So, I have a .quad curve that gives lab-linear readings when given an untagged image of linear RGB patches and I print using No Color Management. To print my photos I could (but won't) convert them to use a lab-linear ICC profile (adjusting the pixel values) and then print with No Color Management again.
Or, I could (and will if this makes sense) follow suggestion 4 and take readings of my final target print, construct an appropriate file from the readings, drop it onto QTR-Create-ICC and make an ICC Profile that characterises the non-linearities left in my curve. I would then select Application Managed color management and select the created ICC profile. My photo files could be tagged with any profile after that.
Is my new understanding correct? If not I am probably more confused than ever :)
Just to clarify another point; is the reason that the generic "QTR driver Gray Matte (or Photo) Paper" ICC profiles are sub-optimal because they won't characterise the non-linearities in my curve and they don't know what my dmax is when working out a Y -> L* conversion?
Also, at one stage I installed a program called ColorFaker which replaces the default OSX sRGB ICC profile to allow some programs to work properly (can't remember why I installed it - google told me to). I was accidentally printing an untagged image with QTR and Print-Tool with No Color Management with ColorFaker turned on (the profile replaced) and the results were different than with it turned off. I was wondering where in the flow of things the system sRGB profile came into it.
Thanks again!