Sorry Nick. I don't read Japanese. But, I wouldn't believe that Sony's processing in the H9 is the first to take approach for a split second.
I've seen far too much evidence of raw files being pre-processed in the past, and this is not a recent thing (IMO, it's been done for at least 4 years now, based on my examining the raw files from a little Konica pocket camera I've got with Dave Coffin that when he was attempting to duplicate the results of the in camera processing while adding this model to dcraw.c).
Camera manufacturers are not stupid. They've got teams of engineers trying to figure out the best place in the image processing pipeline to reduce averaging errors due to the demosaic process, reduce rounding errors from multipliers being applied to RGB channels for color correction, reduce posterization by avoiding bit depth constraints of variables unique to the final file formats, avoid color space constrants by processing prior to those constraints, and more.
Manufacturers may *think* they've got a unique approach to something, when they don't I've seen that over and over in the past, too. Heck, I can remember laughing at defendor's of Nikon's decision to encrypt metatadata related to WB (implying that the storage techniques were proprietary in the way they have a matrix designed to account for more than one color temperature in the image). Not. It's not unique (Dave Coffin reverse engineered the matrixes for the multipliers by using a variety of camera settings and lighting temperatures and determined how it worked by what was changing in the raw files, and made it obvious to me that they were not unique in that approach.
My conclusion, it's wishful thinking on the part of Sony's marketing staff if they really claim to be first at any kind of preprocessing of raw data, especially since the way the data is being processed is not published by other manufacturers.