[Zubersoft]

There are pros and cons to each approach. Using a CSV and sharing the same PDF over all the songs has benefits in terms of disk space, but due to some existing shortcomings in the CSV import, this can cause issues for some users, as the original PDF will be duplicated as the directory it is copied to is based on the title of the first selected song in the CSV import dialog. If you import all of the songs you need at once through the CSV import, then you get the maximum benefit, but if you import the same csv multiple times in a row, and select different songs each time, the original PDF is going to be duplicated, taking up additional disk space. I plan to address this by detecting if the original PDF is already in use in the library, and using the existing file in the library if possible instead of copying the file to a new location. 

If you instead split off the PDF into many smaller pieces, each piece is truly independent of the others, which can nice in some scenarios (like if you want to embed annotations without impacting other songs in the library that might share the same PDF). It just depends on how the library is setup and what features are needed

Mike