Such a ratio (and, obviously, it's calculation) is fundimentally flawed as a representational concept. I've got @ 1000 CDs and half that many more LPs, but I can only think in terms of getting my mind around any one piece of music through listening at any one time. Furthermore, if I only had one piece of recorded music, I would still want to hear it reproduced in the best way possible.
Conversely, an individual can get the hardware to listen to a piece of recorded music in what anyone would objectively consider "High Fidelity" for under $1,000USD in a headphone system or under $1,500USD is a speaker based system in the U.S. today (much less if good pre-owned equipment was obtained). With such a hardware system, the listener could enjoy that self same library of 1500 musical recordings to 93% of their capability until the cows come home.
Two very different ratios (and pictures) emerge; audio hardware and software can be related in vastly different ways without successfully reducing their relation for any individual to such a ratio.
Consequently, I would conclude that you are going to need another metric to make the type of measurement you apparently want.
For what it's worth, I'm presently at about 2.5 favoring hardware, but my software is in a constant state of churn and storage space for it has some limits.