ALYSIA: Automated LYrical SongwrIting Application
Welcome to ALYSIA's home page. ALYSIA is a fully data-driven system that utilizes machine learning techniques to make original songwriting accessible to everyone.
Songs created with ALYSIA
The following is a song made with ALYSIA using Dickinson's poem, "Hope" is the Thing with Feather. Arranged by Joshua Palkki, produced by Ronen and Maya Ackerman, and vocals by Maya Ackerman. Read more about the creation of this song.
Here is another song utilizing a poem by Emily Dickinson, I Took My Power In My Hand. Vocals and production by Margareta Ackerman. Read more about the creation of this song.
Relies on a variation of our system, ROBOCCINI, trained on Puccini music. James Morgan used ROBOCCINI to create the melody. Vocals my Margareta Ackerman and production by Nicolas Mauthes. More about this aria.
Why Do I Still Miss You is the very first song ever made with ALYSIA. The lyrics, production, and vocals are by Margareta Ackerman. Read more about pop songs with ALYSIA.
Here is another pop song made with ALYSIA. The lyrics, production, and vocals are by Margareta Ackerman.
We experiment with using ALYSIA for genre translation. Starting with Joseph McCarthy's lyrics from the Vaudeville song I'm Always Chasing Rainbows, we used ALYSIA to create
a pop song. Read more about the creation of this song.
How does it work?
ALYSIA works by taking any lyrics, and providing arbitrarily many melodic accompaniments to which the lyrics can be sang. The user can then choose among these melodies, and combine them to create a complete song. This requires no musical expertise and makes songwriting easy.
We've designed a fully data-driven system that produces pitches and rhyme based for any given text. The system utilizes random forests which are trained on a corpus from a uniform musical style, using features of both the music and lyrics.
This project is part of a long-term research plan to create the first automated singer/composer, capable of independent original composition and songwriting. Plans also include the challenge of enriching the system with a voice, enabling it to render professional-quality, emotional and varied performances of its compositions.