I tried the first Tritest over the weekend. I didn't have time to go exhaustively through a lot of planes or any testing to determine the why's of the behavior.
First, overall, it is MUCH better than before. It responds to the X, Y, Z axis and with the X axis actually moving with head movement. Z axis seems kinda wierd, either not moving at all, then moving a lot and somewhat unpredictably. Y axis is the strangest of all. Head movement side to side on the X axis seems to be very predicated on Y axis height. A small change on the Y axis can limit X axis a lot as if the canopy were stepped like a ziggurat and not a smooth curve.
The strangest stuff was in changing planes. It worked well with the P51D with just some oddnesses. If for any reason I took the headset off, say to step away from the computer for a bit, on returning the view would lock strangely for a second or two, then move very notchily. This would go on for a minute or two and then smooth out, as if it was reacquiring the headset.
When I tried the Fw190D, it moved very strangely. The movement lagged behind actual head movement and was slower than in the pony. Then it would bounce when it reached the end, rebounding slightly backward in the direction it had come.
Strangest of all was flying the P47M. I would center the view and then fly along looking around. After about a minute of looking around, my view would gradually sink down until I could barely look over the lower edge of the cockpit and my virtual head was lower than the bottome of the gunsight. I'd re-center and it would pop back up, only to gradually lower as I looked around.
I'll try to give the new file a try somewhere during the week and report back.