Upload & Sell: On
My plans for this weekend included taking a bunch of sample images with my various R lenses and writing up a nice report about the differences that I found. I took some sample images this morning off my balcony to get a feel for what I should specifically be testing. What I found was both disappointing and pleasing:
They're all good, but none of them stand out as obviously better and none of them are obviously worse. While I didn't keep track, I think every lens was best at something and worst at another.
I only shoot digital with my NEX cameras, so the Speed Booster or crop throws a wrench into things, adding another reason for not completing my project and providing lots of sample images here. If there's interest, I might repeat the process should I be tempted by the next round of Sony cameras...although that will mostly depend on their ability to deliver a rangefinder-style viewfinder, which my left-eye dominance prefers.
The 21-35 is technically the least interesting of the bunch. Quite a bit of field curvature away from the camera toward the corners leads to blurring at high resolution. The MTF curves tell the story; astigmatism is an issue here. However, at low resolution, the image is quite nice out to the corners and the optical flaws never get in the way of the story of the subject. So I'd say that this is a great documentary or storytelling lens rather than a landscaper's tool. This remains my favorite lens, simply due to its ability to provide a good range of wide-angle views and to get out of the way of the subject matter.
The 28-70 doesn't deserve its poor reputation. Apparently, this later version (which has the same body style as the 35-70/4) was slightly adjusted for better performance at the wide end. Regardless, take a close look at the engineering diagram of the 28-70 versus the 28-90 and you'll see that they employ the same mechanical characteristics. The 28-90 is the successor of the 28-70, and it makes no sense to me that they receive such disparate reviews. In my tests, the 28-70 has the flattest focal plane of my four 35's and far gentler bokeh than the 35-70. There is a midzone dip due to field curvature, but that is only apparent when shooting flat subjects, which tend to make poor images; it has noticeably higher resolution into the corners with flat subjects than any of the other lenses at 35mm. Indeed, if I were to shoot architecture at 35mm, this would be my lens of choice. Bokeh is quite smooth. There's a bit of yellowish tint to the lateral and spherical chromatic aberration, but in general it performs better than the 35-70 as far as outright aberrations go. Yes, the 28-70 distorts; no, that's not really an issue with modern software, although I have yet to generate Lightroom profiles for it, and that's because I haven't found it to be that big of a problem for the kind of shooting I do. I bought this as a foul-weather lens, but it is actually quite nice across the spectrum.
The 35 is absolutely brilliant in the center and with out of focus subjects, but it has a strong to moderate fall-off in resolution toward the edges and then the extreme corners reveal pronounced field curvature. The Summicron 35 II has more gradual field curvature that is quite useful, but the Elmarit's curvature is just annoying. For the rest of the frame, it does very little that one can point to as wrong, it just simply fades away from the excellence of its axial performance. Iím curious to find out how much of this field curvature issue is caused by the combination of Speed Booster and NEX-7.
The 35-70/4 is my newest lens and I haven't had much "real world" experience with it yet. I acquired it for general purpose duty with an emphasis on close focusing (as my lady likes flower shots). Iím not surprised that it isnít categorically superior to the 28-70, but Iím still figuring out their differences. Field curvature is mild; rather flat through most of the frame, then gently bending away from the camera to the corners. Better midzone resolution on flat subjects than the 28-70, but not enough to make it superior for general purpose imagery. The major trait of this lens is its bokeh. It renders blur circles with a highlight on the edge of the circle closest to the image center. This generally avoids double-lined bokeh, but more interestingly is that it creates a false sense of resolution as a blurred line will in fact render a sharp line along the edge of the bokeh. This one-edged bokeh is intriguing to me and will probably be what I focus on over this summer. In terms of color, it feels ďtruerĒ than the 28-70, though I wouldnít notice in images not side-by-side, and it has considerably more purple fringing (though still relatively little in absolute terms). So, with subjects that benefit from hard edges, the 35-70 is preferable; for subjects that benefit from gentle resolution fall-off, the 28-70 is what Iíd select.
The 50 is well known. Lots of character wide open due to strong spherical aberration. Stops down to be my highest-resolving lens in the center at f/4 to f/8, but with substantial fall-off toward the edges. Like with the 21-35, this doesnít interfere with storytelling, but where the subject is along the edge of the frame, as when the photo is of the dawn/dusk sky and the horizon is along the bottom, either zoom will do better at large print sizes. For high contrast lines at f/4 throughout the frame, however, the Summicron will be nearly flawless while the zooms will struggle with fringing. Again, relatively low problems compared to more affordable lenses, but in comparison the difference is noticeable.
The 90 is also well known. As my oldest lens, thereís no surprise that it has lower overall contrast than anything else mentioned here. It behaves like a mild-mannered sibling of the 50; both strengths and weaknesses are similar, just less strong.
Anyhow, I did that assessment and wrote it up to help me clarify my gear strategizing, and Iím sharing it in case anyone finds it interesting. Sorry about not posting images in an image thread. :-)