Back in January in Sony CMOS 44 X 33mm Medium Format Sensor: Oh, the Possibilities, I commented on the new Sony 44 X 33mm 50-megapixel sensor. Now, Sony Alpha Rumors suggests that there might actually be a Sony camera in the works to use that sensor.
What a 44 X 33mm sensor might mean?
Some overarching thoughts on Sony’s strategic position
Sony is very strong technically and could come to dominate, but is very weak in other areas like understanding and supporting professional users.
- Sony is at the vanguard of innovation in mirrorless. A medium format and relatively compact mirrorless camera would be like a steamship at a time of slack sails on a digital sargasso sea. A super premium offering with no competition in either thought or tangible reality makes a ton of sense for a company looking to extend its leadership in the mirrorless arena (heck , Fujifilm is still stuck on APS-C for at least another year or so it looks).
- Soporific Nikon and Canon designs are 3 to 5 years stale, with those companies asleep at the helm of their respective DSLR Titanics. Shove a little iceberg in their direction.
- The Sony RX1/RX1R offerings are strangely lacking in an upgrade: where is the 36MP option, bumped up incremental improvements, etc? So it would make sense to bump up those models along with a medium format premium offering later this year, thus offering a highly attractive model line with a clear camera-lust path.
- It might be that for all of the above, Sony is waiting for a key new part, such as a 4MP EVF, a faster camera CPU, production yields, etc. Frame rates and processing speed and 4K video are all challenges.
A medium format digital mirrorless seems unlikely before late 2014, but that is plenty of time to upstage rivals.
Technical and cost issues
Given Sony’s wide-ranging parts bin and in-house development of sensor and EVF and CPU and so on, a $7000 price tag for a 44 X 33mm sensor mirrorless camera would be an aggressive price point, using the Sony RX1R as a reference. That would be a bargain compared to the cost of the PhaseOne and Hasselblad systems. The EVF would surely be built in, since the camera would have to be relatively large.
Update April 11: I wrote that $7000 figure above a few weeks before the Pentax 645Z pricing showed up, apparently $8500. I still figure that the $7000 price range is about right, though some readers insist on the admirable fantasy of under $3000.
- Sony expertise and volume might be able to get the sensor cost down to $1500 or so if yields can be high and dead pixels and similar are allowed at a less rigorous level. Presumably the best sensors go to PhaseOne and Hasselblad (higher quality requirements), those that miss the cut go into Sony cameras. Grade B sensors can cut the cost down quite a bit, but a cost below $1000 seems unlikely unless somehow Sony really can manufacture in very large quantities (10,000 or more units). That’s a lot of cameras; it would be interesting to know how many RX1R cameras were sold.
- A camera price of $5000 is unrealistically low unless Sony decides it is a loss leader for brand image. Even $7500 is aggressive given sensor and lens costs, at least if imaging quality is the goal. The lens could be a major cost, even more than than sensor cost because to cover a 44 X 33mm sensor with high quality means a lens that would have to cost somewhere around $3000. Built-in and with a relatively slow aperture that cost might come down to $1500. Less than that, and one can forget about peak quality.
- A 44 X 33mm sensor has a 55mm diagonal versus 43.3mm for full-frame. This is substantially larger and makes an ƒ/2 lens unrealistic; it will have to be ƒ/2.8 or even ƒ/3.5 to mitigate cost/quality/size issues. Realistically the lens would have to be about twice the size even being a stop slower, and it will cost at least 2X if not 3X the price to achieve similar quality across the much larger area (keeping focus shift, field curvature and similar issues in check is hard).
- A focal plane shutter would be a disaster for vibration. Hence a fixed lens solution as in the RX1R with a leaf shutter allows for zero vibration, a lens optimized for the sensor and recessing of the lens into the camera body, which keeps the size reasonable.
- It is not realistic to have interchangeable lenses (ILC) with high performance at costs most users would find acceptable (think $4K to $6K to do so, per lens). Unless something 'gives': optical quality or lens speed (think ƒ/4). Moreover, these would imply a focal plane shutter unless the lenses were leaf shutters, which would likely add $1000 to the cost of each lens. Hence it is much more realistic to see a fixed-lens solution.
- PhaseOne and Hasselblad might not be keen on seeing an ILC competing directly. Hence the fixed lens solution is a way to avoid that conflict while simultaneously allowing higher image quality at a lower price in a more compact package.
- To be taken seriously, this new sensor needs lossless 14-bit files, not lossy 11+7 bit compromised images.
The sensor size of 44 X 33mm is 68% larger in area than a full frame sensor. Many pro shooters like that “medium format look” which is often thought of as a focal length issue but derives primarily from lens design (as the Zeiss 55mm f/1.4 APO-distagon shows).
To have the total appeal, a full-on Zeiss lens design for a 44 X 33mm sensor made to medium format professional quality standards would be the camera-lust scenario. Such a design would not be inexpensive, but by making a camera a fixed lens mirrorless camera with a lens optimized for the sensor, certain design obstacles drop out of the way.