Ratings for Current Nikon-Based Digital SLRs

updated: 7/26/07 (note that some of previous ratings have changed to keep up with state-of-the-art)

An email I received this week made me realize that, while I have a lot of specific information about Nikon-based DSLRs on the Web site, I've never really summarized my thoughts about the current crop of bodies in one place. (If you want specifics, go see the review for the particular model elsewhere on this site.) I won't vouch for 100% consistency in opinion across the site as firmware updates and added shooting experience keep making many of my opinions moving targets; it's also difficult to synchronize everything now that the site has sprawled to the size it is. If you see something that confuses you, drop me an email and ask for elaboration.

The tables below are filled with rating words that run in graded ranges, such as this:

BEST, very low, low, moderate, high, very high, WORST

or

BEST, excellent, very good, good, fair, poor, very poor, WORST

If you see a + following a rating, that means it's a bit better than the word would indicate, but not good enough to make the next better 'grade'; while if you see a - after a rating, that means that it's not quite that good, but not bad enough to make the next lower grade.

Some of the terminology in the image quality tables needs description:

Sharpness versus detail: Sharpness is the ability to clearly define edges, while detail is the ability to resolve small objects. Obviously, they interact, but in general you need sharpness before detail, and there are clear differences in the final image when these items vary.

Color fidelity versus saturation: fidelity means the ability to resolve a color as it really is, while saturation is the ability to record a color with contrast and "snap."

Artifacts can be all sorts of things, but in general would primarily be moire, color fringing, stair stepping, and compression artifacts.

The "Overall" ratings are not an average of the other individual ratings, but my assessment of the impact of all the individual nuances taken together. For the intended use, I would choose to use a camera rated good before I would use one rated fair, for instance.

Let's deal with the current Nikon DSLRs first:

Shooting JPEG files, here's a quick and dirty summary of the image quality differences as I see them:

Overall, if someone told me I had to shoot JPEG images for publication tomorrow, I'd pick the S5 Pro first, despite some issues. First, the S5 Pro has very pleasing (but not accurate) color, and the film simulations allow you to create very nice out-of-camera JPEGs easily. None of the Nikon bodies do a poor job at JPEGs. Indeed, there's very little to distinguish between them in most categories. The D200 does have a tendency to produce slightly soft JPEGs (though they sharpen up fine). If you don't need resolution, the D2hs also produces remarkably nice JPEGs, though with a bit of color drift. The D2xs is great at JPEGs up to ISO 400, but noise becomes a factor abve that. Personally, my casual JPEG camera of choice is the D40x, which is a nice light, convenient, and high-quality product that doesn't cost much.

For RAW images, the table changes:

RAW is what you want to shoot if you can. The trouble is that RAW workflow can end up more elaborate and time consuming. In terms of trying to produce the best possible image, however, you have no choice--in my opinion RAW is the only way to go. Here, my first choice is a pair: the D2xs and the S5 Pro. I'm consistently startled about what I can get out of the D2xs, at least up to ISO 400 (above that and having to do noise reduction starts to impact several of the variables). At ISO 100 I'll stack the results out of my D2xs NEFs against cameras that have higher pixel counts, such as the 1DsII. I've written elsewhere that good D2xs images just have a 3D like quality to them, and the only way to see that is to use great shot discipline, good lenses, and shoot raw at ISO 100 or 200. The S5 Pro also gets a nod and a place in my bag because of its unique ability to handle wide dynamic ranges with aplomb. Curiously, overexposing an S5 Pro raw file is probably the right thing to do, and with the right exposure and conversion process, you can get stunningly good images out the camera. All the other Nikon bodies come out in a narrow range of very good. I'm not afraid to shoot with any of them; they just don't quite get to the level I can achieve with my first two choices.

For Handling issues, things change yet again:

You'll wonder about several things in the chart I just showed. The hot temps issue with some of the Nikons has to do with the propensity to produce hot pixels. I can also say that I've used the D2 bodies for weeks at a time at under 0 degrees farenheit and can safely say they are remarkable at low temps.

Finally, for a few other important abilities:

Now let's deal with the older Nikon DSLRs for those of you looking for a used bargain (I've thrown in the Canon 20D as a comparison point):

Shooting JPEG files, here's a quick and dirty summary of the image quality differences as I see them:

Overall, if someone told me I had to shoot JPEG images for publication tomorrow, I'd pick the S3 Pro first and the D70 and 20D a close second (if the D2h had more pixels, it would be my second choice). None of the others would be fully acceptable to me in this respect, as I'd be giving up something significant. For example, with the D100, sharpness is a real issue shooting JPEG, while with the Pro 14n the write speed can actually be slower than RAW, the buffer size is the same, and the image quality often produces artifacts that aren't in carefully managed RAW files. The S2 Pro can generate dreaded color fringing on very fine detail, while the D1x can require post processing for exposure/contrast if you don't use Custom Curves. I expect the D2x to be better than the existing bodies for JPEGs when it appears.

For RAW images, the table changes:

RAW is what you want to shoot if you can. The trouble is that RAW workflow can end up more elaborate and time consuming. In terms of trying to produce the best possible image, however, you have no choice--in my opinion RAW is the only way to go. Here, my first choice is less clear. The D70, D1x, S2 Pro, S3 Pro, and Canon 20D are all very good performers, and the Pro 14n (now SLR/n) can, in the right circumstances do even better. In practice, I tend to choose the D70 and 20D over the others, but that's usually for handling and other reasons than absolute image quality (the Nikon 1/500 flash sync speed is sure nice on bright days when I need fill flash, for example, and the Canon 20D is just a nice body all-round). Again, I expect the D2x to be the camera of choice in this list once it appears. Unfortunately, there's another issue that intersects with RAW use, which is the RAW converter you use. I've prepared a separate article summarizing my thoughts there; just remember you can't simply pick the best camera without also considering your converter and computer platform.

For Handling issues, things change yet again:

There's a couple things in that chart you might wonder about. For example, why is the Pro 14n only rated fair at hot temps? Well, that's because I find that noise can rear it's ugly head in hot temps and the camera will sometimes trigger multiple Recalibrating as it warms up (especially true if you use a Microdrive). The delay that causes can be a nuisance to some types of photographers, though other than added noise, it doesn't bother me much. Another one that might raise eyebrows is the S3 Pro is poor for verticals. But they added a vertical release! Unfortunately, there are no controls on that release and it won't wake the camera from sleep, which makes it completely ineffective. Sensor cleaning on the Nikon models is problematic due to the frame that extends over to the sensor edge--it's a tight space and easy to dislodge grit off the frame onto the CCD you're cleaning!

Finally, for a few other important abilities:

Some of these ratings will be controversial, so I'd better explain them. Autofocus is easy: the D1 series has a better autofocus system than the N80-derived bodies, period, and the D2h beats everything. Motor drive is a similar situation, but the S2 Pro doesn't seem to make it up to the claimed frame rate, in my experience, the D100 does a little better than an N80, the S3 Pro drags down when RAW or dynamic range is increased, and the Pro 14n seems dead on to the N80 (all D1 series do better, the D1h particularly so). The S2 Pro's ability to cancel pre-flash puts it squarely above all but the D2h and D70, though note that you can only reliably use the S2's flash in TTL at ISO 400 or lower. The D70's flash capabilities are the best and most accurate of all, though, assuming you're using SB-600 or SB-800 flash units (the D2h is a very close second, being dinged only for not having an internal flash that can control external wireless TTL flashes). You'll note that I've rated the S2 Pro's exposure accuracy as "varies with body." I've now had experience with almost a dozen S2 Pro bodies, and there seems to be no consistency amongst them. Moreover, my own body came back from being fixed by Fujifilm and now shoots at almost exactly one stop different than it did before it started to have problems. Fortunately, all the bodies tend to be consistent onto themselves, but I really don't like have to keep my exposure compensation set to -1EV on my S2 Pro as my "zero point."

Mac users should be particularly careful about selecting cameras. Two issues come up: tethered shooting/downloading and quality/performance of the supplied Macintosh software. In general, Firewire cameras do better on Macs than USB cameras. Moreover, Nikon's Macintosh software is actually pretty mature, and Kodak's seems to have been developed there first and then brought to Windows. Fujifilm's Macintosh offerings seem poor in comparison. Overall, software performance tends to suffer on Macs, even for Nikon's offerings.

As I noted before, if you want more about a particular camera, see the review for that particular camera.