Here's a question that perhaps only PE can answer (but anyone is welcome to try):
The curves below show two XTOL strips averaged together (red), and two strips from the formula used for the wedding (green and blue):
Note the right quarter of the graphs. The red XTOL graph rises only slighly in the right quarter, but the two wedding graphs have a more pronounced rise there. My earlier developers with less buffering (and AA) don't do this. But I do see it in one earlier dev with more buffering (and AA), as well as the latest two strips. So I think that right-rise is real. I'm guessing it's caused by either (1) boosted buffering, or (2) boosted ascorbic acid (AA) which is a secondary developer. This developer is rather dilute, so I'm surprised to see a boost in highlights instead of compensation that one normally expects with dilution.
Any idea why boosting buffering or AA would cause highlights to have higher slope than shadows and midtones?
Can the curve be linearized?
I was thinking of adding KBr, but that's not soluble in PG. Maybe benzoic acid? Other ideas?
I know the right-rise is small, but I'm a perfectionist who'll go for linear if possible.