Singing Voice Synthesis

• History or general overview: Wikipedia
• Video explainer on how it works: junferno
• If you want to create one without coding anything, my recommendation is OpenUtau.
• Academic review of the state of the art: https://arxiv.org/html/2601.13910

terminology

• SVS - singing voice synthesizer - software or hardware that generates a sound that mimics a singing human voice (for example, Vocaloid by Yamaha)
• articulation
• formant
• source-filter
• FFT
• fundamental frequency
• utterance
• interpolation - gradually shifting from one state to another (as opposed to suddenly snapping into the new state from the old one)

All pages