[sciurius]

... my code is still set up to dynamically determine the delimiter. It just looks for whatever character is used to separate the columns on the first line.

That is a rock solid approach, given that MSPro requires the headings line.

Summarizing.

The CSV is a text file, UTF-8 encoded.
Records (lines) are separated using any of the established line endings (LF (ASCII 10), CR (ASCII 13) or CR+LF) but only one of these line endings should be used for all records. The last record of the file should have a line ending too.
Fields are separated by a separator character that defaults to comma (ASCII 44), but will be dynamically determined while processing the first (headings) record. Fields that need to contain the separator character should be quoted using double quotes (ASCII 34). If a quoted fields need to contain quotes itself, see RFC 4180.
The fields of the first record (headings record) designate the names (interpretation) of the fields of the rest of the records.

As a minimum, each record should contain the name of a song title and the page range.

Given that page ranges frequently contain commas, the choice for comma as separator is suboptimal since it will require many page ranges to be quoted. Therefore it is advised to use semicolon (ASCII 59) instead.

@aspiers: Would you consider to include the above in your README?