Originally posted by monochrome I have no reason to doubt you - Nikon reportedly has a custom sensor in the D810 - but for Sony and Ricoh to come to an agreement for a custom sensor either 1) the sensor is where Ricoh spent it's money, which might explain the little missing bits the trolls are obsessing on; or 2) Ricoh expects to sell one very large boatload of these sensors in K-1's (possibly over a long term in many models, such as with the 16Mp APSc sensor); or both.
In any event,
I expect people to be surprised by the IQ. I anticipate with glee all the grudgingly positive reviews the images will get before the reviewers move on to criticizing USB2.0, the video and whatever else is not-industry-standard.
I wholeheartedly agree with the bolded, but I would disagree that getting different performance out of the same processor as the D810 uses can only come at a high cost to Ricoh.
A little knowledge may a bad thing, but I happen to have a little knowledge of digital circuitry, so here I go, for bad or good. The output of the same sensor assembly can be tweaked or customized by changing instructions for the processor that reads the sensor. It isn't a one-way process, the bias voltage applied to photo-detectors can changed, active reset of photo-detectors implemented, the sampling rate changed and tolerances adjusted for digitizing the analog output of photo-detectors. Even if Sony only provides a static interface to its sensors, Sony's engineers can modify the embedded programming to suit different customers. It doesn't require different dies or x-ray masks; depending on the type of embedded processor(s) used, the programming can be changed for every processor that gets connected to a sensor on the assembly line without requiring tooling or process changes.
You could ask why Sony can't necessarily get the same results out of a particular sensor that other camera manufacturers use; I'm sure it comes down to what final performance characteristics are deemed to be most important and how engineers think they can exploit the observed physical characteristics of the sensor to get the performance they are looking for. Nikon's engineers probably have different objectives and different experiences with Sony sensors than Pentax engineers do. To me, it's a strange process, for better quality control they are trying to eliminate any variability in how a particular sub-system works and to differentiate the finished product they are trying to find a "recipe" that takes advantage of variability in that sub-system.