I got this to use with my Nikon N90 in 1994, when I first went into autofocus. It had a very high reputation then and since, but some of my professional contacts on mailing lists reported not-so-good performance from newer and supposedly better Tokina zooms.

I was considering sticking with this as my main lens for a while after I got the D700 (which brought me back into full-frame photography), but this required testing.

Results seemed pretty conclusive to me.

28mm f/2.8

Those corners are really pretty appalling. This was a reshoot, because I was worried I’d somehow done it wrong in the initial shoot.

 

 

 

 

28mm f/4

Some improvement; but more, some reason to believe that the f/2.8 result wasn’t a mistake somehow.

 

 

 

 

50mm f/2.8

Considerable improvement in the corners; not so much in the center, I almost wonder if that was me? This is why sane people use tripods for lens testing, I believe. But that one was at 1/1000 sec., it really doesn’t seem likely it was camera shake.

 

 

 

 

50mm f/5.6

There we go, that’s about where it’s really supposed to be.

 

 

 

 

70mm f/2.8

Funny-looking at the long end, too; more in the middle really. Coma maybe? I haven’t done serious pixel-peeping at these yet.

 

 

 

 

70mm f/5.6

Nice in the center, okay at the corner now.

 

 

 

 

Conclusion

So, as you may have guessed by now, I’ve ordered the Nikkor AF-S 24-70 f/2.8.

My first-ever “major brand” zoom lens (though in fact Tokina, Tamron, and these days Sigma are “major” by any reasonable standards). So my first zoom lens made by a company that makes camera bodies. No, wait, Sigma makes camera bodies (and Vivitar did, back when they were “major”, may still). Anyway, my first zoom lens carrying the same brand as my camera does.

I got this because the Tokina 28-70/2.8 and the Tamron 28-105/2.8 were too long at the wide end to be comfortable as my main lens on a DX (1.5x crop factor) body. I had the Nikkor 18-70 kit lens for a while (bought used), but I got too much flare in situations that shouldn’t really have given me flare, and I fairly quickly bought this and sold the 18-70 on. This was a fine lens for me, and I regret selling it. I’ll regret it even more if I end up with another DX body after all; but I can’t afford to hold it unused while spending the money for the 24-70/2.8 that I’ve ordered.

Testing this is aimed at providing good examples.

I shot these tests on the D700 in DX mode, so the originals are only 5.1 megapixel images.

17mm f/2.8

This ought to be the worst case, everything at its widest. Not bad, eh?

 

 

 

 

17mm f/8

Still not truly wonderful in the corners; in fact not much improved from f/2.8. But I think that’s more to the credit of f/2.8 in this case.

 

 

 

 

55mm f/2.8

 

 

 

 

55mm f/8

Mmmm, sharp. Well, still far from perfect in the corner.

 

 

 

 

Officially it’s “AT-X 124 AF PRO DX”. Their first shot at a crop-factor DSLR lens, and quite a big success; lots of people on the net like it for example.

A previous round of test shots when I first got this lens is here.

And why am I testing this when I’m leaving DX and about to sell it? Well, I needed to test my old 28-70, and I decided I should include other lenses as sanity-checks, so I’m looking at more than one set of results.

 

 

 

 

12mm f/4

So not so shabby really, I don’t think.

 

 

 

 

12mm f/8

There really isn’t anything much wrong with this by f/8.

 

 

 

 

24mm f/4

The 24mm end is pretty fuzzy in the corner too, at f/4.

24mm f/8

Pretty nice, especially in the center. Possibly better than the Nikkor at f/5.6, even.

 

 

 

 

 

 

I’ve had this since the Australia trip; is that 1983?  I’d gotten out to 28mm with the Vivitar Series 1 28-90 zoom, and was finding myself using the wide end of that a lot (I’d previously had a 28mm for my Miranda Sensorex, which I never liked and rarely used; I might have had one in the Pentax system too). It’s an old manual-focus lens, but it was thought to be very good at the time.  How will it stand up on modern digital?

These are somewhat cluttered — the wall wasn’t really wide enough (I guess I should have walked in closer really), so there’s stuff overlaying the bricks in the edge shot. Try to ignore the out-of-focus foliage!

And I clearly wasn’t considering this carefully, because I only shot an f/5.6 test of this one.

The results aren’t stellar; definitely soft, even at f/5.6, even in the center. Well, it’s only one test image, perhaps I focused badly or held unsteadily. Or not. On the mental list for consideration.

Apparently I should test my lenses semi-carefully more often. I don’t actually think I’ve particularly damaged this one in the last decade. This would explain a sort-of malaise about the more ordinary photos from much of that period.

On the other hand, it does make the decision about upgrading my main zoom pretty simple. So I’ve just placed that order. Yikes, modern good lenses are expensive! And either my memory is going, or it cost me $200 to wait for the D700 to arrive and for me to run some tests on the Tokina 28-70 ATX-Pro f/2.6-2.8 before ordering its replacement.

Philosophy

My lens-testing philosophy: don’t try to make meaningful quantitative measurements. Do try to take some comparable pictures under moderately controlled conditions. Be prepared to repeat experiments as necessary. Don’t expect sets of test shots from one year to be much use in making comparisons to independently-shot tests a few years later. And, the new part: do it more often!

Unless I’m testing for some very specific situation, I like the general approach of taking pictures of a flat surface that I can reasonably easily line the camera up flat to. It should have considerable high-contrast detail all over it. It should be big enough that I can shoot from 10 feet or more away, usually (close-focus is sometimes a problem area; then again for some lenses that may be exactly what I’m testing for). Something with rectangular detail all over makes it fairly easy to line up square to it. Brick walls are good, and newspaper pages taped to the walls (for closer-range work).

Look at the center, and the corners.  Sometimes other places on the edges, especially if looking for problems like whether the sensor plane is parallel with the subject.

If I were better at looking critically at my normal pictures and seeing the sources of problems with them, I might not need to do this.

For most of these, I’m posting a link to the full test shot, full-size but in JPEG format; the center shot, at 100% pixels (the file; what you initially see here is a thumbnail), in PNG or jpeg format, and the corner shot the same. All of these were shot in raw, processed in Bibble Pro, with no sharpening and light density correction. And I see that some of the photo sizes have been messed up by the upload process. I have to say that this WYSIAYG editor and the “convenient” automatic upload procedure for attachments and so forth really makes it tremendously much harder to write a complex post like this.

Given the risks of editing, the complexity, the time, and all that, I’m going to do this as separate posts for each lens. This will make some comparisons across lenses harder; so it goes.