L. Gabrielli, C. E. Cella, F. Vesperini, D. Droghini, E. Principi, and S. Squartini, “Deep Learning for Timbre Modification and Transfer: An Evaluation Study,” in Proc. AES Convention 144, May 2018, Paper 9996. [Online]. Available: https://aes.org/publications/elibrary-page/?id=19513
Gabrielli L, Cella CE, Vesperini F, Droghini D, Principi E, Squartini S. Deep Learning for Timbre Modification and Transfer: An Evaluation Study. In: AES Convention 144. Audio Engineering Society; 2018. Paper 9996. Available from: https://aes.org/publications/elibrary-page/?id=19513
@inproceedings{Gabrielli2018_19513,
author = {Gabrielli, Leonardo and Cella, Carmine Emanuel and Vesperini, Fabio and Droghini, Diego and Principi, Emanuele and Squartini, Stefano},
title = {{Deep Learning for Timbre Modification and Transfer: An Evaluation Study}},
booktitle = {AES Convention 144},
note = {Paper 9996},
year = {2018},
month = may,
publisher = {Audio Engineering Society},
url = {https://aes.org/publications/elibrary-page/?id=19513}
}
TY - CPAPER
TI - Deep Learning for Timbre Modification and Transfer: An Evaluation Study
AU - Gabrielli, Leonardo
AU - Cella, Carmine Emanuel
AU - Vesperini, Fabio
AU - Droghini, Diego
AU - Principi, Emanuele
AU - Squartini, Stefano
T2 - AES Convention 144
M1 - Paper 9996
PY - 2018
DA - 2018/05/06
UR - https://aes.org/publications/elibrary-page/?id=19513
PB - Audio Engineering Society
LA - en
AB - In the past years, several hybridization techniques have been proposed to synthesize novel audio content owing its properties from two audio sources. These algorithms, however, usually provide no feature learning, leaving the user, often intentionally, exploring parameters by trial-and-error. The introduction of machine learning algorithms in the music processing field calls for an investigation to seek for possible exploitation of their properties such as the ability to learn semantically meaningful features. In this first work we adopt a Neural Network Autoencoder architecture, and we enhance it to exploit temporal dependencies. In our experiments the architecture was able to modify the original timbre, resembling what it learned during the training phase, while preserving the pitch envelope from the input.
ER -
