I guess that I see what you see: CCT is 500K higher in ColorHCFR and gamma is approx. .2 lower in ColorHCFR.
However, I'm also a bit concerned that what we are looking at here (the data from ColorHCFR) are not comparable to the data from the LaCie software if the ICC profile is not activated for playback of AVCHD when measuring with ColorHCFR.
This could explain the lower CCT in the LaCie calibration report *if* the windowed applications (e.g. the LaCie test patterns) do not use the same ICC profile as full screen applications (like most DVD/BD players).
Also we cannot be sure that the grayscale behaviour is identical even if the same or no ICC profile is used for windowed *and* full screen modes (but hopefully it is).
Apparently it is not possible to use the LaCie software with other test patterns sources (I have browsed through the manual as well) and even if we can reset the ICC profile we know very little about the test patterns generated by the LaCie software.
I am quite certain that the patterns generated by the LaCie software will not produce the same result as e.g. the APL patterns for AVCHD as the panel load (and therefore the light output) will likely vary with luminance diffferences in the patterns produced by the software.
To find out more we really need some well-defined test patterns to be used with your PC using the same ICC profiles and screen mode as the LaCie software.
Maybe you could playback AVCHD in windowed mode (rather than full screen mode) of the PC with the LaCie software when using ColorHCFR to measure while attempting to load the panel similarly to the LaCie software (e.g. by scaling the player window)?
You might need to take into account that black levels are probably different between the two sources (windowed mode will probably use PC levels (0-255) while AVCHD uses video levels (16-235) which the windowed player may or may not be able to map to PC levels).
This way you *might* get more insight into the differences between LaCie and ColorHCFR. At least the whitebalance/CCT data should be comparable. If the whitebalances differ between the two software packages we might need to investigate more.