We propose a machine learning approach based on hinge-loss Markov random fields to solve the problem of applying reverb automatically to a multitrack session. With the objective of obtaining perceptually meaningful results, a set of Probabilistic Soft Logic (PSL) rules has been defined based on best practices recommended by experts. These rules have been weighted according to the level of confidence associated with the mentioned practices based on existent evidence. The resulting model has been used to extract parameters for a series of reverb units applied over the different tracks to obtain a reverberated mix of the session.

Audio samples for generated mixes as well as examples for the PSL templates can be found under the 'Downloads' tab.

Link to ongoing listening test (created with the Web Audio Evaluation Tool): https://goo.gl/mVQZCp

Related publications

A. L. Benito and J. D. Reiss, “Intelligent Multitrack Reverberation Based on Hinge-Loss Markov Random Fields,” in 2017 AES International Conference on Semantic Audio, 2017.
[More Details] [BIBTEX] [URL (ext.)]


Manager: Adan Benito