Engine dyno software. Anything good on the Internet?

Just in terms of the camshaft inputs -
Look at a lobe profile as measured. It will have the take up ramp, the exact duration at each point over the lift and closing ramps from zero lift to max. If you graphed it using points at .010" intervals, you'd have a picture of a hill with gradual increasing rate of lift from "flat", and a rounded top. That's reality.
Now look at what you supply DD: total duration (doesn't include the take up ramp or closing ramp), and duration at .050". Now plot those two points, plus max lift. Now you have a trapazoidal figure consisting of a straight line from the base circle to .050 lift, and a triangle from .050 lift to max lift.
The 2nd is what the cheaper simulators use and that's why the differences are muted between similar grinds.

Edit - adding to what you've noted in #23 - it simply means the changes you're making fall into the percentage of error of the program.