Since starting singing lessons with Enrica, I have relied on Finale to help me study and understand the music. Since Enrica is not a pianist, I need to have some sort of backing track to practice with, otherwise I shall have great difficulty when I need to rehearse with a real live accompanist. For songs with a complex accompaniment, I need to know both my own part and the accompaniment, to avoid being totally confused about where I should sing and where I should be silent. Some of the music books come with backing tracks, and sometimes these are usable. For instance, the standard accompaniments to the Vaccai exercises are very good for practising, because they are faithful to the score. I can adjust pitch or tempo in Audacity when necessary.
When the aria allows for more freedom of interpretation, using a "canned" accompaniment doesn't work so well anymore. During lessons, things get added to the music, such as where to slow down a little, where to breathe and where to pause. A case in point is Verdi's aria "Caro Nome", from Rigoletto. It contains a fiendish cadenza and several other passages that are open to interpretation, and no two singers sing it exactly alike. So I bravely or foolishly thought I would make my own rehearsal track in Finale. Foolishly, because there are a great many notes to enter, and technical problems that require workarounds. It is a big job that requires enormous attention to detail. That isn't totally bad, because in entering the notes for the vocal part and the accompaniment you need to think about them, and so you learn. Finale has a feature called "human playback" which sometimes does amazingly well, and at other times produces very strange results. That means customizing the score, the articulation marks and the interpretive text, to ensure that the result is the way we want it.
When it comes to the cadenza passages, entering the many tiny notes are a problem, because I find that the note values in the measure don't add up to the time signature, and so you get to a point where you can't add any more. If you make them all grace notes, the playback sounds all wrong. You have to enter them as real notes to get the correct effect. That means changing the time signature for that bar, then carefully working out the note values for the cadenza as it will be sung. If you compare the performances of different sopranos, you will find that they only use the score as a rough guide for the cadenzas, and assemble something from the traditional options, some of which are written up in Ricci's book, according to their interpretation. It's a very individual thing. I allow myself to be guided by my teacher.
My way around entering this idiosyncratic music into Finale is to change the time signature for the cadenza measure to something really generous, like 20 over 4, and then enter the notes as I sing them. Sometimes the runs don't fit and i need to introduce tuplets that aren't in the score. Once the music sounds correct to my ear, I delete the extra rests and change the time signature again, just for that bar. The piano accompaniment may need to be adjusted to ensure that the chords correspond to the correct notes.
This has been a steep learning curve, but well worthwhile.