Open AccessSelf-recovery scheme for audio restoration using auditory masking
Author(s) -
Alejandra Menendez-Ortiz,
Claudia Feregrino-Uribe,
José Juan García-Hernández
Publication year - 2018
Publication title -
plos one
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.99
H-Index - 332
ISSN - 1932-6203
DOI - 10.1371/journal.pone.0204442
Subject(s) - computer science , masking (illustration) , auditory masking , speech recognition , distortion (music) , audio signal , embedding , scheme (mathematics) , signal (programming language) , frequency domain , speech coding , artificial intelligence , telecommunications , bandwidth (computing) , mathematics , computer vision , art , amplifier , mathematical analysis , visual arts , programming language
Self-recovery schemes identify and restore tampering, using as a reference a compressed representation of a signal embedded into itself. In addition, audio self-recovery must comply with a transparency threshold, adequate for applications such as on-line music distribution or speech transmission. In this manuscript, an audio self-recovery scheme is proposed. Auditory masking properties of the signals are used to determine the frequencies that better mask the embedding distortion. Frequencies in the Fourier domain are mapped to the intDCT domain for embedding and extraction of reference bits for signal restoration. The contribution of this work is the use of auditory masking properties for the frequency selection and the mapping to the intDCT domain. Experimental results demonstrate that the proposed scheme satisfies a threshold of -2 ODG, suitable for audio applications. The efficacy of the scheme, in terms of its restoration capabilities, is also shown.