The APS I have uses Logo CMS and tables to create profiles from the measured data that is similar to Profile Maker 5. The details on the exact differences are not disclosed to anyone I know, nor to myself. Yet working with Logo/Gretag-X-Rite for many years I do know where to look, even if I don't have the same regard as a colour scientist would.
The HP proprietary profiler is a quite interesting application with a lot of hybrid math going into it to make very acceptable profiles for a medium number of patches with calculations for expected deviation and or phenomena encountered with specificities of HP inks. Quite remarkable actually.
Now the APS is a Logo thing. OR let's call it X-Rite as they are the owners of the intellectual property since buying Gretag MacBeth.
X-Rite do things quite differently than HP in this regard. The tables the grey builds, tags, smoothing, optimisations, etc etc.
Which ones are better? Well it depends. I found that the HP app or Color Center if you must is better at non photo papers. IS it more accurate or more precise than APS? Uh no but neither are accurate.
The more diffusion you have the more you are changing the difference between a 2º observer reference when you are actually looking at the print. So it comes down to a much more practical call: subjective analysis of the rendering through the entire system.
Photo media. Well it's not too hard to run a loop to see which one is more probable at reducing overall dE. APS is much better in all the tests I did.
Yet what are tests good for? Just get yourself a bunch of images with colours that approach the boundaries of the working space, and the printer. Print a lot, look closely. IF you can't find any differences keep looking. OR if you're satisfied then stop and have fun; print.
I did have a hard time with the soft proofing A2B 0 tables in the Color Center profiles though. Couldn't use them for soft proofing. X-Rite aren't that much better.
In the end the ideal would be they program Argyll into the printer since after all it is Unix. Someone in Canada was trying to set up Argyll with a SDK for the Z . No idea where that went.
Next best thing would be Monaco Profiler Gold.
Yet for the price of a nice i1D2, monitor and printer, and outsourced automated profile measure and create, the APS package is very reasonable for what it allows you to do.
While it may be true that the optimised HP media probably do profile better than non optimised media, it is simply not true that on a similar quality paper and surface there is anything special going on with a media brand agnostic spectrophotmeter.