(Isn't that a mouthful for a name?)
About a year ago, before I started with the tests you see here, I tested all my lenses of the time on resolution targets. This is when I realized that testing lenses was not as trivial as I thought. The Tamron 28-75 (excuse me to shorten the name) caused me lots of trouble. At first, its performance was mediocre at full aperture. Only after adjusting the AF on the Sony A900 did the measured performance become acceptable.
On the tests published in this blog, however, the performance of the Tamron was less than good. At first I thought that maybe the lens had become maladjusted and refrain from publishing the tests. The I retook the pictures of the resolution targets and things became even more confusing. Until I finally realized that the Tamron AF is simply not reliable. The following pictures show what you can get from the lens at 28mm and f/2.8:
Quite a difference isn't it?
For this reason, I decided to publish the results from this lens on this special page. You still can compare the results with the other lens if you open the pages in different tabs, but the results here are not exactly comparable: they have been selected amongst the best out of three different set of pictures, adjusting AF between takes. Note that they are not necessarily the best pictures that the lens can take, I selected the best pict out of 3, but it is quite possible that AF could still be wrong (especially for 50 and 75mm, which are relatively poor).
Here the center pictures:
f-stop | 28mm | 35mm | 50mm | 75mm |
f/2.8 |  |  |  |  |
f/4 |  |  |  |  |
f/5.6 |  |  |  |  |
f/8 |  |  |  |  |
And here the corner pictures:
f-stop | 28mm | 35mm | 50mm | 75mm |
f/2.8 |  |  |  |  |
f/4 |  |  |  |  |
f/5.6 |  |  |  |  |
f/8 |  |  |  |  |
(You can click the pictures to open them in another window, I needed to reduce the images to display 4 pictures in a row...)
So what are my conclusions about this lens from Tamron? In particular, how does it compare with the competition, the Sigma HSM or the CZ24-70?
I won't say much about the Sigma, which I have never held in my hands. All I can say is that, checking pictures taken with it on full-frame cameras on flickr, the corners appear to be very unsharp (check it yourself). On a reduced size sensor, the problem disappear, of course.
I had the occasion to compare the Tamron with the CZ. Better, there is another comparison on artaphot which started a heated discussion on various forums.
In short, the CZ is better (it should be at the price!), but far from perfect. The Tamron is much more compact and lighter (and that counts too).
The facts are that CZ is better: it is sharper full open, has much better contrast and colors and a more reliable AF. The Tamron has much more veiling full open (probably spherical aberration). Corner performance is relatively poor for the two lenses full open. Closed down to f/5.6 or f/8, the two lenses start to be difficult to tell apart on sharpness (but they still have a different color rendition).
Thepart which is more difficult to interpret is the lack of AF reliability of the Tamron: while this is less of a problem with 3 dimensional subjects taken 3 to 10 meters away, it is for landscape pictures. In theory, one could use the microAF adjust function of the A900 to compensate (as I have done here), but it seems that you would need a slightly different adjustment for each focal length, or even depending on focus distance (when I adjust the AF for perfect focus 2 meters away, subject at the horizon are out of focus for example). Even then, you still get out of focus images occasionally. Worse, the lens has the same i.d. as the 24-105 for the A900, so any adjustment is reported to the other lens... I suppose that these AF problems may explain the differences between the various reviews of that lens.
I am still investigating this lens. Any further findings will be posted here.
