A user-experience centered reviewNikon Z8 Real-World Usability ReviewPart 1: Photography

Based on firmware v1.01
Updated: September 2023

(If you’re looking for Z8 review for wedding videography, see part 2 of this review.)

Welcome to a Nikon Z8 review with a spin: we’ll spend little time on technical specifications and pixel peeping, instead focusing on usability of the camera in the field – with emphasis on wedding photography and videography. (Which is just another way of saying we’re looking for a camera that does everything equally well!)

While it’s good to start with the technical stuff to see if the camera will fit into with your requirements, too often reviews don’t focus enough on how it is to actually use the camera. What good is the cutting-edge implementation of function X if it’s buried deep in the menus and hence inaccessible?

I’m obsessed with ergonomics and user-experience, which is why I’ll focus on what it’s like to actually shoot with the Nikon Z8 for prolonged periods in demanding, stressful conditions (looking at you, wedding photography and videography).

A short intro to my background if you haven’t read my other reviews: I shoot candid, documentary style of weddings which means no re-dos and requires split-second reactions. I’m shooting wide(ish) prime lenses which means a lot of moving around and doing so at wide open apertures, requiring quick and reliable AF.

I spent years shooting with likes of Canon 5D Mk3 and R5, Fuji X-T and X-Pro series, Nikon D850 and Z6, so I’ll be drawing from this experience and comparing what it’s like to shoot Z8 compared to all of these.

All photos in this review were taken with Nikon Z8 and edited in Lightroom.

Intro

Okay, so in case you’ve missed all the big announcements, here’s the quick(ish) rundown.

Z8 is the second camera in the new sub-class of professional mirrorless cameras that we could now comfortably start calling ‘shutterless’.

‘Mirrorless’ category was created by dropping the mirror from the camera body, thereby removing one big and loud moving part, allowing the back lens element to squeeze in closer to the sensor, making everything (theoretically) smaller and lighter.

Nikon has now released a second camera (after Z9) that removed the shutter mechanism completely! Advantages should be obvious to any experienced photographer who ever had to replace the shutter on a camera, or a wedding photographer who was told at the venue their cameras shouldn’t make any noises.

Basically, only part of camera where you now have moving parts is in the AF mechanism and iris of the lens. Everything related to exposing the sensor to light is now fixed, meaning you don’t cause any wear and tear by snapping a photo, and also that it’s completely silent (unless you want the camera to play a sound each time you take a photo).

Now might be a good time to quickly discuss the implications of this, and how we got here. (Click here to skip the explanation of how & why of electronic shutters and continue with the review below.)

The ‘why’ of mechanical shutters

Many mirrorless cameras have ‘silent’ modes which use the electronic shutter instead of the mechanical one, so it’s completely silent when you take a photo. With this you apparently have the same situation as with Z8, but not quite.

Without going too technical, the issue with electronic shutters in current crop of mirrorless cameras is that it introduces some compromises in the photo you’ve just taken.

Basically, shutter is used to create darkness over sensor, allowing the camera read the photo off it after it’s been captured – without light interfering with it. The whole sensor can’t (yet) be read out neither simultaneously nor instantaneously, which is why a physical dark curtain helps stop unwanted light changing your photo as it’s being read and converted into a digital image.

As sensor readout times have improved, manufacturers started offering these electronic-shutter or ‘silent’ modes, which capture a photo without ever being in complete darkness. Resetting (emptying) the sensor pixel buckets before the exposure is simple enough, but reading them sufficiently quickly is still tricky. It’s a slowish process, so the sensor isn’t actually exposed to light all at the same time because it needs to be read in segments (or lines, more precisely).

So, exposing and reading the sensor line-by-line is reminiscent of mechanical shutters at higher shutter speeds (think 1/300s or faster), where due to limitations of allowable mechanical stress to the shutter, sensor is at no time exposed to the light in full.  (You can actually tell the point where this starts to happen by your camera’s declared flash sync speed, usually around 1/200s-1/250s for pro cameras. More on this later.)

Instead, the first and second shutter curtain travel over the sensor simultaneously, with a small gap or slit between them which lets the light through.

This could be called a rolling shutter (as it rolls over the sensor), for reasons that are now hopefully obvious. (This name should remind you of something we actually dislike in videography, which is the jello-effect or straight lines becoming slanted or wobbly when shooting video, due to an effect now called rolling shutter.)

The consequence of all of this is that the whole sensor at no point in time sees the same whole picture in front of it. Top of the sensor sees image at one point in time, middle of the sensor sees image at a slightly later point in time, and the bottom sees the latest version of the image. Now, at shutter speeds of 1/8000s or similar, this still practically freezes whatever is going on in front of your lens because most of things in this world aren’t that fast, so noticing the rolling mechanical shutter slit in your photos is practically impossible.

However, remove that mechanical shutter, as you do for shooting video or in case of electronic-only shutters in mirrorless cameras, and you’re left with just the electronic simulation of mechanical shutter, which is nothing more than resetting the sensor pixels line by line, allowing a bit of time to expose it to light and then reading them off the sensor again line-by-line.

This is effectively equivalent to the mechanical shutter with one important caveat: this process is much slower than the actual mechanical shutter travelling at breakneck speeds (e.g. 1/8000s is 0,000125 seconds).

And when I say much slower, I really mean much slower. In cameras like Canon R5 and R6 it’s estimated between 1/30s and 1/60s, depending on exact mode.

Now, if I did a good enough job explaining the role of shutters in cameras, you’re beginning to see the problem. If the whole sensor doesn’t see the whole image at the same time, and doesn’t have the luxury of darkness the mechanical shutter creates at the end of each exposure, reading out sensor this slowly will begin to create issues for fast moving subjects (or moving photographers). The images will become visibly distorted, e.g., a vertical line will become a slanted, almost diagonal line.

This can create all sorts of problems, because fast moving object aren’t the only thing that will create undesired artifacts. Flickering lights will also cause issues usually much worse than moving objects as they flicker at higher frequencies, from 50 Hz to much more, depending on your luck. This can cause ugly lines over your image, banding and similar.

You will also have issues when shooting with flash, as flash tends to last a very short time (1/1000s – 1/15.000s ballpark).

In order to make electronic shutter mode practical, most manufacturers try to increase the sensor readout speed by lowering the image quality somehow. So, while they use the full quality when using the mechanical shutter, switching to electronic shutter will often reduce bit-depth of captured photo to help alleviate the described problems by making readout faster.

This isn’t always acceptable. For capturing contrasty scenes you’ll want your full dynamic range, only available when the sensor is quantized at full bit depth (usually 14-bit), and for shooting under flickering light sources or when using flash you’ll still need the mechanical shutter, which is why they still linger on in most cameras even though cameras would be cheaper, lighter and more durable if they could be removed altogether.

And this winding story finally brings us to the Nikon Z9 and Z8, as the first professional cameras that have dropped the mechanical shutters completely.

What allowed Nikon to do this is using a Sony sensor built upon a new architecture. You have heard of ancient CCDs, replaced by CMOS sensors, then upgraded to BSI-CMOS (backside illuminated). This is more or less current generation of sensors, with Canon still using the older CMOS type of their own design, and many others (often based on Sony sensors) having switched to BSI technology which, in short, offers better light-gathering capabilities.

None of this addressed the long-standing issue of slow readout, which is why stacked-CMOS has been developed.

Without going too deep, they add more electronics directly to the sensor, stacking it with individual pixels, so when you see a stacked-CMOS think ‘faster readout’. Simple as that.

So, while Canon R5 will read out its huge 45MP sensor at around 1/60s, Nikon Z8/Z9 will be able to deliver speeds at around 1/280s, an almost fivefold increase for the same megapixel count!

This means that any rolling shutter problems will be five times less visible. Nikon thought this was enough to drop the mechanical shutter completely, as you’ll notice it approaches the magical and universally accepted limit for flash-sync speeds (which was usually around 1/200-1/250s), allowing the camera to shoot with flash, reasonably shoot under flickering light and stopping the motion at hopefully sufficient 1/280s.

We’ve dropped it, now what?

Let’s start here with noting that Nikon has done this without dropping the quality of the sensor readout, so you still get full 14-bit color depth when shooting raw, which is a pretty cool achievement. This allowed them to create a robust system around this new approach, first of which was implemented in Z9. No small feat, as this class of cameras is on the cutting-edge of pretty much everything and their users forgive no mistakes.

Quick sensor readout is a godsend for more than a few things, not just silent exposures and mechanical durability.

If you know what sensor is looking at during each fraction of a second, you can do stuff with it. Like, display it in the viewfinder at high refresh speeds, getting very close to optical viewfinder experience.

You can also analyze the photo many more times each second, which means things like much more capable autofocus and object/face detection and tracking.

Since you’re reading off the sensor already, you can also write that image to card many more times each second, offering huge burst speeds.

It also allows you to capture video at high resolution and framerates, like 4K120 and 8K60 in raw!

You can also do pre-capture: temporarily saving images in a buffer and writing them to card if user presses the shutter button, thereby capturing photos in the past, before the photographer reacted!

One of my favorites? It allows for no-blackout image capture. It’s a beautiful and paradigm-shifting experience, as there are no surprises with moving objects – even while shooting photos you see what’s happening, who’s moving where and all that with full metering and autofocus.

Finally, let’s repeat it – camera is completely silent when capturing a photo, which is instantly addictive. Once you try it, you don’t want to go back.

Cat ate Z8’s tongue?

Okay, I promised a user-experience centered review of Z8 in wedding context, so it’s time to start talking about what it’s like to use.

Let’s deal with the first question I had after having seen the camera announcement in May 2023 – how will I know whether or not I captured an image if there is no blackout, if I have constant live-feed in the viewfinder, no shutter sound and presumably no image review to spoil all that smoothness?

A good question.

You won’t.

It’s odd, but first time I touched the Z8 shutter button I captured 5 photos because I just wasn’t sure when the camera captured the image. But not to fret, there are tools to help you.

You can choose between several options of a visual and audible indicators.

First is the quick flash of white lines around the inside borders of the viewfinder image, indicating a released shutter. I didn’t notice it in the first instant, but once you know what to look for it’s fairly obvious and very useful. (You can choose whether it’s shown all around the picture frame, or just on the sides.)

Second, you can tell the camera to simulate a viewfinder blackout each time a photo is triggered (you can choose how long it lasts). The viewfinder blackout experience is universally familiar and recognizable to all photographers and won’t require any getting used to (like the flash of white around the frame from first case). Nikon also reminds us that this can aid in panning shots.

My personal preference was to get used to white lines inside the picture frame, because it was a shame not to take advantage of a non-blackout viewfinder now that we finally have an opportunity to do so. I much prefer being able to keep an eye on developments in front of my lens at all times, even during exposure.

Third, you can set camera to play a shutter sound, choosing pitch and volume. I really liked this, because I could set a sound so soft that only I can hear it. My clients can’t hear the camera, so they’re not bothered or distracted by constant clicking, nor they know when I captured what.

I feel all this will turn out to be game changing and I’ve been wishing for a silent camera for years now (ignoring you for the moment, Leica and Fuji X100, with your quiet and durable leaf shutters).

With candid approach to photographing anything, subjects not getting distracted and easily forgetting about your presence is a huge win. Not to mention you won’t spoil sound recording of a ceremony when working along a videographer (or distract the priest/officiant for that matter).

Body and build

The Z8 is the mirrorless monster flagship of Z9 squeezed into the D850 body.

If this sounds promising, you’re right, but if you’re expecting the lightness of a Z6/Z7… don’t.

It’s heavy, it’s substantial, it’s robust and on the large side for a mirrorless. It’s weather sealed and has plenty of buttons in a familiar layout.

The LCD is a tilting type, but with two axes of movement: you can tilt it up and down both in landscape and portrait orientation. If you read my Canon R5 review, you will know that in my personal use-case scenario this is very much preferable to a flip-out LCD. It’s quick to tilt or click back into place, and it also doesn’t stick out, creating a wider camera footprint and a fragile pivot point just waiting to be snapped in hectic excitement of a wedding day shoot. (More on this in the videography-centered portion of the review.)

The rest of the camera body is close enough to D850 and Z9 that I don’t feel I need to get into more details.

I’ll just mention that in my preliminary testing I got about 900 photos on a single charge of EN-EL15c battery, using a mix of EVF and LCD with increased brightness (around +4).

Let’s move on to the more interesting stuff.

Viewfinder (EVF)

While the viewfinder isn’t on the cutting edge of resolution (actually featuring the same pixel count found on Z6/Z7), once you activate high refresh rate, the viewfinder will immerse you in a lag-free shooting experience resembling the optical viewfinders of DSLRs.

If increased battery consumption isn’t a worry for you, you will absolutely love the high-fps display!

There are a few other nice features.

EVF information overlay rotates when you hold the camera vertically, making it more readable and rearranging the info in a more sensible way.

Like R5, you can pick what is displayed on which info screen, which I loved.

You can also reduce the active area of the EVF, making it easier to see the whole display if you can’t put your eye all the way to the viewport (like when wearing glasses).

The ‘DSLR simulation’ mode is still there, allowing DSLR-like behavior of the EVF: it’s only showing the live feed from sensor, and menus and photos for review are shown on the LCD. You can also use the eye sensor to activate other combinations of EVF and LCD and turn off ones you don’t want.

Customizability

When I reviewed Canon R5, I said it was the best mirrorless user-experience for a wedding photographer I’ve had so far. Canon really thought through most of the ergonomics and made a camera that’s customizable and enjoyable, thereby a good companion for long stressful days shooting weddings.

Well, Nikon Z8 gives it a run for its money.

Before I continue, a disclaimer: This review aims to be objective. I’m not a fanboy of any camera brand. Every few years when considering new cameras, I consider all major manufacturers and pick the objectively best one. I have no qualms selling all of my equipment and switching systems if I feel it will help my photography/videography. So far, I made a switch from Canon to Fuji, then to Nikon, most recently to Canon again before selling the R5 and switching back to Nikon after it released the Z8.

Okay, with that out of the way, I will say that the ergonomics and customizability of both Z8 and R5 are class-leading, with a slight edge going to Nikon due to sheer number and usefulness of available functions. Neither of them gets all things perfectly, but they’re both very enjoyable to use and adaptable to different use-cases and shooting styles.

The main critique is that both Canon and Nikon use their own opaque logic of which functions to make easily available and assignable to custom keys, and which ones to keep less available by keeping them inside menus. It would be overwhelming to allow all functions to enter the custom control layout roulette, so they had to pick only a small sub-set… and did so according to their internal logic.

A few examples?

• Canon allows you to map a key to temporary boost screen brightness to maximum. Extremely useful when shooting something critical in bright sunlight and using LCD. Nikon has no such thing and requires multiple steps to change the brightness.
• Nikon allows you to customize all functions assigned to the Q-menu. Canon allows zero-customizations in this menu!
• Canon will let you map a key to toggle between AF-S and AF-C instantaneously. Nikon will require you to hold the AF button and turn a dial.
• Nikon allows you to map a key to stop flash from firing while depressed. It’s amazing for quickly switching between flash and no flash in tricky evening conditions, capturing both styles and deciding what looks good later.
• Nikon offers direct exposure +/- compensation to be mapped to a pair of easily accessible buttons in the front, which allow for quick and silent exposure compensation during video recording at all times.
• On Canon you can pick a button to toggle eco mode, allowing easy access to aggressive power saving (and cooling) when you can afford a bit sleepier camera.
• Nikon allows a single press toggle of FX/DX image crop in video mode, which is infinitely convenient, especially when shooting with primes.

(All of this is based on firmwares available in August 2023.)

Both Canon and Nikon are extremely possessive of some buttons, so for things like shutter button, SET/OK button or press of the joystick, only a tiny number of functions is allowed. Prime example of this is shutter button on Z8 during video. You can choose either “start/stop recording” or “no function”.

I’m not sure why it would hurt to allow a wider set of assignable functions for the shutter button which is pretty useless in video as it stands?

Presumably, they didn’t want the user to go wild and end up confusing themselves later during the shoot, but that’s really underestimating most serious professionals and enthusiasts who are perfectly capable of customizing the camera to tiniest detail and then remembering what they did while shooting (or at least remembering where to revert the changes).

Okay, moving on – I want to share an example of custom key mappings I developed for Z8. Maybe it inspires you for reconsidering your own settings, like setting up the R5 for videography changed my way of thinking about shooting with Z6 and Z8.

Keep in mind this is highly specific for shooting weddings and attempts to help in dealing with constantly moving subjects and changing lighting conditions during documentary-style shooting.

I have the following available mapped directly to a custom key in photo mode:

• deactivate flash while pressed (fn1)
• activate 3D tracking mode with face detection (fn2)
• exposure lock (hold) (AF-ON key)
• shooting menu bank selection + dial (fn3)
• revert to center AF point (joystick press or OK button)
• silent shutter mode (video record button)
• change AF mode (AF button, but with dial functions switched)

In video mode, just a few things are set up differently:

• exposure compensation +/- (fn1 and fn2)
• toggle FX/DX mode (joystick press)
• set mic level + dial (ISO button)

Autofocus in photo mode

The moment you initiate the AF you’ll know Z8 is top of the class. It’s basically the camera you were waiting for to finally switch to mirrorless, a camera that should be able to handle all the D850 could (and hopefully more).

To be fair, when talking just about pure ‘dumb’ autofocus (i.e. no object/face detection), both Z8 and R5 are very reliable, quick and powerful, leaving little to be desired.

My pain point with many mirrorless cameras (likes of Fuji X-T series and Nikon Z6) was the pretty unreliable AF in darkness of many reception venues. Combined with the general inability of mirrorless cameras to use the low-impact red focus assist beams of speedlite units, this was a recipe for disaster when you most need it (try capturing a bouquet toss on a candle-lit beach wedding with a few years old mirrorless!).

D850 was just around the range of where it could focus reliably enough in most conditions that you could be relatively confident on 98% of weddings. Its dedicated AF sensor was rated as working down to -4 EV light levels (but only for the central AF point and with f/2.8 glass or brighter).

For comparison, Z7 has a rating of -2 EV, and Z6 of -3.5 EV (less megapixels = bigger pixels = better on-sensor PDAF), with Z6 II going down to -4.5 EV.

While it seems like Z6/Z6 II is practically the same as D850 in this regard, in reality it was often struggling (and probably using CDAF instead of PDAF after some point of light getting dimmer, which meant very slow AF effectively not able to focus on moving targets).

A bit newer generation of cameras like Canon R5 and R6 work down to -6 and -6.5EV, respectively. (With disclaimer that YMMV depending on the aperture of lens used; this is for f/1.2 glass.)

Nikon Z8 claims sensitivity of -7 EV with f/1.2 lens (or -9 EV in very slow speed starlight AF mode, not practically usable for weddings). This is a significant upgrade compared to Z6, and even considering it may drop to CDAF in dimmest of conditions, it probably has enough headroom to be using PDAF in most realistic conditions you may find on a wedding day, without resorting to using bright and distracting AF assists like white or green LEDs.

My experience supports expectations of good AF under low light, but I’ll feel more comfortable with these claims once I’ve shot 10 candle-lit beach weddings. I’ll get back to you on that one!

AF usability in practice

What’s more important than these numbers is the usability of the particular AF system implementation. As I described in the R5 videography portion of the review, even the best AF with cleverest face detection can be rendered almost useless by a bad implementation.

The good news is that in this regard Z8 absolutely shines!

I just love its dedicated AF button, as keeping it depressed offers a quick switch between AF modes (AF-S, AF-C, AF-F) using one dial, and AF-area modes (like pinpoint, single point, dynamic area, wide area, 3D tracking, auto-area) using the other dial. You can also switch dial functions, which I did immediately.

Dedicated AF button is something I enjoyed on D850 and was greatly missing on Z6, so it’s great to see its return.

As on all mirrorless cameras, the idea is to use a size of AF area that’s as small as necessary (for precision) and as large as possible (for reliability).

The bigger part of sensor you allocate to AF, more quickly and reliably it will work as it will have more possible targets to catch on to. In good light and with decent contrast on your subject this won’t matter that much, but it will be a huge deal once the light gets dim (or when your target contrast gets low).

Bigger boxes here will help a lot, which is why it’s great that on Z8 you have a lot of choice of AF target box size, with two wide area modes where you can basically pick your own size and shape.

It’s worth mentioning that you can also deactivate any area modes you don’t feel you’ll ever be using (or create custom modes with different settings), which means that using the AF key shortcut you can directly (or at least quickly) jump between AF modes or target sizes depending on what your shooting needs of the moment are.

This is neither new nor unique to Nikon (Canon also offers this), but Nikon has a nice twist which I really like. When you disable an area mode on Canon, it’s completely inaccessible.

However, when you disable it on Nikon, it just removes it from the quick selection using the AF key + dial, but you can still access it through the menus.

This makes it easy to leave just your favorite modes on for quick access, but still having the flexibility when you unexpectedly need it. Example? That rare shot through a fence or thick leaves needing the pinpoint area mode (which is usually the first thing I disable).

Face detection and subject tracking

Let’s now talk about face detection for a moment. Nikon had some big claims during the Z8 unveiling in May 2023 and I was eager to put them to the test.

In short – they are true! Face detection is the best I’ve seen so far. It’s lightning quick, it’s very sticky and it’s reliable. It isn’t (and probably never will be) a 100% perfect, but it’s getting to the point where I’m comfortable using through most parts of the wedding.

One of the reasons why it’s so enjoyable to use is the fast sensor readout speed made possible by the stacked sensor design, which coupled with a new DSP easily outperforms both the R5 and Z6.

If a face is oriented towards the camera, it will really be able to pick it up when it’s quite far away (Nikon claims when taking up only 3% of frame area). I didn’t do an exact measurement, but I was impressed, faces getting picked up as far away as I though would even be reasonable.

Equally impressive is the head-detection from behind. While not as reliable as face detection from the front, I didn’t even expect the camera to do any detection from behind so it’s a nice bonus that can sometimes be useful.

One segment where the performance is less stellar is face detection of faces seen completely from one side (in profile). It was hit and miss if the face was smaller in the frame.

However, this wasn’t really an issue in practice because of Nikon’s implementation of face detection.

Whereas many cameras will force you to pick between an ‘area’ AF mode (where anything inside of a box is a target) and intelligent face detection modes looking for faces all over the frame, Z8 will combine both in an intuitive way where once the camera can’t detect a face, the usual AF target box will take precedence and will focus on whatever is inside.

This is great because you don’t need to trust the face detection working 100% of time to use it, fearing it will let you down in the worst possible moment (like the Z6 implementation often does) and without a quick workaround.

You can use it while it works, and when it doesn’t, just use the joystick to move the AF area where it needs to be, not missing a beat. The AF area is visible all the time even when a face is detected and you’re free to move it around. This is why you can hit the ground running if detection misses a face.

To elaborate, AF area modes (of various sizes) support face detection within their boxes, so you can guide the camera to where it should look for faces, or more importantly for weddings, which areas it should ignore. This way you can use face detection with a smaller target size to pick out one or two faces among many others, which is immensely useful when a lot of people are around (think ceremony or a dinner reception). This is especially important as Z8 will recognize faces that are reaaaaally far away and small in the frame!

This exact combination of reliability and usability was something I was waiting for a long time.

I need to say this still isn’t perfect, because if you use a larger AF area and two faces are detected inside, you don’t really have a quick way to select the face you want to focus on. In auto-area AF mode you can use the joystick to move between multiple faces (if they’re all properly detected), but in wide-area modes the joystick is used to move the focus box. So I guess we’re out of directional keys at this point.

The only workarounds I found were: (1) using 3D tracking with its precise small box, (2) using an area mode with a smaller box, or (3) moving the box away from the undesired face so only your target face falls inside.

The latter takes some time though, and can backfire if people are moving a lot, so you could miss the shot while fiddling with the controls.

Of course, you can completely switch focus modes and use dynamic area AF without face detection, which isn’t hard thanks to the dedicated AF key.

Talking about the dedicated key, Nikon allows you to do one more very useful thing: you can pick a key that will activate any one of the AF-area modes you’ve pre-selected (point, small, big, custom sized, 3D tracking, auto… anything!). You can also choose whether this will just change the area mode or also start the AF drive.

As long as you keep the button depressed, the new AF area mode will be active (and focusing, if you opted for that in settings).

This is just amazing. I can’t say how long I’ve waited for this on a camera!

Here’s a use case – I can be shooting with reliable and predictable dynamic area AF (small or medium is in my experience best compromise between precision and reliability), but any time I think face/eye detection and tracking would help I just press my custom key to jump to 3D tracking and it’s active for as long as I need it. Any issues, I let go of the key and have my reliable expanded focus points ready.

Of course, you can set this any way you like; perhaps have a wider area with face detection on by default and have a single point ready as backup on your custom key, or any other combination you like.

Note: you can’t set a key to temporarily jump between AF-S and AF-C modes, only between AF area modes. Canon supports the temporary AF-S / AF-C switch via custom key ever since days of EOS 6D, which I finally don’t miss anymore with all the next-level AF stuff on Z8.

All in all, an amazing AF implementation that leaves little to be desired.

For comparison, R5 has a different approach. It will let you manually move its AF area, and when you’re near a target face, use a custom key to initiate face detection with tracking. (Note: Canon officially only allows you to map eye detection to a custom key, but thankfully it will still detect faces even if it can’t detect an eye.)

This will track just the person you picked as they move throughout the frame. Of course, if tracking loses the face it’s a bit tricky, but with a custom key mapped to activating the face detection you just need to let go of it to instantly revert back to manually positioning the AF area.

The fact that face detection starts around your existing focus point but then tracks the face around the frame means something else: you can break your framing for a moment and catch the face by moving the camera so your point falls on the desired subject (as opposed to moving the AF area), then activate face detection with tracking and quickly re-framing the shot before shooting.

Despite of how it sounds, this process is usually quicker than manually moving the AF area with your joystick.

On Z8, you have a mode that will work in exactly this way – 3D tracking. Face detection in other area modes (except auto-area AF) doesn’t track them outside of predefined AF area, so you can’t move the frame to ‘catch’ a face and reframe before shooting as the AF area would lose the face once you reframed.

Z8 has amazing, game-changing 3D tracking, which is essentially object/subject tracking activated when its small and precise focus area is hovering over an object. (The size of area is predefined and fixed, around the size of a single default focus area ‘unit’.)

Half-press of the shutter button will start the tracking all around the frame and keep it in focus as long as you keep the button half-pressed. If a face is detected under or near the focus area, it will focus on the eye or the face. If it loses the face, it reverts to regular tracking, which is very useful as you don’t have to rely on face recognition working 100% of the time, camera will continue tracking no matter what it thinks it is.

When moving the object around the frame a lot, camera can drift away from it or drop it completely if the object is changing shape or appearance quickly. However, this happened very rarely in my testing and the workaround was quick – just letting go of the shutter button to stop tracking, and re-initiating it from the last point where the focus area was set.
I could see this being a problem in certain critical situations (like shooting with a strong backlight, within crowds of people or repeating patterns), but more testing will tell.

3D tracking quickly became my favorite AF mode. Besides its reliability, the combination of instant face detection with a fallback to object recognition is just amazing. Most importantly, I get the focus-and-reframe functionality with full AF-C and tracking, where I previously had to manually move a point in AF-S with focus-and-reframe on DSLRs.

In short, I’m in love with 3D tracking.

Finally, we have the auto-area mode which uses the whole frame to find faces. Among the detected faces, you can pick the one you want to track using the joystick. I tend to steer away from such modes even with a small number of subjects in front of the lens as no face detection (so far) works flawlessly 100% of the time, so when it loses a face from its sights there is no way to tell the camera where it should keep focusing.

Wide area modes are much better in this regard as they automatically revert to a focus area which you can move wherever you want, all the while focusing without any interruption.

Pixel peeping

I promised no pixel-peep talk here, so this will be short. The 45MP sensor produces very high quality raws that leave little to be desired. It’s not quite on the level of the D850 just looking at the measurements, but you’ll be hard pressed to notice any difference.

Knowing the underlying difference between DSLRs and mirrorless cameras when it comes to AF sensors (i.e. mirrorless cameras having to integrate PDAF pixels into the main sensor itself), it’s almost surprising that the image quality is on the same level!

The measured dynamic range is tiny bit smaller on Z8 at the base ISO of 64, theoretically giving you slightly less forgiving photos when needing to do extreme exposure pushing after the capture. However, in practice I could push the shadows as much as I wanted and still didn’t notice image falling apart. I also found no evidence of the ugly and problematic banding noise that the Z6 exhibits when pushed +5 stops.

When considering how much easier it is to capture proper exposures with mirrorless cameras in the first place (with exposure preview and live histogram), all this headroom you have will be even less utilized when editing.

In summary, the sensor is still ISOless for all intents and purposes, which is truly great news!

DPReview has completed their studio scene measurements, which I recommend to anyone wanting to pixel peep.

I compared Nikon Z8 and D850, Canon R5 and Sony A1. If you just want the key takeaways:

• R5 has the least amount of noise as ISO gets higher, followed by D850, Z8 and then A1. (Note: A1 has a 50MP sensor, compared to 45MP in Z8, D850 and R5). Looking at ISO 12.800 difference is visible, but not decisive.
• Image sharpness on lowest ISOs is best with Nikons (D850 slightly ahead of Z8), followed by Sony and then Canon. It’s probably due to the low-pass filter, as Canon has the least amount of moiré. However, the loss of sharpness in Canon is barely noticeable, in contrast to the moiré which is more noticeable with Nikons (and Sony to a certain degree).
• There’s a lot of visible chromatic aberrations with D850, somewhat less with R5 and none with Z8. This has nothing to do with sensor, of course; it’s probably a result of automatic software adjustment despite being shot in raw (possibly lens CA information passed down to the raw converter). Even so, it’s amazingly well controlled on Z8 and if it means not having to do it manually later, I’m sold.

I’ll give the final mention to light metering, where Z8 excels. Initially, I was worried because Nikon has removed the option of a custom shortcut that would meter only for highlights while pressed, but the metering has really impressed me and I’m starting to think I won’t need highlight biased metering at all. I generally have to adjust exposure compensation much less frequently, even when shooting backlit subjects.

Part of the ‘trick’ in this particular case is simple – when a face is detected in the focus area, exposure is biased to properly expose for it, even if it means turning up the exposure by 2 or 3 stops.

Overheating

I’d wager that many photographers who don’t read photography news every day first heard about Canon R5 after stumbling upon articles with titles like “Picture Perfect Meltdowns: Surviving the Sizzle with an Overheating R5” or “Say Cheese, but Watch for the Heat Wave: A Hotter-than-Hot Canon R5 Review”.

To be fair, R5 just happened to be the first major camera to hit upon the wall of physics which prevents us from sticking a tiny piece of silicon into a compact camera body and driving it at 101% of its capability without causing significant heat emissions.

Capturing 8K video or very high frame-rate 4K video will get things hot and there’s just no escaping it with current technology.

You can use larger bodies which dissipate heat more efficiently (like Nikon Z9) or stick active cooling behind the sensor (like Canon R5C), but for a compact body you’re pretty much out of options if you want very high performance.

All of this goes to say that despite Nikon having three years more to develop Z8, it still suffers from potential overheating issues like the R5 does.

To be fair, Z8 performance is much more in line with R5 and its 1.8.x firmware which allowed much longer shooting/recording times, than the flaming hot potato of the camera original R5 v1.00 was.

From anecdotal evidence, Z8 performance is pretty variable depending on whether you shoot photos, video, how hot it is, whether you’re in direct sunlight, whether you use two cameras or just one, it varies depending on how much burst photos you snap and what your video frame-rate is; it will change if you use high-fps viewfinder mode, don’t allow camera any pauses between shots, have longer standby times etc. etc.

I will refrain from reviewing this part of the camera usability until I’ve had many more weddings in different conditions. You can find static video recording tests in many other reviews, but I just don’t feel that shooting a static scene in the shade (even with hot weather) is a realistic test for a wedding photographer/videographer. (First off, we often shoot in direct sunlight, but we also don’t shoot long clips for the most part. Therefore, we have things working both with and against us, so we just need a realistic test — preferably, many of them!)

So far I’ve only gotten to a CFExpress hot card warning during a 15 minute 4K25 clip in cloudy weather at around 30C/86F. A tiny issue in pretty reasonable summer weather.

What I want to see is how it will handle 4K100 clips and thousands of photos at 40C/104F in direct sunlight. You’ll be the first to know!

Overall impressions of Z8 for wedding photography

In one sentence, Nikon created a monster-camera for wedding photographers which – once you first depress the shutter button – is instantly addictive. It’s enjoyable to use, it’s very speedy and instantly reacts to your commands, it’s silent (if you wish it to be) and offers very high image quality.

The biggest problem I have is that I tend to shoot much more photos with it, filling up my cards faster and creating problems when having to cull all those images later.

Knowing that there is no mechanical shutter coming one tiny step closer to failure with each photo I snap, combined with no audible sound when taking a photo as a cue something was just captured, it’s very, very easy to get carried away and overshoot.

Whether you’re currently a Nikon user or not, you’re likely to love how capable and intuitive the Z8, whether it’s for wedding photography or pretty much anything else.

In part two of this review, I’ll talk about Z8 for wedding videography.

Technical specifications

Many reviews focus on tech stuff, so I didn’t repeat most of it here if it wasn’t crucial to the user experience. Here’s some of the reviews I used when deciding to buy the camera and confirming the specs.

A detailed overview of all Z8 technical specifications is on Nikon’s website.

DPReview did a nice preliminary review of the Z8 here.

Gerald Undone did a quick video review of the camera and measured few things I haven’t seen elsewhere, like dynamic range and sensor readout speed.

( Part 1: Photography | Part 2: Videography )