Guitar Effects Recognition and Parameter Estimation With Convolutional Neural Networks

Comunità M, Stowell D, Reiss JD. Guitar Effects Recognition and Parameter Estimation With Convolutional Neural Networks. J Audio Eng Soc. 2021;69(7/8):594-604. doi:10.17743/jaes.2021.0019

@article{Comunita2021_21124,
  author = {Comunità, Marco and Stowell, Dan and Reiss, Joshua D.},
  title = {{Guitar Effects Recognition and Parameter Estimation With Convolutional Neural Networks}},
  journal = {Journal of the Audio Engineering Society},
  volume = {69},
  number = {7/8},
  pages = {594--604},
  year = {2021},
  month = jul,
  publisher = {Audio Engineering Society},
  doi = {10.17743/jaes.2021.0019},
  url = {https://doi.org/10.17743/jaes.2021.0019}
}

TY  - JOUR
TI  - Guitar Effects Recognition and Parameter Estimation With Convolutional Neural Networks
AU  - Comunità, Marco
AU  - Stowell, Dan
AU  - Reiss, Joshua D.
T2  - Journal of the Audio Engineering Society
J2  - J. Audio Eng. Soc.
VL  - 69
IS  - 7/8
SP  - 594
EP  - 604
PY  - 2021
DA  - 2021/07/06
DO  - 10.17743/jaes.2021.0019
UR  - https://doi.org/10.17743/jaes.2021.0019
PB  - Audio Engineering Society
LA  - en
AB  - Despite the popularity of guitar effects there is very little existing research on classification and parameter estimation of specific plugins or effect units from guitar recordings. In this paper, convolutional neural networks were used for classification and parameter estimation for 13 overdrive, distortion, and fuzz guitar effects. A novel dataset of processed electric guitar samples was assembled, with four sub-datasets consisting of monophonic or polyphonic samples and discrete or continuous settings values, for a total of about 250 hours of processed samples. Results were compared for networks trained and tested on the same or a different subdataset. We found that discrete datasets could lead to equally high performance as continuous ones while being easier to design, analyze, and modify. Classification accuracy was above 80%, with confusion matrices reflecting similarities in the effects timbre and circuits design. With parameter values between 0.0 and 1.0, the mean absolute error is in most cases below 0.05, while the root mean square error is below 0.1 in all cases but one.
ER  -