Diffusion-Based Denoising of Historical Recordings

Miranda BV, Deslandes RA, Irigaray I, Biscainho LWP. Diffusion-Based Denoising of Historical Recordings. J Audio Eng Soc. 2025;73(4):220-230. doi:10.17743/jaes.2022.0198

@article{Miranda2025_22814,
  author = {Miranda, Bernardo V. and Deslandes, Rafael A. and Irigaray, Ignacio and Biscainho, Luiz W. P.},
  title = {{Diffusion-Based Denoising of Historical Recordings}},
  journal = {Journal of the Audio Engineering Society},
  volume = {73},
  number = {4},
  pages = {220--230},
  year = {2025},
  month = apr,
  publisher = {Audio Engineering Society},
  doi = {10.17743/jaes.2022.0198},
  url = {https://doi.org/10.17743/jaes.2022.0198}
}

TY  - JOUR
TI  - Diffusion-Based Denoising of Historical Recordings
AU  - Miranda, Bernardo V.
AU  - Deslandes, Rafael A.
AU  - Irigaray, Ignacio
AU  - Biscainho, Luiz W. P.
T2  - Journal of the Audio Engineering Society
J2  - J. Audio Eng. Soc.
VL  - 73
IS  - 4
SP  - 220
EP  - 230
PY  - 2025
DA  - 2025/04/07
DO  - 10.17743/jaes.2022.0198
UR  - https://doi.org/10.17743/jaes.2022.0198
PB  - Audio Engineering Society
LA  - en
AB  - In the context of audio restoration, the need to remove background noise from historical music recordings is a recurring problem, for which traditional signal processing and supervised deep learning methods have been previously applied. In this work, a generative approach that adapts conditional diffusion sampling for removing perceptually distributed noise is investigated, using the particular case of background noise removal from solo classical piano recordings as a proof of concept. The proposed method uses a set of noise examples to simulate perceptually distributed noise with specific characteristics throughout conditional diffusion sampling. Experiments with real historical 78 RPM recordings and clean recordings with added 78 RPM noise and tape hiss demonstrate that diffusion-based audio denoising performs comparably to state-of-the-art deep learning methods.
ER  -