Présentations > Conférence plénière

Conférence plénière par Gaël Richard

Jeudi 06/04 - 11h - Centre Pompidou, petite salle

Audio Signal Processing in the 21st century

Audio signal processing has seen many landmarks in its development as a research topic. Historically, the vast majority of work was driven by the objective to build models that capture the essential characteristics of the analysed audio signal and to represent it with a limited set of parameters or components. The field has now evolved beyond the essential characteristics explored in the past. For instance, a wide variety of speech/audio signal models have since been proposed and in particular around signal decomposition/factorisation models and sparse signal representations. Nevertheless, the general trend of the entire research domain that has recently emerged is a paradigm shift towards data-driven methods based on machine learning, and especially deep learning. In many applications, such data-driven models obtain state-of-the art results if appropriate data is available to train the models. Without aiming for exhaustiveness, this presentation will provide a view of the important outcomes of the field in the last 25 years illustrating also this emergence of pure data-driven models [1]. The presentation will cover in particular the research tackled in novel signal models and representations, in the modeling, analysis and synthesis of acoustic environments and acoustic scenes, in signal enhancement and separation, in Music Information Retrieval (MIR) and Detection and Detection and Classification of Acoustic Scenes and Events (DCASE).

[1] Gaël Richard, Paris Smaragdis, Sharon Gannot, Patrick A. Naylor, Shoji Makino, Walter Kellermann, Akihiko Sugiyama, Audio Signal Processing in the 21st century, submitted to a special issue of the IEEE Signal Processing magazine (2023)