AR model for low latency packet loss concealment for wireless sound zones at low frequencies

Pedersen CS, Zhou M, Møller MB, de Koeijer NEM, Østergaard J. AR model for low latency packet loss concealment for wireless sound zones at low frequencies. In: AES Convention 154. Audio Engineering Society; 2023.pp. 10651:1-10. Available from: https://aes.org/publications/elibrary-page/?id=22058

@inproceedings{Pedersen2023_22058,
  author = {Pedersen, Christian Sejer and Zhou, Mo and Møller, Martin Bo and de Koeijer, Niels Evert Marius and Østergaard, Jan},
  title = {{AR model for low latency packet loss concealment for wireless sound zones at low frequencies}},
  booktitle = {AES Convention 154},
  pages = {10651:1-10},
  url = {https://aes.org/publications/elibrary-page/?id=22058},
  year = {2023},
  month = may,
  publisher = {Audio Engineering Society}
}

TY  - CPAPER
TI  - AR model for low latency packet loss concealment for wireless sound zones at low frequencies
AU  - Pedersen, Christian Sejer
AU  - Zhou, Mo
AU  - Møller, Martin Bo
AU  - de Koeijer, Niels Evert Marius
AU  - Østergaard, Jan
T2  - AES Convention 154
SP  - 1
EP  - 10
M1  - Paper 10651
PY  - 2023
DA  - 2023/05/06
UR  - https://aes.org/publications/elibrary-page/?id=22058
PB  - Audio Engineering Society
LA  - en
AB  - Personal sound zone systems use multiple loudspeakers to deliver individual audio experiences to several people in a room. At low frequencies, the use of distributed woofers and sound field control is advantageous for creating sound separation (contrast) between the zones. Wireless streaming to the woofers is convenient from a practical point of view but packet losses significantly decreases the sound quality and the sound separation between the zones. In this study autoregressive (AR) models are evaluated for concealing the effects of wireless packet losses both with regard to contrast performance and sound quality using several objective sound quality models. According to these models, for mild independent and identically distributed (i.i.d.) packet loss to a single woofer, the AR model can lead to imperceptible performance degradations, but for more bursty packet losses the AR model is less effective - though significantly better than silence substitution.
ER  -