It's all about the lenses, right? Well, no, it's about making images... but lenses have something to do with that. I know fast autofocus can be useful, and rangefinder lenses have a following, but old lenses designed for manual-focus 135-format SLRs are particularly appealing. Why? Because a wide range of such lenses are optically good, mechanically impressive, readily available, and reasonably priced.
The key to using most of those lenses is having a very short mount-flange-to-sensor distance so you can fit a glassless adapter and still focus to infinity; DSLRs can't do it, but mirrorless cameras can. The really cool thing about the Sony A7/A7R is that either can cost-effectively put a high-quality full-frame 36x24mm sensor behind nearly any old SLR lens. Of course, you can sort-of do that on an APS-C mirrorless camera using a focal reducer such as the original Speed Booster (SB) or the cheaper Lens Turbo (LT). As I write this, a new 24MP APS-C NEX-7 body and LT can be had for about $1100, while a 24MP A7 body is around $1700. In fact, a 20MP APS-C A3000 body and LT can be had for about $500. Those price differences could buy some really nice old lenses....
For SLR lenses, is an A7 really a big step up from a NEX-7 with an optional LT? To find out, I've done a little informal evaluation of the same full-frame SLR lenses on:
- A7 full-frame 24MP
- NEX-7 APS-C 24MP with Lens Turbo (LT)
- NEX-7 APS-C 24MP
- A7 APS-C crop 10MP
- A100 APS-C 10MP
The ancient A100 DSLR is in there because it has a 10MP APS-C sensor, matching the sensel size and APS-C crop resolution of the A7. It doesn't have the high ISO performance nor video modes of newer CMOS sensors, but according to DxOMark, the A100's CCD is otherwise shockingly competitive with modern APS-C cameras like Canon's 70D. I'm not saying anyone should rush out to buy an A100, but it's a good test case here.
You might be wondering why I'm so interested in comparing APS-C crops. Well, it turns out that most lenses have image quality (IQ) drop significantly as you move off axis. A lens that makes beautiful image centers can have horrific image quality in the corners. Then again, some lenses produce image quality that is quite consistent across the whole frame. For some time, I've been wondering how often the "sweet spot advantage" of taking the the center APS-C crop might produce superior image quality (at the same total pixel count) as actually capturing the intended full frame. Testing just a few lenses will not give the definitive answer, but it is at least a start.
The specific lenses used for these quick, preliminary, tests were a 35mm f/2 Super-Multi-Coated Takumar, 50mm f/1.4 Super-Takumar, and 70-210mm f/3.5 Vivitar Series 1 zoom. These are well-known and well-respected lenses that easily can be bought at prices well below any new lenses of similar focal length and aperture. My total cost for these three lenses was under $200, and even a bad shopper should pay less than $400. Thanks to their M42 mounts, all three can focus to infinity on E and A mount bodies via adapters; most other old SLR mounts cannot be adapted to the A-mount A100. I tested with a glassless M42->A, glassless M42->E, and FD->E LT focal reducer behind a glassless M42->FD adapter. Here are the contenders:
|Left-to-right: A100 + 70-210mm, A7 + 50mm, and NEX-7 + Lens Turbo + 35mm|
It is worth noting that the test exposures were all made as ISO 100 camera JPEGs. The images were scaled down to 900 pixels wide for posting here. With the lenses wide open, exposure was near the brightest the cameras could handle, producing color and exposure differences that are more due to deliberate operator error than camera flaws. The stopped-down exposures look much more consistent across cameras.
On with the testing!