Sound Field Indicators for Hearing Activity and Reverberation Time Estimation in Hearing Instruments

Streich AP, Feilner M, Stirnemann A, Buhmann JM. Sound Field Indicators for Hearing Activity and Reverberation Time Estimation in Hearing Instruments. In: AES Convention 128. Audio Engineering Society; 2010.pp. 7985:1-11. Available from: https://aes.org/publications/elibrary-page/?id=15282

@inproceedings{Streich2010_15282,
  author = {Streich, Andreas P. and Feilner, Manuela and Stirnemann, Alfred and Buhmann, Joachim M.},
  title = {{Sound Field Indicators for Hearing Activity and Reverberation Time Estimation in Hearing Instruments}},
  booktitle = {AES Convention 128},
  pages = {7985:1-11},
  url = {https://aes.org/publications/elibrary-page/?id=15282},
  year = {2010},
  month = may,
  publisher = {Audio Engineering Society}
}

TY  - CPAPER
TI  - Sound Field Indicators for Hearing Activity and Reverberation Time Estimation in Hearing Instruments
AU  - Streich, Andreas P.
AU  - Feilner, Manuela
AU  - Stirnemann, Alfred
AU  - Buhmann, Joachim M.
T2  - AES Convention 128
SP  - 1
EP  - 11
M1  - Paper 7985
PY  - 2010
DA  - 2010/05/06
UR  - https://aes.org/publications/elibrary-page/?id=15282
PB  - Audio Engineering Society
LA  - en
AB  - Sound field indicators (SFI) are proposed as a new feature set to estimate the hearing activity and reverberation time in hearing instruments. SFIs are based on physical measurements of the sound field. A variant thereof, called SFI short-time statistics SFIst^2, is obtained by computing mean and standard deviations of SFIs on 10 subframes. To show the utility of these meature sets for the mentioned prediction tasks, experiments are carried out on artificially reverberated recordings of a large variety of sounds encountered in daily life. In a classification scenario where the hearing activity is to be predicted, both SFI and SFIst^2 yield clearly superior accuracy even compared to hand-tailored features used in state-of-the-art hearing instruments. For regression on the reverberation time, the SFI-based features yield a lower residual error than standard feature sets and reach the performance of specially designed features. The hearing activity classification is mainly based on the average of the SFIs, while the standard deviation over sub-window is used heavily to predict the reverberation time.
ER  -