During the studies for my Masters degree, I was deeply engrossed in the confluence of my undergraduate studies, classical music performance, and the work I was currently undertaking in Computer Science topics, in particular with functional programming with a significant emphasis on Haskell. It turned out that Haskell is particularly well-suited for the task of expressing musical constructs, primarily due to its type system programming and features, which allowed an almost one-to-one translation between Western classical notation and runnable Haskell code.
One library that is commonly used in Haskell for producing music, and the foundation for work in my thesis is Euterpea 1. Euterpea is a domain-specific language (DSL) for music composition that provides a rich set of abstractions for music composition, including notes, chords, rhythms, and melodies. It also provides a range of functions for manipulating and transforming musical structures, such as transposing, inverting, and reversing notes and chords.
Euterpea is a powerful tool for producing music, but it can also be challenging to learn, especially for those who are new to Haskell or functional programming.
An interesting and interactive way to get started with Euterpea (and functional programming with Haskell) is the Haskell School of Music (HSOM) 2. The HSOM textbook provides a comprehensive introduction to Haskell and Euterpea. It covers a wide range of topics, including basic Haskell syntax, music theory, and the Euterpea library itself. The textbook includes a range of exercises and examples for which for some of them I’ve published the answers.
Enamoured with this language, my Masters thesis 3 sort to make use of Euterpea and the expressive syntax of Haskell to experiment with various AI techniques. Rhythmgn 4 is a probabilistic rhythm generator that produces outputs for a given rhythmic ensemble, such as a set of drums or percussive instruments. The program takes input in the form of Euterpea code and processes it to produce a continuous and varying output. The output can then be used to drive a synthesiser or drum machine.
Rhythmgn uses a number of techniques from probability theory to generate its outputs. For example, it may use Markov chains to model the transitions between different rhythmic patterns, or it may use stochastic processes to add variation and randomness to the rhythms. These techniques can produce highly complex and interesting rhythms that would be difficult or unlikely to produce by hand.
In order to use Rhythmgn, you will need to have some basic knowledge of Haskell and Euterpea. You will also need to have some knowledge of music theory and rhythm, as the program relies heavily on these concepts. However, with some practise and experimentation, you should be able to produce some interesting and unique rhythms using Rhythmgn.
Overall, Haskell and Euterpea provide a powerful set of tools for producing music, and it is well worth spending a weekend or two learning the ropes and experimenting with the language to produce fun and creative outputs.