Wikia

UTAU wiki

Setup and Configuration

6,067pages on
this wiki
Talk0

Generating freqmaps

Frequency maps (.frq files) are needed to make an UTAUloid sing. When you render a song, the program checks the presence of the freqmaps, and generates them if it hasn't found them.

To generate them before hand, go to Tools→Voicebank Settings.

Fullscreen capture 10272012 71131 PM.bmp

This window should pop up:

Fullscreen capture 10272012 71624 PM.bmp

To generate a .frq file, right click and select "Initialize Frequency Table"

Fullscreen capture 10272012 71143 PM.bmp

This window will pop up:

Fullscreen capture 10272012 71201 PM.bmp


To generate multiple frqs at a time, click "Select Multi" and select all the voicebank files before clicking "Generate Frequency Table."


Making The OTO.ini

Remember this window?

Fullscreen capture 10272012 71624 PM.bmp

This is also the oto'ing screen.

Name: The file name.

Alias: The name of the file in either hiragana or romaji.

Offset: This part cuts off silence or noise before a sample. Appears as a purple higlight in the oto editing screen.

Consonant: The part of a sample UTAU doesn't stretch. Should contain the consonant and up until the vowel becomes stable. Appears as a pink highlight in the oto editing screen.

Cutoff: Marks the end of the sample. Appears as a purple higlight in the oto editing screen.

Preutterance: Marks the ending of the consonant. Should be put right inbetween the consonant and vowel. Appears as a red line in the oto editing screen.

Overlap: Marks where the sample overlaps into the previous sample. Appears as a green line in the oto editing screen.

Vowels (あ/a, い/i, う/u, え/e, お/o, ん/n) should have a lot of overlap, and little to no preutterance.

Fullscreen capture 10272012 72942 PM.bmp

Hard consonants (b, ch, d, g, j, k, p, t) should have little, or negative, overlap.

Fullscreen capture 10272012 72932 PM.bmp

Soft consonants (f, h, l, m, n, r, s, sh, v, w, y, z) should have some overlap.

Fullscreen capture 10272012 73005 PM.bmp

All samples should be cut off before the vowel tapers off.

Fullscreen capture 10272012 73719 PM.bmp

Octave Shifts

Banks with octave shifts, or multiple pitches, have a much larger range than monotone UTAUs. It is recommended that users who want to make multi-pitch banks use at least three pitches, though banks such as Ritsu Namine and Laru Mine use more. To get UTAU to use these multiple pitches, users use prefixes and suffixes to denote the pitch change. Below is a chart of common symbols used.

Pre/Suffix Meaning
Higher than base pitch.
Lower than base pitch.
Lower than base and ↓ pitch.

High than base and ↑ pitch.

+ Higher than base pitch, commonly used by those who cannot use Japanese symbols.
- Lower than base pitch, commonly used by those who cannot use Japanese symbols.
A3, C4, D4 etc... Denotes the sample's pitch, commonly used in banks with four or more pitches.


To get UTAU to use these pre/suffixes automatically, you must set a prefix map.

Go to Tools→Voicebank Settings and click "Edit Prefix Map"

Fullscreen capture 10272012 74718 PM.bmp

This window should come up:

Fullscreen capture 10272012 74737 PM.bmp

You select the pitch or pitches, then enter either the prefix or suffix used. Then click set.

Voice Information

To show information on your UTAU, first create a "character.txt" file with the following:

name=(UTAU name, voicebank name if applicable)
image=(The name of a 100x100 .bmp or .jpg file, such as "character.jpg"
author=(Your name)

Then, create a "readme.txt" with other information on your UTAU. Save both these files in the voicebank folder, and when you open UTAU and click on the voicebank pitch, something like this should come up:

Fullscreen capture 10272012 75215 PM.bmp

Around Wikia's network

Random Wiki