Transformée en scattering sur la spirale temps-chroma-octave
This addresses sound analysis problems for audio processing researchers, but it appears incremental as it builds on existing scattering and neural network concepts.
The paper introduces a scattering representation for sound analysis and classification, which is translation-invariant, stable to deformations, and captures harmonic structures, applied to study deformations in the source-filter model.
We introduce a scattering representation for the analysis and classification of sounds. It is locally translation-invariant, stable to deformations in time and frequency, and has the ability to capture harmonic structures. The scattering representation can be interpreted as a convolutional neural network which cascades a wavelet transform in time and along a harmonic spiral. We study its application for the analysis of the deformations of the source-filter model.