The internet is abuzz with rumors of a full-frame Sony NEX, a development that seems likely. I share a few brief thoughts here on that very near term prospect.
Sony sensors offer sensational image quality. With the Nikon D800 sensor already excellent, what has another two years done for sensor quality at high resolution?
Might we see a 32 megapixel or 36 megapixel full-frame NEX, perhaps one done right and optimized to operate without an anti-aliasing filter and relatively thin sensor cover glass to avoid off-center ray angle issues?
The lens quality would have to be exceedingly high to deliver to such a sensor, which probably means f/2.8 lenses, f/2 and f/1.4 being a serious challenge and perhaps unacceptable in size and weight and cost, at least if high performance is a goal for full performance at the first two aperture stops.
Given the performance seen with the Sony RX1R / RX1 (whose design is now over a year old), Leica might consider the emergence of a full frame NEX to be an existential question*. But don’t expect Leica M lenses to perform well on a full-frame NEX sensor; ray angle issues are likely to cause severe image quality deterioration.
* The neck jewelry market aside, Leica should stick to core competencies: lenses. Let Sony and others make cameras for Leica lenses.
The Sony RX1R (and RX1) show that a lens developed for a specific sensor can offer quality equal to and perhaps exceeding the very best lenses available. Several factors contribute to the high quality:
- The lens can be optimized for specific sensor characteristics: sensor resolution, thickness of sensor cover glass, the presence or absence of an anti-aliasing filter.
- By being integral to the camera, the rear of the lens can be recessed to within a millimeter or so of the sensor and/or the edges of the available space. This keeps the apparent lens size smaller as well as reducing certain physical constraints on optimizing the lens design. Nor is any external lens barrel needed for the portion of the lens inside the camera.
- Reliability can be high: no contacts to be made between lens and camera; no gaps and alignment tolerances to the mount; it's all built together and presumably can be tested for quality together; no issues with a million lens samples on a million bodies, each of which can vary slightly and/or change over time slightly.
Which brings us to a full-frame NEX with interchangeable lenses: so far, no interchangeable lens projects into the camera body, the entire lens requires a proper lens barrel, and there are tolerance issues. And possibly optical compromises too: the flange offset, unknown sensors with unknown properties, anti-aliasing filter or not, and so on.
So it is an open question whether (for example) Zeiss could design a 35mm f/2 lens for a full-frame NEX that matches the performance of the 35mm f/2 Sonnar built into the Sony RX1R. My guess is that such a lens would need to be f/2.8 to meet the quality and size requirements. A new flange and mount standard might ease certain design issues, but high performance won’t come cheap. Still, AF lenses (instead of focusing helicoid) and flange/mount friendliness might allow prices that are significantly lower than DSLR Zeiss lenses.
It remains my assertion that two fixed lens cameras offers the near ideal situation over one camera body with interchangeable lenses of variable optical quality: for most applications two focal lengths is plenty while providing redundancy. In the field, I have never found more than three lenses to be particularly useful; two tends to work really well. More is less.
I like the results from the existing APS-C Sony NEX-7, but I am not thrilled with the ergonomics or the idle lens gyration focusing behavior (a distracting nuisance), or arbitrary limitations on exposure time or App Crap that cannot be removed and so on—things that make it feel unserious, things that get in the way. Will Sony address the “dissatisifier” issues and will we avoid tacky advertising lingo printed right on the camera body?
The addition of the latest Sony EVF technology integrated into the body would be highly appealing and highly appropriate for a full-frame NEX. Let us hope that a separate EVF is possible but not required.
Image stabilization akin to the Olympus E-M1 / E-M5 on full frame would be a compelling advance, a “killer” feature for some uses. Imagine super high ISO quality combined with several stops of in-body image stabilization, including for video. Now that’s something exciting for every user: no optical degradation when in-body IS is used, allowing the best possible lens performance with any and all lenses.
Would the camera body be pro quality like the new Olympus E-M1? Or some intermediate consumer grade build? A body costing something in the $2500 range might be disappointing as a medium and long term investment if it is akin to the NEX-7 build quality (which is good, but far from ideal).