Audio Texture Synthesis with Scattering Moments

Abstract

We introduce an audio texture synthesis algorithm based on scattering moments. A scattering transform is computed by iteratively decomposing a signal with complex wavelet filter banks and computing their amplitude envelop. Scattering moments provide general representations of stationary processes computed as expected values of scattering coefficients. They are estimated with low variance estimators from single realizations. Audio signals having prescribed scattering moments are synthesized with a gradient descent algorithms. Audio synthesis examples show that scattering representation provide good synthesis of audio textures with much fewer coefficients than the state of the art.

5 pages, 2 figures

References in corpus (1)

• Deep Scattering Spectrum

Cited by in corpus (16)

• Super-Resolution with Deep Convolutional Sufficient Statistics
• Scattering Networks for Hybrid Representation Learning
• Joint Time-Frequency Scattering
• Audio style transfer
• Joint Time-Frequency Scattering for Audio Classification
• Texture Synthesis Using Shallow Convolutional Networks with Random Filters
• Compressing the Input for CNNs with the First-Order Scattering Transform
• Scaling the Scattering Transform: Deep Hybrid Networks
• Phase retrieval of bandlimited functions for the wavelet transform
• Stable Phase Retrieval in Infinite Dimensions
• Fatigue monitoring and maneuver identification for vehicle fleets using a virtual sensing approach
• Geometric Scattering for Graph Data Analysis
• Geometric Scattering on Manifolds
• The Shape of RemiXXXes to Come: Audio Texture Synthesis with Time-frequency Scattering
• wav2shape: Hearing the Shape of a Drum Machine
• Machine olfaction using time scattering of sensor multiresolution graphs