MERCURY MUSIC
A new framework for Computer-Assisted Composition
Deep learning, generative models, neural networks, transformes, Markov models
Composing by means of AI techniques.
Improve your creativity with Mercury.
Friendly intuitive user interface.
Mercury offers a simple and intuitive user interface for stablishing the values of the different calculation parameteres. The music material generated is shown on the screen and can be played through MIDI. The music is imported/exported into MusicXML format.
Thematic bridging.
Mercury allows you to generate intermediate music materials -thematic bridging- between two melodies, chord sequences or rhythmical patterns by means of fuzzy c-means clustering or evolutionary computing, generating ongoing transitions and providing user with a huge amount of new music material.
Fuzzy c-means clustering.
The fuzzy c-means clustering allows to compare structures with very different number of elements. The convergence process generates intermediate states that can be used as transitions from the initial material to a final desired one.
Evolutionary composition.
Mercury implements new definitions for music simmilarity that can be used as fitness functions in evolutionary algorithms applied to music.
Markov models.
Mercury offers the posibility to generate new music material using machine learning techniques like Markov Models. The markov models have been trained with corpora of several composers like Bach, Monteverdi, Jazz, etc. The results are short ternary-formed pieces in wich a melody is harmonized in the style of the selected composer.
“Mercury allows me to search and filter thousands of new musical ideas. Whatever my goal is, the software helps me in the compositional proces.”
— Brian Martínez