I had an eye on this one for some time because of the wide aperture, good sample images (sharpness, bokeh, loCA), compact IF design, and relatively long working distance. The idea was that it should replace the 50/1.8 in my bag.
I'd waited for proper tests to show up; unfortunately, the Photozone test revealed some odd things in the MTF - the sharpness results looked too good to be true both at the wide and the stopped-down ends, indicating perhaps sth fishy about the aperture. And the Canon mount version MTF looked much different. Several users have complained about underexposure. Me not likes buying lenses likely bound to return.
Then, a few months have passed. I asked a local dealer the specific question and they asked at Tamron. The response was that the issue was found and corrected. So I bought my specimen on Friday hoping for the best.
After some tests, the impression is that the response was not an honest one and that the issue may have been covered up but not solved. I am not sure though, so I'm asking you for advice in this matter.
QUESTION 1: Are deviations like this at max aperture a common thing?
QUESTION 2: How about the deviations at intermediate stops (like f/5.6 - there are more like this also in the half-stops)?
Comment: the 50/1.8 looks poor but it has a thing with veiling flare in bright light - it would look a bit better with lower exposure and the hood. Could somebody help with the effective aperture values for the 50/1.8 on 56mm extension?
QUESTION 3: Do you think my speculations make sense, or you think they are flawed, or you have alternative ones?
QUESTION 4: Keep or return?
QUESTION 5: The colours look a bit washed out wide open, just as with the 50/1.8... Are good portrait lenses supposed to behave differently?
Tnx, Sem
I'd waited for proper tests to show up; unfortunately, the Photozone test revealed some odd things in the MTF - the sharpness results looked too good to be true both at the wide and the stopped-down ends, indicating perhaps sth fishy about the aperture. And the Canon mount version MTF looked much different. Several users have complained about underexposure. Me not likes buying lenses likely bound to return.
Then, a few months have passed. I asked a local dealer the specific question and they asked at Tamron. The response was that the issue was found and corrected. So I bought my specimen on Friday hoping for the best.
After some tests, the impression is that the response was not an honest one and that the issue may have been covered up but not solved. I am not sure though, so I'm asking you for advice in this matter.
- A quick impression is that I mostly have to use a stop higher exposure than with Nikon AF 50/1.8. Generally I would not consider this a serious problem.
- In the M mode at inf focus (as the MTF test), the lens is not clean f/2, there is no more light at f/2 than at f/2.4, and there is just a bit more at f/2.4 than at f2.8. The aperture moves obviously; when metering it is even a bit wider open than at f/2. I "played scales" against a white wall with constant (not completely even) illumination:
QUESTION 1: Are deviations like this at max aperture a common thing?
QUESTION 2: How about the deviations at intermediate stops (like f/5.6 - there are more like this also in the half-stops)?
- One could expect this to cause more fiddling with the exposure when changing to/from wide-open aperture values with the same scene. However, the change in the metering at large apertures seems to compensate for the lack of light. So when one uses A mode and changes from f/8 to f/2 the resulting images are exposed similarly, looking at the resulting histogram. I guess this is a kind of a trick that covers the issue up a bit and makes work with such a lens easier, but the results of shapness measurments such as the Photozone MTF stay the same. This could be changed from earlier samples of the lens, if they have actually changed something.
- I have not noticed anthing fishy in the near-diffraction region at f/16 and f/22 (where the MTF seems to sharp).
- The exposure behaves differently at 1:1, and the first half stop of the aperture already has some effect. This time I played scales using SB-800 at constant 1/180 s against a tilted mm grid paper at approx 45 deg, so that one can get an impression of the sharpness, perspective, contrast, and loCA, compared to the 50mm/1.8 on two tubes (with a similar magnification). Following: the histograms, the thumbnails and the pixel-peeping crops.
Comment: the 50/1.8 looks poor but it has a thing with veiling flare in bright light - it would look a bit better with lower exposure and the hood. Could somebody help with the effective aperture values for the 50/1.8 on 56mm extension?
QUESTION 3: Do you think my speculations make sense, or you think they are flawed, or you have alternative ones?
QUESTION 4: Keep or return?
QUESTION 5: The colours look a bit washed out wide open, just as with the 50/1.8... Are good portrait lenses supposed to behave differently?
Tnx, Sem
