[quote name='karlmera' timestamp='1312804546' post='10498']
But is that logical?
[/quote]
Yes, that is the nature of MTF tests. The test target must have a certain specific size on the image frame for the measurements to make any sense. Everyone does it that way, just the nature of the measurement beast. And macro lenses of course are not only used for macro photos.
It is not as if macro lenses get to be less sharp all of a sudden when you focus closer...
There are a few exceptions, which are not that consistent. Like my Nikkor 55mm f3.5 micro 1:2 macro (1st generation). It is said to be one of the, if not the, sharpest Nikon SLR lens, but only at macro distances. Its performance breaks down when used as non macro lens. More modern macro lenses have a design which prevents that.
08-08-2011, 01:23 PM
(This post was last modified: 08-08-2011, 01:38 PM by Klaus.)
[quote name='mst' timestamp='1312806569' post='10499']
Yes, if you want tests that are comparable within the whole test system.
I know you're interested in macro performance, but as Klaus already said, we simply cannot do these tests, mainly because we'd have to find someone who is able to produce a decent test chart. That chart would have to be 24x36mm in size (or maybe 48x72mm if you can lie with 1:2 MTF data) and printed at almost infinite resolution.
-- Markus
[/quote]
In theory such a "chart" could be custom-created by the semiconductor industry. We are talking about a mask equivalent to about 20x of the resolution of an image sensor ... just consider the costs for this. Assuming the pixel density of the upcoming NEX7/A77 sensor we would talk about a gigapixel charts printed on 36x24. In practice we would face all sorts of alignment and focusing problems. Not to mention centering issues. The margin of error would be extreme.
Regarding our findings - I think its viable to assume that they are at least valid till a max. magnification of 1:5 for a macro lens (with floating elements). That should be good enough anyway.
[quote name='Klaus' timestamp='1312809799' post='10503']
In theory such a "chart" could be custom-created by the semiconductor industry. We are talking about a mask equivalent to about 20x of the resolution of an image sensor ... just consider the costs for this. Assuming the pixel density of the upcoming NEX7/A77 sensor we would talk about a gigapixel charts printed on 36x24. In practice we would face all sorts of alignment and focusing problems. Not to mention centering issues. The margin of error would be extreme.
Regarding our findings - I think its viable to assume that they are at least valid till a max. magnification of 1:5 for a macro lens (with floating elements). That should be good enough anyway.
[/quote]
I have a target that outresolves my camera by a big margin at 1:1... focusing is indeed very porblematic, let along trying to determining what the results would be.
[quote name='Brightcolours' timestamp='1312806797' post='10500']
There are a few exceptions, which are not that consistent. Like my Nikkor 55mm f3.5 micro 1:2 macro (1st generation). It is said to be one of the, if not the, sharpest Nikon SLR lens, but only at macro distances. Its performance breaks down when used as non macro lens. More modern macro lenses have a design which prevents that.
[/quote]
Zeiss provides MFT curves at both 1:10 & 1:2
http://applications.zeiss.com/C12578620052CA69/0/B06F47F367D29FB3C1257864004D344F/$file/mp_2_100_en.pdf
http://flickr.com/ephankim
[quote name='youpii' timestamp='1312817089' post='10508']
Zeiss provides MFT curves at both 1:10 & 1:2
[url="http://applications.zeiss.com/C12578620052CA69/0/B06F47F367D29FB3C1257864004D344F/$file/mp_2_100_en.pdf"]http://applications....mp_2_100_en.pdf[/url]
[/quote]
Nicely computer generated stuff probably.
[quote name='Klaus' timestamp='1312817671' post='10509']
Nicely computer generated stuff probably.
[/quote]
And very close results too for both distances...
[quote name='Klaus' timestamp='1312809799' post='10503']
In theory such a "chart" could be custom-created by the semiconductor industry. We are talking about a mask equivalent to about 20x of the resolution of an image sensor ... just consider the costs for this. Assuming the pixel density of the upcoming NEX7/A77 sensor we would talk about a gigapixel charts printed on 36x24. In practice we would face all sorts of alignment and focusing problems. Not to mention centering issues. The margin of error would be extreme.
Regarding our findings - I think its viable to assume that they are at least valid till a max. magnification of 1:5 for a macro lens (with floating elements). That should be good enough anyway.
[/quote]
Assuming that you /Klaus and Marcus / are reach they ready to pay 50 000 Euro for such target.
Beside of aligment problems I wonder how you can keep the target clean. I hope that Klaus is not going to arrange clean room :-)
Yes Klaus,
I have forgotten to mention that cammera must be mounted of stages used in semiconductor lithography equipment.
The are even more expensive than target.
To fight with environmental vibration you need a specal concrete floor and own building. Not cheap!
[quote name='mst' timestamp='1312797599' post='10492']
I can only speculate why this doesn't work with the 150/2.8. Maybe it has to do with the effective aperture (with a 2x TC added, the lens is f/5.6 at large distances only).
[/quote]
That is my guess also. If you use a 1.4x on it, the focus range is limited to a shorter region than without, and the combination from memory is about 1:1 magnification at that point. I guess with a 2x they could have limited it further out again to maintain enough effective aperture, but for whatever reason they chose not to.
<a class="bbc_url" href="http://snowporing.deviantart.com/">dA</a> Canon 7D2, 7D, 5D2, 600D, 450D, 300D IR modified, 1D, EF-S 10-18, 15-85, EF 35/2, 85/1.8, 135/2, 70-300L, 100-400L, MP-E65, Zeiss 2/50, Sigma 150 macro, 120-300/2.8, Samyang 8mm fisheye, Olympus E-P1, Panasonic 20/1.7, Sony HX9V, Fuji X100.
[quote name='miro' timestamp='1312790032' post='10480']
"Accurate manual focusing a bit tricky though because the focus path is relatively short for a macro lens. "
1. How much is the MF rotating angle?
2. Do you mean that you have MF problem during the target test or when you perform target/resolution test?
The non OS version rotates 270 degree and I find it just good enough for MF at macro distances. The infinity setting is different story.
PS Nor of macro lenses are good here. E.g. tokina 100macro rotates more than 270 degree while the difference between 3m ant infinity is just few degree.
Sigma 150 non OS is just a little bit better. With a little care I can focus maunaly for porthraits and landscapes shoots.
I don't mentioned canon 100 macro non L - it rotates just 170 degree from MFD to INF.
[/quote]
160 degrees here I'd say.
