Article Index


Most wide-to-normal(ish) zooms like Vario-Sonnar 24-70/2.8 ZA are fairly good performers. Canon's 24-105/4L is one of my favorite lenses when I travel light and choose to drag along a Canon body. Vario-Sonnar fills in a similar role for situations when I use Sony camera. But the word 'good' is obviously relative here - good compared to what? My reference here is other zoom lenses in similar focal range, but not fixed focals. Because of much more complex optical designs that manufacturers have to employ in zooms, comparing say Vario-Sonnar to Distagon 24/2 is not gonna be very fair. Take a look for example at the Imatest MTF plot below. If we try to make our decision about resolving capabilities of this lens solely based on this test chart, we might conclude that the lens is not particularly good. and if we compare the said lens at 24mm to the Distagon 24/2, we would simply render a verdict that this lens sucks plain and simple (at least @ 24mm). The Imatest chart claims that the lens performs quite well in the center across all focal ranges, but under-performs around borders. 24mm is particularly challenging where resolution seems to be the lowest and does not seem to improve much with stopped down apertures. 35mm to 70mm results are better, but still exibit lower relative resolution compared to the center. Stopping down to f/4 starts to produce good results everywhere except @ 24mm borders, which require another stop-down to get to more or less decent, per Imatest, results. But Imatest's numbers describe only one dimension - how well (or poorly in this case) the image performs with a slanted edge discrete MTF algorithm and completely ignores any visual aspects.


24mm (Sony a850)


35mm (Sony a850)


50mm (Sony a850)


70mm (Sony a850)


The crops of the imaging target below compare image quality around corners at f/2.8 and f/8. The shots here were taken at the focusing distance of 3m for 24-35mm and 6m for 50-70mm. This is what was fed into Imatest - note that as Imatest claims, f/2.8 corners are indeed softer pretty much across the zoom range. But visually, 35mm, 50mm and 70mm borders at f/2.8 are not too bad, even if you compare them side by side with f/8. It's the 24mm borders that are quite noticeably worse off. Ok, so what does this mean for a user?


24mm (Sony a850)


35mm (Sony a850)


50mm (Sony a850)


70mm (Sony a850)



Well, the image quality of Vario-Sonnar 25-70/2.8 ZA in real life applications is quite good actually. Throughout the years of using this lens, I have yet to discard an image from Vario-Sonnar 24-70/2.8 because of its quality, specifically sharpness. Yes, borders are softer - we a;ready determined that, particularly at wider apertures, where distortion claims its toll (you can observe that in the crops below, even without peeking at the images at high magnification). But guess what  - image quality, most notably at 24mm, is somewhat better when the lens is shot at longer focusing ranges (infinity in the examples below), than what we're led to believe by Imatest. Also, remember that I mentioned that comparison should be made in the appropriate scope, that is we should be comparing Vario-Sonnar to other 24-70mm lenses? Well, I own Nikon's 24-70/2.8G and used Canon's 24-70/2.8L in the past, and while comparing these three lenses is not strictly speaking possible because of the variance in cameras, processors, sensors etc., subjectively speaking, Vario-Sonnar does not produce results that are much off compared to the above mentioned competition. Nikon seems to have better resolution at 24mm, but that gets lost when you use these lenses at long focusing distances.The biggest issue with Vario-Sonnar's 24mm is the somewhat higher distortion and field curvature, which wreck chaos at close focusing distances. This is where Distagon 24/2 clearly bests Vario-Sonnar 24-70/2.8 - it's sharper, has less distortion and less field curvature, but again, that's not a fair comparison. And that's is - if we were to focus only on the Imatest's results, we would not have managed to piece together a complete picture (pardon the pun) - Vario-Sonnar's Achilles heel, the 24mm at wide apertures, is mitigated to some degree at longer focusing distances, and Imatest misses that completely.


24mm (Sony a850)


35mm (Sony a850)


50mm (Sony a850)


70mm (Sony a850)



On the final note, take a look at the graphs below, which acerage MTF results for various focusing distances and aperture settings. Note that the 24-35mm were tested only in 1m to 5m focusing distances, while 50 -70mm ranges were tested in 1m to 8m distances.



Color & Rendering

Overall, color reproduction of the Vario-Sonnar 24-70/2.8 is pretty good. The color gamut is shifted slightly towards reds, highlighting warmer tones better - this is a third Sony mount Carl Zeiss lens I have tested and all three had somewhat higher emphasis on warmer colors - Planar 85/1.4 ZA was the warmest and Sonnar 135/1.8 ZA and Vario-Sonnar 24-70/2.8 have more or less similar characteristics. I don't really understand why that is the case, considering that all carry the T* coating designation (arguably all same) and considering that most original Zeiss primes emphasize blues over reds, which in turn gives images a cooler feel to them. Just o be clear, Vario-Sonnar 24-70/2.8 does not have Canon's 'over the board reds', but it is also not as neutral as Nikon.


The lens  underexpose images at wider apertures quite consistently, mostly because of the pretty heavy vignetting at wider apertures. If you're like me, and have a habitual preference to underexpose images slightly, to capture more tonality in shadows and mid-tones, than you should be fine, otherwise you'd probably need to try fiddling with the exposure compensation a bit, setting it to +0.5/+1. Just don't forget to reset it when shooting in the 50-70mm range, where the underexposure is not that pronounces.

Tonality reproduction with the lens is quite good overall, but does vary slightly across the zoom range. 24-35mm range produces slightly higher contrast and does compress shadows more than the longer end of the zoom. 50mm is well balanced, while 70mm has longer mid-tone tail, at some expense of highlights and shadows. Global contrast is also lower at 70mm, particularly at close focusing distance, but not completely trashed - those of you who like contrastier images would find it easy to add a little bit of contrast to your images, without flattaning shadows too much.

Chromatic aberration is reasonably well controlled in this lens. While Imatest clocks ~1.1px of lateral CA at f/2.8 and 24mm, and much lower at longer focal ranges, in real life I have yet to see a picture taken with the Vario-Sonnar where CA is easily noticeable under normal shooting conditions. If you eyeball pictures hard enough at 100% magnification, you will probably notice some fringing around borders, but it is doubtful that CA would have any major impact on overall image feel. Longitudinal CA is also well controlled - there are minor traces that I have seen once in a while at 24mm and wider apertures, but disappear at longer end of the zoom.

CA Borders Sony a850 (24Mp)



DOF & Bokeh


DOF and bokeh testing on zoom lenses is a brutal exercise. Not only beccause the number of possible test combinations proliferates very quickly, but mostly because the rendering characteristics  change often quite drastically across different zoom levels. The wide to medium zooms like Vario-Sonnar 24-70/2.8 ZA are the biggest troublemakers - at wide angles DOF is not particularly shallow and in many cases bokeh rendering is rather harsh, while longer end of the zoom range behaves much better and opens up a world of creative possibilities. Having owned this lens for ~2 years now, I do find muself gravitating away from 24mm whenever I am trying to create a clear isolation between the subject and the background - there is simply too much DOF at 24mm. 35mm is slightly better, but 70mm is what I use most often, even if it means stepping away from a subject a little bit in order to fit it into the frame.

Let me be clear here - I don't see anything wrong in the bokeh at 24mm (or as a matter of fact at any of the other focal ranges). There are no easily observable artifacts, and rendering of out of focus backgrounds is fairly neutral as there are not signs of spherical aberration throughout the zoom range. It's just shorter focal lengths leave too much definition even with wide open apertures. Examples of that are below - as is customary in the reviews these days, here are series of shots iterating over different zoom and aperture levels. All shots were made at close to MFD (~50cm). The distance to the farthest building in the background is over 50m. 

DOF @ 24mm

thumb-dof4-cz-variosonnar-24-70mm-f28za-iso100-s2000-f28-24mm thumb-dof5-cz-variosonnar-24-70mm-f28za-iso100-s1000-f4-24mm thumb-dof6-cz-variosonnar-24-70mm-f28za-iso100-s250-f8-24mm

DOF @ 35mm

thumb-dof10-cz-variosonnar-24-70mm-f28za-iso100-s2000-f8-35mm thumb-dof11-cz-variosonnar-24-70mm-f28za-iso100-s1000-f4-35mm thumb-dof12-cz-variosonnar-24-70mm-f28za-iso100-s250-f8-35mm

DOF @ 50mm

thumb-dof7-cz-variosonnar-24-70mm-f28za-iso100-s1600-f28-50mm thumb-dof8-cz-variosonnar-24-70mm-f28za-iso100-s800-f4-50mm thumb-dof9-cz-variosonnar-24070mm-f28za-iso100-s200-f8-50mm

DOF @ 70mm

thumb-dof1-cz-variosonnar-24-70mm-f28za-iso100-s1250-f28-70mm thumb-dof2-cz-variosonnar-24-70mm-f28za-iso100-s640-f4-70mm thumb-dof3-cz-variosonnar-24-70mm-f28za-iso100-s200-f8-70mm

 As one would expect, at equal focusing distances, the bokeh feels much smoother at longer zoom ranges, giving us nice fade-in/fade-out transitions from foregrounds into backgrounds. At close focusing distances the lens produces slightly lower contrast, characteristic to most lenses, not just this particular one. Combined with muted backgrounds this gives a very nice feel  to images. Background highlights are rendered in neutral way, even at 24mm, although as mentioned above, the bokeh feels cluttered because of the longer DOF and higher definition it brings to backgrounds.