This is the project website & blog of MIRLCAuto: A Virtual Agent for Music Information Retrieval in Live Coding, a project funded by the EPSRC HDI Network Plus Grant - Art, Music, and Culture theme.
Tutorials
Check out our new video tutorials!
MIRLCa Set up #
[NEW] In this video, you will find out how to set up the SuperCollider extension MIRLCa.sc and related extensions MIRLC2.sc, flucoma (official link coming soon) and the Freesound quark.
MIRLCa Performance Mode #
Similar, Random and Tagged Sounds Using Machine Learning #
Coming soon.
Similar Sounds Using Machine Learning #
In this video, you will discover how to use MIRLCa in performance mode. This video relates to the first version released in December for the workshops, which only has the function "similar" implemented with machine learning and a single path for sound downloads, sound credits list generation and loading the JSON model files. It is still useful to watch to understand the workflow of the program.
MIRLCa Training Mode #
Training Using Sounds by Random and by ID #
[NEW] In this video, you will get exposed to a full training session retrieving sounds by random and by ID, which gives you more control and flexibility in the training session.
Training Using Sounds by Random #
In this video, you will get familiar with the training workflow and how to train a model retrieving random sounds only. This video relates to the first version released in December for the workshops.
Code Repository #
Here you can find a link to our GitHub repository: https://github.com/mirlca