I've just read the review - nicely written, and it's what a field review should look like. There are two question I'm wondering about, though:
1) Sharpness beyond f/16 reported to be poor. I'm surprised that there's no statement like "this is to be expected due to diffraction" or something like that. As is, it sounds like a flaw that could be avoided by better lens design.
2) Have you (Darwin) tested whether the poor corner performance of the 16-35 at wider apertures, as tested on a flat subject, is due to field curvature?
I think these are two bits of information that could make the good review even better. What are your thoughts on the subject?
Best regards & thanks to everybody writing articles!
Basil, you are right that these may reveal more differences between the lenses, but what I appreciated about this review is that, unlike others which focus on lab tests, it is about using the lenses in the field and for prints - a real world comparison the way I (and probably the vast majority of owners) will use such a lens in our work.
Performance wide open and beyond f/16 for an ultra-wide lens will tell me nothing useful since I will likely never use these settings.
For what it's worth, I did a similar test with my 17-40 and a friend's 16-35 II. I could not tell the differences between the files (shot at at f/16).
This is exactly the kind of review most of us need. There are a lot of reviews, threads and discussions that are on the pixel peeper level that have little to do with real world performance. Unfortunately, many will obsess about them and forget that photography is about images not pixels or test charts. Thanks Darwin.
This is the reason I have owned the TS-E 24mm, 45mm & 90mm since moving to digital! Not as convenient but much sharper! They can also be used with the EF1.4x Extender giving you 6 lens choices!
I have been contemplating buying the 16-35 II or the 17-40 for a few weeks since I broke my 24-135. I think I'll get the 17-40 for backpacking. Thanks Darwin!
Yes poor lens performance beyond f16 is due to diffraction - I did not explain this in the review so thanks for mentioning it here. All 35mm lenses I have ever tested have poor performance at f22 (or whatever the smallest aperture is on the lens), enough that I can see a difference in prints. I no longer use any of my lenses stopped down all the way but at one aperture open from the smallest aperture (e.g. f16) - for those who want more info see:
I tested edge sharpness of the lens using flat surfaces - newspapers tapped to wall, graffitti on a wall etc - and the results were the same as the rock photos posted (which is also a flat surface) - so field curvature should not have been a factor in the results. The simple fact is that both lenses suck when it comes to edge sharpness on a full frame camera especally at wider apertures. I posted the rock photos as most of us shoot nature and not newspapers.
In the end shooting real scenes (landscapes), I could not see any real benefit of the 16-35. Remember there is reportedly much variation in the quality of the same model of lens - sample variation so to speak. There are poor versions of both lenses out there. I don't understand why this should be the case shouldn't quality control be expected with lens manufacturing?? Darwin
Thanks for replying, Darwin. I agree with you that QC should take care of the unacceptably poor copies, particularly if the subject is a multi-1000$ L-grade (gold ring in Nikon land) lens. I guess that the manufacturers' definitions of what is unacceptable are not what we think they should be.
About field curvature - I'm not sure whether we're talking about the same thing (might be a language issue on my part, I'm not natively English speaking). What I'm referring to is that the focus field of many wide angle lenses is not flat, but curved towards the camera at the corners. So if you "overfocus" beyond the optimum point for center sharpness, the corners should become sharper if the reason for corner softness is the shape of the focus field. The Nikon 17-55/2.8 DX has a reputation for it, and some suggest to focus it beyond infinity to improve corner sharpness when shooting lanscapes.
Maybe my hypothesis is plain silly, though. I'm not so familiar with Canon lenses, while you definitely are.
Basil, I now understand what you mean. I did notice 'curvature' effects in both lenses. Even on a flat surface I could use LIVE VIEW to manually focus on the center and the edges would be soft (as in the tests). And then I could focus on the edges and the center would soften up. This curvature does not happen with the 24mm TSE lens though.
I have done some photos with the 17-40 and focused on the center of a composition at f11, and then refocused for important edges (e.g. foreground) and shot again at f11 for optimal sharpness and then blended the best parts in Photoshop. The easier way to get acceptable performance edge-to-edge without having to do two pixtures is to stop down to f16 but not beyond with both wide zooms. Or use the zooms on a small sensor camera to forget the edge problems! Darwin
Discussion
7 replies1) Sharpness beyond f/16 reported to be poor. I'm surprised that there's no statement like "this is to be expected due to diffraction" or something like that. As is, it sounds like a flaw that could be avoided by better lens design.
2) Have you (Darwin) tested whether the poor corner performance of the 16-35 at wider apertures, as tested on a flat subject, is due to field curvature?
I think these are two bits of information that could make the good review even better. What are your thoughts on the subject?
Best regards & thanks to everybody writing articles!
Basil
Performance wide open and beyond f/16 for an ultra-wide lens will tell me nothing useful since I will likely never use these settings.
For what it's worth, I did a similar test with my 17-40 and a friend's 16-35 II. I could not tell the differences between the files (shot at at f/16).
Guy
I have been contemplating buying the 16-35 II or the 17-40 for a few weeks since I broke my 24-135. I think I'll get the 17-40 for backpacking. Thanks Darwin!
Yes poor lens performance beyond f16 is due to diffraction - I did not explain this in the review so thanks for mentioning it here. All 35mm lenses I have ever tested have poor performance at f22 (or whatever the smallest aperture is on the lens), enough that I can see a difference in prints. I no longer use any of my lenses stopped down all the way but at one aperture open from the smallest aperture (e.g. f16) - for those who want more info see:
http://www.luminous-landscape.com/tutorials/understanding-series/u-diffraction.shtml
http://www.bobatkins.com/photography/technical/diffraction.html
http://www.naturephotographers.net/articles1207/bc1207-1.html
I tested edge sharpness of the lens using flat surfaces - newspapers tapped to wall, graffitti on a wall etc - and the results were the same as the rock photos posted (which is also a flat surface) - so field curvature should not have been a factor in the results. The simple fact is that both lenses suck when it comes to edge sharpness on a full frame camera especally at wider apertures. I posted the rock photos as most of us shoot nature and not newspapers.
In the end shooting real scenes (landscapes), I could not see any real benefit of the 16-35. Remember there is reportedly much variation in the quality of the same model of lens - sample variation so to speak. There are poor versions of both lenses out there. I don't understand why this should be the case shouldn't quality control be expected with lens manufacturing?? Darwin
About field curvature - I'm not sure whether we're talking about the same thing (might be a language issue on my part, I'm not natively English speaking). What I'm referring to is that the focus field of many wide angle lenses is not flat, but curved towards the camera at the corners. So if you "overfocus" beyond the optimum point for center sharpness, the corners should become sharper if the reason for corner softness is the shape of the focus field. The Nikon 17-55/2.8 DX has a reputation for it, and some suggest to focus it beyond infinity to improve corner sharpness when shooting lanscapes.
Maybe my hypothesis is plain silly, though. I'm not so familiar with Canon lenses, while you definitely are.
Basil
I have done some photos with the 17-40 and focused on the center of a composition at f11, and then refocused for important edges (e.g. foreground) and shot again at f11 for optimal sharpness and then blended the best parts in Photoshop. The easier way to get acceptable performance edge-to-edge without having to do two pixtures is to stop down to f16 but not beyond with both wide zooms. Or use the zooms on a small sensor camera to forget the edge problems! Darwin
Basil