has a button.  Picture controls already has a button.  (I shoot raw but I do like these for previews, starting point for editing, assisting in raw exposures, and for video).  Great start to the buttons.

What do I use my extra buttons for?  Easy: unlike the Z6, the Z8 allows us to assign AF modes + AF-On to some buttons, which means temporarily changing AF mode only while holding that button.  So I’m going to make AF-On into 3D Tracking+AF-On.  Maybe joystick press for AF-On (only), since I can move the AF box to where I want and then just push it to focus, in an easy spot for my thumb already.  And maybe Fn2 for auto-area in a pinch?  Now why do I have a dedicated AF/MF button again?  I really missed it on the Z6, but think I might not need it as much now.  Can I customize it?  Nope.  Though I can now also speed up how fast my AF points move with the joystick.

3D tracking is back and it’s brilliant!  Just like my DSLRs but better.  And much better than subject tracking on the Z6, both in interface and performance.  It’s very responsive and sticks like glue.  Even for fast and erratic pans or subject movement.

Let’s look at more button customizations.  As I said earlier, I don’t know if I need the mode button.  Can I customize it?  Nope.  I’ll rarely–if ever–use bracketing.  But I can’t customize it either.  Because none of the top 4 buttons can be customized.  The trash button is non-functional during shooting, so it’s also a wasted button that should be able to be customized (but isn’t).  And as I said earlier, metering mode should be dual purpose with EC like it was on DSLRs, but I can’t make this happen either.  There are also a few other buttons that have limited customizability, like the OK button; so this is going to be reset AF point to center.  Despite all this, I generally like what I can customize and how little I need to since there are a few extra dedicated buttons for things I would normally assign.  I felt like I had to waste Fn buttons on the Z6.  On the Z8, I don’t feel limited, but being able to reassign every button would be nice to have.  And formatting without obscure menu diving is back!  But every time you use it, you seem to end up back at that menu item when you re-enter the menu…

After I customize my buttons, I try them out for a while and see how I like them.  Adjust, rinse, and repeat until they feel right.  Same for the i-menu–but after buttons.  Usually takes a week or two, with time away in-between.  If I can come back to the camera and just intuitively push the button I want over time, I’m done customizing.

My biggest fear right now is that my Z6 and Z8 will diverge in how I assign buttons and make it hard to switch between the two.  When Nikon releases a Z6iii, I hope they keep some consistency with the Z8 button layout and assignment options.

Image Quality

This is a pointless section.  So far, IQ is really good, like every camera today.  It’s a modern full-frame.  And it’s 45MP, which is far more than the 24MP that I (still) personally find ideal for my own workflow.

Viewfinder & Bursts

Before I even try to use the Z8 in practice, I want to see this whole blackout-free situation.  I’ve heard mixed messages on when it works, but I know it must be impossible slower than 1/250.

Before we get into details, the Z8 has a nice viewfinder!  It’s very bright and clear and big.  By default, it appears to be maybe 30FPS?  However–and this is true for all Z’s and perhaps all mirrorless cameras–the moment you start autofocusing, the sensor and viewfinder clearly refresh faster (60FPS or 120FPS).  This is optimal, and it’s to avoid wasting battery life and build of heat.  And this is how the Z8 works too.

Anyway, let’s do a quick menu dive to see if there are related settings.  There are.  I’d tell you which ones, but I’d like to encourage you to stop being lazy and read the manual.

Blackout test time. click-click-click…

Ohhh.  I see.  Easy.  And it makes sense why people are writing inconsistent things.  Ok so here it goes (and this stuff is not in the manual):

First, let’s talk a bit about the sensor & silent/electronic shutter.

The Z8–like the Z9–uses a stacked sensor that does a full pixel readout at ~1/270 (according to the internet).  This is roughly 10-20x faster than most other mirrorless cameras that don’t use stacked sensors.  Most mirrorless cameras–including the Z6 & Z7–have full readouts somewhere in the range of 1/10 – 1/30.

But 1/270 is also almost as fast as most mechanical rolling shutters that most full-frame cameras have.  And this fact significantly complicates comparisons.  Mechanical shutters need to accelerate and decelerate, meaning their speeds–and consequently exposures–are inconsistent within a frame.  The top and bottom (during acceleration and deceleration) might expose slightly darker, with the horizontal center of the frame exposing more.  And because this involves exposure, this fact also impacts on things like measured dynamic range, shutter lag, rolling shutter effects, etc.  So overall, things are complicated, and some measurements that might seem simple are not as easily or directly comparable.  Most numbers on cameras are simplified estimated averages.

Also remember that during viewfinder display, the sensor does not do a full pixel readout–it doesn’t need to (you don’t need all 45MP to display a 1.2MP / 3.69 dot / 1280×960 viewfinder).  Traditionally, sensors have dropped to a lower bit-depth like 12-bit or 10-bit (if available), and they also line-skip and/or pixel bin, which allows for a much more rapid readout and processing and also saves battery life.  For reference, this is how the Z7–which has a much slower full pixel readout closer to 1/15ish–can refresh its viewfinder at 60FPS during autofocus:  the sensor is only scanning a fraction of the data.

And aside from the viewfinder, when a sensor today is in the middle of exposure, it cannot actually display an image until after the exposure is complete:  each pixel has to finish measuring the light before it reports back.  And this means it cannot display a frame in the viewfinder while it is exposing–only after it’s done.

So after all that technical context, how does “blackout-free” work??

It’s a very clever trick by Nikon:  the camera consistently and rapidly switches sensor readout modes for the viewfinder in-between the pictures you are taking–and it sends each to different memory and processing streams so that they aren’t slowed down.  Nikon’s dual stream controller–which is almost certainly outside of the sensor–is like 2 commanders, alternately telling the sensor to take a shot.  It is so rapid that you don’t notice.  Since the stacked sensor is fast enough to switch modes rapidly and get ready for the next shot near instantaneously, it will take a 1/240 (or faster) line-skipped exposure just for the EVF, 120 times per second.  And it also has another 1/240 left over in-between frames (since 1/120 – 1/240 = 1/240).  Because the stacked sensor can also do a very fast full pixel readout in less than 1/240, the sensor has time to take a separate full image exposure in-between the EVF shots before the EVF needs the sensor to expose for the next frame.  And notably, the camera can also concurrently perform an autofocus calculation immediately after any of the frames, including the 120FPS frames used for the EVF.  In other words, the blackout/exposure is so small, that it’s in-between the regular viewfinder frames.

Conceptually, the blackout-free sensor reads during a single shot would look something like this:

Those black frames in between in the top EVF stream don’t get displayed in the viewfinder.  And as a note:  the viewfinder frames theoretically don’t even have to be at 120FPS.  They could even be at 60FPS or 30FPS, as long as the full exposure has time in-between.

Technically 100% blackout.  Practically 0% blackout.

Technically, you never see the moment of the shot that you actually took.  You just see the top row.  It displays the moment 1/250 (4ms) before and 1/250 (4ms) after; and you see these moments 120 times per second, consistently.  For perspective, the blink of a human eye lasts about 100ms; and human reaction time is around 200ms.  All this is why it appears blackout-free.  And in this case, appearance is what matters.  Because appearance is the entire point of a viewfinder.

So that is the “blackout-free” mode; and it currently only works at shutter speeds of 1/250 or faster.

For the next section, I’m assuming setting D12: View All in Continuous Mode is set to On.  This setting only seems to apply between (as low as) 1/8 and 1/200.  But please, read the manual.

Using a slower shutter speed also seems almost blackout free, but more stuttered?  What’s happening?  Simple.  Between shutter speeds of 1/8 and 1/200 the camera does a hybrid of the above–it switches between showing both the top and bottom rows.  Because now that the Image Capture exposure takes longer than half of the 120FPS, the camera cannot refresh EVF in-between every shot.  But also, since it’s only shooting those Image Capture exposures at 20FPS, it can do some EVF shots at 120FPS.  And as a note:  since the image capture frames will also show but take longer, they’ll be delayed.  So you are seeing the EVF frames, then the shot you took, then EVF frames, then the shot you took, then EVF frames, etc.  This introduces a slight lag and stutter when you see the actual frames, but this effect is not really prominent until slower shutter speeds.  This mode is appears to be how cameras like the Sony A1 or Canon R5 shoot bursts.

Once your shutter speed gets below 1/20, you can obviously no longer get 20FPS (and there is no time for any in-between viewfinder frames).  So by this point, the camera burst rate slows down and simply displays only the frames you took after you take them, in a pure slideshow mode (just like the Z6 & Z7).  So this 1/8 to 1/200 range is a spectrum:  at 1/8 you get a pure slideshow; and at 1/200, you get almost blackout free at 120FPS, interrupted intermittently by a slideshow at 20FPS.

I think in theory, Nikon could have kept blackout-free at 30FPS or 60FPS viewfinder.  My guess is they didn’t either due to complications in designing the dual stream controller for consistency and low shutter latency or for tracking & autofocus performance.

This “slideshow” mode is how the Z6 & Z7 (ii’s) do it.  Except, on the Z8, this is an exceptionally fast slideshow at worst.  It feels more like my Nikon 1 V3–shooting super fast with a fast slideshow in the viewfinder and little electronic shutter sounds at 20FPS.  Meanwhile, my Z6, displays a more delayed slideshow, and at around half the speed, with mechanical shutter sounds.  It appears that not only is the slideshow on the Z6 slower, but the processor might take a moment before it displays, and the sensor might take a moment to switch readout modes.

Earlier, I said setting D12 should be on.  If it’s off, instead of the frame you just took, you’ll see the viewfinder but with a blackout frame at 20FPS for each shot you take, roughly equal in duration to your shutter speed.  Which brings us to the next section:

Once you get to 1/6 or slower (configurable), the camera completely blacks out the viewfinder during an exposure with no slideshow and then shows a few EVF frames in-between shots.  Just like if you had set D12 to off, it no longer shows each shot you just took.

Oh, and then there’s also starlight view, which essentially lengthens the duration of the EVF shutter speed and slows down the refresh rate, so that you can use the viewfinder better in pitch-black conditions (and I haven’t tested extensively, but could even use a different readout mode).

So to summarize:

• 1/250 or faster:  the viewfinder appears completely blackout-free.  Better than a DSLR.

• 1/8 – 1/200:  if D12 is on, the viewfinder is partial blackout-free, partial slideshow.  The degree of which depends on shutter speed & burst rate.  Just like the fastest mirrorless cameras.

• 1/6 or slower (or 1/200 & slower if D12 is off): the viewfinder is completely black during exposure, and the EVF shows a live view in-between shots.  Like a DSLR, but instantaneous for less blackout.

This is a very clever idea by Nikon.  This is the best viewfinder experience I’ve ever had on any camera, including both mirrorless and DSLRs.

I’ve now tried this a few times, including at a soccer (/football) match and a few other athletic events.  Hands down much better than my Z6 and my DSLRs (that I said I’d still be using but haven’t seriously used in years).  It’s easier to follow the action through the viewfinder, quickly choose what to focus on, have the camera focus perfectly, and take faster bursts for longer.  It’s also so intuitive–I immediately stopped remembering that my Z6 has blackout/slideshows until I started using it again.

I wanted to spend a lot of time on this topic so that people can understand how to optimize it for their experience.  And also so we can think about the possibilities and expectations about how and when this could come to other cameras.

Buffer & card slots (…and the SD controversy…)

I’m adding in this section after I’ve observed complaints about the buffer and SD card.  And what I’ve observed in these complaints is that they often seem to come from a place of not thinking; but rather, perhaps from ease of complaining after drawing extrapolated conclusions from watching limited Youtube test videos without having done adequate exploration on options for solutions in practice.  And probably a degree of not understanding basics in how buffers work.  But I think there’s more to the story.  Sometimes, the solution is up here and not over there.  (I just pointed to my brain and then to a YouTube tab, but you can’t see it because this is a written article and not a YouTube video).

I’ve been on the fence with whether I prefer the 2nd slot being SD; but I don’t mind it.  While the SD is significantly slower and a 2nd type of card to buy, it is also much more universal and more readily available essentially everywhere.  This is important to me since I travel so much and sometimes (for many reasons), I might need to buy or borrow another card in a pinch while out and about.  But buffer for continuous shooting is also important.  So how do we balance these factors on the Z8?

First, let’s briefly walk through what and why a buffer is.  A buffer is just fast temporary memory–when you take a picture, the picture gets temporarily stored in the camera’s internal buffer memory (which is very fast and reliable), and then the camera writes that picture to your card (which is slower than the buffer memory).  This buffer allows the camera to shoot continuously for longer–while you are shooting a burst, the buffer images are also being offloaded onto the memory card.

Since these are directly related to buffer performance, let’s look at some file size estimates:

• ~60MB = raw, lossless compressed
• ~35MB = raw, high efficiency*
• ~24MB = raw, high efficiency
• ~35MB = JPEG, fine*
• ~9MB = JPEG, basic*

(Note that there are also other options, including an even smaller basic (no star) and Normals, which are halfway between the basic and fines).

That’s quite a range.  And let’s also throw up a handful of card rating speeds:

• 1725MB/s = Delkin Black CFExpress
• 300MB/s = Delkin Black SD

NikonRumors has articles on speed tests that includes more options.

So here are a few observations from these data points:

• CFExpress cards are rated ~5x faster than the SD cards
• The smallest sizes here (JPEG basic*) are ~7x smaller than the highest (lossless raw)
  ^(These two should provide a clue of where we are going with this…)
• High-efficiency raws are roughly half as large as lossless raws
• JPEG fine* are roughly the same size as high-efficiency* raw
• High efficiency raws are smaller than JPEG fine*
• JPEG basic* are ~4x smaller than high-efficiency* raw
• JPEG basic* are ~4x smaller than JPEG fine*
• If the Z8 has ~1.5GB of buffer memory, this is ~25 lossless raws or ~167 JPEG basic* (but of course, these are constantly being written to cards at the same time, which is why the buffer is almost always larger than these figures)

Multiply each file size you use by 20 (FPS) to estimate the rate at which the buffer will fill (eg. lossless + JPEG-fine* is ~2000MB/s).  Look up your card to estimate the rate at which the buffer will empty.  And I haven’t done any tests, but I’d assume the Z8 has somewhere around a 1.5-2GB buffer, give or take.  The continuous shooting duration or number of continuous burst frames can be calculated by using the size of the buffer, minus input rate (FPS * image sizes in MB/s), plus the output/card speed (MB/s).  But this is complicated and not 100% consistent, so tests are usually better.

If set to backup, the Z8 will write both versions of the image to each card and wait for both to be written before offloading the next image.  The issue is that when writing images to cards (eg. raw to CFE, JPEG to SD), the buffer will be limited by the slower ratio of card speed to size, reducing the burst speed after the buffer has filled.  Frequently, the SD card could become a bottleneck.  Note “could”… (…because this is foreshadowing).

There are also numerous buffer tests out there; but they always seem to test JPEG fine* for some reason.  And I’m going to challenge this practice.  Because years ago, I decided to explore other JPEG options (to improve buffer and file space); and back then, I came to the conclusion that JPEG basic* was fine for backup images during bursts, which is how they are often used.  Let’s see if that works for the Z8.

Because on the Z8, a JPEG basic* is still 45MP, and it’s still a 9MB JPEG, which is quite a large image file.  And we are exploring this because we might be able to improve the buffer by more than 4x relative to JPEG-fine* when writing to an SD card if JPEG-basic* is sufficient.  Or another way to think about this:  the ratio of CFE to lossless raw is ~29; and the ratio of SD to JPEG basic* is ~33; so in theory, if shooting lossless raw to CFExpress and JPEG-basic* to SD card, the CFExpress card could sometimes be the bottleneck.  And if you want to do even better, take advantage of the camera’s technology:  high-efficiency raw to the CFE and JPEG-basic* to the SD card, which will also allow each set of shots to take less space in the camera’s buffer.

But JPEG-basic*?  Eww!  Disgusting, right?  Let’s just take a quick look at one of these from the Z8 (or rather, a screenshot of a 45MP, unedited JPEG basic* from the Z8):

I don’t see any issues with JPEG compression at this size.  Shall we zoom in?

Still looks “fine” to me.  Let’s go further:

And now, I can start to see some compression artifacts in the finest details of the hair and in the texture of the shirt.  But honestly, it still looks fine even at this level.

But this is a JPEG.  What about tonality and malleability?  Let’s try tonal compression on this JPEG.

Oh.  Seems that the JPEG basic* still has lots of tonality baked in too–likely since this is 9MB and also has 45MP.  So now we are pixel peeping and doing extreme HDRs just to see if a JPEG basic* provides terrible, unusable quality from a Z8.  And when we actually explore the images, it turns out that it doesn’t.

So now let’s take a step back again, away from the camera specs, and away from the limited YouTube tests.  Let’s start using our brains, just for a moment.

Why are we shooting double?  Why JPEGs to the SD card and raw to the CFExpress again?  The purpose in this scenario is for a backup, in case the CFExpress card fails.  Maybe it gets lost, or physically damaged, or corrupted with irrecoverable data loss.  Think about the small probability of that happening to your CFExpress card, and where we would be forced to use our JPEGs instead.  Because almost always, we are going to throw these JPEGs away–they’ll never see the light of a monitor.  My out-of-camera JPEGs usually just take up hard drive space until I go back and delete them.

And after seeing the above examples of both the detail and tonality in JPEG basic* from the Z8, think about what would happen if *gasp* you were forced to use JPEG basic* for that rare occasion of CFExpress failure.  Would you die?  Would your clients die?  Would you get sued because your clients pixel-peeped your 45MP HDR wedding portrait?  Would you be forced to quit photography forever, shunned by society for reducing the pixel-peeping quality of those precious memories?  Would the memories fade even away?

No.  Everyone will be fine.  The pictures will be fine too.  Basic is fine.  Be fine.  Be basic.  Use JPEG basic* for the SD card if you need both a backup and lots of buffer when you shoot 20FPS continuous bursts.

Or better yet, make some time to assess your priorities and explore the options and test them out; and then pick the one that’s best for your circumstances.  Is using JPEG-fine* for a backup that you’ll throw away more important than getting a shot during a burst in the first place?  Would JPEG-basic* or JPEG-normal* work better?  Should you even be taking hundreds of shots at 20FPS during a prolonged burst for this circumstance?  Should you perhaps consider JPEG-basic* as backup for some types of action bursts, while you use raw backups or for slower shooting backups such as landscapes or some portraits?  Could this be one reason that settings banks and the i-menu exists on these cameras?

This is what settings options are for.  This is what skills and experience are for.  This is what brains are for.  Use them.

We don’t need to blindly follow someone whining on the internet about the Z8 buffer and SD cards if we can be efficient to do what we need to do.  Oh, and speaking of blindly following some random person on the internet, I forgot to actually give you numbers so that you can blindly follow what I write.  And why don’t we have fun and go closer to worst case…?

Using an ancient Sandisk Extreme Pro 95MB/s UHS-I V30 SD card that I was surprised I still had lying around and 20FPS on a single burst, I was able to get:

• JPEG-basic* ~160 shots (~8 seconds).  And the buffer refilled almost immediately.
• JPEG-fine*: ~40 shots (~2 seconds), with ~8 seconds to refill the buffer.
• Lossless raw:  ~30 shots (~1.5 seconds).
• Lossless raw + JPEG-fine*:  ~25 shots (~1 second), with ~30 seconds to refill the buffer

^ Take another look at those card specs.  I can buy a 64GB version of that card (new) for $30 today; and a used one for $12.  That’s a very slow card by today’s standards.  And despite its already adequate performance, I would expect significantly better buffer performance on a modern UHS-II V90 SD card.  I wouldn’t even be surprised if I were able to achieve near unlimited buffer for practical purposes.  And so I’ve got a newer Delkin Black SD card on the way, and I can repeat the test & update in the comments in case anyone is interested.

But suffice to say:  I don’t plan on ever having the buffer–or the SD card–be a problem for me in practice.

Autofocus

This is an important subject, but it’s been beaten to death, and it’s so tedious and difficult to quantify and compare properly that most don’t and just give a subjective opinion or feeling.  Here’s mine:  autofocus works really well.  And the button customization makes it even better.  In my brief time with the camera so far, I’ve not had any problems with the autofocus.  It just works, and it just works far better than any other camera I’ve ever used.  In both stills and in video.

There are the usual weaknesses that almost all mirrorless cameras have today.  Horizontal lines (eg. focusing on a horizon) is a (rare) weakness for landscapes–some simple remedies are slightly tilting the camera so that the PDAF system can actually compare parallax; or using the zoom function to force the camera into CDAF.  Subject recognition (though less, autofocus) in pitch black environments can be a weakness, as it is on all cameras.  Those are the only weaknesses I’ve come across.  In all other cases, the AF is just brilliant.

And just to qualify this:  I did a few shoots with the Z8 already in some of the worst conditions I could have imagined–moving subjects in the dark where the only ambient lighting was bright movies being projected onto both the subject and the background.  This shoot was mainly for video–I was doing a lot of running around with a gimbal, but I took a few stills too,mainly just to test out the camera.  The camera sometimes struggled a bit with subject identification (though autofocus was usually fine) from around ISO 12800 through Hi2 (ISO 102400), all while at F/1.8.  My remedy when this happened–inconsistently–was either Fn2 (which I’ve assigned as “emergency” auto-area autofocus) or AF-On / 3D tracking to put the box on the subject and recompose–this instantly got me perfectly focused shots.  To get a sense of what I mean by these conditions, here is one shot at ISO 12800, at F/1.8 and 1/50.  The camera subject detection worked in circumstances like this, but it was more inconsistent than usual, sometimes working, sometimes not.

(And the lighting and projections were constantly changing–just a moment earlier, the scene looked like this):

And here are some video screenshots, also using very high ISOs.  (It was darker than it appears here–in person, the walls and floor looked almost black):

These were challenging conditions, to say the least.  But didn’t appear to be much of a challenge for the Z8.  (And the Z6 AF did better than I expected, but not as good as the Z8).

Video & Audio

I do quite a bit of video and always prefer a small setup, and this was a big factor for me choosing the Z6 and now the Z8.  And why I’d have preferred a Z6iii.  There are a few things I don’t particularly like on the Z8.  Some of these are minor, and some are pretty major for me:

• No native DCI resolutions–they are all UHD.  Is 4096×2160 too much to ask for?
• No true 24FPS.  23.976FPS only
• No BRAW (for DaVinci Resolve).  Minor.
• Slight VF lag during oversampled video, but much better than the Z6.  (This is because the camera is actually reading and processing all of the pixels during video, unlike during stills mode where the EVF is line-skipping).  Minor.
• No open gate to support 2x anamorphic lenses.  Minor.
• Apparently the same audio preamps, which I’m not a fan of on the Z6.  I’ll have to do some tests on the Z8.
• The list of video formats in the menu can get to be long because it shows combinations of resolution and frame rates.  This may even limit our available options.  For example, there are 4 items for 1080 (HD), each with a different frame rate.  Better would be one submenu item for resolution and a separate submenu item for framerate, which would allow for easier navigation through more combinations.  I’d like to see something more like (and note that not all of these exist on the Z8, but they could ideally be added via firmware):
  • Resolution:
    • 8256 x 4644
    • 8192 x 4320 (8K DCI)
    • 7680 x 4320 (8K UHD)
    • 4096 x 2160 (4K DCI)
    • 3840 x 2160 (4K UHD)
    • 1920 x 1080 (HD)
  • Frame Rate:
    • 120 FPS (subsampled)
    • 100 FPS (subsampled)
    • 59.94 FPS
    • 59.94 FPS (subsampled)
    • 50 FPS
    • 50 FPS (subsampled)
    • 48 FPS
    • 29.97 FPS
    • 25 FPS
    • 24 FPS
    • 23.976 FPS

That above list would be up to 66 menu items using Nikon’s current menu system.  So Nikon:  as you get more serious about video, please split these and give us more resolutions and framerates.  Like how other manufacturers do.

There are also a few things I really like about video on the Z8:

• Good choice of internal codecs, including 10-bit Prores 422 HQ
• 24-bit, 48KHz linear PCM audio
• High bit-depths
• 4K oversampled from 8K up through 60FPS.  This means each output pixel in 4K has full color information as well, unlike a native 4K sensor.
• Good video-oriented frame-lines, including for 2.4:1.  Yes, for cinema people, the camera will draw a 2.4:1 box to help you frame while you’re shooting 16:9.
• Very customizable viewfinder/display, including removing most icons and info.  Lets you focus on the framing and feels like you’re watching a movie rather than a viewfinder.
• Live USB-C power!  Not just charging when the camera is off.
• Good heat management (so far).  (I’ve written about this in the past, but this was expected: some compare to the Canon R5, but that uses dual pixels, which significantly increases heat generated).  I have yet to have the camera overheat, and I’ve done hours of video, including outside in heat.
• (minor):  Full sized HDMI for external connections.  Makes it easy to use a standard cable connect to a television to immediately see footage, instead of that one cable you’ve “misplaced.”

Then, there are things that I think I’m supposed to like about video on the Z8 but really don’t care about:

• Ability to digitally zoom 2x, and still output 4K.  Maybe I’ll use this one day?
• Raw video.  It’s a cool spec, but I don’t plan on ever using it.  But I like that Nikon offers it, primarily because I hate Red’s business practices.

I think if Nikon could specifically add support for 4K DCI and true 24FPS in ProRes HQ 422, it would enable more use cases for anyone with a cinema-type application.  The closest thing currently is 4K UHD at 23.976 in ProRes HQ 422–close but not quite there.  The camera is technically capable of this–there’s no technical reason it couldn’t be deployed via firmware.

The internal video is clearly better than on my Z6, which is already good.  Due to the combination of codecs, bit depths, and oversampling methods, video quality (detail and tonality) is much improved, even just comparing 4K.  This makes less of a difference if one is producing video that will be recompressed to be streamed online through something like Youtube, so it’s overkill for that (unless one is doing heavy grading during editing).  On the Z6, I was primarily using picture controls to have the camera do most of the grading for me.

In case you’re wondering, I’ll personally usually be using oversampled 4K in Prores HQ 422 (10-bit), which looks stunning.  And I’m doing tests on using NLog vs my custom Picture Controls.

In the context of other future cameras, I think Nikon could theoretically offer most of the above–except 8K–in a Z6iii, as long as they choose a sensor that is fast enough and leverage the Expeed 7 found in the Z9 & Z8.  I think a practical solution would be a 6K / 24MP sensor that can do 12-bit scans at 60FPS–and the more the better.  Everything else–other than rolling shutter–is a processor limit.  And the current Z6 sensor is not too far off of this.  Using a slower or higher resolution sensor would likely limit the video quality, probably through cropping, line skipping, or pixel binning and would be a downgrade relative to the current Z6’s, so I hope Nikon doesn’t follow Sony in a pointless megapixel race for a Z6-series camera.

Summary & Impact

I hate the Z8.  Because it is so good that I love it.  So far, it already feels like the best and most capable camera I’ve ever owned.  It feels like it can do anything.  Other than size–of lossless compressed raws, Prores HQ 422 video, and slightly larger physical size–I cannot see anything that it doesn’t do far better–or at minimum just as good–as other cameras I’ve owned.  There’s not much more to say than that.

I hope we also get some firmware updates for DCI resolutions and true 24FPS framerates.  And I hope Nikon keeps up with firmware updates for quite some time.  I feel like there are quite a few easy features, capabilities, and refinements that are not bottlenecked by the hardware (like those video modes, menu updates, and button customizations).  But even without these, it’s a brilliant offering from Nikon.

It’s hard to imagine Nikon–or anyone–making a “better” camera than this.  It will happen eventually…it’s just hard to imagine.  For example, it’s hard for me to imagine everyone owning wall-sized 8K TV’s too.  But for now and the foreseeable future, I cannot see being limited by the Z8.  It’s an absolute technological leap over the Z6 for many of my circumstances.  And I’m really interested and excited to see how Nikon leverages some of the Z9 & Z8 in their future cameras.

– RC Jenkins, a formally educated, world-renowned expert in all subjects & internet user.

If you have an interesting idea for a guest post, contact me here.

Nikon Z8 pre-order links: