Remember when instrument sampling started? Technicians recorded every note from a specific guitar or piano, flute or trombone. Then they would build a virtual instrument from all those recordings of the notes. This could be used with MIDI for improved, life-like, very realistic results. Samples sound like a real instrument, not like a synthesizer or computer game.
The same is happening with the human voice. Many companies are now giving us the same ability with voice. It’s not at all like listening to a robot because it is made from many recordings of a real person singing. You can even do this with your own voice but the samples you sing must be perfectly clear.
There is a product called ACE Studio which got permission from good singers to use their voices in this way. You type or paste lyrics under the MIDI notes in your melody track. The pronunciation is very good but might require minor editing.
For best results, save the vocal and bring it into a DAW. The person singing will sound as good as the accompanying instruments. There are 140 male and female singers. You can also do duets, harmonies, and songs in 8 languages.
This technology is new since 2023 but is amazing already. It is rapidly improving with each new update. You feel like you’re working offline because you use a downloaded application. There are many additional features I have not described. ACE Studio has a 14-day free trial.
(Another product called Suno used celebrity singers without permission and they are in legal trouble about it.)
That’s a really interesting parallel to instrument sampling. It makes perfect sense that building a model from individual vocal performances would create something so much more nuanced than just synthesized tones.