MVA Material for the course on Audio Signal Processing

Slides of the course

• Part I : Introduction (download)

• Part II : Analogic signal/Digital signal (download)

• Part III : Filtering (download) 

• Part IV : Stochastic signal processing (download) 

• Part V : Speech processing (download)

• Part VI : Denoising (download)

• Part VII : Time-frequency analysis (download)

Registration to the course :

• Send me an email at emmanuel.bacry@cnrs.fr before January 31st to notify your registration

For the validation of the course :

• A report on a paper (to choose among the list below or choose your own and send me an email for validation)
• Send me an email at least at least 2 weeks before the oral exam indicating the paper you chose
• Send me the report before 8pm the day before the oral exam
• This is an individual work
• The report should be structured in the following way :
• A part corresponding to a summary of the paper
• A part (more personal) corresponding to your critics (positive/negative), your numerical experiments and/or extensions
• A bibliography (you are encouraged to talk about other papers than the reference paper in the second part of your report)


• A 15min oral exam
• 10min : Presentation of your report
• 5min : I will ask you some technical questions about the course (in relation to the subject of your report)
• WARNING : simply reading the slides of the course is NOT sufficient to understand the course

Possible Papers for the report :

• Audio Denoising by Time-Frequency Block Thresholding (2008) (download)
• Elimination of the Musical Noise Phenomenon with the Ephraim and Malah Noise Suppressor (1994) (download + the reference paper : download)
• A modeling and algorithmic framework for (non)social (co)sparse audio restoration (2017) (arXiv:1711.11259)

• Improved Phase Vocoder Time-Scale Modification of Audio (1999) (download)
  • Time contraction/dilation
  • Using an adaptive window size (2015) (1 more paper : download)
• A comparison of recent neural vocoders (2019) (study in detail one of them in another paper to be found) (download)
• Additive Model (1 chapter of a thesis : download + bibliography)
  • Transformation of sounds (time contraction/dilation, transposition)
  • Improvements of the model (noise, transitories,...)
• Percussive sound synthesis (download)
• Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders (2017) (download)

• Multipitch estimation of piano sounds using a new probabilistic spectral smoothness principle (2010) : download
• A pitch salience function derived from mharmonic frequency deviations for polyphonic music analysis (2014) : download
• Combining Spectral and Temporal Representations for Multipitch Estimation of Polyphonic Music (2015) : download
• A Discriminative Model for Polyphonic Piano Transcription (2007) : download
• Chord detection using deep learning (2015) (download)
• Pitch recognition using NMF (2021) (arXiv:2107.11250)

• Source separation with Gaussian source model (2005) (download)
• Multichannel Nonnegative Matrix Factorization in Convolutive Mixtures for Audio Source Separation (2010) (download)
• A Source/Filter Model with Adaptive Constraints for NMF-based Speech Separation (2016) (download)
• Unsupervised Source Separation via Self-Supervised Training (2022) (arXiv:2202.03875)

• Multi-Feature Beat Tracking (2014) (download)
• Sigma/Delta (2001) (download)
• MPEG2 AAC (focus on psycho-acoustic modelization ) (an example of paper)
• Summarization of music (2016) ( download)
• The Application of Hidden Markov Models in Speech Recognition (2008) (download)