English | 中文

Github | Gitee | Gitlab | Bitbucket | Coding | Jihulab

Hitsuboku Kumi (筆墨クミ) is a UTAU virtual singer developed by Cubialpha. This project ports Hitsuboku Kumi Chinese CVVC voicebank to deepvocal. This is the first open-source deepvocal voicebank on Github.

According to Hitsuboku Kumi's Terms of use, redistribution of voicebanks, editing of oto.ini and audio samples, and porting to other synthesis softwares are allowed without permission, so long as the name "Hitsuboku Kumi" is unchanged, "Cubialpha" is credited, a link to this website is included, and it is clearly stated that you have made edits.

If you like this voicebank, please give her a star.

If you find a bug, please create an issue to report it.

(花欺~沉睡在浮空的花海, Original Music:Cube^3, Lyric:雨狸, midi:卦者那啥子靈風)

Github Release

Tencent Weiyun

Aliyun Drive (After downloading, please manually change the filename extension from ".docx" to ".7z", and decompress it.)

• 4 pitches (A3, D4, A4, D5)
• using Voicemith reclist
• provide dvtb1.1.6 and dvtb2.1.0 builds
• tail symbol: R
• vowel pause from Hitsuboku Kumi Japanese Act4: "!a" "a_!" etc
• vocal fry from Hitsuboku Kumi Japanese Act4: "'a" "'a_'" etc
• standalone consonants (only available in the dvtb1 build): p, f, t, k, q, x, c, s, ch, sh, r

See feature list for details

• The volume is Normalized to -9db

Follow this instruction only if you want to modify this voicebank. If you just want to use this, see the Download section.

1. Download this project with the following command:

git clone https://github.com/oxygen-dioxide/hitsuboku-kumi-chn-deepvocal

2. Download and install deepvocal toolbox

3. Open kumi.dvtb with deepvocal toolbox. Click "Function→Build voice bank"

4. Click "Build Voice Model Files" to compile this voicebank.

5. Click "Build Voice Bank" to pack this voicebank. The packed voicebank will be output to .\bin folder.

Deepvocal Official Site

Deepvocal toolbox manual

Deepvocal Mark Tool (DVMT) repo

Hitsuboku Kumi Official Site

Cubialpha's bilibili channel