You may have not heard of it yet but NEUTRINO is a voice synthesizer that uses a neural network as its primary framework, created by SHACHI.
Features like timing, pitch, voice quality, and blurring of voice from music score are all created using this neural network. Speech is synthesized by a vocoder based on the aforementioned estimated paramaters.
AI and Moto Neural Network (NN) based singing voice synthesis uses music information to find out what timing/pitch/voice quality/voices will be and then vocoders it.
And the best of all? NEUTRINO is free to use thanks to it being freeware!
NEUTRINO’s Key Components
This software identifies itself as “Natural and Smooth” where its features are directly extracted from the data obtained of the singers and their songs. It is able to reproduce not only the voice quality of the singer, but also all of their characteristics like tonality and singing habits.
It is also very easy to “Install and Use“. Just open the lyrics and melodies that you like and the software will do the rest. It’s almost as if you were producing songs with a real actual singer! With just one click, you’ll be able to set up the software for all your musical needs.
Start quick and enjoy a world you’ve never seen before with this revolutionary vocal synthetizer!
NEUTRINO’s Singer Library
As of February 2020, there are two singers available in NEUTRINO. These are called Tohoku Kiritan and Yoko which you can check out below:
Here are some test samples made by 96s_kM4osM from km4osm.com/!
A tutorial on how to install and use NEUTRINO is available on the official website here. (Japanese)