Preserve a Voice While It Can Still Sing
Every memorial song starts with whatever recordings happened to exist. Voice preservation is the other order: recording deliberately, while the person is here and able to sing, so what your family keeps is a clean, complete set rather than a noisy fragment from a wedding video. Some people do it for a grandparent. Some do it for themselves.
My Voice Studio
Train your permitted voice once, then reuse it for covers
Step 1
Train the voice
Add clean recordings of one voice you own or have permission to use.
Step 2
Create the cover
Choose a source performance. Its melody, words, timing, and phrasing stay in place.
Source performance
YouTube · upload · microphone
Why people bank a voice now
Voices change. Age softens them, surgery interrupts them, and conditions that affect speech — ALS among them — can take singing before anything else. Recording early is simply easier than recording late, and the difference in the material is significant.
It is also a nicer thing to do while everyone is well. A grandparent recording the songs they grew up with is an afternoon of singing, not a medical task, and the family ends up with both the recordings and the model.
How to bank a singing voice
The aim is coverage: enough of the voice, cleanly captured, that a model has something to learn from.
What a preserved voice can do later
A trained model is reusable, so the recordings stop being an archive nobody opens and become something the family can actually use: a lullaby for a grandchild who was not born yet, a song at an anniversary, a duet recorded across a generation.
The voice remains the singer’s. Preservation only makes sense on a consented voice, and the person who recorded it should be the one who decides what it is used for and who may use it.
Record on purpose. Train while the recordings are fresh.