The University of Nantes is seeking a postdoctoral research associate in machine learning for audio signal processing as part of the MoDiBE "Model based Differentiable Bandwidth Extension" research project. This project will develop differentiable sinusoidal models of musical sounds for innovative applications in audio restoration. Specifically, the research associate will address the bandwidth extension problem: i.e., to recover a broadband version of an audio signal from a narrowband input.

Qualified applicants must have a PhD in Audio and Music Technology, Machine Learning, or a related area. They also should have a solid experience in Python programming for audio signal processing and adhere to best practices in open-source software development. Your role will involve:

• an extensive bibliographical work on bandwidth extension and related tasks
• the implementation a bandwidth extension system for monophonic audio via differentiable harmonic sinusoidal models
• the extension of the proposed model to polyphonic and inharmonic musical signals
• the application to a related real-world problem: namely, the recovery of near-field audio from a far-field recording
• the communication of scientific results in conferences (IEEE ICASSP, WASPAA, ISMIR) and journals (IEEE TASLP, EURASIP JASMP) in the field.

Apply by: December 15th, 2021

Starting date: Feb 1th, 2022 or later

Duration: 15 months

Location: LS2N, Centrale Nantes, France

Advisors: Mathieu Lagrange (https://mathieulagrange.github.io) and Vincent Lostanlen (https://www.lostanlen.com) 

Salary: 2177 € Net

The  postdoctoral research associate will join the SIMS "Signal, Image and Sound" team of the LS2N "Digital Science Lab of Nantes". For more details and to apply, please contact mathieu dot lagrange at ls2n dot fr. To apply, please send a cv and a cover letter.