Evaluation of Tube Screamer Simulation using Machine Learning

Autor: Miquel Royo
Přispěvatelé: Frederic Font, Xavier Lizarraga
Jazyk: angličtina
Rok vydání: 2022
Předmět:
DOI: 10.5281/zenodo.7116346
Popis: In this paper a comparison between the original analog circuit that represents theground truth and the simulation of nonlinear analog circuits using neural networks isperformed. Traditionally the white box approach has provided some good results interms of accuracy but implies an important computational demand to emulate digitalaudio effects (DAFx). Newer approaches using neural networks provide a black boxapproach that can be more efficient, accessible and potentially obtain similar or evenbetter results. In this work I build the original electronic clone analog circuit as wellas the equivalent digital circuit using the state of the art technologies using neuralnetworks and perform an evaluation methodology to obtain some insights.
Databáze: OpenAIRE