Get the Olympus OMD E-M5 Mark II digital camera at B&H Photo.
Whether or not you have or plan to have an Olympus OMD E-M5 Mark II with its hi-res sensor shift mode, my latest EM5 Mark II hi-res mode evaluation should be intensely interesting for the promise it holds out for ultra-high resolution images from any brand that might implement sensor shift technology. Barring something horrible: some sort of patent standoff that keeps other players out, and thus deprives us of this tech.
BTW, I don’t much like the form factor of the E-M5 Mark II: the grip is just lousy for my relatively large hands; I much prefer the grip of the E-M1. An E-M1 Mark II would solve that.
Sensor shift already has existed for some years now in very high quality and at very high cost: medium format Hasselblad digital backs, intended for studio work).
But it is Olympus that deserves kudos for delivering a version of the technology at a lowball $1099 price point. The Olympus results do not deliver the quality of the Hasselblad system by any means. But the writing is on the wall—more and better implementations are sure to come. And it is far easier technologically than to linearly double or quadruple sensor resolution.
And so I now consider two technologies must-have features that ought to find their way into all cameras: (1) sensor image stabilization along with (2) ultra high-res sensor shift technology. Consider the following if implemented only as far as Olympus does it:
- 24MP full-frame sensor could generate 96MP raw files delivering something approaching 60 megapixels of detail under ideal conditions.
- 36MP full-frame sensor could generate 144MP raw files delivering something approaching 90 megapixels of detail under ideal conditions.
- 50MP full-frame sensor could generate 200MP raw files delivering something approaching 125 megapixels of detail under ideal conditions.
The above is only true for tripod based shooting with no camera or subject movement; it is not a general solution. However, with smart enough hardware, even handheld operation could be possible with powerful image processing hardware (semi-randomized shifts by some camera movement handheld, combined for one better image, this is actually already implemented in specialized software programs, but I for one want the camera to deliver to me one raw file).
With appropriate design, the camera itself could be smarter than what Olympus delivers, e.g., delivering smaller raw files large enough to include the extra detail, but not wastefully large files containing much less detail than the numbers would imply. By using the camera itself to process appropriately. Better to get 95% of the gains with 50% of the file bloat in my view.
Since the Sony A7 II already has image stabilization and a Sony A7R II ought to have the same stabilization, the possibility of a firmware update for the A7 II which supports sensor shift technology is intriguing. Actually, it’s startling that Sony has not already done so, but perhaps there is a hardware limitation or some other reason. Also, any A7R II would have to move to an EFC shutter and eliminate the sharpness-destroying shutter vibration.
The sensor-shift approach used by Olympus is a type of oversampling, an idea that I have long advocated for higher image quality.
Oversampling using a double-resolution sensor would be much better than shifting the sensor, but this happens slowly over time. However, an ideal existence proof for the validity of oversampling will soon present itself: comparing the 50-megapixel Canon 5Ds R against the current 23-megapixel 5D Mark III in June or July 2015, depending on when Canon delivers.