Have you ever tried to imagine the outcome of a battle between two armies whose weapons and military strategies are separated by a century of development? Having recently got my hands on Nikon’s first full-frame flagship DSLR, the famous D3, I decided to try a similar experiment. For its competitor, I chose the Nikon Z9. True, these two rivals are separated by 14 years rather than centuries. But at the current rate of technological development, it might as well be!
So, what are the similarities and differences between the two cameras? First of all, they both represent the best of what Nikon is, and was, capable of. The 12.0MP Nikon D3 launched in August 2007 at a price of $4999. It wasn’t until the following year that the Nikon D3X saw the light of day. The latter had a gargantuan 24.4-megapixel resolution (well, for its time!) and an equally high price tag of $7,999. The Nikon Z9 launched more recently, in October 2021, at a price of $5,500.
Taking inflation into account, the Z9’s price is pretty similar to that of the original D3. What is dramatically different is the resolution, which has increased 3.8 times in fourteen years. Such an increase can be compared to the increase in the capabilities of the human brain over the last few million years, and its decrease over the last ten :)
To make my comparison really tell the story of where technology has gone in those 14 years, I’ve equipped each camera with a wide-angle lens from its era. On the Nikon D3, I used the Nikon 16-35mm f/4.0 AF-S G ED VR lens, launched in 2010. The Nikon Z9 got a similar, but 9 years younger, Nikon Z 14-30mm f/4 S.
But let’s not waste any more time and head out into the field to see how it all works. For my short test, I chose the Jizera Mountains, a remote corner in the north of the Czech Republic. The well-preserved spruce forests, mountain climate, and peat bogs here give the landscape a magical and somewhat mysterious character. On the map, I picked out a small stream away from the tourist trails. After a tortuous descent, during which I refreshed myself with blueberries, I stood at the bottom of a magical canyon. At its bed, a beer-colored stream roared through the rocks. How typical of my country.
My test could begin. The design was simple in principle. Not the sophisticated scientific approach you are used to in our reviews. I set up the composition with one camera, took several variations of the same image, and then repeated the same thing with the other camera. The aim was to get pairs of photos where the main difference would be the camera/lens pair used.
In reality, there were three other variables at play. Firstly, the minimum ISO of the two cameras is different. On the D3 it’s ISO 200, while on the Z9, it’s a much lower ISO 64. This affected the shutter speed and therefore the amount of blur in the water and vegetation. There was also a significant difference in the quality of the polarization filters I used. On the D3 with a 16-35mm f/4 (77mm diameter) I used my high quality VFFOTO filter with 14 anti-reflective layers. On the Z9, I used a borrowed cheap no-name filter because the 14-30mm f/4 requires an 82mm diameter. The difference between the two would make a whole article. (So much so that after the first two shots, I preferred to simply hold a quality filter in front of the 14-30mm lens.) The last major factor that was difficult to control was the weather. Sun alternated with heavy clouds, wind with windlessness. But the biggest problem was the heavy rain that started towards the end of the day.
Whipped by the rain and gusts of wind, I stood on high ground overlooking the bog. As soon as the wind calmed, I removed the lens cap and took a few shots. After a few seconds, the front of the lens became waterlogged, so I dried it off while shielding it from the rain and repeated the process. For this foggy scene, using a polarizing filter didn’t make sense. Thanks to this, the sharpness and contrast were not affected by the bad filter, nor by the vibrations caused by holding the good one in front of the lens. This is where the differences between the two cameras really became apparent.
Well… maybe at web resolution, the differences aren’t apparent at all. Instead you can judge from the 100% crops:
A 100% crop on one camera is not the same as a 100% crop on a lower-resolution body! Perceptually, the sharpness isn’t all that different, but the greater number of pixels on the Z9 means that we are seeing that sharpness much more zoomed in.
And here are two more 100% crops from the same scene to show what I mean:
After returning to my cabin and drying my wet clothes and cameras by the fire, I went out again. It was late at night, and the supermoon had just risen in the sky. Not exactly the ideal situation for star photography, but at least the rain had stopped, and there were some gaps in the clouds.
In the torchlight, animal eyes in the forest shone out at me from the darkness. A fox was running not far from me. A doe, which would have fled during the day, walked around with dignity. A toad was sitting on a stump and two deer were grazing in a meadow surrounded by tall trees. I didn’t notice this until I was processing the photos. It was so dark in the forest that I couldn’t see the tip of my nose. See if you can find the deer in the Z9 photo! Without a lot of brightening in post-production, I expect you can’t.
And here are the crops, if you’re interested:
At a pixel level, there is a bit more noise in the Z9 photo, but there is also far more detail in the leaves and grass. I remember how impressed I was with how well the Nikon D3 handled high ISOs at the time. It’s still impressive from a noise standpoint, but a lot of details are lost in the low resolution. If both photos were printed at the same size, the noise levels would be similar on both cameras, but the detail would favor the Z9.
Two more crops:
For a camera that would be old enough to get its driver’s permit, the D3 holds up really well, but this was never a fair fight and of course the Z9 comes out ahead.
Summary
The Nikon D3 was a fantastic camera in its time, and it still is. Holding it is still a pleasure, and a look through the viewfinder will make you reminisce about the old golden days of photography. But technological evolution has worked its magic, and modern high-resolution sensors like the one in the Z9 are clearly better. Not to mention the other benefits that come with time, like processing speed and video features. Does it make sense to buy a D3 now? Pragmatic arguments are hard to find, except for price. It is amazing how you can get a used full-frame professional camera like the D3 for about $400 used – less than most entry-level mirrorless cameras.
The same goes for the lenses. Especially with wide-angle ones, the move from the somewhat narrow F-mount to the XXL-sized Z-mount has meant an overall improvement in lens quality. However, the real benefits are only apparent at wide aperture values and large prints, or if you expect to do a lot of cropping. Lenses also go obsolete more slowly than cameras. The 16-35mm f/4G that I mentioned is selling used, today, for almost exactly the same price as the D3 is. That’s still a huge decrease from the original price of $1260, but it’s not a 90% decrease like we see with the D3.
What about you, do you have an old digital camera sitting in a drawer? A mate from the old days? Then take it out and stretch its shutter. The mirror will still clap happily, the shutter will rattle, and you might remember your first digital steps. Isn’t it fantastic that we lived through those times?
I have my D3s still with me along with a 14-24 and a 24-70. The camera was a gem and still a gem. Its color reproduction is only comparable with canon original 5D which I had long ago. Now I use sony A7RM4. From film to dslr, natural color has lost its essence. In this digital era, we are all getting ahead towards the fake surroundings which may have shiny and gorgeous from outside but have no true feeling. Eventually AI will get us all.
I still own the D3 (great camera), but have since moved away to Sony mirrorless with the A1 and will never be going back to pointlessly massive cameras. On that topic the XXL sized mount of the Nikon Z cameras has long since proven to have no correlation with lenses image quality potential.
I think a comparison between those cameras should include the same angle of view, exposure time and aperture with an ISO variance of no more than 5%.
Comparing 100% views does not take sensor development into account.
Otherwise, a nice comparison that encourages me to ‘de-mothball’ my old treasures again
I’m completely not surprised that the D3 still looks great with that 12mp sensor from that time. This is a weird comparison of the evolution of flagship camera’s though because we’re shooting static shots. The real difference is not at all image quality, the F mount and D3 are good enough there for 99% of peoples needs today, but rather in autofocus and ability to shoot fast. Let’s ignore video for now but the D3 is obsolete for that and it’s a modern requirement in cameras. Back to stills. I would love to see a comparison here of trying to get action shots or birds in flight. Let’s see a comparison of kingfisher keeper shots and sports keepers from a week of shooting between the D3 and F mount and the Z9 and Z mount and see what the end results look like. If technological progress is to be believed the Z9 should get many x the moments to choose from and the end result should be a larger volume of “keepers” to choose from and hopefully more superstar images consistently. Thats what a professional would really be paying for in all these updates and upgrades. Image quality has been great for a decade plus now.
Have to disagree about the golden age of Photography. That was a Nikon F2a 50mm 1.4 with motor drive Trix Film Develop and Print it yourself awesome
Unfortunately, because update cameras are very expensive, I could not buy a full-frame camera. And I had to buy a D3 camera. Of course, the camera is low shot. Now, reading your article, I hoped for my camera. Thanks
The new technology is not everything. Z9 have more features, but doesn’t have the same character and the feel while you are out and shooting. Z9 is cold electronic box and images are very flat. I do own nearly every single DSLR Nikon ever produced and I do have complete collection of D the Pro line from D1 to D6. My most loved cameras are D3, D3s, D4s. Even today, I do use them for my wedding work. In my opinion this was the pick from the Nikon, when they produced the best sensors with the best image quality and body build. My two D850’s are benched for a third year now. Just because of high resolution sensors and the image quality is nowhere near to D3, D3s, D4s..Will I upgrade to mirrorless? Probably never.
Those Z bodies are not the same and I don’t think they ever will.
What about d3x?
Very good article Andy and comparison. Love the shots presented and see little if any difference between the 2 cameras. What we see mostly is improvements in features and ISO range.
I followed your test but used a D30 vs R7. Fortunately both use seamlessly the same lens with all functions of the lens working 100% perfectly on either camera. I saw a good improvement in dynamic range and resolution but colors were indistinguishable.
And yes it was fun looking through the optical viewfinder. But I have been spoiled by the EVF and will never go back.
Oh I want a Z9 or a Z8. It is just so hard to justify the expense.
Although I spent many hours every week taking photos, it is only because I love the experience of doing it. It is my favourite way to pass time.
Maybe 1% or less are shared on Flickr or Instagram. I will buy a Z6 III the moment it comes out (if it has the expected improvements in focus tracking). But I do love my D3s D800 and Z6.
Interesting but not really a meaningful comparison. The settings are a mile apart, with the D3 set for saturated images with shadows evident and the Z9 set to look very HDR-like (which is probably why there are posters indicating that they prefer the look of the D3 images).
And also a technical error – the Z9 doesn’t have quite 2x the resolution of the D3. It takes 4x the pixel count to double linear resolution. You compared pixel totals, not the difference in resolution.