[mst]

Since the aperture of the lens has been successfully calibrated by Tamron Europe, all affected lab tests were redone and graphs updated.



The verdict remains unchanged, however. And this is surprisingly also true for the way too good resolution fully stopped down.



http://www.opticallimits.com/nikon--nikkor-aps-c-lens-tests/481-tamron_60_20_nikon



-- Markus

[Brightcolours]

[quote name='mst' timestamp='1294061346' post='5317']

Since the aperture of the lens has been successfully calibrated by Tamron Europe, all affected lab tests were redone and graphs updated.



The verdict remains unchanged, however. And this is surprisingly also true for the way too good resolution fully stopped down.



http://www.opticallimits.com/nikon--nikkor-aps-c-lens-tests/481-tamron_60_20_nikon



-- Markus

[/quote]

Not all lenses show exactly the same amount of softening due to diffraction at the same f-stops, I think that has to do with the position of the aperture (not sure). So in theory, it might just be that this Tamron design is just exceptionally good in this area.



However, we have to leave that idea when we look at how the same lens performs on another camera:



Nikon (D200):





Canon (EOS 50D):





We can see here that it is NOT the optical design... so it must be that the aperture is not correct. Have you checked exposure durations?

[mst]

[quote name='Brightcolours' timestamp='1294071231' post='5321']

We can see here that it is NOT the optical design... so it must be that the aperture is not correct. Have you checked exposure durations?

[/quote]



That was my idea, too. In fact, the aperture not fully closing is the only reason that seems to make sense. However, I checked durations and "unfortunately" they are flawless:



f/5.6: 1/40s

f/8.0: 1/20

f/11: 1/10

f/16: 1/5

f/22: 1/2.5



The resulting files show no shift in exposure (f/22 files aren't any brighter).



Honestly, I'm absolutely clueless about the reason for the high f/22 resolution. Will be interesting to see if the lens shows the same behaviour on the D7000.



-- Markus

[Brightcolours]

[quote name='mst' timestamp='1294076651' post='5322']

That was my idea, too. In fact, the aperture not fully closing is the only reason that seems to make sense. However, I checked durations and "unfortunately" they are flawless:



f/5.6: 1/40s

f/8.0: 1/20

f/11: 1/10

f/16: 1/5

f/22: 1/2.5



The resulting files show no shift in exposure (f/22 files aren't any brighter).



Honestly, I'm absolutely clueless about the reason for the high f/22 resolution. Will be interesting to see if the lens shows the same behaviour on the D7000.



-- Markus

[/quote]

Quite puzzling. But there might be a (strange) explanation.



With Canon cameras, aperture settings are only done in the lens, there is no stuff going on between body and lens (well there is, but more on that later). With Nikon, that is different. We all know that with focussing closer, the bellows factor makes macro lenses less "fast": The closer we focus, the smaller the real aperture is. With Canon, the camera keeps on saying "f2.8", where on Nikon we can see the aperture figure in camera go to f3.5, f4 even, even though we do not change the aperture setting.



This means that, while the aperture mechanism does not change, the camera and lens together determine a different f-value.

And with Canon... only the user set f-value is being "used", not the "real" aperture.



There is something else going on too, "recently" discussed on internet. Wide aperture lenses seem to be too fast for sensors to deal with, and there seems to be tampered with ISO amplifications to make the exposure times match the f-value.

For instance, if I put my 35mm f2 lens on my 450D normally, the photo taken is a bit lighter than when I mount the lens in such a way that the lens/camera electrical contacts are not connected.

In other words... same lens, same lens opening, so same exposure, but different amplification of the signal.



This not only can be observed with Canon, also Nikon (and Pentax and others).



Now, combining these two things, one can imagine that maybe Tamron's lens does strange things on Nikon because "it can", closing down less and letting the camera "adjust ISO" accordingly?



It can not be the optical design, as that is identical for the different mounts, with identical aperture placement.

[Klaus]

Regarding the characteristic of the curve the Canon variant is more "valid" though. At f/22 there isn't really much quality left. This is also in sync with other macro lens tests.



Only the Tamron in Nikon mount is the exception to the rule here.

[mst]

[quote name='Brightcolours' timestamp='1294105340' post='5329']

This means that, while the aperture mechanism does not change, the camera and lens together determine a different f-value.

And with Canon... only the user set f-value is being "used", not the "real" aperture.[/quote]



That's just a matter of reporting, isn't it? Canon cameras display the physical aperture, while Nikon cameras show the effective aperture. Of course, both (have to) use the effective aperture for metering.



Anyway: no issue here. The Tamron already starts to report slightly different effective apertures at the distance we measure MTF at, but I made sure to use the correct physical aperture settings. f/22 in the charts is reported as an effective aperture of f/25 by the lens (which means "fully stopped down").



[quote name='Brightcolours' timestamp='1294105340' post='5329']There is something else going on too, "recently" discussed on internet. Wide aperture lenses seem to be too fast for sensors to deal with, and there seems to be tampered with ISO amplifications to make the exposure times match the f-value.[/quote]



Yes, I know, that LL masterpiece ... but the total amount of magnification will probably hardly make any impact in our measurements. And of course, as I understand it, this happens at large apertures, not at f/22.



-- Markus

[Guest]

Markus, what about blur discs? I think because it is a macro lens, you can probably create large blur discs despite f/22 by using a large magnification and distant highlights. The blur discs have to shrink along the closing aperture proportionally.



Christian

[mst]

[quote name='TheChris' timestamp='1294133284' post='5333']

The blur discs have to shrink along the closing aperture proportionally.

[/quote]



I'll check that.



-- Markus

[Brightcolours]

[quote name='TheChris' timestamp='1294133284' post='5333']

Markus, what about blur discs? I think because it is a macro lens, you can probably create large blur discs despite f/22 by using a large magnification and distant highlights. The blur discs have to shrink along the closing aperture proportionally.



Christian

[/quote]



Look again how the same lens behaves on Canon 50D. Same lens, but sensible measurements.

[Brightcolours]

[quote name='mst' timestamp='1294131755' post='5332']

That's just a matter of reporting, isn't it? Canon cameras display the physical aperture, while Nikon cameras show the effective aperture. Of course, both (have to) use the effective aperture for metering.



Anyway: no issue here. The Tamron already starts to report slightly different effective apertures at the distance we measure MTF at, but I made sure to use the correct physical aperture settings. f/22 in the charts is reported as an effective aperture of f/25 by the lens (which means "fully stopped down").







Yes, I know, that LL masterpiece ... but the total amount of magnification will probably hardly make any impact in our measurements. And of course, as I understand it, this happens at large apertures, not at f/22.



-- Markus

[/quote]

How does the reporting happen then? Yes, when we set aperture to "f22" on the camera, the aperture is still at f2. The lens is not reporting anything. When we make a photo, the lens gets closed to "f22" on Canon.



So what goes on with Nikon? We set an aperture, but the lens remains wide open. Yet, when we focus closer, we see the aperture "reporting" a smaller set aperture. The aperture mechanism is not reporting that. So how does that happen? What is driving that? The lens is not closing its aperture mechanism more, and if you have an unknown lens, the reported aperture is not changing obviously.



It is merely the fact that it changes that is interesting.



So, now back to the DXO/LL wide open trickery. We now also know that cameras apparently can cheat with amplification.



Now I am combining the two... apparently, lenses can get a Nikon camera to report a different aperture than set, probably from measurements done by the manufacturer during design. So, with macro lenses and the bellows factor for instance, they can get the camera to show smaller apertures than originally set, when we focus closer.



When we also take into account that ISO-related amplification can be used to make exposure match what we are expecting, we could imagine a scenario here where the Tamron on Nikon does not really close to f22, but that the amplification is being used to let the exposure time match a real f22.



I know, this sounds like a big "why would one want to do that", especially since smaller f-values with macro photography are not used for longer shutter times, but rather for more DOF. It would have been done with resolution tests in mind, to show better results on review sites....



Anyway, what is the most curious here is the difference of measurements between the Canon EOS 50D (measurements are the way we expect) and the Nikon D200 (measurements make little sense).