For some, possibly most, needs of using raw having yuv444, 14+ (or 16+) bits of precision and lossless or near-lossless (with strong guarantees of maximum error on individual pixels) compression is supposedly acceptable. I base this interpretation mostly on digital photography discussions I have seen on the internet. I am not a photographer myself.
(12 bits of component already requires adapting to the current lighting and can supposedly slow down professional photography.)