This page publishes a past answer by Rei Yumesaki to an inquiry about the Tsukuyomi-chan Corpus as reference information.
このページでは、「つくよみちゃんコーパス」について過去にいただいたお問い合わせに対する夢前黎からの回答を参考資料として公開しています。
The inquiry concerned the use of synthesized voice based on the Tsukuyomi-chan Corpus's voice quality as the output audio for self-made conversational AI software in which a character other than Tsukuyomi-chan speaks.
「つくよみちゃんコーパスの声質を使用した合成音声を、つくよみちゃん以外のキャラクターが喋る自作の会話AIソフトウェアの出力音声として使いたい」というご相談でした。
Past Answer by Rei Yumesaki
Local testing falls under Category 1.
The demo video falls under Category 2.
Examples of the required credit attribution can be found at the link above.
Does your plan fall under Category 3?
This includes any act that enables third parties to use the voice quality of the Tsukuyomi-chan Corpus.
If your use falls under Category 3, the following points for confirmation are also important.
Will the avatar using Tsukuyomi-chan's voice be Tsukuyomi-chan herself, or a different character?
In Category 3, if Tsukuyomi-chan's voice quality is used for another character's voice synthesis program, you must consult me and obtain my permission.
Regarding your case, I request you to make it clear to everyone that another character is speaking using Tsukuyomi-chan's voice quality.
"Everyone" includes software users who do not read the manual and viewers who watch the screen capture video of the application posted by users.
One possible solution is to always display the credits on the screen.
For example, as shown in the image below:
Voice Synthesis Dataset: Tsukuyomi-chan Corpus (つくよみちゃんコーパス) ※Please change "Voice Synthesis Dataset" to the appropriate term for your usage.
https://tyc.rei-yumesaki.net/material/corpus/
Licensed by Rei Yumesaki
This is also effective for displaying the copyright information for the avatar character and preventing various problems.
In addition to the standard credit display (such as the credit display template example provided in the Category 3 terms) in the manual or on the distribution page, please also display the credits on-screen as shown in the image above.
If this is not possible, please consult with me again.
I aim to develop Tsukuyomi-chan into a Free Material Character with global recognition.
I am concerned that if her voice becomes widely associated with other characters before that happens, the original Tsukuyomi-chan might later be wrongly perceived as lacking uniqueness.
I would appreciate your understanding regarding the necessity of proper credit attribution.
For your information, I welcome you to use Tsukuyomi-chan's 3D model to make her speak.
For example, the 3D model below, which I created, is explicitly permitted for commercial use and integration into software.
Tsukuyomi-chan Official 3D Model Type A
https://tyc.rei-yumesaki.net/material/avatar/3d-a/
In Category 3, please note that there are prohibitions that must be carried over and credit attributions that are required.
In particular, please make sure that the following prohibitions are reflected in the terms of use for synthesized voice using the Tsukuyomi-chan Corpus.
(If you do not do so, you must obtain separate permission from me.)
- Criticizing and attacking. The definitions of "criticizing and attacking" shall follow the Tsukuyomi-chan Character License ( https://tyc.rei-yumesaki.net/about/terms/#condition3 ).
- Calling for support for or opposition to a particular political position, religion, or ideology.
- Publishing strongly stimulating or sensitive content without appropriate zoning.
- Publishing generated audio in a form that permits others to reuse it as material.
These are not the complete Category 3 terms.
Your software or service must comply with all applicable conditions in the Category 3 terms, including requirements not listed above.
I appreciate your cooperation.