Patent · US Active

Perceptually-based loss functions for audio encoding and decoding based on machine learning

US11817111B2 · kind B2 · utility

2Cited by
7References
19Claims
0Family size

Assignee

Inventors

Key dates

Filing dateApr 10, 2019
Grant dateNov 14, 2023
Priority date
Expiry dateDec 2, 2040

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG10L19/00
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

Computer-implemented methods for training a neural network, as well as for implementing audio encoders and decoders via trained neural networks, are provided. The neural network may receive an input audio signal, generate an encoded audio signal and decode the encoded audio signal. A loss function generating module may receive the decoded audio signal and a ground truth audio signal, and may generate a loss function value corresponding to the decoded audio signal. Generating the loss function value may involve applying a psychoacoustic model. The neural network may be trained based on the loss function value. The training may involve updating at least one weight of the neural network.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.