Logged In: YES
user_id=616697

Perhaps a better approach would be to have multiple
independent question fields (front, front1, front2, ...)
each with its own 'Alt spelling' field for the audio clip,
and multiple independent answer fields (back, back1, back2,
...).

By default, everything works as originally intended, but if
a custom mapping template is activated then the user will
have any combination of question fields displayed on the
Front side of the card, and on the Back, one of the answer
fields available.

So, the standard template would be:

Q A


front back
-------
example

But then, if you wish, you can put together a custom
template with front=Introduction, front1=Definition,
front2=Case and then have a choice of grouping:

Q A


front back
front1 --------
front2 example

Or, perhaps,

Q A


front back2
front2 --------
example

or whatever.

A Profile Editor would allow the user to create/edit various
profiles by listing what range of fields is available and to
choose from.

Granule will play audio clips of each of the `field' in Q in
a sequence.

This would seem to solve your problem, and in addition, give
a chance to all Japanese (kanji, kana, romaji) students to
use their learning material effectively with granule.

I'm only concerned with two things: 1) should I store
profiles along with each cardfile or make them a part of
general preferences? 2) where to find some free time to do
all of this?

I haven't given a thought to the scoring system you've
proposed - I definitely don't want to change the way it
works right now.