Options
Audio Inpainting based on Self-similarity for Sound Source Separation Applications
Author(s)
Date Issued
2020-09-24
Date Available
2024-05-08T12:16:29Z
Abstract
Sound source separation algorithms have advanced significantly in recent years but many algorithms can suffer from objectionable artefacts. The artefacts include phasiness, transient smearing, high frequency loss, unnatural sounding noise floor and reverberation to name a few. One of the main reasons for this is due to the fact that in many algorithm, individual time-frequency bins are often only attributed to one source at a time, meaning that many time-frequency bins will be set to zero for a separated source. This leads to an impressive signal to interference ratio but at the cost of natural sounding resynthesis. Here, we present a simple algorithm capable of audio inpainting based on selfsimilarity within the signal. The algorithm attempts to use the non-zero bin values observed in similar frames as substitutes for the zero bin values in the current analysis frame. We present results from subjective listening tests which show a preference for the inpainted audio over the original audio produced from a simple source separation algorithm. Further, we use the Fr´echet Audio Distance metric to evaluate the perceptual effect of the proposed inpainting algorithm. The results of this evaluation support the subjective test preferences.
Sponsorship
Science Foundation Ireland
Other Sponsorship
Insight Research Centre
Type of Material
Conference Publication
Publisher
IEEE
Copyright (Published Version)
2020 IEEE
Web versions
Language
English
Status of Item
Peer reviewed
Journal
2020 IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP)
Conference Details
The IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP 2020), Tempere, Finland, 21-24 September 2020
This item is made available under a Creative Commons License
File(s)
Loading...
Name
Audio Inpainting based on Self-similarity for Sound Source Separation Applications.pdf
Size
1.79 MB
Format
Adobe PDF
Checksum (MD5)
93439965157261050803ae683e204962
Owning collection
Mapped collections