WAY OFF. Here’s another one from Thomas Knoll: "Actually, to create a camera filter set that is "perfect", it is not required to exactly the match the human cone responses (or the XYZ responses). All that is required is the filter responses be some linear combination of the human cone responses. If that is the case, then a simple 3 by 3 matrix can be used in software to recover the exact XYZ values.”
Thanks for entering the thread. Yes, I remember that answer by Mr. Knoll, which was in response to a question that I had posted. I had presumed that the digital camera filters would get best results when they matched the response of the human cones, but fortunately this stipulation is not needed and camera makers have some latitude in the filters. Unfortunately, he added that current digital camera filters are not linear and so we don't get perfect matches.
From my work with DCRaw, I saw that the source code contains this note: Thanks to Adobe for providing these excellent CAM -> XYZ matrices! The coefficients in the matrix are listed for various camera models. I presume that when one calibrates ACR with the Fors script or manually by Bruce Fraser's method, one is refining the values of these coefficients which describe the characteristics of the camera filters.