In Black Mirrors S05E03, a comatose artist is still able to record and perform music when her producers take control over her mind. They use a device to extract the music that she is “composing” in her mind (note by note) and extract it into playable music. They obtain lyrics and melodies that they (in true producer fashion) - alter to sound more positive and appealing to their audience.
Though the context in the episode is one of taking away the conscious creative control from the human, I think it has some really interesting concepts of auralizing, that could be used to give a musical voice to the differently abled. There is also a key commentary of the artist’s existing voice recordings being used to recreate a synthetic version of her voice. This can be extended to creating a database of voices for everyone - voice actors, celebrities, family members, deceased or otherwise. Apart from the exciting possibility of listening to your favorite songs in the voice of your favorite artists (or your own), it also suggests a scary question of public domain data and its direct connection to identity and an individual’s rights over it.
Key pieces of SciFi tech:
1. Extracting music from brain (they show a note level extraction of music from comatose artist’s brain/dreams)
- Trainor, Laurel J., Kelly L. McDonald, and Claude Alain. "Automatic and controlled processing of melodic contour and interval information measured by electrical brain activity." Journal of Cognitive neuroscience 14.3 (2002): 430-442.
2. Modifying amount of positivity of content (Changing lyrics and pitch contours to an affective slider)
3. Database of voice for singing voice synthesis
-references – services offered by companies like Voctrolabs, LyreBird inc.
-creating similar databases of unique voices and owning the rights to your voice.
-listening to your favorite song in your favorite artists voice.
4. Holographic concerts by singing voice (Inspired from reality – Hatsune Miku, etc.)
It would be really exciting to be able to build the voice database for each unique person, starting with people whose voices are in public domain. Apart from releasing music in other customizable voices, one could even automate the process of voice over recordings for film and other media. Home assistants could have the voice of someone you love, or your favorite actor. What would this database of voices look like and how would it be regulated? The owner of the voice will be the natural owner of the copyright to their voice.