Based on firmware v1.01
Updated: September 2023
(If you’re looking for Z8 review for wedding photography, see part 1 of this review.)
I’ll preface this videography Z8 review with a short intro to my shooting style, which is important as it guides the imperatives of user experience in my eyes.
As a wedding videographer, I shoot documentary run’n’gun style. I don’t use second shooters or additional cameras, I shoot handheld, move a lot, use wide lenses (mostly 28mm f/1.4) at wide open apertures. I use a lot of aids the camera offers, like aperture-priority exposure mode, autofocus with tracking (if it’s any good), auto white-balance and any kind of mechanical stabilization offered by the camera or the lens.
Reliability and predictability of all of these is crucial, as there are no do-overs and no second shooters with me, so this is something I’ll pay a lot of attention to in this Z8 review.
This review will in part also compare Canon R5 and Nikon Z8 and talk about the different user experiences.
All photos in this review were taken with Nikon Z8 and edited in Lightroom.
Without going into too much boring details, Z8 can record anything from 1080p120 and 4K120 to 8K60. It always records sound regardless of the mode (looking at you R5!)
Supported codecs are h.264, h.265, Apple ProRes, ProRes RAW and Nikon Raw (N-RAW). Not all combinations are supported, so consult the documentation if this information is critical.
Of more interest is the fact that it supports 8-bit and 10-bit color depth (or even 12-bit with raw video formats). In h.265 it supports 4:2:0 chroma subsampling and if you opt for ProRes you can get 4:2:2.
It’s possible there is some baked-in noise reduction in all video modes except N-RAW, based on a test Gerald Undone did, but knowing what I know about quantization noise and video codecs, I’m not convinced. In his test it looked very much like classic perceptually encoded macroblock dragged out from the shadows by adding gain and not any kind of noise reduction, but this portion of the video was short and superficial, so I’ll refrain from hard conclusions.
Based on my own practical tests of the h.265 codecs, I’ve seen plenty of fine noise in 4K HQ mode which wouldn’t be present in its finely detailed form if there was any kind of noticeable noise reduction. (And the video noise reduction option is disabled in log-gamma mode, so there’s a hint as well.)
I will mention here that the R5 has comparable video modes, although not all. It goes up to 8K30 instead of 60, it does support 4K120 but without recorded sound and encapsulated into a 30fps container (so it plays in 4x slow motion by default).
However, R5 can record 4:2:2 subsampled footage in h.265, while Z8 only supports this if you use ProRes HQ (which is recorded at a pretty high bitrate, not really practical for multiple hours of shooting weddings); otherwise all h.264 and h.265 is recorded in 4:2:0. How important this is depends on your use case. For non-critical color work like weddings, no one will probably notice any difference, least of all your clients.
Note: I’ll now switch to talking about 25/50/100 fps modes of PAL (European) system, instead of NTSC (American and Japanese) modes of 29,97/59,94/119,88 fps. They will be closely equivalent in corresponding modes so you can follow no matter which system you’ll end up using.
I opted for detailed testing of h.265 N-log in since I intend to record in this format. I found it’s a great compromise between practicality and quality, and having hardware decoding of h.265 on my M1 Mac, editing is as smooth as any other editing-centered codecs like All-I h.265 or Apple ProRes.
Available bitrates in N-log in 4K modes go from around 200Mbps in 4K25 to around 400Mbps in 4K100. On a 256GB card this will approximately give you from 1 to 3 hours of recording time.
Let’s now mention first usability detail that has been bothering me with Nikon ever since I started shooting video with it. In most cameras I used, you will pick your video system (PAL or NTSC) and you will only see compatible framerates selectable for recording. Logical. You usually know where your target audience/client live and don’t need the other system.
However, Nikon doesn’t ask you which system you prefer and always offers all framerates when selecting recording modes. So you get double the amount of modes, which means double the amount of scrolling, possibility of tapping the wrong one in a hurry and wasting precious time all around.
With Canon R5 this isn’t a problem as it offers only compatible modes like other cameras. Canon also allows you to pick three GOP encoding types for most frame sizes/rates: All-I, IPB and IPB light (IPB with lower bandwidth). I really liked these options as they offered flexibility depending on your circumstances. Another point for Canon.
However, where Canon lost me is the 4K100 mode which forces you to use All-I encoding with a monstruous bitrate close to 2 gigabits per second! That means less than 20 minutes of recording time on a 256GB card and you also need an extremely fast card (it cannot be the SD card).
I end up using the 100fps mode for about 10-20 minutes of footage per wedding, which means that with Canon I’d need another expensive and fast 256GB card just for this portion of the shoot! A huge win for Nikon which uses ‘only’ 400Mbps in this mode and can fit 1.3 hours of footage on the same card.
I also did a quick test with Z8 recording 4K100 to a cheap V30 SD card which worked normally for the duration of the test, so it’s possible you don’t even need the expensive CFExpress type B card.
Finally, this may be a good place to mention dual-card redundancy, which Canon has and Nikon doesn’t have in video mode. Canon can record compatible (lower bitrate) modes on both CFExpress and SD card simultaneously, while Nikon doesn’t offer anything like this.
Caveat is that once this is turned on, Canon will refuse to record anything in incompatible video modes, until you either turn off dual-card recording or change the video mode to something else.
This is honestly mind-boggling. If you get caught out with this at inopportune time, you will miss shots while fiddling with the camera. The worst part is that these settings are universal, i.e. not linked to custom shooting modes, so you can’t create e.g. custom 4K100 mode which goes to one card and another custom 4K25 mode which gets written to both cards.
In practice, it means I ended up not using dual-card redundancy with Canon even though this would be a huge load off my mind, knowing I’d have something even if one card fails. Another example of hardware capability being there without clever support in software. It’s a shame.
Resolution, oversampling, details…
Let’s talk about the fact Z8 has a 45MP sensor, just like R5, D850 and Z9. It’s not a coincidence they all have 45MP sensors, as it naturally gives you the option of 8K output.
Even if you have no interest in shooting 8K, this is still good news because you can get high quality 4K footage created from native 8K output. This has various names: oversampled 4K, 4K HQ or extended oversampling. (I’ll use Canon nomenclature of 4K HQ because it’s short and clearer than Nikon’s slightly cryptic ‘extended oversampling’.)
Hundreds of articles have been written about this, so I’ll just remind you of the basic idea: a big photo that you make smaller increases in quality, apparent sharpness and even dynamic range (as noise is reduced while downsampling). Conversely, a small photo gets worse as you enlarge it.
Since we have an 8K sensor and require 4K output, we can do one of two things.
One, we can read out every second pixel (8K / 2 = 4K) which is also called line-skipped 4K.
Or two, we can read full 8K and use image processing to downsample it using high quality algorithms. Other way of looking at this is that we have oversampled our frame by reading out 8K of pixels even though we need just 4K, which is why it’s called ‘oversampled 4K’.
Line skipping is quick and efficient. Quick means fast readout times and little rolling shutter artifacts. Efficient means power-saving: your battery will last longer and your camera won’t overheat (as easily).
The bad news is that this is of course just one part of a compromise and the thing you’re sacrificing may turn out to matter the most: image quality. Line skipped 4K lacks sharp details and readily produces other artifacts (like aliasing or moiré).
Oversampled 4K, on the other hand, is very demanding for the camera. It has to read out 45 million pixels many times each second, process them and encode them. This means slow readout speed (around 3 times slower on Z8!), much more pronounced rolling shutter and a lot of heat and battery consumption.
The good news of course is that this is the classic case of having a big photo we want to reduce in size, thereby increasing its quality.
Okay, that’s all theory. What about practice?
Simply put, it matters – a lot.
Standard (line-skipped) 4K modes on both Z8 and R5 leave much to be desired. I wouldn’t even call them true 4K for their lack of details and very visible moiré on fine patterns. Diagonal lines are also visibly jagged (aliased). Whether this is good enough for you depends on what you’ll be using this footage for.
The oversampled footage in both cameras is nuanced and detailed, but not over-sharpened, so there’s still room for improvement in post.
DPReview has completed their studio scene measurements, which I recommend to anyone wanting to pixel peep. Here are some key takeaways for video shooters:
- Amount of detail in 4K oversampled (HQ) modes in Z8 and R5 is comparable. Both images look beautiful with plenty of details, but Nikon has both more apparent sharpness and jaggier edges of straight lines.
- Since we’re talking about re-compressed JPEG outputs here, it’s difficult to untangle influences of downsampling algorithm, sharpening, video compression and photo compression (for the final stage). It’s likely user settings, like sharpning setting, can influence the output in these regards.
- In high frame-rate 4K capture (100+ fps for Nikon and 50+ fps for Canon), quality of details from both cameras drop dramatically. It’s no wonder since they barely have the time to read their sensors, let alone process the images.
- The results from Nikon and Canon are… different. They’re both equally bad compared to oversampled 4K, but in different ways.
- Canon has blurrier details, so if the goal is to attain maximum resolving power, Nikon would win easily. However, Canon is probably blurrier because it’s using a softer downsampling method, possibly bilinear or maybe even nearest neighbour with some gaussian blur. Looking at some aliasing artefacts in Canon’s footage, I think this is easily possible.
- Nikon appears to be using just the nearest neighbour method (basically line skipping), which produces images with more apparent detail, as the results tend to have sharper edges. Nikon’s approach, however, results in all kinds of artefacts like aliasing and moiré, whereas Canon’s blurry image covers those up a bit better.
- If the goal is to have a more natural picture, Canon wins, but if you’re after details (both real and aliased), Nikon is better.
Now that I’ve seen the difference between oversampled and line-skipped footage, I can’t say I’ll ever be using the latter unless my camera has a nearly dead battery or is severely overheating.
Where the two cameras differ more are small (but important) details.
Z8 does oversampling automatically in 4K25. Setting the ‘extended oversampling’ option on or off makes no difference. Your choice only applies in 4K50. Again, in 4K100 it doesn’t apply as camera is forced to do line-skipping due to high readout speed requirements.
Canon R5 can’t do oversampling in 4K50, which is a shame. The sensor is just too slow to do 8K50 which it will then downsample, which is also why it doesn’t offer an 8K50 mode.
I use 50fps mode a lot during the wedding. Not only do I like to slow down some romantic moments, I also like to use it for shots of details, panning and establishing shots because it looks much smoother when shot handheld, as I do. So having 4K50 HQ mode is great!
Finally, even if you don’t care about the amount of details that much, if you need to do additional processing in post (especially cropping, straightening or digital stabilization), sharper starting image will always end up as a sharper final image, so beginning with HQ is always a good idea.
What I don’t know as of yet is how taxing the HQ mode is. How much faster will it drain the battery and what about overheating?
For now, I did a full day of 4KHQ recording shorter clips at around 25°C temperature with no issues. I only got a ‘hot card’ warning during a longer clip (around 12 minutes of 4K25 HQ).
I still didn’t get to do a proper test on a 40°C wedding, but others did extensive 20°C testing. While it won’t mean much for performance in really hot weather and direct sunlight, we can extrapolate some trends, so let’s dig into it.
4K25 HQ mode doesn’t appear to cause overheating at all (which is probably why Nikon decided to record 4K25 HQ by default, without option to turn off the oversampling).
In the 4K50, HQ mode will at least halve the recording time due to overheating if shot continuously.
If you’re shooting shorter clips, it appears Z8 will be able to shoot for a long time and overheating may not be a factor at all as it appears to be pretty good at cooling down quickly, but I’ll get back to you on this one at the end of the summer when I’ve recorded through a few long heat-wave wedding days.
In regards to overheating, Z8 appears to be slightly better than R5 with the 1.8.1 firmware. This is probably due to more efficient stacked sensor design, but heat management may also be in question (after all, Nikon has had 3 more years to solve R5’s problems).
Two more quick mentions here.
Nikon and Canon both offer FX (1x) and DX (1.6x) crops in 4K. However, Nikon also has 2.3x crop, which is a nice addition.
What I like the most is that Nikon has a customizable shortcut key which toggles between FX and DX modes with a single press, which makes it immensely useful. With Canon your best option is digging through the Q-menu or camera menus. While you can turn it on, it can’t be a spur of the moment thing. With Nikon you can basically change your mind every clip and it will take just a fraction of a second to switch. Huge win for Nikon.
Final detail: Z8 has something called high-res zoom. It’s quite simple: since 4K output is taken from an 8K sensor, we obviously have more pixels than we need. If we don’t do oversampling, we can use those extra pixels for simulating a zoom lens without losing much quality (i.e. it’s not destroying image sharpness like usual digital zoom does). It’s basically just a variable crop between 50% and 100% of sensor output. It can be done smoothly and even mapped to Z-lens control rings or custom keys for better experience.
Caveat is that it only works in 4K30 or 1080p modes, only with certain focus modes and without face detection.
Recording interface and tools
Let’s take a short break and get back to usability.
One of my pet peeves for years has been the fact that most recording devices and cameras have a single small indicator that recording is in progress – usually a red dot, sometimes accompanied by a bit of text.
In theory that’s all well and good. But in practice that’s something like designing a bridge to withstand the usual amount of traffic, or having just one parachute, or recording to just one card. It will work for a long time, but then, something will fail and everything will blow up in your face.
Who hasn’t heard a story (or had it happen themselves) of a videographer who thought he was recording… when he didn’t?
I had it happen more than once. Sometimes it’s your fault, sometimes you press the button but maybe not quite completely, sometimes you just press it at the wrong time, when camera is in-between states (often a problem on slowish Z6), or you get distracted, or the camera flashes an error and doesn’t start rolling… it could be anything.
As unexpected things happen on a wedding, your reflexes guide your hand to press the record button and your brain is then on autopilot, starting to consider other recording parameters. And not noticing…. that no recording has begun! Weddings are full of these moments and I’ll be the first to admit I occasionally had misunderstandings with my Z6 in this regard.
(Not to digress too much, but Z6 has this very protruding eyepiece which under some angles exactly covers the red dot indicator on the screen, e.g. when looking from above with screen tilted out. It’s so bad that I had to remove the eyepiece permanently. By the way, Z8 has no such eyepiece, it’s pretty flush with the camera body.)
I’m so determined to avoid the not-recording issue in critical moments that I have a simple rule for recording ceremonies: I press record at the very beginning, make sure I’m rolling multiple times and never stop recording until the end, even if I’m changing position, shooting b-roll or just walking around, lens pointed towards the floor. Under no circumstances will I stop rolling!
All of this brings me back to the recording indicator. Couldn’t the manufacturers help us out and make the indicator… more obvious? Like a bigger red dot? Or maybe two dots, one in each opposing corner?
Or… let’s go a bit overboard here… maybe add a red frame around the whole screen? (This was actually my idea a few years ago.)
Well, imagine my surprise when it turned out that Nikon is actually listening to user feedback and that the Z8 has implemented this very thing – an (optional) red frame around the whole screen!
Not only that, but they were so proud of this (rightfully!) that they mentioned this function in their Z8 announcement multiple times!
Let’s move on. As you noticed by now, I used to record with R5, but not anymore. I sold it. And if you want to know one of the two dealbreakers it had, we’ve just reached the point to talk about it.
R5 has it in photo mode and has it in video stand-by mode, but press record, and it’s gone with no way to turning it on. I won’t go into more details (read the second part of my R5 review if you want to know more), but my pet theory is that Canon did this deliberately to preserve R5C differentiation.
That’s all good Canon, but basically all other cameras offer a live histogram so I don’t know what your thinking was here.
The good news is that Z8 has not only a live histogram, but a waveform monitor as well! You can also pick in which display modes it will show and you get to pick small or big waveform. I’m in love.
In addition to having the live histogram, Z8 has zebras as well and they can be displayed at the same time as the histogram. Both cameras have dual-zebras (for highlights and for custom portion of the mid-tones), but only R5 can display both types at the same time. With Z8 you have to pick which ones you want displayed (which can be mapped to a custom key).
Both cameras has customizable zebras, with Nikon offering slightly more precision in setting them up.
Another thing that was driving me crazy on Z6 was the sound level meter. Since Z6 has fixed information screens, I could either display the screen where I could see the sound levels without the live histogram, or vice versa.
So, if you’re shooting a ceremony in tricky lighting (which happens to me around 100% of the time) and want to monitor sound levels along with the exposure info, you need to pick (or constantly switch) between the too.
Z8 has blessed me with not having to pick anymore as it can finally display both!
More generally, both R5 and Z8 have customizable information displays and this is very useful. You can pick amount of information layouts you want to cycle through, you can pick which groups of elements you want visible and you can decide what applies to LCD and what to EVF. So, I can populate my own screen with a waveform, sound levels and anything else I find desirable.
Finally, a short note an useful tool for focusing. You know how you can zoom into the live view to check focus before you shoot or record? Well, I finally have in my hands a camera which can zoom into the picture while recording!
I was very happy when I found this, because I can now check my focus during long takes without risking re-activating AF and having it hunt or wander off for whatever reason, spoiling the shot.
I love this new functionality and much prefer it to focus peaking, as it allows me razor-sharp focus on practially anything that has any kind of texture or contrast. When you zoom in, it’s easy to fine-tune your focus with manual focus ring, zoom out and continue recording like nothing happened. I tried this in 4K25, 50 and 100 and it always worked!
I’ve been waiting for 10-bit log-gamma internal recording for a long time. Even though Z6 could theoretically offer it, it was only with an external recorder, which wasn’t really compatible with my run’n’gun style.
The basic idea here is that shooting documentary style with constantly changing conditions and no control over the events you want some leeway and flexibility in color grading because you’ll never get everything perfect. Some fleeting events occur and disappear by the time you’ve figured out your exposure and started inputting the manual settings, which is why I mostly work in aperture mode to let the camera do the thinking, while I focus on capturing what’s in front of me.
This doesn’t work perfectly all the time so I don’t give camera a free reign; I guide the meter with exposure compensation and use exposure lock (hold) available via a handy button (AF-ON) which lets me work in a pseudo-manual mode, where the exposure is locked and constant for however long I need it to stay so. I’m always just a key tap away from getting back to full auto if anything surprising starts happening behind my back.
But no matter what, you’ll still make mistakes, which is why the flexibility of 10-bit log footage is so desirable.
I was in love with Canons C-log3 gamma profile because it was easy to grade and looked beautiful. With N-log you get roughly similar specs, their curve also encodes around 12 stops of dynamic range, like C-log3.
Initially, I had a more difficult time with grading N-log. Thing that confused me is the Nikon’s N-log LUT (for log-gamma to Rec.709 conversion) which produces a very bright, almost burnt-out, punchy, saturated and contrasty image when applied. It’s very much completely too much and I thought I was doing something wrong.
Even worse, it makes the log-noise more pronounced and doesn’t tolerate any overexposure in the footage because it easily blows up highlights and anything resembling them (overexposure would help fight the shadow noise).
I’m not sure what the deal is here. While I don’t have extensive experience working with log-profiles, I had no issues converting and grading C-log3 and D-log footage with their corresponding input LUTs.
Anyway, I quit on the Nikon’s N-log LUT because I prefer to overexpose my log footage by about 2 stops to control the shadow noise, and overexposed footage absolutely falls apart with N-log (and it’s almost impossible to bring it back after applying the LUT). I found a really nice, color-accurate LUT from Gamut.io which works much better than Nikon’s to start the grading process.
Grading this footage with creative log LUTs is a different matter. It’s nice to work with, quite forgiving (especially if you’re coming from 8-bit world) and looks great when graded properly. In this regard there aren’t huge differences between C-log3 and N-log in my book.
I’ll just mention that both sensors in R5 and Z8 have roughly the same measured dynamic range in video (in the ballpark of 12 stops, depending on the codec used), and both C-log3 and N-log are well suited to encoding that range. (Original C-log has around 11 stops and a hard cutoff in the shadows, so I preferred the 12 stops of C-log3 and it’s softer shadow roll off into noise.)
On the topic of noise, log-gamma will make your shadow noise more visible, as it does by definition, so let’s talk about noise generally.
First, a note: I did all noise testing with a cold camera. Shoot in the heat for prolonged time and it will get worse.
Z8 has solid noise performance. Shooting video (4K HQ h.265 10-bit N-log) at ISO 25600 produced usable footage.
Noise was very visible (it’s a very high ISO, after all), but it was the acceptable type of fine noise (high-frequency). In preliminary testing, I didn’t really see the dreaded low-frequency noise, banding or blotchy medium-frequency noise.
Dropping the ISO to 8000-10000 area produced really good results which I would use in editing without any noise reduction.
With that said, here we have one of those significant differences between R5 and Z8 – noise reduction in log-gamma profile. Canon supports it, while Nikon doesn’t.
This isn’t really ideal for Nikon. Noise reduction can be done post-hoc, in editing, but it’s paralyzingly slow and inefficient. You can get any clip to look good with plugins like NeatVideo, but it takes a huge amount of processing and many denoising plugins/effects aren’t even viewable in real time. Once you apply them, your edit is pretty much cemented. It’s tedious to do further edits and exporting takes forever.
Since weddings are about dealing with a lot of material in a reasonable amount of time, I like the Canon’s option of in-camera noise reduction in log-profile. It is available in 4 levels and the middle ones did a terrific job of reducing noise while preserving details. I was happy to use it and the results were great, coming into the edit with already cleaned up footage and no slow-down penalty.
However, in the end, given the Z8’s amount of noise at high ISOs and the fact wedding videos are for most part viewed over streaming services with solid compression, I think this probably won’t turn out to be a huge deal.
Under this somewhat unusual subtitle I’d like to talk about the fact that when using log-gamma curve the camera will increase its base ISO. I’ve discussed the reasons for this in the R5 review, but in summary, this is just a trick by camera manufacturers that mostly does nothing to camera’s sensor and amplifiers, used to guide your exposure settings into preserving more highlights (which the log curve is well poised to capture).
In case of both N-log and C-log3, your new base ISO will be 800. So if you decided to listen to camera’s metering after jumping to log-profile, you may be using ISO 800 or higher to shoot a sunlit scene. Now let’s say you desire to shoot wide open, for example at f/1.4 as I do, and you may start noticing a problem.
You’ll need a really short exposure!
With Z8, you won’t have any issues with that as at allows going up to 1/32.000s in video. Canon here was a big disappointment however, allowing only 1/4000s – which is really low. I’m not sure what the reason is, my guess was that their slow sensor readout speed limited what they could offer here. A big plus for Nikon.
AF and face detection
The good news is that most of the AF system and modes work similarly in the video mode. You can refer to the photography part of the review for more info, so I’ll keep it short.
Wide-area AF modes still all offer a focus box with option to detect faces inside. Great.
3D tracking also works and is fairly sticky to its subjects. Very much liking it for panning and tracking shots right now, but time will tell how good it holds up. It’s not perfect so I probably won’t use it in critical moments (like tracking a bride’s face as she comes down the aisle, among dozens of other faces), but it will make my life easier in many other non-critical moments during the day.
The amazing thing is that subject recognition and 3D tracking work even in 4K100 mode, as opposed to Canon’s R5.
There are however two differences between photo and video mode on Z8 which I wish weren’t there.
When using 3D tracking (called subject tracking in video mode), there’s a focus box you can move around and you start tracking by half-pressing the shutter button. To end tracking and resetting the focus box position in photo mode, all you need to do is let go of the shutter button.
In video mode, you will need to press the OK button to end tracking.
This isn’t a huge deal, but it’s an extra step I’ll need to get used to when switching from photo to video. Maybe the idea was to make it easier to track a subject if you adjust/lose your grip during rolling? For this it will help, probably.
In photo mode, we could also map a custom key to jump into any AF area mode we wished (while it’s pressed). There is no such option in video, so you’ll have to use your AF button to switch modes using a dial, or save different AF modes into different shooting banks (which I did).
It’s a shame camera doesn’t behave in the same way in both photo and video mode. I don’t really see a technical reason for it, it’s just a bit of software.
However, it’s still lightyears ahead of Canon’s AF implementation in video, where you essentially need to choose between AF-S and continuous, constant, unending full time focusing which continues even when you’re not recording (called AF-F on Z8). If it sounds ridiculous, it is. It’s one of the main reasons why I sold the R5 after only a month. For more details, refer to the R5 review.
Anyway, the video AF implementation on Z8 is great, it’s flexible and pretty enjoyable to use, even if it isn’t perfect. In contrast to Canon, you can pick between AF-S, AF-C and AF-F modes and any AF area you’d like, so you can really set it up according to the needs of the moment.
I tried testing face detection at very low light, and to my surprise it worked pretty well when face was moderately large in the frame. We’re talking about ISO 25600, f/1.4 and 1/25s darkness!
I will continue testing this in the field as it can depend on many factors, but so far it’s more promising than I dared hope.
Shooting handheld has its pros and cons. It’s more organic, allows you to move more, be quicker and nimbler, react faster to events… but it also produces shakier footage than using some mechanical stabilization system. This is why camera stabilization plays a big role in this style of shooting.
Z8 on paper offers similar stabilization as the R5 (around 6 stops in CIPA testing).
In my testing, this stabilization was visibly better and more stable than what Z6 offered. However, Canon paired with RF lens with optical IS was better still, as it had IBIS working in concert with lens IS. And since that RF lens has a big imaging circle, IBIS can really get to work and move the sensor to further extremes than would be possible with adapted lenses with smaller imaging circles (suitable for EF or F mounts).
With Z-lenses you get more stabilization than with adapted glass, but for non-IS Z lenses this will be just a modest increase in performance.
Various ergonomic observations
Coming from Z6 and R5, the Z8 is fast to start up and fast to start/stop recording. It’s very responsive.
Z6 was sluggish in this respect, so from stopping current recording to being able to start a new one would take some time. If you pressed the record button too soon, nothing would happen so more precious time was wasted. R5 was somewhere in-between.
Z8 also retains the good old power switch around the shutter button design. It’s great for one-handed operation, especially if you turn off camera often to save battery life or help it cool down. R5 was terrible in this regard, requiring two hands to pick it up and turn it on. I thought I’d get used to it, but I didn’t.
Another thing I like on Z8 is the fact that it’s the first mirrorless I’ve had where you don’t see an exposed sensor when changing lenses!
It’s really nice to know you have the option of turning on the sensor shield when camera is off, so any dust or objects don’t fall directly on your sensor while exposed. (Activating the shield does make the startup time a tiny bit slower. It’s barely noticable, though.)
Moving on; with R5 I had somewhat of a conundrum to solve. How do I adjust exposure compensation during video recording silently and without shaking the camera?
My solution was to map this to the RF lens control ring, but it was still clicky, thin and kind of hard to reach, depending on what kind of grip I had at the moment and how close the focusing ring was on that particular lens.
With any Nikon Z camera you can use fn1/fn2 buttons in pair for this purpose. One increases and the other decreases the exposure compensation. They’re soft, completely silent and perfectly placed under fingers of your right hand while holding the camera’s grip. It means they’re easily accessible and pressing them skilfully won’t introduce more shake into your footage. While I’d like to use them for some other functions as well, this is so convenient it’s irresistible.
While on topic of customization, Z8 allows a custom key for setting the mic recording level (using a dial to change the level). I just love this, makes my life so much easier.
On Z6 I had to go into the Q-menu, which while recording usually produces visible shaking no matter how careful I am. R5 was similar, but its Q-menu isn’t customizable so you can’t conveniently place mic level function where you’d want it for easiest access (like in the first spot).
Z6, Z8 and R5 all have different LCD screens. Canon’s is fully articulated, flip-out on a single hinge. Z6 has a tilt-screen on one axis, while Z8 can tilt its screen on two axes (helpful for vertical shooting).
For my particular handheld shooting style, Canon’s screen is a nightmare.
I use the camera strap stretched tightly around my neck to provide the third point of contact with the camera, which helps with getting more stable footage. With Nikon cameras, this puts LCD in front of my face. I tilt it so I can view it from above – perfect.
With Canon, if you want the screen tilted it needs to be flipped out to the side first. So, at that point you have a screen to the left of your face, sticking out on a fragile hinge with camera straps just waiting to get tangled with it, and much worse – the camera strap is going over the middle of the LCD, so you can’t see what you need to see while recording.
Unless you’re a vlogger who needs to turn the screen 180° around for viewing from the front, a big usability win goes to Nikon.
It’s been a very long review, so let’s keep this short.
Z8 is an amazingly capable camera which basically comes as a wedding professional’s wish-list of things that Nikon has heard and delivered. (Although they took their sweet time getting here.)
The shutterless design is game changing and probably class-creating in event photography. It will also make camera more durable as there are few moving parts, so forget the dreaded shutter replacements every few years.
Autofocus is the best I’ve tried on a camera, both in performance and implementation of its detection and tracking features. It’s easy to activate and switch modes at moment’s notice, which makes everything even more useful.
Image quality is very high in photo mode, almost on par with legends like D850. You’d be hard pressed to notice differences in noise and dynamic range during real life use.
Video quality is also great, with oversampled 4K being on by default in 25/30fps and the option to turn it on for 50/60fps capture. N-log looks great and is easy to capture because of available live histogram, waveform monitor and zebras.
I also feel like I’m future-proofed with internal 8K60 capture and 12-bit RAW video formats available. It may be on overkill for weddings, but you never know when you may need them for commercial work!
The camera is endlessly customizable (although not completely) and you can really make it work as an intuitive extension of yourself. Almost nothing of note is missing in this department.
The cherry on top of this is Nikon’s promise it will use firmware updates to improve and add new features in the upcoming years. Finally, Fuji’s applied Kaizen philosophy is starting to rub off on both Canon and Nikon.
In conclusion, a wholehearted recommendation for Nikon Z8 for both wedding photography and videography!
( Part 1: Photography | Part 2: Videography )