Recap:
- Where did the 1.2 power value come from in the original default?
- @nick : we could look at the half-float quantization for the upper limit to make sure invertibility is functional
- @Thomas_Mansencal: we should put that nuke script and collab for the default values in the repo for tracking
- @matthias.scharfenber will calculate the effective primaries of a few DSLRs from the DNG matrices to see if they would make any difference to the limits
- @sdyer checked with @Alexander_Forsythe about the zero-clamp in the ACES to ACEScg CTL - Alex doesn’t have too much more information on the history - we all agree it should be gone. Scott is going to push a hotfix now, and talk with Alex about official release (along with the gamut mapping work?)
- @sdyer noted that the CC24 RGB values in the spreadsheet Nick used are the ones up for inclusion in the revision of SMPTE 2065-1-004.
- The group seemed fine with the default choice of 3 decimal places of precision on threshold and distance limits