Noise reduction in analog tape audio recordings with deep learning models

Irigaray I, Rocamora M, Biscainho LWP. Noise reduction in analog tape audio recordings with deep learning models. In: AES Conference: AES 2023 International Conference on Audio Archiving, Preservation & Restoration. Audio Engineering Society; 2023. Paper 1. Available from: https://aes.org/publications/elibrary-page/?id=22138

@inproceedings{Irigaray2023_22138,
  author = {Irigaray, Ignacio and Rocamora, Martin and Biscainho, Luiz W.P.},
  title = {{Noise reduction in analog tape audio recordings with deep learning models}},
  booktitle = {AES Conference: AES 2023 International Conference on Audio Archiving, Preservation \& Restoration},
  note = {Paper 1},
  year = {2023},
  month = jun,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=22138}
}

TY  - CPAPER
TI  - Noise reduction in analog tape audio recordings with deep learning models
AU  - Irigaray, Ignacio
AU  - Rocamora, Martin
AU  - Biscainho, Luiz W.P.
T2  - AES Conference: AES 2023 International Conference on Audio Archiving, Preservation & Restoration
M1  - Paper 1
PY  - 2023
DA  - 2023/06/06
UR  - https://aes.org/publications/elibrary-page/?id=22138
PB  - Audio Engineering Society
LA  - en
AB  - This work addresses the problem of noise reduction in tape recordings using a deep-learning approach. First, we build a data set of audio snippets of tape noise extracted from different functional tape equipment — comprising open reel and cassette. Then, we adapt and train an existing deep-learning architecture originally proposed to remove noise from 78 RPM gramophone records. The model learns from mixtures of the noise snippets with clean audio excerpts at different SNRs. Experimental results validate the approach, showing the benefits of using real tape recording noise in training the model. Furthermore, the data set of tape noise snippets and the trained deep-learning models are publicly available. In this way, we encourage the collective improvement of the data set and the broad application of the denoising approach by sound archives.
ER  -