"breathe" is in kanji 息 (reading is "iki").
Usage of "a -" and "a R" markings (aliasing) variates voice bank to voice bank but nowadays using "a R" for "a's ending breathe" is semi-established ("a R" is usually sounding like a strong "AHH!" sound)
The best way to record natural sounding breathe samples for own UTAU is that...
1) ...you open Audacity and you start to sing an song as acappella
(any song is ok and don't worry about your singing skills).
2) After that, select breathe sample only and then from File => Export Selection...
3) Export your breathe sample as 16 Bit wav file (and name file as "breathe.wav/br.wav/息.wav/anything you wish )
4) If you want to have more than one breathe sample, you can re-do steps 1-3. You can record one "normal breathe", second as "long breathe", third as "fast breathe" and so on.
I have a made reeeeeally long time ago a short of tutorial of how to make breathe samples work in UTAU without crash it down but it's little bit outdated
holy my English grammar back then, I really need to re-write it some day buuuut the tutorial is still working. If you're interests in, you can read it
here!
[doublepost=1490300538][/doublepost]I advise personally to add breathe samples afterwards in mixing stage (aka breathe samples are not rendered throughout UTAU), so breathe will sound for sure "natural". UTAU sometimes makes some breathe samples sound OK and sometimes, errhhh.... NOT.
So it's easier to add breathe afterwards at mixing stage than trying to use in UTAU but it's up to you. :v