A Machine Learning Approach to Detecting Sound-Source Elevation in Adverse Environments

O'Dwyer H, Bates E, Boland FM. A Machine Learning Approach to Detecting Sound-Source Elevation in Adverse Environments. In: AES Convention 144. Audio Engineering Society; 2018. Paper 9968. Available from: https://aes.org/publications/elibrary-page/?id=19485

@inproceedings{ODwyer2018_19485,
  author = {O'Dwyer, Hugh and Bates, Enda and Boland, Francis M.},
  title = {{A Machine Learning Approach to Detecting Sound-Source Elevation in Adverse Environments}},
  booktitle = {AES Convention 144},
  note = {Paper 9968},
  year = {2018},
  month = may,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=19485}
}

TY  - CPAPER
TI  - A Machine Learning Approach to Detecting Sound-Source Elevation in Adverse Environments
AU  - O'Dwyer, Hugh
AU  - Bates, Enda
AU  - Boland, Francis M.
T2  - AES Convention 144
M1  - Paper 9968
PY  - 2018
DA  - 2018/05/06
UR  - https://aes.org/publications/elibrary-page/?id=19485
PB  - Audio Engineering Society
LA  - en
AB  - Recent studies have shown that Deep neural Networks (DNNs) are capable of detecting sound source azimuth direction in adverse environments to a high level of accuracy. This paper expands on these findings by presenting research that explores the use of DNNs in determining sound source elevation. A simple machine-hearing system is presented that is capable of predicting source elevation to a relatively high degree of accuracy in both anechoic and reverberant environments. Speech signals spatialized across the front hemifield of the head are used to train a feedforward neural network. The effectiveness of Gammatone Filter Energies (GFEs) and the Cross-Correlation Function (CCF) in estimating elevation is investigated as well as binaural cues such as Interaural Time Difference (ITD) and Interaural Level Difference (ILD). Using a combination of these cues, it was found that elevation to within 10 degrees could be predicted with an accuracy upward of 80%.
ER  -