Comments on: The Sigma 70-200mm f2.8 EX DG HSM Macro II

By: mhhd

mhhd — Fri, 18 Dec 2009 18:21:58 +0000

Dear David,

I read your comment about the dpreview test of this lens with Nikon body. I read this one first and was diaapointed about the lenses performance. So after your test I gained some hope, but unfortunatelly I am not that much into the technical details. Does this mean that nikon with its 5.6 sensor will not achieve good results in general with this lens? or is there a chance? My girlfriend is intereested in this lens for her D700 (sorry, mo Sony ).
Thanks for clarification

Regards,

mhhd

By: admin

admin — Mon, 21 Sep 2009 11:00:34 +0000

The Alpha 100 AF is only slow with screw drive focusing. With the HSM, it becomes as fast as any other body – the HSM AF has a big effect on the 100, real improvement will be seen.

David

By: poncho93

poncho93 — Wed, 16 Sep 2009 04:36:15 +0000

Does anyone have experience with this lens and an Alpha 100? I’m debating on whether or not to get this lens or upgrading my camera body to an a700 before they are discontinued. For now I can only pick one or the other. Let me know if a lens like this would be a waste with a slow AF body like the a100. Thank you.

By: admin

admin — Sat, 25 Jul 2009 17:17:07 +0000

A tilted sensor will still be out of focus in some zones, and perfect alignment is the only reasonable QC goal. I don’t have a web link for FOTO, there are many mags with a similar name, it is the Swedish one. I don’t keep the copies I get as we end up submerged in paper that way.

I am able to revise the on-line report here, the BJP articles have to be just so many words, and I also have to think in terms of their market.

David

By: plevyadophy

plevyadophy — Sat, 25 Jul 2009 13:06:23 +0000

Sir,

Thanks VERY much for your VERY prompt response.

In response to your first answer, does that mean for focusing off of the sensor in live view mode it matters not whether the sensor is perfectly aligned?

Secondly, that FOTO magazine you talk of, I think I may have stumbled across it before but not too sure. Do you by chance have a web link for them? Thanks.

Thanks for link to Anders Uschold. I am sure I recall you mentioning him before either over at the PhotoClubAlpha forum or in the BJP (by the way, I have read your Sigma 70-200 twice now, as it appeared in the BJP this week but the write up here is more informative)

Regards,

By: admin

admin — Sat, 25 Jul 2009 11:55:41 +0000

With on-sensor live focus, the point focused is generally perfectly focused – wherever it happens to be. It can be affected by aperture related issues, but many live AF systems operate at the working aperture, not wide open. With the Nikon cameras having LV focus and also offering micro-AF adjustment for the phase detect sensor system, the corrections only apply to phase detect AF, and you use the LV enlarged focus to fine tune the micro adjustment until the two match. With the A900, you have to take shots and review them.

2 – first, the Zeiss testing facility, which was set up originally with Hasselblad and still supplies lens test data to some TIPA members (Technical Image Press Association). Swedish FOTO magazine for example uses the Zeiss source data to create the technical part of their tests. As far I know, it’s still all running the same way – multiple samples of the lens must be supplied for testing, figures are averaged, no comment is ever made – they just provide the readings.

Secondly, Anders Uschold. He seems to have stopped providing material to the British Journal, I guess they never mastered the art of translating his stuff properly. He’s very good at showing how the combination of lens+sensor is what matters. His purely technical language means he can make strong criticisms of lenses without sounding as if he is trashing them.

http://www.uschold.com

David

By: plevyadophy

plevyadophy — Sat, 25 Jul 2009 10:14:43 +0000

Sir,

Thank you for a very good review, and for an extremely educaational write up.

Two things.

1)
I note that you point to the issues of sensor alignment, and AF-sensor module needing to be perfectly collimated.

How then do these issues, of uneven focus across a frame, play out with cameras such as the Panasonic G1 which is an interchangeable lens mirroless design?

And of course, a related issue would be, if ever Sony produces a camera with off-sensor live view, will there be potential for a difference between the focus results obtained by way of live view and that of focusing traditionaly?

2)
You also said:

“Simple practical results can tell you more than any amount of procedural testing. That is not to say you should ignore chart-testing, and I hold a couple of German tech-lab test experts in high regard.”

Would you be so kind as to share with us the details of those German tech-labs? I have come across one that I find interesting, Photozone.de.

Regards,

By: admin

admin — Thu, 16 Jul 2009 16:50:50 +0000

I have added a little paragraph at the end of the original story, with a diagram showing how field curvature combined with sensor alignment error can spook results. If your 80-200mm had a dead flat focus plane, it might have fallen within acceptable limits even with a slightly out of parallel sensor, and the falloff would have looked more or less symmetrical if the centre focus was perfect.

Take a lens with a curved focus field, add some FF or BF error, some sensor alignment error – I’m sure you can see the potential for various combinations of fairly ugly, asymmetrical result.

David

By: vbpholaw

vbpholaw — Thu, 16 Jul 2009 16:16:17 +0000

Thanks for the further comments David. I would note with respect to parts of an image being sharp and others not, that for my “brick wall test” I was at the time comparing the 70-200 SSM to my 80-200/2.8 HS APO (which I still had), and the latter lens either did not show a similar issue wide open, or if it did it was in a different portion of the frame (I don’t recall, and would have to try to find the photos and look at them to be certain – what I do remember is that I concluded it was a lens rather than camera issue; if it was the camera/sensor then the same problem should have appeared for both lenses). If I recall correctly, it also was not the same at different focal lengths (just as one lens performed ever so slightly better at one focal length, while the other might eke out a superior showing at a different focal length). This also suggested to me that it was not the sensor but the lens. As for what characteristic of the lens caused it I won’t speculate (don’t know enough to do that ).

When one thinks about how critical is the alignment of various components in the focusing chain, it is sort of amazing how the whole system works so well (generally speaking). The elements of the lens, the lens mount and lens mounting flange, the mirror assembly, the location of the focusing screen/AF sensor and the location of the imaging sensor. All need to be in proper alignment and calibration for the system to work right. No doubt there are some tolerances (however small) for each of these things, but if you have a camera and/or lens for which some or many of these components are slightly off, it can be difficult to figure out where the problem lies and get the fix right. But that’s really getting off topic.

Thanks again.

By: admin

admin — Thu, 16 Jul 2009 08:59:01 +0000

I am not able to provide direct comparisons with the Minolta as I sold that within a few days of trying it on the Alpha 900 with the 2X APO and on its own. The pic at the end was added in an edit, after realising that I had commented on the curved field of the Minolta at close range, but provided no real evidence. I do have other pix on file of course but I never used the Minolta 70-200mm as much as I have been using the Sigma. My Minolta also tended to be softer on one side of the image than the other, but the focal plane alignment accuracy of the D7D sensor (on which it got most use) is poor compared to the A900. The A900 has about the best parallel/plane setup I have come across. It still shows some error with the Sigma 12-24mm at 12mm and I have to remember to change my vertical shooting orientation to keep the ground and distance sharp – wrong choice and the tops of buildings start to blur slightly. But it’s still far better in terms of setup than the Canon 5D MkII I tested with a 16-35mm f/2.8 MkII – that showed a really noticeable ‘one end soft’ issue.

In my experience, these problems are very rarely the lens (though test reports nearly always say they are). They are to do with sensor flatness and focal plane parallelism/collimation, plus AF front and back focus errors, plus lens field curvature and aperture related focus shift. The combination of all factors can result in the point of focus being just acceptable centrally, one side of the frame being better focused but with more spherical aberration sharpness loss (so looking just acceptable as well), and the opposite side of the frame being much worse focused and having the same aberrations leading to a slightly smeared blur.

The alignments required are critical, especially with ultrawides and wide aperture lenses. Most DSLRs are just not well enough set up to perform well. My A900 is one of the best I’ve ever used. The Nikon D3 (3 bodies) we have used have been excellent, D700 not quite so. Canon, very variable. The specialist repair company Fixation, in London, offer a service to align and centre sensors on Canon and Nikon to higher standards than the factory spec. They don’t do Sony or any sensor-stabilised cams.

David