Studio Tests

We’ve been quite impressed with this lens when using it in the field, and the studio tests show why. The resolution results may not appear wonderful - the lens isn't all that sharp at 16mm, and the corners are soft wide open at all focal lengths - but this is offset by excellent characteristics in other areas, with little in the way of chromatic aberration, falloff or barrel distortion. This lens is, in fact, an excellent example of why absolute sharpness is not always the Holy Grail of optical design.

Sharpness Decidedly indifferent at 16mm, where it needs to be stopped down to F11 for best results, the lens improves markedly at longer focal lengths, and is at its best around 28mm. Image quality improves significantly on stopping down , with F8 and F11 the sweet spot; at smaller apertures diffraction takes its toll, as usual. Fortunately apertures smaller than F22 aren't available.
Chromatic Aberration A fairly respectable performance here; Pentax have generally limited chromatic aberration to the least problematic blue/yellow type. As always the effect is most pronounced at wideangle, where the red-cyan component is sufficiently large as to be visually intrusive; at 28mm and above there's little to worry about.
Falloff This is an area where the 16-45 really shines. Due perhaps to the generous front element dimensions and unambitious zoom range,  you’ll struggle to see any perceptible falloff (i.e. beyond 1 stop) except wide open at 16mm. Well done Pentax.
Distortion Distortion is (as always) most pronounced at wideangle, but with a mere 1% barrel at 16mm, this is unusually well-corrected for a wideangle zoom; an impressive performance indeed. As usual, the distortion isn't 'pure' barrel, but instead a more complex 'wave' type, with re-correction towards the corners. At 20mm the lens is almost perfectly corrected, but at longer focal lengths it shows distinct pincushion distortion, maxing out at -1.2% at 45mm.

Macro Focus

A very creditable performance here, with an image area of 78mm x 52mm, equating to a maximum magnification of 0.31x, unexpectedly rather better than Pentax’s specified 0.26x. This is achieved at 45mm focal length and a focus distance of 25cm, giving a working distance of about 10cm between the subject and the front of the lens. 

Image quality is very good indeed; slightly soft at F4, the centre is very sharp at F5.6, with corner performance close behind. There's just mild pincushion distortion, plus a little blue/yellow chromatic aberration. Impressive stuff.
Macro - 78x52 mm coverage
Distortion: Mild pincushion
Corner softness: Very low
Focal length: 45 mm (68mm equiv)

Specific image quality issues

Chromatic Aberration

Overall chromatic aberration is fairly well controlled, but is (as usual) most problematic at wide angle. At 16-20mm, green/magenta CA can become intrusive towards the edge of the frame (see example below). At longer focal lengths, our studio tests show clear blue/yellow fringing, but in real-world shooting this turns out to be essentially a non-issue; it certainly doesn't have the destructive visual impact of red/cyan or green/magenta fringing.

16mm, F8.0 100% crop

Portait Format Softness

One clear issue with our sample was extreme softness at the edge of the frame when shooting at wide angle and in portrait format, especially with the lens angled sharply upwards (perhaps not a common shooting situation, but distinctly noticeable in some architectural shots). We can only speculate that this could be a decentring phenomenon due to the play in the lens barrel when fully extended at 16mm.

16mm, F5.6 100% crop
16mm, F5.6 100% crop

Flash Shadowing

With its unusual ‘reverse zoom’ design, this lens shows severe shadowing of the built-in flash at wide angle settings. The effect will be dependent upon the specific camera body used, and how far the flash lifts above the lens axis; with the K10D, the manual states that ‘vignetting may occur when the focal length is shorter than 28mm, or when the focal length is 28mm and the shooting distance less than 1m’. Our own testing shows these to be reasonable, although somewhat conservative guidelines.

Here's a picture of the tasteful dpreview office carpet, showing the flash shadowing using our K10D test body with the lens set to 16mm. This is close to being as bad as it gets, but most modern lenses show no shadowing at all, and clearly a real-world image showing this effect would be pretty unusable.