So, you want to record a Japanese UTAU.
And why wouldn't you? UTAU was developed especially for Japanese voicebanks; most songs with USTs are Japanese. It's simply the most popular language to use in the program.
The problem is that most of the overseas UTAU fandom doesn't know Japanese, and finding reclists can be difficult when one doesn't know what to look for. Even hiragana charts can miss some of the useful stuff, and the fandom ends up with voicebanks that either lack necessary sounds... or one that has useless ones. While the latter fills unnecessary space on a computer, the former can render an UTAU voice useless.
This guide aims to demonstrate which Japanese sounds are critical to a well-rounded voicebank, as well as which ones are just handy to have to deal with loanwords. It will also go over which aspects of Japanese phonology are outdated or brand new to the language, and how necessary each one is.
This reclist is open to discussion, feedback, and recommendations. If you feel there needs to be additional or updated information, let me know!
Here is a list of all the sounds with the following notation:
- Green: All these sounds are absolutely necessary. If your UTAU doesn't have these, it can't do Japanese. No ifs, ands, or buts.
- Light green: These sounds are in the Japanese language... but can be made by combining other syllables. Still, it's recommended to have these in the voicebank.
- Yellow: These sounds do not exist in standard Japanese, but are used frequently in really common loanwords, and are thus useful to have, but not necessary.
- Orange: These sounds used to exist in Japanese, or are very rarely used in loanwords, but will likely not be used at all. Include these only if you have specific intentions for them.
- Red: Please do not do these. They are either redundant, experimental, or just plain wrong. If you see these in a reclist that claims to be Japanese-only, run.
- Blue: There may be a reason to include these. They'll be explained.
Onto the syllables!
These are listed by their hiragana version followed by romaji.
あ い う え お ん --- a i u e o n
か き く け こ --- ka ki ku ke ko
さ し す せ そ --- sa shi su se so
しゃ しゅ しょ --- sha shu sho
た ち つ て と --- ta chi tsu te to
ちゃ ちゅ ちょ --- cha chu cho
な に ぬ ね の --- na ni nu ne no
は ひ ふ へ ほ --- ha hi fu he ho
ま み む め も --- ma mi mu me mo
や ゆ よ --- ya yu yo
ら り る れ ろ --- ra ri ru re ro
わ を --- wa wo *
が ぎ ぐ げ ご --- ga gi gu ge go
ざ じ ず ぜ ぞ --- za ji zu ze zo
じゃ じゅ じょ --- ja ju jo
だ で ど --- da de do
ば び ぶ べ ぼ --- ba bi bu be bo
ぱ ぴ ぷ ぺ ぽ --- pa pi pu pe po
* "wo" is considered an outdated sound in modern Japanese, instead being pronounced "o"... but singers use it all the time. You can either choose to use "を" to represent "wo" (which is the standard way of doing it) or with "うぉ" (which is more phonetically accurate) and let "を" be an alias for "お/o". Both sides have their points, and both are equally valid.
きゃ きゅ きょ --- kya kyu kyo
にゃ にゅ にょ --- nya nyu nyo
ひゃ ひゅ ひょ --- hya hyu hyo
みゃ みゅ みょ --- mya myu myo
りゃ りゅ りょ --- rya ryu ryo
ぎゃ ぎゅ ぎょ --- gya gyu gyo
びゃ びゅ びょ --- bya byu byo
ぴゃ ぴゅ ぴょ --- pya pyu pyo
すぃ --- si
しぇ --- she
てぃ とぅ --- ti tu
ちぇ --- che
つぁ つぃ つぇ つぉ --- tsa tsi tse tso
ほぅ --- hu
ふぁ ふぃ ふぇ ふぉ --- fa fi fe fo
ふゅ --- fyu
いぇ --- ye
うぃ うぇ --- wi we
ずぃ --- zi
じぇ --- je
でぃ どぅ --- di du
てゅ --- tyu
でゅ --- dyu
ヴぁ ヴぃ ヴ ヴぇ ヴぉ --- va vi vu ve vo
クァ クィ クェ クォ --- kwa kwi kwe kwo *
グァ グィ グェ グォ --- gwa gwi gwe gwo *
カ゜ キ゜ ク゜ ケ゜ コ゜ --- nga ngi ngu nge ngo **
* These were historically used in Japanese, and still appear in some Japonic languages. They are (extremely rarely) utilized in modern Japanese as an antiquity.
** "ng" represents the phoneme /N/ which is different from /n/. Think of the difference of the n sounds in "sin" versus "sing"; /N/ is the latter.
"ng" is sometimes used at the end of a Japanese word ending in "n", for example, "nihon" sounds like "nihong" on occasion.
Even though it's usually used at the end of a sentence, there are kana representing ng followed by vowels.
Any consonant followed by a w; for example, mwa, fwo, zwe, swi.
The only exceptions are the possible "kwa kwi kwe kwo" and "gwa gwi gwe gwo".
Any consonant followed by a "ye"; kye, gye, bye, pye, etc. are all completely unused.
Any consonant followed by a "yi"; kyi, gyi, byi, pyi, etc. are all redundant.
sya, syu, syo
shya, shyu, shyo
chya, chye, chyo
tsya, tsyu, tsyo
jya, jyu, jyo
tya, tyi, tye, tyo
dya, dyi, dye, dyo
zya, zyu, zyo
づぁ づぃ づ づぇ づぉ --- dza dzi dzu dze dzo *
ぢ --- dji **
ジャ ジ ジュ ジェ ジョ --- zha zhi zhu zhe zho (written in katakana as opposed to hiragana) ***
la, li, lu, le, lo ****
ra, ri, ru, re, ro ****
tha, thi, thu, the, tho ****
* In Japanese, the kana づ represents "dzu", but in most dialects this sound has been blended with the "zu" sound. However, it can still be pronounced the old way in order to have a little more accent variety. This is optional and only for flair.
** This kana represents "dji", and is similar to the "dzu" kanji because it is usually pronounced exactly like "ji" in most dialects. It can, too, be recorded for a little variety.
*** In Japanese, the "j" sound either sounds like the English j (as in joke) or a "zh" sound (as in pleasure). If you seek to have both variations in your bank, you can represent the second kind in either katakana (as seen above), with a 2 (so, じ2), or some other mark.
**** These English consonants are the only major ones missing from Japanese, and many contemporary Japanese songs include English in their lyrics. Including these strings in a reclist can increase its Engrish capabilities significantly, but are definitely not needed.
Once again, if there's something you feel needs to be elaborated on further or be included/excluded from this resource, please let your voice be heard. Enjoy recording your Japanese!