Originally Posted by Doug Blackburn
I know you don't want to give away any details about AutoCal, and I understand we have the option to tighten the error limit for AutoCal, but do we have any other contol over AutoCal? For example, is there any way for us to change what AutoCal uses to compute errors? Experienced calibrators know that varions dE calculations produce different results for the same measured set of coordinates. Because of that, some of us look at 2 or 3 different dE calculations as we do manual calibrations to insure that ALL of them are as low as possible for each measurement. Is there any user control over what error calculations (thinking more of graycale and 6-point CMS here, this doesn't necessarily apply to the Cube/matrix calibration) are used by AutoCal. For example, If I choose dEuv for error calculations shown to me in CalMAN, does that cause AutoCal to use dEuv (instead of the default dE2000)? If there is no user control over the error calculations AutoCal does, is it doing anything internally to insure it doesn't get into any situations where, say dE2000 produces an error of 15 while dE 1994 produces an error of 3, while dE 1931 produces an error of 6.
We are looking more at the actual light measurements and trying to get as close as possible to the target.
The threshold we quit at for early termination is entirely dependent on the dE formula you use.
When we can't hit the dE, or the dE is set to 0, we basically run till every control is one click over or under the target.
The AutoCal does typically use a 0.3 step size. That's because it is an 8bit change at that point and not all displays respond to smaller changes. Also with the step size at 0.3 we get dE's that are sub 0.5, so we do get to the point where the measured value is less than 1 8bit step away from the ideal target value. I know we are always looking to do better, but I think when we get well below human perception, it's good enough.
Of course you can always go back in and tweak the DDC controls if you want to use the native step size and a dE of 0.4 isn't good enough for you.Edited by sotti - 10/10/12 at 7:44pm