Estimation of Late Reverberation Characteristics from a Single Two-Dimensional Environmental Image Using Convolutional Neural Networks

Kon H, Koike H. Estimation of Late Reverberation Characteristics from a Single Two-Dimensional Environmental Image Using Convolutional Neural Networks. J Audio Eng Soc. 2019;67(7/8):540-548. doi:10.17743/jaes.2018.0069

@article{Kon2019_20492,
  author = {Kon, Homare and Koike, Hideki},
  title = {{Estimation of Late Reverberation Characteristics from a Single Two-Dimensional Environmental Image Using Convolutional Neural Networks}},
  journal = {Journal of the Audio Engineering Society},
  volume = {67},
  number = {7/8},
  pages = {540--548},
  year = {2019},
  month = jul,
  publisher = {Audio Engineering Society},
  doi = {10.17743/jaes.2018.0069},
  url = {https://doi.org/10.17743/jaes.2018.0069}
}

TY  - JOUR
TI  - Estimation of Late Reverberation Characteristics from a Single Two-Dimensional Environmental Image Using Convolutional Neural Networks
AU  - Kon, Homare
AU  - Koike, Hideki
T2  - Journal of the Audio Engineering Society
J2  - J. Audio Eng. Soc.
VL  - 67
IS  - 7/8
SP  - 540
EP  - 548
PY  - 2019
DA  - 2019/07/06
DO  - 10.17743/jaes.2018.0069
UR  - https://doi.org/10.17743/jaes.2018.0069
PB  - Audio Engineering Society
LA  - en
AB  - In augmented-reality (AR) applications, reproducing acoustic reverberation is essential for the immersive audio experience. The audio components of an AR system should simulate the acoustics of the environment that is experienced by the users. Earlier, in virtual–reality (VR) applications, sound engineers could program all of the reverberation parameters for a particular scene in advance or when the user is at a fixed position. However, adjusting the reverberation parameters using conventional procedures is difficult because the unlimited range of such parameters cannot be programmed for AR applications. Therefore, it is necessary to dynamically estimate the reverberation characteristics based on the environments in which the users move. Considering that skilled acoustic engineers can estimate the reverberation parameters using the images of a room without performing any measurements, we trained convolutional neural networks to estimate the reverberation parameters using two–dimensional images. The proposed method does not require the simulations of sound propagation using 3D reconstruction techniques.
ER  -