With the 14-24mm f/2.8 G, Nikon established a benchmark for wide-angle zoom performance. This lens is still considered to be the best on the market, a better choice for architecture than, for example, a 20mm f/2.8, and Nikon full-frame camera users have been abandoning prime wide-angles for this all-in-one solution. It has some disadvantages, however, and one of these is the deep curvature of the front element, which relies on petal-wing lens hood protection and cannot easily accept front-mounted filters.

The specification of the new 16-35mm f/4 G ED VR sounds attractive. Such a lens might be a fraction of the size of the 14-24mm, it’s a stop slower, and 2mm extra length on the wide end must reduce the complexity. But when it arrived, I was surprised. It looks not so far off the size of the 24-70mm f/2.8 companion to the 14-24mm, at 125mm long with a 77mm front filter thread.

And there lies one of the main reasons to buy the 16-35mm instead of the 14-24mm – you can fit any good quality slimline or wide-angle design polariser, or use a suitable adaptor for graduated ND and similar system filters. The 14-24mm is very large indeed, with a 98mm overall diameter, though any extra length compared to the 16-35mm comes from its built-in lens hood. With its lens hood fitted, the 16-35mm is 158mm long compared to the 14-24mm’s 131.5mm.

There is a smaller Nikkor with a similar range, the 17-35mm f/2.8 IF-ED SWM, but it dates from 1999 and the pre-digital era. It lacks Nano Crystal coating, the use of new glass types, and the VRII vibration reduction of the 16-35mm.

Nikon’s new preference for this extremely long, slim optical train is clear from all the newer lenses, compared to those of 10 years ago. The design is able to deliver better quality to the outermost zones of the full-frame digital sensors.

So, this is not a small lens – in fact, it’s surprisingly large, and a bit daunting for the subject. And it’s not a fast lens at f/4, but more than adequate for normal situations. It is light, as the optical unit is compact within the mount. The key question is whether this lens is another 14-24mm – a class leader.

During the test period I was in contact with a professional who felt the results were slightly softer than some of his trusted Nikkor lenses. With the benefit of raw files from three samples of the lens, we pinned it down to micro AF adjustment, sample variation and the general look and feel of D700’s 12 megapixel images.

My sample was just a notch below the 14-24mm quality, but both the D3x and D3x are very different cameras from the D3 supplied for this test, and it was on these later bodies that I last worked with the 14-24mm.

Nikon appears to have optimised this lens to be used at any focal length and all apertures from f/4 to f/11 with high performance across as much of the full frame as possible. If the focus is perfect at f/4, there is little improvement on stopping down, even if the image never has the maximum bite associated with some Nikkor designs.

The 16-35mm range may be only just over 2x zoom, but in the wide-angle range, that's a big difference. Image (c) David Kilpatrick.

Distortion is well optimised – much better than the 16-35mm f/2.8 designs from other makers. Corner sharpness suffers sudden death at 16mm wide open, just as it does with these other f/2.8 lenses, but it’s a clean death and affects only the extreme couple of millimetres. Once stopped down a bit, you can barely detect this softening.

Mechanical vignetting has been eliminated. Beyond optical laws that can’t be overturned, there is no hint of corner shading. Contrast is not especially high, despite excellent control of flare.

I appreciated the 16-35mm better after comparing it with a 17-35mm f/2.8-4 I use occasionally. At 17mm and f/4, this cheaper lens is marred by aberrations that extend well into the DX 1.5× crop format central zone (it’s not really usable at f/2.8). The Nikon lens is free from visible aberrations, even over full frame, at 16mm, at f/4. I needed no CA correction during raw conversion.

In addition to testing the lens on a D3, I checked it on the 12 megapixels of the D5000, which demands more of resolution. Interestingly, it didn’t look any softer on the D5000 than it did on the D3. So it should cope with a 30 megapixel full-frame sensor.

The VR function gives a great deal of freedom. With such a wide angle of view, there are interesting shots to be taken with extreme depth-of-field around f/16. Switching on VR means you can handhold down to a half second in theory (that’s four stops relative to a “safe” 1/30s).

Since no previous wide-angles have had VR, I tested the sharpness at a high shutter speed wide open with VR on and off. If anything, the results with VR on had the edge. With this lens, VR is not harming sharpness.

It’s a dilemma. The 16-35mm is not much less than the 14-24mm; it doesn’t have the ultra-wide short end, but 16mm on full frame is already hard to handle well. It takes filters, it’s light, but it is not pocketable. It has VR, and no other wide-angle or wide zoom offers that.

Had this lens been introduced before the 14-24mm, it would have won that “best in class” rosette; if the price was lower it would be easier to recommend. As it stands, around £995 for this lens versus £1295 for the 14-24mm, I would try to find the extra £300. Then again, the comparable 12-24mm f/4 G for the DX format (equal to an 18-36mm full frame) is more than £800 and the 16-35mm seems entirely in proportion with that.

And I would need to think hard about filters. I’ve almost abandoned using them because of the inability to fit them to my wide-angle choices, either in any way, or without visible cut-off corners from the slimmest filter. The Nikkor 16-35mm is filter-system friendly, and that may be its strongest unique selling point.