Skip to Main content Skip to Navigation
Conference papers

Effects of audio coding on ICA performance: an experimental study

Abstract : In this paper, we study the influence of lossy audio coding on the performance of Independent Component Analysis (ICA). In particular, we derive two compression scenarios from practical implementations. In the first case, we consider the situation when the sources are independently compressed, decompressed and then mixed. In the second, we consider the situation when mixtures of sources are jointly compressed. We experimentally show that the modification of the spectrotemporal diversity due to compression has almost no effect on the performance of ICA methods. We also show that the two tested stereo encoding strategies have a major effect on the performance of ICA, especially when the mixed signals are compressed at low bit rates. As these strategies have been extended to audio systems involving much more than two channels, our work suggests that ICA will not be able to successfully process the high number of observations provided by modern immersive audio systems.
Complete list of metadata

Cited literature [14 references]  Display  Hide  Download
Contributor : Emmanuel Vincent Connect in order to contact the contributor
Submitted on : Wednesday, April 24, 2013 - 9:50:09 AM
Last modification on : Thursday, November 25, 2021 - 8:22:28 AM
Long-term archiving on: : Monday, April 3, 2017 - 9:06:31 AM


Files produced by the author(s)


  • HAL Id : hal-00817202, version 1


Matthieu Puigt, Emmanuel Vincent, Yannick Deville, Anthony Griffin, Athanasios Mouchtaris. Effects of audio coding on ICA performance: an experimental study. 11th IEEE Int. Workshop on Electronics, Control, Measurement, Signals and their application to Mechatronics, Jun 2013, Toulouse, France. ⟨hal-00817202⟩



Les métriques sont temporairement indisponibles