How lenses are tested
new Tokina AT-X 70-200mm f/4 PRO SD IF FX Lens
We use an image-analysis application called
Imatest Pro to test the resolution of lenses. Photos of specially designed test
charts are taken in controlled and consistent conditions at a range od focal
lengths and apertures. These are analysed for sharpness at the centre and edge.
The average of these two figures is used for the final resolution figure.
Resolution is quoted in ‘line widths/picture height’, which is the number of
lines that can be resolved within the height of the picture. This is different
to ‘line per millimetre’ figure, because they only apply to a single negative
size/format (typically 35mm). Digital cameras have different sensor sizes,
which mean different enlargement ratios for prints and hence different ‘lpm’
requirements for lenses. Line widths/picture height measurements sidestep this
and relate resolution to the final image instead.
Lenses don’t just ‘stop’ resolving
progressively finer detail – they resolve it at lower and lower contrast. In
the past, photographers have disagreed about when detail becomes too ‘soft’ to
count. Imatest gets round this by using modulation transfer frequency (MTF)
analysis and can define a cut-off point for resolution. This is called ‘MTF50’,
or the point where the contrast falls to 50%.
The resolution figures are also dependent
on the camera used. Different SLRs have different resolutions, different
strength low-pass (anti-moire) filters over the sensor and different processing
and sharpening algorithms. The same lens tested on different cameras will yield
One of the lens’s strengths is the
lower-than-average barrel distortion, but it is still noticeable with straight
lines at the borders of the images.
At 17mm this lens has good contrast and
colour, the descriptive power is impressive. It’s a slightly different story at
the other end of the zoom range though.
Ghosting and flare
With sun striking the front element
obliquely, the 17-35mm is prone to heavy veiling glare (flare), and ghosting.
The supplied hood is a must.
The convex elements and built-in lens hoods
of the faster f2.8 versions defeat usual filters. The modest 17-35mm is more
practical, though the filter ring is 82mm.
The Tokina has excellent central resolution
but the edges fall behind even when stopped down. Viewed at normal distances,
this lens has pleasing sharpness.
Ultra-wide-angle lenses are prone to
lateral chromatic aberration, while this lens is well-corrected there is still
some coloured fringing.
Model: Tokina AT-X 17-35mm f4 Pro FX
Web: www.kenro.co.uk or www.tokinalens.com
Angle of view: 103.9-64.74 degrees (Full frame)
Max aperture: f4
Min aperture: f22
Min focus distance: 0.28mm
Mount: Nikon F (G). Canon EOS (EF)
Filter size: 82mm
Metal and plastic components are used for
the barrel and control rings but the build quality and design is hard to
beat. Focusing and zooming are both internal, so there are no extending
The f4 maximum aperture is adequate but
it is constant and on a par with rivals. There are the usual aspherical and
SD (low dispersion) glass elements but it lacks the fancy image stabilization
of the rival Nikon and has a shorter zoom range than the Canon
Quality of results
Good performance at the wide-end,
especially in terms of contrast and central sharpness, though it’s a
different story at 35mm where the image lacks contrast and resolution
The 17-35mm is strong at the shorter
focal lengths – contrast is particularly impressive but there are some
trade-offs in performance
Both the zoom and focus controls are
well-damped, but autofocus is far from smooth or quiet. The 0.28m minimum
focus distance offers a reasonably versatile close-up capability at 35mm
Few can argue with the capabilities at
the wider-end. While cheaper than the Nikon rival, it’s not as clear-cut for