Это видео недоступно.
Сожалеем об этом.
The Great Z7ii Wildlife Test
HTML-код
- Опубликовано: 14 июн 2021
- I test the Nikon Z7ii for ISO, large and medium file image quality, different focus modes and even how it performs in the rain.
All that and "through the lens" view lets you see what I see.
You're right. I shoot the Z 7ii, now after the Z 7. The Mk ii is a mirage in the sense that I have a vision of firmware updates still to come, that would make me benefit more from the second processor. It is marginally better than the Z 7, today, IMO. Whatever the details, it is a great camera. A couple notes, still, to add to the conversation.
As to megapixels, two things: (1) a sensor does NOT have pixels and (2) MP are an area unit.
(1) Pixels have RGB, each of them and the sensor is analogue and colour-blind. We get colour from that sensor through a filter grid overlay that has either an R, or a G or a B filter over the individual photosites. meaning that a photosite is monochrome because of that filter. What the naive call colour science is in the filter specification to begin with - cut-off wavelengths and rate of filter decay from the cut-off wavelengths. The analogue photosite measurements by the camera get converted into digital in the Analog-to-Digital (AD) circuit. That AD was discrete from the sensor and Nikon-proprietary in an old D700, but comes bundled (a.k.a. stacked) with the sensor these days. But even a stacked AD circuit runs a camera manufacturer AD conversion function (function, as in maths). This is where we have 14 bits gradation resolution. We get pixels from raw processing. With all the artefacts it can cause if done sub-optimally.
Note that increasingly, both influencers, shops and sensor baker are referencing sensor resolution as if these have pixels and this is a total corruption. Also note that "raw processing" is based on colour guessing by interpolation of photosite [x,y] data with its neighbours - but the photosites along the edge of your raw file do not have neighbours. This is either soled in hardware with additional sensor rows and columns that are used for deBayerization, but do not end up in your shots, or the raw processing must have a separate edge algorithm. This is where the term "effective photosites" comes from that defines what ends up in the image visible to you.
(2) Already before WW2, "photography" knew that human perception of sharpness (detail resolution) relates linearly (*) to image/frame size. That is why, e.g. in the 1970s, a film or lens resolution test was specified in a linear unit: linePairs per millimetre (lp/mm).This means for digital cameras with the same aspect ratio, to compare their resolution potential, we have to compare one side of their sensors. If we do that with a 24MP Z 6 and a 45.6MP Z 7, then the difference between a 6 and 7 makes the latter only 37% better - which perfectly explains why we can be disappointed by what we gain from the additional MP. Or it explains why much lower MP cameras can result in surprisingly good photos. Simply put, to get 2x the resolution of a 24MP "6", you would need 96MP. And this is under the ceteris paribus assumption, assuming glass and tech that is good enough for the best of two cameras. There are two factors that colour such a 6-to-7 comparison: the 6 has a fuzzy filter, the 7 has not, which gives the 7 more contour sharpness, but more difficult raw processing. And there is the digital veil over our images in the sense that "digital" easily "invents" pixels when we blow up beyond 100%, creating an illusion of resolution (and how well this is done, is a software thing in the first place, a camera/lens thing second).
(2.1) How are MP relevant? Well, they predict file size and indicate the load to the camera's electronics. Also, it predicts the processing power we need for raw processing. When my powerful workstation PC died on me, I tried to use a 2,000 currency notebook. Ninth gen i7, 16 GB RAM, discrete GPU with its own 4GB RAM and 4K display plus ability to drive two 4K displays. Well, Lr creates previews based on horizontal display resolution and from the "7" files to the 7,680 of two 4K displays, there is an awful lot of processing. The notebook (Windows) did not cut it. It would crash or Lr could not run when some other application was already running. I could forget for Lr to take a smart object into Ps. Under 32GB RAM and under 6GB video RAM you will suffer. So I built a new workstation: 64GB RAM and 11GB video RAM with an 8 core/16 thread 10th gen i7 that can run at over 5 GHz. Working with that is a breeze and importing files from CFexpress cards over Thunderbolt 3 is very fast too. Which is to say that stepping up to D850 or Z 7(ii) resolution may cause you some consequential financial damage.
(*) "Linear" as in linear magnification is the same thing that underlies the "square law" of perspective and the "inverse square law" of flash exposure. Consider a subject frame of 1 x 1 at distance 1 (which sets an image angle). Now move to distance 2 and your frame is at 2 x 2 in the subject. At distance 3, the framed subject has size 3 x 3. The area of the framed subject went from 1 to 4 to 9. MP are like that 1, 4, 9 "area" thing and human "linear" perception of resolution is like the single side of the frame, 1, 2, 3.
Thank you for all that information and seriously well written, even I understood it. That’s a fascinating point about the PCs not coping with the larger files.
@@seancameronphotographic3740 - thank you. It's not "the larger files" but the combination with the 2 4K displays. Raw processing for a 1080p display (1920 horizontal) is a lot less work than for 2*4K at 7680. I guess that computational effort increases exponentially when correct details become more important. I mentioned "deBayerization" and this is the first step in raw processing - the interpolation colour guessing. This is where artefacts are created. Like Moiré or "noise". A second pass called "demosaicking" then tries to remove the artefacts. If you look at older JPEGs where two blobs of different colour meet, and zoom in on the border, then you'll see cross-bleeding of colour from one blob over the edge into the other blob. That's very bad demosaicking. Then there's aliasing that needs to be dealt with. Where deBayerization is mathematically precise and repeatable, it is just wild-assed guessing. If edges of blobs are detected, then this could stop the cross-bleeding. And that edge detection needs some sort of AI like convolution. Which adds to the computational effort. Where Lr processes my raw files on my PC workstation in the blink of an eye, DxO PhotoLab 4 with its DeepPRIME AI needs seconds to deal with what they call "noise" but actually may be raw processing.
I use single point, continuous focus with back button focus both on my D850 and Z72. Never fails on both of them. In my opinion everybody should be shooting with back button focus. 😀
Great point(s) about the focus modes and letting the camera do the work...ive never really been in to all these eye focus setupos , although i can see the value in certain situations...i suppose coming from the school of hard knocks of film photog ,and early digital, ive got used to 'controlling' where my focus is...also using Hasselblad systems where you aim at the eye and recompose...im justn looking at moving sideways to the Z7ii from D850 and XT4 before a Z9.. great Vid Sean
Well done! I find myself switching between 3 or 4 focus modes consistently. Agree with your bit on "the perfect photo."
Thank you Chris
Great photos
Many thanks
Yeh found same, on D850 it was simply easier to get what i want and faster, however Z i find i am swapping Focus points, we know we got updates to come from Z9, so lets see this in a few months time when the firmware changes trickle down to the Z6II/7II. I hope the Z9 delivers for us all in the brand, as the market needs there to be 3-4-5 brands to keep the prices down and competition up.
I don't think any camera has had so much hinging on its performance out of the blocks than the Z9. With Nikon saying that the D6 was their last pro DSLR they've gambled hugely. People like myself, that trade in our top cameras every few years due to the work they get, will face a real dilemma if the Z9 is not as good as the D6 /D5. Sell up now and jump ship whilst the gear is still worth something to trade in, or keep buying older technology whilst it's still available and hope.
I love that all us Z camera owners are waiting for the promised firmware update, I hope it's all we are promised.
As a lifelong Nikon shooter and hoping Nikon would catchup in the mirrorless world - it has not, I do keep hoping. So, I switched to Canon and have been blown away to the point of selling all my Nikon stuff except my old D800, 300 F2.8 and 500 F4. I keep searching reviews hoping that that Nikon is in some way redeemable in the mirrorless world so watched this video.... and they are not. Great video, but again reinforced why I switched. Even at ISO 6000+, the Canon files are clean. I was thinking that I may be wrong and went out with me D850 and felt like I was carrying an anchor. I LOVED that camera until I picked up the Canon 90D. Again fantastic video, and although I feel guilty about leaving Nikon whos color I still prefer, but we all want more tech and Nikon is still stuck so far in the past. I don't think the Z9 will be a catch up either.
As I keep saying, the Z9 needs to be a great camera otherwise a lot more people will abandon them unfortunately. I share your pain at the moment and await the news from Nikon with a huge amount of trepidation. We need that Z7ii / Z6ii firmware update too.
@@seancameronphotographic3740 Sean, thank you for the reply!! The leaked specs on the new Z9 are impressive, but not compared to the the R5 or even the A1 (I cant stand Sony color). I was shooting landscapes with my old D800 this weekend - so amazing, but for wildlife, fast action color and amazing eye detection..... Canon. And, at a lower price! I think the Z9 will be over priced for most that can get an R5 or R6 that will have the same or better performance. Ah well...... I am having trouble selling all my old Nikon stuff now too.
Very informative
Thank you, it was fun to do. Now to see if the Canon and Sony are as fun
Excellent content very well presented. Can you please elaborate in the rain issue? Also hard to tell on RUclips what was your perception on Z6ii image large vs z7ii image at medium? Thanks looking forward to more videos like this!
Thank you Karen, yes the rain caused the sub-selector (as the manual calls it) to malfunction slightly. When I was changing focus modes, switching right caused me to leave the menu. I thought I was being clumsy so repeated it a few times and then got another photographer to try, with the same result.
I had a similar issue in the rain with a D4s a few years ago which resulted in a repair to the said sub-selector. This time luckily it started functioning properly after a wipe down and a tea in the cafe, (the camera then me :)
I have a Z6ii and can hardly remember all of the focus modes, never mind devise a strategy to evaluate each and choose a favourite subset. If you develop such a strategy, would you please do a video? I’d be grateful.
Now there's a challenge Edward. Let me have a think
Hello could you please tell us what your computer system is with how much storage you have for your files? Thanks.
No problem, I have three computers, an old Acer T3-710 for emails etc, A Concept D running through a SAMSUNG - C34J791 Quad HD 34" Curved LED Monitor and a MacBook Pro running through a 27" Thunderbolt Display monitor.
Finally, all my files are stored on pairs of external hardrives
Z7 has 64 ISO and Better Res and No Anti Alias Filter and the Z7II has 900 Secs long exposure over the Z7 (My Camera 30 Secs ). I am not into Video and so the Z6 family was never considered
Yes there is a slight res difference but it’s virtually identical, I don’t think either versions have the Anti Alias Filter but that’s fascinating info on the long exposure Peter, thank you.
A bit Confusing about the tracking mode. The z cameras cannot and Nikon don't claim they can track except for Human faces dogs and cats. You were lucky that it thought the Lemur was a dog or a cat. My z6ii thought the road wheel in a car in my viewfinder was a face with the wheel studs which were at 10 & 2 oclock were the eyes. Had to smile at that. Your technique of using single point af is I think the best way to try and photograph wildlife that is more or less still.
To be honest Edward, I'm amazed that it did so well with the meerkats, it's a clever system.
I love the story about the wheel nuts, that made me laugh and I think we agree that the best way to focus on (relatively static) wildlife is single point.
As for tracking, even the Z7ii manual talks about animal "subject tracking" but I concede that there is a difference between that and eye recognition.
I hope you agreed with me towards the end that this is not infallible, and it would be pretty boring and unfulfilling if it were.
@@seancameronphotographic3740 I do understand your point about making photography boring etc. But there are times when all the bells and whistles make a difference. For one example I went out to do some wildlife down near the river where I live and I had been after an in flight photo of a Heron that frequents the area. I walked round some bushes into the perfect spot to see it on finals to one of its spots and I duly tried to get a shot and out of about 15 or so shots not one was in focus. I did at that moment wish I had my D500 as that would have nailed it. Maybe Nikon will fix that issue hopefully soon for BIF.
Absolutely, I actually said in one of my previous videos that the eye tracking sometimes got images that even my D5 couldn't, such as a dog suddenly grabbing a ball or tug toy with a narrow depth of field. Often the eyes aren't quite sharp but the ears are. It was easier with the Z6ii. However I see that as a bonus, I wouldn't right the camera off because it could get it perfectly every time.
Interestingly, I have the same issue with the wheel nuts. Trying to photo my 6 year old with the Z7ii. The eye recognition often chooses wheels over humans.
Nikon z7ii is not d850 good I stay with my Sony A7riii