Section: New Results

Axis 4: Real-time Audio Sources Classification

Participants : Christophe Biernacki, Maxime Baelde.

This work addresses the recurring challenge of real-time monophonic and polyphonic audio source classification. The whole power spectrum is directly involved in the proposed process, avoiding complex and hazardous traditional feature extraction. It is also a natural candidate for polyphonic events thanks to its additive property in such cases. The classification task is performed through a nonparametric kernel-based generative modeling of the power spectrum. Advantage of this model is twofold: it is almost hypothesis free and it allows to straightforwardly obtain the maximum a posteriori classification rule of online signals. Moreover it makes use of the monophonic dataset to build the polyphonic one. Then, to reach the real-time target, the complexity of the method can be tuned by using a standard hierarchical clustering preprocessing of sound models, revealing a particularly efficient computation time and classification accuracy trade-off. The proposed method reveals encouraging results both in monophonic and polyphonic classification tasks on benchmark and owned datasets, even in real-time situations. This method also has several advantages compared to the state-of-the-art methods include a reduced training time, no hyperparameters tuning, the ability to control the computation - accuracy trade-off and no training on already mixed sounds for polyphonic classification. This work is now published in an international journal [16] and Maxime Baelde defended his PhD thesis on this topic this year [11].

It is a joint work with Raphaël Greff, from the A-Volute company.