Character Driven Text-To-Speech Software from Z Vocal Project
As part of our research into vocal synthesis, we have released an open Beta for a text-to-speech program based on Piper. This program allows for basic speech synthesis in a variety of languages including English, Spanish, French, and Chinese. We are also working on getting Japanese working.
The program was made in only about a week, and there are a few more things we are going to try adding to it in the next couple months. The biggest thing being supporting a mouth open image for each voicebank’s banner. Patch notes can be read here.
Z Idol Talker’s output is meant to be used for funny internet videos and should not be used in large productions in place of a real voice actor. The program must not be used with voices trained on samples you don’t have the rights to, and the use of the program should be noted whenever possible.
You can not only use your own Piper models in the program but customize the UI to fit whatever theme you desire.
An example voicebank with the default UI assets is available on the download page. A series of text and video tutorials will be released as soon as possible so please stay tuned!
Download Z Idol Talker now! It’s completely free for non-commercial use. Some Z Vocal Project voicebanks allow for very limited commercial use, so be sure to check each license.
Main Application + ZVP Voicebanks
These voicebanks were trained by Z Vocal Project on publicly available corpuses. Please check license files for more information. Password is “ZIdolTalker”. Download Here.
As of January 2024 we have decided to take down voicebanks not trained by Z Vocal Project to give us more freedom when it comes to the engine, and to ensure all voicebanks we distribute were sourced ethically.
This means that the external piper models we formatted for use in Z Idol Talker have been taken down. If you downloaded any of these voicebanks, we kindly ask you to delete them instead of trying to preserve them.