Concord

Description

Please note that Concord is no longer under active development and will not be updated to support UE versions past 5.1. I apologize for any inconvenience this may cause.

Concord lets you create and play intricate musical dice games! Its powerful probabilistic modelling features let you create generative music systems with full control over random behavior. You can also train models on pattern datasets with Concord’s machine learning capabilities. With Concord you can import your musical ideas from MIDI files and create probabilistic models to generate endless variations thereof.

At its core, Concord is about creating, editing and sampling from factor graphs over discrete random variables. The resulting distributions can be affected by dynamic parameters and observations. After running a sampler to find a likely configuration for the values of the discrete random variables, musical pattern data can be computed from these values and fed to a sound generator such as MetaSounds.

Factor graphs defining a distribution over patterns can have a large number of random variables, parameters, factors and outputs. A Concord model is an abstraction of a factor graph that enables designers to reason about multidimensional arrays of these respective elements and allows for composition and instancing of models. An essential part of the Concord toolset is the graph editor used to specify Concord models. Nodes are connected in a directed, acyclic graph to build up the model in a similar way to other graph editors in Unreal Engine like Blueprint for visual scripting or MetaSounds for the creation of sound sources and effects.

Check out the Concord tutorial playlist on YouTube for an introduction to the functionalities: https://youtube.com/playlist?list=PLKchFlXMNqi5Dwlnq4IugVnf0bt25QG_c